Research

Prostate cancer in Western Australia: trends in incidence and mortality from 1985 to 1996 Timothy J Threlfall, Dallas R English and Ian L Rouse MJA 1998; 169: 21-24 For editorial comment, see McCredie & Cox Abstract - Introduction - Methods - Results - Discussion - References - Addendum - Authors' details - - ©MJA1998 Abstract Objective: To measure trends in recorded incidence and mortality rates of prostate cancer in Western Australia from 1985 to 1996 and to relate these to prostate-specific antigen (PSA) testing for prostate cancer. Design: Descriptive study based on data from the Western Australian Cancer Registry, the Australian Bureau of Statistics and the Health Insurance Commission. Data: All newly diagnosed cases of prostate cancer and all deaths from prostate cancer in Western Australia from 1985 to 1996. Main outcome measures: Recorded incidences and mortality rates for prostate cancer. Results: After increasing steadily from 42 per 100 000 person-years in 1985 to 61 in 1992, the recorded incidence more than doubled to 134 per 100 000 person-years in 1994, then fell sharply to 87 in 1996. Among men aged 50 years or more, those aged 50-54 years had the largest annual increases: 14% (95% confidence interval [CI], 10%-19%) from 1985 to 1992 and 108% (95% CI, 84%-134%) from 1992 to 1994. They also had the smallest annual decline between 1994 and 1996 (8%; 95% CI, 1% increase to 16% decrease). The mortality rate showed no sudden increases or decreases. In men aged 60 years or older, the mortality rate increased annually by 2.9% (95% CI, 2%-4%) from 1985 to 1996. The number of Medicare reimbursements for PSA tests increased until May 1995, then fell. There was a significant correlation between the monthly number of PSA tests and new cases of prostate cancer (P < 0.01). Conclusions: Following a period of steady increase, the recorded incidence of prostate cancer increased dramatically in 1992 because of screening by PSA testing. From 1994, these incidence figures declined almost as sharply, partly because of reductions in testing. The mortality rate has not shown any systematic deviation from its long-term trend. Introduction In the early 1990s the recorded incidence of prostate cancer increased dramatically in Australia,1,2 several years after a similar increase in the United States.3-7 Testing for the disease among asymptomatic men by measuring plasma prostate-specific antigen (PSA) is believed to be responsible for the increases.1,3,4,7We report here on trends in incidence figures and in the mortality rate for prostate cancer in Western Australia from 1985 to 1996, and their relationship to PSA testing. Methods Data sources We obtained data on prostate cancer cases and deaths from the Western Australian Cancer Registry, and population estimates from the Australian Bureau of Statistics (ABS).8 We standardised rates to the World Standard Population and calculated the risk of men developing prostate cancer before the age of 75 years.9 Because mortality rates based on coding of cause of death by the ABS were not available for 1996, we used 1996 rates from the Registry, which began coding causes of death in 1990; from 1990 until 1995, the two mortality rates were almost identical. Data relating to the number of PSA tests reimbursed by Medicare in Western Australia were obtained from the Health Insurance Commission in July 1997. Socioeconomic status To investigate any effect of socioeconomic status (SES) on recorded incidence we used an index, derived from the 1991 census, in which each census collection district (about 50 households) is assigned a score.10 For Perth patients, addresses at the time of diagnosis were mapped to collection districts for the 1991 census using MapInfo.11 Geographical coordinate data were provided by the Western Australian Valuer General's office and the Department of Land Administration. The SES index was divided into quarters of its distribution. Because population data were available at the collection district level for census years only (ie, 1986 and 1991), we could not calculate SES-specific incidence. To determine whether any changes in numbers of cases by SES might be the result of different changes in population size in areas of different SES, we compared numbers of cases of prostate cancer and lung cancer. Trends in age-specific rates We used Poisson regression in EGRET12 to model age-specific recorded incidences and mortality rates. Analyses of incidence were restricted to men aged 50 years or older and analyses of mortality rate to men aged 60 years or older because there were few events in younger age groups. We analysed recorded incidence for each of the periods January 1985 to December 1992, January 1992 to December 1994 and January 1994 to December 1996. (Because the periods overlap, the results were not independent.) A single analysis of mortality rate was conducted. Likelihood ratio tests were used to obtain P values. We fitted age group as a categorical (ie, factored) variable and year of diagnosis as a continuous variable. The coefficient for year was exponentiated to give an annual percentage increase in the rate (eg, a coefficient of 0.35 when exponentiated is 1.42, equivalent to an annual increase of 42%). Age was added first, followed by the year of diagnosis, and then the interaction between the two. The interaction was fitted with age as a categorical variable and as a continuous variable, and the difference between these models was tested. Fitting the interaction with age as a categorical variable tests whether the secular trend was the same for all age groups; fitting it with age as a continuous variable tests whether there was a greater increase in younger men than in older men (or vice versa). In the analyses of incidence, P values for comparison of the two types of interaction were 0.17 for 1985-1992, 0.32 for 1992-1994 and 0.70 for 1994-1996. As there were no significant differences, the results reported for recorded incidence are from models in which the interaction involved age as a continuous variable. In all analyses, the final models provided good fits to the data -- the smallest P value for goodness-of-fit (for incidence in the period 1992-1994) was 0.09. Prostate-specific antigen testing Medicare began identifying PSA tests, together with prostatic acid phosphatase (PAP) tests, as a separate item during November 1993. Fewer than 1% of these tests would be PAP tests (Dr Glen Edwards, Chemical Pathologist, Western Diagnostic Pathology, personal communication). We plotted the number of tests reimbursed by Medicare each month from December 1993 to December 1996. To investigate the trends in the numbers of tests, we fitted a curve through the data. Results Incidence From 1985 until 1992, the age-adjusted recorded incidence increased steadily from 42 to 61 per 100 000 person-years (Figure 1). In the next two years it more than doubled to 134 per 100 000 person-years, but then fell almost as sharply to 87 in 1996. The risk of prostate cancer before age 75 years was one in 23 in 1985, one in six in 1994 and one in nine in 1996. Age: In men aged over 50 years, the recorded incidence of prostate cancer increased annually by 5% (95% confidence interval [CI], 3%-6%) between 1985 and 1992 (trend, P < 0.001). The largest relative increases in recorded incidence between 1985 and 1992 occurred in the youngest men (Table 1; interaction between year and age, P < 0.001). The largest relative increases between 1992 and 1994 were also seen in the younger age groups (interaction between year and age, P < 0.001), and from 1994 to 1996 the decline was greatest in the oldest men (interaction between year and age, P < 0.001). As a result of the different relative changes in different age groups, the differences in age-specific rates in 1996 were smaller than in earlier years. In 1985, the recorded incidence for men in the oldest age groups was close to 1000 times higher than for men aged 50-54 years, but by 1996 the relative difference was about 100-fold. We also examined the absolute changes in recorded incidence between 1992 and 1996 (Table 2). Between 1992 and 1994, the largest absolute increases were in men aged 65-79 years. Between 1994 and 1996, the largest absolute decreases were in men aged 70 years or older, so that between 1992 and 1996 the overall increases were greatest in men aged 60-69 years. The overall changes in men aged 55-59 years and 70-74 years were similar. The mean age at diagnosis was 73 years in 1985, 74 years from 1986 until 1990, 73 years in 1991 and 1992, 72 years in 1993, 70 years in 1994 and 69 years thereafter. Place of residence: We examined age-standardised recorded incidence for prostate cancer separately for the Perth metropolitan region and the rest of Western Australia. Before 1992, the two rates were similar in all years. During the sudden rise and fall, these incidences were, respectively, 1992: 65 per 100 000 person-years (Perth), 51 per 100 000 person-years (non-metropolitan areas); 1993: 141 per 100 000 person-years, 107 per 100 000 person-years; 1994: 141 per 100 000 person-years, 110 per 100 000 person-years; 1996: 87 per 100 000 person-years, 84 per 100 000 person-years. Socioeconomic status: We mapped 94% of lung cancer and prostate cancer cases in the Perth metropolitan area to a 1991 census collection district, with no apparent trend over time in the proportion mapped. Before 1993, the numbers of prostate cancer cases in the four SES groups were similar (Figure 2). However, the increase in cases in 1993 and 1994 was greatest in areas of highest SES, and the largest declines in numbers of cases from 1994 to 1996 were also in these areas. Over the same period, there was no consistent change in the distribution of lung cancer cases by SES (Figure 2). Mortality rate The age-adjusted mortality rate from 1985 to 1996 showed no sudden increases or decreases (Figure 1). In men aged 60 years or older, the estimated annual increase from 1985 to 1996 was 2.9% (95% CI, 2%-4%; trend, P < 0.001). Adding quadratic (P = 0.78) or cubic (P = 0.92) terms for year did not improve the fit of the Poisson model. Furthermore, models using year as a continuous or a categorical variable fitted equally well (P = 0.14), indicating that year-to-year variations in the trend of the age-adjusted rates could be a result of chance alone. The increase differed across age groups (interaction between age as a categorical variable and year of death, P = 0.03), but the trends by age were inconsistent (Table 1; interaction between age as a continuous variable and year of death, P = 0.96). Prostate-specific antigen testing On average, there were 5337 tests reimbursed each month in Western Australia. The numbers of tests initially increased before falling, although there was substantial monthly variation (Figure 3). A cubic curve fitted the data well (R2 = 0.46) and provided a better fit than a quadratic curve (P < 0.001) or a linear model (P < 0.001). The fitted maximum monthly number of tests occurred in May 1995. Spearman's rank correlation between the monthly number of PSA tests and new cases of prostate cancer was 0.48 (P < 0.01). Discussion After increasing steadily during the 1980s, the age-adjusted recorded incidence of prostate cancer more than doubled between 1992 and 1994. By 1996, it had fallen to a level about 40% higher than that in 1992. In contrast, the mortality rate increased by about 3% per year, with no sudden increases or decreases. The youngest age groups showed the largest relative increases in recorded incidence up to 1994, and the oldest age groups showed the greatest decrease after 1994, causing a substantial compression of the range of age-specific recorded incidence in Western Australia. When the rates rose steeply in 1992, the absolute increases were greatest in men aged 65-79 years. However, men aged 70 years or older had the greatest absolute falls from 1994, so that, between 1992 and 1996, the largest absolute increases were in men aged 60-69 years. The increase appeared first in Perth and the peak was higher in Perth. However, by 1996, Perth and the rest of Western Australia had similar recorded incidences of prostate cancer. Within Perth, the changes were greatest in areas of high SES. Most observers have attributed the increases in recorded incidence of prostate cancer during the 1980s to improved case detection, particularly following transurethral resection of the prostate for benign prostatic hypertrophy.1,13 The sudden rise in incidence figures in about 1993 was observed in all Australian States.1 In the United States, similar dramatic increases were observed first in 1989.3 These increases are almost certainly the result of the introduction of screening by PSA testing.1,3 In Western Australia, free PSA testing during Prostate Awareness Week would have contributed to the increase. Each October from 1993, about 1100 men attended Prostate Awareness Week in Perth for PSA tests (Mr M D'Antuono, Biostatistician, Urological Research Centre, University of Western Australia, personal communication). These tests do not appear in the Medicare figures, although the peaks in Medicare-funded PSA tests in November 1994 and November 1995 might be the result of publicity surrounding Prostate Awareness Week. However, substantially fewer tests were performed at the screening venues than were reimbursed by Medicare each month in Western Australia. Thus, Prostate Awareness Week is unlikely to have greatly increased the recorded incidence in Western Australia. Rapid declines in recorded incidence, such as we found in Western Australia since 1994, have also been reported in some US States.4,6,7,14 Part of the Western Australian decrease is probably the result of reduced screening activity. The most likely explanation for the decrease in Medicare reimbursements that started in 1995 is a reduction in PSA testing for screening. In fact, the reduction in screening tests is probably greater than Figure 3 suggests, because PSA testing is also used for surveillance of men with prostate cancer. Therefore, its overall use will decline less rapidly than its use for screening. Widespread publicity in the media about the controversy surrounding screening for prostate cancer may have contributed to this decline, and the greater declines seen in the oldest men may be because of concerns that screening is unlikely to benefit those with a short expectation of life.15 Recorded incidence is expected to decrease even if screening activity remains constant. After the introduction of screening, recorded incidence increases because the time of diagnosis is advanced and some cases may be diagnosed that would not have become symptomatic. When all prevalent cases are detected, incidence figures will fall until new cancers develop, whereupon they will rise again. If screening detects only those cancers that would eventually have been diagnosed anyway, the recorded incidence will stabilise at its pre-screening level. Otherwise, it will stabilise at a higher level.16 Before 1980, the mortality rate from prostate cancer in Australia was stable for some time.17 Since then it has increased, but more slowly than recorded incidence, indicating that short-term survival, at least, has improved over time. Increasing diagnosis of disease with low potential for metastasis is one explanation for the discrepancy. What changes in mortality rate can we expect? Gann has argued that if screening is effective and covers enough of the population, the mortality rate should eventually decrease.16 It is too early for any effect of screening on mortality rate to be seen, and by the end of 1996 there had been no new trend in the mortality rate in Western Australia. Because of the effect of lead time (the time by which screening advances diagnosis), changes may not occur for some years. We have witnessed extraordinary changes in the recorded incidence of prostate cancer in Western Australia, and we have strong evidence that these changes are the result of medical practice rather than intrinsic changes in the incidence. Surveillance of incidence and mortality rates may help answer the question of whether screening has benefit, although more rigorous scientific evaluations are also needed. References McCredie M. A rising incidence of prostate cancer in Australia: effect of "screening" or more disease? Cancer Forum 1995; 19: 7-12. McCaul KA, Luke CG, Roder DM. Trends in prostate cancer incidence and mortality rates in South Australia, 1977-1993. Med J Aust 1995; 162: 520-522. Lu-Yao GL, Greenberg ER. Changes in prostate cancer incidence and treatment in USA. Lancet 1994; 343: 251-254. Gilliland FD, Welsh DJ, Hoffman RM, Key CR. Rapid rise and subsequent decline in prostate cancer incidence rates for New Mexico, 1989-1993. Cancer Epidemiol Biomarkers Prev 1995; 4: 797-800. Merrill RM, Brawley OW. Prostate cancer incidence and mortality rates among white and black men. Epidemiology 1997; 8: 126-131. Polednak AP. Trends in prostate carcinoma incidence in Connecticut (1988-1994) by age and race. Cancer 1997; 79: 99-103. Newcomer LM, Stanford JL, Blumenstein BA, Brawer MK. Temporal trends in rates of prostate cancer: declining incidence of advanced stage disease, 1974 to 1994. J Urol 1997; 158: 1427-1430. Australian Bureau of Statistics. Estimated Resident Population by age and sex in statistical local areas, Western Australia. Canberra: AGPS, 1996. (Catalogue No. 3203.5.) Parkin DM, Muir CS, Whelan SL, et al, editors. Cancer incidence in five continents. Vol VI. IARC Scientific Publications No. 120. Lyon: International Agency for Research on Cancer, 1992. Australian Bureau of Statistics. Information Paper: 1991 Census socioeconomic indicators for areas. Canberra: AGPS, 1993. (Catalogue No. 2912.0.) MapInfo [computer program]. Version 3. New York: MapInfo Corporation, 1995. EGRET [computer program]. Version 1.02. Seattle: Statistics and Epidemiology Research Corporation, 1995. Potosky AL, Kessler L, Gridley G, et al. Rise in prostatic cancer incidence associated with increased use of transurethral resection. J Natl Cancer Inst 1990; 82: 1624-1628. Stephenson RA, Smart CR, Mineau GP, et al. The fall in incidence of prostate carcinoma. On the down side of a prostate specific antigen induced peak in incidence -- data from the Utah Cancer Registry. Cancer 1996; 77: 1342-1348. Chodak GW, Thisted RA, Gerber GS, et al. Results of conservative management of clinically localized prostate cancer. N Engl J Med 1994; 330: 242-248. Gann PH. Interpreting recent trends in prostate cancer incidence and mortality. Epidemiology 1997; 8: 117-120. Holman CD, James IR, Segal MR, Armstrong BK. Recent trends in mortality from prostate cancer in male populations of Australia and England and Wales. Br J Cancer 1981; 44: 340-348.(Received 30 Sep 1997, accepted 31 Mar 1998) Addendum Since this article was submitted in September 1997, data on the recorded incidence and mortality rate of prostate cancer in the first eight months of 1997 have become available. The age-adjusted recorded incidence of prostate cancer was 64 per 100000 person-years, while the mortality rate was 16 per 100000 person-years. Thus, the recorded incidence was almost the same as in 1992, before the rapid rise and fall. In 1997, men aged 50-69 years had higher recorded incidences of prostate cancer than in 1992, but older men had lower recorded incidences. Authors' details Health Department of Western Australia, East Perth, WA. Timothy J Threlfall, MB BS, MPH, Senior Medical Officer, Western Australian Cancer Registry; Ian L Rouse, PhD, General Manager, Health Information Centre. Department of Public Health, University of Western Australia, Nedlands, WA. Dallas R English, PhD, Senior Lecturer. Reprints: Dr D R English, Department of Public Health, University of Western Australia, Nedlands, WA 6907. E-mail: dallasATdph.uwa.edu.au Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". <URL: http://www.mja.com.au/>

Timothy J Threlfall · Dallas R English · Ian L Rouse

Self-reported prevalence of prostate-specific antigen testing in South Australia: a community study Carole B Pinnock, David P Weller and Villis R Marshall MJA 1998; 169: 25-28 For editorial comment, see McCredie & Cox Abstract - Introduction - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - ©MJA1998 Abstract Objective: To determine the prevalence and other characteristics of self-reported blood testing (prostate-specific antigen [PSA]) for prostate cancer in the community. Design: Interview-based prevalence study. Participants and setting: 695 men aged 40 years and over in the Spring 1996 South Australian Health Omnibus survey of a probability sample of 3011 households, weighted to reflect the age and sex distribution of the South Australian population. Outcome measures: Number of men who had a PSA test in the preceding 12 months; number of first tests; the person initiating and performing the test; knowledge of the next step if test result abnormal; number of men visiting doctor for lower urinary tract symptoms in the preceding 12 months. Results: 20.3% of participants reported having a PSA test in the preceding 12 months; 62.1% were first tests. Prevalence of testing was highest in the age group 70-79 years (35.8%). Most tests were initiated by the general practitioner (41.2%) and by patients (35.7%). Of those tested, 45.3% had inadequate knowledge of the next step. Visiting a doctor for urinary symptoms was significantly associated with PSA testing (P < 0.001): 47.7% of men who visited a doctor for urinary symptoms had been tested compared with only 17.4% of those who did not visit the doctor for this problem. Only age and visiting a doctor for urinary symptoms were significant independent predictors of having a PSA test. Conclusions: Investigation of lower urinary tract symptoms contributed substantially to PSA testing, and those tested did not adequately understand the consequences. Our findings suggest a need for a better framework for PSA testing in general practice, including all important elements of decision-making, such as evidence and patient preference, as well as the means to ensure adequate patient counselling before testing. Introduction Screening for prostate cancer remains a topic of widespread debate. Recent systematic reviews have concluded that there is currently insufficient evidence to establish an overall benefit of prostate cancer screening, and major health agencies have issued conflicting recommendations.1-3 Randomised controlled trials examining this question have begun in the United States and Europe,4,5 but results will not be available for some years. Nevertheless, prostate-specific antigen (PSA) testing of asymptomatic men is thought to be common in Australia,6,7 and the recent doubling in recorded incidence of prostate cancer in this country8,9 has been attributed to PSA testing. Little is known about the circumstances of such testing -- whether it occurs in general practice or other specialties, whether men are truly asymptomatic, whether the test is initiated by the patient or the doctor, and how much information about the test is provided to the patient. Most authors agree that men seeking a test should be informed about the risks as well as the benefits of taking the test,10,11 and that individual preferences should be taken into account.12-15 A community survey of self-reported participation in PSA testing confirmed high rates in the community (25% of men 40 years and over with no history of prostate cancer) and an even higher rate of intention to test (53.9%).16 The strongest predictor of past testing was a visit to a doctor for urinary tract symptoms, and the strongest predictor of intention to test was perceived vulnerability to prostate cancer.16 This study was undertaken in 1996 to expand these findings by establishing the prevalence of self-reported prostate cancer testing (PSA testing) over the preceding 12 months, the incidence of PSA testing (first tests), who initiated the test, who performed it, the association with investigation for urinary symptoms, and whether participants had understood the immediate consequences of taking the test. Methods Survey Questions were included in the Spring 1996 South Australian Health Omnibus survey, a multiple-user household interview survey undertaken on behalf of a number of healthcare organisations in South Australia. The sampling method provides a probability sample of the South Australian population.17,18Participants were asked whether they had visited the doctor for troublesome urinary symptoms in the past 12 months (as in a previous survey17), ever been diagnosed with prostate cancer or had a blood test for prostate cancer (PSA test) in the previous 12 months. Those who had been tested were asked who had initiated the test, who performed it and what they understood to be the next step if the test were abnormal. Question alternatives were derived from a previous qualitative study (who initiated test)19 and expert clinical opinion (who performed it, next step). Questions were tested for comprehensibility and acceptability in 50 pilot interviews before the survey. Analysis Data were weighted by household size, age, sex and geographical region to benchmarks derived from the estimated resident population at 30 June 1995 (Australian Bureau of Statistics). Because of the clustered nature of the sample, confidence limits were calculated after allowing for a design effect of 1.1, calculated using the method of Kish,20 which inflates the standard error. We analysed data for men aged 40 years or older who had not had a diagnosis of prostate cancer, using SPSS for Windows.21 Statistical significances were determined by Pearson c2 tests or Fisher's exact test, and logistic regression by the forced entry method (the contribution is evaluated after removal of effects of all other variables). The contribution of demographic and urinary symptom variables was examined in a logistic regression model to identify independent predictors of having a PSA test in the preceding 12 months. Results Sample In unweighted numbers, from the sampling frame of 4081 households, 3011 interviews were conducted, giving a response rate of 73.8%. The reasons for not participating were refusal (548), no contact could be established (301), selected respondent away for duration of study (108), illness/mental incapacity (65) and respondent unable to speak English (48). Of the 3011 people interviewed, 695 were men 40 years or older and 642 of these had not had a diagnosis of prostate cancer. The average age of these 642 men was 58.3 years (SE, 0.50 years; range, 40-91 years). After weighting, 6.0% (95% confidence interval [CI], 4.45%-7.86%) of men aged 40 years or more reported being diagnosed with prostate cancer. The weighted sample size of men aged 40 years and over with no reported diagnosis of prostate cancer was 716. PSA testing and demographic factors Overall, 20.3% of men older than 40 years reported having a blood test for prostate cancer (PSA test) in the preceding 12 months (Table 1). This varied with age (chi-squared = 48.1; df = 4; P < 0.001), with the highest proportion (35.8%) being in the age group 70-79 years. Nearly two-thirds of tests (62.1%) were first tests, and this proportion was high across all age groups. Men were slightly more likely to be tested if they lived in metropolitan (20.8%; 95% CI, 17.5%-24.1%) than in rural (19.0%; 95% CI, 15.8%-22.1%) areas, but the difference was not significant. Educational attainment was not associated with testing. Lifetime occupation also showed no clear trends (eg, drivers and plant operators had similar rates to managers and administrators), nor did marital status and country of birth. Lower urinary tract symptoms Of 715 men aged 40 years and over who had not had a diagnosis of prostate cancer, 65 (9.1%) had visited a doctor for troublesome lower urinary tract symptoms (LUTS) in the preceding 12 months (46 visited their general practitioner). Of these 65 men, 31 (47.7%) had had a PSA test, compared with only 113/650 (17.4%) of those who did not visit the doctor for such a problem (P < 0.001, Fisher's exact test). A first visit to a doctor for LUTS was significantly associated with a first PSA test (P = 0.018, Fisher's exact test): 15/17 (88.2%) men who had a first visit for LUTS in the previous 12 months also had a first test in that time, compared with only 6/14 (42.9%) men for whom it was not the first visit. Who initiates and performs the test Most testing (70.6%) was performed in general practice, and most tests were initiated by the general practitioner (Table 2). However, patients initiated 35.7% of tests. Tests performed by "other doctor" may reflect those done by pathology services, but could also have been included in other types of medical assessments. Doctors initiated 25 (80.6%) of the tests on the 31 men who visited for LUTS and were tested, but initiated only 51 (45.1%) of the tests on the 113 who were tested but had not visited a doctor for LUTS (chi-squared = 12.3; df = 1; P < 0.001). Knowledge of the "next step" There was wide variability in understanding of the next step if the test were abnormal. The most frequent response was "do not know" (34.0%). Other responses were "operation on the prostate" (10.7%), "no further action" (0.6%), "referral to a specialist" (27.2%), "repeat the test" (6.3%), "biopsy of the prostate" (4.8%) and "other" (16.7%). The first three (do not know, operation on the prostate, no further action), totalling 45.3%, may be considered to reflect poor understanding. Independent predictors of testing The logistic regression model (chi-squared = 89.8; df = 22; P < 0.001) included age, rural/metropolitan residence, country of birth, educational attainment, marital status, lifetime occupation, and doctor visit for LUTS in the past 12 months. Of these, only age and visit to a doctor for LUTS were significant independent predictors of having had a PSA test in the past 12 months (Table 3). Discussion We found a high level of self-reported PSA testing among men in South Australia. Testing was associated with lower urinary tract symptoms and was initiated by the patient or his GP. Nearly half the respondents had an inadequate knowledge of the appropriate next step if the test result were abnormal. Our study was a community-based survey of self-reported testing. Such studies suffer from limitations in that individual recall of past events and comprehension of the research questions may vary. Nevertheless, representative community sampling makes it possible to derive estimates of incidence and prevalence, a benefit over general practice-based studies. The level of self-reported rates of PSA testing observed in this study is higher than that reported in New South Wales across all age groups.7 This may result from a real difference between South Australia and New South Wales or from differences in survey method. The sampling method differed between the two studies (probability sample weighted to reflect age and sex structure of South Australian population versus random telephone number selection from Sydney metropolitan white pages). Older men were under-represented in the New South Wales study compared with 1991 Census data. The interview method (face to face in South Australia versus telephone in New South Wales) may also have contributed to the differences. In our study, the number of men tested increased with age and peaked in the age group 70-79 years, with nearly half the tests in this age group being first tests. The choice to begin PSA testing at this age is of concern, as those least likely to benefit are men who can anticipate less than a decade of life.1 We found no association between PSA testing and socioeconomic factors such as education and occupation, although in the United States participation in prostate cancer screening is reported to be sensitive to socioeconomic factors.22,23 However, our result is consistent with other Australian data.7,16 It is also consistent with the view that testing is initiated more by doctors than patients in this community. There was only a small difference between rural and metropolitan testing rates, which suggests that access to services does not influence the number of men seeking or being offered testing. The role of LUTS in prompting testing for prostate cancer has been reported previously,7 and is important because of the high prevalence of such symptoms in the community. In a recent South Australian survey, 26.4% of men over 18 years reported experiencing troublesome urinary symptoms in the past 12 months, and 10.2% had visited a doctor for this reason. A similar number (8.6%) were substantially dissatisfied with their urinary function.17 The relative proportions of men and their doctor initiating prostate cancer testing have not been reported previously. We found that, in the absence of urinary symptoms, this is roughly equal. However, among those men who had visited a doctor for LUTS, 80% of tests were doctor-initiated. This strong association between doctor as test initiator and LUTS suggests that, while testing may be for case-finding or screening purposes in asymptomatic men, its use may be investigational for those with LUTS. We cannot know, from these and other data,7 whether other indicators for investigational PSA testing (such as abnormal or suspicious digital rectal examination, family history of prostate cancer, or complicated LUTS) were present, and therefore whether testing was appropriate in these cases. The current clinical guidelines for uncomplicated LUTS do not recommend testing for prostate cancer.24 The American College of Physicians maintains that, as no association between LUTS and prostate cancer has been demonstrated, testing in men both with and without LUTS consistent with benign prostatic enlargement constitutes screening.10 Current guidelines focus on when not to use the PSA test, but not when it is appropriate.24 There is no framework for doctors and patients which includes all the important elements of decision-making in this area, including evidence, patient preference and medicolegal issues. PSA testing in general practice is driven in part by concern of patients and in part for investigational purposes. In addition, there is anecdotal evidence that general practitioners are concerned that if a PSA test is not offered, and prostate cancer is later diagnosed, they may be seen as negligent. Clinical guidelines or other measures promoting the appropriate use of the PSA test in general practice need to give consideration to all of these factors. Undoubtedly, providing information about testing for prostate cancer is more complex than for breast cancer or cervical cancer screening. The finding that a large proportion of men tested did not understand the immediate consequences of testing is thus not surprising, but of concern. We cannot tell from these data whether information was not given, not understood, or not recalled. Nevertheless, it is reasonable to assume that if this very basic information is not understood effectively, then more complex information (such as the likelihood of a false positive result, side effects of a biopsy, effectiveness of treatment for early-stage prostate cancer, and the risks of impotence and incontinence resulting from radical surgery for localised cancer) would also not have been understood. Yet this has been suggested as basic information needed by a patient to understand the implications of a PSA test.11,14,15 We therefore see a need for closer examination of the exchange and uptake of information in consultations that result in a PSA test, and the provision of resources for GPs which aid effective counselling. Such resources may need to include longer consultations. Acknowledgements The study was funded through a collaboration of Repatriation General Hospital, Daw Park, the Anti-Cancer Foundation of Australia, and the Department of Evidence Based Care and General Practice, Flinders University of South Australia, under the auspices of the Collaborative Centre for Prostate Health. References Australian Health Technology Advisory Committee. Prostate cancer screening. Canberra: AGPS, 1996. National Health Service. Screening for prostate cancer: effectiveness matters. York, UK: NHS Centre for Reviews and Dissemination, 1997. Mettlin C, Jones G, Avarette H, et al. Defining and updating the American Cancer Society guidelines for cancer-related checkup: prostate and endometrial cancers. CA Cancer J Clin 1993; 43: 42-46. Denis L, Middelheim A. To screen or not to screen? Prostate 1992; Suppl 4: 63-70. Gohagan J, Prorok P, Kramer B, et al. The prostate, lung, colorectal and ovarian screening trial of the National Cancer Institute. Cancer 1995; 75: 1869-1873. Parkes A, Killer G. Prostate-specific antigen -- is it already being used as a screening test? [letter]. Med J Aust 1994; 161: 722-733. Ward JE, Hughes A-M, Hirst GHL, Winchester L. Men's estimates of prostate cancer risk and self-reported rates of screening. Med J Aust 1997; 167: 250-253. McCredie M. A rising incidence of prostate cancer in Australia: effect of "screening" or more disease? Cancer Forum 1995; 19: 7-12. McCaul K, Luke C, Roder D. Trends in prostate cancer incidence and mortality rates in South Australia. Med J Aust 1995; 162: 520-522. American College of Physicians. Screening for prostate cancer. Clinical Guideline: Part III. Ann Intern Med 1997; 126: 480-484. Hirst GH, Ward JE, Del Mar C. Screening for prostate cancer: the case against. Med J Aust 1996; 164: 285-288. Hahn DL, Roberts RG. PSA screening for asymptomatic prostate cancer: truth in advertising. J Fam Pract 1993; 37: 432-436. Woolf SH. Should we screen for prostate cancer? Men over 50 have a right to decide for themselves [editorial]. BMJ 1997; 314: 989-990. Flood AB, Wennberg JE, Nease RF Jr, et al. The importance of patient preference in the decision to screen for prostate cancer. Prostate Patient Outcomes Research Team. J Gen Intern Med 1996; 11: 2342-2349. Wolf AM, Nasser JF, Wolf AM, Schorling JB. The impact of informed consent on patient interest in prostate-specific antigen screening. Arch Intern Med 1996; 156: 1333-1336. Weller D, Pinnock C, Silagy C, et al. Prostate cancer testing in South Australian men: influence of sociodemographic factors, health beliefs and lower urinary tract symptoms. Aust N Z J Public Health 1998; 22: 400-402. Pinnock C, Marshall V. Troublesome lower urinary tract symptoms in the community: a prevalence study. Med J Aust 1997; 167: 72-75. Wilson D, Wakefield M, Taylor A. The South Australian Health Omnibus Survey. Health Prom J Aust 1992; 2: 47-49. Pinnock C, O'Brien B, Marshall V. Older men's concerns about their urological health: a qualitative study. Aust N Z J Public Health 1998; 22: 368-373. Kish L. Estimates of unit variance: design effect. In: Survey sampling. New York: John Wiley and Sons, 1965; 257-263. SPSS for Windows [computer program]. Version 7.5. Chicago, Ill: SPSS Inc, 1996. Robinson SB, Ashley M, Haynes MA. Attitudes of African Americans regarding screening for prostate cancer. J Natl Med Assoc 1996; 88: 241-246. Diefenbach PN, Ganz PA, Pawlow AJ, Guthrie D. Screening by the prostate-specific antigen test: what do the patients know? J Cancer Educ 1996; 11: 39-44. National Health and Medical Research Council. Clinical practice guidelines. The management of uncomplicated lower urinary tract symptoms in men. Canberra: Commonwealth of Australia, 1997. (Received 7 Nov 1997, accepted 7 May 1998) Authors' details Division of Surgery, Repatriation General Hospital, Daw Park, SA. Carole B Pinnock, PhD, Principal Research Scientist. Flinders University of South Australia, SA. David P Weller, FRACGP, FAFPHM, Senior Lecturer, Department of Evidence Based Care and General Practice; Villis R Marshall, MD, FRACS, Professor of Surgery, Department of Surgery, Flinders Medical Centre, and Head, Division of Surgery, Repatriation General Hospital, Daw Park, SA. Reprints will not be available from the authors. Correspondence: Dr C R Pinnock, Division of Surgery, Repatriation General Hospital, Daws Rd, Daw Park, SA 5041. E-mail: spinncbATrgh.edu.au - Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". <URL: http://www.mja.com.au/>

Carole B Pinnock · David P Weller · Villis R Marshall

Abstract Objective: To assess the feasibility of offering the choice of prescribing injectable heroin (diamorphine) or injectable methadone to opiate-dependent injecting drug users and to assess whether there are health and social gains associated with prescribing injectable opiates. Design: A protocol-driven prospective observational study. Type of injectable opiate received was based on self-selection. Setting: A large west London drug clinic. Patients: Fifty-eight patients admitted to the clinic between 1 June 1995 and 31 December 1996, who were long term opiate-dependent injecting drug users, who had previously tried and failed oral methadone and who were apparently unable or unwilling to give up injecting. Main outcome measures: Retention in treatment, illicit drug use, HIV risk behaviour, criminal activity, social functioning, health and psychological status as measured by self-report, urinalysis and doctors' ratings. Results: Thirty-seven patients (64%) chose heroin and 21 (36%) chose injectable methadone. Fifty (86%) were retained in treatment after three months, 40 (69%) after six months and 33 (57 %) after 12 months. Among those in treatment at three months, there were significant reductions in illicit drug use, illicit drug-injecting risk behaviour, and criminal activity, and significant improvements in social functioning, health status and psychological adjustment. Generally, these gains were sustained between three, six and 12 months. Doctors' ratings of health and urinalysis results further supported these findings. Conclusions: Injectable heroin is not always the drug of choice. This intervention retained most patients in treatment with substantial benefits to both patients and the community. Prescribing injectable opiates to long term injecting drug users is a feasible treatment option. Introduction Opiate dependency is a major public health and social problem. Oral methadone treatment is the most common form of treatment for opiate dependency, and its effectiveness has been well demonstrated.1 However, Britain is one of the few countries where doctors can legally prescribe pharmaceutical heroin (diamorphine) and injectable methadone, and injectable opiates have been prescribed for opiate dependency since the 1920s in the case of heroin, and since the 1970s in the case of methadone.2Despite the legality of this practice, numbers of patients receiving such treatment remain small. Methadone accounts for 96% of all opiate prescriptions for treating drug dependency in the United Kingdom;3 injectable heroin accounts for only 2% of the total number of prescriptions for opiates;4 and methadone ampoules make up 9% of all methadone prescriptions.3 While injectable methadone is prescribed by doctors attached to specialist drug services, general practitioners and private doctors, a special licence from the Home Office is required to prescribe injectable heroin, and most doctors with such licences work in specialist drug services.5 Injectable opiate prescribing has not been guided by clinical protocols, but has been characterised by a high level of flexibility in decisions about eligible patients and dosage. The prescribing of injectable opiates has been the source of international controversy and debate. A large research trial has been undertaken in Switzerland6 and, in Australia, a heroin trial was designed to determine the impact of offering the choice of an injectable opiate prescription.7 However, the Australian Federal Government has decided that this trial will not proceed. It has been argued that injectable opiate prescribing may attract resistant opiate-dependent users into, and retain them in, treatment, with potential health and social benefits, including a reduction in crime,8 but there is a lack of scientific evidence on which to base these claims. The Swiss trial found that heroin prescribing retained drug users in treatment, with reductions in crime and improvements in health status.6 However, results of the most influential research carried out in the UK -- a randomised controlled trial comparing maintenance with oral methadone with heroin -- were inconclusive.9 In that study, the heroin group were better retained in treatment, but they continued to inject regularly and use illicit opiates in small amounts, which could lead to increasing clinic caseloads. On the other hand, the oral methadone group were more likely to either drop out of treatment or become abstinent. There were higher arrest rates and higher levels of drug involvement and criminal activity among those who did not become abstinent. Other studies have also found conflicting results.10-12 In this study we aimed to assess the feasibility of offering opiate-dependent injecting drug users the choice of treatment with injectable heroin (diamorphine) or injectable methadone, and to examine possible health and social gains associated with prescribing injectable opiates. Methods This pilot study was carried out at a west London drug treatment clinic. The clinic's protocol aimed to reduce illicit drug use and HIV risk behaviours; to improve physical, psychological and social functioning; and to move patients on to oral methadone treatment en route to abstinence. Eligible patients met the following criteria: aged over 21 years; dependent on opiates; unable or unwilling to give up injecting (defined by injecting for a minimum of three years with consistent injecting over the previous nine months and evidence of injecting over the past three months); previously failing oral methadone treatment (defined by regular continued use of illicit opiates while receiving oral methadone, continuing to inject regularly and receiving doses of oral methadone in excess of 80 mg/day); and problems relating to drug use in areas of health, social functioning or crime. Patients chose treatment with either injectable diamorphine (heroin) or injectable methadone. After a one-month induction period, they were required to stay with their drug of choice. A ceiling dose of 200 mg/day of either drug was set, and the treatment dose was achieved through tolerance testing over one week and stabilisation over the first month; after this time doses could only be reduced. Drugs were dispensed at the clinic daily (Monday to Friday with weekend doses taken home) for the first few weeks and then less frequently (a few times a week or weekly). After their initial tolerance test, patients were not permitted to inject on-site. To reduce the risk of injectable opiates being diverted to others, patients had to return used ampoules (batch numbers were checked) before receiving further ampoules. Criteria for disciplinary discharge from the study were evidence of "double-scripting" (receiving an additional prescription for opiates for drug dependency outside the clinic), dealing in the injectable opiates prescribed in this study, violence in the clinic, continual and persistent evidence of illicit drug use, consistently failing to return used ampoules, and injecting outside the clinic in the hospital grounds. The Australian Opiate Treatment Index (OTI)13 -- a multidimensional scale measuring illicit drug use, HIV risk behaviour, criminal activity, social functioning, physical and psychological health -- was administered to patients at entry and at three-monthly intervals. This instrument, which has been validated in both Australia13 and the UK,14 has a high correlation with doctors' and nurses' reports and with urine results at opiate treatment settings.14 Clinic doctors used standardised instruments to rate patients' health and psychological well-being at entry and at three-monthly intervals. These instruments included a physical rating scale for recording doctors' opinions of patients' physical health, including cardiovascular, respiratory, gastrointestinal, central nervous system and injecting-related health (Tallack F, Metrebian N, The Centre for Research on Drugs and Health Behaviour, London, 1996), and the Brief Psychiatric Rating Scale15 for recording doctors' judgements of patients' psychopathology. In addition, illicit drug use was measured through random urine tests throughout the treatment. Statistical analysis Tests of significance included t tests for related samples, and McNemar's chi-squared tests for dichotomous and ordinal data. Results from parametric and non-parametric tests were compared where normal approximations to the underlying distributions were suspect. We considered a P value of 0.05 to be significant. Results Fifty-eight subjects were recruited to the study. Forty-two (72%) were male, 50 were white (86%), and their median age was 38 years (range, 24-49 years). Their median duration of injecting heroin was 19.5 years (range, 4-30 years), and they had been in opiate treatment a median of four times previously (range, 2-17 times). Thirty-seven (64%) chose to receive diamorphine, 21 (36%) chose methadone. A higher mean dose of diamorphine was prescribed over three months (diamorphine, 181.43 mg/day [SD, 22.2; range, 120-200] v. methadone, 148.18 mg/day [SD, 45.1; range, 100-200) and over 12 months (185.24 mg/day [SD, 15.7; range, 150-200 v. 161.25 mg/day [SD, 46.4; range, 90-200]). Fifteen of the patients prescribed heroin (71%) reported experiencing night-time withdrawal symptoms as a result of heroin's shorter duration of action and were thus given an additional prescription for oral methadone. The mean dose of additional oral methadone was 24 mg/day (SD, 5.2; range, 20-30). A number of patients were found loitering in the vicinity of the clinic after their prescriptions were dispensed, and one was found injecting the prescription while still on hospital premises. No other major postdispensing problems were reported in the surrounding community. Fifty patients (86%) were still in treatment after three months, 40 (69%) after six months and 33 (57%) after 12 months. Reasons for discontinuing treatment are given in Box 1. Patients still in treatment at three months had significantly reduced their consumption of illicit drugs (with the exception of amphetamines, which did reduce, but not significantly), and made positive changes in all health and social domains. Self-reported criminal behaviour was initially low and further significantly reduced. There were significant improvements in social functioning (employment, housing, relationships and involvement in drug-using networks), health status and psychological adjustment. HIV risk behaviour (injecting and sexual risk behaviour -- the OTI considers any injecting as risk behaviour), and illicit drug-injecting risk behaviour (frequency of injecting and sharing of illicit drugs) reduced significantly. Although not significant, levels of sexual risk behaviour reduced slightly. There were no significant differences in measures of health and social behaviour between three and six months (data not shown), except for illicit drug injecting, which increased, although at six months it was still significantly less than at entry (1.13 [SD, 2.06] v. 5.66 [SD, 4.11]; P < 0.0001). Between six and 12 months, there were significant reductions in HIV risk behaviour, illicit drug-injecting risk behaviour and sexual risk behaviour, but no other significant changes. Results of urinalysis suggest that there were (non-significant) reductions in tranquilliser, amphetamine, and cocaine use between entry and three months, and between three and six months, which were sustained between three and 12 months. Few patients receiving injectable methadone were using illicit opiates. Measures of health and psychological status as reported by clinic doctors suggest health and psychological well-being had significantly improved at three months and these improvements were generally sustained. These findings were similar to measures of self-report. All measures reported at six and 12 months had significantly improved or reduced compared with measures obtained for the same sample at entry, with the exception of HIV risk, sexual risk and doctors' rating of psychological well-being, which had not significantly reduced or improved at six months but did reach significance again at 12 months. There were no differences found in drug use, health or social status reported at entry between those leaving treatment before 12 months and those remaining. Discussion These pilot study findings showed that opiate-dependent injecting drug users with long injecting careers (most started between 1970 and 1982) and for whom opiate treatment had failed multiple times previously were attracted into and retained by therapy with injectable opiates. Compared with drug users in a national UK study of oral methadone maintenance programs,16 our patients were older (38 v. 29 years) and had been injecting for longer (19.5 v. 9 years). While some clinicians17 and drug users' forums perceive a high demand for heroin treatment, our findings suggest that heroin is not always the drug chosen by users, with over one-third choosing injectable methadone. There has been much discussion about the correct dose of heroin.18 Over 12 months doses remained within the limit of 200 mg/day. In the Swiss study patients were stabilised on much higher doses of 500-600 mg/day.19 While interpretation of our findings is limited by the absence of a control group receiving oral methadone therapy and by reliance on self-report data, this is one of few studies to systematically examine the use of injectable opiates in treating opiate dependence, and the results will be used to inform a multicentre randomised controlled trial. In our study, at six months, 31% of patients had left treatment, and 40% of the 25 who left over the full 12 months were discharged because they violated the study protocol. Only two (16%) requested a move to oral methadone therapy and one became abstinent. There was one fatality (from hepatic failure) but no overdoses. By comparison, in the Swiss study of 366 patients receiving injectable heroin,20 18% had left treatment at six months and, of these, 48% switched to another treatment modality (mostly methadone maintenance) and 25% were excluded for threat of violence or other inappropriate behaviour. Four died and there were no overdoses. The retention rate in our study was higher than that reported by a national study of oral methadone maintenance programs at one month15 (86% v. 78%), and similar at six months (69% v. 67%).21 This suggests that the long-term opiate-dependent drug users in our study were well retained in treatment. Other studies have found similar high retention levels.9-11 There are concerns that prescribing injectable opiates might encourage drug users to continue injecting and discourage them from accepting oral methadone treatment or becoming abstinent.9 It is impossible to know whether our patients would have been more likely to move towards abstinence had they received oral methadone. However, these patients were long-term opiate-dependent drug users who had had a median of four previous opiate treatments and had tried and failed at least two oral methadone treatments without achieving abstinence. Our patients made significant health and social gains and experienced significantly reduced drug-related harm in the first three months. These gains were generally sustained between three and six, and six and 12 months. At entry, there were no significant differences in health and social status between patients later discharged from treatment and those who remained in the study. There has been some concern that patients' health might deteriorate while receiving prescribed injectable drugs.10 Previous research has shown little improvement in health and social wellbeing.9-11 The study by Battersby et al10 of drug users at high risk from serious physical illness, including HIV, found one individual developed a life-threatening illness (cervical spine osteomyelitis resulting from intravenous drug use) but survived, several injected into their femoral vein (a highly dangerous practice which carries a high risk of causing deep venous thromboses) and the health of 20% deteriorated. However, the authors concluded that "it was not possible to determine the nature of risk taking that would have occurred ... in the absence of the present treatment intervention". By contrast, we found that, although some patients were also injecting into their femoral vein, two moved to oral methadone because of poor health related to injecting and one died, for those remaining in treatment significant improvements were made in health, psychological adjustment and social functioning. These improvements were seen between entry and three months, and sustained between three, six and 12 months. There was also a reduction in illicit drug-injecting risk behaviour. Doctors' ratings of patients' health and psychological well-being verified these improvements. The problems of vein care associated with injectable opiates need to be weighed up against benefits in client attraction and retention, and in the provision of clean pharmaceutical drugs and injecting equipment, and advice on safer injecting practices and healthcare. Our findings do not support the suggestion that prescribing injectable opiates eliminates illicit drug use and criminal activity,22-24 as both declined significantly, but neither was eliminated. Similar results have been obtained by other studies.9-11 Results of urinalysis corroborated self-reported non-opiate drug use. However, it was not possible to corroborate self-reported opiate use as we could not differentiate between illicit and prescribed heroin. Diversion of prescribed opiates to others is of particular concern when there is no observed on-site injecting. While the clinic attempted to reduce the risk of such diversion, ways of corroborating self-reported compliance with prescription and illicit opiate use are needed. Prescribing injectable opiates is one of many options in a range of treatments for opiate-dependent drug users. In showing that it attracts and retains long term resistant opiate-dependent drug users in treatment and that it is associated with significant and sustained reductions in drug use and improvements in health and social status, our findings endorse the view that it is a feasible option. Further research is needed to examine the potential benefits of this treatment at both an individual and community level. Randomised controlled trials comparing alternative treatments and their relative cost effectiveness are required to fully assess this treatment option. Acknowledgements The Centre for Research on Drugs and Health Behaviour is core funded by the North Thames Office of the NHS Executives Research and Development Directorate. Riverside Mental Health Trust Substance Misuse Service, Ealing, Hammersmith and Hounslow and Kensington, Chelsea and Westminster Health Authorities funded this study. Thanks to all the clients who took part in the research, and to the clinic staff: Colin Small, Mark Lee, Victor Mtutu, Sue Byers, Movena Lucus, Gail Jones, Nicky Meux and Sam Nyeck. References Farrell M, Ward J, Mattick R, et al. Methadone maintenance treatment in opiate dependence: a review. BMJ 1994; 309: 997-1001. Strang J, Gossop M, editors. Heroin addiction and drug policy: the British system. Oxford: Oxford Medical Publications, 1994. Strang J, Sheridan J, Barber N. Prescribing injectable and oral methadone to opiate addicts: results from the 1995 national postal survey of community pharmacies in England and Wales. BMJ 1996; 313: 270-272. Strang J, Sheridan J. Heroin prescribing in the "British System" of the mid 1990s: data from the 1995 national survey of community pharmacies in England and Wales. Drug Alcohol Rev 1997; 16: 7-16. Sell L, Farrell M, Robson P. Prescription of diamorphine, dipipanane and cocaine in England and Wales. Drug Alcohol Rev 1997; 16: 221-226. Uchtenhagen A. Summary of the synthesis report. In: Uchtenhagen A, Gutzwiller F, Dobler-Mikola A, editors. Programme for a medical prescription of narcotics: final report of the research representatives. Zurich: Institute for Social and Preventive Medicine, University of Zurich, 1997. Bammer G. Should the controlled provision of heroin be a treatment option? Australian feasibility considerations. Addiction 1993; 83: 467-475. Metrebian N, Shanahan W, Stimson GV. Heroin prescribing in the UK: an overview. Eur Addict Res 1996; 2: 194-200. Hartnoll RL, Mitcheson MC, Battersby A, et al. Evaluation of heroin maintenance in controlled trial. Arch Gen Psych 1980; 37: 877-883. Battersby M, Farrell M, Gossop M, et al. "Horse trading": prescribing injectable opiates to opiate addicts. A descriptive study. Drug Alcohol Rev 1992; 11: 35-42. McCusker C, Davis M. Prescribing drug of choice to illicit heroin users: the experience of a UK Community Drug Team. J Subst Abuse Treat 1996; 13; 6: 521-531. Stimson GV, Oppenheimer E. Heroin addiction treatment and control in Britain. London: Tavistock Publications, 1982; 229-252. Darke S, Hall W, Wodak A, et al. Development and validation of a multi-dimensional instrument for assessing outcome of treatment among opioid users: the Opiate Treatment Index. Br J Addict 1992; 87: 733-742. Adelekan M, Green A, DasGupta N, et al. Reliability and validity of the Opiate Treatment Index among a sample of opioid users in the United Kingdom. Drug Alcohol Rev 1996; 15: 261-270. Overall JE, Gorham DR. The brief psychiatric rating scale. Psych Rep 1962; 10: 799-812. The Task Force to Review Services for Drug Misusers. Report of an independent review of drug treatment services in England. London: Department of Health, 1996; 61-64. Marks J. Who killed the British System? Druglink 1995; 10: 21. Strang J, Farrell M. Treatment of heroin users [letter]. BMJ 1994; 308: 718. Farrell M, Hall W. The Swiss heroin trials: testing alternative approaches [letter]. BMJ 1998; 316: 639. Uchtenhagen A, Dobler-Mikola A, Gutzwiller F. Medical prescription of narcotics. Eur Addict Res 1996; 2: 201-207. Gossop M, Marsden J, Stewart D, et al. The National Treatment Outcome Research Study in the United Kingdom: six month follow-up outcomes. Psych Addict Behav 1997; 11: 324-337. Beckett D. Prescription of controlled drugs to addicts. BMJ 1983; 287: 127. Connell PH. Drug dependence in Great Britain: a challenge to the practice of medicine. In: Steinberg, editor. Scientific basis of drug dependence. London: Churchill Livingstone, 1969. Marks J. Opium the religion of the people. Lancet 1985; i: 1439-1440. (Received 9 Sep 1997, accepted 24 Apr 1998) Authors' details The Centre for Research on Drugs and Health Behaviour, Department of Social Science and Medicine, Imperial College School of Medicine, London, United Kingdom. Nicky Metrebian, BA(Hons), Research Fellow; Gerry V Stimson, PhD, Professor, and Director. Chelsea and Westminster Drug Treatment Unit, Riverside Mental Health Trust Substance Misuse Service, London, United Kingdom. William Shanahan, MRCPsych, Lead Clinician, and Clinic Director. Riverside Mental Health Trust, London, United Kingdom. Brian Wells, MRCPsych, Trust Medical Director. Reprints: Ms N Metrebian, The Centre for Research on Drugs and Health Behaviour, Department of Social Science and Medicine, Imperial College School of Medicine, 200 Seagrave Road, London SW6 1RQ, United Kingdom. E-mail: n.metrebian@cxwms.ac.uk Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia ".

Nicky Metrebian · William Shanahan · Brian Wells · Gerry V Stimson

Abstract Objective: To define recent trends (1993-1996) in incidence of endstage renal disease (ESRD) among Australian Aboriginal people in the Top End of the Northern Territory (NT). Design: Analysis of hospital and clinical records of the Darwin-based ESRD treatment program from 1993 to 1996 and comparison with data accumulated since 1978. Participants: All people entering the ESRD treatment program from 1978 to 1996. Main outcome measures: Number of patients treated for ESRD; their ethnicity, age and sex; comorbidities in Aboriginal patients; treatment methods and outcomes. Results: More Aboriginal people presented with ESRD between 1993 and 1996 (87) than in the previous 15 years of the program (68). The incidence of ESRD in Aboriginals reached 838 per million in 1996, and is doubling every 4 years. Aboriginal people presenting with ESRD are younger than non-Aboriginal people with ESRD, and, in contrast to non-Aboriginals, ESRD rates are higher in women than men. The numbers and proportions of Aboriginal ESRD patients who have hypertension, type 2 diabetes and cardiac disease are rising. Haemodialysis remains the most common form of treatment, and the number of dialysis treatments is doubling every 2.5 years. Only 9% of Aboriginal patients entering the program in 1993-1996 were treated with chronic ambulatory peritoneal dialysis and only 3% received transplants. Despite their younger age, survival of Aboriginal people on dialysis is low (median 3.3 years v. 6.5 years in non-Aboriginals), and graft survival after transplant is poor (37% at 5 years v. 88% in non-Aboriginals). Survival has not improved in the past 4 years, with fewer deaths from infection offset by more deaths from cardiovascular disease. Conclusions: The predicted doubling of ESRD incidence among Aboriginal people by the year 2000 will add an enormous burden to limited resources. Risk factors for renal disease underlie all the excess morbidity and mortality in NT Aboriginal adults, and arise out of accelerated lifestyle changes and socioeconomic disadvantage. Better living conditions and education, robust and integrated primary healthcare programs, and systematic screening for early renal disease and treatment of those with established disease are all matters of urgency. Introduction The high rate of endstage renal disease (ESRD) in Australian Aboriginal people is attracting increased attention.1-4 The Northern Territory (NT) faces particular challenges in dealing with this problem. It has the lowest population (179 000 people) and tax base of any Australian State, and the highest proportion of Aboriginal people (22% of the population, compared with 5%-6% for Queensland and Western Australia -- the States with the next-highest proportion). It has an area of 1 346 200 km2, and only two dialysis centres, 1500 km apart (one in Darwin and one in Alice Springs), which can scarcely accommodate the increasing numbers of Aboriginal people needing treatment. In a previous article,1 we reported the rising rates of ESRD among Aboriginal people in the Top End of the NT (Figure 1), with an average annual incidence of 440 per million from 1988 to 1993. Aboriginal patients were younger than their non-Aboriginal counterparts and, in contrast to the non-Aboriginal population, more women than men were affected. There was little use of chronic ambulatory peritoneal dialysis (CAPD), and the few transplants gave poor results. Comorbidities were high and there was a shorter survival time on treatment than in non-Aboriginal people. Tiwi people living on Bathurst and Melville islands had especially high rates of ESRD, but for other communities there were too few cases to allow calculation of population-adjusted rates. We present data for ESRD treatment in the Top End of the NT for the period 1993-1996, thus extending the period of analysed data to 19 years. The expanded data allow calculation of regional population-specific rates, and illuminate trends in diagnoses, comorbidities, treatment methods and survival. Methods Study population The total study population comprised all people entering the Top End ESRD treatment program from January 1978 to December 1996 inclusive. Our data sources were dialysis unit records and hospital files. We used Australian Bureau of Statistics aggregate and regional data from the 1987, 1991 and 1995 population censuses for the denominators for calculating ESRD rates for the 4-year intervals 1985-1988, 1989-1992, and 1993-1996.5The 1991 Aboriginal community census was used to calculate individual community rates for the three 4-year intervals. People moving to the NT from other countries or other States and becoming residents are potentially included in the census and thus the denominators. Statistical analysis We calculated age-standardised rates of ESRD with Epi-Info6 using aggregate Australian non-Aboriginal data from the 1996 Australian and New Zealand Dialysis and Transplant Registry (ANZDATA) as the reference population.7χ2 tests with Yates' correction were used to examine differences between groups. Kaplan-Meier survival analyses were performed by SPSS for Windows,8 and the log-rank test was used to compare the differences between the groups. Results In the period 1993-1996, 104 people entered the ESRD treatment program, making a total of 214 for the period 1978-1996. Of the total, 155 (72%) were Aboriginal, and 59 (28%) were non-Aboriginal. Incidence of treated ESRD Figure 2a shows the dramatic increase in incidence of ESRD in Aboriginal people over the three 4-year periods, with an average doubling time of about 4 years. More Aboriginal people (87) presented for treatment in the 4 years 1993-1996 than in the previous 15 years of the program (68). Within this most recent 4-year period, rates have continued to rise at an average of 22% per year, peaking at 838 per million9 in 1996 (compared with 39 per million in non-Aboriginal people in the NT). With age adjustment, the 1993-1996 average annual rate represents a 15-fold increase and the 1996 rate a 21-fold increase over ESRD rates in non-Aboriginal Australians nationwide. NT rates for non-Aboriginals, which did not change over this time period, are lower than Australian aggregate rates (62 per million from 1993-1996) because the NT has a younger population. All regions of the Top End are experiencing the same phenomenon, although current rates vary. Figure 2b shows the changes in incidence in Aboriginal people in the five regions with the highest current rates. The number of Aboriginal people receiving ESRD treatment (the prevalence), including those with functioning transplants, peaked at 2871 per million in 1996 versus 377 per million for non-Aboriginals. Finally, the number of dialysis procedures (which accrue most of the costs) is rising by 28% per year, or doubling every 2.5 years.9 Sex and age distribution of ESRD patients Ninety-three (60%) of the 155 Aboriginals with ESRD were women and 62 (40%) were men, compared with 23 (39%) women and 36 (61%) men among the 59 non-Aboriginal patients. Figure 2c shows the higher ESRD rates in Aboriginal women than men at each time period, resulting, in 1993-1996, in an age-adjusted relative risk of 31 in women versus 16 in men. On average, Aboriginal people were 5 years younger than non-Aboriginal people on entering the program (44 v. 49 years). However, the age distributions were quite different, with Aboriginal people most commonly presenting between the ages of 30 and 49 years, and non-Aboriginals between 50 and 69 years. Figure 3 shows that ESRD rates in Aboriginal people rose in most age groups over the life of the program, so that the average age and age distribution did not change appreciably. The age-specific incidence of ESRD in Aboriginal people peaked at age 50-59 years, and in non-Aboriginals at over 70 years. Renal failure causes and comorbidities in Aboriginals The Table compares the distribution of "causes" of ESRD in the two most recent 4-year intervals in those patients with documented causes. The proportion of ESRD in Aboriginal people attributed to (but not always solely due to) diabetes has almost doubled, that classified as glomerulonephritis has fallen by more than half, and the proportion in the "unknown" category has increased markedly. Aboriginal people with ESRD are more likely than non-Aboriginal people to have type 2 diabetes (48% v. 24%; P = 0.002) and hypertension (52% v. 32%; P = 0.01). Furthermore, these proportions have increased recently: for diabetes from 37% pre-1993 to 60% in 1993-1996 (P < 0.005); and for hypertension from 43% to 57% (P = 0.1); and for people with both diabetes and hypertension from 19% to 38% (P = 0.011). Treatment for ESRD Fifty-one per cent (30) of non-Aboriginal people, but only 17% (26) of Aboriginal people, have received transplants since 1978, and only 3% in the period 1993-1996. Potential reasons for this discrepancy include medical ineligibility, discouraging earlier results, distance to the transplant centre (in Adelaide), underservicing and difficulties in supervision of care in remote areas, lack of living related donors, and difficulties in HLA matching with donor organs. In 1996, only 11% of Aboriginal people with ESRD had functioning transplants versus 65% for non-Aboriginal people. Chronic ambulatory peritoneal dialysis (CAPD) rates in Aboriginal patients remain low, with only 5% of incident cases pre-1993 and 9% in 1993-1996 treated in this manner. Patient and graft survival Despite the younger age of Aboriginal patients with ESRD, their "integrated" survival (with all forms of treatment, ie, dialysis and transplantation) was significantly worse than that of non-Aboriginal people, with median survival times of 3.6 versus 12.3 years (P = 0.0025). This difference was reflected in those with and without diabetes. While survival on dialysis (CAPD and haemodialysis), as shown in Figure 4, tended to be lower in Aboriginal people (median 3.3 v. 6.5 years; P = 0.34), both graft and patient survival after transplantation were clearly worse (Figures 5 and 6). Patient survival at 1 and 5 years after transplantation was 92% and 60% for Aboriginals, compared with 97% and 93% for non-Aboriginals (P < 0.001), and graft survival at 1 and 5 years was 73% and 37% for Aboriginal people, compared with 97% and 88% for non-Aboriginal people (P < 0.001). There has been no improvement in integrated survival, or in dialysis or transplant survival separately, in the period 1993-1996. Causes of death The causes of death in Aboriginal people have changed for the period 1993-1996. Deaths due to cardiac disease have increased from 33% before 1993 to 51% after 1993. Deaths due to voluntary withdrawal from treatment are unchanged (24% and 25%, respectively). In contrast, only one non-Aboriginal patient has ever withdrawn from treatment, and none in the past decade. Discussion These data for 1993-1996 confirm and further define the rise in incidence of renal failure among Aboriginal people in the Top End of the Northern Territory. No community is spared, and this pattern is repeated in Central Australian Aboriginals, who have even higher rates (1400 per million in 1996).4 Current incidence rates are now comparable with those of Afro-Americans and Native Americans (800 and 744 per million, respectively, in 1995); however, age-adjusted rates are much higher in Australian Aboriginals (due to their very youthful population) and are increasing much faster, with a doubling time less than 4 years versus 10 years for the US minority groups.10 These alarming figures are nonetheless underestimates, because some Aboriginal people, especially older people and those from very remote communities, decline treatment for ESRD or are medically unsuited. Most of the increase is real. It is not due to ageing of the Aboriginal population, as it is reflected across every age group; and it is not due to improved ascertainment, at least in the major communities, as awareness has been high since the mid 1980s. The disability and the personal, family and community disruption are great, and the resource requirement will be truly formidable if current rates of increase, which project up to 500 new cases between 1997 and 2004 (exponential progression), are sustained. The data confirm the younger age and female predominance of Aboriginal people with ESRD. The latter might be due, in part, to lower birthweights and the relatively higher adult body weights in women, with more marked insulin resistance and earlier onset of type 2 diabetes.11,12 Both occur also in Central Australia.4 Some of the change in attributed cause of ESRD over the most recent 8-year period reflects subjectivity in assignment criteria, but much reflects reassignment of the common finding of bland glomerulomegaly with absent or minimal inflammation13-15 from the "glomerulonephritis" category to the "unknown" category. The rise in renal failure attributed to diabetes is compatible with the dramatic increase in rates of diabetes and its complications in all Aboriginal communities,16 but clinical and biopsy data show that diabetes is more often a facilitating factor for disease expression and progression rather than the prime or sole cause of the underlying nephropathy.11,14 The increased incidence in all communities of Syndrome X (obesity, hypertension, dyslipidaemia, dysglycaemia and predisposition to cardiovascular disease), which is attributed to insulin resistance, likewise explains the increasing proportions of Aboriginal people presenting for ESRD treatment with diabetes and hypertension as comorbidities.11,16-18 These conditions are already generating more heart attacks, strokes, coronary angioplasties, coronary artery bypass graft procedures and cardiovascular deaths among Aboriginal ESRD patients. As the epidemic grows and the Aboriginal population ages, these complications will become more common, and further increase costs, complicate treatment and compromise survival. The persistently high rate of withdrawal of Aboriginal people from ESRD treatment reflects difficulties with chronic disability, the complex treatment regimen, and loss of social and family support and "land identity" associated with relocation from their community to Darwin for treatment.19,20 Initiatives to move treatment closer to home include attempts to promote CAPD, which has low rates of technical failure and peritonitis, but major problems with exit-site infections;20 a renewed focus on transplant (seven Aboriginal people received transplants in 1997); construction of the first community-based dialysis unit with seven stations on the Tiwi islands; and the possibility of placing haemodialysis stations in clinics in high risk remote areas. However, deliberations about efficient and equitable delivery of ESRD treatment must no longer be allowed to dominate the dialogue. Dialysis in the Top End currently costs $496 per treatment, which is about $75 000 per patient per year. This cost does not include medicines, relocation and housing, transportation and hospitalisations.9 The allocation of resources of this magnitude to people with a median life expectancy of 3.3 years must be balanced by serious and sustained community-based initiatives to prevent and ameliorate the underlying problem. Most renal disease in Aboriginal communities is marked by albuminuria, and all renal failure arises in people with a history of progressive overt albuminuria.11Risk factors for renal failure include low birthweight and infant malnutrition, infections (scabies, poststreptococcal glomerulonephritis), increasing adult weight, high blood pressure, increasing glucose levels, insulin resistance, dyslipidaemia, and heavy drinking.11 Several risk factors can operate simultaneously, progressively compounding the decline in renal function that accompanies increasing age. The current epidemic is probably explained by the confluence of many risk factors over a short time period, associated with dramatic lifestyle changes and serious socioeconomic disadvantage. Ironically, the great fall in infant mortality between the late 1950s and late 1970s, a consequence of better hospital management of sick babies, means that those low birthweight babies now surviving to adult life are at high risk for renal and other chronic diseases.11,21 This multifactorial perspective on renal disease necessitates a rethinking of renal disease classifications.11 There is a need for a stronger focus on community and individual risk factor profiles and on pathophysiological interactions, and some de-emphasis of categorical definitions. It justifies a general preventive health services model, which will also reduce the diabetes, hypertension, cardiovascular disease, chronic lung disease and infections that contribute to the excess mortality in NT Aboriginal adults.17 In addition, screening programs to recognise early and established renal disease, and treatment (including angiotensin-converting enzyme inhibitors) to arrest disease progression, must be incorporated into regular adult healthcare in every Aboriginal community as a matter of urgency.22,23 Acknowledgements This study was supported by the National Health and Medical Research Council of Australia, and the Australian Kidney Foundation. We thank Dr David Pugsley, who led the way with renal services in the Northern Territory, and established the basis of these observations, and Dr Diane Howard and Dr Sid Selva-Nayagam, who have had the longest tenure of care of renal patients. We thank the staff of the Nightcliff Dialysis Unit and the Renal Unit at Royal Darwin Hospital for their excellent care and cooperation. Dr Zhiqiang Wang assisted with statistical analyses, and Ms Susan Jacups with clerical and graphic support. References Hoy WE, Mathews JD, Pugsley DJ. Treatment of end-stage renal disease in the Top End of the Northern Territory: 1978-93. Nephrology 1995; 1: 307-313. Hoy WE, Silva D. NT Top End Aboriginal end stage renal disease data and projections. Darwin, NT: Renal Strategy Committee, Territory Health Services, February 1997. Hoy WE. Renal disease in Aboriginal Australians. Med J Aust 1996; 165: 126-127. Renal disease in Central Australia -- challenges and opportunities for better health. Health Strategies, Deakin. Melbourne: Deakin University, September, 1997. Australian Bureau of Statistics. 1996 Census of population and housing. Aboriginal community profiles, small area data. Canberra: ABS, 1996. (Catalogue No. 2020.0.) Dean AG, Dean JA, Coulombier D, et al. Epi-Info, version 6: a word processing, database and statistics program for epidemiology on microcomputers. Atlanta, Ga: Centers for Disease Control and Prevention, 1994. Disney APS, editor. ANZDATA Report 1996. Nineteenth report of the Australian and New Zealand Dialysis and Transplant Registry. Adelaide: Queen Elizabeth Hospital, 1996. SPSS -- Statistical Package for the Social Sciences [computer program], release 6.1, standard version. Chicago: SPSS Inc, 1988. Margetts C, Morris E. Cabinet submission to the NT Legislature for renal services funding. Darwin: Royal Darwin Hospital, February 1998. United States Renal Data System (USRDS). Annual Data Report. Bethesda, MD: The National Institutes of Health, NIDDK, April 1997. Hoy WE, Mathews JD, McCredie DA, et al. The multidimensional nature of renal disease: findings in a high risk Australian Aboriginal community. Kidney Int 1998. In press. Hoy WE, Norman RJ, Hayhurst B, G Pugsley DJ. A health profile of adults in a Northern Territory Aboriginal community, with an emphasis on preventable morbidities. Aust N Z J Public Health 1997; 21: 121-126. Lloyd ML, Moore L, Pugsley DJ, Seymour AM. Renal disease in an Australian Aboriginal population: a pathologic study. Nephrology 1996; 2: 315-322. Howard DM, Davis J, Pugsley DJ, et al. Morphologic correlates of renal disease in a high risk Aboriginal community. Proceedings of the 31st Annual Scientific Meeting, Australian and New Zealand Society of Nephrology; Perth, WA; March 5-8, 1996. Sydney: Australian and New Zealand Society of Nephrology, 1996. Bertram JH, Young RJ, Seymour AE, et al. Glomerulomegaly in Australian Aborigines. Nephrology 1998. In press. Markey P, Weeramanthri T, Guthridge S. Diabetes in the Northern Territory. Darwin: Diabetes Australia (Northern Territory), 1996. Cunningham J, Condon J. Premature mortality in Aboriginal adults in the Northern Territory, 1979-1991. Med J Aust 1996; 165: 309-312. Reaven GM. Insulin resistance and compensatory hyperinsulinemia: role in hypertension, dyslipidemia, and coronary artery disease. Am Heart J 1991; 121: 1283-1288. Willis J. Fatal attraction: do high technology treatments for end stage renal disease benefit Aboriginal people in Central Australia? Aust J Public Health 1995; 19: 603-609. . Snelling P. I hit beautiful serves -- but my partner keeps putting the ball in the net: renal transplantation in Aboriginal patients. Proceedings of the 10th Dialysis and Transplant Workshop of the Australian Kidney Foundation and the Australian and New Zealand Society of Nephrology; Launceston, Tasmania; October 1997. Sydney: Australian and New Zealand Society of Nephrology, 1997: 62-63 Barker DJP. Mothers and babies and disease in later life. London: BMJ Publishing Group, 1994. Hoy WE. Markers for cardiovascular and renal morbidity: expectations for an intervention program in an Australian Aboriginal community. Clin Exp Pharmacol Physiol 1996; 23 (Suppl 1), S33-S37. Hoy WE. Guidelines for screening and treatment of renal disease in Aboriginal communities. Nephrology 1998. In press. (Received 5 Jan, accepted 10 Mar, 1998) Authors' details Menzies School of Health Research, Darwin, NT. Janine L Spencer, FRACP, DTM&H(Lond), Paediatrician, Australian Kidney Foundation Research Fellow. Wendy E Hoy, BScMed, FRACP, NHMRC Senior Research Fellow, Director of Renal Program; and Renal Community Services Specialist, Territory Health Service. TVW Telethon Institute for Child Health Research, Perth, WA. Desiree T Silva, FRACP, MPH, Paediatrician. Royal Darwin Hospital, Darwin, NT. Paul Snelling, FRACP, Nephrologist, Royal Darwin Hospital, and Territory Health Service. Reprints will not be available from the authors. Correspondence: Dr W E Hoy, Menzies School of Health Research, PO Box 41096, Casuarina, NT 0811. E-mail: wendyATmenzies.su.edu.au Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia ".

Janine L Spencer · Desiree T Silva · Paul Snelling · Wendy E Hoy

Abstract Objective: To determine the prevalence of the Wernicke-Korsakoff syndrome (WKS) in Australia and compare this with previous studies. Design and setting: Prospective autopsy study at the New South Wales Institute of Forensic Medicine, 1996-1997. Methods: Brains of deceased people (aged over 15 years) derived from 2212 sequential autopsies performed between 1 January 1996 and 31 December 1997 were studied macroscopically and microscopically to identify cases of WKS. Main outcome measures: Standard histological criteria for WKS and any available clinical data. Results: Twenty-five cases of WKS were identified (prevalence, 1.1%), mostly among the 5.9% of the 2212 people who had a history suggestive of alcohol abuse. Only four cases (16%) had been diagnosed during life. Conclusions: There has been a significant reduction in the prevalence of WKS in Australia since the introduction of thiamine enrichment of bread flour. While the prevalence is still higher than in most other Western countries, further research is needed before adding thiamine to alcoholic beverages can be recommended. Introduction Wernicke-Korsakoff syndrome (WKS) is a potentially fatal neurological disorder caused by thiamine (vitamin B1) deficiency. It is easily treated by thiamine supplementation, with most people showing dramatic clinical improvement. Biochemical studies in the 1980s showed that 10%-20% of Australians had a low thiamine intake, with those who used alcohol excessively being particularly "at risk",1 and previous autopsy studies had shown that Australia had the highest recorded prevalence of WKS in the world, with most cases occurring among those who abused alcohol.2-5 To address these problems, the Australian Health Ministers' Council recommended the mandatory enrichment of bread flour with thiamine. Consequently, since 1991 bread flour in Australia has contained not less than 6.4 mg of thiamine per kilogram.6 Human thiamine requirements have been estimated to be 1.0-1.5 mg daily.7 The major food sources of thiamine are cereal products. Although thiamine is naturally present in the grains used to produce flour, much is removed during some types of processing, and thiamine enrichment of flour has been practised in a number of countries for many years.8,9 It is a mandatory requirement in the United Kingdom, Canada, and Denmark, while in the United States enrichment is not mandatory, but most flour is enriched. Since the introduction of thiamine enrichment of rice and flour, thiamine deficiency is said to have virtually disappeared in countries such as Japan and the US.10,11 Our aim was to establish whether this safe and simple public health measure has reduced the prevalence of WKS in Australia, and hence to determine whether we need to consider alternative public health measures such as adding thiamine to all flours used in food preparation or even to alcoholic beverages.12 Methods We prospectively studied all sequential autopsies of people aged over 15 years (as WKS is rare in a younger population in Western countries) from the NSW Institute of Forensic Medicine in Sydney during 1996 and 1997. In line with standard autopsy procedures, brains were fixed in 10% formol saline for at least two weeks. After careful external examination and removal of the brainstem and cerebellum, the brain was cut in the coronal plane, the first cut being made at the level of the anterior (ventral) tip of the mammillary bodies. The hemispheres were then cut at 10 mm intervals. The brainstem and cerebellum were separated by dividing the right and left cerebellar peduncles. The cerebellum was sectioned in the sagittal plane through the vermis and through the lateral hemispheres. The brainstem was sectioned in the horizontal plane at 3 mm intervals. The brain slices were examined, paying particular attention to the mammillary bodies and the regions around the third ventricle and the floor of the fourth ventricle. It has been shown that the mammillary bodies are abnormal on microscopic examination in 99%-100% of cases of WKS.3,13 The mammillary bodies were measured in three dimensions and brain slices were photographed if any macroscopic abnormality was noted. Blocks of tissue were taken for paraffin embedding and the preparation of histological sections. The principal block for this study incorporated both of the mammillary bodies and the walls of the third ventricle. Blocks from other regions of the brain were taken according to routine neuropathological protocols of the Institute of Forensic Medicine and the Neuropathology Department.4 These varied depending on the clinical history, cause of death, and general autopsy findings. Sections were cut at 10 mm and stained with haematoxylin and eosin. Brains were stored until each case was completed, so that, if necessary, additional blocks of tissue could be taken for further examination. This was particularly important if an abnormality was identified in the section of the mammillary bodies. A number of other diseases can affect the mammillary bodies, but these can usually be easily differentiated from WKS.5,14 All cases were examined microscopically by a neuropathologist (C H) without knowledge of the relevant clinical history or the macroscopic findings. In all cases in which WKS was suspected, the mammillary body section was also stained with reticulin, Weil myelin, glial fibrillary acidic protein, Perl's iron stain, Bodian silver impregnation, and Nissl stain for neurones. The purposes of these stains are summarised in Box 1. Clinical histories and general pathological data were correlated with the neuropathological findings. Statistical comparisons were made with data from a previous, similar study3 using chisquared statistics with Yates' correction factor and relative risk with 95% confidence intervals. Results We diagnosed 25 forensic cases of WKS from 2212 brains, giving an overall prevalence of 1.1%. Box 2 compares the prevalence data with those of a similar Australian study (including both forensic and hospital cases of WKS) undertaken by one of the authors (C H) from 1973 to 1981,3 and shows that there has been a statistically significant reduction in the prevalence of WKS. The mean age at death of the people with WKS was 55 years; 23 (95%) were men. Hospital medical records, located for 18 of the 25 deceased people with WKS, showed that alcohol appeared to have been a predisposing factor for WKS in 16 (89%). Of the remaining two, one was a man with paraplegia who was totally dependent and required gastrostomy feeding, and the other man had both lung and bowel cancer. Information available at autopsy indicated that 5.9% of the 2212 brains studied were from people with a history which suggested an alcohol problem, giving a prevalence of WKS in an Australian forensic population with a history suggestive of alcohol problems of 19%. A clinical diagnosis of WKS (Korsakoff's psychosis) during life had been made in only four (16%) of the 25 deceased whose WKS was identified at autopsy. However, two others had a diagnosis of alcohol-related brain damage. The hospital records located for 18 deceased people with WKS showed that four had severe amnesia (Korsakoff's psychosis), two others had memory problems and three had ataxia or unsteady gait. None were reported to have ophthalmoplegia or nystagmus, which are commonly described clinical signs in WKS.15 Three had a history of epilepsy. Eight (44%) had been given thiamine (vitamin B1) supplements during one or more of their hospital admissions. The diagnosis of WKS in this study was based on typical pathological abnormalities, which will not be described in detail as this information is available in current neuropathological texts.6 In brief, lesions are seen in a characteristic distribution -- in the mammillary bodies and around the walls of the third and fourth ventricles. Macroscopic findings vary depending on the stage of the disease (see Box 3). There were two acute cases, six acute-on-chronic cases (evidence of both acute and chronic damage to mammillary bodies), and 17 chronic cases. These data are compared with those of a previous study3 in Box 4. Discussion Our findings show a significant reduction in the prevalence of WKS in Australia compared with the findings of a similar study of forensic cases between 1973 and 1981 (see Box 2).3 The two forensic populations were similar, even though these two studies were conducted in different Australian States (New South Wales and Western Australia) and at different times. Laws relating to deaths which must be reported to the Coroner and undergo forensic autopsy are almost identical in these two States. Moreover, one of the authors (J H) was the Director of the Forensic Department in Perth, WA, at the time of the first study and is currently the Director of the NSW Institute of Forensic Medicine, where the current study was carried out; he commented that the profiles of cases in the two autopsy studies were very similar. The high prevalence of WKS at autopsy in Australia2-5 has been reflected in clinical studies, with, for example, 170 cases of WKS being identified among 1100 total inpatients at Queensland's largest hospital for the mentally ill.12 However, a retrospective study of records from 17 major Sydney hospitals from 1978 to 1993 showed that the number of acute cases of WKS was lower in 1992 and 1993 than in any of the other years.16 This suggests that dietary enrichment of bread flour with thiamine may have had an impact on the occurrence of acute cases of WKS. Our finding of fewer cases of acute WKS compared with the earlier WA study also supports this (see Box 4, above). As patients with acute WKS who are treated appropriately with parenteral thiamine respond within days, and signs and symptoms usually resolve completely, not all cases will progress to chronic disease. The development of chronic disease is not fully understood, but many chronic cases of WKS are likely to be the result of recurrent episodes of either clinical or subclinical thiamine deficiency.17 WKS is said in the medical textbooks to have a characteristic clinical picture of mental changes (confusion, obtundation), ataxia, and eye signs (nystagmus, ophthalmoplegia). However, as shown in this study and previously, analysis of clinical signs and symptoms of patients diagnosed with WKS post mortem reveals that a minority of cases have the full clinical picture and about a third exhibit only mental changes.15 Thus, the diagnosis can be easily overlooked. Many authors agree that WKS can develop as a result of repeated "subclinical" episodes of thiamine deficiency,18 so that studies of the prevalence of this disorder based on clinical findings are unsatisfactory and may underestimate the true incidence.15 Given that most of the cases of WKS we identified were chronic, some explanation must be sought for the significant reduction in the prevalence of this disease within six years of the introduction of thiamine enrichment of bread flour. There are four possible explanations: 1. A true reduction in the number of new (acute) cases of WKS as a result of the increase in dietary thiamine, and reduced numbers of "at risk" cases. 2. An improvement in the clinical status of patients who already have WKS and a reduced occurrence of further clinical or "subclinical" episodes of thiamine deficiency.17 This could result in an increase in the longevity of patients with existing WKS -- for example, through a reduction in the sudden unexpected deaths seen among chronic alcoholics. 3. A general improvement in the health status of the Australian population, independent of the introduction of thiamine supplementation. 4. Publication of the high prevalence of WKS in Australia may have led to an increased awareness, particularly among health professionals, of the necessity for thiamine treatment in "at risk" patients. Almost half of the 25 people whose WKS was identified at autopsy had been treated with thiamine in hospital. This increased awareness and therapeutic intervention may have played a role in the reduced prevalence. It is likely that each of these explanations played a part in reducing the prevalence of WKS. However, the relative importance of each will only be clarified with time and further study as the enrichment of bread flour with thiamine, has only been in place for six years. It should be noted that the results of this study may in fact underestimate the beneficial effect of the enrichment of bread flour with thiamine as many of the cases of chronic WKS will have developed before the enrichment program commenced. Further prevalence studies should be carried out in, say, another five years, and this information, together with biochemical studies of the thiamine status of the Australian population, will enable rational decisions to be made with regard to other public health programs, such as fortification of alcoholic beverages with thiamine. Our findings neither support nor confirm the prediction of Price and Theodorous that many people who abused alcohol would not be protected by a flour enrichment program because alcoholic beverages are, "for a hazardously long period of their lives, their only form of caloric intake".12 However, they had suggested, on the basis of this prediction, that thiamine supplementation of alcoholic beverages might be a more effective preventive measure, and in reviewing this proposal Connelly and Price noted that fortification of beverages, rather than bread flour, would be more cost-effective by a factor of 20-40-fold.19 Although there has been a significant decrease in the prevalence of WKS in Australia since the introduction of thiamine enrichment of bread flour in 1991, the rate is still higher than in most other Western countries5 and there is a need to maintain vigilance in the management of patients, particularly those with known or suspected alcohol problems. Acknowledgements This research was supported by the Australian Brewers' Foundation and by the National Health and Medical Research Council (Grant number 943302). We are grateful to all the staff at the NSW Institute of Forensic Medicine for their assistance and cooperation. References Wood B, Breen KJ. Clinical thiamine deficiency in Australia: the size of the problem and approaches to prevention. Med J Aust 1980; 1: 461-464. Harper C. Wernicke's encephalopathy: a more common disease than realised. J Neurol Neurosurg Psychiatry 1979; 42: 226-231. Harper CG. The incidence of Wernicke's encephalopathy in Australia -- a neuropathological study of 131 cases. J Neurol Neurosurg Psychiatry 1983; 46: 593-598. Harper CG, Gold J, Rodriguez M, Perdices M. The prevalence of the Wernicke-Korsakoff syndrome in Sydney, Australia: a prospective necropsy study. J Neurol Neurosurg Psychiatry 1989; 52: 282-285. Harper C, Fornes P, Duyckaerts C, et al. An international perspective on the prevalence of the Wernicke- Korsakoff syndrome. Metab Brain Dis 1995; 10: 17-24. Harper C, Butterworth R, editors. Nutritional and metabolic disorders. 6th ed. London: Arnold, 1997: 601-664. Freeman RM, editor. Rational use of vitamins in practice. Toronto: JB Lippincott, 1979: 115-122. Axford DWE, Williams DA. Flour enrichment around the world. Br Flour Milling Baking Res Assoc 1981; 4: 156-163. Bauerfeind JC. Nutrification of food. In: Shiks ME, Young VR, editors. Modern nutrition in health and disease. 7th ed. Philadelphia: Lea and Febiger, 1988. Figueroa WG, Sargent F, Imperiale L, et al. Lack of avitaminosis amongst alcoholics. J Clin Nutrition 1953; 1: 179-199. Sebrel WH. Enrichment: good gift of yesterday. Cereal Science Today 1966; 11: 228-230. Price J, Theodorus MT. The supplementation of alcoholic beverages with thiamin: a necessary preventative measure in Queensland. Aust N Z J Psychiatry 1979; 13: 315-320. Victor M, Adams RD, Collins GH. The Wernicke- Korsakoff Syndrome. Philadelphia: Davis, 1989. Schubert T, Friede RL. Transneuronal mammillary atrophy. J Neurol 1979; 221: 67-72. Harper CG, Giles M, Finlay-Jones R. Clinical signs in the Wernicke-Korsakoff complex -- a retrospective analysis of 131 cases diagnosed at autopsy. J Neurol Neurosurg Psychiatry 1986; 49: 341-345. Ma JJ, Truswell S. Wernicke-Korsakoff syndrome in Sydney hospitals: before and after thiamine enrichment of flour. Med J Aust 1995; 163: 531-534. Lishman WA. Cerebral disorders in alcoholism. Syndromes of impairment. Brain 1981; 104: 1-20. Lishman WA. Alcohol and the brain. Br J Psychiatry 1990; 156: 635-644. Connelly L, Price J. Preventing the Wernicke-Korsakoff syndrome in Australia: cost-effectiveness of thiamin-supplementation alternatives. Aust N Z J Public Health 1996; 20: 181-187. (Received 7 Aug1997, accepted 16 Feb 1998) Authors' details Department of Pathology, University of Sydney, and Royal Prince Alfred Hospital, Sydney, NSW. Clive G Harper, MD, FRCPA, Professor; Donna L Sheedy, BA, Research Assistant; Ana I Lara, Technical Assistant. Network for Brain Research into Mental Disorders, Prince of Wales Hospital, Sydney, NSW. Therese M Garrick, RN, BNurs, Research Assistant. New South Wales Institute of Forensic Medicine, Sydney, NSW. John M Hilton, MB ChB, FRCPA, Director; Jack Raisanen, MD, Forensic Neuropathologist. Reprints: Professor C G Harper, Department of Pathology, University of Sydney, NSW 2006. E-mail: clivehATpathology.su.oz.au Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia ".

Clive G Harper · Donna L Sheedy · Ana I Lara · Therese M Garrick · John M Hilton · Jack Raisanen

Abstract Objective: To determine the health-related quality of life (HRQOL) of Australian men after radical prostatectomy. Design: Cross-sectional study. Setting: Private and public practices of three urologists in south-east Queensland, July 1989 to June 1995. Participants: 140 men with no evidence of disease recurrence 1 to 6 years after radical prostatectomy. Main outcome measures: Voiding and erectile potency and HRQOL. Recall of preoperative status and status at survey were established by an independently administered multi-item questionnaire. Results: 112 men (80%) completed the study questionnaire. Difficulty with bladder control before the operation was reported by 25 (22%; 95% confidence interval [CI], 15%-31.2%), and the incontinence rate after treatment was 22/112 (20%; 95% CI, 12.7%-28.2%). Men with incontinence after operation were more likely to recall preoperative urinary symptoms. Eighty-four (75%) men were happy or coping with their sexual function after radical prostatectomy despite an erectile potency rate of only 12% (95% CI, 7%-20%). Twenty-eight (25%) had tried penile injections and three have had penile prostheses since their operation. Impotence was reported more frequently (40%) as the treatment-related problem most affecting life, followed by "concern about cancer" (12%) and incontinence (8%). Impotence was also the most common cause given for diminished HRQOL. Conclusions: Loss of sexual function after radical prostatectomy is more commonly perceived as a major problem and is more likely than urinary incontinence to adversely affect HRQOL. Loss of sexual function and its effect on HRQOL needs to be given greater emphasis in counselling before radical prostatectomy. MJA 1998; 168: 483-486 Introduction Prostate cancer is the second most common cause of cancer-related death in Australian men, and in 1989 became the most common cancer in men in New South Wales.1 This increase may be partly attributed to a more health-conscious, ageing population, as well as greater use of "routine" digital rectal examination (DRE) and prostate-specific antigen (PSA) tests.2 PSA tests, together with DRE and transrectal ultrasound-guided prostatic biopsies, have enabled the diagnosis of potentially curable early-stage prostate cancer.3 In particular, increased efforts have been made to identify early-stage prostate cancer in men under 70 years of age, even though a significant survival advantage has yet to be demonstrated. Screening and case detection remain controversial. There is argument about whether the tests are sufficiently sensitive and specific for effective screening, and whether screening affects outcomes enough to be cost effective.4,5Opinions also differ regarding the optimal management of localised prostate cancer.6 In men over 70 years of age, or in those with appreciable co-morbidity, a conservative approach is generally accepted. Healthy younger men are more likely to live long enough to experience progression of their disease, so radical prostatectomy and radiotherapy, as well as "watchful waiting", are options in this group. Judging by current published studies, these options may provide similar outcomes in selected patients: up to 10 years after diagnosis, similar survival rates are seen in patients treated immediately with surgery or radiotherapy, and in patients initially watched and then treated with androgen ablation, transurethral resection or radiotherapy if the disease progresses.7 This has created a significant dilemma for both doctor and patient when selecting appropriate treatment.6 The lack of a clearly superior treatment option makes the impact of treatment on health-related quality of life (HRQOL) of greater importance. The difficulty in selecting appropriate treatment is compounded by a lack of Australasian data on the effect of treatment on HRQOL. Radical prostatectomy has been increasingly used in treating patients with localised prostate cancer and is considered an appropriate option for men who have a life expectancy in excess of 15 years.8 Despite improved surgical technique, postoperative impotence and incontinence may still occur, although the reported incidence varies considerably.9 Our study was undertaken to record patients' perception of HRQOL after radical prostatectomy, to assist patients, families and doctors in their discussions about treatment expectations and outcomes. Methods Sample Surgical audit data were collected prospectively on all patients having radical prostatectomy between July 1989 and June 1995. Three of us (P S H, A N J and L C T) performed the operations. We all receive referrals from other specialist urologists, and perform most of the radical prostatectomies in Queensland. We believe that the patients studied are likely to be representative of the Australian population, although there are no data currently available from other States or Territories. We selected for review patients who had no evidence of recurrent or residual disease to avoid any effect treatment failure may have had on HRQOL. No evidence of disease was defined as a PSA level less than 0.1 µg/L and no abnormal signs and symptoms at the last clinic visit. Men operated on less than a year before the survey were excluded as complications may still resolve during this period.9 As complications are likely to remain stable after one year, we believe the cohort to be homogeneous for the purposes of studying the effects of radical prostatectomy on HRQOL. To minimise recall bias, we included only patients operated on less than six years earlier. Non-surgical factors, such as co-morbidity and ageing, were thought to be unlikely modifiers of HRQOL during this period, so that most of the changes seen could be attributed to the prostatectomy. We decided that post-hoc subgroup analysis of a cross-sectional study with small numbers was of limited value so we included patients in the survey regardless of whether nerve sparing (which may affect postoperative potency) was contemplated or performed. Comparisons with non-surgical therapies may be facilitated by this approach. Definitions We defined urinary incontinence as the need to wear incontinence pads regularly, and defined erectile potency as the ability to achieve an erection firm enough for sex more than once a month. Questionnaire As higher complication rates are usually reported in studies in which patients are reviewed independently of their treating physician,10 our questionnaire was administered independently (by R J B). Each patient was telephoned before the questionnaire and a letter of explanation was mailed. They were assured of confidentiality, that the questionnaire was being administered independently of their treating doctor and that their answers would have no impact on management of their condition. Patients who had not replied within one month were sent one reminder. As there is no current internationally validated HRQOL questionnaire for patients with prostate cancer,11 two of us (D L N and R J B) developed the questionnaire. The initial questions collected demographic data; there were 16 items about bladder and sexual function, with similar questions to check for internal consistency; and another five items assessed postoperative therapies (such as penile injections and prostheses) and satisfaction with treatment. As our questionnaire is a new tool, only some sections have been validated.10,12-14 The questionnaire assessed men's perceptions of their urinary, sexual and overall function during the month before receipt of the questionnaire. (This is a recognised method used in other validated scoring systems.15,16) Urinary symptoms and erectile function, at the time of the survey and before surgery, were assessed on a five-point scale and included severity, effect and bothersomeness of symptoms. Satisfaction with treatment and willingness to have the same treatment again were also assessed on a five-point scale. An edited version of the Functional Assessment of Cancer Therapy Scale was used to assess health, social life and satisfaction with life.14 Statistical analysis We calculated exact confidence intervals (CI) and Fisher's exact tests using the STATA statistics package.17 Results Respondents Of 185 men having radical prostatectomy, 140 had no evidence of disease at last review. One of these 140 died in a motor vehicle accident and three more were lost to follow-up. Completed questionnaires were received from 112 (80%) men whose ages at the time of survey were normally distributed around the mean of 64 years (range, 54-73). Urinary incontinence There were 22 (20%; 95% CI, 12.7%-28.2%) respondents with postoperative urinary incontinence (Box). Those with more severe incontinence were more likely to report urge incontinence or mixed stress and urge incontinence. Twenty-five respondents (22%; 95% CI, 15%-31.2%) recalled "trouble with bladder control" before surgery, although none required pads. Seventeen of the 25 (68%; 95% CI, 46.5%-85%) were incontinent after the operation, compared with only 5 of the 87 (6%; 95% CI, 2%-13%) who did not recall having problems before surgery. That is, men with urinary incontinence after radical prostatectomy were much more likely to recall preoperative urinary symptoms. There was no statistically significant association between age and postoperative continence, nor was there any association with level of education or area of residence (Fisher's exact test). Urinary symptoms did not interfere with daily activities in 93 (83%; 95% CI, 75%-85%) respondents, and 89 (79%; 95% CI, 71%-86%) were either very happy or happy with their present bladder function. Erectile potency Of the 112 respondents, 99 (88%; 95% CI, 81%-94%) recalled preoperative erectile potency, but only 14 (12%; 95% CI, 7%-20%) described erectile potency at survey. Although only 14 were potent, 23 (20.5%; 95% CI, 13.5%-29%) were happy and 61 (54.5%; 95% CI, 45%-64%) were coping with their level of sexual function. Since surgery, 28 of the respondents (25%) have tried penile injections and three (2.7%) now have a penile prosthesis. Of the 13 men who were impotent before surgery, two have tried penile injections and one has had a penile prosthesis. There was no statistically significant association between postoperative impotence and age, level of education or place of residence (Fisher's exact test). Health-related quality of life Most respondents enjoyed a high HRQOL -- 104 (93%) were satisfied with their life and with their social life. Nearly all respondents reported good general health. The most common problem affecting their lives was impotence (44 men; 40%), followed by "concern about cancer" (13 men; 12%) and "bladder problems" (9 men; 8%). Despite the high prevalence of impotence, 104 (93%; 95% CI, 80.5%-97%) respondents were satisfied with their treatment and 98 (88%; 95% CI, 80%-93%) would opt for the same treatment again. Impotence was the most common reason given for treatment dissatisfaction (7/8) and reluctance to have the same treatment again (8/14). A final section asked about "any other problems related to your surgery which affects your quality of life". One man had a problem with a lack of pad-disposal facilities in golf club toilets, and one, although potent, said that loss of ejaculation left him unsatisfied. Discussion Our questionnaire was designed to examine specific problems of incontinence, impotence and patients' perception and satisfaction with treatment and feeling of well-being after radical prostatectomy. We found impotence to be the most common cause of diminished HRQOL, followed by "concern about cancer" and then incontinence. The preoperative urinary difficulty and impotence rates in our study are similar to those in other published prospective series,18,19 and our postoperative results lie within the range of the results of other published studies,10 suggesting that our questionnaire and study design are valid. Differences between our results and those of others may be explained by our case selection and independent data collection. In Australia, initial presentation of men with prostate problems during the study period was usually prompted by lower urinary tract symptoms. Men with troublesome urinary symptoms are more likely to have detrusor instability, a factor that commonly predisposes them to incontinence after radical prostatectomy.20 Our respondents reported a 20% incontinence rate and 22% recalled preoperative urinary difficulties. In contrast, Steiner et al reported an 8% incontinence rate after radical prostatectomy,21 but the more widespread screening for prostate cancer in the United States2-5 makes it likely that their patients were referred after screening and may not have had incontinence problems. The methods by which data are collected may also affect reporting of incontinence. Other studies that also used independent data collection15,22 have reported higher rates of incontinence (31% and 47%, respectively). We found that men with postoperative incontinence were much more likely to recall preoperative symptoms. However, we advise caution in interpreting this apparent strong association because of the limitations of cross-sectional studies (such as only measuring subjects' status once and not taking account of variation in patients' condition, including only those with successful treatments, and not yielding true relative risks) and the potential effects of recall bias. Patients with incontinence after surgery may have thought more about their predicament and been more likely to recall preoperative urinary symptoms; and, conversely, those continent after the operation may have had urinary symptoms before the operation but did not recall being troubled by them. Nevertheless, specific enquiry regarding urinary symptoms is advised when discussing treatment options with patients with localised prostate cancer, and those with urinary symptoms need to be informed of the association with urinary incontinence after radical prostatectomy. Lack of preoperative symptoms, however, does not guarantee postoperative continence as 6% of this group were incontinent after the operation. The preoperative impotence rate in our patient group is similar to that reported by Jonler et al. Men in their study were of similar age, and data were collected prospectively in a community setting.19 After the operation 12% of our respondents were potent, which is similar to the rates Jonler et al22 and Fowler et al10 reported (16% and 11%, respectively), but lower than the 70% found by Quinlan et al.23 Recall bias, case selection and independent data collection, as discussed for incontinence, are also plausible explanations for these differences. Impotence was much more likely to be reported as a major quality-of-life problem than incontinence, which is consistent with other series.12,24 Loss of potency as a cause of diminished HRQOL is not specific to radical prostatectomy, as Jonler et al concluded in their study of 1680 men attending a cancer screening program "impotent men have a lower QOL than potent men".19 Many men are prepared to trade off survival for sexual potency,25 so some men may choose a treatment with possibly lower long-term survival to increase their chance of remaining potent. The impact of radical prostatectomy and other treatments on potency should be discussed in detail when counselling patients with localised prostate cancer before therapy. However, erectile potency and a happy sex life do not go hand-in-hand, as 75% of respondents were happy or coping with their sexual function but only 12% claimed postoperative potency. This may be because people who have made a treatment decision are likely to believe, and want others to believe, that they have made the right choice,24 especially if they are disease free, as this group were. References Coates M, McCredie M, Armstrong BK. Cancer in New South Wales. Incidence and mortality, 1993. Sydney: NSW Cancer Council; 1996. McCredie M, Coates M, Churches T, Rogers J. The rising incidence of prostate cancer in Australia -- a result of "screening"? J Epidemiol Biostat 1996; 1: 99-105. McCaul KA, Luke CG, Roger DM. Trends in prostate cancer incidence and mortality rates in South Australia, 1977-1993. Med J Aust 1995; 162: 520-522. Hirst GHL, Ward JE, Del Mar CB. Screening for prostate cancer: the case against. Med J Aust 1996; 164: 285-288. Kaye KW. Prostate cancer: enthusiasm for screening. Med J Aust 1995; 162: 540-541. Whitmore WF Jr. Management of clinically localized prostatic cancer -- an unresolved problem [editorial]. JAMA 1993; 269: 2676-2677. Chodak GW, Thisted RA, Gerber GS, et al. Results of conservative management of clinically localized prostate cancer. N Engl J Med 1994; 330: 242-248. Freedman G, Hanlon M, Lee W, Hanks G. Young patients with prostate cancer have an outcome justifying their treatment with external beam radiation. Int J Radiat Oncol Biol Phys 1996; 35: 243-250. Madsen F, Bruskewitz R. Functional results of radical prostatectomy. Curr Opin Urol 1995; 5: 246-248. Fowler JF Jr, Barry MJ, Lu-Yao G, et al. Patient-reported complications and follow-up treatment after radical prostatectomy. The National Medicare Experience: 1988-1990 (updated June 1993). Urology 1993; 42: 622-629. Borghede G, Karlsson J, Sullivan M. Quality of life in patients with prostate cancer: results from a Swedish population study. J Urol 1997; 158: 1477-1486. Brickman AL, Soloway MS. Quality of life 12 months after radical prostatectomy. Br J Urol 1995; 75: 48-53. Herr HW. Quality of life of incontinent men after radical prostatectomy. J Urol 1994; 151: 652-654. Cella DF, Tulsky DS, Gray G, et al. The Functional Assessment of Cancer Therapy Scale: development and validation of the general measure. J Clin Oncol 1993; 11: 570-589. O'Leary MP, Barry MJ, Fowler FJ Jr. Hard measures of subjective outcomes: validating symptom indexes in urology. J Urol 1992; 148: 1546-1548. Barry MJ, Fowler FJ Jr, O'Leary MP, et al. Correlation of the American Urological Association symptom index with self-administered versions of the Madsen-Iversen, Boyarsky and Maine medical assessment program symptom indexes. J Urol 1992; 148: 1558-1563. STATA statistics package [computer program]. Version 5.0. Texas: Stata Corp; 1996. Diokno A, Brock BM, Brown M, Herzog A. Prevalence of urinary incontinence and other urological symptoms in the non-institutionalised elderly. J Urol 1986; 136: 1022-1025. Jonler M, Moon T, Brannan W, et al. The effect of age, ethnicity and geographical location on impotence and quality of life. Br J Urol 1995; 75: 651-655. Goluboff E, Chang D, Olsson C, Kaplan S. Urodynamics and the etiology of post prostatectomy urinary incontinence: the initial Colombia experience. J Urol 1995; 153: 1034-1037. Steiner MS, Morton RA, Walsh PC. Impact of radical prostatectomy on urinary continence. J Urol 1991; 145; 512-515. Jonler M, Messing EM, Rhodes RR, Bruskewitz RC. Sequelae of radical prostatectomy. Br J Urol 1994; 74: 352-358. Quinlan DM, Epstein JI, Carter BS, Walsh PC. Sexual function following radical prostatectomy: influence of preservation of neurovascular bundles. J Urol 1991; 145: 998-1002. Litwin MS, Hays RD, Fink A, et al. Quality of life outcomes in men treated for localized prostate cancer. JAMA 1995; 273: 129-135. Singer PA, Tasch E, Stocking C, et al. Sex or survival: trade-offs between quality and quantity of life. J Clin Oncol 1991; 9: 328-334. (Received 7 Apr 1997, accepted 24 Feb 1998) Authors' details Princess Alexandra Hospital, Brisbane, QLD. Peter S Heathcote, FRACS, Urologist; Peter N Mactaggart, FRACS, Urologist; Robyn J Boston, MB BS, Urology Registrar; Leslie C Thompson, FRACS, Urologist; David L Nicol, FRACS, Urologist. Royal Brisbane Hospital, Brisbane, QLD. Anthony N James, FRACS, Urologist. Reprints will not be available from the authors. Correspondence: Dr D L Nicol, Department of Urology, Princess Alexandra Hospital, Ipswich Road, Brisbane, QLD 4102. E-mail: D. NicolATmailbox.uq.edu.au Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia ".

Peter S Heathcote · Peter N Mactaggart · Robyn J Boston · Anthony N James · Leslie C Thompson · David L Nicol

Abstract Objective: To ascertain the effectiveness of a home vaccination service for children behind in their vaccination schedule. Design: Randomised controlled trial of nurse-administered vaccination at home. Children were allocated at random to the intervention or the control group before any contact with the parents was made. Setting: 10 council areas in north-west metropolitan Melbourne defined by 56 postcode zones. Six-week intervention period from November 1996. Participants: 405 children -- all those in the study area (n = 2610) 90 days late (age 9 months) for their third diphtheria-tetanus-pertussis/poliomyelitis/Haemophilus influenzae type B (DTP/OPV/Hib) vaccination, or 120 days late (age 16 months) for their measles-mumps-rubella (MMR) vaccination, according to the Australian Childhood Immunisation Register. Main outcome measures: Number of children completing DTP/OPV/Hib or MMR during the intervention period, and number up to date before intervention. Results: Verification of vaccination status with the parents revealed that 123 (60%) of the children in the intervention group and 113 (56%) of those in the control group were up to date with their vaccinations, leaving a study population of 81 (intervention group) and 88 (control group). Vaccination was achieved in 46 (57%) intervention children and 24 (27%) control children (risk ratio [RR], 2.08; 95% CI, 1.4-3.1; P < 0.001). For DTP/OPV/Hib, 18/32 (56%) intervention children and 12/36 (33%) control children were vaccinated (P = 0.06). For MMR, 28/49 (57%) and 12/52 (23%) children were vaccinated, respectively (P < 0.001). Home vaccinations were completed with 26 families (including five siblings). The average cost per child vaccinated as a result of the home program was $92.52. Conclusion: Home vaccination for children behind in their immunisation schedule is an effective, acceptable and relatively cheap method of completing recommended vaccinations. We recommend that a home vaccination program be widely implemented and made available, particularly for disadvantaged families. Introduction In Australia, incomplete immunisation of children under 2 years of age remains an important public health problem. Many children never complete their immunisation schedule or are many months overdue.1 Uptake rates are lowest for measles-mumps-rubella vaccination at age 12 months and the diphtheria-tetanus-pertussis booster at 18 months.1-3 Being late for a primary course is predictive of being late for or missing subsequent vaccinations.3,4Parents' beliefs about the seriousness of the vaccine-preventable illnesses and the safety and efficacy of vaccines are important predictors of vaccination uptake.5 However, there are other barriers to children being vaccinated -- frequent minor childhood illnesses, parental forgetfulness, and advice from health providers to delay vaccinations.6-10 A current strategy to overcome barriers to vaccination is to make vaccination more accessible, but parents are still required to bring their children to be vaccinated. For some families, however, it may be more effective to take the vaccination service to the child. We have explored (i) the effectiveness of offering a home vaccination service to children at greatest risk of not completing their immunisation schedule by age 2 years; and (ii) the usefulness of the Australian Childhood Immunisation Register (ACIR) for identifying these at-risk children. Method Participants The Australian Childhood Immunisation Register provided identifying information for children living in 10 local council areas in north-west metropolitan Melbourne (defined by 56 postcode zones) who were either born January 1996 and 90 days late for their third diphtheria-tetanus-pertussis/poliomyelitis/Haemophilus influenzae type B vaccination (DTP/OPV/Hib; 1st milestone), or born June 1995 and 120 days late for their measles-mumps-rubella vaccination (MMR; 2nd milestone). The intervention period comprised six weeks from November 1996. Making contact to verify vaccination status before randomisation would have in itself constituted an intervention. Therefore, before any contact was made with the parents, the children were allocated at random (by computer) to the intervention or the control group. Contact We made initial contact with the intervention group by letter, then by telephone one to three weeks later to verify vaccination status, to organise an appointment, and to administer a pre-vaccination health check as recommended by the National Health and Medical Research Council's Australian immunisation handbook.11 This health check was to ensure that the child did not require special medical attention and could be vaccinated at home. If no telephone contact could be made, two follow-up letters were sent. As local councils and maternal and child health nurses provide a substantial number of childhood vaccinations in Victoria and maintain vaccination records, we checked these (possible) providers for vaccination details if parents could not be contacted. Children were confirmed as either overdue for vaccination or up to date with vaccination if parents, the local council or the maternal and child health nurse provided a record of the vaccination. Intervention The study was approved by the Royal Children's Hospital Ethics in Human Research Committee. Parents signed a consent form to participate in the study and a standard State Government vaccination consent form. A nurse administered vaccination in the child's home at a time convenient to the parents. Siblings were also vaccinated if they were due for vaccination. The nurse providing the vaccination had completed a standard Victorian Government Department of Human Services immunisation course. A resuscitation kit (including adrenalin) was taken on each home visit, and the cold chain was maintained by transporting vaccines in a temperature-monitored car refrigerator. Before vaccination, the nurse administered a pre-vaccination health checklist11 to confirm the child's medical history, as obtained during the initial telephone contact, and to assess the child's health on the day of vaccination. Vaccines that were due were verified from the parent-held Child Health Record. The child's temperature was taken if he or she was hot or appeared unwell (a temperature > 38.58c precludes vaccination11). Paracetamol was offered to all children before vaccination. The nurse remained with the family for more than 20 minutes after vaccination. The visits included time for parents to complete questionnaires about immunisation service use, reasons for the delay in vaccination, education level, family size and whether the family had a Health Care Card. (A Health Care Card is a Federal Government card available to low income families, including those receiving government pensions, to obtain concessions for health and medical expenses; ie, it is an indicator of disadvantage or poverty.) Neither written consent nor sociodemographic information was obtained from parents who chose to have their child vaccinated by another provider, or whose child was up to date with immunisation, or who refused to take advantage of the home service. Follow-up of control children Two months after the intervention period, and based on updated information from the Australian Childhood Immunisation Register, we sent letters to parents of control children for whom neither the Register nor local councils had recorded a third DTP/OPV/Hib or an MMR vaccination. We followed the letters with a telephone call to verify vaccination status and to offer, in this case, vaccination at the Royal Children's Hospital. Parents of control children were also informed of local vaccination services offered by the maternal and child health nurse or of the schedules of mobile vaccination vans provided by local councils. No sociodemographic information was collected from the parents of control children. Cost analysis Costs included travel, estimated at $0.50 per kilometre, nursing time at $25 per hour, consumables (excluding vaccines) as charged by the Royal Children's Hospital and clerical work at $17 per hour for 18 days. Statistical analysis Sample size was estimated assuming that 35% of intervention children would accept vaccination and 6% of control children would be immunised. This would require 30 in each group, with a set at 0.05 and statistical power 80%. Statistical associations were assessed with chi-squared tests. Confidence intervals and risk ratios were calculated with the STATA program.12 Results Subjects There were 2610 children born in June 1995 or January 1996 in the study area and registered with the Australian Childhood Immunisation Register. Of these, 416 children (16%) were identified by the Register as overdue for their third DTP/OPV/Hib or MMR. The Figure shows the number of children on the Register meeting the study criteria, the number excluded and the vaccination status of the children in the intervention and control groups at the time of contact. On verification of vaccination status with parents, 123 (60%) of the intervention children and 113 (56%) of the control children were confirmed as being up to date with their immunisation schedule, and therefore were ineligible for the intervention, leaving 81 children in the intervention group and 88 control children. Those whose status could not be verified were assumed for analysis to be unvaccinated. In total, 2430 (93%) children were up to date with their vaccinations at the beginning of the study period: 1219 (95%; 95% confidence interval [CI], 93.6%-96.0%) 9-month-old children and 1211 (92%; 95% CI; 90.8%- 93.8%) 16-month-old children. Intervention Table 1 shows the number of children vaccinated during the intervention period. To estimate the effect of the intervention on uptake for the full cohort, the cohort was divided into two equal groups (n = 1305 each). The number of children immunised in the group with the intervention children increased from 1220 (93.5%) before intervention to 1266 (97%) after intervention. The group with the control children increased from 1210 (93%) to 1234 (95%). Using similar logic, but dividing for type of vaccine, in the group with intervention children the rate for 1st milestone vaccination increased to 98% and for 2nd milestone to 97%, whereas the rates for the group with control children increased to 96% and 94%, respectively. The mean (SD) age for DTP/ OPV/Hib vaccination for intervention and control children was 10 (0.2) months and 11.5 (0.3) months, respectively (which was significantly different; P < 0.001), compared with 7 (1.3) months for children having DTP vaccination before study commencement. The mean (SD) age for MMR vaccination for intervention and control children was 17.2 (0.1) months and 19 (0.3) months (P < 0.001), compared with 14 (1.7) months for those having MMR vaccination before study commencement. In the intervention group 26 children were immunised by the study nurse and 19 by their doctor or local council service. One child who had a severe egg allergy was vaccinated at the Royal Children's Hospital Immunisation Adverse Events Clinic. Ten children due for MMR were also given their 18-month DTP/Hib boosters and five siblings were brought up to date with their vaccination schedule. In all, 82 vaccines were administered to study children and siblings. On the day of vaccination 13 children had colds or were taking antibiotics; none had a fever. All were vaccinated as arranged. Two families refused the service because they were against immunisation and 22 families preferred to use their own doctor. As mentioned, 19 did so within the study period. One mother changed her mind about home vaccination because of concern about her child's egg intolerance. The child was vaccinated a month after the intervention by her doctor. Table 2 summarises the demographic information of those immunised at home and major reasons given by parents for delayed vaccinations. Costs The mean cost per child vaccinated in the intervention group was $92.52, and the mean cost per visit per vaccine was $52. These costs excluded cost of visits to the general practitioner by those being vaccinated by their own doctor. Fifty-one per cent of the cost was attributable to clerical time needed to verify vaccination status. Travel costs were 12% of total costs and 33% of nurses' costs. Discussion We have shown that offering home vaccination is an effective method of bringing children (and their siblings) up to date with their immunisation schedule. Importantly, we used information from a population-based register, and thus provided vaccinations for children in socially disadvantaged families. Such families have been identified previously by the Australian Bureau of Statistics13 and others14 as being most at risk of not completing the scheduled childhood immunisations. A major finding of this study was the unexpectedly high proportion of children already vaccinated at the commencement of the study. This proportion was substantially higher than expected from previous statewide estimates -- 95% v. 84% for 1st milestone vaccination and 92% v. 78% for 2nd milestone vaccination2 -- and exceptionally high given that our study was conducted in a socially disadvantaged area. Data for our study on children's vaccination status were from Australian Childhood Immunisation Register enrolments, which are derived from Medicare data and miss about 2% of children; however, this would have had a minimal impact on these vaccination estimates. It is also unlikely that substantial misclassification of vaccination status occurred. While we did not formally cross-check vaccination dates, when dates were obtained from two sources 85% matched. When dates differed it cannot be assumed that Register dates were correct. In some cases vaccinations reported to the Register by us were incorrectly recorded or missing. Thus, we found the usefulness of the Australian Childhood Immunisation Register as a source of accurate information to be limited. However, our study was undertaken in the first year of the Register's existence and it is expected that accuracy of the Register will improve. Limitations of this study arise from the need to randomise the population sample before verification of immunisation status. This has the potential to introduce bias because of the possibility of differences in response between the intervention and control groups. Another limitation was the number of children in each group with whom no contact was made. However, these limitations are unlikely to have caused substantial bias. In both groups a similar number of control (56%) and intervention children (60%) were excluded because they were up to date with vaccinations, and likewise the proportion of control (15%) and intervention children (14%) who could not be contacted to determine immunisation status was similar. There is also no reason to suppose that any differences in vaccination rates between these two uncontacted groups would be sufficient to bias the estimate of the intervention effect. Assuming that 50% of the children in each group who could not be contacted were vaccinated, the risk ratio for vaccination would be 1.67 (95% CI, 1.3-2.2; P < 0.001). Itinerancy is a risk factor for incomplete and late vaccination,3,15 making it not surprising that a considerable number considered overdue for vaccination could not be contacted. We have shown that those who can be contacted can be vaccinated. A similar program that accesses vaccination information at a local level may be more efficient at targeting families who move frequently. It is obvious that a home service will cost more than mass vaccination programs. The cost per vaccine, taking into account only nurses' time and travel costs, was about $23, which compares favourably with the £8 reported by an outreach program in the United Kingdom.16 The costs of the service would be reduced with improved accuracy of Australian Childhood Immunisation Register information (clerical costs would be reduced by 50%), by offering a local rather than a centralised service (travelling costs would be reduced by 30%), and by incorporating the vaccination service into a broader home visiting program to promote child health and support disadvantaged families in this endeavour. As indicated by many studies,17-22 a barrier to age-appropriate immunisation is often not parental unwillingness to have their child vaccinated, but immunisation providers failing to provide a service. About a third of the parents of children behind in their vaccinations reported having recently consulted a doctor. In almost all these cases the child could have been vaccinated at that time. To prevent diseases such as measles, immunisation rates need to exceed 95%.23 Even the high uptake rates found at the commencement of our study are below this level. Innovative and proactive methods are necessary to attain these high levels and have been found to be effective.15,24 Acknowledgements This study was funded by National Health and Medical Research Council (NHMRC) project grant number HS371. Lyndal Bond was funded by an NHMRC Scholarship. We would like to thank the research nurse (Michelle Wills), the Royal Children's Hospital Immunisation Adverse Events Clinic for providing a service for children in the intervention and control groups, and local council health departments for their cooperation. References Australian Bureau of Statistics. Children's immunisation Australia. Canberra: ABS, April 1995, 1996. (Catalogue No. 4352.0.) Lester R, Norris P. 1994/5 Pre-school immunisation coverage. Notifications for Victoria, July-September 1995. Health Protection Update (Public Health Division, Department of Human Services, VIC) 1995; 3: 10-13. Hanna CJ, Wakefield JE, Doolan CJ, Messner JL. Childhood immunisation factors associated with failure to complete the recommended schedule by two years of age. Aust J Public Health 1994; 18: 15-24. Li J, Taylor B. Factors affecting uptake of measles, mumps, and rubella immunisation. BMJ 1993; 307: 168-171. Peckham C, Bedford H, Senturia J, Ades A. The Peckham Report National Immunisation Study: factors influencing immunisation uptake in childhood. Horsham: Action Research For The Crippled Children, 1989. Jones K, Fasher B, Hanson R, et al. Immunisation status of casualty attenders: risk factors for non- compliance and attitudes to "on the spot" immunisation. J Paediatr Child Health 1992; 28: 451-454. Miles TA, Merrell WH. An outbreak of measles in the Hunter area of New South Wales. Aust J Public Health 1992; 16: 302-304. New SJ, Senior ML. "I don't believe in needles": qualitative aspects of a study into the uptake of infant immunisations in two English health authorities. Soc Sci Med 1991; 33: 509-518. Harding GC. How can the uptake of vaccines be increased? J Inst Health Educ 1984; 22: 5-11. Adjaye N. Measles immunisation: some factors affecting non-acceptance of vaccine. Public Health 1981; 95: 185-188. National Health and Medical Research Council. The Australian immunisation handbook. 6th ed. Canberra: AGPS, 1997. Stata Statistical Software [computer program], Release 5.0. College Station, Tex: Stata Corporation, 1997. Australian Bureau of Statistics. 1989-90 National Health Survey children's immunisation, Australia. Canberra: AGPS, 1992. Bazeley P, Kemp L. Childhood immunisation: the role of parents and service providers. A review of the literature. Canberra: National Immunisation Programme, Commonwealth Department of Human Services and Health, AGPS, 1994. Pearson M, Makowiecka K, Gregg J, et al. Primary immunisations in Liverpool II: is there a gap between consent and completion? Arch Dis Child 1993; 69: 115-119. Jefferson N, Sleight G, MacFarlane A. Immunisation of children by a nurse without a doctor present. BMJ 1987; 294: 423-424. MacIntyre R, Nolan T. Attitudes of Victorian vaccine providers to pertussis vaccine. Med J Aust 1994; 161: 295-299. Burgess MA. Pertussis vaccine -- time to stop the confusion. Med J Aust 1994; 161: 293-294. Begg NT, Nicholl A. Immunisation. BMJ 1994; 309: 1073-1075. Stevens D, Baker R, Hands S. Failure to immunise against whooping cough. Arch Dis Child 1986; 61: 382-387. Kinder J, Teare L, Rao M, et al. False contraindications to childhood immunisation. Br J Gen Pract 1992; 42: 160-161. Askew GL, Finelli L, DeGraaf J, et al. Beliefs and practices regarding childhood vaccination among urban pediatric providers in New Jersey. Pediatrics 1995; 96: 889-892. Nolan T. Measles -- eradication or procrastination? Med J Aust 1990; 152: 449-450. Birkhead GS, LeBaron CW, Parsons P, et al. The immunisation of children enrolled in the special supplemental food program for women, infants and children (WIC). JAMA 1995; 274: 312-316. (Received 29 Sep 1997, accepted 6 Apr 1998) Authors' details Clinical Epidemiology and Biostatistics Unit, Department of Paediatrics, University of Melbourne, Royal Children's Hospital, Melbourne, VIC. Lyndal M Bond, MA(ApplPsych), NHMRC Scholar. Terry M Nolan, PhD, FRACP, FAFPHM, Head, Clinical Epidemiology and Biostatistics Unit. Department of Human Services, Melbourne, VIC. Rosemary A Lester, MB BS, MPH, FAFPHM, Head, Infectious Diseases Unit. Reprints: Ms Lyndal Bond, Clinical Epidemiology and Biostatistics Unit, Department of Paediatrics, University of Melbourne, Royal Children's Hospital, Parkville, VIC 3052. E-mail: bondATcryptic.rch.unimelb.edu.au Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia ".

Lyndal M Bond · Terry M Nolan · Rosemary A Lester

Asthma and other atopic diseases in Australian children Australian arm of the International Study of Asthma and Allergy in Childhood Colin F Robertson, Marita F Dalton, Jennifer K Peat, Michelle M Haby, Adrian Bauman, J Declan Kennedy and Louis I Landau MJA 1998; 168: 434-438 Abstract - Introduction - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - ©MJA1998 Abstract Objective: To determine the prevalence of asthma, eczema and allergic rhinitis in Australian schoolchildren using the protocol of the International Study of Asthma and Allergy in Childhood (ISAAC). Design: Questionnaire-based survey. Setting: Melbourne, Sydney, Adelaide (in winter-spring, 1993) and Perth (in winter-spring, 1994). Subjects: All children in school years 1 and 2 (ages 6-7 years) or in year 8 (ages 13-14 years), attending a random sample of 272 schools, stratified by age and city. Main outcome measures: Parent-reported (for 6-7 year olds) or self-reported (for 13-14 year olds) symptoms of atopic disease in the previous 12 months, or ever; treatment of asthma; and country of birth. Results: 10 914 questionnaires were completed for 6-7 year olds and 12 280 for 13-14 year olds (84% and 94% response rates, respectively). Prevalence of wheeze in the past 12 months was 24.6% for the 6-7 year olds and 29.4% for the 13-14 year olds, and, among 6-7 year olds, was significantly higher in boys (27.4%) than girls (21.7%). Children born in Australia were more likely to report current wheeze than those born elsewhere (6-7 year olds: odds ratio [OR], 1.82; 95% confidence interval [CI], 1.55-2.15; and 13-14 year olds: OR, 1.88; 95% CI, 1.68-2.11). Prevalences of current eczema and allergic rhinitis were 10.9% and 12.0%, respectively, for the 6-7 year olds, and 9.7% and 19.6%, respectively, for the 13-14 year olds. Asthma, eczema and rhinitis coexisted in 1.8% of 6-7 year olds and 2.8% of 13-14 year olds. Conclusion: This study provides evidence that asthma prevalence in Australian schoolchildren is continuing to increase and is higher among Australian-born children than among those born elsewhere. Asthma, eczema and allergic rhinitis coexist to a lesser extent than expected. These results form the basis for future Australian and international comparisons. Introduction There is now substantial evidence that the prevalence of asthma and other atopic disorders is increasing worldwide.1,2 While the prevalence of asthma has been documented in the past 30 years, variation in methods and lack of uniform diagnostic criteria make direct comparison between studies difficult. Little is known about the prevalence of the other atopic disorders -- eczema and allergic rhinitis -- both throughout the world and particularly in Australia. The International Study of Asthma and Allergy in Childhood (ISAAC) is a collaborative project which has developed a standardised methodology to describe the prevalence and severity of asthma, rhinitis and eczema in children throughout the world.3 Such data will provide a framework for aetiological research into lifestyle, environmental and genetic factors affecting these disorders. Phase 1 of ISAAC is to determine the prevalence of the disorders throughout the world. Phases 2 and 3 will be more comprehensive, using more detailed questionnaires and objective measures to confirm the differences seen in Phase 1 and to identify important aetiological factors. Our study was part of Phase 1 of ISAAC. It aimed to determine the prevalence of asthma and other atopic diseases in Australian schoolchildren, to determine the burden of atopic disease in this country, and to provide a basis for international comparison. Methods We used the protocol of ISAAC3 to survey two age groups: 6-7 year olds (school years 1 and 2) and 13-14 year olds (school year 8). Subjects were all children in the relevant years of a random sample of primary and secondary schools. The sample comprised about 10% of all government, Catholic and independent schools in the metropolitan areas of Adelaide and Perth; the area within a radius of 20 km from the GPO in Melbourne; the area within a radius of 10 km from the GPO in Sydney for primary schools (school years 1 and 2); and the Western Region of Sydney for secondary schools (school year 8). Previous studies have shown these areas of Sydney and Melbourne to be representative of the metropolitan areas of these cities.4,5 A five-page questionnaire was issued by teachers for completion by parents of the 6-7 year olds, and by the 13-14 year olds in the classroom under examination conditions. The questionnaires contained the three standard ISAAC modules, asking about symptoms of asthma, eczema and allergic rhinitis3 (see Box 1 for definitions), an additional module about treatment of asthma, and two extra questions about the children's and mothers' country of birth. No translations were provided. If the first questionnaire was not returned by the 6-7 year olds, a second was issued. A second visit was made to the secondary schools, if necessary, to recruit students absent at the initial visit. Data were analysed with the statistical package SPSS-X.9 Results were adjusted for cluster effect, and chi-squared tests were used to compare prevalences, while significance of odds ratios (OR) was assessed with 95% confidence intervals (CIs). Results Details of schools and subjects surveyed are shown in Box 2; 201 primary schools and 71 secondary schools participated, comprising 7%-42% of schools in the sampling area; 9% of schools selected declined to participate. A total of 12 952 questionnaires were issued to the 6-7 years age group (response rate, 84%) and 13 078 to the 13-14 years age group (response rate, 94%). Prevalence of atopic diseases in the two age groups is shown in Box 3. Asthma Prevalence of current wheeze was 24.6% for the 6-7 year olds (95% CI, 23.8-25.4), and 29.4% for the 13-14 year olds (95% CI, 29.1-29.7) (Box 3). In the younger group, current wheeze was significantly more common in boys than in girls (OR, 1.36; 95% CI, 1.25-1.49), but this sex difference was reversed in the older group (OR 0.82; 95% CI, 0.76-0.89). Figure 1 (below) compares the prevalence of atopic diseases between the four cities. For the 6-7 year olds, there was no significant difference in prevalence of current wheeze between cities, but for the 13-14 year olds prevalence was slightly higher in the western cities (Adelaide and Perth: 32.3%) than in the eastern cities (Sydney and Melbourne: 25.9%) (OR, 1.37; 95% CI, 1.26-1.48). There was a similar difference between west and east in percentage of 13-14 year olds who had had more than 12 episodes of wheeze per year (4.1% versus 3.1%) and who had attended the emergency department (3.5% versus 2.9%) (data not shown). The prevalence of current wheeze was generally higher in the older age group. The spectrum of asthma among children who reported current wheeze is shown in Box 4. While most children in both age groups reported only one to three asthma episodes in the previous 12 months, 8.0% of 6-7 year olds and 12.2% of 13-14 year olds reported more than 12 episodes. Sleep disturbance due to asthma was common, with 11.2% of 6-7 year olds and 9.8% of 13-14 year olds reporting sleep disturbance on one or more nights per week. About 7% of both age groups reported a hospital admission for asthma in the previous 12 months. Patterns of asthma treatment are shown in Box 5. Regular b2-agonists were taken as sole therapy by 5.5% of 6-7 year olds and 7.4% of 13-14 year olds with current wheeze, while regular inhaled steroids were taken by 21.1% of 6-7 year olds and 14.6% of 13-14 year olds, rising to 49.7% and 36.9% for those with more than 12 episodes per year. While overall 26.5% of 6-7 year olds with current wheeze and 15.8% of 13-14 year olds had a written asthma management plan, this increased to 46.5% and 25.9% in those who reported 12 or more attacks in the past 12 months. Most children attended a doctor at least once during a wheezy episode throughout the year, but only 42.2% of 6-7 year olds and 31.3% of 13-14 year olds visited a doctor for a regular check-up. Eczema Prevalence of current eczema did not vary significantly between the cities (Box 3). Eczema was less common in boys than in girls in both age groups (6-7 year olds: OR, 0.81; 95% CI, 0.72-0.92; 13-14 year olds: OR, 0.57; 95% CI, 0.51-0.65). Sleep disturbance due to itching was common among those with current eczema; it was reported to occur at least weekly by 7.9% of 6-7 year olds and 13.4% of 13-14 year olds, and at a lesser frequency by 27% of 6-7 year olds and 30.4% of 13-14 year olds. Allergic rhinitis The prevalence of current allergic rhinitis was significantly higher in Adelaide and Perth than in Sydney and Melbourne (6-7 year olds: OR, 1.62; 95% CI, 1.44-1.82; 13-14 year olds: OR, 1.53; 95% CI, 1.40-1.68). Like wheeze, rhinitis was more common in boys than girls in the younger group (boys versus girls: OR, 1.19; 95% CI, 1.06-1.33), while this sex difference was reversed in the older group (boys versus girls: OR, 0.64; 95% CI, 1.40-1.68). Among those with current rhinitis, 71% of 6-7 year olds and 76% of 13-14 year olds reported that it interfered with their daily activity to some extent (troublesome rhinitis), while 18.5% of 6-7 year olds and 19.1% of 13-14 year olds described this interference as moderate to "a lot". Atopic disease and country of birth Children born in Australia were more likely to report current wheeze than those born elsewhere (6-7 year olds: OR, 1.81; 95% CI, 1.54-2.14; 13-14 year olds: OR 1.89; 95% CI, 1.69-2.12). This trend was similar for children whose mothers were born in Australia compared with those whose mothers were born elsewhere (6-7 year olds: OR, 1.29; 95% CI, 1.18-1.42; 13-14 year olds: OR, 1.58; 95% CI, 1.45-1.71). When children born outside Australia were analysed by region of birth (United Kingdom, Central Europe, South-East Asia or the Middle East), there was no difference in the prevalence of wheeze between regions. Eczema and rhinitis were also more common in children born in Australia than those born elsewhere. For eczema the OR was 1.31 (95% CI, 1.06-1.63) for 6-7 year olds and 1.36 (95% CI, 1.14-1.61) for 13-14 year olds. For rhinitis, the OR was 1.79 (95% CI, 1.42-2.26) for 6-7 year olds and 1.5 (95% CI, 1.32-1.70) for 13-14 year olds. The proportion of children born outside Australia was higher in the eastern cities among 13-14 year olds (23%) than in the western cities (15%). Similarly, the proportion of mothers born outside Australia was higher in the eastern cities (54%) than in the western cities (39%). When the odds ratio comparing prevalence of wheeze among 13-14 year olds in western versus eastern cities was adjusted for child's country of birth, it fell from 1.37 to 1.25 (95% CI, 1.15-1.36). Interrelations of atopic diseases Figure 2 (below) shows the overlap of asthma, eczema and allergic rhinitis. While 35.2% of 6-7 year olds reported having at least one of these conditions in the past 12 months, only 1.8% reported having all three. Corresponding figures for 13-14 year olds were 41% with at least one condition and 2.8% with all three. Among those with current wheeze, only 19% of 6-7 year olds and 18% of 13-14 year olds reported coexistent current eczema, with no apparent age effect in the relationship. Discussion This study describes the burden of atopic disease in Australian schoolchildren. The prevalence of current wheeze was similar to that reported in recent epidemiological studies in Australia.10 However, comparison with results of a similar questionnaire given to Melbourne schoolchildren in 1990 suggests that, although the spectrum of asthma remains unchanged, the prevalence of recent wheeze has increased from 23.1% in 1990 (95% CI, 21.7-24.5)4 to 27.2% in 1993 (95% CI, 25.6-28.8) (P < 0.01). The rate of increase (1.4% per annum) is similar to that reported in an earlier Australian study (1.24%)10 and higher than that reported in European studies (0.1%-0.4%).1 Morbidity due to asthma remains significant, with high levels of symptoms, emergency department attendances and hospital admissions. Asthma is the second most common reason for admission to a paediatric hospital bed in Victoria (after otolaryngological conditions), with a rate in children of 738 per 100 000 population in 1994-1995.11 The total annual cost to the community associated with asthma management in Australia was estimated in 1989 as $627 million, or $769 per asthmatic person.12 These costs are likely to have increased because of the increases in medication costs and asthma prevalence. There was a significant difference in the prevalence of current wheeze and current rhinitis between the eastern and western States. A possible explanation is the difference in patterns of immigration, with more children in the eastern cities born outside Australia than in the western cities. Indeed, the odds ratio comparing prevalence of wheeze among 13-14 year olds in western versus eastern cities fell from 1.37 to 1.25 after adjustment for country of birth. Internationally, ISAAC has collected data on over half a million children from 120 centres in 48 countries. Australia ranks third-highest in prevalence of current wheeze for 13-14 year olds and second-highest for 6-7 year olds.13 For "current rhinitis", Australia ranks fifth and, for eczema, eleventh. Australia's high ranking for asthma prevalence is supported by data for asthma mortality. This was not collected by ISAAC, but comparison of available data from 11 developed countries shows Australia had the highest mortality rate due to asthma in 1990.14 We found evidence from throughout Australia for continuing lack of effective treatment of asthma. Among children with more than 12 episodes of wheeze per year, only 64% of 6-7 year olds and 43% of 13-14 year olds were taking regular preventive treatment. Further, 5.5% and 7.4% of those reporting "current wheeze" used regular b -agonists in the absence of any preventive therapy, despite the cumulative evidence against the practice. Sodium cromoglycate was used by 19% of the 6-7 year olds and 11% of the 13-14 year olds who reported taking regular preventive therapy, showing some support for the Australian paediatric asthma guidelines, which recommend cromoglycate as first-line therapy for mild to moderate persistent asthma.15 We also found eczema and rhinitis to be common and to cause significant morbidity among Australian schoolchildren. Eczema was less common in boys than in girls in both age groups, a trend seen throughout the world.16 It is not life-threatening, but may cause considerable physical and psychological disability (including discomfort from itching, which may result in sleep loss and secondary infection, as well as the psychological effects of a visible skin disease). Treatment can be expensive and time consuming. Recent Australian estimates of the cost to the family were $330 to $1255 a year, depending on eczema severity.17 Additional costs to the community for consultations ranged from $209 to $642 a year for each child. Allergic rhinitis also carries significant morbidity. The effect on quality of life of perennial rhinitis has been estimated to be similar to, or worse than, mild to moderate asthma.18 In adults, hayfever is estimated to cause, on average, the loss of a third of a day from work each year, in addition to loss of productivity through symptoms or the sedating effects of some drug treatments.18 There are no precise estimates for the cost of therapy, as many sufferers do not consult a medical practitioner,8 and most treatment is available "over the counter". The higher prevalence of "current wheeze" found among 13-14 year olds compared with 6-7 year olds should be interpreted with caution, as the respondents differed between the two groups (parents for the 6-7 year olds and the children themselves for the 13-14 year olds). In an earlier study of Melbourne 7-year-olds and 15-year-olds, in which parents completed the questionnaire for both age groups, prevalence of "current wheeze" was lower among the 15-year-olds (18.6%) than among the 7-year-olds (23.1%).4 Further, comparison of adolescent and parent responses to an Australian asthma morbidity questionnaire showed that the adolescents reported a higher incidence of symptoms than their parents.19 The correlation between the three atopic diseases was less than anticipated. Atopy is usually associated with increased serum levels of IgE and positive skin reactivity to common allergens and has a strong genetic basis. The factors that determine the phenotypic expression of atopy and direct it to asthma, eczema or hayfever are unclear. This diverse expression of the genotype needs to be considered when studying the genetics of asthma. In conclusion, Australia has a high prevalence of atopic disorders, ranking among the highest in the world. Our study, part of a much larger international study, provides an opportunity to gain new insights into the causes and natural history of these disorders. Acknowledgements We would like to thank the schools, parents and children who participated, the research assistants who helped collect the data, and the State departments of education that approved the study. In Adelaide, the study was supported by Rotary, in Perth by the Asthma Foundation of Western Australia, and in Melbourne and Sydney by internal department funds. References Magnus P, Jaakkola JJK. Secular trends in the occurrence of asthma among children and young adults: critical appraisal of repeated cross sectional surveys. BMJ 1997; 314: 1795-1799. Wuthrich B. Epidemiology and natural history of atopic dermatitis. Allergy Clin Immunol Int 1996; 83: 77-82. Asher I, Kiel U, Anderson HR, et al. International study of asthma and allergies in childhood (ISAAC): rationale and methods. Eur Resp J 1995; 8: 483-491. Robertson CF, Heycock E, Bishop J, et al. Changes in prevalence of asthma in Melbourne schoolchildren over 26 years. BMJ 1991; 302: 1116-1118. Peat JK, Toelle BG, Gray EJ, et al. Prevalence and severity of childhood asthma and allergic sensitisation is seven regions of New South Wales. Med J Aust 1995; 163: 22-26. Jenkins MA, Clarke JR, Carlin JB, et al. Validation of questionnaire and bronchial hyperresponsiveness against respiratory physician assessment in the diagnosis of asthma. Int J Epidemiol 1996; 25: 609-616. Williams HC, Burney PGJ, Pembroke AC, Hay RJ. Validation of the UK diagnostic criteria for atopic dermatitis in a population setting. Br J Dermatol 1996; 135: 12-17. Sibbald B, Strachan DP. Epidemiology of rhinitis. In: Busse WW, Holgate ST, editors. Mechanisms in asthma and rhinitis: implications for diagnosis and treatment. Oxford: Blackwell Scientific Publications, 1994: 32-43. Norusis MJ. SPSS/PC+ Advanced Statistics. V5.0 [computer program]. Chicago, Ill:SPSS Inc, 1992. Peat JK, van den Berg RH, Green WF, et al. Changing prevalence of asthma in Australian children. BMJ 1994; 308: 1591-1596. Information Analysis Unit, Acute Health, Victorian Department of Human Services. Victorian inpatient mordibity database. Melbourne: Department of Human Services. Sighted Oct 1997. Toelle BG, Peat JK, Mellis CM, Woolcock AJ. The cost of childhood asthma to Australian families. Pediatr Pulmonol 1995; 19: 330-335. Beasley R, Keil U, von Mutius E, et al. Worldwide variation in the prevalence of symptoms of asthma, allergic rhinoconjunctivitis and atopic eczema: the international study of asthma and allergies in childhood (ISAAC). Lancet 1998. In press. Robertson CF, Sennhauser F, Mallol J. The change in prevalence and severity of asthma in developed and developing countries. Phelan PD (ed). Baillieres Clin Paediatr 1995; 3: 253-275. National Asthma Campaign. Asthma management handbook. 3rd edition. Melbourne: National Asthma Campaign, 1996. Williams HC, Robertson CF, Stewart AW, et al. Worldwide variation in the prevalence of symptoms of atopic eczema in the International Study of Asthma and Allergies in Childhood. J Allergy Clin Immunol 1998. In press. Su JC, Kemp AS, Varigos GA, Nolan TM. Atopic eczema: its impact on the family and financial cost. Arch Dis Child 1997; 76: 159-162. Juniper EF. Measuring health-related quality of life in rhinitis. J Allergy Clin Immunol 1997; 99: S742-S749. Bishop J, Robertson CF, Caust J, et al. Concordance between adolescent and parent response to an asthma morbidity questionnaire. Am Rev Respir Dis 1993; 147: A373. Received 30 Oct 1997, accepted 10 Mar 1998 Authors' details Department of Thoracic Medicine, Royal Children's Hospital, Melbourne. Colin F Robertson, MSc, FRACP, Deputy Director; Marita F Dalton, Assoc Dip Med Rec, Research Assistant. Department of Medicine, University of Sydney, Sydney. Jennifer K Peat, PhD, Senior Research Fellow; Michelle M Haby, MAppSc, Research Assistant. School of Community Medicine, University of New South Wales, Sydney. Adrian Bauman, PhD, FAFPHM, Associate Professor. Department of Respiratory Medicine, Women's and Children's Hospital, Adelaide. J Declan Kennedy, MD, FRCP, Physician. Department of Respiratory Medicine, Princess Margaret Hospital for Children, Perth. Louis I Landau, MD, FRACP, Professor of Paediatrics. Reprints will not be available from the authors. Correspondence: Dr C F Robertson, Department of Thoracic Medicine, Royal Children's Hospital, Flemington Road, Parkville, VIC 3054. E-mail: cfrobATcryptic.rch.unimelb.edu.au Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". <URL: http://www.mja.com.au/> © 1998 Medical Journal of Australia.

Colin F Robertson · Marita F Dalton · Jennifer K Peat · Michelle M Haby · Adrian Bauman · Louis I Landau

Severe cervical spinal cord injuries related to rugby union and league football in New South Wales, 1984-1996 Tai R Rotem, James S Lawson, Stephen F Wilson, Stella Engel, Sue B Rutkowski and Chris W Aisbett MJA 1998; 168: 379-381 For editorial comment see Yeo Correction note: This article was published online with a typographic error that was corrected on 15 June 1998: A P value of 0.06 was misprinted as 0.6. Jump to corrected par. Abstract - Introduction - Methods - Results - Discussion - Acknowledgement - References - Authors' details - - ©MJA1998 Abstract Objective: To determine the frequency and circumstances of serious cervical cord injuries associated with rugby union and league football in New South Wales. Design: Retrospective review of patients with rugby football-related cervical spinal cord injuries. Setting: The two central spinal units in NSW, from January 1984 to July 1996. Outcome measures: Admission to spinal units; injury resulting in permanent tetraplegia. Results: During the review period, 115 rugby football players (56 union and 59 league) were admitted to the spinal units because of cervical spinal cord injuries. 49 patients had resultant permanent neurological deficits (complete tetraplegia [quadriplegia]) -- 26 associated with rugby union and 23 with rugby league. Two patients died of injury sequelae within two weeks of admission. There was no significant change in the rate of football-related admissions to spinal units for either code. There was a small decline in the number (from 15 in 1984 to 1987 to 7 in 1992 to 1996) and incidence (from 1.2 to 0.5 per 10 000 participants) of patients with tetraplegia associated with rugby union. When this decline was tested as a trend over the years, it was found to be statistically significant (P = 0.06). No significant trend was found in the tetraplegia data associated with rugby league. Cervical spinal cord injuries leading to complete tetraplegia were most commonly associated with scrum-like plays in union and with tackles in league. Conclusion: Serious cervical spinal injuries associated with both codes of rugby continue to occur in NSW. Rugby football in its various forms is still an inherently dangerous game. Introduction Injury to the cervical spinal cord sustained while playing rugby union or league football has tragic personal consequences, with subsequent high demands on community resources.1 During the 1970s and early 1980s there was an approximately twofold increase in the incidence of such injuries in many countries,2-4 possibly because of the development of more powerful scrums and tackles (see Box for terms). In response, between 1984 and 1996, new rules were introduced intermittently, existing rules were more strictly enforced, and "safety" programs were offered -- these included exercises for strengthening players' neck muscles and the banning of players with long thin necks from taking dangerous positions, such as the front row of scrums. We report the findings of a survey of rugby union- and league-related cervical spinal cord injuries in New South Wales (NSW) for the period 1984 to 1996. Methods We retrospectively reviewed the medical records of all patients admitted to the two central spinal units in NSW (at the Royal North Shore Hospital and Prince Henry Hospital, Sydney) because of cervical spinal cord injuries associated with rugby union and rugby league football from January 1984 to December 1996. Patients transferred from outside NSW were excluded from the analysis. Almost all patients (more than 95%) with spinal injuries that occur in NSW are admitted to these two units. Additional information about the circumstances of the injuries, including eyewitness reports of the events leading to them, was obtained from the NSW Sporting Injuries Committee -- a statutory authority which administers an insurance scheme for people permanently injured while playing sports. We also estimated the number of participants in rugby union and league football in NSW per year in the review period, based on the Committee's data5 plus surveys of NSW schoolboy football participants.6 As some of these data were incomplete and there was some duplication, these estimates are broad approximations only. Statistical analysis The number of cervical spinal cord injury cases is small, so any statistical analysis necessarily lacks power. To develop valid conclusions from this study, we analysed cases of cervical spinal cord injuries according to: total number of admissions to the two spinal units; permanent neurological deficits (ie, complete tetraplegia); and trends during the review period. When there are large numbers of participants and the risk of cervical spinal cord injury to any particular player is very low over a short time period -- and the injury is permanent and will not allow a return to the game -- the number of injuries in each year is assumed to follow a Poisson distribution. In the absence of data on hours of playing football or training for football, we assumed that the total number of hours devoted to the game for each code and grade did not vary from year to year. The null hypothesis for each test we performed was that there was no change in aggregated risk of cervical spinal injury against the alternative that the risk had decreased. All tests were based on the null-hypothesis assumption that the best estimate of annual rate of injuries was the average rate of injuries. We then fitted a simple linear regression, of time, to a probability transformation of the injury data. The number of injuries for each year was converted to the probability of observing, at most, that number of injuries in a particular year. These values were then converted to their corresponding standard normal score. The tests were based only on the regression slopes, which were tested to see if they were negative. The critical level of significance for these tests was 0.1. All calculations were performed using Microsoft Excel. The function POISSON was used to calculate the cumulative probabilities; the function NORMINV was used to calculate the standard normal scores; the function LINEST was used to calculate the slope and its standard error; and the function TDIST (one-tailed 11 degrees of freedom) was used to obtain the significance of the results. The Random Number Generator was used in a small simulation study to confirm the size of the testing procedure. Results During the 13-year period of the review, 117 rugby football players (56 union and 59 league players, plus one "backyard" and one "touch" player) were admitted to the spinal units because of cervical spinal cord injuries. The injuries were associated with permanent neurological deficits which led to complete tetraplegia (quadriplegia) in 26 union and 23 league players and in both the "backyard" and the "touch" football players. (The touch and backyard football players were not included in this analysis because they were not playing according to official union or league rules [Table 1].) The 23 "unspecified" admissions shown in Table 1 could not be included in the formal analysis because the football code was omitted from their clinical records. Two patients died of pneumonia within 14 days of sustaining the cervical spinal cord injuries. The age range of the injured players from both codes was 15-37 years (median, 22 years); all were male. The number of patients admitted per year to the spinal units because of cervical spinal cord injuries varied widely: from one to nine per year for league and one to eight per year for union. The number of players who had permanent tetraplegia varied from none to five per year for league and none to six for union (Table 1). Incidence and trends There were about 106 000 participants in rugby league and about 31 000 in rugby union in NSW during each year of the review.5,6 Trends in rates of admissions and incidence of permanent neurological deficits per 10 000 participants are shown in Table 2. Using the data shown in Table 1, there was no decline in the number (and hence the rate) of admissions for union (P = 0.21) or league (P = 0.33) players. There was a small but significant decline for union in the number and hence the incidence (P = 0.06) of players with permanent neurological deficits, but no significant decline for league (P = 0.16). Type of play associated with severe cervical spinal cord injuries Although detailed self-reports and eyewitness accounts of the events surrounding the injuries were available for 43 of the 49 patients with complete tetraplegia, no consistent patterns of play or events could be identified other than that the injury occurred in scrum-like plays (scrums, mauls and rucks; n = 23) or tackles (n = 26). Repeatedly, eyewitnesses observed that a player was found paralysed on the ground, without an obvious causal explanation, while the general play moved on. This lack of a specific, out-of-the-ordinary event was supported by two video recordings of cervical spinal cord injuries that occurred during scrums. The injuries were most common in scrum-like plays in union and in tackles in league. The most common level of spinal cord damage was C4-5 (Table 3). Discussion Despite its limitations as a retrospective study (such as the use of records created for other purposes), the findings of this review are disturbing. While for union players there was a small reduction in both the number and incidence of those with permanent neurological deficits leading to tetraplegia, there were no such changes for league players. Nor were there any changes for either code in the number of players admitted or rates of admission to spinal units. While any fall in the number and incidence of permanent neurological deficits among union players is pleasing and may well be a consequence of rule changes and safety measures,7 the unchanged rate of admissions to spinal units suggests that such improvements may not be sustained. In those players admitted to spinal units who do not experience permanent neurological deficits, the injuries can be regarded as "near misses" (ie, possible damage to the vertebral column without spinal cord injury).7 While evacuation procedures for suspected cervical and other spinal cord injuries may vary, such "near misses" included in the admission rates offer a useful indicator of the risks of such injuries associated with rugby union and league football.8 Our observation that cervical spinal cord injuries are more common in scrum-like plays in union and tackles in league is similar to recent findings in New Zealand.7 However, there is insufficient information in this review and in the New Zealand7 and other reviews9 of rugby football-associated spinal cord injury upon which to base sound recommendations for detailed changes to the rules or other safety measures. These reviews confirm the obvious -- that both rugby union and league are inherently dangerous games. There appear to be several options: Maintain the status quo and accept that each year several union and league players in NSW (and elsewhere) will suffer cervical spinal cord injuries which will leave them permanently paralysed below the neck; For rugby union, change the rules so as to substantially reduce the number of scrum-like plays and change the organisation of the scrum; For rugby league, introduce changes to the rules aimed at substantially altering the nature of the tackle. Acknowledgement This study was conducted with a grant from the NSW Sporting Injuries Committee. Disclaimer of conflict of interest: Apart from providing access to their own case files, the NSW Sporting Injuries Committee was in no way involved with the collection or analysis of data and did not have the right to disapprove or influence the contents of the manuscript. References Yeo JD. Prevention of spinal cord injuries in an Australian study (NSW). Paraplegia 1993; 31: 759-763. Taylor TKF, Coolican MRJ. Spinal cord injuries in Australian footballers, 1960-1985. Med J Aust 1987; 147: 112-118. Silver JR. Injuries of the spine sustained during rugby. BMJ 1984; 288: 37-43. Silver JR, Stewart D. The prevention of spinal injuries in rugby football. Paraplegia 1994; 32: 442-453. NSW Sporting Injuries Scheme. Annual reports of the NSW Sporting Injuries Committee. Sydney 1984 to 1995. Sydney: NSW Sporting Injuries Committee, 1984-1995. Northern Sydney Area Health Service. NSW Youth Sports Injury Report. July 1997. Sydney: Northern Sydney Area Health Service, 1997. Armour KS, Clatworthy BJ, Bean AR, et al. Spinal injuries in New Zealand rugby and rugby league -- a twenty-year survey. N Z Med J 1997; 110: 462-465. Noakes T, Jakoet I. Spinal cord injuries in rugby union players: How much longer must we wait for proper epidemiological studies? BMJ 1995; 310: 1345-1346. Kew T, Noakes TD, Kettles AN, et al. A retrospective study of spinal cord injuries in Cape Province rugby players, 1963-1989. S Afr Med J 1991; 80: 127-133. (Received 18 Sep 1997, accepted 27 Jan 1998) Authors' details School of Health Services Management, Faculty of Medicine, University of NSW, Sydney, NSW. Tai R Rotem, BSocSci, Research Fellow; James S Lawson, MD, MHA, Professor; Chris W Aisbett, BSc, Visiting Fellow. Department of Aged Care and Rehabilitation, Royal North Shore Hospital, Sydney, NSW. Stephen F Wilson, MB BS, FAFRM(RACP), Senior Staff Specialist. Department of Rehabilitation, Prince Henry Hospital, Sydney, NSW. Stella Engel, DPRM, FAFRM(RACP), Director. Spinal Injuries Unit, Royal North Shore Hospital, Sydney, NSW. Sue B Rutkowski, MB BS, Director. Reprints will not be available from the authors. Correspondence: Professor J S Lawson, School of Health Services Management, Faculty of Medicine, University of NSW, Sydney, NSW 2052. - Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". <URL: http://www.mja.com.au/>

Tai R Rotem · James S Lawson · Stephen F Wilson · Stella Engel · Sue B Rutkowski · Chris W Aisbett

Attitudes to and use of a modified prescription form by general practitioners and pharmacists Merilyn J Liddell and Sue P Goldman MJA 1998; 168: 322-325 For editorial comment see Roberts Abstract - Introduction - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - ©MJA1998 Abstract Objective: To evaluate the rate of use and acceptance of a new prescription form designed to provide more information to pharmacists and patients. Design: Prospective descriptive study. Setting: A semi-rural community outside Melbourne, Victoria, in November 1994. Participants: GPs and pharmacists working three or more sessions per week in the locality, and 21 consumers who formed two consumer focus groups. Intervention: An education session for GPs and pharmacists, followed by a one-month trial of new prescription forms which included notations to facilitate interprofessional communication. Main outcome measures: Rate of use and acceptability of the new prescription notations. Results: Eighteen GPs and 10 pharmacists participated; 3600 forms were issued and 2521 prescriptions, including 3464 prescription items, were analysed. GPs and pharmacists used at least one of the new prescription notations in 45% (1559) of prescription items; 35% of prescription items (1222) were notated with the general purpose of the medication. Qualitative findings suggested that GPs, pharmacists and consumers considered the new prescription form to be beneficial and useful. Conclusions: A modified prescription form to increase communication between GPs and pharmacists is acceptable in clinical practice. Introduction Over recent years professional standards for prescribing have emphasised the need for a more patient-centred approach,1 as shown in such documents as the National Health and Medical Research Council's guidelines for medical practitioners on providing information to patients,2 the Policy on the Quality Use of Medicines developed by the Commonwealth Government,3 and regulations on the provision of written consumer information with dispensed medication.4 Research on how to provide more patient-centred pharmaceutical information and counselling shows that closer collaboration between general practitioners (GPs) and pharmacists would help to ensure that medication information is both comprehensive and relevant for consumers.5,6 A joint working party of the Royal Australian College of General Practitioners (RACGP) Victoria Faculty and the Victorian Branch of the Pharmaceutical Society of Australia (PSA) was formed in 1993 to consider how to improve communication between GPs, pharmacists and consumers. A series of recommendations were developed and endorsed by the national executives of the RACGP and the PSA, and published in a joint statement.7 The major recommendation was the use of a new prescription form to aid communication between GPs, pharmacists and patients. Here, we describe a pilot study to determine the rate of use of, and the attitudes of GPs, pharmacists and patients to, a modified prescription form for use in normal clinical practice. No similar study has been reported in the literature (based on the following searches: MEDLINE, January 1982 - June 1997; International Pharmaceutical Abstracts, January 1970 - December 1996; HEALTH, January 1975 - December 1996; EMBASE, January 1994 - December 1996; Pharmaceutical News Index, January 1974 - December 1996). Methods Ethical approval for this study was received from the Monash University Standing Committee on Ethics in Research on Humans and all subjects gave informed consent for their par ticipation. The study was conducted in November 1994 in a semi-rural community on the outskirts of Melbourne with a population of 25 000. All 24 GPs and 13 pharmacists working three or more sessions per week in the locality were invited to participate; 18 GPs and 10 pharmacists agreed. Twenty-one consumers from existing heart and arthritis support groups at the local community health centre were recruited for two focus groups of eight and 13 participants. Before the intervention, the 18 participating GPs and 10 pharmacists completed separate questionnaires (29 questions for pharmacists and 28 for GPs), which covered demographics and assessed (on a scale of 1 to 5) attitudes to professional roles, interprofessional collaboration and consumers' needs for medication information and counselling. Intervention The intervention began with an educational session for GPs and pharmacists to introduce the modified prescription form on which GPs could notate the purpose of prescribed medications, changes to existing medications and specific patient needs. Some of these notations were unprompted, requiring freehand additions to the prescription, while others were prompted, appearing as abbreviations on the form which could be circled when appropriate (see Box 1). A glossary of the new prescription notations was provided. Box 1 shows the glossary which was provided to pharmacists. During the education sessions, the new notations were explained as follows. The major change was that GPs were encouraged, where they considered it appropriate, to indicate the purpose of each prescribed medication as part of the written directions. For example, -adrenoceptor blocking agents might be notated as "for arrhythmias" or "for blood pressure". Pharmacists could then ensure that their medication counselling was relevant to that particular indication and would include the medication's purpose on the medication label for the patient's information. There was provision for GPs to request that the purpose not be written on the medication label by the pharmacist in certain circumstances (such as to preserve patients' privacy if the nature of their illness was not known to family members). GPs were also encouraged to indicate to the pharmacist if therapy with an ongoing medication was to cease (enabling pharmacists to update their computer records to show that a drug was no longer to be taken, and attempt to withdraw any unpresented repeats from circulation), or if the directions intentionally specified an unusual dosage or quantity to be dispensed. Finally, GPs could circle abbreviations printed on the prescription form to indicate specific patient needs or changes in the medication regimen. Pharmacists could then check the patient had understood changes to directions and amend the computer records accordingly. At the end of the educational session, 200 triplicate prescription forms were issued to each GP for use over one month. The third copy of each prescription dispensed (minus identifying patient details) was collected to determine the use of the new notations. After one month, individual one-hour open-ended, semi-structured interviews were conducted with all participating GPs and pharmacists, covering their overall views on the reformatted prescription, specific issues of practical application and suggested changes. Two separate one-and-a-half-hour focus group discussions with consumers dealt with aspects of the use of the reformatted prescription which would directly affect patients. Data analysis For frequency analysis of the questionnaires and of the prescription data, we used, respectively, the statistical package Genstat8 and the computer database program Access.9 Transcripts of interview data and focus group discussions were coded according to emerging themes around the prescription format and particular notations. The analysis, with the NUDIST program for qualitative analysis,10 concentrated on direct testimonies of views held about the new prescription form. Results The median age of the 18 GPs was 35 years (range, 20-29 to >70 years) and 14 were male. Their median workload was 175 patients per week (range, 50-99 to >200) and 13 had postgraduate qualifications. The GPs' age and sex distribution was similar to that of respondents in a national general practice population survey.11 The median age of the 10 pharmacists was 40 years (range, 20-29 to 50-59 years) and eight were male. Their median workload was 125 prescriptions dispensed per day (range, 50-99 to >200) and three had postgraduate qualifications. Attitudes to interprofessional communication In the questionnaire, 16 GPs (89%) and all of the pharmacists stated that they were in favour of more collaboration with one another. GPs had a reasonably high regard for the local pharmacists -- 13 (72%) did not agree that pharmacists' advice conflicted with their own; 17 (94%) agreed they had a good working relationship with the pharmacists, and none believed that pharmacists contacted them unnecessarily. Seven pharmacists (70%) stated that they enjoyed a comfortable working relationship with GPs; only three (30%) agreed that GPs were not always receptive to inquiries about prescriptions, and five (50%) expected the GP to be responsive to contact about a potential drug interaction. Before the intervention, the concept of including the general purpose of a medication on the prescription was not recognised by seven GPs (39%) and seven pharmacists (70%) as useful or necessary for the pharmacist. However, nine pharmacists (90%) stated that to counsel patients adequately they often needed to ask a patient for more information than was on the prescription. Use of the new prescription form Of a total 3600 prescription forms distributed to the GPs, 70% (2521 prescriptions for 3464 individual prescription items) were collected and analysed. Forty-five per cent of individual prescription items (1559) included one of the new notations. The rate of use of the new notations by the GPs is shown in Box 2. The purpose of the medication was notated for 35.3% of all prescription items (1222), with use of this notation by individual GPs ranging from 1% for the lowest user to 92% for the highest user. In only 0.3% (range, 0-4.5%) of cases where this notation was used did doctors request that the information be omitted from the label of the medication container. Attitudes to the new prescription form Overall, the new system was positively received by most GPs and pharmacists, and they suggested it should be widely implemented. Also, most GPs and pharmacists saw indicating the purpose of the medication on the prescription as the most substantial and important focus of the new prescription notations, and suggested that this would enhance the quality of information provided to consumers and minimise irrelevant or inappropriate advice. Further, some suggested that compliance may be improved because patients would receive consistent advice from GPs and pharmacists. Most GPs and pharmacists considered it useful to note when therapy with a long term medication was to be ceased, as patients may forget verbal instructions by the GP. Although the need did not often arise in practice, both GPs and pharmacists were in favour of specifically informing the pharmacist when an unusual dosage or a special quantity of medication was being prescribed. Most GPs and pharmacists thought it appropriate to indicate a new treatment, primarily for ongoing management. Noting a change of directions was considered useful by doctors and pharmacists alike, particularly with elderly patients or those who might still have unused prescriptions. The pharmacist could then be sure the change was intended, check the patient had understood the change, amend any unpresented repeats available, and note the change in the pharmacy computer record. Many GPs, pharmacists and consumers regarded preprinted abbreviations for specific patient needs as unnecessary; it was suggested that longhand could always be used for these or other specific requests to the pharmacist. The abbreviation for noting when separate written instructions had been given to the patient was considered a simple and useful quality control mechanism by doctors, pharmacists and consumers. The consumer focus groups indicated that pharmacists needed to know why someone was being prescribed a medication to enable them to provide appropriate counselling. They expressed reservations about conditions of a very personal nature; in such cases they expected the doctor to either ask their permission or omit the information from the prescription. In general, they considered it both appropriate and helpful to have the purpose of the medication included on the medication label. Discussion In this study the questionnaire findings gave basic data about pre-existing attitudes of the GPs and pharmacists to their respective roles, while the qualitative data explored, in much more depth, the strengths and weaknesses of the new innovation. A 70% return rate of prescriptions was satisfactory, as non-redemption of prescriptions can range from 5%-20%.12 Our most notable finding was the utilisation rate of the new notations, with 35% of prescription items indicating the purpose of the medication to the pharmacist. This was somewhat unexpected, as such a strategy had not previously been considered necessary by GPs or pharmacists. Further, such a high utilisation rate requires considerable behaviour change, and behaviour change strategies generally have much more modest outcomes (eg, 6% for minimal intervention smoking cessation programs13 and 8%-25% for group counselling14). The benefits of asking GPs to indicate the purpose of the medication on the prescription were supported by the qualitative data suggesting that when this strategy was employed it was highly valued. Both GPs and pharmacists indicated that the new prescription form would improve the quality of information given to patients by ensuring that information given by pharmacists was consistent with the advice already given by the GP. It would also enable more information to be given to patients on the medication label. It was of particular interest that, while some of the GPs and pharmacists had very high workloads, they were able to use the new prescription conventions easily in their normal daily practice. There was a high level of utilisation of the new prescription form as a whole, with some notations being used frequently, and others barely at all. This is to be expected as the appropriate circumstance for the use of some notations (such as unusual dosage) would be infrequent. The notation to pharmacists that drug therapy was to be ceased was not widely used, possibly because this affected only a small number of patients, but more likely because there was no specific reminder on the prescription form itself and doctors may have simply forgotten to use it. While some variation in the use of notations may have related to differing perceptions of their value, it was probably also a result of the difficulty of altering ingrained behaviour. Some participants stated that they would try to increase their use of the strategies if they were implemented in the future. Indicating the purpose of the medication is easily misinterpreted as being the same as including the diagnosis, and thereby providing unnecessarily detailed and confidential information to the pharmacist. Participating GPs, however, reported that they provided general information at an appropriate level for the pharmacist and, subsequently, the patient. From the pharmacists' perspective, information about the purpose of a medication was most useful for drugs with multiple indications. It enabled them to target their information appropriately without asking questions to ascertain the precise indication for a particular patient. Generally, consumers thought that including the purpose of the medication on the container label was likely to enhance patients' understanding of their own medication. Confidentiality needs always to be considered, but there are clear advantages in having a drug's purpose included on the label in most situations. As we used only two small consumer focus groups, they could provide only a limited indication of what consumers may think of the proposed system. However, consumers did suggest that the new system would help them gain access to appropriate information and advice. Any future implementation should be made with adequate consumer input, to ensure attention to their requirements. In conclusion, this study showed that it is possible to modify the prescription form to include more information in a way that is acceptable to GPs and pharmacists in their daily practice. GPs, pharmacists and consumers found such changes worthwhile and believed they could lead to better medication management and patient care. Acknowledgements We thank the Royal Australian College of General Practitioners and the Victorian Branch of the Pharmaceutical Society of Australia, and the members of the joint working party -- Val Constable, John Daffey, David Dammery, Chris Hogan, Alistair Lloyd, Mary Murray and Denise Ruth -- as well as the GPs, pharmacists and consumers in the study. Finally, we thank the Department of Health, Housing, Local Government and Community Services for financial support for the project. References Liddell M. Rational prescribing and professional standards. Med J Aust 1994; 160: 564-567. Working Party of the Health Care Committee, National Health and Medical Research Council. General guidelines for medical practitioners on providing information to patients. Canberra: National Health and Medical Research Council, 1993. Commonwealth Department of Health, Housing and Community Services, in conjunction with the Pharmaceutical Health and Rational Use of Medicines (PHARM) Working Party. A policy on the quality use of medicines. Canberra: Commonwealth Department of Health, Housing and Community Services, 1992. Part 2A, Schedule 12 of the Therapeutic Goods Regulations under the Therapeutic Goods Act 1989, Section 63. Murphy B, Ruth D, Murray-Hodge M. The use of qualitative research in the development of the HEARTWISE program for general practitioners. Med J Aust 1993; 158: 626-628. Ruth D, Hodge M, Murphy B. Improving the relationship between general practitioners and pharmacists. Aust Fam Physician 1994; 23: 1536-1540. Royal Australian College of General Practitioners and the Pharmaceutical Society of Australia. General practitioners' and pharmacists' interprofessional communication. Aust Fam Physician 1994; 23: 1544-1546. Genstat [computer program]. Version 5 Release 3.1. Lawes Agricultural Trust (Rothamstead Experimental Station, UK), 1993. 9. Access [computer program]. Version 2.0. Redmond, Wa.: Microsoft Corporation, 1994. Richards T, Richards L. The NUDIST qualitative data analysis system. Qual Sociol 1991; 14: 307-325. Bridges-Webb C, Britt H, Miles D, et al. Morbidity and treatment in general practice in Australia 1990-1991. Med J Aust 1992; 157 Suppl Oct 19: 14. Beardon P, McGilchrist M, McKendrick A, et al. Primary non-compliance with prescribed medication in primary care. BMJ 1993; 307: 846-848. Kottke T, Battista R, DeFrieze G, et al. Attributes of successful smoking cessation interventions in medical practice: a meta-analysis of 39 controlled trials. JAMA 1988; 259: 2882-2889. Curry S, Marlatt G, Gordon J, et al. A comparison of alternative theoretical approaches to smoking cessation and relapse. Health Psychol 1988; 7: 545-556. (Received 5 Jun, accepted 10 Nov, 1997) Authors' details Department of Community Medicine and Department of General Practice, Monash University, Melbourne, VIC. Merilyn J Liddell, MB BS, FRACGP, Senior Lecturer; Sue P Goldman, BEd, GradDipSoc, Research Fellow. Reprints will not be available from the authors. Correspondence: Dr M J Liddell, 867 Centre Road, East Bentleigh, VIC 3165. E-mail: merilyn.liddell AT med.monash.edu.au Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". <URL: http://www.mja.com.au/>

Merilyn J Liddell · Sue P Goldman

Use of systematic reviews of randomised trials by Australian neonatologists and obstetricians Christopher F C Jordens, Penelope Hawe, Les M Irwig, David J Henderson-Smart, Margaret Ryan, Deborah A Donoghue, Roger G Gabb and Ian S Fraser MJA 1998; 168: 267-270 For editorial comment see Phillips Abstract - Introduction - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - ©MJA1998 Abstract Objective: To determine what proportion of Australian neonatologists and obstetricians report using systematic reviews of randomised trials. Design: Cross-sectional survey using structured telephone interviews. Setting: Australian clinical practice in 1995. Participants: 103 of the 104 neonatologists in Australia (defined as clinicians holding a position in a neonatal intensive care unit); a random sample of 145 members of the Royal Australian College of Obstetricians and Gynaecologists currently practising in Australia. Main outcome measures: Information sources used in clinical practice; reported awareness of, access to and use of systematic reviews, and consequent practice changes. Results: Response rates were 95% (neonatologists) and 87% (obstetricians); 71 neonatologists (72%) and 55 obstetricians (44%) reported using systematic reviews, primarily for individual patient care. Databases of systematic reviews were used with a median frequency of once per month. Among neonatologists, systematic reviews were used more commonly by those who were familiar with computers, attended professional meetings, and had authored research papers. Among obstetricians, they were used more commonly by those who were familiar with computers, had less than 10 years' clinical experience, attended more deliveries, and were full-time staff specialists in public hospitals. Of neonatologists who reported using systematic reviews, 58% attributed some practice change to this use. For obstetricians, the corresponding figure was 80%. Conclusions: There is evidence that Australian neonatologists and obstetricians use systematic reviews and modify their practice accordingly. Dissemination efforts can benefit from knowledge of factors that predict use of systematic reviews. Introduction Randomised controlled trials are widely accepted as the best method of evaluating the effectiveness of medical interventions. However, their findings are slow to change medical opinion and practice.1,2 The communication of clinically important research findings is hampered by the volume and geometric growth of the medical literature.3 Reviews address this problem, but conventional review methodology is unscientific4,5 and open to selection biases inherent in the publication process.5,6 This has led to the development of systematic reviews. A properly conducted systematic review begins with an exhaustive search for published and unpublished randomised trials addressing a well defined research question. The second step is to identify trials of adequate quality to contribute to decision-making. The results of included trials are then pooled, or "meta-analysed", to arrive at a quantitative estimate of the benefits and harms of treatment. Systematic reviews published by the Cochrane Collaboration are also continuously updated by specialist review groups.7 Systematic reviews in pregnancy and childbirth have been available for some years. In 1988 a comprehensive trial register became commercially available in database form as the Oxford database of perinatal trials.8 Effective care in pregnancy and childbirth, 9 a textbook based on this research synthesis, was published the following year, with a guide for non- specialist readers.10 A companion textbook, Effective care of the newborn infant,11 appeared in 1992. From 1993, systematic reviews relating to pregnancy and childbirth were available electronically in the Cochrane pregnancy and childbirth database.12 All these reviews are currently being updated for The Cochrane library.13 These new publications represent prototypes for future publications in other fields of practice. However, their impact on clinical practice remains uncertain, despite several surveys to ascertain the extent of their dissemination and uptake.14-18 To make a practical difference, systematic reviews must be readily available to clinicians who are aware of them, who use them, and who implement their findings. To ascertain whether this is occurring in Australia, we conducted a survey of neonatologists and obstetricians with the aims: To gauge awareness of, access to, and use of the Oxford database, Cochrane database, and Effective care textbooks; To place these resources in the context of other information sources that inform practice in obstetrics and neonatology; and To ascertain how often and why these resources are consulted, predictors of their use, and whether their use has led to reported changes in clinical practice. Methods Participants Clinicians who held a position in one of Australia's 23 neonatal intensive care units were eligible for the survey. The Australian and New Zealand Neonatal Network (ANZNN), which covers all neonatologists, provided a current listing. The Royal Australian College of Obstetricians and Gynaecologists (RACOG) drew a simple random sample of 20% of Fellows from its membership. Those sampled were eligible if they were currently practising obstetrics. Any clinicians not currently practising in Australia, not on the telephone network, or who were involved in designing this survey, were ineligible. Eligible clinicians were sent an introductory letter, telephoned, and invited to schedule a confidential, 10-minute telephone interview about information sources used in clinical practice. Systematic reviews were not mentioned until the interview was under way. Interviews An existing interview schedule15,16 was modified in consultation with the ANZNN and the RACOG. Respondents were first asked to name the three main sources of information they used for keeping up with new developments in their field, and for dealing with uncertainty about a specific treatment decision. They were also asked which of the three information sources they considered the most important or useful for each task. Respondents were then asked about their access to and use of computers, and asked directly whether they had heard of, had access to, and used the Cochrane database, Oxford database, and Effective care of the newborn infant (neonatologists) or Effective care in pregnancy and childbirth (obstetricians). Those who reported using any of these resources were asked what a systematic review was, how often they consulted that resource, what they used it for, whether they thought using it had made any difference to their clinical practice and, if so, whether they could name a treatment policy they had altered in response to evidence from a systematic review. To count as users of systematic reviews, respondents had to mention randomised trials or meta-analysis in their description of a systematic review. The final questions sought demographic information on clinical experience, place of education, attendance at professional meetings, research publications, academic appointments, and qualifications. A research degree was defined as a doctorate, relevant master's degree, or bachelor's degree of science in medicine, held in addition to basic medical and specialty qualifications. Reports of authorship were verified by searching MEDLINE and by screening abstracts. For neonatologists it was noted whether they held full- or part-time positions, whether they worked in a hospital with an obstetric unit, whether they headed a neonatal unit, and in which unit they worked. For obstetricians, it was noted whether they held a position as a full-time staff specialist in a public hospital. They were asked how many deliveries they attended each year, and whether they worked at any hospital with an accredited RACOG training post. The interview schedule was modified after a pilot study with seven neonatologists in New Zealand. The main survey was conducted between August and December 1995, with ethics approval from the ANZNN, RACOG and the University of Sydney. All interviews were conducted by the first author (C F C J). Analysis Confidence intervals for the proportion of obstetricians who reported using systematic reviews were calculated with a finite population correction. Confidence intervals were not calculated for the corresponding proportion of neonatologists, as this was ascertained for all Australian neonatologists. Descriptive and c 2 analyses were conducted. Then, using a backwards stepwise modelling procedure, a multivariate logistic regression analysis was used to determine which study factors predicted the reported use of systematic reviews. With the neonatology data, this modelling procedure was repeated using binomial generalised estimating equations to adjust for clustering of respondents within neonatal units.19 Results Of 106 listed neonatologists, three were ineligible (one had retired, one was not practising in Australia, and one was involved in designing this survey). Of the 103 eligible clinicians remaining, 98 completed interviews (95% response). From a sample of 210 obstetricians, 65 were ineligible for the survey (51 no longer practised obstetrics, 13 were not practising in Australia, and one could not be located on the telephone network). Of 145 eligible obstetricians, 126 completed interviews (87% response). Respondent characteristics are shown in Box 1, with comparative data on computer use. For the purpose of keeping up with new clinical developments, respondents favoured journals, conferences and meetings, colleagues and MEDLINE (in that order) over systematic reviews. For the purpose of clinical problem-solving, colleagues, MEDLINE, journals, and other printed medical literature were favoured over systematic reviews as sources of information. Although the percentages favouring each source varied between the two specialties and according to whether the respondent was simply nominating the resource or nominating it as the most important resource, the rank order of the resources remained consistent. In each specialty, 21% of respondents mentioned systematic reviews without prompting from the interviewer when naming the information sources they used, or when describing their computer use. When questioned directly about systematic reviews, 71 neonatologists (72%) and 55 obstetricians (44%; 95% confidence interval, 36%-51%) reported using them in either electronic database or textbook form. The confidence interval indicates the reliability of the sample prevalence as an estimate of the population prevalence. The Figure shows reported levels of awareness of, access to and use of systematic reviews by specialty. Predictors of use of systematic reviews Results of the multivariate analysis are shown in Box 2. Among neonatologists, three factors significantly (P<0.05) and independently predicted use of systematic reviews: attendance at meetings of the Australian Perinatal Society, authorship of at least one research paper, and familiarity with computers (our index of this was use of a computer for word-processing). Results from cluster analysis agreed with those from logistic regression. Among obstetricians, four factors sig nificantly and independently predicted use of systematic reviews: familiarity with computers, mode of practice (full-time staff specialists were more likely to use systematic reviews than others), clinical experience (recent graduates were more likely to use systematic reviews), and number of deliveries per year (likelihood of using systematic reviews increased with this number). Descriptors of use of systematic reviews Both neonatologists and obstetricians reported using databases of systematic reviews with a median frequency of once per month. The main purpose for using systematic reviews (irrespective of their format) reported by respondents from both specialties was individual patient care. Systematic reviews were also said to be used for (in rank order of frequency of reporting) teaching, preparing a pre sentation, reviewing current management or developing evidence-based protocols or guidelines, settling disputes, background information, reviewing a topic, research, as a source of references, and patient information. Of the 71 neonatologists who reported using systematic reviews, 58% said that this had changed their clinical practice in some way, and 44% gave at least one example of a treatment policy they had altered in response to a systematic review. The most common examples were treatment of respiratory disorders, and use of steroids and indomethacin. Corresponding percentages for obstetricians who reported using systematic reviews were 80% (reported a change in practice) and 71% (gave at least one example). The most common example was use of steroids in management of preterm rupture of membranes. Discussion We found that 72% of neonatologists and 44% of obstetricians reported consulting systematic reviews, primarily for the purpose of individual patient care. Databases of systematic reviews were used at a median frequency of once per month. Systematic reviews were used more commonly by those who were familiar with computers in both specialties, by those who attended professional meetings and had authored research papers among neonatologists, and by those who had had less than 10 years' clinical experience, attended more deliveries, and who had a position as a full-time staff specialist among obstetricians. The findings of this survey are based on self-reporting. Although the survey was designed to minimise over-reporting, it was still subject to imprecision in respondents' recall about (for example) the frequency with which they used systematic reviews, and to difficulties in attributing practice changes to their use. Nevertheless, this was the first study of the use of systematic reviews by Australian clinicians, and the findings have immediate and practical relevance for organisations such as universities, government agencies and specialty colleges that are trying to improve access to the best available evidence and to promote its use.20-23 The association between use of systematic reviews and attendance at Australian Perinatal Society meetings among neonatologists could reflect greater receptiveness to innovations among clinicians who attend professional meetings. However, it also supports the perception (reported elsewhere13 ) that professional organisations play an important role in disseminating research findings. Future dissemination efforts could usefully concentrate on these organisations. In both specialties, familiarity with computers predicted use of systematic reviews. Better access to digital information technology and training in its use is therefore likely to enhance uptake of these reviews. This might include improved access to The Cochrane library and the Internet, and training workshops for searching specialist databases. The remaining predictors of use of systematic reviews, as well as the overall contrast between neonatology and obstetrics, suggest that specialists practising primarily in public hospitals are more likely to use systematic reviews than those working primarily in private practice. Special efforts are therefore needed to reach the latter. As systematic reviews do not appear to rank highly among the information sources used in clinical practice, it is important to stress their relative advantages. Uptake of this innovation appears more likely among recent graduates, with senior clinician researchers acting as "product champions".24 Although debate persists as to whether practising clinicians accept the innovations of evidence-based medicine,25 our survey found evidence that Australian neonatologists and obstetricians use systematic reviews, and appear to modify their practice accordingly. Efforts are needed to enhance the use of these reviews, and to conduct further evaluations of their influence on clinical practice. In attempting to improve practice standards, it is important to pursue methods which have been shown to be effective.26 Acknowledgements We would like to thank the many clinicians who gave up their time to participate in this survey. Infrastructure support for this project was provided by the Department of Public Health and Community Medicine at the University of Sydney. Special thanks are due to Petra Macaskill, who conducted the cluster analysis, and Jeanette Ward and Mary Osborne, from the Central Sydney Area Health Service Needs Assessment and Health Outcomes Unit, who provided helpful advice on survey procedures. References Antman EM, Lao J, Kupelnick B, et al. A comparison of results of meta-analyses of randomized control trials and recommendations of clinical experts: treatments for myocardial infarction. JAMA 1992; 268: 240-248. Stross JK, Harlan WR. The dissemination of new medical information. JAMA 1979; 241: 2622-2624. Warren KS. From papyrus to parchment to paper to pixels: information technology and the future of biomedical publishing. In: Lock S, editor. The future of medical journals. London: BMJ, 1991: 127-146. Mulrow CD. The medical review article: state of the science. Ann Intern Med 1987; 106: 485-488. Light RJ, Pillemer DB. Summing up: the science of reviewing research. Cambridge: Harvard University Press, 1984. Dickersin K, Berlin JA. Meta-analysis: state-of-the-science. Epidemiol Rev 1992; 14: 154-176. Sackett DL. The Cochrane Collaboration. ACP J Club 1994; 120 Suppl 3: A-11. Chalmers I, editor. Oxford database of perinatal trials. Version 1.2, disk issue 7 (Spring). Oxford: Oxford University Press, 1992. Chalmers I, Enkin M, Keirse MJNC. Effective care in pregnancy and childbirth. Oxford: Oxford University Press, 1989. Enkin M, Keirse MJNC, Chalmers I. A guide to effective care in pregnancy and childbirth. Oxford: Oxford University Press, 1989. Sinclair JC, Bracken MB. Effective care of the newborn infant. Oxford: Oxford University Press, 1992. Enkin MW, Keirse MJNC, Renfrew MJ, Neilson JP, editors. Cochrane pregnancy and childbirth database. Oxford: Update Software, 1993. The Cochrane Library [database on disk and CD-ROM]. The Cochrane Collaboration. Oxford: Update Software; 1996. Updated quarterly. Lomas J. Retailing research: increasing the role of evidence in clinical services for childbirth. Milbank Q 1993; 71: 439-475. Stocking B. Implementing the findings of effective care in pregnancy and childbirth in the United Kingdom. Milbank Q 1993; 71: 497-523. Paterson-Brown S, Fisk NM, Wyatt JC. Uptake of meta-analytical overviews of effective care in English obstetric units. Br J Obstet Gynaecol 1995; 102: 297-301. Paterson-Brown S, Wyatt JC, Fisk NM. Are clinicians interested in up to date reviews of effective care? BMJ 1993; 307: 1464. Hyde C. Who uses the Cochrane pregnancy and childbirth database? BMJ 1995; 310: 1140-1141. Zeger SL, Liang K-Y. Longtitudinal data analysis for discrete and continuous outcomes. Biometrics 1986; 42: 121-130. Rychetnik L. Evidence-based medicine in the GMP (Graduate Medical Program). Focus: graduate medical degree news and events. Sydney: 1995: 1-2. Liddle J, Williamson M, Irwig L. Method for evaluating research and guideline evidence. Sydney: NSW Department of Health, 1996. National Health and Medical Research Council. Clinical practice guidelines: the management of early breast cancer. Canberra: NHMRC, 1995. Commonwealth Department of Human Services and Health. Guidelines for the development and implementation of clinical practice guidelines. Canberra: the Department, 1995. Rogers EM. Diffusion of innovations. New York: MacMillan, 1983. Miles A, Bentley P, Polychronis A, Grey J. Evidence-based medicine: why all the fuss? J Evaluation Clin Pract 1997; 2: 83-85. Grimshaw JM, Russell IT. Effect of clinical guidelines on medical practice: a systematic review of rigorous evaluations. Lancet 1993; 342: 1317-1322. (Received 11 Jun, accepted 11 Sep, 1997) Authors' details University of Sydney, Sydney, NSW. Christopher F C Jordens, MPH, Postgraduate Student, Department of Public Health and Community Medicine; now Researcher, Centre for Values, Ethics and the Law in Medicine, Department of Surgery, University of Sydney; Penelope Hawe, MPH, Senior Lecturer, Department of Public Health and Community Medicine; Les M Irwig, FFPHM, PhD, Associate Professor, Department of Public Health and Community Medicine; David J Henderson-Smart, PhD, FRACP, Professor, and Director, NSW Centre for Perinatal Health Services Research, and Department of Neonatal Medicine, Royal Prince Alfred Hospital, Sydney, NSW; Deborah A Donoghue, RN, BSocSc, Senior Research Assistant, Australian Institute of Health and Welfare National Perinatal Statistics Unit; Ian S Fraser, MD, FRACOG, Professor in Reproductive Medicine, Department of Obstetrics and Gynaecology, Queen Elizabeth II Research Institute for Mothers and Infants. Royal Australian College of Obstetricians and Gynaecologists, Melbourne, VIC. Margaret Ryan, MSW, PhD, Research Officer. Centre for Professional Development, Victoria University of Technology, Melbourne, VIC. Roger G Gabb, PhD, Professor, and Director. Reprints: Mr C F C Jordens, Department of Surgery, Blackburn Building D06, University of Sydney, NSW 2006. E-mail: cjordens AT surgery.usyd.edu.au Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". <URL: http://www.mja.com.au/>

Penelope Hawe · Les M Irwig · David J Henderson-Smart · Margaret Ryan · Deborah A Donoghue

Smoking behaviours of Australian adults in 1995: trends and concerns David J Hill, Victoria M White and Michelle M Scollo MJA 1998; 168: 209-213 For editorial comment see Gray Abstract - Introduction - Methods - Results - Discussion - References - Authors' details - - ©MJA1998 Abstract Objectives: To estimate the prevalence of smoking among Australian men and women in 1995 and to examine trends in smoking prevalence in Australia over the past 10 years. Design: A representative sample of adults participated in face-to-face interviews conducted by a large market research company. Participants: 2819 men and 2880 women over the age of 16. Main outcome measure: Self-reported smoking behaviours assessed by standard questions. Results: Overall, 27.1% of men and 23.2% of women were smokers of tobacco (factory-made cigarettes, pipes, cigars or roll-your-own cigarettes). This difference in smoking prevalence of men and women was significant. More men (32.1%) than women (21.7%) were past smokers and more women (53.4%) than men (39.3%) had never been regular smokers. On average, male smokers smoked about 20 factory-made cigarettes a day, while women smoked about 18. Occupation and education levels were inversely related to smoking prevalence. Comparisons with earlier data suggest that the decline in smoking prevalence seen in previous surveys has ceased. However, the number of cigarettes consumed daily decreased between 1992 and 1995. In the period between 1983 and 1989, when per capita expenditure on adult antismoking campaigns rose, smoking prevalence declined, but levelled off thereafter in a period when expenditure on campaigns fell. Conclusion: Failure to find a continuing decline in prevalence of smoking among the Australian population is of great concern and indicates the importance of continuing and extending antismoking programs. Introduction Cigarette smoking is still the most important preventable cause of premature death in Australia. Reducing the prevalence of smoking rightly remains a high priority on Australia's public health agenda.1,2 As such, it is important to collect and publish estimates of the prevalence of smoking and its sociodemographic correlates at regular intervals, as indicators of progress towards controlling the virtual epidemic of tobacco-related disease,3 and as a basis for public decisions on action to take. We have published data on national smoking prevalence for the years 1974,4 1976,5 1980,6 1983,7 1986,8 19899 and 1992,10 using a standard method of data collection. Here we report data for 1995. Methods Survey method: A large market research company carried out the sampling and interviewing as part of an omnibus survey in September/October 1995. Procedures were the same as in our previous surveys. In brief, interviews were conducted on the weekend, with respondents within a census collector's district selected at random within specified strata, including State and rural or urban divisions. Within each collector's district an individual residence was chosen at random for the first contact and the adjacent house was contacted next. Further adjacent households were approached until the required number of interviews for that collector's district were obtained (usually eight from about 200 households in each census collector's district). Response rate: Of the households contacted 47% agreed to participate in the survey, 42% refused, and in the remaining 11% either the residents were too old or ill or could not speak English, or call-back was unsuccessful. The response rate for this survey was similar to the rates for the 1992 and 1989 surveys, but it was lower than response rates in earlier years. Thus, if variable response rates influence reported prevalences, this could only have occurred for the period before 1989; recent comparisons are not confounded by response rates. Data collected: Respondents indicated their smoking status by choosing a category from the following list: Current smoker -- cigarettes only, cigarettes plus cigars or pipes, cigars only (ex-cigarettes), pipes only (ex-cigarettes), cigars only (never cigarettes), or pipes only (never cigarettes); Past smoker -- of cigarettes only, of cigarettes plus pipe or cigar, of cigars or pipes only; and Never smoker -- those who had never smoked regularly. Current smokers of factory-made cigarettes indicated the brand usually smoked, the size of the pack usually bought and the number of cigarette packets usually smoked in a week. Respondents' sex, age (in five-year age groups), country of birth, highest level of education achieved, and occupation were recorded, as well as the occupation of the household's main income earner. Analysis of data: We calculated 95% confidence intervals (95% CI) associated with estimates of smoking prevalence among men and women and the difference between estimates. Confidence intervals around prevalence estimates for particular subgroups (eg, men aged 50-59 years) are not given. We used chi-squared tests of association to examine differences in the prevalence of smoking between groups, and logistic regression analyses to explore whether associations between smoking prevalence, educational level, occupation and, for women, country of birth were independent of associations between age and smoking. We used analyses of variance and t tests to examine whether the mean number of cigarettes smoked differed between groups. Finally, to compare the prevalence of smoking across recent years, we age-standardised the data from the 1995, 1992 and 1989 surveys to the age distribution of the 1986 sample. Results Smoking status: Of the 5699 participants, 51% (2880) were women and 49% (2819) were men. A comparison of the distributions of age, occupation, education and country of birth in the sample data with those in the census data indicated no bias in the socio demographic variables in the dataset. Eighty-two per cent of respondents were interviewed on the first visit to their home. The smoking status of these respondents was similar to that of the more hard-to-reach respondents (those who were interviewed on the second or third contact). Cigarettes dominated tobacco use, with only 1% of men and less than 1% of women indicating that they smoked only pipes or cigars. Only 8% of men and 4% of women had smoked roll-your-own cigarettes in the month before the survey. The smoking status of the males and females in the sample is shown in Table 1. Smoking prevalence: In 1995, the estimated prevalence of smoking among Australian men aged 16 years and over was 27.1% (95% CI, 25.6%-28.7%), while among women the prevalence was estimated to be 23.2% (95% CI, 21.7%-24.7%); this difference was significant (3.9%; 95% CI, 1.6%-6.2%). The prevalence of smoking peaked between the ages of 25 and 29 in men (34.7%) and in women (35%), after which age smoking generally decreased among both men and women. Past smoking or non-smoking: More men (32.1%; 95% CI, 30.4%-33.8%) than women (21.7%; 95% CI, 20.2%-23.2%) were past smokers, a highly significant difference (10.4%; 95% CI, 8.1%-12.7%). More women (53.4%; 95% CI, 51.6%-55.2%) than men (39.3%; 95% CI, 37.5%-41.1%) had never been regular smokers, and again this difference was highly significant (14.1%; 95% CI, 11.5%-16.7%). Men and women differed in their age-related patterns of past smoking. Among men, the proportion of past smokers increased with age to reach a peak of 54.4% in those over 70. Although among women the proportion of past smokers peaked at 26% for those over 70, the proportion of past smokers in the other age groups showed little variation from the overall rate of 21.7%. The proportion who had ever smoked was calculated by adding the percentage of those who had smoked in the past to the percentage of current smokers. Overall, 59.2% (95% CI, 57.4%-61.0%) of men and 44.9% (95% CI, 43.1%-46.7%) of women had smoked at some stage in their life. The association between age and ever having smoked differed for men and women. While among men the proportion who had ever smoked increased with increasing age, for women the proportions of ever smokers began to decrease after the age of 35. Quit proportions: The quit proportion (ie, proportion of ever smokers who had given up smoking in each age and sex group) is also shown in Table 1. The overall quit proportion for men (0.54) was slightly higher than that among women (0.48). However, as Table 1 shows, there is little difference in the quit proportions for men and women in most age groups, except for the 16 to 19 years and 60 to 69 years age groups. Factory-made cigarettes: The mean number of factory-made cigarettes smoked per day by smokers was 19.7 (SD, 11.6) for men and 18.1 (SD, 10.8) for women; this difference was statistically significant (t=2.55, df=1238, P=0.01). In 1995, the cigarettes smoked by women had a lower average tar content than the cigarettes smoked by men (t=5.56, df=1022, P<0.001). The average tar content of cigarettes smoked by men was 8.2 mg (SD, 2.8), while for women it was 7.2 mg (SD, 3.0). Education: As in previous reports of this survey series, the prevalence of smoking and the mean number of cigarettes smoked per day differed for people in various sociodemographic groups. As Table 2 shows, smoking prevalence decreased with increased education, so that only 17.0% of men and 14.2% of women who were university graduates smoked. The association between education level and smoking prevalence was significant for both sexes, but was stronger for men (chi-squared=91.3, df=6, P<0.001) than women (chi-squared=35.5, df=6, P<0.001). The proportions of ex-smokers in the different education levels indicate that quitting was common in all groups. However, the greater proportion of never smokers in the better-educated groups indicates that the lower prevalence of smoking among this group was due to their never having smoked in the first place. For both men and women, smokers with more years of formal education smoked fewer cigarettes per day than did those who had fewer years of education (men: F3,618 = 3.15, P<0.05; women: F3,614=8.92, P<0.001). Occupational levels: Respondents were classified into one of four occupation levels based on the occupation of the household's main income earner. Occupational levels differed in the level of skill required for the job, such that unskilled workers (eg, labourers) were classified as "lower blue collar", while skilled workers (eg, plumbers) were classified as "upper blue collar". As occupation status increased, the prevalence of smoking decreased (Table 2). Among men from "upper white collar" households, 18.7% smoked compared with 40.9% of men from "lower blue collar" households. The association between occupation status and smoking was significant for both men (chi-squared=102.3, df=6, P<0.001) and women (chi-squared=46.8, df=6, P<0.001). Among men, the proportion of ex-smokers was slightly lower among "lower blue collar" households than other groups . Among women, the proportion of ex-smokers was similar among all occupation groups. The proportion of never smokers was greatest among higher occupation levels. This pattern of results indicates that the lower prevalence of smoking seen in higher occupation groups is due to the relatively lower rate of taking up smoking among these groups rather than their greater success at quitting. The differences in the mean number of cigarettes consumed per day between occupation groups was not statistically significant for men (F3,618 = 2.0, P=0.12) or women (F3,614=1.4, P=0.24). Country of birth: The prevalence of smoking was lowest among both men and women born in Asian countries. However, while the prevalence of 19.8% among Asian-born men was not significantly different from that found for groups born elsewhere, the prevalence of smoking among Asian-born women was significantly lower than the prevalence of smoking among women born in Australia (chi-squared=7.2, df=1, P<0.01) or the United Kingdom (chi-squared=6.6, df=1, P<0.01). The number of cigarettes smoked per day showed little variation according to place of birth. The overall association between place of birth and cigarette consumption was not significant for either men (F4,617=1.3, P=0.26) or women (F4,614=1.0, P=0.40). However, Asian-born men consumed significantly fewer cigarettes per day than did men born elsewhere (t=2.14, df=620, P<0.05). Logistic regression analyses of demo graphic data: In separate analyses, age was entered before the predictor variable of education level, occupational status or country of birth (for women only), and in each case the association between smoking prevalence and the predictor variable remained significant. Pack size and number of cigarettes smoked: The cigarette packet size most commonly used was 25 (by 36% of smokers), followed by packets of 30 and 40 (19% each), 50 (17%), 20 (6%) and 35 (4%). The number of cigarettes smoked per day was related to packet size (F5,1205=34.93, P<0.001). The mean number of cigarettes smoked per day by those who used packets of 20 was 13, for packets of 25 it was 16, for packets of 30 it was 17, for packets of 40 the mean was 23, and for packets of 50 it was 25. Except for women from lower blue collar households, the packet of 25 cigarettes was the most popular size for all occupation and education groups. After collapsing pack-size categories and combining sexes, 55% of blue collar smokers used pack sizes of 20, 25 or 30 compared with 68% of white collar smokers, and 32% of white collar smokers used packs of 35, 40 or 50 compared with 45% of blue collar smokers (chi-squared=20.0, df=1, P<0.001). Comparisons with previous years Smoking prevalences in Australia from 1986 to 1995 are shown in Figure 1, with a line of best fit superimposed. Of concern from a public health point of view is that the observed prevalence in 1995 is no lower than that in 1992, necessitating an upward adjustment of the projection to 2001 compared with our projection based on the four triennial surveys up to 1992. It was expected that the national prevalence targets for the year 2000 would be bettered.10 Now the trends suggest they will not be met, either for men or women. The results of the 1995 survey also show 5% more men and 3% more women smoking than would have been expected on the basis of the trends to 1992, and, by extrapolation, this means that in 1995, 356 000 more men and 384 500 more women were smoking in Australia than expected . The mean number of cigarettes smoked by men in 1995 was lower than the 22.1 cigarettes smoked per day by men in 1992 (t=3.39, df=1348, P<0.001). Among women, however, the average number of cigarettes smoked per day in 1995 was not significantly less than the number smoked in 1992 (19.1) (t=1.59, df=1322, P=0.113). The legislative and other activity to restrict advertising and promotion of tobacco, as well as expenditure on adult-directed antismoking campigns, between 1989 and 1995 are outlined in Box 1. Discussion This is the first time in eight successive surveys that the reported prevalence of smoking in men was no lower than the previous survey and the first time since 1983 that this was also the case for prevalence of smoking in women. If these data signal an underlying change and the previously falling prevalence of smoking has indeed stabilised, a major public health response is indicated. Attainment of what is arguably the nation's primary public health target1 -- reducing the prevalence of smoking in men and women to 20% by 2000 -- is clearly under threat. Given that this target was and still may be seen as modest, this would be a major public health failure, as the following calculations show. For every percentage point smoking prevalence in Australia exceeds the national goal, nearly 140 000 people are smoking who, had the goal been met, would not have been. So, if the goal is missed by 3%, which would be the case if the prevalence remains stable, nearly 420 000 more people than expected will be smoking. If they remain smokers, according to estimates of Doll et al,15 210 000 will die prematurely as a result of their smoking. Against the above fairly alarming observations must be set some auspicious trends -- smokers are smoking less and are probably less exposed to inhaled carcinogens and this should flow through to modest public health gains. As well as benefiting themselves, it seems plausible that a lower daily consumption by smokers is reducing the passive exposure of others to cigarette smoke, as forgone cigarettes may be some of those previously smoked at work or in public places where others would be exposed to sidestream smoke. Assuming these data indicate a slow-down or stalling in the previous reduction in smoking prevalence, why has this occurred and what needs to be done? Figure 2 suggests one explanation for influences on smoking prevalence. At times when antismoking activity is high, whether it be in the form of policy or programs, smoking declines, but when these abate prevalence stagnates. Clearly, these data call for more extensive and rigorous analysis which might incorporate, in a multivariate analysis, other factors such as price, regulations and pro-cigarette promotions. This could determine the extent to which program expenditure affects smoking prevalence. The trends are extremely worrying. Stasis in public policy and prevention programs paralleled by static smoking levels suggest the importance of continuing to extend antismoking programs in order to restart the reduction in smoking prevalence. The experience of some Scandinavian countries shows that public health authorities cannot rest on their laurels.16 Much remains to be done (Box 2). References Commonwealth Department of Human Services and Health. Better Health Outcomes for Australians. Canberra: AGPS, 1994. Australian Institute of Health and Welfare. Tobacco use and its health impact in Australia. Canberra: AIHW, 1996. English D, Holman CD, Milne E, et al. The quantification of drug caused morbidity and mortality in Australia. Canberra: Commonwealth Department of Human Services and Health, 1995. Gray N, Hill D. Patterns of tobacco smoking in Australia. Med J Aust 1975; 22: 819-822. Gray NJ, Hill D. Patterns of tobacco smoking in Australia II. Med J Aust 1977; 20: 329-330. Hill D, Gray N. Patterns of tobacco smoking in Australia III. Med J Aust 1982; 1: 23-25. Hill D, Gray N. Australian patterns of tobacco smoking and related health beliefs in 1983. Community Health Stud 1983; 8: 307-316. Hill D. Australian patterns of tobacco smoking in 1986. Med J Aust 1988; 149: 6-10. Hill D, White V, Gray N. Australian patterns of tobacco smoking in 1989. Med J Aust 1991; 154: 797-801. Hill D, White V. Australian adult smoking prevalence in 1992. Aust J Public Health 1995; 19: 305-308. Chapman S, Wooward S. Australian court rules that passive smoking causes lung cancer, asthma attacks and respiratory disease. BMJ 1991; 302: 943-945. Schollam v Dept of Health (NSW). District Court (NSW). 25 May 1992, Case 40830/86. Borland R, Mullins R. The increasing prevalence of workplace smoking bans in Victoria. J Occup Health Safety Aust N Z 1994; 10: 35-40. Borland R, Morand M, Mullins R. Prevalence of workplace smoking bans in Victoria. Aust N Z J Public Health 1997; 21: 694-698. Doll R, Peto R, Wheatley K, et al. Mortality in relation to smoking: 40 years' observations on male British doctors. BMJ 1994; 309: 901-911. Rimpela A. Critical analysis of the Finnish Tobacco Act: implementation and legitimacy 1977-89. Tobacco Control 1992; 1: 285-292. (Received 11 Aug 1997, accepted 27 Jan, 1998) Authors' details Centre for Behavioural Research in Cancer, Anti-Cancer Council of Victoria, Melbourne, VIC. David J Hill, PhD, Director; Victoria M White, MA, Behavioural Scientist; Michelle M Scollo, BBSc, GradDipCommHlth, Public Health Consultant. Reprints will not be available from the authors. Correspondence: David J Hill, Director, Centre for Behavioural Research in Cancer, Anti-Cancer Council of Victoria, 1 Rathdowne Street, Carlton South, VIC 3053. E-mail: davidh AT accv.org.au Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". <URL: http://www.mja.com.au/>

David J Hill · Victoria M White · Michelle M Scollo

Prevalence of tuberculosis infection in Melbourne secondary school students Paul D R Johnson, John B Carlin, Catherine M Bennett, Peter D Phelan Michael Starr, Jane Hulls and Terry M Nolan MJA 1998; 168: 106-110 Abstract - Introduction - Methods - Statistical analysis - Results - Discussion - Acknowledgements - References - Authors' details - - ©MJA1998 Abstract Objective: To estimate the prevalence of asymptomatic Mycobacterium tuberculosis infection in Melbourne secondary school students. Design: Cross-sectional Mantoux testing of a partly random and partly targeted sample of secondary school students, designed to enable estimation of prevalence by region of birth. Setting: Fifty-one State and Catholic secondary schools in metropolitan Melbourne during 1995. Participants: Australian and overseas-born students in Years 9 and 10. Outcome measures: Proportions of students with positive Mantoux reactions (defined as induration at 48 hours of ≥5 mm with a history of recent exposure; ≥10 mm and no prior BCG vaccination; ≥15 mm and prior BCG vaccination). Results: Of 2586 students potentially eligible for testing, evaluable results were obtained from 1274 (49%). The overall prevalence of infection for Melbourne students in Years 9 and 10 was 2.5% (95% CI, 1.1-3.9%). Main predictors of a positive test were birth overseas and number of years residing overseas. Prevalence varied considerably by region of birth, and was very low in students born in Australia (0.7%), "other developed countries" (0.7%), and Southern Europe (0). The highest rates were observed in students born in Indochina (15.9%), other countries in South East Asia (10.2%), and Eastern Europe (10.2%). Conclusions: The risk of a young person becoming infected with M. tuberculosis while living in Melbourne is very low. Our results do not indicate a need for the reintroduction of mass screening in Victorian schools. If targeted screening were to be considered, the group most likely to benefit would be recently arrived migrants from Indochina. Introduction Human infection with Mycobacterium tuberculosis is usually clinically silent, and may then only be detectable by a positive Mantoux skin test. A minority of infected individuals develop active tuberculosis (TB), and those with pulmonary disease are the major source of new human infections. For M. tuberculosis to persist within a community over time, each person with pulmonary TB must infect an average of 20 others.1,2 In developed countries, the transmission rate is much lower than this, and the incidence of tuberculosis has declined steadily for many years.2 In the United States, a 32-year trend of declining TB notifications was reversed in the mid 1980s.3 This resurgence is thought to have resulted in at least 51 000 unexpected cases of the disease, and has been attributed to the impact of AIDS, urban decay, homelessness and high levels of migration from regions where TB is endemic.4 Recent increases in incidence have also been reported from other developed countries, including Denmark, Italy, the Netherlands, Spain, Switzerland, France and the United Kingdom.5 In Australia, the incidence of TB has remained constant during the 1990s (rates per 100 000 population of 5.95 in 1990 and 5.75 in 1995).6 In Victoria, the annual incidence of TB per 100 000 declined from 47 in 1954 to 6.2 in 1992;7 it was 6.35 in 1995.6 In 1970 40% of new cases of active tuberculosis in Victoria were in people born overseas, but by 1990 this figure had risen to 80%. Currently, most new TB patients are migrants from Indochina and South East Asia.7 Although intending adult migrants are screened by chest x-ray before their arrival in Australia, migrant children are generally not screened, partly because childhood tuberculosis is not normally transmissible. In a recent inner-Sydney study, 27% of foreign-born Year 8 students (159 of 580), compared with 2% of those born in Australia (20 of 1221), had positive Mantoux reactions.8 As adolescents appear to have an increased risk (compared with children over three years and adults) of developing active tuberculosis,9 this group may become a source of new, locally acquired infection. During 1995, we conducted a Mantoux survey of healthy metropolitan Melbourne secondary students in Years 9 and 10. Our aim was to estimate the prevalence of asymptomatic TB infection, and to identify specific groups of students by region of birth who may benefit from future targeted screening and intervention programs. Methods The study was a cross-sectional survey of secondary school students in the metropolitan region of Melbourne (population, 3.1 million, with people aged 12 to 17 years comprising approximately 8% [Australian Bureau of Statistics, 1996 Census]). To recruit enough overseas-born students for estimating prevalence by region of birth, we used a combination of targeted and random sampling, aiming to include schools where at least 4% of students were born overseas, plus a 5% random sample of all other schools. The Australian Bureau of Statistics (ABS) provided numbers and country of birth of Melbourne residents aged 12 to 17 years from the 1991 national census. The Directorate of School Education and the Catholic Education Office provided numbers of students at each school speaking a language other than English at home. These data were combined and used to select schools with high enrolments of overseas-born students. Independent schools were not included as they comprised only 18% of all Year 9 and Year 10 students (Directorate of School Education, personal communication), and we assumed that they would have a low proportion of recently arrived overseas-born students. Of 50 purposely selected and seven randomly selected schools approached, six and one, respectively, declined. The randomly selected school was replaced by a neighbouring school, so that 51 schools participated. Parents of all students in Years 9 and 10 at each participating school were provided with an explanatory letter and a short survey (in English and, if appropriate, one of 12 translations). The survey was part of a separate study on asthma, but included a question on country of birth and sought permission to approach students a second time for the TB study. The initial (asthma) survey was distributed to all 13 020 Year 9 and Year 10 students at the 51 schools; 9794 usable responses were obtained (75%). Respondents comprised 85% from purposely selected and 15% from randomly selected schools. From the returned surveys we created 12 notional region-of-birth groups, based partly on geography and partly on numbers of respondents to the first survey, and an additional group comprising Australian-born students with both parents born in Asia (Box 1). We aimed to enrol approximately 200 students from each group to allow us to estimate prevalence within each with a 95% confidence interval of ± 1.9% if the true prevalence were 2%, and ± 6.4% if the true prevalence were 30%. Selection for Mantoux testing was random, except when the number of respondents in a category was less than 200, in which case all respondents from that region were included. Parents of students selected for Mantoux testing were sent information letters, consent forms and a TB survey (in English and one of 14 translations). This survey sought parental consent for Mantoux testing and included questions on parents' country of birth, date of arrival in Australia and history of BCG vaccination. Mantoux testing (by two experienced nurses, with two assistants from the Victorian Tuberculosis Program) was performed at school. Responses on each TB survey were checked for completeness and each student's deltoids, forearms and scapulas inspected for the presence of BCG vaccination scars before testing, which involved intradermal injection of 0.1 mL of a 100 IU/mL solution of purified protein derivative (PPD; CSL Limited, Parkville, Vic.) to the volar aspect of the student's left forearm. A single batch of PPD was used throughout. At between 48 and 72 hours, the extent of transverse palpable induration was measured by ruler and recorded in millimetres. The study was approved by the Ethics in Human Research Committee of the Royal Children's Hospital. Statistical analysis We used the exact binomial method for confidence intervals and the chi-squared test for comparisons between groups. Logistic regression was used to estimate prevalences, adjusted for differences in duration of residence in Australia. As we had deliberately selected schools with high concentrations of foreign-born students, we estimated overall prevalence of infection by direct standardisation to the population distribution of region of birth in students aged 12 to 17 years in metropolitan Melbourne (1991 Census data, ABS). Results Of 2586 students (the results of our efforts to create the 13 groups) sent the TB survey, 620 (24.0%) did not return it, and 692 (27%) returned completed surveys but declined to be tested or were away on the day of testing. Test results were therefore available for 1274 students (49%). To explore the potential for response bias, survey response rates and acceptance of testing were compared between subgroups defined by age, sex and time since arrival in Australia (Box 2). Younger students and females were statistically more likely to return their surveys than older students and males (data not shown), but refusal to be tested did not vary between these subgroups. Proportions of students with Mantoux reactions that were positive by National Health and Medical Research Council criteria (≥5 mm with a history of recent exposure; ≥10 mm and no prior BCG vaccination; ≥15 mm and prior BCG vaccination8,10 ) were compared by region-of-birth group (Box 3). Birth overseas, number of years resident outside Australia (Box 4) and past BCG vaccination were predictors of a positive result. The crude prevalence was 5.3%. After standardisation by region of birth, we estimated that 2.5% of all students in Years 9 and 10 in metropolitan Melbourne had positive Mantoux results (95% CI, 1.1-3.9%). Students born in Australia, "other developed countries" and Southern Europe had the lowest rates (0.7%, 0.7% and 0, respectively). The highest rates were observed in students born in Indochina (15.9%), other countries in South East Asia (10.2%), and Eastern Europe (10.2%) (Box 3). Differences in mean number of years resident overseas between the groups prevented direct statistical comparison, so logistic regression was used to estimate the odds ratio for increase of risk for each year lived overseas (odds ratio per year, 1.15; 95% CI, 1.08-1.24), and to standardise rates by region so they could be compared directly. After this adjustment, statistically significant differences in prevalence by region persisted, but the ranking of some regions, most notably Eastern Europe, was altered (Box 3). Prevalence of TB in the group of Australian-born students with Asian-born parents was 3.6%. None of these students had received a prior BCG vaccination. This prevalence was fivefold higher than the background rate of 0.7% for other students born in Australia, but this difference did not reach statistical significance (P = 0.17, two-tailed Fisher's exact test). The 244 students with a result ≥10 mm and the two with a result ≥5 mm plus a history of exposure to someone known to have TB were referred to a special clinic at the Royal Children's Hospital for chest x-ray and clinical review. However, 10 students with indurations ≥10 mm declined to attend the clinic, five of whom had positive Mantoux reactions. Of the 236 students who attended the clinic, 174 had at least one BCG vaccination scar and/or documentary evidence of BCG vaccination and a Mantoux reaction < 15 mm; these were therefore considered to have negative results. Students with results positive by NHMRC criteria8,10 were offered isoniazid preventive therapy, provided there was no evidence of active disease and they had not been previously treated. Sixty-two students had positive Mantoux reactions. Of these, seven had previously been prescribed isoniazid preventive therapy, two had previously been treated for TB disease, and one (a recently arrived refugee from East Timor) had active pulmonary TB. Five students refused isoniazid therapy, and five were not offered therapy as they were considered to have a reduced risk (recent BCG). Forty-two students with positive Mantoux reactions were offered isoniazid preventive therapy; 38 (90%) completed six months of treatment. Multiple BCG scars (evidence of previous vaccinations) were common in students from Eastern Europe (mean number, 1.6; 33% of students with ≥2 scars) and the Middle East, whereas those from Indochina generally had only one (mean number, 0.9; 11% of students with ≥2 scars). Discussion The risk of a young person becoming infected with M. tuberculosis while living in Melbourne appears to be very low. The major determinant of the size of the Mantoux reaction in this study was birth overseas. However, most students with reactions to testing had also received at least one prior BCG vaccination, which complicates interpretation. Most Australian-born students had not been vaccinated, and the prevalence of infection in this group was very low (0.7%, giving a calculated annual risk of infection of 0.04% per year). The higher prevalence in Australian-born students with Asian-born parents may be the result of low-level transmission within migrant communities, although the apparent difference could have been due to chance. For those born overseas, the number of years spent outside Australia correlated positively with Mantoux results, indicating that the risk of infection increases with duration of residence in an endemic region. The prevalence of infection in Year 9 and Year 10 students in metropolitan Melbourne appeared to be approximately half that identified in a recent survey in Year 8 students in inner Sydney, both overall and within specific migrant subgroups.8 The authors of the Sydney study commented that the prevalence they identified was higher than previously reported in Australia and have since found a slightly lower prevalence in a further survey of younger students.11 Our lower rate may be the result of differing patterns of migration between Melbourne and Sydney, or the selection of our sample from the whole metropolitan region instead of just the inner city. Although only half of eligible students in our study were tested, we do not believe that our results are systematically biased in a way that would have led us to grossly underestimate prevalence. In particular, we found no association between recent arrival in Australia and the likelihood of refusing to be tested, and the group with the lowest participation rate (Eastern Europe) showed the second-highest prevalence. In retrospect, acceptance of testing may have been improved by selecting whole classes rather than individuals within a class for testing, and our two-stage study design allowed ample opportunity for students to withdraw. However, any future targeted screening program would need to select individuals from within a larger group, and part of our study rationale was to investigate the acceptability of such programs. Even if it were assumed that non-participants had twice the prevalence of those tested, the true prevalence would only be 50% greater than our estimates. There is controversy about the influence of BCG vaccination on the results of subsequent Mantoux testing. In countries with a high prevalence of TB infection, a single BCG vaccination is often given shortly after birth, but this is unlikely to influence the result of a Mantoux test 15 years later.12 By contrast, BCG given to older children or given several times during childhood probably does influence Mantoux reaction size.13 Although the current NHMRC guidelines make some allowance for past BCG vaccination, they could be further refined.13 For example, for recently arrived migrants who have received a single BCG vaccination early in life and who have lived for many years in a region of high prevalence it may be appropriate to use ≥10 mm to indicate a positive reaction, while for migrants from countries that routinely give three BCG vaccinations in childhood but have a lower prevalence of infection ≥20 mm may be more appropriate.13 Whether or not such refinements are introduced, the distinction between positive and negative reactions will remain somewhat arbitrary; a more reliable test is urgently required. Mass screening of secondary school students by Mantoux test was discontinued in Victoria over 10 years ago. Our results do not suggest that such programs need to be reintroduced, and recent overseas studies suggest that mass screening at school is unlikely to be cost effective.14-16 However, 24% of overseas-born students in our study had Mantoux reactions ≥10 mm, one of whom had active pulmonary disease and one-quarter of whom were considered eligible for isoniazid preventive therapy. If reactions of ≥10 mm for students born in a high-risk region with a history of having received a single BCG vaccination in infancy were considered positive, the number of students eligible for preventive therapy would increase further. If targeted screening were introduced, the group most likely to benefit would be recently arrived students from Indochina. We offered isoniazid therapy to students who tested positive because adolescents have an increased risk of developing active tuberculosis.9,17 However, we were mindful that isoniazid therapy is not entirely without risk even in young people,18-20 and that the risk of infection progressing to disease in an affluent society with a low prevalence of HIV infection may be much lower than the 10% often quoted.21 We cannot therefore be completely confident that wider use of isoniazid in this way would result in a net benefit to the Australian community. Acknowledgements This study was supported by a grant from the John Burge Estate administered by the Victorian Department of Human Services. We gratefully acknowledge the assistance of the principals, coordinators, teachers, students and parents at the participating schools. We also wish to thank the following individuals: Mary Randall, Mary McColl and staff of the Victorian Tuberculosis Program; Marita Dalton, Colin Powell, Department of Thoracic Medicine; and Susan Sawyer, Centre for Adolescent Health, Royal Children's Hospital. References Bates JH, Stead WW. The history of tuberculosis as a global epidemic. Med Clin North Am 1993; 77: 1205-1217. Styblo K. Recent advances in epidemiological research in tuberculosis. Tuberc Res 1980; 20: 1-63. Centers for Disease Control and Prevention. Tuberculosis morbidity -- United States, 1992. MMWR Morb Mortal Wkly Rep 1993; 42: 696-704. Snider DE Jr, Raviglione M, Kochi A. Global burden of tuberculosis. In: Bloom BR, editor. Tuberculosis: pathogenesis, protection and control. Washington DC: ASM Press, 1994: 3-12. Raviglione MC, Sudre P, Rieder HL, et al. Secular trends of tuberculosis in western Europe. Bull World Health Organ 1993; 71: 297-306. Oliver G, Harvey B. Tuberculosis notifications in Australia, 1995. Commun Dis Intell 1997; 21: 261-269. MacIntyre CR, Dwyer B, Streeton JA. The epidemiology of tuberculosis in Victoria. Med J Aust 1993; 159: 672-677. Alperstein G, Fett MJ, Reznik R, et al. The prevalence of tuberculosis infection among year 8 schoolchildren in inner Sydney in 1992. Med J Aust 1994; 160: 197-201. Comstock GW, Livesay VT, Woolpert SF. The prognosis of a positive tuberculin reaction in childhood and adolescence. Am J Epidemiol 1974; 99: 131-138. National Health and Medical Research Council. Tuberculosis in Australia and New Zealand into the 1990s. Canberra: AGPS, 1990. Alperstein G, Morgan KR, Fett MJ, et al. Prevalence of tuberculosis infection among primary school-entry children in Sydney. A N Z J Pub Health 1996; 20: 123-128. Menzies R, Vissandjee B. Effect of bacille Calmette-Guerin vaccination on tuberculin reactivity. Am Rev Resp Dis 1992; 145: 621-625. Ildirim I, Hacimustafaoglu M, Ediz B. Correlation of tuberculin induration with the number of Bacillus Calmette-GuŽrin vaccines. Ped Infect Dis J 1995; 14: 1060-1063. Driver CR, Valway SE, Cantwell MF, Onorato IM. Tuberculin skin test screening in schoolchildren in the United States. Pediatrics 1996; 98: 97-102. Mohle-Boetani JC, Miller B, Halper M, et al. School-based screening for tuberculous infection: a cost-benefit analysis. JAMA 1995; 274: 613-619. Starke JR. Universal screening for tuberculosis infection: school's out! JAMA 1995; 274: 652-653. Katz J, Kunofsky S. Logistics of chemoprophylaxis of tuberculosis. Chest 1971; 59: 600-605. Gal AA, Klatt EC. Fatal isoniazid hepatitis in a child. Ped Infect Dis J 1986; 5: 490-491. Israel HL, Gottlieb JE, Maddrey WC. Perspective: preventive isoniazid therapy and the liver. Chest 1992; 101: 1298-1301. Millard PS, Wilcosky TC, Reade-Christopher SJ, Weber DJ. Isoniazid-related fatal hepatitis. West J Med 1996; 164: 486-491. Haas DW, Des Prez RM. Mycobacterium tuberculosis . In: Mandell GL, Bennett JE, Dolin R, editors. Principles and practice of infectious diseases. 4th ed. New York: Churchill Livingstone, 1995: 2213-2243. (Received 23 May, accepted 27 Sep, 1997) Authors' details Clinical Epidemiology and Biostatistics Unit, Royal Children's Hospital, Melbourne, VIC. Paul D R Johnson, FRACP, PhD, Research Officer, Royal Children's Hospital Research Institute (also, Infectious Diseases Physician, Department of Infectious Diseases and Clinical Epidemiology, Monash Medical Centre); John B Carlin, BSc(Hons), PhD, Deputy Head (also, Associate Professor, University of Melbourne, Department of Paediatrics); Catherine M Bennett, BSc(Hons), Research Officer; Jane Hulls, RN, Research Nurse; Terry M Nolan, PhD, FRACP, Head (also, Associate Professor, University of Melbourne, Department of Paediatrics). University of Melbourne, Department of Paediatrics, Royal Children's Hospital, Melbourne, VIC. Peter D Phelan, MD, FRACP, Stevenson Professor, and Head (currently, Emeritus Professor of Paediatrics). Department of Microbiology and Infectious Diseases, Royal Children's Hospital, Melbourne, VIC. Michael Starr, MB BS, FRACP, Paediatrician. Reprints: Dr P D R Johnson, Department of Infectious Diseases and Clinical Epidemiology, Monash Medical Centre, Clayton, VIC 3168. E-mail: Paul. Johnson AT med.monash.edu.au Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". <URL: http://www.mja.com.au/>

Paul D R Johnson FRACP, PhD · John B Carlin · Catherine M Bennett · Peter D Phelan · Michael Starr · Jane Hulls · Terry M Nolan

Apolipoprotein screening in Australian children: feasibility and the effect of age, sex, and ethnicity Judith F Lynch, Michelle D Marshall, Xing L Wang and David E L Wilcken MJA 1998; 168: 61-64 Abstract - Introduction - Methods - Results - Discussion - References - Authors' details - - ©MJA1998 Abstract Objectives: (i) To evaluate the feasibility of detecting adverse lipid profiles in schoolchildren by measuring capillary dried blood spot apolipoprotein levels, and (ii) to assess the effect of age, sex and ethnicity on apolipoprotein levels. Design: We measured capillary dried blood spot apolipoproteins B and A-I (apo B and apo A-I); assessed levels in relation to age, sex and ethnicity; and recalled children with elevated levels for a full lipid profile measurement. Participants and setting: 6992 children (3501 boys and 3491 girls), aged 5-13 years, from schools in eastern Sydney, 1991-1995. Main outcome measures: Capillary blood levels of apolipoproteins B and A-I, and serum total cholesterol level. Results: Of the 6951 children who provided an adequate fingerprick blood sample, we recalled 1465 children (21.1%) (640 boys [43.7%] and 825 girls [56.3%]) with elevated apo B levels and/or apo B : apo A-I ratios for further testing, either by us or by their family doctor (overall estimated compliance rate up to 70%). Among the 458 children who returned to us, there was a 90% positive predictive value for a total cholesterol level of over 4.5 mmol/L in those with both elevated apo B levels and high apo B : apo A-I ratios. Girls had higher apo B levels and apo B : apo A-I ratios than boys (P < 0.00001 for both), and in both sexes there was a trend downwards for apo B and upwards for apo B : apo A-I ratio over the age range tested, but levels were relatively stable between the ages of 6 and 10 years. Indian children (1.5% of the screened population) had the highest apo B levels, followed by white children (71.1%); Asian children (9.2%) had the lowest (P < 0.00001 compared with Indian and white children). Conclusions: The high positive predictive value of capillary blood apolipoprotein levels for an adverse lipid profile in children suggests that measuring apolipoprotein levels by this method is a useful initial approach to cardiovascular risk assessment. Introduction Despite a recent decline in cardiac deaths, coronary disease is still the largest single cause of premature death in Australia.1 It is known that atherogenesis, the underlying pathological process, may begin in childhood;2,3 that relevant risk factors may track from childhood to adulthood (particularly elevated levels of total and low-density lipoprotein [LDL] cholesterol and increased body weight); and that these factors tend to aggregate within families.4,5 With these considerations in mind, we explored the feasibility of screening for adverse lipid profiles in a target group of primary schoolchildren, with the aim of implementing family-based coronary prevention by secondarily identifying any parents at risk. We measured levels of apolipoprotein B (apo B), the carrier protein for the atherogenic LDL cholesterol, and apolipoprotein A-I (apo A-I), the principal carrier protein for the antiatherogenic high density lipoprotein (HDL) cholesterol, in capillary dried blood spots from 6992 primary schoolchildren aged 5-13 years. There is increasing evidence that high apo B and low apo A-I levels are as reflective of cardiovascular risk as are their respective lipoproteins.6 We describe here the positive predictive value of high apo B and/or apo B : apo A-I ratio for elevated total cholesterol levels and the effects of age, sex and eth nicity on lipid levels in Australian children. Methods School and population demographics We received permission from 70 schools in eastern Sydney (public primary schools, 45%; Catholic primary schools, 29%; and private fee-paying schools, 26%) to contact parents and offer them the option of their child participating in a "Heart Health Screening Program". Children were given consent forms to take home to their parents. On the forms, we asked for the country of origin of each family, and this information was confirmed by a family history questionnaire (sent to all parents within a week of the testing at school) asking each parent to identify their country of origin. Only 1% of the total declined to answer this question (the "unknown" group) (Table 1). Blood collection and apolipoprotein measurement Testing took place between 1991 and 1995. From each child we obtained a capillary blood sample by fingerprick, which was spotted directly onto filter paper (No. 903: Schleicher & Schuell Inc., Keene, NH, USA), allowed to air dry and stored at 2 70¡C until analysis. Concentrations of dried blood spot apo B, apo A-I and lipoprotein (a) (Lp(a)) were measured within two weeks by our previously described enzyme-linked immunosorbent assay methods,7-9 and from these results an apo B : apo A-I ratio was also calculated. These methods allowed us to carry out all three measurements on one small (minimum, 20 µL sample) capillary dried blood spot. We report here the results obtained for measurement of apo B and apo A-I only. Recall In the first school tested, we established the frequency distribution for apo B levels and apo B : apo A-I ratios and set the level for recall for each of these as at or above the 90th percentile. These cut-off points were adjusted from the apo lipoprotein results obtained from each of the next seven schools. However, a stable cut-off level was reached by the third school. We defined three categories of children for recall: those with an ele vated apo B level only (Criterion 1), those with a combination of elevated apo B level and apo B : apo A-I ratio (Criterion 2), and finally those with an elevated apo B : apo A-I ratio only (Criterion 3). We notified the parents of the children for recall by letter, explaining the results and suggesting that both parents, the child, and any other siblings might come to us for further lipid testing, or attend their family doctor. In those who returned to us, we measured total cholesterol level, as well as HDL cholesterol and triglyceride levels, in a venous blood sample by standard methods. LDL cholesterol level was calculated by the Friedewald formula,10 and we also measured serum apo B, apo A-I and Lp(a) by the methods outlined above.7-9 The parents of children with apo B levels and apo B : apo A-I ratios below the established cut-off points were also sent a letter explaining that their children's levels currently fell below the recall criteria. These families were sent a healthy lifestyle pamphlet specifically directed to maintaining a low level of cardiovascular risk; they were also asked if they would like to volunteer as a healthy control family and have complete lipid profiles measured as above. Statistical analysis As the distributions of age, apo B and apo A-I levels, and apo B : apo A-I ratios were normal, analyses by parametric methods were appropriate. We used one-way analysis of variance (ANOVA) for comparisons of measured quantitative variables when there were more than two groups; for example, when children were divided according to high apo B, high apo B : apo A-I ratio, or both, as categorical variables, and levels of total cholesterol and LDL cholesterol, and total cholesterol to HDL cholesterol ratio as the continuous outcome variables. When the comparison was between two groups -- for example, when lipid values were compared in girls and boys as categorical determinants -- Student's t test was used.11 We employed a chi-squared comparison for assessing associations between categorical variables, and used the SPSS statistical software package for all analyses.12 Ethical approval The study was approved by the ethics committee of the University of New South Wales. Results Initial testing The parental consent rate to initial testing was 63.9%. There were 6992 children aged 5-13 years available for testing, 3501 (50.1%) boys and 3491 (49.9%) girls. Only 41 children (0.6%) failed to provide an adequate finger-prick blood sample on the day, leaving a total population of 6951 children tested. Recall population The cut-off point for apo B level was set at 0.550 g/L or more and for the apo B : apo A-I ratio 1.200 or more. We recalled 1465 children (or 21.1% of the population tested). Of these recalls, 65.3% were for elevated apo B only (Criterion 1), 19.6% for elevated apo B and apo B : apo A-I ratio (Criterion 2), and 15.1% for elevated apo B : apo A-I ratio only (Criterion 3). In all, 458 of the children recalled returned with their families to our laboratory for lipid studies (31.2% compliance rate). To obtain an assessment of how many may have been seen by their family doctor (the other option we suggested), we contacted 205 of the remaining 1007 families. Among these, 118 had seen their family doctor for follow-up, and 89 had either moved or were not interested in further testing. This suggested an overall compliance rate of about 70%, if this sample was representative. Of the non-recalled "normal" population of 5486 (78.9%), only 151 families (2.7%) volunteered for lipid testing, an insufficient number for meaningful statistical analysis. Predictive value We used the results of the 458 children who returned for a full lipid profile to assess the positive predictive value of our screening method. The National Heart Foundation of Australia recommends a total cholesterol level of 4.5 mmol/L as the upper limit for children in this age range.13 With a level of over 4.5 mmol/L as the endpoint, Criterion 1 (elevated apo B only) had a positive predictive value of 84% (95% confidence interval [CI], 80%-88%), and with Criterion 2 (elevated apo B level and apo B : apo A-I ratio combined) the predictive value increased to 90% (95% CI, 85%-95%). However, Criterion 3 (elevated apo B : apo A-I ratio only) yielded only 66% (95% CI, 54%-77%), possibly reflecting children with low HDL cholesterol and normal total cholesterol levels. Raising the cut-off point did not substantially improve the positive predictive value of our screening method. When the cut-off point for apo B was increased from 0.550 g/L to 0.700 g/L, for example, the positive predictive value of Criterion 2, the most strongly predictive, only increased to 94% (95% CI, 91%-97%). ANOVA confirmed the association between high apo B levels from analysis of dried blood spot and an adverse lipid profile, and also that a combined high apo B level and apo B : apo A-I ratio was a stronger predictor. In this analysis, criteria established by high apo B, high apo B : apo A-I ratio and a combination of both were determinant categorical variables, and the total cholesterol and LDL cholesterol levels and total chol es terol to HDL cholesterol ratio were the continuous outcomes. Total (P < 0.0006) and LDL cholesterol (P < 0.0014) levels and the total cholesterol to HDL cholesterol ratio (P < 0.0006) were all significantly higher in Criterion 2 than in Criterion 1 or Criterion 3 (Table 2). Sex, age and ethnicity There was a highly significant difference in the apo B and apo B : apo A-I ratio levels between boys and girls (P < 0.00001 for each) (Figures 1 and 2), but no difference (P = 0.6) in apo A-I levels (data not shown). The girls had consistently higher apo B levels and apo B : apo A-I ratios over the age range studied (Figures 1 and 2). This resulted in a significant difference in the recall rate between girls and boys (chi-squared = 23.362; P < 0.00001). This sex difference was also reflected in the total cholesterol levels of children returning for further lipid tests (data not shown). As shown in Figures 1 and 2, there was a downward trend in apo B level and an upward one in apo B : apo A-I ratio in boys over the age range screened, while the opposite was observed in girls; but levels were relatively consistent in both between the ages of 6 and 10 years. Most of the 6951 children screened were white (71.1%); the remainder were of diverse ethnic origins (Table 1). The mixed-race group (8.3% of the total) included white/Asian (28.2%), white/Arabic (17.8%), white/South American (10.7%), and other combinations (56.7%). As age was a significant contributor to apo B concentrations, we compared the "age adjusted apo B levels" for both boys and girls to assess potential differences between major ethnic groups (Figure 3). Indian children had the highest apo B levels, followed by white children. Asian children had the lowest levels and these were significantly different from those of the Indian, white, mixed-race and South American children (P < 0.00001 for each). We did not include black, Australian Aboriginal or Pacific Islander ethnic groups in this analysis because of small numbers and highly skewed age distributions. In all ethnic groups, girls had higher levels than boys (Figure 3). Discussion We have established previously that apolipoprotein measurements provide a convenient and effective approach to the detection of dyslipidaemia,14 and that apo B levels in children are correlated with the occurrence of coronary events in their grandparents, which highlights the relevance of measurements in children to assessing risk of vascular disease in older family members.15 We have also established a clear-cut association between increased apo B levels and apo B : apo A-I ratios and body mass index in Australian children.16 Here, we extend these find ings by demonstrating the feasibility of screening of schoolchildren's capillary blood apolipoprotein levels as an approach to family-based primary coronary prevention. Although we do not have adequate data to determine the sensitivity or specificity of our screening method, the presence of both an elevated apo B level and raised apo B : apo A-I ratios had a positive predictive value of 90% for the detection of elevated total serum cholesterol level. Raising the apolipoprotein cut-off points only improved the predictive value to 94%. This method identified elevated apolipoprotein levels in 21.1% of our population, of whom 90% had elevated total cholesterol levels. Screening for apolipoprotein levels may be more relevant than screening for total cholesterol levels alone, as total cholesterol level does not show the interrelations between levels of the atherogenic LDL and the antiatherogenic HDL cholesterol.17 In a small number of children (0.48%) apo B levels were elevated without elevation of total cholesterol level.18 These families were also provided with dietary and lifestyle advice as this may also be associated with increased cardio vascular risk. Our findings clearly show highly significant differences between boys and girls at this age, with girls having higher apo B levels and apo B : apo A-I ratios than boys. We suspect this biological difference, which was evident before the onset of puberty in both boys and girls, to be hormonally based, although we have no data to support this. The vari ation in apolipoprotein levels within the age range studied could also be hormonally based. Whatever the mech anisms, these same age-related sex differences have been identified in several other studies in children.19,20 The consistency of these results clearly indicates a need to have different cut-off points in boys and girls for elevated apolipoprotein levels. Our results also define the age range (between 6 and 10 years) when levels are most stable and most appropriate for screening, findings which will be incorporated into future studies. There were uniform sex-related differences in apolipoprotein levels in each of the racial groups in our population, as well as very significant differences in levels between the various ethnic groups, as has also been found in previous studies.21,22 In our population, Indian and white children had the highest apo B and apo B : apo A-I ratios and Asian children had the lowest. It is well established that Chinese and Japanese populations have much lower cholesterol levels than whites and a correspondingly lower prevalence of coronary artery disease. Indian populations have a higher prevalence of cardiovascular disease.23 While these variations may relate to the genetic background of each ethnic group, diet and lifestyle undoubtedly make major contributions, and this is particularly evident in non-Western groups who have moved to live in a Western society. The higher prevalence of coronary disease in Indians living in the United Kingdom, for example, is well documented.24 Ethnic differences may become blurred with time if dietary habits become more uniform, and this has occurred among Asians emigrating to the United States and the United Kingdom; within a generation they acquired local lifestyles and a correspondingly higher prevalence of coronary artery disease than that in the communities they had left.25,26 In conclusion, our study demonstrates the feasibility of conducting a program of measuring apolipoprotein levels in capillary blood samples to assess lipid profiles in children to facilitate family-based coronary prevention. It documents wide acceptance by both schools and parents. Implicit in such a study is the potential for a multiplier effect in that, if a child has an elevated apolipoprotein level and therefore an adverse lipid profile, it is very likely that this will also be seen in at least one parent and other siblings.27 Our ongoing screening program requires the establishment of a concurrent intervention program to improve dietary and lifestyle habits of affected children and their parents. This is of more immediate relevance to affected parents as they are approaching the age of overt coronary disease. However, the program provides an ideal opportunity to establish healthy lifestyles in young children at a time of easy acceptance, and with the potential for preventing a disorder that may have its origins in childhood.28 References Heart and Stroke Facts. A report by the National Heart Foundation of Australia. Canberra: NHFA, 1996. Stary HC. Evolution and progression of atherosclerotic lesions in coronary arteries of children and young adults. Arteriosclerosis 1989; 9(1 Suppl): I19-I32. Tracy RE, Newman WP, Wattigney WA, Berenson GS. Risk factors and atherosclerosis in youth autopsy findings of the Bogalusa Heart Study. Am J Med Sci 1995; 310(Suppl 1): S37-S41. Berenson GS, Wattigney WA, Bao W, et al. Rationale to study the early natural history of heart disease: the Bogalusa Heart Study [review]. Am J Med Sci 1995; 310(Suppl 1): S22-S28. Sanchez-Bayle M, Gonzalez-Requjo A, Ruiz-Jarabo C, et al. Serum lipids and lipoproteins in Spanish children and adolescents: a 5 year follow-up. Acta Paediatr 1996; 85: 292-294. Bao W, Srinivassan SR, Berenson GS. Tracking of serum apolipoproteins A-I and B in children and young adults: the Bogalusa Heart Study. J Clin Epidemiol 1993; 46: 609-616. Wang XL, Dudman NP, Wilcken DE. Enzyme-linked immunosorbent assay of apolipoprotein B in blood spotted onto filter paper, suitable for neonatal screening. Clin Chem 1989; 35: 1000-1004. Wang XL, Dudman NP, Blades BL, Wilcken DE. Changes in the immunoreactivity of apo A-I during storage. Clin Chem 1989; 179: 285-293. Wang XL, Wilcken DE, Dudman NP. An indirect sandwich ELISA for Lp(a) in serum and dried blood spots. Clin Chim Acta 1992; 207: 73-86. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972; 18: 499-502. Norusis MJ. General factorial analysis of variance. SPSS for Windows Advanced Statistics, release 6.0. Chicago: SPSS Inc, 1993: 31-56. McNeil D. Statistical methods II: one-way analysis of variance. In: McNeil D, editor. Epidemiological research methods. New York: John Wiley & Sons, 1996: 66-71. National Diet and Heart Disease Advisory Committee, National Heart Foundation of Australia. Guide to plasma lipids for doctors. Curr Therap 1992; 10(Suppl 1): S1-S8. Wang XL, Wilcken DE, Dudman NP. Apolipoprotein A- I and B and the B/A-I ratio in the first year of life. Pediatr Res 1991; 30: 544-549. Wilcken DEL, Wang XL, Greenwood J, Lynch JF. Lipoprotein (a) and apolipoproteins B and A-I in children and coronary vascular events in their grandparents. J Pediatr 1993; 123: 519-526. Wilcken DE, Lynch JF, Marshall MD, et al. Relevance of body weight to apolipoprotein levels in Australian children. Med J Aust 1996; 164: 22-25. Taylor CJ, Olpin S, Rattenbury J, et al. Familial hyper cholesterolaemia: pilot study to identify children at risk. J Clin Pathol 1993; 46: 730-733. Grundy S. Small LDL. Atherogenic dyslipidaemia and the metabolic syndrome. Circulation 1997; 95: 1-4. Resnicow K, Morley-Kotchen J, Wynder E. Plasma cholesterol levels of 6585 children in the United States: results of the know your body screening in five states. Pediatrics 1989; 84: 969-976. Christensen B, Glueck C, Kwiterovich P, et al. Plasma cholesterol and triglyceride distributions in 13,665 children and adolescents: the Prevalence Study of the Lipid Research Clinics Program. Pediatr Res 1980; 14: 194-202. Zhang W, Evans AE, Cambien F, et al. Distribution of lipid variables in subjects in Belfast, Northern Ireland and Taiyuan, PR China. Atherosclerosis 1993; 102: 175-180. Freedman DS, Lee SL, Byers T, et al. Serum cholesterol levels in a multiracial sample of 7,439 preschool children from Arizona. Prev Med 1992; 21: 162-176. Enas EA, Mehta J. Malignant coronary artery disease in young Asian Indians: thoughts on pathogenesis, prevention, and therapy. Coronary Artery Disease in Asian Indians (CADI) Study [review]. Clin Cardiol 1995; 18: 131-135. Bhatnagar D, Anand IS, Durrington PN, et al. Coronary risk factors in people from the Indian subcontinent living in west London and their siblings in India. Lancet 1995; 345: 405-409. Egusa G, Murakami F, Ito C, et al. Westernized food habits and concentration of serum lipids in the Japanese. Atherosclerosis 1993; 100: 249-255. Robinson D, Kawamura T, Hinohara S, Sakamoto Y. Levels of cardiovascular risk factors in Japanese people living in the UK. J Cardiovasc Risk 1995; 2: 449-458. Shaukat N, de Bono DP, Jones DR. Like father like son? Sons of patients of European or Indian origin with coronary artery disease reflect their parents' risk factor patterns. Br Heart J 1995; 74: 318-323. Resnicow K, Cross D, Lacosse J, Nichols P. Evaluation of a school-site cardiovascular risk factor screening intervention. Prev Med 1993; 22: 838-856. (Received 15 Apr, accepted 15 Oct, 1997) Authors' details Department of Cardiovascular Medicine, Prince Henry and Prince of Wales Hospitals, Sydney; and Community Health Services and Programs, South Eastern Sydney Area Health Service, Royal South Sydney Hospital, Sydney, NSW. Judith F Lynch, HTech, Heart Health Education Program Co-ordinator; Michelle D Marshall, BSc, Scientific Officer; Xing L Wang, PhD, Research Fellow; David E L Wilcken , MD, FRACP, Visiting Professor of Medicine. Reprints will not be available from the authors. Correspondence: Professor D E L Wilcken, Department of Cardiovascular Medicine, Room 163, Clinical Sciences Building, Prince Henry Hospital, Little Bay, Sydney, NSW 2036. Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". <URL: http://www.mja.com.au/>

Judith F Lynch · Michelle D Marshall · Xing L Wang

The impact of catalytic converters on motor vehicle exhaust gas suicides Virginia H Routley and Joan Ozanne-Smith MJA 1998; 168: 65-67 Abstract - Introduction - Methods - Results - Discussion - Acknowledgments - References - Authors' details - - ©MJA1998 Abstract Objective: To examine the trends in motor vehicle exhaust gas suicides since 1970 and to investigate the impact of catalytic converters. Design: Australia-wide database analyses and a retrospective stratified series of 100 Victorian cases. Data sources: Australian Bureau of Statistics, 1970-1995; Australian Institute of Health and Welfare, National Injury Surveillance Unit, 1991/92-1995/96; Victorian Coroner's files, 1994-1996. Results: There were 509 motor vehicle exhaust gas suicides in Australia in 1995, representing 22% of total suicides. Since the 1986 requirements for reduced carbon monoxide emissions from new vehicles (and thus the use of catalytic converters), the absolute numbers and rates of such suicides have increased, and they have come to represent a larger percentage of total suicides. Of 75 Victorian victims' vehicles traced, 36% were manufactured during or after 1986, showing that exhaust gas suicides have occurred in vehicles with catalytic converters. Blood carboxyhaemoglobin levels did not differ between victims using vehicles with or without catalytic converters. Between 1976 and 1991 exhaust gas suicides increased at a faster rate than motor vehicle registrations. Australian hospital admissions for exhaust gas suicide attempts have increased substantially since 1991-1992. Conclusion: Catalytic converters and the associated lower CO emission limits of 9.3 g/km had not, by 1995, resulted in a reduction in numbers, rates or percentages of exhaust gas suicides in Australia. Introduction Suicide is a major problem in Australia, with approximately 2000 deaths from suicide annually in recent years. Since 1990 suicide has become a more common cause of death than motor vehicle crashes. In 1995, suicide by motor vehicle exhaust gas accounted for 509 deaths and was second only to hanging. It was most commonly used by males (85%), especially those aged 20-50 years. Rates of motor vehicle exhaust gas suicide per 100 000 population in Australia were 4.69 for males, 1.02 for females and 2.85 total (Australian Institute for Suicide Research and Prevention, 1996; unpublished data). Further, motor vehicle exhaust gas related self-harm represented a smaller proportion of patients hospitalised for attempted suicide than of successful suicides (2.0% v. 21.6%; Australian Institute of Health and Welfare, National Injury Surveillance Unit, 1996; unpublished data), indicating the high lethality of this means of suicide. Motor vehicle availability has been found to be related to exhaust gas suicides,1 and in Australia the rates of such suicides in different States vary to some extent with motor vehicle registration rates. For example, the Northern Territory has the lowest rates for each, while Western Australia has relatively high rates (Australian Institute of Health and Welfare, National Injury Surveillance Unit, 1996; unpublished data).2 Carbon monoxide (CO), the most toxic component of motor vehicle exhaust gas, is colourless, odourless and tasteless and is produced from the incomplete combustion of organic fuels. Mild CO poisoning produces headache and drowsiness; more severe poisoning leads to collapse, coma and death.3 Other effects can include myocardial infarction, deterioration of personality and impaired memory.4,5 Environmental regulations have required carbon monoxide emission levels from vehicle exhausts to be reduced over past decades. The required maximum levels have been 24.3 g/km from July 1976 (Australian Design Rule [ADR] 27A), 9.3 g/km from 1986 for new passenger vehicles (ADR 37-00) and 2.1 g/km for new models (ADR 37-01) from 1997 and for all new passenger vehicles from 1998. To cope with the unleaded petrol legally required since 1986 to reduce environmental air pollution from exhaust emissions, vehicles usually require catalytic converters (ADR 37-00; Australian Standard 2877). These convert carbon monoxide and other pollutants into the by- products, carbon dioxide and water. The purposes of this study were to examine overall trends in exhaust gas suicides, and to investigate the impact of the 1986 reduction in allowable carbon monoxide emissions (or the effects of catalytic converters) on that suicide rate. Methods To examine the trends in exhaust gas suicide between 1970 and 1995, we analysed Australia-wide data for mortality from motor vehicle exhaust gas suicides and for total suicides by age, sex and State of residence obtained from the Australian Bureau of Statistics (ABS). We examined Australian hospital admissions data from the Australian Institute of Health and Welfare, National Injury Surveillance Unit for admissions related to motor vehicle exhaust gas suicide attempts. Rates for suicide by motor vehicle exhaust gas were obtained from the Australian Institute for Suicide Research and Prevention. Data on vehicle numbers, rates per head of population and year of manufacture were obtained from ABS publications. Australian exhaust gas suicides and motor vehicle registration rates were compared for the ABS Motor Vehicle Census years since 1970. To examine the effects of catalytic converters, we undertook a retrospective stratified case series study of 100 exhaust gas suicides in Victoria between 1 January 1994 and 31 December 1996, noting the year of manufacture of the motor vehicles involved and victims' blood levels of carboxyhaemoglobin (COHb) -- a reflection of carbon monoxide levels. We compared COHb levels in victims whose suicides involved vehicles manufactured before 1986 with those whose vehicles were manufactured from 1986 (when CO emission levels were reduced). Cases of exhaust gas suicide in 1994 were obtained from a Victorian State Coroner's Office publication,6 while those in 1995 and 1996 were obtained from the Victorian Coroner's Facilitation System database. Results Since 1975 there have been approximately 7000 motor vehicle exhaust gas suicides in Australia. Suicides show an increasing trend, with dips in 1979 and 1993 (R2 = 0.58), while motor vehicle registrations follow an exponential trend (R2 = 0.88), showing a steady rise (Figure 1). Between 1979 and 1991 such suicides increased at a faster rate than motor vehicle registrations. Motor vehicle exhaust gas suicides have steadily increased as a percentage of total suicides since at least 1970 (7% in 1970, 13% in 1980 and 22% in 1995). Hospital admissions for Australian motor vehicle exhaust gas suicide attempts have increased steadily in recent years (1994/95 data are not available), with the data showing an exponential trend (R2 = 0.99). Deaths also increased, but show no clear trend (Figure 2). The 100 cases of exhaust gas suicide obtained from the records of the Victorian Coroner comprised 33 of a total of 96 in 1994, 33 of 140 in 1995, and 34 of 136 in 1996. The year of vehicle manufacture had been recorded for only 75 of these cases (25 in each of 1994, 1995 and 1996). Of these 75, 20% in 1994, 56% in 1995 and 32% in 1996 involved vehicles manufactured in 1986 or later. In total, 27 of the 75 vehicles (36%) were 1986 models or later. This proportion is not significantly different (P = 0.62) from the 39% of vehicles manufactured in 1986 or later in the total Victorian fleet of 2 799 310 vehicles in May 1995 (ABS 1995 Motor Vehicle Census).2 For the 74 cases in which both the victim's COHb level and the age of the vehicle was known, 26 involved vehicles manufactured in 1986 or later and 48 involved earlier models. The median COHb level of victims in these two groups was identical (77.1%). Discussion Motor vehicle exhaust gas suicide is a major public health problem in Australia, accounting for 22% of suicides in 1995. Despite the introduction of catalytic converters and the 1986 reduction in CO exhaust limits, motor vehicle exhaust gas suicides have increased in number, rate and as a proportion of total suicides since 1986. Hospitalisations, however, have doubled, suggesting an increase in failed attempts, and the rate curve for successful suicides also appears to be flattening in the 1990s. In 1995, 43% of registered Australian motor vehicles were manufactured in 1986 or later.2 If exhaust gas suicide had been made impossible by catalytic converters and the 1986 reduced CO emission standard of 9.3 g/km, then rates for such suicides should have reduced from 2.08 to about 1.19 per 100 000 population between 1985 and 1995 (assuming a similar distribution of vehicles to that of the Australian fleet and an unchanged number of attempts). In fact, rates per 100 000 population have been about 2.5 for the past five years and peaked at 2.85 in 1995. It is clear from our investigation of the sample from the Victorian Coroner's files that vehicles manufactured since 1986 have been used in exhaust gas suicides. Our finding that there was no difference in median COHb levels between victims who used vehicles with and without catalytic converters is consistent with the results of a New South Wales study.7 Given that CO emissions from vehicles with catalytic converters are lower, it can be deduced that time (and possibly other factors) need further investigation. The few overseas studies which have examined the relationship between suicides and motor vehicle exhaust gas have found that imposing emission controls reduces the incidence of such suicides.8-10 Vehicle emission limits for CO in the US have been lower than those in Australia, being at our 1986 limit (9.3 g/km) as early as 1975 and at just under our 1997 limit (2.1 g/km) in 1981.11 Thus, US trends should shed some light on the Australian situation, especially as the test conditions are similar. However, it should be noted that the average age of vehicles in the US is seven years, compared with 11 years in Australia, which suggests that there would be a longer lag time for exhaust emission changes to affect suicide rates in Australia. Motor vehicle exhaust gas is a less popular method of suicide in the United States than in Australia (1991 rates per 100 000 population being 0.73, compared with 2.78 for Australia).12 Data available for the US to 1991 show a decline in the exhaust gas suicide rate from the years 1963 to 1979, an increase from 1981 to 1987, followed by a decline to below the pre-1982 rates. Exhaust gas suicides in the US have decreased from 8.9% of suicides in 1970 to 7.4% in 1980 and 5.6% in 1991.8,12 Despite some conflicting evidence in Australia and the US as to the reductions in CO exhaust emission levels ne cessary for influencing exhaust gas suicide rates, these emission reductions do not appear to have produced the success suggested by early reports in the international literature.8-10 There are a variety of possible explanations for this. Firstly, exhaust gas suicides usually involve a hose or pipe being used to feed exhaust gas from an idling vehicle into the sealed interior of a vehicle. Clearly, CO emission levels set for environmental reasons may not be particularly relevant in this situation, and if suicides are to be prevented it would be more appropriate to set CO emission limits in relation to each vehicle's cabin volume. Further, testing vehicles for compliance with environmental standards involves three phases, none of which involve engine idling (during which CO emission could be higher) only. Secondly, CO emissions from vehicles with catalytic converters may exceed the legislated limit. Specific examples are: the engine idles from a cold start, with a delay of 1.5-3 minutes before the catalytic converter has warmed up and is operating efficiently (as blood COHb concentration rises most rapidly when first exposed to CO the initial absorption rate would be particularly high3 ); the condition of the catalytic converter has deteriorated or the engine may require tuning (under ADR 37-00 a catalytic converter is required to operate effectively for 80 000 km or five years, whichever occurs first); and for reasons that are unclear, the pollution performance of 4-9-year-old cars has been found to deteriorate faster than that of older cars.13 A 1989 US report noted that, even with reduced CO emission limits and catalytic converters, it is still possible to use car exhaust for suicide.10 If exposure is prolonged, residual CO content would eventually cause death, or suffocation might occur. Also, destruction of the engine management system, by the use of leaded petrol and neglect, or physical disconnection of the engine management system, could increase the CO content, and thus the lethality of this method for suicide.8 Further research needs to be undertaken in this technically complex area. A study of the vehicles used for unsuccessful exhaust gas suicide attempts and of other factors contributing to their failure would assist in understanding the relationship between vehicles and exhaust gas suicides. New regulations in Australia have not yet reduced suicides from motor vehicle exhaust gas. There is a need to monitor the situation, especially in the light of the 1997 reduction in CO to 2.1 g/km, and to review the regulations, vehicle design and testing methods. Acknowledgments The study was funded by the Victorian Health Promotion Foundation through its funding of the Victorian Injury Surveillance System, a project of Monash University Accident Research Centre. We thank the Victorian State Coroner's Office for providing access to relevant records, Associate Professor David Ranson (Victorian Institute of Forensic Medicine), David Lester (Richard Stockton College of New Jersey), Jerry Moller and Stan Bordeaux (Australian Institute of Health and Welfare, National Injury Surveillance Unit) for providing data, Christine Chesterman and Voula Stathakis (Monash University Accident Research Centre) for assisting with data collection and analysis, respectively, and Dr Ella Sugo (formerly of the NSW Institute of Forensic Medicine), Brian Hobsbawn (Environment Australia) and Jerry Moller for providing valuable comment. References Lester D. Car ownership and suicide by car exhaust in nations of the world. Percept Motor Skills 1994; 79: 898. Australian Bureau of Statistics. May 1995, Motor Vehicle Census. Canberra: ABS, 1996. (Catalogue No. 9309.0.) Ernsting J, King P. Aviation medicine. 2nd ed. London: Butterworths, 1988. Willis Hurst J, Schlant R, Rackley C, et al. The heart. 7th ed. New York: McGraw-Hill, 1990. Smith JS, Brandon S. Morbidity from acute carbon monoxide poisoning at three-year follow-up. BMJ 1973; 1: 318-321. Unnatural deaths, 1993/94. Melbourne: Victorian State Coroner's Office, December 1995. Sugo E, Duflou J, Sercombe J, Brown J. Suicidal inhalation of carbon monoxide -- a reappraisal of variables affecting lethal levels. Paper presented at the Annual Scientific Meeting of the Royal College of Pathologists of Australasia;1996 Sep 16-19; Sydney. Sydney: The College, 1996. Lester D. Changing rates of suicide by car exhaust in men and women in the United States after car exhaust was detoxified. Crisis 1989; 10: 164-168. Clarke R, Lester D. Toxicity of car exhausts and opportunity for suicide: comparison between Britain and the USA. J Epidemiol Community Health 1987; 41: 114-120. Lester D, Abe K. Car availability, exhaust toxicity and suicide. Ann Clin Psychiatry 1989; 1: 247-250. Code of Federal Regulations. Title 40: Protection of environment. Part 86 -- control of air pollution from new and in-use motor vehicles and new and in-use motor vehicle engines: certification and test procedures. Washington, DC: Office of the Federal Register National Archives and Records Administration, 1997. Centers for Disease Control and Prevention, National Center for Health Statistics. Vital statistics of the US. Vol. II. Mortality. Part A. Washington: Public Health Service, May 1996. Federal Office of Road Safety. Motor vehicle pollution in Australia. Report on the national in-service vehicle emission study. Canberra: AGPS, 1996. (Received 18 Mar, accepted 27 Sep, 1997) Authors' details Monash University Accident Research Centre, Melbourne, VIC. Virginia H Routley, BEc, GradDipSocStats, Research Fellow; Joan Ozanne-Smith, MB BS, MPH, Professorial Fellow. No reprints will be available. Correspondence: Ms V H Routley, Monash University Accident Research Centre, Wellington Road, Clayton, VIC 3168. Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". <URL: http://www.mja.com.au/>

Virginia H Routley · Joan Ozanne-Smith

Breast cancer mortality trends in Australia: 1921 to 1994 Catherine L Smith, Anne Kricker, Bruce K Armstrong MJA 1998; 168: 11-14 Abstract - Introduction - Methods - Results - Discussion - Conclusions - Acknowledgements - References - Authors' details - - ©MJA1997 Abstract Objective: To analyse breast cancer mortality trends in Australia and to see if mammographic screening has yet led to a reduction in mortality. Design: Retrospective analysis of trends in mortality rates from breast cancer in Australian women between 1921 and 1994, and in potentially explanatory variables such as fertility, body size, age at menarche, and screening. Results: Changes in breast cancer mortality in Australian women could not be explained by chance variation alone. Mortality rose steadily (average annual increase, 1.0%) to 1940-1944, fell to the 1960s and early 1970s, and rose (average annual increase, 0.3%) to the late 1980s. Between 1985-1989 and 1990-1994, breast cancer mortality fell by 3.2% in women 50-69 years of age (the target age group for mammographic screening) and by 4.2% in women 25-49 years of age. There was almost no change ( - 0.2%) in breast cancer mortality in older women in this period. The proportion of women screened in all age groups increased substantially between 1988 and 1994; nearly 65% of women in the target age group had had at least one mammogram by 1994. Decreases in fertility were followed by increases in mortality, and vice versa. Conclusions: Trends in breast cancer mortality have probably been influenced by changing fertility, nutrition and body-size increases among Australian women. Improvements in stage at diagnosis and treatment have probably moderated the upwards pressure on mortality caused by an increasing incidence. Recent falls in mortality could be expected to continue as more women participate in the mammographic screening program. This trend should be more clearly evident in the second half of the 1990s. Introduction Breast cancer mortality in women in most developed countries, including Australia, has levelled off or fallen in recent years.1 In Britain, an important fall in mortality since 1989, considered too early to be the result of mammographic screening, is thought to be the result of improved treatment.2 Similar observations have been made in the United States.3 We examined trends in breast cancer mortality and breast cancer risk factors in Australia between 1921 and 1994 to see if we could explain recent mortality changes and, in particular, to see if mammographic screening has yet led to decreased breast cancer mortality in the target age group for screening (50-69 years). Methods Age-specific and age-standardised mortality rates from breast cancer per 100 000 woman-years, standardised to the world population, were calculated for one four-year period (1921-1924) and 14 five-year periods (1925-1929 to 1990-1994) from annual mortality and population data from the Australian Bureau of Statistics. Rates for 1995 have subsequently been released. Estimates of the number of women who had had at least one mammogram by the end of each year from 1984 to 1994 were obtained as follows: Annual numbers of women by age having a first or subsequent bilateral mammogram (bilateral mammograms being most likely for screening purposes) under Medicare from 1984 to 1994 were estimated from data of the Commonwealth Department of Health and Family Services; Annual numbers of first screens in 10 pilot projects of the national screening program were estimated by age from 1 January 1988 to 30 June 19914-7 (BreastScreen Australia, personal communication) on the assumption, where necessary, that all screens were initial and that rates were constant within and over time periods and age groups; and Similar estimates were made for BreastScreen Australia from 1 July 1991 to December 1994 by applying the age distribution for all screens, initial and subsequent, in Victoria8 (BreastScreen Australia, personal communication) and New South Wales9 to published national screening estimates,9 and by estimating the fraction that were initial screens from State-based proportions of initial screens. These three sets of estimates were totalled to estimate numbers of women who had had at least one mammogram by the end of each year from 1984 to 1994. Total fertility rates by age in Australian women were obtained by year from 1921 to 1994.10,11 Rates of first births by age and year, available only for married women, showed patterns similar to those for total fertility. Results Mortality Breast cancer mortality (Box 1a) increased in Australia between the periods 1921-1924 and 1940-1944 (average annual increase, 1.0%), fell to a post-war low around 1960, and increased slowly again to a peak in 1985-1989. The average annual increase in rates from 1970-1974 to 1985-1989 was 0.3%. Thereafter, rates fell by 0.5% a year between 1985-1989 and 1990-1994. The recent fall in annual age-standardised rates has been from 20.7 per 100 000 woman-years in 1990 and 1991, to 19.6 per 100 000 in 1992, 1993 and 1995, and 19.2 in 1996; in 1994 the mortality rate was 20.3 per 100 000 woman-years. Age-specific mortality rates for women aged 25-49 years showed little change between 1921 and 1994 (Box 2). In women 50 years of age and older, breast cancer mortality rose steadily to reach initial peaks at different times between 1935-1939 and 1950-1954. These peaks occurred in women born between 1865 and 1885. Mortality in women 50 years and older then fell to a post-war low between 1950-1954 and 1985-1989 depending on age group and occurring in women born around 1890 to 1905. Thereafter, mortality rates again rose and appeared to peak in 1980-1984 or 1985-1989 for each age group of women from 50-54 years to 70-79 years. There was little evidence of this later mortality rise in women older than 80 years. In the target age group for mammographic screening (50-69 years), the age-standardised mortality rate increased by 4.0% (95% CI, - 0.1 to 8.3) from 1980-1984 to 1985-1989 and then fell by 3.2% (95% CI, - 6.9 to 0.6) to 1990-1994. In women 25-49 years of age, rates increased by 7.5% (95% CI, 0.4-15.0) from 1980-1984 to 1985-1989 and then fell by 4.2% (95% CI, - 9.9 to 1.9) to 1990-1994. Women 70 years of age and older showed only weak evidence of change from the early to the late 1980s (1.3%; 95% CI, - 3.1 to 6.0) and almost no evidence of a fall in the 1990s ( 2 0.2%; 95% CI, - 4.2 to 4.0). Mammographic screening Our analysis of mammographic screening data is summarised in Box 3. The percentages of women screened between 1988 and 1994 increased in all age groups, most notably in women aged 50-69 years. The beginning of most pilot projects in late 1988 and early 1989 and of the National Program in 1991 led to a substantial change in the age distribution of screening mammograms. There was a nearly sevenfold increase in the cumulative numbers of women in the target age group who had been screened once or more, from around 138 000 in 1988 to nearly one million by 1994. The estimated number of women under 50 years of age who had been screened (which was twice the number in the target group in 1988) increased only 3.7 times (to an estimated 1 140 000) by 1994. By 1994, about 70% of women in their 50s and 58% of women in their 60s had had an initial screen. In all, 54% of women in their 40s and 22% of women younger than 40 were estimated to have been screened once by 1994. In older women, the proportion screened reached 25% at 70-74 years of age and 6% at 80 years and older. Fertility Fertility fell at all ages from 1921-1924 or 1925-1929, to reach a minimum in most age groups in 1930-1934 or 1935-1939 (Box 1b), and then rose to reach a peak between 1945-1949 and 1970-1974. Rates again fell from these peaks. Minima were reached in 1975-1979 or 1980-1984 in those aged between 30 and 44 years, with subsequent rises to 1990-1994. There was little evidence that any of these trends related better to cohort of birth than time period. Discussion Compared with many other countries, breast cancer mortality in Australia has varied little over the past 75 years.1 However, the clear trends we found cannot be explained by chance fluctuation. The recent fall in the early 1990s, while apparently quite definite in women aged 40-69 years, could be a chance occurrence and will need to be shown for several more years for it to be declared a "real" trend. There are three broad classes of explanation for these trends, represented under the subheadings below. Changes in how cause of death is registered and coded can influence reported cause-specific mortality rates, but no substantial changes of this kind are known to have occurred in Australia.1,12 We have specified "real" incidence (below) because recent apparent increases in breast cancer incidence have probably been the result of increased screening.3,13 These increases will not cause increases in mortality because they reflect either earlier diagnosis of breast cancers, or the diagnosis of lesions that would otherwise never have been detected.13 Changes in determinants of real incidence Incidence trends: Whether changes in incidence caused changes in mortality would be most easily determined by comparing incidence and mortality trends. Incidence of breast cancer in New South Wales (representing about one-third of Australian women) changed little from 1972 to 1983, but increased steadily from 1984, and by 1995 was nearly 50% higher than it was in 1983.14 The greatest increase was in women in the target age group for mammographic screening (50-69 years). As there was no parallel increase in mortality during that period, the observed increase in incidence has probably been caused by screening. Incidence was not measured in Australia before 1972. However, it is most likely that incidence rates in Australia, as in several other countries,15-17 were increasing and thus underlie the steady increase in mortality from 1921 to the peak of the mid 1940s. Incidence rates in Australia were probably also increasing in the post-war period, as in other populations of European origin,18,19 when Australian mortality rates were falling or stable. Fertility: Breast cancer is associated with late age at first birth, childlessness and low parity.13 Box 1 shows that Australian trends in rates of mortality from breast cancer moved in the opposite direction from those of fertility rates, with changes in mortality occurring a few years after those in fertility. Australian women born in the 1840s were at the forefront of a transition to lower family sizes in English-speaking countries.10 This falling fertility could have produced the increasing mortality from breast cancer from 1921 onwards. The peak breast cancer mortality in the 1940s occurred in women born before 1885; the highest proportions of unmarried and childless women seen in Australia up to the 1940s were among those born in 1871 to 1876.10 The upward trend in fertility after 1935, which peaked in 1955-1964, started about 10 years before mortality began to fall in the mid 1940s. Fertility again fell in the late 1960s and 1970s to a new low in the 1980s, with a pronounced shift during that period to later childbearing; mortality began to rise again in the 1980s. Body size and age at menarche: Each 5-cm increase in average height in adult women has been estimated to increase breast cancer risk by 10%.13 From the early 1900s to about 1980, net increases of 8-9 cm in height and 10 kg in weight20,21 could have contributed appreciably to increases in mortality in women born from about 1895 to 1935 (evident in overall mortality from breast cancer between 1970-1974 and 1985-1989). A fall in age at menarche is also associated with an increase in breast cancer incidence.13,22 Trends in Australia, probably similar to the 2-3 months' fall per calendar decade seen in the United Kingdom and United States in the 100 years to about 1950,13,22,23 would have been expected to increase breast cancer rates. Younger age at menarche is very likely caused by increased height and body mass index, perhaps because menarche depends on attainment of a critical body mass.22,24 Diet: Dietary changes may have affected breast cancer rates by way of changes in body size, and possibly by other means.25 Alcohol consumption is associated with increased risk of breast cancer, being 35% higher in women who have 2-4 drinks and 67% higher in those who have more than four drinks a day compared with women who drink little or no alcohol.26 The high alcohol consumption among Australian women of the early 1800s was not equalled again until 1989, when more than 50% of women over 18-20 years were consuming up to two drinks a day.26,27 However, it has been estimated that no more than 3% of breast cancers in Australia in 1990 were the result of drinking more than two drinks of alcohol a day.26 Physical activity may reduce risk of breast cancer.27 The proportion of Australian women who participate in any recreational exercise (around 70%) appears not to have changed in recent times.28 Changes over time in the physical activity associated with running a household and in paid employment have not been measured. Migration: Breast cancer rates vary six-fold internationally.29 Migration of women from countries of higher (UK) and lower (eastern Europe, southern Europe, Asia) breast cancer mortality than in Australia has varied, but the net effect of migration on breast cancer mortality rates has probably been small. Oestrogen use: Use of the oral contraceptive pill increased rapidly in Australia after its introduction in 1961.30 If the Pill has caused an increase in breast cancer mortality, it would have done so mainly in younger women who were current or recent users.31 The increase in mortality in the late 1980s, however, was mostly in women over 50 years of age. Falling mortality from breast cancer in the generations of women who first used oral contraceptives in the US, UK and Sweden also suggests no major effects of the Pill on breast cancer rates.32,33 The use of oestrogen replacement therapy, which may increase breast cancer risk, has probably not affected breast cancer incidence appreciably as long term use has been uncommon in Australia. Changes in determinants of stage at diagnosis Trends to smaller breast cancers and fewer axillary node metastases over nearly 100 years34-36 are probably the results of increased access to and use of care. The trend to more localised disease is probably continuing because of increased screening.3,13,37 Such trends would have caused downward pressure on breast cancer mortality. Mortality from breast cancer in Australian women in the target age group for screening (50-69 years) fell by 3.2% between 1985-1989 and 1990-1994. This may be the result of screening, although the same or larger falls in mortality also occurred in younger women, for whom there is little evidence that mammographic screening reduces breast cancer mortality.38 Changes in determinants of probability of survival after diagnosis at a particular stage The effectiveness of radical mastectomy as the primary treatment for breast cancer has probably changed little in the past 100 years. However, advances in anaesthetics and operating conditions38 as well as the recent use of adjuvant chemotherapy and hormonal therapy3,39 have almost certainly increased survival. Conclusions While no certain conclusions can be drawn about the causes of changes in breast cancer mortality since 1921, it is probable that the increase to the mid 1940s was caused mainly by rapidly falling fertility in the latter part of the 19th and the early 20th centuries. Subsequent increased fertility, and earlier diagnosis, may have contributed to the fall in mortality from 1940-1944 to 1960-1964. The increasing mortality in women born between 1895 and 1935 was probably caused by nutritional factors leading to increases in body size and resultant earlier age at menarche. However, this incidence-driven increase in mortality was probably moderated by increasing survival with earlier diagnosis and, more recently, improved treatment. Improved treatment is probably the reason for cross-sectional falls in mortality between 1985-1989 and 1990-1994 in women up to 69 years of age. Early effects of mammographic screening may have contributed to these falls, but should be more clearly evident in the second half of the 1990s. Acknowledgements Australian Bureau of Statistics data on mortality from breast cancer from 1921 to 1994 were supplied by Mr Paul Jelfs from the national mortality database at the Australian Institute of Health and Welfare, Canberra. References Hermon C, Beral V. Breast cancer mortality rates are levelling off or beginning to decline in many western countries: analysis of time trends, age-cohort and age-period models of breast cancer mortality in 20 countries. Br J Cancer 1996; 73: 955-960. Beral V, Hermon C, Reeves G, Peto R. Sudden fall in breast cancer death rates in England and Wales [letter]. Lancet 1995; 345: 1642-1643. Chu KC, Tarone RE, Kessler LG, et al. Recent trends in US breast cancer incidence, survival, and mortality rates. J Natl Cancer Inst 1996; 88: 1571-1579. Essendon Breast X-Ray Program Collaborative Group. A mammographic screening pilot project in Victoria 1988-1990. Med J Aust 1992; 157: 670-673. Rickard MT, Lee W, Read JW, et al. Breast cancer diagnosis by screening mammography: early results of the Central Sydney Area Health Service Breast X-ray Programme. Med J Aust 1991; 154: 126-131. Robinson JI, Crane CEB, King JM, et al. The South Australian Breast X-Ray Service: results from a statewide mammographic screening programme. Br J Cancer 1996; 73: 837-842. Australian Health Ministers' Advisory Council. Breast Cancer Screening Evaluation Committee. Breast cancer screening in Australia: future directions. Canberra: AGPS, 1990. Victorian Breast Screening Program. Annual statistical report 1994. Melbourne: Victorian Breast Screening Program, 1996. Smith D, Oudod V, Supramaniam R, et al. BreastScreen NSW Statistical Report 1991 to 1995. Sydney: NSW Cancer Council, 1996. McDonald P, Ruzicka L, Pyne P. Marriage, fertility and mortality. In: Vamplew W, editor. Australians: historical statistics. Sydney: Fairfax, Syme and Weldon, 1987: 44-61. Australian Bureau of Statistics. Births, Australia. Canberra: AGPS, 1995. (Catalogue No. 3301.0.) Fleming NT, Armstrong BK, Sheiner HJ, James IR. Occurrence of breast cancer in Australian women. Med J Aust 1981; 1: 289-293. Ursin G, Bernstein L, Pike MC. Breast cancer. Cancer Surv 1994; 19-20: 241-264. Cancer Control Information Centre. Breast cancer incidence, 1995. Sydney: NSW Cancer Council, 1996. Wigle DT. Breast cancer and fertility trends in Canada. Am J Epidemiol 1977; 105: 428-438. Stevens RG, Moolgavkar SH, Lee JA. Temporal trends in breast cancer. Am J Epidemiol 1982; 115: 759-777. Ewertz M, Carstensen B. Trends in breast cancer incidence and mortality in Denmark, 1943-1982. Int J Cancer 1988; 41: 46-51. Tulinius H, Sigvaldason H. Trends in incidence of female breast cancer in the Nordic countries. In: Magnus K, editor. Trends in cancer incidence. Washington: McGraw-Hill, 1982: 235-247. Nab HW, Mulder PG, Crommelin MA, et al. Is the peak in breast cancer incidence in sight? A study conducted in the southeastern Netherlands. Eur J Cancer 1994; 30A: 50-52. May GM, O'Hara VM, Dugdale AE. Patterns of growth in Queensland schoolchildren, 1911 to 1976. Med J Aust 1979; 2: 610-614. Hitchcock NE, Maller RA, Gilmour AI. Body size of young Australians aged five to 16 years. Med J Aust 1986; 145: 368-372. Henderson BE, Bernstein L. The international variation in breast cancer rates: an epidemiological assessment. Breast Cancer Res Treat 1991; 18 Suppl 1: S11-S17. Frisch RE. Body weight, body fat, and ovulation. Trends Endocrinol Metab 1991; 5: 191-197. Petridou E, Syrigou E, Toupadaki N, et al. Determinants of age at menarche as early life predictors of breast cancer risk. Int J Cancer 1996; 68: 193-198. Prentice RL, Sheppard L. Dietary fat and cancer: consistency of the epidemiologic data, and disease prevention that may follow from a practical reduction in fat consumption. Cancer Causes Control 1990; 1: 81-97. English DR, Holman CDJ, Milne E, et al. The quantification of drug caused morbidity and mortality in Australia, 1995 edition. Canberra: AGPS, 1995. Willett WC, Trichopoulos D. Nutrition and cancer: a summary of the evidence. Cancer Causes Control 1996; 7: 178-180. Bennett SA, Magnus P. Trends in cardiovascular risk factors in Australia. Results from the National Heart Foundation's Risk Factor Prevalence Study, 1980-1989. Med J Aust 1994; 161: 519-527. Parkin DM, Muir CS, Whelan SL, et al. Cancer incidence in five continents. Vol. VI. Lyon: International Agency for Research on Cancer, 1992. (IARC Scientific Publications No. 120.) Santow G. Trends in contraception and sterilization in Australia. Aust N Z J Obstet Gynaecol 1991; 31: 201-208. Collaborative group on hormonal factors in breast cancer. Breast cancer and hormonal contraceptives: collaborative reanalysis of individual data on 53 297 women with breast cancer and 100 239 women without breast cancer from 54 epidemiological studies. Lancet 1996; 347: 1713-1727. Beral V, Hermon C, Reeves G, et al. Breast cancer trends in women in Sweden, the UK, and the USA in relation to their past use of oral contraceptives. In: Proceedings of the Second International Symposium on Hormonal Carcinogenesis. Berlin: Springer Verlag, 1996: 99-106. dos Santos Silva I, Swerdlow AJ. Recent trends in incidence of and mortality from breast, ovarian and endometrial cancers in England and Wales and their relation to changing fertility and oral contraceptive use. Br J Cancer 1995; 72: 485-492. Snaedal G. Cancer of the breast. A clinical study of treated and untreated patients in Iceland 1911-1955. Acta Chir Scand 1964; Suppl 338. Joensuu H, Toikkanen S. Comparison of breast carcinomas diagnosed in the 1980s with those diagnosed in the 1940s to 1960s. BMJ 1991; 303: 155-158. Buchanan EB. A century of breast cancer surgery. Cancer Invest 1996; 14: 371-377. Kricker A, H¿yer AP, McCredie M, Porter LA. Breast cancer in NSW women: a shift in tumour size. Med J Aust 1995; 163: 79-81. Glasziou PP, Woodward AJ, Mahon CM. Mammographic screening trials for women aged under 50. A quality assessment and meta-analysis. Med J Aust 1995; 162: 625-629. Early Breast Cancer Trialists' Collaborative Group. Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy. Lancet 1992; 339: 1-15. (Received 29 Jan, accepted 21 Jul, 1997) Authors' details National Health and Medical Research Council National Breast Cancer Centre, Sydney, NSW. Catherine L Smith, MPH, Statistician; Anne Kricker, PhD, Epidemiologist. Cancer Control Information Centre, NSW Cancer Council, Sydney, NSW. Bruce K Armstrong, DPhil, FRACP, Director. Reprints will not be available from the authors. Correspondence: Dr A Kricker, NHMRC National Breast Cancer Centre, PO Box 572, Kings Cross, NSW 2011. Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". <URL: http://www.mja.com.au/> © 1997 Medical Journal of Australia.

Catherine L Smith · Anne Kricker · Bruce K Armstrong

Funding Australia's health and medical research A commitment to improving health through high quality research MJA 1997; 167: 608-609 In 1996, the National Health and Medical Research Council's research expenditure was approximately $142 million.1 According to official estimates, this is about a quarter of Australia's total outlay on health research.2 With this level of expenditure, it is reasonable to ask what we gain from this investment. In this issue of the Journal, Bourke and Butler show that in 1995 Australian research provided about 2.5% of new knowledge in medicine and the health sciences (Australian research produced 18390 publications, which were cited over 88000 times), a contribution that has strengthened in recent years.3 This is tangible evidence of our commitment to improving the health of people not only in Australia but in other countries. In one sense, this health research commitment is similar to our other international commitments, such as preserving world heritage areas or reducing emissions of gases causing ozone depletion. Apart from this international contribution to knowledge and new discoveries to improve health, research also benefits Australia itself. For example, health care practitioners with research training are of key importance in ensuring that Australian clinical practice is effective and of high quality because they can access, sift and digest new research findings here and elsewhere. Researchers active in public health are needed to provide informed advice on the benefits and costs of proposed public health measures; for example, the detection and prevention of cancer and cardiovascular disease. The strong base of Australian medical research is also helping to build our biotechnology industry. Finally, and less tangibly, Australians derive satisfaction and pride through the achievements of their health researchers, just as they are proud of high achievers in sport and the arts. The NHMRC now provides research funds across the spectrum of health and medical science via a single Research Committee formed in 1997 by a merger of the Medical Research Committee and the Public Health Research and Development Committee. As the largest single provider of peer-reviewed health research funds for the public good, the NHMRC has the responsibility of ensuring that Australia's health research effort is of high quality: the Research Committee provides funds on the basis of excellence (as judged by peer review), strives for fair, open and competitive processes, considers funding applications in all areas relevant to health, and attempts to ensure that research findings are translated into better health and health care delivery. Health research is conducted mainly in universities, hospitals and medical research institutes (Box 1, above). About 30% of total NHMRC research support is provided to medical research institutes, about half of which is via block funding to five institutes. Box 2 shows the support won by institutes receiving over $1 million, those which are block funded, and the proportion of institute funding going to each State (for institutes receiving over $100 000). NHMRC funding to the major universities is shown in Box 3. The Government provides additional support for hospital and university research via salaries for researchers doing academic or clinical research as part of their teaching and clinical duties, but this does not occur for all research institutes. The institutes also play a valuable role in gaining additional support for health research from private corporations and individuals. The report by Bourke and Butler3 shows that the medical research institutes produce some of our most highly regarded research. Ten of the 12 most cited Australian biomedical research articles are associated with institutes, and the overall citation rate for publications from institutes is impressively high: medical research institutes publish about 13% of all Australian health research articles, and these account for about 21% of the total citations of Australian health research. Underfunding the institutes would jeopardise this high quality output. Bourke and Butler comment that their results raise issues of the efficacy of block funding,3 but stress that conclusions about funding methods require further bibliometric studies. The NHMRC's Research Committee is actively reviewing all aspects of research support, including issues in block funding, such as how best to review institutes and how to increase competition between institutes. Some will argue that the data of Bourke and Butler show that substantial, longer term research support (e.g., block funding) is required in order to achieve the highest impact research. Others will point to examples such as St Vincent's Institute for Medical Research in Melbourne to argue that institutes can achieve outstanding publication results without block funding. The NHMRC supports a pluralistic system because the aims of health research extend well beyond quality research publication. It currently supports three-year project grants (about 48% of the current total NHMRC funding), and five-year program grants (9% of total NHMRC funding) and block grants (13% of total NHMRC funding) to provide for longer term research commitments. New researchers are supported via scholarships and postdoctoral awards (about 9% of funding), and outstanding individuals are supported by Fellowships (11% of funding, exclusive of Fellows on Programs and at block-funded institutes). Nothing in science or medicine is immune to change. The NHMRC's research support system must accord with the ways in which research is currently conducted and be relevant to contemporary health needs. One exciting plan for the triennium 1997-1999 is the introduction of multidisciplinary Health Research Networks in important health areas. This scheme will involve researchers at different locations working collaboratively on research in areas of major health importance. Part of the aim is to capitalise on the merger of the Medical Research Committee and the Public Health Research and Development Committee by creating teams of biomedical, clinical and public health researchers. The Networks will focus on effective delivery of health care or prevention. They will thus be integrated research teams (as in the NHMRC Program Grants Scheme), but multisite and multidisciplinary. Interestingly, Bourke and Butler's article3 shows that research collaborations, both local and international, are growing strongly. The NHMRC-sponsored Networks scheme acknowledges this trend and will help to build new collaborations. Bibliometric data are important for monitoring Australian research. We also need better indicators of the other outcomes of research (such as discoveries which benefit health or reduce health costs, better-informed medical and public health decision making, a vibrant biotechnology industry, and new generations of trained researchers in all areas) so that we can assure the Australian community that its financial investment in health and medical research is well placed. Warwick P Anderson Professor and Chairman, Research Committee (Public Health and Medical) National Health and Medical Research Council, Canberra, ACT National Health and Medical Research Council. 1996 Annual Report. Canberra: AGPS, 1997. Australian Institute of Health and Welfare. Health Expenditure Bulletin No 13, July 1997. Bourke PF, Butler L. Mapping Australia's basic research in the medical and health sciences. Med J Aust 1997; 167: 610-613. - ©MJA 1997 Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". <URL: http://www.mja.com.au/> © 1997 Medical Journal of Australia.

Warwick P Anderson

Mapping Australia's basic research in the medical and health sciences Paul F Bourke and Linda Butler The Institute for Scientific Information indexes most of the major international basic research journals in science in the Science Citation Index (SCI). Australia's presence in the medical and health sciences journals in the SCI and the citations its published research receives in these journals show that Australia's basic medical research has high international "visibility". Mapping the source of the most highly "visible" Australian medical research articles shows high impact research coming from several different sectors (research institutes, universities, hospitals, etc.), but with a concentration in the member institutions of the Australian Association of Medical Research Institutes (AAMRI). Published research from the AAMRI is cited at a rate two-thirds higher than the Australian average for medical and health sciences. (MJA 1997; 167: 610-613) Introduction - Data source - To what extent is Australian basic research in the medical and health sciences covered by the REPP database? - Where is the research being conducted? - Are patterns of medical authorship changing? - How "visible" is Australian medical research? - Where is Australia's most "visible" medical research being conducted? - Where are Australia's most highly cited medical articles produced? - Discussion - References - Authors' details - - - ©MJA1997 Introduction A very high proportion of the articles reporting the results of basic research in the medical and health sciences are in the journals indexed by the Institute for Scientific Information (ISI) in the Science Citation Index (SCI). The SCI is therefore an excellent tool for identifying the sources of Australian published research in this field, while the citations received by (or references to) these articles can also be used to identify the source of the most-cited research. We have focused on Australia's basic research output, not its applied research. Data source The Research Evaluation and Policy Project (REPP) at the Australian National University (ANU) has constructed a database of all Australian research published in ISI-indexed journals for the period 1981-1995. This database has been well documented in several of our published studies, and a description is available on the Internet.1,2 Our analysis of this database is based on the addresses shown on the publications, which we have "cleaned" down to the level of university department. This has been done by ensuring all variants of a departmental, faculty or institutional address (sometimes running into hundreds) are grouped together and given the same "standard" address. The database also contains details of the citations in ISI journals received by these Australian research articles; that is, the number of times Australian articles are in the reference list of other ISI journal articles. We also code the sector -- universities, hospitals, medical research institutes, government institutions, and "other" -- of each address. The medical research institutes sector comprises the 25 members of the Australian Association of Medical Research Institutes (AAMRI). Medical research in the Government sector comes primarily from the Commonwealth Scientific and Industrial Research Organisation (CSIRO) and the State and Federal departments of health. The "other" sector includes industry and non-profit organisations. We acknowledge "grey areas" at the margins of these assignments to sectors. For example, where a research group based in a hospital with a university connection lists the university in the address, we consistently assign the publication to the university. Another example is the John Curtin School of Medical Research (JCSMR), a full-time medical research institute (though not a member of AAMRI), which is also a research school of the ANU funded from that university's operating grant. Its publications are assigned to the ANU and hence fall within the universities sector. Most publications can be unambiguously assigned, but some addresses are not precise guides to the source where the research was conducted. This arises particularly for researchers who hold conjoint or adjunct appointments in two institutions. If such an author nominates a single institution, we accept that this is the location of the research leading to that publication. If an author specifies two separate addresses, two records are created for that publication showing the two institutions (double-counting created by this procedure is removed for sectoral and national analysis, but remains when the individual institution is the focus of the analysis). If the address itself has multiple components, the publication can not be split into multiple records and has to be assigned to the apparent primary institution. Very few publications fall into this latter category, and their existence has little effect on the map of national, sectoral or institutional sites of research. We have used the standard bibliometric practice of ascribing an article to a particular field of research on the basis of the classification of the journal in which it appears. This procedure is not without problems, particularly in the case of analysis at the subfield level, but experience from our other studies has shown that the results are accurate if the data are being used to map research in a large field.3 To what extent is Australian basic research in the medical and health sciences covered by the REPP database? In this field, at least 70% of published research output from universities and medical research institutes is in the form of journal articles, and, of these, at least 70% appear in SCI journals.4,5 We therefore estimate that, as a minimum, the REPP database covers 50% of the published output in these two sectors. SCI coverage of research from the sectors hospitals and government institutions may be less complete in terms of their total research output, but their contributions to basic research will be well represented. In this article, we are using the REPP database to answer questions about the map of medical research in Australia, concentrating primarily on publications from 1990 onwards, but also introducing some time-series analysis. Where is the research being conducted? The sectoral distribution (Figure 1) shows that the bulk of Australia's basic medical research is located in universities and hospitals. Most of the research articles from the universities sector (75%) come from the 10 teaching medical schools and the JCSMR at the ANU. Our data also show that little has changed in the sectoral location of research over the past 15 years: the share of publications from the hospitals sector has remained constant, there has been a small drop in the universities share, and there has been a corresponding increase in the share from the medical research institutes. Are patterns of medical authorship changing? The REPP database enables us to distinguish publications by type of authorship, viz: Single author -- one author only (i.e., no collaboration); Group -- more than one author but sharing the same departmental address; Institutional -- more than one author from different departments/faculties; National -- more than one author from different institutions in Australia; and International -- more than one country listed in the author addresses. Figure 2 shows how the type of authorship of medical research articles has changed in the past 15 years. The two authorship types showing marked decline are "single author" articles, and "group" articles. Medical and health sciences in Australia appeared slow to exhibit the "internationalisation" of research apparent in other fields since the early 1980s, but the period since 1987 has seen a dramatic change. The proportion of publications involving international collaboration nearly doubled between 1987 and 1995. Collaboration with other Australian institutions has also become more common, increasing from 19% to 26% over the 15-year period. (A detailed analysis of Australia's international collaboration in basic research may be found in a monograph we prepared for the Australian Research Council.6 ) How "visible" is Australian medical research? We assess "visibility" by the number of research articles published and the citations those articles receive. Figure 3 plots Australia's share of "world" publications (i.e., of the total in the SCI) in the medical and health sciences and its share of all citations in SCI journals. The chart also plots Australia's Relative Citation Impact (RCI), which is calculated by dividing its share of "world" citations by its share of "world" publications. The most notable feature of this Figure is that Australia's share of publications in SCI medical journals increased by 25% between 1986 and 1995. The average RCI for the whole period was 1, indicating that Australian publications are attracting appropriate notice. Australia's RCI has not changed significantly over time, but remained at or marginally above 1. This is a strong performance as citation rates are influenced principally by publications from the major research centres of America and Europe. Where is Australia's most "visible" medical research being conducted? A standard measure used to compare the visibility of research in different sectors is the average number of citations received per publication (cpp). In Box 1 (below) we relate the number of publications produced by each sector in the period 1991-1995 to the number of citations those publications attracted in the same period. The leading position of the institutes making up the AAMRI is consistent with Richard Smith's impression of Australia's research institutes.7 Some of the differences in cpp rates can be attributed to the varied research profiles of different institutions. Data supplied by ISI enable us to quantify this, as we can calculate cpp rates for sets of journals. For example, articles in immunology journals for the same period attracted citations at the average rate of 6.21, while those in clinical sciences journals averaged 3.81. We would therefore expect hospitals, with their strong clinical focus, to have a lower cpp rate than those AAMRI establishments with a strong presence in immunology. However, the difference in cpp rates apparent in Box 1 (above) cannot be explained fully by differences in field concentrations; we calculate that differing citation rates across fields account for only a third of the gap in cpp rates between the AAMRI institutions and other sectors. The remainder is a measure of the differing visibility and impact of the research. Not all institutions within a given sector have similar visibility. In Box 2 we list the top five institutions (in terms of cpp rates) in each of the four sectors active in medical research. In the universities sector, we looked specifically at the teaching medical schools and for this reason have excluded JCSMR. The institutions in Box 2 are those with more than 100 SCI publications in the period. In this instance, we count publications in multidisciplinary journals such as Nature, Science and Proceedings of the National Academy of Sciences together with publications in medical and health sciences journals, on the assumption that for the listed institutions and faculties these articles would almost certainly relate to medical research. Sorting the institutions in Box 2 by cpp rates within sectors has the effect of allowing volume of publications to be moderated by impact. The consequences of the choice of the measure on which to rank institutions can be seen by a closer examination of the medical research institutes. While the Walter and Eliza Hall Institute of Medical Research has the largest number of publications and citations, and would be ranked first if these were the measures used, the Ludwig Institute for Cancer Research is top-ranked on the basis of average cpp rates. All three are measures of impact and visibility, but cpp rates take institutional size into account. We looked at the research focus of institutions to determine if the higher cpp rates of some institutions resulted from differing fields of concentration. Again, we found that while an institution's cpp rate was affected by the relative impact of the fields in which it was active, this accounted at most for only 30% of the variation in cpp rates between institutions in any given sector. Where are Australia's most highly cited medical articles produced? We identified a very small group of 12 articles, published since 1990, which have attracted more than 200 citations (Box 3). This Box does much to explain the ranking of the AAMRI institutes in Box 2. The Ludwig Institute for Cancer Research had only 135 publications satisfying our criteria of publication date and journal, yet three of these have attracted more than 200 citations. St Vincent's Institute of Medical Research had even fewer publications (130), but claimed authorship of the most highly cited publication for the period, with 551 citations, and another article with 256 citations. Seven of the listed articles, including the most highly cited publication, were "wholly" Australian; the other five articles involved international collaboration. Discussion Studying ISI journals in isolation does not permit conclusions to be drawn about quality. However, an analysis of citations can provide a guide to the source of Australia's most visible research, and there is a well established positive association8 between high visibility in ISI-indexed journals and research judged on other grounds, such as via peer evaluation and esteem measures, to be of high quality. The limitations of bibliometric analysis are well documented.9 Publications can attract large numbers of citations because they contain error, or because they report a new technique with wide application. Citation data are also highly skewed. Many publications attract no citations at all, and most of those that do receive only one or two. These and other problems are of little consequence when the focus is at the national or sectoral level, involving large numbers of publications.10 Our data provide only one approach to constructing a profile of Australian medical research. However, bibliometric or literature-based analysis cannot stand in isolation from historical and other kinds of evaluative judgements, and should not be used in a policy setting apart from those perspectives. That said, we believe that the information reviewed here does allow some interesting points to be made. The most encouraging inference we draw from our study is that, using the measure of Relative Citation Impact, Australian medical research stands relatively high in terms of international visibility. As Box 3 makes clear, Australian-based researchers publish in international journals and have well established links to collaborative projects in the major centres of work in the field.6 The bulk of Australia's basic research in the medical and health sciences comes from the universities and hospitals, but Australia's medical research institutes, the members of AAMRI, have the highest international profiles. Research from these institutions has had a major impact on the international community. However, high visibility is not confined to the AAMRI institutions, and, as we have shown, research achieving very high impact also comes from hospitals and universities. One of the most interesting policy issues which these data raise is the efficacy of block funding by comparison with direct project funding of research. Medical research in Australia is undertaken in a pluralist system, with major contributions from hospitals, universities, AAMRI and government institutions. While at first glance our data may appear to argue for block funding through the prominence of several AAMRI institutions funded in this way, the situation is more complex. Many of the AAMRI institutions are block-funded, but in some instances this accounts for as little as 35% of their total income. Any conclusions about funding await the completion of detailed bibliometric studies of the relative performance of medical research, in which we will attempt to identify the output of research supported by the varying modes of research funding. References Bourke P, Butler L. A Crisis for Australian science? Canberra: Performance Indicators Project, Australian National University, 1993. (Monograph Series No. 1.) < http://coombs.anu.edu.au/Depts/RSSS/REPP/repp.htm > Butler L, Bourke P, Biglia B. CSIRO: profile of basic research. Canberra: Research Evaluation and Policy Project, Australian National University, 1997. (Monograph Series No. 4.) National Board of Employment Education and Training (NBEET). Quantitative indicators of Australian academic research. Canberra: AGPS, 1994. (Commissioned Report No. 27.) Bourke P, Butler L. Monitoring research in the periphery. Canberra: Research Evaluation and Policy Project, Australian National University, 1996. (Monograph Series No. 3.) National Board of Employment Education and Training. International links in higher education research. Canberra: AGPS, 1995. (Commissioned Report No. 37.) Smith R. Top of the pile: the institutes. BMJ 1991; 302: 1006-1010. Narin F. Evaluative bibliometrics. Cherry Hill, NJ: Computer Horizons Inc, 1976. Van Raan AFJ, editor. Handbook of quantitative studies of science and technology. Amsterdam: Elsevier Science Publishers, 1988. Garfield E. In: Evered D, Harnett S, editors. Ciba Foundation Conference: the evaluation of scientific research. Chichester (UK): John Wiley & Sons, 1989. (Received 6 Jun, accepted 29 Sep, 1997) Authors' details Research Evaluation and Policy Project, Research School of Social Sciences, Australian National University, Canberra, ACT. Paul F Bourke, PhD, FASSA, Head of Research Evaluation and Policy Project; and Professor of History. Linda Butler, BEcon, Research Officer. Reprints: Professor P F Bourke, Research Evaluation and Policy Project, Research School of Social Sciences, Australian National University, ACT 0200. E-mail: paulb AT coombs.anu.edu.au - ©MJA 1997 Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". <URL: http://www.mja.com.au/> © 1997 Medical Journal of Australia.

Paul F Bourke · Linda Butler

The Institute of Medical and Veterinary Science Brendon J Kearney Situated in the grounds of South Australia's principal teaching hospital, the IMVS provides a unique blend of statewide pathology services with comprehensive basic and clinical research programs. MJA 1997; 167: 614-617 Introduction - Growth of the IMVS - Core pathology services - Research at IMVS - The Hanson Centre for Cancer Research - The future of IMVS - References - Authors' details - - - ©MJA1997 Introduction The Institute of Medical and Veterinary Science (IMVS), in Adelaide, South Australia, began in 1938 as a development from the Royal Adelaide Hospital (RAH) laboratories largely because Sir Trent Champion de Crespigny, an eminent physician, Medical Superintendent of the RAH and Dean of Medicine from 1929 to 1947, had a vision for an institute which combined laboratory services, teaching and research.1,2 In this respect, the IMVS differs from pure research institutes and has continued to emphasise an integration of service, teaching and research in clinical medicine. Growth of the IMVS The steady development of the IMVS was interrupted in the late 1970s by a series of enquiries into management efficiency, services provided and internal policies, as well as a critical review of the care and control of animals used in scientific experiments. The result was a major redefinition of core services: the Division of Forensic Pathology and the Division of Biology became the Forensic Science Centre (and relocated elsewhere in Adelaide) and the Division of Veterinary Pathology was transferred to the South Australian Department of Agriculture. In July 1982, a new Act (with a new Council) defined and strengthened the relationships of the IMVS with the RAH and the University of Adelaide, allowing it to concentrate on the mainstream areas of medical pathology and their application to research, teaching and service. In 1986, an internal report (Review of Research) on the IMVS/RAH campus by Professors David Penington and Gustav Fraenkel outlined a framework for the development of research. The integration of service, teaching and research with the outcome of high quality care provided by clinicians actively engaged in research was endorsed in the report, but they also suggested the creation of separate basic science laboratories to support research and avoid disparate but duplicated facilities on the campus. They recommended that this development should be part of the IMVS/RAH campus because the benefits of a separate research centre would be diminished by the separation of research activities from service functions. The IMVS and RAH accepted the challenge, and by 1990 funds had been raised and the Hanson Centre (now the Hanson Centre for Cancer Research) had been established. Over the past decade, the IMVS has organised itself along business lines and become reliant on its ability to provide high quality patient services, attract grants and contract for service and research. About 80% of its $55 million expenditure annually is derived from contracts or grants (for public and private pathology services on a fee-for-service basis); the remaining 20% is derived from direct grants from the South Australian Government for statewide or specific pathology services. Therefore, the IMVS differs from pure research institutes that rely on government grants, research grants and donations. While this adds a complexity to its management, it is also a strength, ensuring that the IMVS remains competitive, efficient and relevant to the services for which it contracts. The strength of the IMVS is its collocation on the RAH and University of Adelaide medical school campus. This allows easy collaboration, integration and coordination with both the hospital and University, which has a symbiotic effect on achievement. This integration between hospital, university, service and research is exemplified by the organisational structure of the IMVS. Core pathology services The provision of a high quality general medical pathology service to the RAH remains the core service of the IMVS, and The Queen Elizabeth Hospital laboratories recently joined the IMVS. The IMVS also provides pathology services to metropolitan and rural areas in South Australia through a network of 12 country and regional laboratories. These comprehensive, regionally based laboratory facilities have ensured that a high quality pathology service is available to most communities in South Australia. In addition, services are provided to the Northern Territory and some areas in Victoria. The integration of the IMVS with a teaching hospital and university medical school has allowed the IMVS to develop a comprehensive teaching and training program in pathology and to become involved in education and training in collaboration with clinical disciplines. Research at IMVS Examples of the symbiotic relationship between pathology services, clinical services and research are illustrated by the research endeavours occurring within the Divisions of the IMVS. Division of Tissue Pathology The Division of Tissue Pathology, developed by Professor Barrie Vernon-Roberts, is one of the largest in Australia, with specialists in tissue pathology subspecialising and with the Division integrated into the University of Adelaide and RAH. The Adelaide Head Injury Group incorporates neuropathology from the IMVS, neurosurgery from the RAH and bio mechanical engineering from the University of Adelaide's Road Accident Research Unit. Together, they form a unique combination collaborating in basic laboratory and clinical research into the effects of head injury caused by road trauma. The research objective of the Adelaide Head Injury Group is to reduce the severity of brain injury from accidents by prevention and early intervention. Until recently, brain injury was believed to be instantaneous and irreversible. Work undertaken by the group suggests that axonal damage evolves over hours and may be modified by early interventions. They have developed techniques to identify the pathological changes occurring within an hour of a head injury. The increased production of an amyloid precursor protein is a sensitive indicator of brain injury and can be used to measure the effectiveness of treatments aimed at reducing axonal damage. The Spinal Research Group has won several national and international awards, including the prestigious Volvo Prize (twice), the International Society for Biomechanics Prize, the Amrad Award, and the Smith and Nephew Spine Research Award (twice). It is supported by the specialised bone and histoquantification laboratories of Dr Robert Moore and Dr Nicola Fazzalari. The group collaborates closely with the RAH Spinal Services and RAH Department of Orthopaedics and Trauma. They used a sheep model to reproduce the pathology of human disc lesions and showed how investigational studies of discs can inadvertently introduce bacteria, causing damage to the disc and adjacent bone,3 and the beneficial effect of prophylactic antibiotics to prevent this common complication.4 Other studies have analysed the damage to the disc resulting from cumulative wear and tear to the spine, and have defined the biomechanical and biochemical characteristics of this degeneration. The researchers are currently examining the vascularisation of the disc during ageing, the potential for allograft disc replacement, and the use of cultured osteogenic precursor cells in surgical fusion of the spine. The Bone and Joint Group, in collaboration with the RAH Department of Orthopaedics and Trauma, studies bone structure and remodelling, with an emphasis on osteoarthritis and osteoporosis, using computerised analysis and modelling.5 The findings are relevant to development of preventive treatment for osteoporotic fractures and osteoarthritis.6 The group has also shown that loosening of artificial joints is caused by particles of metal or plastic released from the articulating surfaces.7 Division of Medical Microbiology and Virology The Division of Medical Microbiology and Virology has a long and distinguished reputation. Dr J E McCartney, Medical Bacteriologist at the IMVS in the early 1950s, was a close colleague of Alexander Fleming (the discoverer of penicillin), who gave him some of the original Penicillium mould. Professor Barrie Marmion, Professor of Virology at the IMVS during the 1980s, worked on developing an effective Q fever vaccine and on hepatitis B. Professor Christopher Burrell, the current Head, is acclaimed for his cloning of the hepatitis B genome.8 This Division is the core of the South Australian Infectious Diseases Service. It provides statewide services in food and environmental testing, monitoring of viral diseases and food- borne microbial diseases as well as the State HIV/AIDS Reference Laboratory and national Salmonella Reference Laboratory. A first for the microbiology laboratories was recognition of the association between commercial potting soils and Legionella longbeachae infections.9 They have played a leading role in developing diagnostic procedures for Legionella spp., as well as researching virulence mechanisms and their source and role in infections in the community. The expertise of the Division in applying research into service was shown during the 1995 outbreak of haemolytic- uraemic syndrome in South Australia.10 Three days after receipt of the first specimen, the laboratories had implemented a gene amplification test to detect toxin in food and clinical specimens, identified the toxigenic Escherichia coli strain and traced its source. The tracing of the source of E. coli within days contrasts to the weeks and months needed in outbreaks in Japan and the United States and was significant in rapidly controlling its spread. Infectious diseases research programs. The Division also maintains programs of basic research related to infectious diseases, largely funded by the National Health and Medical Research Council. AIDS Research Laboratory is one of the component units of the National Centre for HIV Virology. Its goal is to define the early molecular events in the replication of HIV in the infected cell.11 This has led to clarification of the events in cell-to-cell transmission of the virus. A related area of research is the use of antisense molecules as a form of gene therapy to inhibit HIV replication. Hepatitis Research Laboratory is studying the mechanism of infection with hepatitis B and related viruses, including patterns of spread within the body, regulation of the carrier state and ways to interrupt it, and the mechanism of protection by vaccines and ways to improve it.12 Herpes simplex research on the pathogenesis of herpes simplex, with emphasis on latent infections within the nervous system and how the host response reacts to infection, is being directed by Dr Tony Simmons. Papillomavirus research into the replication of papillomavirus, the patterns the virus expresses during infection of epithelial surfaces and the factors controlling the different outcomes of the infection (e.g., warts, cancer) is being directed by Dr Geoff Higgins. Adelaide Infectious Diseases Centre. The Division of Medical Microbiology and Virology provides a continuum from patient diagnostic testing, reference test development and public health to basic research. To maximise synergy between these activities, the Adelaide Infectious Diseases Centre has been established, combining the IMVS Division of Medical Microbiology and Virology, the University of Adelaide Department of Microbiology and the infectious diseases services of the RAH and The Queen Elizabeth Hospital. This group is pre-eminent in South Australia in infectious disease surveillance, reference testing, epidemiological studies and disease outbreak investigation. Division of Clinical Biochemistry Bone research group. Professor Christopher Nordin, Associate Professor Allan Need and Associate Professor Howard Morris are addressing bone loss at the menopause, which, together with age-related bone loss, eventually leads to osteoporosis. Studies have established that increased urinary calcium excretion at menopause contributes significantly to bone loss.13 Increased dietary protein and salt exacerbate the urinary calcium "leak" and its effect on bone loss, while dietary restriction of these helps stabilise bone loss. Studies are being undertaken into the changes in the intestinal absorption of dietary calcium at menopause.14 Studies on new therapies for osteoporosis and on the most effective way of monitoring treatment are continuing. A longitudinal study of osteoporosis in men has found that bone loss commences at age 50 and that the pattern and causes of bone loss are different from those found in women.15 The effects of diet, physical activity, hormone and growth factor levels are being evaluated for their effect on the rate of bone loss. A sophisticated rat model of postmenopausal osteoporosis has allowed investigation of the influences on bone at the molecular level. Local factors have been identified which protect bone from the increased cellular activity induced by oestrogen deficiency.16 The Detectacol bowel screening program was one of the earliest Australian voluntary bowel screening programs using an immunologically based test to detect faecal occult blood.17 It has been running for 15 years and has been endorsed through randomised controlled trials showing a reduction in mortality from colorectal cancer as a result of such screening techniques.18 Bone metastasis research. In collaboration with the Department of Surgery of the University of Adelaide, the Division is using a rat tumour model to study how metastatic tumours develop at incision sites after laparoscopic surgery. They will investigate changes in the immune state of the peritoneum and the effect of the different gases used to expand the peritoneum during laparoscopy. Division of Haematology Bone marrow transplantation. Haemopoiesis research began in the early 1980s, led by Dr Chris Juttner and Dr Bik To. They pioneered blood stem cell mobilisation and transplantation, which has now almost replaced autologous bone marrow transplantation.19 The group is currently seeking to define the role of integrin and bone resorption in mobilisation. Work continues on using haemopoietic stem cells purified using immunomagnetic techniques to refine the application in allogeneic and autologous transplantation in leukaemia, lymphoma and myeloma. Ex-vivo expansion is another project for which the Division has a pioneering role as a way of improving the safety of transplantation.20 Dr Tim Hughes is leading a study of chronic myeloid leukaemia and has established a SCID-NOD model for characterising leukaemic stem cells.21 The group is currently involved in 12 clinical trials in transplantation for leukaemia and lymphoma. Research continues into mesenchymal stem cell and cellular therapies, cell death differentiation and oncogenesis and cellular interactions. The stromal group, led by Dr Paul Simmons, has made a major contribution on STRO-1-positive marrow stromal cell percursor and its role in bone formation, haemopoietic regulation and cellular therapies using marrow stromal tissue. Dr Sharad Kumar's group is studying apo ptosis, including the role, activation and regulation of caspases, the function of homologue(s) of C elegans CED-4 and the identification of CED-4M. The group is studying how integrins may modulate cytokine response and function of tetraspan and GPI-linked molecules. The Hanson Centre for Cancer Research The RAH Board and the IMVS Council have strongly supported the development of the Hanson Centre. Through fundraising, the RAH has provided $7.5 million over the past decade for the construction of buildings and provision of equipment for the Centre. The IMVS provides infrastructure funds from its operations and through its commercial company, Medvet Science Pty Ltd, created by the IMVS Council to commercialise the intellectual property and other biomedical activities of the IMVS/RAH campus, and has contributed in excess of $2 million in research grants and infrastructure funding to the Centre. When the Centre began, in 1991, the core research groups were the RAH Division of Human Immunology and the Division of Haematology, but other groups have since joined, including the University of Adelaide Department of Cardiology and the RAH Departments of Gastroenterology, Rheumatology and Thoracic Medicine. These groups have their research support within the Hanson Centre, but are integrated into the service and teaching activities of the IMVS, RAH and the University of Adelaide. Professor Mathew Vadas, appointed 10 years ago to head the Division of Human Immunology, has developed strong basic research groups of excellence which have become known as the Hanson Centre for Cancer Research. This Centre is now the jewel in the research crown of the IMVS. Professor Vadas's research is focused on vascular biology.22,23 Other groups, led by Dr Shannon and Dr Cockerill, have had success in gene regulation,24 and Dr Gonda's laboratory is working on oncogene research.25 The Lipid Research Laboratory has succeeded in understanding the structural principles that govern the assembly of high density lipoprotein.26,27 It was known from epidemiological studies that the human blood high density lipoprotein fraction was protective against atherosclerosis, but the structural and functional basis for this protection was largely unknown. In collaboration with the Vascular Biology Laboratory, they discovered a novel and likely vital function for this particle that directly regulates the function of the vessel wall, and hence the propensity for atherosclerosis.23 Significant progress has been made in understanding cell death. The orderly and timely death of cells is essential in giving shape to the body and renewing its worn parts. A disorder in this process leads to abnormal growth that is a component of cancer. Work in the Molecular Regulation Laboratory, headed by Dr Sharad Kumar, and the Cytokine Laboratory of Dr Angel Lopez has provided novel insights into this process.28 The signals that are involved in "sentencing" and "execution" leading to cell death have been elaborated, and a new therapeutic agent based on an actual growth factor has been discovered that induces this process in blood cells. It is likely that the clinical implications of these discoveries will be considerable. The future of IMVS The IMVS will maintain a business-like approach to the management of operations and to the continued development of Medvet Science. A recent external review of research on the IMVS/RAH/University of Adelaide campus, led by Professor Lawrie Powell, has commented on the excellent productivity of research on the campus. The IMVS intends to further enhance research through expansion of laboratories for basic research while maintaining the strong integration of teaching, patient services and clinical research. New areas of research are being developed in the neurosciences and in molecular pathology, with an emphasis on cancer genetics. References The Institute of Medical and Veterinary Science. Annual Report 1952. Adelaide, IMVS: 1952: 5. The Institute of Medical and Veterinary Science. 50th Anniversary Review 1938-88. Nicholson B, editor. Adelaide, IMVS: 1988: 33-34. Fraser RD, Osti OL, Vernon-Roberts B. Discitis following chemonucleolysis -- an experimental study. Spine 1986; 11: 679-687. Fraser RD, Osti OL, Vernon-Roberts B. Iatrogenic discitis -- the role of intravenous antibiotics in prevention and treatment. An experimental study. Spine 1989; 14: 1025-1032. Fazzalari NL, Crisp DJ, Vernon-Roberts B. Mathematical modelling of trabecular bone structure; the evaluation of analytical and quantified surface to volume relationships in the femur and iliac crest. J Biomech 1989; 22: 901-910. Moore RJ, Durbridge TC, McNeil PJ, et al. Trabecular spacing in postmenopausal Australian women with and without vertebral fractures. Aust N Z J Med 1992; 22: 269-273. Howie DW, Cornish BL, Vernon-Roberts B. Resurfacing hip arthroplasty classification of loosening and the role of prosthesis wear particles. Clin Orthop 1990; 255: 144-159. Burrell CJ, Mackay P, Greenaway PJ, Hofschneider MK. Expression in Escherichia coli of hepatitis B virus DNA sequences cloned in plasmid pBR322. Nature 1979; 279: 43-47. Steele P W, Lanser JA, Sangster N. Isolation of Legionella longbeachae serogroup I from potting mixes. Appl Environ Microbiol 1990; 56: 49-53. Paton AW, Ratcliff RM, Doyle RM, et al. Molecular microbiological investigation of an outbreak of hemolytic-uremic syndrome caused by dry fermented sausage contained within Shiga-like toxin-producing Escherichia coli . J Clin Microbiol 1996; 34: 1622-1627. Li P, Stephenson AJ, Kuiper LJ, Burrell CJ. Double-stranded strong-stop DNA and the second template switch in human immunodeficiency virus (HIV) DNA synthesis. Virology 1993; 194: 82-88. Triyatni M, Jilbert AR, Qiao M, et al. Protective efficacy of DNA vaccines against duck hepatitis B virus infection. J Virol. In press. Morris HA, Cleghorn DB, Need AG, et al. The 5-year reproducibility of calcium-related biochemical variables in postmenopausal women. Scand J Clin Lab Invest 1995; 55: 383-389. Nordin BEC, Morris HA, Need AG, et al. Dietary calcium and osteoporosis. In: Pietinen P, Nishida C, Khaltaev N, editors. Proceedings of the 2nd World Health Organization Symposium on Health Issues for the 21st Century: nutrition and quality of life. Kobe, Japan, 1993. Geneva: WHO, 1996: 181-198. Need AG, Horowitz M, Stiliano A, et al. Vitamin D receptor genotypes are related to bone size and bone density in men. Eur J Clin Invest 1996; 26: 793-796. Morris HA, Porter SJ, Durbridge TC, et al. Effects of oophorectomy on biochemical and bone variables in the rat. Bone Mineral 1992; 18: 133-142. Bone HA, Roder D, McCaul K, Milliken L. Epidemiology of cancer in South Australia. Adelaide: South Australian Cancer Registry, 1993: 57-63. Weller D, Thomas D, Hiller J, et al. Screening for colorectal cancer using an immunochemical test for faecal occult blood: results of the first 2 years of a South Australian programme. Aust N Z J Surg 1994; 64: 464-469. To LB, Haylock DN, Simmons PJ, Juttner CA. The biology and clinical uses of blood stem cells. Blood 1997; 89: 2233-2259. Haylock DN, To LB, Dowse TL, et al. Ex vivo expansion and maturation of peripheral blood CD34 + cells into the myeloid lineage. Blood 1992; 80: 1405-1412. Hughes T, Iverson PO, Lewis ID, et al. A second generation GM-CSF analogue that prevents dissemination and induces remission of human juvenile myelomonocytic leukaemia in engrafted immunodeficient mice. Blood. In press. Korpelainen EI, Gamble JR, Vadas MA, Lopez AF. IL-3 receptor expression, regulation and function in cells of the vasculature. Immunol Cell Biol 1996; 74: 1-7. Cockerill GW, Rye KA, Gamble JR, et al. High density lipoproteins inhibit cytokine-induced expression of endothelial cell adhesion molecules. Arterioscler Thromb Vasc Biol 1995; 15: 1987-1994. Cockerill PN, Osborne CS, Bert AG, Grotto JRM. Regulation of GM-CSF gene transcription by core-binding factor. Cell Growth Differ 1996; 7: 917-922. Gonda TJ, Favier D, Ferrao D, et al. The c-Myb negative regulatory domain. In: Wolff L, Perkins AS, editors. Molecular aspects of myeloid stem cell development. Curr Top Microbiol Immunol 1996; 211: 99-107. Rye KA, Hime N, Barter PJ. The influence of sphingomyelin on the structure and function of reconstituted high density lipoproteins. J Biol Chem 1996; 271: 4243-4250. Rye KA, Hime N, Barter PJ. Evidence that cholesteryl ester transfer protein-mediated reductions in reconstituted high density lipoprotein size involve particle fusion. J Biol Chem 1997; 272: 3953-3960. Kumar S, Kinoshita M, Noda M. Characterization of a mammalian cell death gene Nedd2. Leukemia 1997; 11 Supp 3: S385-S386. Authors' details The Institute of Medical and Veterinary Science, Adelaide, SA. Brendan J Kearney, FRACP, FRACMA, Director. Reprints: Professor B J Kearney, Director, Institute of Medical and Veterinary Science, PO Box 14, Rundle Mall, SA 5000. - ©MJA 1997 Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". <URL: http://www.mja.com.au/> © 1997 Medical Journal of Australia.

Brendon J Kearney

Mammographic screening: results from the 1996 National Breast Health Survey Alexandra L Barratt, Jill Cockburn, Sally Redman, Christine Paul and Janice Perkins For editorial comment, see Glasziou & Irwig This article has been cited in Rodger A, Kavanagh AM, Outcome measures of an Australian breast-screening program, MJA 1998; 169: 179-180 Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au>". Abstract - Introduction - Methods - Results - Discussion - References - Authors' details - - - ©MJA1997 Abstract Objective: To establish the extent of women's knowledge of mammographic screening, particularly in relation to the national screening program, BreastScreen Australia, and to estimate the proportion of women who are participating in screening both within and outside BreastScreen Australia. Design and setting: Validated prospective telephone survey of women aged 30-69 years selected at random from across Australia. Participants: 2935 women with no previous breast cancer diagnosis. Results: The adjusted response rate was 64%. Almost 90% of women had heard of the national program; only 1% correctly stated that screening is for asymptomatic women. 60% correctly identified the current recommended age of starting screening is about 50 years of age; 26% thought screening should begin at about 40 years of age. Approximately 60% correctly reported that the recommended screening interval is every two years; 27% thought screening should be done annually. 55% reported ever having had a mammogram, and 37% reported having had at least one screening mammogram. Among women in the target age group (50-69 years) about 70% reported ever having had a screening mammogram, and about 50% reported having had a screening mammogram within the national program in the last two years. Among women aged 40-49 years, 29% reported ever having had a screening mammogram, and 22% reported having been screened in the last two years. Conclusions: Awareness of the national screening program is high, but some women do not know the purpose of screening, the target age group and the recommended screening interval. Compliance with screening is good among women in the target age group; many women in their 40s are also participating in screening. MJA 1997; 167: 521-524 Introduction Breast cancer is the most common cause of cancer death in Australian women, with approximately 7000 new cases and over 2500 deaths annually.1 Evidence from randomised controlled trials shows that mammographic screening can substantially reduce mortality from breast cancer among women aged 50-69 years,2,3 but high rates of compliance with screening are necessary to achieve these benefits. Since 1991 the Commonwealth and State governments have funded a national program of mammographic screening, BreastScreen Australia, which provides free mammographic screening for women aged 40-79 years, but specifically targets women aged 50-69 years. Clearly, if women are to participate, they need to know about mammographic screening generally, and about the service provided by the national program in particular. In 1996, we undertook a national survey on behalf of the National Health and Medical Research Council National Breast Cancer Centre (NBCC) to establish baseline information about women's knowledge of, and participation in, screening mammography. Limited data about participation within the national program are available from the national program's own evaluation reports. Little is known about participation in screening outside the national program. Our aims were, firstly, to determine Australian women's knowledge of mammographic screening and the national program, and whether knowledge differs by area of residence or education. Secondly, we aimed to estimate the proportion of women who are participating in screening both within and outside BreastScreen Australia. Methods The survey was conducted by telephone by a market research company on behalf of the NBCC in April 1996. The questionnaire was developed and pilot-tested by the NBCC; test-retest reliability was assessed by administering the questionnaire twice to a random sample of 100 women at a one-week interval, with good agreement for key items ( k = 0.89-1.00). Telephone numbers were chosen at random from the electronic white pages telephone directory. Up to five callbacks were made to contact eligible women (those aged 30-69 years whose English was sufficient to complete a telephone interview). The sampling frame was stratified to ensure participation by women in all States and Territories, and in urban and rural areas.4 The outcome of every phone call was recorded. Ethical approval for this study was obtained from the NSW Cancer Council Ethics Committee in March 1996. Validation A random sample of 150 women who reported having been screened by BreastScreen Australia were asked for consent to check their questionnaire answers against BreastScreen Australia records. For those who consented, identifying data and the estimated date of last screen within the program were forwarded to BreastScreen Australia for confirmation. To assess selection bias, one survey question (Have you ever had a mammogram?) was placed in the February 1996 Australian Bureau of Statistics (ABS) Population Survey Monitor, which is conducted quarterly by household interview. The question was asked only of women 30-69 years in the ABS sample, and the results were weighted as for our survey (described below) to give a comparable estimate. Statistical analysis Our data were weighted for area of residence and age based on ABS projected population estimates. As all percentages reported in this paper are weighted results, sample sizes are not routinely given in the results. Significance testing was done by chi-squared tests based on the sample size of the survey; P values of less than 0.05 were considered statistically significant. Ninety-five per cent confidence intervals were calculated from the standard error of the proportion, again based on the sample size of the survey. Education was coded according to ABS categories and collapsed to those with post-school qualifications and those without post-school qualifications.5 Age was classified as younger (30-49 years) and older (50-69 years) to differentiate women within and outside the target age range of the national program. Results The raw response rate was 70.2% (calculated as completed interviews [3014] over contacted eligible women [4291]). After inflating the denominator by the estimated number of eligible women among those who could not be contacted (n = 389), the response rate was 64.4%. Seventy-nine women who had previously been diagnosed with breast cancer were excluded. The demographic characteristics of the remaining women are shown in Box 1. Of the 150 women asked, 127 (85%) consented to having their screening records checked. Of 126 women who reported having been screened within BreastScreen Australia within the last two years, attendance was confirmed for 121 (96%). Twenty-six per cent of women were able to recall the mammogram date accurately to the day, and a further 38% accurately recalled the month. Ninety-two per cent of women reported the mammogram date accurately to within 12 months. The weighted ABS estimate for the question Have you ever had a mammogram? was 50% (95% CI, 46.9-53.2). Responses to specific questions Do you know any ways to detect breast cancer in the early stages? -- 63.2% of respondents nominated mammography without prompting. Those more likely to nominate mammography were older women (chi-squared = 14.99; P < 0.01) and women with post-school qualifications (chi-squared = 12.05; P< 0.01). The Commonwealth and State governments fund a mammographic screening program in Australia known as the National Program for the Early Detection of Breast Cancer. In [respondent's State] it is known as [program name in respondent's State] . Have you heard of this Program? -- 88.7% of women responded that they had heard of the national program. More women in rural areas than in urban areas had heard of the program (92.9% v. 87.0%; chi-squared = 14.82; P< 0.001). More older than younger women had heard of the program (93.1% v. 86.2%; chi-squared = 42.09; P< 0.001). There was no difference in the proportion who had heard of the program by educational level. Which women do you think should attend for mammographic screening; only women who have no symptoms of breast cancer ; only women who have symptoms of breast cancer ; or it doesn't matter whether women have symptoms or not? -- only 1.3% of women correctly replied that screening was for women without symptoms, 2.9% replied that it was for women with symptoms, and 94.8% replied that it didn't matter. Slightly more younger women than older women thought screening was for women with symptoms (3.6% v. 2.0%; chi-squared = 11.70; P< 0.009). There were no differences between women living in urban and rural areas or according to educational level. When does the government program recommend women should start (stop) having mammograms? -- 60.3% of women correctly reported that the recommended commencement age was 50-59 years; 26.0% reported it was 40-49 years (see Box 2). More younger women than older women reported that the recommended commencement age was 40-49 years (28.1% v. 22.5%; chi-squared = 58.73; P= 0.001). Again, there were no differences in these proportions between women living in urban and rural areas or according to educational level. Most women (55.5%) thought screening should never cease, and 19.5% of women were unable to say when screening should stop (Box 2). How often does the government screening program suggest women should have a screening mammogram? -- 58.7% correctly reported that the government program suggests screening should be done every two years, but 26.7% reported that annual screening is recommended. More younger women than older women reported that annual screening is recommended (32.1% v. 17.4%; chi-squared = 178.76; P< 0.001). Referral and access to the government program Of respondents in the target age group (50-69 years), 34.7% reported that their GP had suggested they have a mammogram as part of the government program. This proportion did not vary by area of residence or by education level. Among all respondents, almost all women (95.8%) reported having visited a GP within the last two years. Among women in the target age group, 84.0% knew that a screening mammogram within the national program was free. More women with post-school qualifications knew this (87.9% v. 80.9%; chi-squared = 9.30; P= 0.01), but the proportions did not vary according to area of residence. Most women in the target age group (65.8%) reported being very confident of knowing how to make an appointment for a mammogram within the program, and this proportion did not differ according to area of residence or educational level. Participation in screening mammography A series of questions was used to determine whether women had had a mammogram within or outside the national program. To determine whether it had been a screening mammogram, they were then asked, When you had [that mammogram] was it because you already thought you had some problems with your breasts?. Overall, 55.0% of all women reported ever having had a mammogram (Box 3). The frequency of having had a mammogram rose steeply with age, peaking in the 50-59 years age group at 86.5%. Overall, 36.6% of women reported having had at least one screening mammogram (Box 3). Among women aged 30-39 years, approximately 6% reported having been screened, and this proportion rose to approximately 70% in the 50-69 years target age group. Having had a screening mammogram was not associated with area of residence or with education. In the target age group (50-69 years), 49.5% of women reported being screened in the national program within the last two years, and this was not influenced by area of residence or educational level. Another 12.1% reported having been screened outside the national program within the last two years. Among women 30-39 years of age, only 3% had been screened in the last two years, mostly outside the national program. In the older age groups, most women reported being screened within the national program (Box 4). Discussion Our findings are the first validated, baseline estimates of Australian women's knowledge of, and participation in, mammographic screening since the establishment of the national program. Our adjusted response rate of 64% compares well with those of similar reported surveys.6-8 The estimate by the ABS for the identical question Have you ever had a mammogram? was 50% (95% CI, 46.9-53.2) compared with ours of 55% (95% CI, 53.2-56.8), suggesting that our estimates are likely to be only minimally affected by selection bias. However, we excluded some women, notably non-English speakers and those living in homes without telephones. Also, as it is likely that women who chose to participate were more aware of breast cancer and screening mammography than those who refused, our data may overestimate both women's knowledge of, and participation in, such screening. Awareness of the national program was high (nearly 90%). Although asking a prompted question may have inflated this estimate, the result does suggest that the promotion of the program has been effective. The vast majority of women knew that the national program is free, and that it recommends women start screening at around 50 years of age and have a screening mammogram every two years. Most women also felt very confident in knowing how to access the program. The absence of significant variation in knowledge about screening mammography according to area of residence or educational level suggests that the national program has been successful in widely promoting screening. However, consistent with a previous study,9 most women were not aware that screening is intended only for asymptomatic women. Most women thought symptoms didn't matter, suggesting they appreciate to some extent that the national program is not designed for the evaluation of symptoms. While this may be encouraging, it is still of some concern as people who do not appreciate the distinction between diagnostic and screening tests may be more likely to seek compensation for missed diagnoses.9 As most women reported that their GPs had never recommended they attend the national program for screening, there is a greater opportunity for GPs to recruit more women to screening. Our findings indicate that attendance for mammography is becoming increasingly common. Surveys conducted in 1988 found that only 18% of women spontaneously mentioned mammography as a means of detecting breast cancer in the early stages, and 50% claimed to have heard of it when prompted.10 Only about 10%-20% of women had ever had a mammogram, about half for diagnostic and half for screening purposes.7,11 Clearly, this situation has changed dramatically, with 60% of women now nominating mammography as a means of early detection in an unprompted question, 55% of all women reporting having ever had a mammogram, and 37% of women reporting having had at least one screening mammogram. High proportions of women in the age group targeted by the national program reported having been screened at least once (70%) and having been screened in the last two years (about 60%). Many more of these women are being screened as part of the national program than outside it. Our finding that 49.8% of women in the target age group had been screened within the national program in the last two years is consistent with BreastScreen Victoria's estimated participation rate of 47.7%.12 Recent meta-analyses showing a small reduction in mortality after about 10 years of screening for women aged 40-49 years13 have prompted heated debate over whether younger women should also be targeted for screening. It is interesting to note that we found quite high screening rates among women aged 40-49 years, indicating that many women in this age group are, in fact, already being screened. References Taylor R, Smith D, Hoyer A, et al. Breast Cancer in New South Wales 1972-1991. Sydney: NSW Cancer Council, 1994. Kerlikowske K, Grady D, Rubin SM, et al. Efficacy of screening mammography. A meta-analysis. JAMA 1995; 273: 149-154. Glasziou P, Woodward AJ, Mahon CM. Mammographic screening trials for women aged under 50. A quality assessment and meta-analysis. Med J Aust 1995; 162: 625-629. Department of Primary Industries and Energy, Department of Human Services and Health. Rural, remote and metropolitan areas classification. Canberra: AGPS, 1994. Australian Bureau of Statistics. Labour force status and educational attainment, Australia. Canberra: Australian Bureau of Statistics, 1994. (Catalogue No. 6235.0.) Etzi S, Lane DS, Grimson R. The use of mammography vans by low-income women: the accuracy of self-reports. Am J Public Health 1994; 84: 107-109. Hill D, White V, Borland R, Cockburn J. Cancer-related beliefs and behaviours in Australia. Aust J Public Health 1991; 15: 14-23. Smith W, Chey T, Jalaludin B, et al. Increasing response rates in telephone surveys: a randomized trial. J Public Health Med 1995; 17: 33-38. Cockburn J, Redman S, Hill D, Henry E. Public understanding of medical screening. J Med Screening 1995; 2: 224-227. Cockburn J, Murphy B, Schofield P, et al. Development of a strategy to encourage attendance at screening mammography. Health Ed Res Theory Practice 1991; 6: 279-290. Irwig L, Cockburn J, Turnbull D, et al. Women's perceptions of screening mammography. Aust J Public Health 1991; 15: 24-32. BreastScreen Victoria: 1995 annual statistical report. Melbourne: BreastScreen Co-ordination Unit, 1996. Breast-cancer screening with mammography in women aged 40-49 years. Report of the Organizing Committee and Collaborators, Falun Meeting, Falun, Sweden 1996. Int J Cancer 1997; 68: 693-699. (Received 11 Mar, accepted 12 Aug, 1997) Authors' details Department of Public Health and Community Medicine, University of Sydney, NSW. Alexandra L Barratt, PhD, FAFPHM, Lecturer. Discipline of Behavioural Science in Relation to Medicine, University of Newcastle, NSW. Jill Cockburn, PhD, Head. National Health and Medical Research Council National Breast Cancer Centre, Sydney, NSW. Sally Redman, PhD, Director; Christine Paul, PhD, Consultant Behavioural Scientist. New South Wales Cancer Council Cancer Education Research Program, Wallsend, NSW. Janice Perkins, PhD, Research Associate. Currently at the Discipline of Behavioural Science in Relation to Medicine, University of Newcastle, NSW. No reprints will be available. Correspondence: Dr A L Barratt, Department of Public Health and Community Medicine, Building A27, University of Sydney, NSW 2006. - ©MJA 1997 <URL: http://www.mja.com.au/> © 1997 Medical Journal of Australia.

Evaluation of the ThinPrep Pap test as an adjunct to the conventional Pap smear Jennifer M Roberts, A Marion Gurley, Julia K Thurloe, Ronald Bowditch and Colin R A Laverty For editorial comment, see Wain Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au/>". Abstract - Introduction - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - ©MJA1997 Abstract Objective: To evaluate the ThinPrep Pap test as an adjunct to the conventional Pap smear. Design and setting: Prospectively collected cervical samples were split for independent screening at a large specialised private gynaecological pathology practice in Sydney. Main outcome measures: Detection of additional significant abnormalities (cervical intraepithelial neoplasia 1, or more severe); changed management recommendations from "repeat smear in 12 months" or "...six months" to "colposcopy"; a reduction in unsatisfactory reports. Results: 35 560 paired (split-sample) conventional and ThinPrep slides were prepared. Significant abnormalities were detected in 724 conventional smears (2%). Additional significant abnormalities were found in 85 ThinPrep slides whose corresponding conventional smear was negative or unsatisfactory even after review, representing a 12% increase in the detection of significant abnormalities. As a result of the addition of ThinPrep, management recommendations were changed from "repeat smear in 12 months" or "...six months" to "colposcopy" for 89 of 1669 women whose conventional Pap smears showed minor non-specific changes or papillomavirus. There were 1258 conventional smears (3.5%) that were unsatisfactory compared with 235 ThinPrep slides (0.7%); for only 74 samples (0.2%) were both slides unsatisfactory. Conclusions: The addition of the ThinPrep Pap test improves detection and clinical management of cervical abnormalities, and reduces the number of unsatisfactory samples which would otherwise require repeat tests. MJA 1997; 167: 466-469 Introduction The conventional Pap smear has been the mainstay of screening for cervical precancer for approximately 50 years, without major changes in techniques relating to preparation and interpretation. In recent years there has been increasing pressure to improve Pap smear standards to reduce the occurrence of false negative results, with their associated morbidity and mortality for the women involved, and legal consequences for health professionals.1,2 Many false negative cases are the result of problems of sample transfer and smear quality.3,4 The conventional Pap smear represents only a small subsample of the cellular material collected on the sampling implement,5,6 with the remainder traditionally being discarded. Importantly, this small subsample is a non-random fraction of the material collected and is not necessarily representative of the whole. The traditional smearing technique also produces slides which vary greatly in quality, making it sometimes difficult to find and interpret abnormal cells. Improvements in both subsample selection and slide quality should therefore reduce the number of false negative cases. The ThinPrep Pap test is an automated technique developed with the aim of producing better quality slides.7 Each cervical sample is rinsed into a vial of fixative fluid. In the laboratory the sample is homogenised, and a representative subsample is collected on a single-use filter, transferred to a glass slide, and stained in the routine fashion. This technique results in a better quality preparation, as the problems of poor fixation, uneven thickness of the cellular spread, and obscuring of cells by blood and inflammatory exudate are overcome. We aimed to evaluate the impact of this technology on detection rates of significant abnormalities, accuracy in reporting, and numbers of unsatis - fact ory preparations when it was used in addition to the conventional Pap smear. Methods In a pilot study in the first half of 1996, we evaluated 1051 samples to test the performance of the ThinPrep 2000 machine (Cytyc Corporation, Boxborough, Mass., USA), as previous experience with an earlier model had highlighted some technical problems.8 The pilot study showed slides of uniformly high quality and so, in July 1996, the test was offered to all referring practitioners at a charge to patients of $20.30. For this study, we analysed paired samples collected between 1 July 1996 and 30 May 1997 by practitioners who chose to offer ThinPrep (TP) to their patients as an adjunct test. We used a split-sample protocol in which routine samples of the transformation zone (squamo-columnar junction) -- the area of the cervix where abnormalities are most likely to occur -- were taken with a broom-like device, the Cervex brush (Rovers BV, Netherlands). We insisted on the use of this implement as our earlier study8 had shown that sampling with a spatula was more likely to result in a TP slide with insufficient cells. For postmenopausal women and those who had had previous cervical surgery or ablation, an endocervical brush sample was added. After a conventional Pap smear was made, the sampling implement was vigorously rinsed in a vial of PreservCyt fixative fluid (Cytyc Corporation, Boxborough, Mass., USA). TP slides were prepared from the vial by the ThinPrep 2000 machine and both slides were stained routinely. A duplicate of each request form was made, and the conventional smear and the TP slide were examined independently by different cytotechnologists. Nineteen of our 28 cytotechnologists were involved in reading the TP slides; all 28 read conventional smears. Any slide showing minor abnormalities was reviewed by a second experienced cytotechnologist. If cervical intraepithelial neoplasia (CIN) of any grade was detected, or if a high grade lesion was suspected, the case was referred to a cytopathologist for review. Slide pairs (conventional Pap smear and TP) displaying a significant discrepancy were also reviewed. The final report issued to the referring doctor was structured according to the National Health and Medical Research Council (NHMRC) reporting terminology9 (see Box 1) and contained any relevant information obtained from both slides, together with the relevant management recommendations. For the purposes of analysis, separate final results of the two slides were also recorded. A significant abnormality was defined as one for which the management recom mendation is "colposcopy" (see Box 1). This includes any grade of CIN, adenocarcinoma in situ (AIS), invasive carcinoma or the suspicion of a high grade abnormality ("inconclusive" in NHMRC terminology). Results Approximately 500 referring practitioners chose to offer TP to their patients as an adjunct test. Some of these doctors offered the TP test to all women, and others only to selected women. Thirty-five per cent of women having non-screening cytological tests had TP, whereas only 24% of women having routine cytological tests had TP. Currently, about 30% of all smears received in this laboratory are accompanied by a TP sample. We received 35 560 paired split-sample conventional Pap smear slides and TP samples for analysis in this study. Box 1 presents a comparison of final (i.e., reviewed) results for TP and conventional Pap smear slides. Of the 34 141 paired results that were satisfactory for both slides, 32 195 (94.3%) showed total agreement and 1946 (5.7%) did not. Of the latter group, the TP slide showed a more severe abnormality than the conventional slide in 1194 (61.4%), and the opposite was true in 752 (38.6%; chi-squared = 100.4; df = 1; P < 0.001). For the 1946 paired slides which did not show total agreement, 271 had substantial discrepancies in which colposcopy would have been recommended based on the result of one test, but not the other. Of these, colposcopy would have been recommended on the basis of the TP result alone for 167 (61.6%), and on the basis of the conventional Pap smear result alone for 104 (38.4%; chi-squared = 14.6; df = 1; P < 0.001). Significant abnormalities were detected in 724 conventional smears. An additional 85 significant abnormalities were detected on TP slides for which the corresponding conventional slides, even after review, were either negative (78) or unsatisfactory (7). This represents a 12% increase in the detection of significant abnormalities. There were 14 TP slides that predicted high-grade epithelial abnormality (HGEA), while the corresponding conventional smears were either negative or unsatisfactory. Histological follow-up of 11 of these confirmed HGEA in eight, and low-grade epithelial abnormality (LGEA) in three. There were also 27 TP slides reported as inconclusive, while the corresponding conventional smears were either negative or unsatisfactory. Histological follow-up of 17 of these showed HGEA in seven, LGEA in eight, and inflammation in only two. Further, 15 conventional smears originally reported as negative, and reviewed because their TP counterparts showed an abnormality, were subsequently reported as HGEA or inconclusive. These "screening" false negatives are not represented in Box 1, which shows only final results (i.e., after review). ThinPrep slides not only increased detection of abnormalities, but also influenced management recommendations. On the basis of conventional smear tests, the recommendations for 1669 women (4.7%) with minor non-specific changes or features of human papillomavirus (HPV) effect were "repeat smear in 12 months" or "...six months", respectively. For 89 of these women (5.3%), this recommendation was changed to "colposcopy" as a result of a higher grade abnormality being detected on the TP slide. Histological follow-up concentrated on the predicted HGEA and inconclusive categories. Findings are shown in Box 2 (below). Of conventional smears, 1258 (3.5%) were found to be unsatisfactory, but only 74 (0.2%) were unsatisfactory by both methods, representing a 94% reduction in unsatisfactory reports. TP predicted HGEA in four smears reported as unsatisfactory on the conventional smear; all four have been histologically confirmed to be HGEA. Minor abnormalities (minor non- specific changes/HPV) were detected in 1669 (4.7%) of the conventional smears and 2027 (5.7%) of the TP slides. An endocervical component was absent in 8.3% of the conventional smears and in 20.0% of the TP slides. In 6.4% of cases, neither slide displayed an endocervical component. With respect to glandular abnormalities, 13 cases of adenocarcinoma in situ (AIS) or adenocarcinoma were detected by both methods. There was one histologically confirmed case of adenocarcinoma in which the conventional smear was reported as inconclusive (suspicious of AIS) and the corresponding TP slide was negative. Further TP slides were made from the remaining sample in the vial but none of these contained abnormal cells. Discussion Overall, we found that diagnostic agreement between the conventional Pap smear and the TP Pap test was high. Where there was disagreement, a higher grade abnormality was predicted significantly more often by TP than by the conventional smear. The addition of the TP test resulted in detection of additional histologically confirmed HGEA, leading to a reduction in the number of false negatives and a corresponding increase in sensitivity. The fact that the same proportion of HGEA (83%) was histologically confirmed for both conventional Pap smear and TP shows that TP does not result in undue overreporting, and that, although sensitivity has been increased, the positive predictive value of the test has been maintained. The extra significant abnormalities detected on TP and the increased accur acy of reporting are mainly attributable to improved subsampling of the cervical specimens, resulting in fewer "subsampling" false negatives. Computer-assisted rescreening of conventional smears would have no impact on this area. "Screening" false negatives (although few in this study) will also be reduced by the addition of a slide of much higher quality, and by the independent screening of additional material. That there were cases in which an abnormality detected on conventional smear was not detected on TP is not surprising, as the split-sample protocol favours the conventional smear. In these cases it appears that all abnormal cells in the sample were transferred to the conventional smear, leaving none in the TP vial. A "direct-to-vial" protocol in which 100% of the material collected is rinsed into the fluid fixative should eliminate this problem. However, if only a very small number of abnormal cells are present in the vial, then they may not be represented on any single TP slide. The addition of the TP test resulted in a large reduction in reporting of unsatisfactory preparations. This is a direct result of removal of the many variables associated with the conventional Pap smear technique (poor fixation, uneven thickness of the cellular spread, and obscuring of cells by blood and inflammatory exudate), and represents significant time and cost savings, as the need for a repeat smear is avoided. Significantly, there were four women with histologically confirmed HGEA which was predicted by TP, but whose conventional smears were reported as unsatisfactory. These abnormalities may have been detected by the recommended repeat smear three months later, but there is always a risk that women will not return for follow-up. A very small percentage of TP slides were unsatisfactory and this was always the result of there being insufficient cells in the specimen. In these cases the machine reported a dilute specimen. We believe that a number of these cases were the result of the use of incorrect sampling implements, contrary to the recommended protocol. Of some surprise to us was the increase in reporting of minor non-specific changes. While some of this no doubt represents a genuine increase in detection of minor abnormalities, we feel that other factors must be contributing. In particular, we believe that the improved nuclear morphology obtained in the TP process requires cytotechnologists and cytopathologists to "relearn" subtle criteria used at this end of the diagnostic spectrum. We noted this trend early in the study period and have been addressing it with numerous educational sessions. We continue to monitor reporting in this category. The addition of TP tests resulted in a decreased proportion of combined reports lacking an endocervical compon ent. However, TP alone had a higher rate of absent endocervical compon ent. The reason for this is unclear, but may relate to TP slides being prepared from left-over cellular material. We can only speculate that the "endocervical component present" rate may improve in a direct-to-vial situation. Even though there were more TP slides lacking an endocervical component, there was only one case in which there was a significant glandular abnormality present in the conventional smear and not in the TP slide. As further slides prepared from the TP vial also failed to show abnormal cells, we assume that no abnormal cells were present in the vial. An important advantage of the TP process is the presence of further cellular material in the fluid fixative which can be used to prepare more slides, or for HPV typing which may have clinical relevance in the triage of patients with low grade abnormalities.10 In the United States, the Food and Drug Administration has approved the TP Pap test as a replacement for the conventional Pap smear.11 This decision was based on analysis of data on 7360 paired samples derived from six different centres (Data on file, Cytyc Corporation, Boxborough, Mass., USA). Our data, on a larger sample from a single practice, support the assertion that the TP Pap test performs substantially better than the conventional smear on analysis of significant parameters. While we continue to offer the TP Pap test as an additional proced ure, we anticipate that the machine-made slide may replace the conventional smear in the future. Acknowledgements We thank Tabatha Lovelace, Lisa Wong, Rozanne Van Gramberg and Samira Bounassif for technical assistance. Statement of potential conflict of interest: The Cytyc Corporation (Boxborough, Mass., USA) lent us one of the Thinprep 2000 machines used in this study. The company was not involved in the design of the study, collection of data, analysis of results or preparation of the manuscript. References DeMay RM. To err is human -- to sue American. Diagn Cytopathol 1996; 15: iii-vi. Skoumal SM, Florell SR, Bydalek MK, Hunter WJ. Malpractice protection: communication of diagnostic uncertainty. Diagn Cytopathol 1996; 14: 385-389. Gay JD, Donaldson LD, Goellner JR. False-negative results in cervical cytologic studies. Acta Cytol 1985; 29: 1043-1046. Joseph MG, Cragg F, Wright VC, et al. Cyto-histological correlates in a colposcopic clinic: a 1 year prospective study. Diagn Cytopathol 1991; 7: 477-481. Hutchinson ML, Isenstein LM, Goodman A, et al. Homogeneous sampling accounts for the increased diagnostic accuracy using the ThinPrep Processor. Am J Clin Pathol 1994; 101: 215-219. Goodman A, Hutchinson ML. Cell surplus on sampling devices after routine cervical cytologic smears. A study of residual cell populations. J Reprod Med 1996; 41: 239-241. Zahniser DJ, Hurley AA. Automated slide preparation system for the clinical laboratory. Cytometry 1996; 26: 60-64. Laverty CRA, Thurloe JK, Redman NL, Farnsworth A. An Australian trial of ThinPrep: a new cytopreparatory technique. Cytopathology 1995; 6: 140-148. National Pathology Accreditation Advisory Council, Commonwealth Department of Health and Family Services. Requirements for gynaecological (cervical) cytology. Canberra: AGPS, 1997. Sherman ME, Schiffman MH, Lorincz AT, et al. Cervical specimens collected in liquid buffer are suitable for both cytological screening and ancillary Human Papilloma virus testing. Cancer 1997; 81; 89-97. Communication, 20 May 1996. Rockville, Md.: Center for Devices and Radiological Health, Food and Drug Administration, 1996. (Received 3 Feb, accepted 14 July 1997) Authors' details Dr Colin Laverty & Associates, Pathologists, Eastwood, NSW. Jennifer M Roberts, MB BS, FRCPA, Pathologist; A Marion Gurley, MB ChB, FIAC, Pathologist; Julia K Thurloe, BSc, MEc, Statistician; Ronald Bowditch, BScAg CT(ASC), Senior Cytotechnologist; Colin R A Laverty, MB BS, FRCPA, Principal Pathologist. Reprints: Dr C R A Laverty, Dr Colin Laverty & Associates, Pathologists, 18 Glen Street, Eastwood, NSW 2122. ©MJA 1997 <URL: http://www.mja.com.au/> © 1997 Medical Journal of Australia.

Reduction of inappropriate use of blood products by prospective monitoring of transfusion request forms Annabel Tuckfield, Michael N Haeusler, Andrew P Grigg and Jack Metz For editorial comment, see Allain & Williamson Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au/>". Abstract - Introduction - Methods - Results - Discussion - References - Authors' details - - ©MJA1997 Abstract Objective: To determine the effect of prospective monitoring on appropriateness of transfusions of red cells, platelets and fresh frozen plasma (FFP). Design: Prospective interventional study. Setting: Royal Melbourne Hospital (a tertiary teaching hospital), Melbourne, Victoria, March-May 1996. Intervention: The blood product request form was modified to incorporate indications for transfusion and clinical and laboratory data. Requests were monitored by blood bank laboratory staff for conformation with hospital transfusion guidelines; non-conforming requests were discussed with the requesting medical practitioner by the Haematology Registrar before blood products were issued. In cases of disagreement, blood products were always issued. Subjects: 200 consecutive transfusion episodes for each product (red cells, platelets and FFP). Outcome measures: Appropriateness of transfusion, assessed by a Consultant Haematologist according to hospital guidelines. Rates of inappropriate transfusion episodes after intervention were compared with rates in a previous study. Results: After intervention, rates of inappropriate transfusion episodes fell significantly (red cells, 16% to 3% [P=0.004]; platelets, 13% to 2.5% [P=0.02]; and FFP, 31% to 15% [P=0.02]). Almost all inappropriate FFP transfusion episodes post-intervention were due to failure to demonstrate prolongation of prothrombin or activated partial thromboplastin times more than 1.5 times the control value. Conclusion: Prospective monitoring of request forms can reduce rates of inappropriate transfusions. High rates of inapproriate FFP transfusions possibly reflect uncertainty about appropriate laboratory criteria for FFP transfusion. While results of large prospective randomised controlled clinical trials of FFP transfusions are awaited, current laboratory criteria can be retained, but should be applied with flexibility. MJA 1997; 167: 473-476 Introduction In 1995, we reported results of an audit of blood product use at a tertiary teaching hospital (Royal Melbourne Hospital, Victoria).1 Indications for transfusion were analysed, mainly by retrospective review of medical records. We found that blood product use was inappropriate for 16% of red cell, 13% of platelet and 31% of fresh frozen plasma (FFP) transfusion episodes. In a significant number of episodes, the specific indication for transfusion was not documented in the medical record. Blood products must be used appropriately to minimise patient exposure to potential hazards, to conserve a limited resource and to contain costs. Measures to reduce inappropriate use of blood products have had varying results. While retrospective audit and education do not produce consistent and continuing improvements,2-8 prospective monitoring (monitoring of requests for blood products before issue) has usually been more successful.9-15 Consequently, in March 1996, the Royal Melbourne Hospital introduced prospective monitoring of transfusion requests with the aim of reducing inappropriate use of blood products. We present the results of the first three months of this intervention. Methods Intervention Blood products request form: The blood products request form was redesigned to incorporate the indication for transfusion and relevant clinical and laboratory data. These included: For red cells: haemoglobin level and clinical parameters (pulse rate and blood pressure); For platelets: platelet count and presence or absence of bleeding; For FFP: results of coagulation tests (prothrombin time [PT] or international normalised ratio [INR] and activated partial thromboplastin time [APTT]); and For all blood products: reasons for transfusion and operation (if applicable). In addition, the redesigned request form stipulates that the indication for transfusion must conform with the hospital's transfusion guidelines, which are printed on the reverse of the form (Box 1). Criteria for appropriate transfusions: The criteria were the same as those used in the audit.1 For FFP use, the requirement for coagulation times to be prolonged by more than 50% translated to PT >20 seconds or APTT >40 seconds. For massive blood transfusion (10 or more units of red cells in less than 24 hours), blood products were issued before laboratory results were available, and these transfusions were accepted as appropriate. In view of the potentially devastating effects of minor haemorrhage associated with neurosurgical procedures, these were exempted from the criteria for FFP use.1 Monitoring: On receiving a completed blood product request form, the senior medical laboratory scientist determined that the necessary information had been provided, and telephoned the requesting doctor to obtain any missing information. The scientist then checked that the indications conformed with transfusion guidelines, and if so, blood products were issued. If clinical or laboratory indications did not conform, the request was referred to the Haematology Registrar who consulted the requesting doctor. There were three possible outcomes: they agreed (i) either that the transfusion was not indicated and the blood products were not issued or (ii) that it was indicated and the products were issued or (iii) they disagreed and the products were issued. The consultation was never confrontational, and the blood product was always issued when there was disagreement. However, for the study these cases were subsequently referred to the Consultant Haematologist who reviewed the haematological data, transfusion request and medical record to decide if the transfusion was indeed inappropriate. Analysis of outcomes For each type of blood product (red cells, platelets and FFP), we analysed 200 consecutive transfusion episodes between March and May 1996. As in the initial audit, we excluded FFP transfusions for patients with thrombotic thrombocytopenic purpura, because of the different rationales for FFP use in this condition and in coagulation factor depletion. Rates of inappropriate blood product use after introduction of prospective monitoring were compared with those found during the previous audit (pre-intervention) with two-sided Fisher's exact tests. Results Indications for transfusion of red cells, platelets and FFP after introduction of prospective monitoring are shown in Box 2 (below). They were the same as in the pre-intervention audit.1 Numbers of transfusions referred to the Consultant Haematologist for review were: red cells, eight; platelets, 12; and FFP, 41. Final numbers of transfusion episodes deemed inappropriate were: red cells, six (3%); platelets, five (2.5%); and FFP, 30 (15%). Compared with the pre-intervention audit, rates of inappropriate transfusion episodes were significantly reduced, from 16% to 3% for red cells (P= 0.004), 13% to 2.5% for platelets (P=0.02) and 31% to 15% for FFP (P= 0.02). Of the six patients in whom red cell use was deemed inappropriate, five had a pretransfusion haemoglobin concentration greater than 100 g/L, the threshold for appropriate transfusion (range in these patients, 104-114 g/L). The sixth patient had a haemoglobin concentration of 98 g/L, but no clinical indications to warrant transfusion. In four of the five inappropriate platelet transfusion episodes, the platelet count was greater than 63x109/L (range, 63x109-159x109/L), a level at which platelets are not indicated except in the case of excessive bleeding with cardiac bypass surgery. Although three of these four transfusions were given in association with cardiac bypass surgery, there was no evidence of excessive bleeding. In 30 patients, FFP transfusion was deemed inappropriate as coagulation test results did not meet the criterion of a 50% prolongation of coagulation times. In one of these 30, this was because no coagulation tests had been performed. Discussion We found that the rate of inappropriate transfusion episodes fell significantly after introduction of prospective monitoring, from 16% to 3% for red cells, from 13% to 2.5% for platelets and from 31% to 15% for FFP. The success of prospective monitoring in reducing rates of inappropriate transfusions for red cells and platelets was not matched for FFP. This agrees with results of three previous studies, which have found persistently high rates of FFP transfusions despite monitoring (>40%,8 33%13 and 27%,15 respectively). As in these studies, we found the reason for deeming FFP use inappropriate was almost invariably failure to show sufficient abnormality in coagulation test results. The need to document abnormal coagulation with a PT or APTT greater than 1.5 x the control value, before FFP transfusion is deemed appropriate, rests on the assumptions that abnormal bleeding is unlikely if coagulation times are shorter than this but likely if they are longer, and that, when coagulation times are greater than 1.5 x the control value, FFP transfusion is likely to prevent or control bleeding. Detailed analysis of published data suggests that these assumptions are questionable.16-23 It is of interest that we found that lowering the cut-off for appropriate FFP transfusions from PT >20 seconds (1.5 times the control value) to PT >17 seconds would have lowered the rate of inappropriate transfusions from 15% to 3%, comparable to the rate found for red cell and platelet transfusions. A PT of 17 seconds is still 3.5 seconds longer than the median control value. Further investigation is needed to establish the clinical guidelines for FFP transfusion. Two publications24,25 have pointed to the need for a large multicentre prospective randomised controlled clinical trial to provide a definitive answer for the role of FFP in patients with acquired multiple coagulation defects. In the interim, it would seem reasonable to require docu mentation of abnormal coagulation as a criterion for appropriate FFP use, but the available evidence does not warrant rigid application of a cut-off level, such as 50% prolongation of PT or APTT. Other issues that must be addressed before a system of prospective monitoring can be introduced include exemption criteria. Some protocols exempt operating room patients,6 "desperate situations"14 and haematology and oncology patients.13 Emergency situations are also usually exempted from the requirement for laboratory data if the laboratory cannot provide urgent results at all times. If the criteria are to be applied to emergency situations, laboratory facilit ies must be available to provide the data quickly. In some programs of prospective monitoring, blood products are never withheld, but apparently inappropriate transfusions are later reviewed.11,14 Refusal to issue blood products leads to an adversarial relationship between clinicians and laboratory staff, which may compromise patient care. Refusal might also have medicolegal implications if subsequent patient morbidity or death could be attributed to withholding of the transfusion. For sustained improvement in practice, prospective monitoring must be continued indefinitely.11 This is both time consuming and demanding of staff. The demand might be lessened by computerised audit of transfusion requests; clinical and laboratory data could be entered into a program which flags non-compliant requests for review by blood bank staff.26 In conclusion, prospective monitoring was successful in reducing inappropriate use of red cells and platelets, but only partly successful for FFP. This probably reflects uncertainty about the appropriate clinical guidelines for FFP use, and although current laboratory criteria for FFP use should probably be retained, they should be applied with flexibility. References Metz J, McGrath KM, Copperchini ML, et al. Appropriateness of transfusions of red celIs, platelets and fresh frozen plasma. An audit in a tertiary care teaching hospital. Med J Aust 1995; 162: 572-577. Lam HTC, Schweitzer SO, Petz L, et al. Are retrospective peer-review transfusion monitoring systems effective in reducing red blood cell utilization? Arch Pathol Lab Med 1996; 120: 810-816. Toy PTCY. Audit and education in transfusion medicine. Vox Sang 1996; 70: 1-5. Bamerte RE, Fish DJ, Eisenstaedt RS. Modification of fresh frozen plasma transfusion practices through educational intervention. Transfusion 1990; 30: 253-257. Soumerai SB, Salem-Schatz S, Avorn J, et al. A controlled trial of educational outreach to improve blood transfusion practice. JAMA 1993; 270: 961-966. Solomon RR, Clifford JS, Gutman SI. The use of laboratory intervention to stem the flow of fresh frozen plasma. Am J Clin Pathol 1988; 89: 518-521. Brien WF, Buttier RJ, Inwood MJ. An audit of blood component therapy in a Canadian general teaching hospital. Can Med Assoc J 1989; 140: 812-815. Shanberge JN, Quattrochiocchi-Longe T. Analysis of fresh frozen plasma administration with suggestions for ways to reduce usage. Transfus Med 1992; 2: 189-194. Simpson MB. Prospective concurrent audits and medical consultation for platelet transfusions. Transfusion 1987; 27: 192-195. McCullough J, Steeper TA, Connelly DP, et al. Platelet utilization in a university hospital. JAMA 1988; 259: 2414-2418. Silver H, Tahhan HR, Anderson J, et al. A non- computer dependent prospective review of blood and blood component utilization. Transfusion 1992; 32: 260-265. Brandis K, Richards B, Ghent A, et al. A strategy to reduce inappropriate red blood cell transfusion. Med J Aust 1994; 160: 721-722. Hawkins TE, Carter JM, Hunter PM. Can mandatory pretransfusion approval programmes be improved? Transfus Med 1994; 4: 45-50 Cheng G, Wong HF, Chan A, et al. The effects of a self-educating blood component request form and enforcements of transfusion guidelines on FFP and platelet usage. Clin Lab Haem 1996; 18: 83-87. Marconi M, Almini D, Pizzi MN, et al. Quality assurance of clinical transfusion practice by implementation of the privilege of blood prescription and computerized prospective audit of blood requests. Transfus Med 1996; 6: 11-19. Counts RB, Haisch C, Simon TL, et al. Hemostasis in massively transfused trauma patients. Ann Surgery 1979; 190: 91-99. Braunstein AH, Oberman HA. Transfusion of plasma components. Transfusion 1984; 24: 281-286. Ciavarella D, Reed RL, Counts RB, et al. Clotting factor levels and the risk of diffuse microvascular bleeding in the massively transfused patient. Br J Haematol 1987; 67: 365-368. Harvey MP, Greenfield TP, Sugrue ME, et al. Massive blood transfusion in a tertiary referral hospital. Clinical outcomes and haemostatic complications. Med J Aust 1995; 163: 356-359. Houry S, Georgeac C, Hay J, et al. A prospective multicentre evaluation of preoperative hemostatic screening tests. Am J Surg 1995; 170: 19-23. Gelb AB, Roth RI, Levid J, et al. Changes in blood coagulation during and following cardiopulmonary bypass. Lack of correlation with clinical bleeding. Am J Clin Pathol 1996; 106: 87-99. Murray DJ, Olson JD, Strauss AR, et al. Coagulation changes during packed red cell replacement of major blood loss. Anesthesiology 1988; 69: 839-845. Murray DJ, Pennell BJ, Weinstein SL, et al. Packed red cells in acute blood loss: dilutional coagulopathy as a cause of surgical bleeding. Anesth Analg 1995; 80: 336-342. McVay PA, Toy PTCY. Lack of increased bleeding after liver biopsy in patients with mild hemostatic abnormalities. Am J Clin Pathol 1990; 94: 747-753. Cohen H. Avoiding the misuse of fresh frozen plasma. BMJ 1992; 307: 395-396. Gardner RM, Golubjatnikov OK, Laub RM, et al. Computer critiqued blood ordering using the HELP system. Comput Biomed Res 1990; 23: 514-528. (Received 20 Mar, accepted 1 Jul 1997) Authors' details Department of Diagnostic Haematology, Royal Melbourne Hospital, Melbourne, VIC. Annabel Tuckfield, FRACP, FRCPA, Registrar; Michael N Haeusler, FAIMS, Senior Scientist, Blood Bank; Andrew P Grigg, FRACP, FRCPA, Haematologist; Jack Metz, MD, FRCPA, Head. Reprints will not be available from the authors. Correspondence: Dr A Tuckfield, Diagnostic Haematology, Post Office, Royal Melbourne Hospital, Parkville, VIC 3050. E-mail: depATmis.medrmh.unimelb.edu.au ©MJA 1997 <URL: http://www.mja.com.au/> © 1997 Medical Journal of Australia.

Reported general practitioner vaccination procedures, 1994 and 1996 Ana Herceg, Monica B Johns and Helen M Longbottom For editorial comment, see Kable & Harris Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au/>". Abstract - Introduction - Methods - Results - Discussion - References - Authors' details - - ©MJA1997 Abstract Objective: To determine general practitioner (GP) vaccination procedures during the first two years of implementation of the National Childhood Immunisation Program. Design: Two cross-sectional self-completion mail surveys of Australian GPs, in October/November 1994 (Survey 1) and April/May 1996 (Survey 2). Participants: 1417 systematically selected (Survey 1) and 1482 randomly selected (Survey 2) GPs. Intervention: Mailing of a package of publications on immunisation to all GPs, a community education campaign, and the introduction of the Australian Childhood Immunisation Register. State and Territory and local initiatives on immunisation, changes in vaccine distribution and funding, and media coverage of vaccination issues may also have affected GPs' procedures. Main outcome measures: Reported vaccination procedures consistent with National Health and Medical Research Council (NHMRC) recommendations. Results: Response rates were 77% (Survey 1) and 78% (Survey 2). There were statistically significant rises in the proportions of GPs who reported vaccination procedures consistent with NHMRC recommendations for simultaneous administration of vaccines, injection site, vaccination of children taking antibiotics or those with mild upper respiratory tract infection, and vaccine storage. There was a reduction in reported opportunistic vaccination, and no change in reported procedures regarding pertussis vaccination. Conclusions: Many factors may have contributed to these changes in GP vaccination procedures. While the changes mostly show an improvement in adherence to NHMRC guidelines, further progress needs to be made, particularly in the areas of opportunistic vaccination, pertussis vaccination and vaccine storage. The best method of informing GPs of current vaccination recommendations needs to be explored. MJA 1997; 167: 299-302 Introduction Vaccine-preventable diseases of childhood continue to occur in Australia, with large outbreaks of measles, pertussis and rubella occurring in 1993 and 1994.1 Vaccination coverage rates in Australia are still not high enough to prevent these outbreaks, with reported coverage varying from 60% for pertussis to 92% for measles.2 More than 60% of childhood vaccinations in Australia are given by general practitioners (GPs),2 and vaccination guidelines are published by the National Health and Medical Research Council (NHMRC) in The Australian immunisation handbook, which is updated regularly.3-5 Studies in the United States have shown that health care systems and immunisation provider services contain major barriers to childhood vaccination, including inaccessibility, insufficient staff, insufficient clinic hours, missed opportunities to vaccinate (i.e., not vaccinating children who present for another reason and are due or overdue for vaccination), and misconceptions about true and false contraindications to vaccination.6 The existence or otherwise of these barriers to childhood vaccination in Australian health care has not been well researched, but one study showed that many immunisation providers in Victoria were unaware of the NHMRC recommendations for diphtheria-tetanus-pertussis (DTP) vaccine and combined diphtheria-tetanus (CDT) vaccine.7 Other studies found that GPs frequently do not follow the NHMRC recommendations for site of injection of vaccines and vaccine storage.8,9 The National Childhood Immunisation Program was devised to address the problem of low vaccination cover - age through a coordinated national approach. The Program, which commenced in July 1994, involved Commonwealth, State and Territory health authorities, as well as professional and consumer bodies. Aspects of the Program included improving surveillance and reporting for vaccine-preventable diseases and vaccination coverage; improving vaccine purchasing, distribution and delivery; and educating parents, the general public and vaccination providers about vaccination. We conducted two national surveys of GPs to estimate changes in vaccination procedures during the first two years of the National Childhood Immunisation Program. Methods Sample selection and survey method Providers were systematically (Survey 1) and randomly (Survey 2) selected from GPs registered with the Health Insurance Commission who had lodged more than 50 services in the preceding quarter. A 10% sample was selected from each group, comprising 1843 GPs in Survey 1 and 1888 GPs in Survey 2. The sample for Survey 1 was created by selecting every tenth GP on the list; a modified classic synchronised sampling method was used for random selection for Survey 2.10 Both surveys were anonymous self-completion mail surveys. For each survey, an initial questionnaire and covering letter was sent to each GP, followed a week later by a reminder letter and a duplicate questionnaire. Practitioners were asked to ignore the reminder if they had already completed the questionnaire, and to return only one blank questionnaire if they did not wish to participate in the survey. The two surveys contained identical questions about GPs and their practices, vaccine storage methods and vaccination procedures. Questions on vaccination procedures and vaccine refrigeration were based on the NHMRC recommendations in the fifth edition of The Australian immunisation procedures handbook (1994). No identifying information was collected. Interventions Survey 1 was conducted in October/November 1994, one month before the National Childhood Immunisation Program sent a package of publications to all vaccination providers, including all GPs. The package included The Australian immunisation procedures handbook (fifth edition) and other publications on vaccine provision and storage. Survey 2 was conducted in April/May 1996, after the National Childhood Immunisation Program had been in place for almost two years. Activities of the Program during that time included distribution of the publications mentioned above (November 1994), a community education campaign (April to November 1995), and the introduction of the Australian Childhood Immunisation Register (January 1996). During the period between the two surveys, other activities which may have affected GP immunisation practices included State and Territory initiatives to inform immunisation providers about vaccination procedures, changes in vaccine distribution and funding systems, the development of school-entry vaccination record requirements in some jurisdictions, media coverage, and local promotion of vaccination issues through Divisions of General Practice and other agencies. Statistical analysis Data entry and analysis were done with Epi Info version 6. 11 The chi-squared test for independence in 2 x 2 tables was used to compare the results of the two surveys. Analyses were based on the number of respondents who completed each question rather than the total number of respondents. Results Response rates After exclusion of respondents who indicated they did not provide childhood vaccinations, response rates were 77% for Survey 1 and 78% for Survey 2. Ninety-eight per cent of respondents worked in general practice. The two samples did not differ significantly in type of practice or age groups, but, for number of vaccines given per week, significantly more respondents in Survey 2 gave between one and nine (Box 1, below). Vaccine administration procedures Reported vaccine administration procedures consistent with the NHMRC recommendations are compared in Box 2. The proportion of practitioners who would always opportunistically vaccinate fell from 32% to 28%, although an additional 47% of GPs in 1996 said they would frequently vaccinate children opportunistically. Simultaneous administration of DTP, oral polio vaccine and Haemophilus influenzae type b (Hib) vaccine increased from 49% to 68%. Recording of vaccine batch number and use of recommended injection sites also increased. There were statistically significant rises from 1994 to 1996 in reported willingness of GPs to vaccinate children with mild upper respiratory tract infections or who were taking antibiotics (Box 3), and there was a fall from 94% to 90% in the proportion of GPs who would not vaccinate a febrile child. There was no change between the two surveys regarding when CDT should be used to replace DTP, or vaccination for a child taking inhaled or topical steroids. Reported storage of vaccines according to NHMRC recommendations improved between the two surveys, including use of a maximum/minimum thermometer, storing vaccines only on the top and middle shelves of the refrigerator and not storing food or specimens in the same refrigerator (Box 4). Although reported daily checking of the vaccine refrigerator temperature increased, in 1996 only 27% of GPs reported that this was done. A further 30% reported that the refrigerator temperature was checked weekly. In November 1994, 15% of respondents reported their practices had immunisation recall/reminder systems functioning. In April 1996, 78% of GPs reported always contributing to the Australian Childhood Immunisation Register. Discussion Our study has a number of limitations, including the sampling methods, the number of non-respondents and the fact that it was based on self-reported behaviour. While the sampling methods for the two surveys were not identical, the two groups were demographically comparable, and we believe both samples are representative of the GP population. It is estimated that around 170 GPs would have received both surveys; however, as the surveys were conducted anonymously we have no record of this overlap. Thus, it is possible that a small proportion of the reported change in behaviour was because of this previously surveyed group who would have been more aware of the issues. Finally, self-reported behaviour does not necessarily correlate with actual behaviour. Hence, these surveys probably provide a conservative estimate of behaviour contrary to NHMRC guidelines. Our study showed a significant improvement in several reported GP vaccination procedures between 1994 and 1996. The reasons for these changes are probably multiple, and include the information provided through the National Childhood Immunisation Program, State and Territory immunisation provider information and other professional education sources, as well as community and media interest in the topic. The degree to which each of these sources may have influenced the change cannot be assessed from our study. However, the coordination and information dissemination role of the National Childhood Immunisation Program meant that the information conveyed from health authorities and professional organisations to GPs between the two surveys was consistently based on NHMRC recommendations. Our results indicate, however, that more progress needs to be made in some areas. Opportunistic vaccination, including in hospital emergency departments, has frequently been identified as a way of improving immunisation coverage.12 In 1996, many GPs potentially missed opportunities to vaccinate by infrequently or never checking children's immunisation status when they presented for another reason. The high use of the parent-held Personal Health Record identified in our survey was also found by Bailey et al.13 Recording of batch number on the Record and/or in practice records improved between the two surveys; this is recommended by the NHMRC to ensure that children potentially at risk of illness can be found if a faulty batch of vaccine is identified. Misconceptions about true and false contraindications to vaccination have previously been recognised as barriers to full vaccination; this especially applies to when it is appropriate to replace DTP with CDT.6,7 The NHMRC guidelines have previously not been clear about pertussis vaccination, and this is likely to have contributed to the confusion. Most GPs in our study would replace DTP with CDT if a child had a febrile convulsion or persistent screaming after the last dose of DTP. These are situations which were previously classed by the NHMRC as contraindications to DTP and which, in the 1994 recommendations, "may preclude" another dose of DTP. The 1997 sixth edition of The Australian immunisation handbook provides new, clear guidelines on pertussis vaccination.5 GPs need to be aware of these changes; following the new recommendations should increase pertussis vaccination rates and prevent more disease. Vaccine storage procedures have been previously shown in New South Wales to be inadequate in many immunisation provider practices.9,14 Although there were improvements between our two surveys, in 1996 only 27% of practices checked and recorded the refrigerator temperature daily. Correct vaccine storage is essential to maintain vaccine potency. Our study looked only at private vaccination providers. We currently have no information about immunisation procedures of public providers, who administer around 40% of vaccines, and further studies are needed to assess procedures in this sector. While many GPs have kept abreast of changing recommendations, further improvements are possible. In addition, because of the impending availability of new vaccines and combinations of vaccines, further changes to the vaccination schedule and recommendations are likely in the coming years. Our study has not identified the best way to inform GPs of changes in recommendations, and this must be determined. The 78% participation rate of GPs in the Australian Childhood Immunisation Register five months after its commencement shows that good communication channels with GPs are available. Further improvements in vaccination provider practices will result in more children receiving the correct, efficacious vaccines safely and on time. References Herceg A, Oliver G, Myint H, Andrews G, et al. Annual report of the National Notifiable Diseases Surveillance System, 1995. Commun Dis Intell 1996; 21: 440-464. Australian Bureau of Statistics. Children's immunisation, Australia. Canberra: ABS, April 1995. (Catalogue No. 4352.0.) National Health and Medical Research Council. Immunisation procedures. 4th ed. Canberra: AGPS, 1991. National Health and Medical Research Council. The Australian immunisation procedures handbook. 5th ed. Canberra: AGPS, 1994. National Health and Medical Research Council. The Australian immunisation handbook. 6th ed. Canberra: AGPS, 1997. Cutts FT, Orenstein WA, Bernier RH. Causes of low preschool immunisation coverage in the United States. Annu Rev Publ Health 1992; 13: 385-398. McIntyre CR, Nolan T. Attitudes of Victorian vaccine providers to pertussis vaccine. Med J Aust 1994; 161: 295-299. Rixon G, March L, Holt DA. Immunisation practices of general practitioners in metropolitan Sydney. Aust J Public Health 1994; 18: 258-260. Liddle JLM, Harris MF. How general practitioners store vaccines. A survey in south-western Sydney. Med J Aust 1995; 162: 366-368. Calcino G. Sampling from the HIC data set. Proceedings of the General Practice Evaluation Program 1993 work-in-progress conference, Canberra; 20-22 May 1993. Canberra: Department of Health, Housing, Local Government and Community Services, 1993. Dean AG, Dean JA, Couombier D, Brendel KA, et al. Epi Info, Version 6: a word processing, database, and statistics program for epidemiology on microcomputers. Atlanta, Ga: Centers for Disease Control and Prevention, 1994. Jones K, Fasher B, Hanson R, Burgess M, et al. Immunization status of casualty attenders: risk factors for non-compliance and attitudes to "on the spot" immunization. J Paediatr Child Health 1992; 28: 451-454. Bailey FK, Nossar V, Jeffs DA, et al. An evaluation of the NSW Personal Health Record. Sydney: NSW Health, 1992. Miles T. The integrity of the vaccine cold chain in the Hunter area of New South Wales. Aust J Public Health 1993; 17: 169-171. (Received 19 Sep 1996; accepted 21 Jul 1997) Authors' details National Centre for Disease Control, Department of Health and Family Services, ACT. Ana Herceg, MPH, MAppEpid, Medical Epidemiologist, Surveillance and Management Section. Monica B Johns, BAppSc; formerly, Senior Project Officer, National Childhood Immunisation Program. Helen M Longbottom, MPH, FAFPHM; formerly, Director, Surveillance and Epidemiology Section. Reprints will not be available from the authors. Correspondence: Ana Herceg, Surveillance and Management Section, National Centre for Disease Control, MDP 6, Department of Health and Family Services, GPO Box 9848, ACT 2601. ©MJA 1997 <URL: http://www.mja.com.au/> © 1997 Medical Journal of Australia.

Ana Herceg · Monica B Johns · Helen M Longbottom

Helicobacter pylori infection in an Australian regional city: prevalence and risk factors Hedley G Peach, Dora C Pearce and Stephen J Farish Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au/>". Abstract - Introduction - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - ©MJA1997 Abstract Objective: To investigate the prevalence of Helicobacter pylori infection and potential risk factors for infection in an adult Australian population. Design: Cross-sectional study. Setting: Ballarat, a major regional city in Victoria (population, 78 000; 92% born in Australia), November 1994 to July 1995. Participants: 217 adults randomly selected from the electoral roll. Main outcome measures: H. pylori IgG antibody status by enzyme immunoassay; amount of dental plaque; sociodemographic and other potential risk factors; odds ratios for risk factors determined by logistic regression analysis. Results: Age-standardised prevalence of H. pylori infection was 30.6%. After adjustment for age, sex and socioeconomic index, positive H. pylori status was significantly associated with increasing number of tooth surfaces with a high plaque score (odds ratio [OR], 1.7; 95% confidence interval [CI], 1.1-2.7), increasing number of years in a job with public contact (OR, 1.7; 95% CI, 1.3-2.3), blood group B antigen (OR, 3.1; 95% CI, 1.1-9.1), and having lived in a household with more than six members during childhood (OR, 2.5; 95% CI, 1.1-5.5). Negative H. pylori status was significantly associated with increasing education, having ever lived on a farm, and having teeth scaled less than once a year. Conclusions: H. pylori infection is common. Dental plaque may be a reservoir for H. pylori, which is probably transmitted by person-to-person contact, and blood group B antigen may predispose to infection. Community education about effective oral hygiene and adoption of good hygiene practices by those with regular public contact may be important to prevent acquisition and transmission of H. pylori . MJA 1997; 167: 310-313 Introduction The presence of Helicobacter pylori in humans is now recognised as a chronic infection which in most cases persists indefinitely and is involved in the development of several serious diseases, including peptic ulcer, gastric cancer, and possibly cardiovascular disease.1 It is uncertain from where and how H. pylori is acquired, but most investigators seem to favour person-to-person transmission.1 In developed countries, including Australia, the prevalence of H. pylori infection increases with age at the rate of about 1% per annum, to reach 40% or more at age 60 years.2 In contrast, in developing countries prevalence is 40% or more at age 10 years.2 Seroconversion in childhood also seems more likely in particular populations in developed countries; it was found to occur among people of lower socioeconomic background in Northern Ireland3 and England 4 and in a rural town in Western Australia.5 Suspected childhood risk factors for H. pylori infection in developed countries include having a single parent, lower parental social class, greater housing density or proportion of rented housing in school catchment area, overcrowding, bed-sharing, and absence of a fixed hot water supply.6 In adults, previously investigated risk factors include race-ethnicity, socioeconomic status, education, dental prosthesis and irregular dental treatment, public contact, working with animals, smoking status, alcohol consumption, homosexuality, sexually transmitted diseases, and having an infected child or partner.1,2,7-10 To assess the importance of these and other risk factors for H. pylori infection in Australia, we investigated the H. pylori status and risk factors of a sample of adults in Ballarat, a regional city in Victoria. Methods Setting and population Ballarat is a major regional city in Victoria (population, 78 000; with 92% of residents born in Australia according to the 1991 census).11 The study population was drawn from the population sample used in a 1992 survey of cardiovascular disease risk factors. This survey used the procedures of the National Heart Foundation for its surveys of Australian capital cities.12 A sample of 501 adults randomly selected from the electoral roll was invited to attend a screening clinic; 338 (68%) attended and gave informed consent. Between November 1994 and July 1995, 217 of the 262 participants who were still contactable (83%) participated in our study of risk factors for H. pylori infection, after again giving informed consent. The study was approved by the Ballarat Base Hospital Ethics Committee. Investigations Plasma samples were tested for H. pylori IgG antibodies with the Pyloriset EIA-G Kit (Pyloriset, Orion Corporation, Orion Diagnostica, Espoo, Finland). This kit uses one of the two original Marshall and Warren strains of H. pylori13 for antigen preparation. A cut-off antibody titre of 500 was used to classify subjects as positive or negative, as recommended by the manufacturer (sensitivity, 92.5%; specificity, 84.3%).14 Dental plaque was also assessed for the 156 dentate subjects. The mouth was rinsed with water before and after fluorescein application, and plaque score was recorded on a modified Dental Health Services of Victoria treatment chart.15 Research personnel were blind to the H. pylori status of each subject. A high plaque score (≥2) was defined as a clear line of plaque along the gingival margin or a larger coverage of the tooth surface with plaque.15 After examination, the subject was given toothpaste and a toothbrush to remove residual fluorescein. A questionnaire on socio demographic and potential risk factors was administered by one of three trained research personnel. Job title and duties were used to classify jobs held over a subject's lifetime, without knowledge of his or her H. pylori status, as mainly involving contact with the general public or otherwise. Subjects identified the 1991 Census Collector's District in which they lived from a map of Ballarat; this was classified according to socioeconomic index.16 Blood group was determined if unknown. Statistical analyses The sample-based estimate of the prevalence of H. pylori infection was standardised to the age distribution of Ballarat's population.12 This accounted for any imbalances between respondents and the general population caused by under-representation of younger or more mobile residents in the electoral rolls or differential response rate by age. Odds ratios (ORs) and 95% confidence intervals (CIs) for suspected risk factors for H. pylori infection were produced by logistic regression analyses. Collinearity was first assessed through a series of multiple linear regression analyses, using each predictor variable in turn as the dependent variable and all others as independent variables. Dependent variables with an R2 greater than 0.95 were excluded. All remaining variables were entered into the logistic regression model, and then excluded manually in a stepwise manner if they were not significantly associated with H. pylori status and had an OR in the range 0.95-1.05. Edentate subjects were assigned the mean plaque score of subjects with teeth, as recommended by Thompson.17 This avoided having to analyse edentate and dentate subjects separately, with the inevitable loss in precision when estimating risk associated with other variables. The effect of response to the study on ORs was assessed by the method of Criqui.18 The associations found using H. pylori status as a dichotomous variable, based on a cut-off antibody titre of 500, were confirmed using H. pylori antibody titres (logarithmically transformed) as a continuous variable in a multiple linear regression model. Results The prevalence of H. pylori infection in Ballarat was estimated as 30.6% after standardising to the age distribution of the city's population. The proportion of subjects who were H. pylori-negative in the H. pylori study (66.8%) was similar to that in the original cardiovascular study (69.6%). H. pylori-positive subjects in the H. pylori study had similar age, sex distribution, education and median antibody titres to those in the original cardiovascular study. H. pylori-negative subjects in the two populations were also similar for these variables. Therefore, the ORs for suspected risk factors were unlikely to be affected by the response rate.18 The association between positive H. pylori antibody status and possible risk factors in the final logistic regression model is shown in the Box. In the final model, which adjusted for age, sex and socioeconomic index of residence, the following were significantly associated with positive H. pylori status: increasing number of buccal and lingual tooth surfaces with a plaque score ≥2 (adjusting for the absence of natural teeth and number of teeth), increasing years in a job with public contact, presence of blood group B antigen, and having lived in a household of six or more people during the first 16 years of life. Washing of hands only sometimes or rarely after using the toilet and visiting a dentist less than once a year had a large but not statistically significant association with positive H. pylori status. Negative H. pylori status was significantly associated with increasing education, having ever lived on a farm, and having teeth scaled less than once a year. Univariate analysis showed that dentate subjects who visited a dentist at least annually had a significantly lower prevalence of H. pylori infection if they had their teeth scaled less than once a year (three out of 25) compared with more often (20 out of 56) (chi-squared = 4.78; P = 0.03; 1 degree of freedom). In addition, among those who had their teeth scaled less than once a year, visiting a dentist at least annually was protective (three infected out of 25) compared with visiting a dentist less often (27 infected out of 78) (chi-squared = 4.69; P = 0.03; 1 degree of freedom). The following variables were initially included in the logistic regression model, but excluded from the final model because of lack of a significant association with H. pylori status: dentures, history of gum infections, frequency of sharing toothbrush or cup, travel to high prevalence countries, smoking status, lifetime alcohol consumption, vegetarian diet, eating unwashed home-grown vegetables, washing hands before eating, years with occupational or recreational contact with animals, having visited farms, marital status, height, weight, and variation between interviewer/oral health observers. The results of multiple regression analysis confirmed those of the logistic regression analysis for all variables except increasing education, which was negatively, but not significantly, associated with H. pylori titre. Discussion To our knowledge, this is the first time in Australia that the prevalence of H. pylori infection has been determined in a random sample of the adult population. Our crude age-specific prevalences were among the lowest of such figures published for European countries and white Americans.19 Nevertheless, the considerable proportions of gastroduodenal ulcers, peptic ulcer bleeds and gastric cancers attributable to H. pylori make an infection rate of 30.6% an important public health issue.20 The cost-benefit ratio for eradicating H. pylori infection with antibiotics in asymptomatic adults will not be known for some time.20 It is therefore important to identify factors associated with the acquisition and transmission of H. pylori. This is the first time that Public contact through employment; Moderate or heavy accumulation of dental plaque; Frequent scaling of teeth; Only rarely or sometimes washing hands after using the toilet; and Having blood group B antigen have been found to be positively associated with H. pylori infection. Further, it is the first time that having lived on a farm, as a third of our sample had done, has been identified to be protective against H. pylori infection. We confirmed that living in an overcrowded household during childhood and visiting a dentist less than once a year are positively associated with H. pylori infection in Australia, although the latter association was not significant. The association between H. pylori infection and public contact through employment was not surprising, as overcrowding in childhood has led to speculation that close personal contact may promote transmission of the organism.7 The protective effect of having lived on a farm in our study may be a consequence of low population density in rural areas. Rural China has a significantly lower prevalence of H. pylori infection, despite poorer standards of hygiene, compared with urban China.2,21 There is other evidence to corroborate our finding that moderate or heavy accumulation of dental plaque is significantly associated with H. pylori infection. H. pylori has been cultured from dental plaque, and identical H. pylori ribotypes were found in the mouth and gastric antrum of ulcer patients.22,23 Moreover, H. pylori has been detected by reverse transcription polymerase chain reaction in only moderate to heavy accumulations of plaque.22 Our finding that visiting the dentist less than once a year increased the risk of H. pylori infection is consistent with the protective effect of regular dental treatment found by Gasbarrini et al.10 If H. pylori inhabits plaque, scaling may lead to ingestion of the organism and inoculation of the stomach. A decreased risk of H. pylori infection from having teeth scaled less than once a year is therefore plausible. In agreement with other studies, we found no significant association between H. pylori infection and alcohol consumption. We also found no significant association with smoking, although 52% of our sample were current or ex-smokers. Others have found significantly more smokers among patients with no abnormalities on endoscopy who had H. pylori infection,8 and the difference could be caused by confounding by oral hygiene. Smokers have poor oral hygiene and visit their dentist and brush their teeth less often than non-smokers.24 Further, in our study, dentate subjects who had ever smoked had a significantly larger mean number of tooth surfaces with a high plaque score than non-smokers, after adjusting for number of teeth (P = 0.02). Although the risk of H. pylori infection was increased for people who washed their hands only sometimes or never after using the toilet, because of the small number of these subjects the odds ratio had a wide confidence interval and the association was not significant. Faecal-oral spread of the organism is known to occur, but our data imply that it is relatively uncommon between adults, as others have suggested.1 Nevertheless, this result underlines the importance of good hygiene practices and the provision of handwashing facilities both indoors and out-of-doors. Eating from communal bowls has been postulated as a mechanism of H. pylori transmission in China.21 Our study found that sharing of cups and cooking utensils was not common in Ballarat and not a risk factor. Genetic effects have been found to influence the acquisition of H. pylori.1 However, the importance of the association we found between the presence of blood group B antigen and H. pylori infection must await further research into the mechanisms of genetic susceptibility to H. pylori infection. Our results imply that good hygiene practices are essential for those in frequent contact with the public to prevent H. pylori transmission. As moderate to heavy dental plaque emerged as a clear risk factor, reducing plaque may have an important role in preventing acquisition of H. pylori. This is another reason for educating the public, adults and children about effective tooth-brushing and other oral hygiene techniques. Regular dental treatment may need to be encouraged, while avoiding the ingestion of plaque debris which could occur during tooth scaling. In view of these potentially important public health implications, confirmatory evidence of our results should be sought. Acknowledgements We thank Mrs Wendy Ross and Mrs Cheryl Potter for assistance with the interviews and oral examinations, Dr John Williams for advice on the examinations, Mr Christopher Pearce (Dorevitch Pathology, Ballarat Base Hospital) for technical assistance, State Data Centre, Ballarat, for mapping the Census Collector's Districts, Colgate for toothbrushes and toothpaste, and Chiron Diagnostics for serological advice. The study was funded by the Shepherd Foundation, Melbourne, VIC. References Hunt RH. Helicobacter pylori : from theory to practice. Am J Med 1996; 100: 1S-64S. Taylor DN, Blaser MJ. The epidemiology of Helicobacter pylori infection. Epidemiol Rev 1991; 13: 42-59. McCallion WA, Ardill JES, Bamford KB, et al. Age dependent hypergastrinaemia in children with Helicobacter pylori gastritis -- evidence of early acquisition of infection. Gut 1995; 37: 35-38. Banatvala N, Mayo K, Megraud F, et al. The cohort effect and Helicobacter pylori . J Infect Dis 1993; 168: 219-221. Cullen DJE, Collins BJ, Christiansen KJ, et al. When is Helicobacter pylori infection acquired? Gut 1993; 34: 1681-1682. Patel P, Mendall MA, Khulusi S, et al. Helicobacter pylori infection in childhood: risk factors and effect on growth. BMJ 1994; 309: 1119-1123. Whitaker CJ, Dubiel AJ, Galpin OP. Social and geographical risk factors in Helicobacter pylori infection. Epidemiol Infect 1993; 111: 63-70. Bateson MC. Cigarette smoking and Helicobacter pylori infection. Postgrad Med J 1993; 69: 41-44. Aceti A, Attanasio R, Pennica A, et al. Campylobacter pylori infection in homosexuals. Lancet 1987; 2: 154-155. Gasbarrini G, Pretolani S, Bonvicini F, et al. A population based study of Helicobacter pylori infection in a European country: the San Marino Study. Relations with gastrointestinal diseases. Gut 1995; 36: 838-844. HealthWIZ [computer file]: national social health database. Canberra: Department of Health and Family Services. National Heart Foundation. Risk factor prevalence survey. Canberra: National Heart Foundation, 1989: 133-136. Marshall BJ, Warren JR. Unidentified curved bacilli in the stomach of patients with gastritis and peptic ulceration. Lancet 1984; 1: 1311-1314. Pearce DC, Peach HG, Farish SJ. Helicobacter pylori antibody titres in serum, plasma and successively thawed specimens: implications for epidemiological and clinical studies. J Clin Pathol 1996; 49: 1017-1019. Muhlemann HR. Introduction to oral preventive medicine. Carol Stream, Ill: Quintessance Publishing Company, 1976: 124. Ross R, Farish S, Plunkett M. Indicators of socio-economic disadvantage for Australian schools. Geelong: Deakin Institute for Studies in Education, 1988. Thompson WD. Statistical analysis of case-control studies. Epidemiol Rev 1994; 16: 33-50. Criqui MH. Response bias and risk ratios in epidemiologic studies. Am J Epidemiol 1979; 109: 394-399. Goodman KJ, Correa P. The transmission of Helicobacter pylori . A critical review of the evidence. Int J Epidemiol 1995; 24: 875-887. Axon A, Forman D. Helicobacter gastroduodenitis: a serious infectious disease. BMJ 1997; 314: 1430-1431. Mitchell HM, Li YY, Hu PJ, et al. Epidemiology of Helicobacter pylori in Southern China: identification of early childhood as the critical period for acquisition. J Infect Dis 1992; 166: 149-153. Nguyen A-MH, Engstrand L, Genta RM, et al. Detection of Helicobacter pylori in dental plaque by reverse transcription-polymerase chain reaction. J Clin Microbiol 1993; 31: 783-787. Khandaker K, Palmer KR, Eastwood MA. DNA fingerprints of Helicobacter pylori from mouth and antrum of patients with chronic ulcer dyspepsia. Lancet 1993; 342: 751. Hardo PG, Tugnait A, Hassan F, et al. Helicobacter pylori infection and dental care. Gut 1995; 37: 44-46. (Received 18 Feb, accepted 21 Jul, 1997) Authors' details Department of Public Health and Community Medicine, University of Melbourne, Ballarat Health Services Base Hospital, Ballarat, VIC. Hedley G Peach, PhD, FFPHM, Professor; Dora C Pearce, BAppSc(Biol), Research Fellow. Epidemiology and Biostatistics Unit, Department of Public Health and Community Medicine, University of Melbourne, VIC. Stephen J Farish, MEd, Senior Lecturer. Reprints: Professor H G Peach, Department of Public Health and Community Medicine, University of Melbourne, Ballarat Health Services Base Hospital, PO Box 577, Ballarat, VIC 3353. E-mail: a.temperley AT phcm.unimelb.edu.au ©MJA 1997 <URL: http://www.mja.com.au/> © 1997 Medical Journal of Australia.

Hedley G Peach · Dora C Pearce · Stephen J Farish

Early discharge and risk for postnatal depression Anthea R Hickey, Philip M Boyce, David Ellwood and Allen D Morris-Yates For editorial comment, see Buist Readers may print a single copy for personal use. No further reproduction or distribution of the articles should proceed without the permission of the publisher. For permission, contact the Australasian Medical Publishing Company Journalists are welcome to write news stories based on what they read here, but should acknowledge their source as "an article published on the Internet by The Medical Journal of Australia <http://www.mja.com.au/>". Abstract - Introduction - Methods - Results - Discussion - References - Authors' details - - ©MJA1997 Abstract Objective: To determine whether early discharge ( < 72 hours) after childbirth increased the risk for women developing postnatal depression. Design: Prospective cohort design consisting of an initial interview, and six-weekly assessments for 24 weeks using a self-report questionnaire and the Edinburgh Postnatal Depression Scale (EPDS). Women discharged within 72 hours were compared with the remaining women. Setting: Tertiary referral hospital in western Sydney, New South Wales, 1993. Participants: All 749 women delivering over a three-month period were recruited. Of the 522 participants, 425 women completed the study. Main outcome measures: Women scoring > 13 on the EPDS on two or more occasions were considered potential "cases" of postnatal depression. The diagnosis was confirmed using the Structured Clinical Interview for DSM-III-R disorders (SCID). Results: Of the 153 women (36%) discharged early, 22 women (14.4%) developed postnatal depression over the study period compared with 20 of the 272 women (7.4%) who had standard length of stay. Women who were discharged within 72 hours had a significantly increased risk for developing postnatal depression (odds ratio [OR], 2.12; 95% confidence interval [CI], 1.07-4.21). This risk persisted when other sociodemographic, obstetric and psychosocial risk factors were controlled for in a logistic regression analysis (OR, 3.06; 95% CI, 1.22-7.69). Conclusion: Women planning early discharge after childbirth should be carefully assessed before discharge and follow-up should be rigorous. The potential to develop postnatal depression should be considered in all women choosing early discharge from hospital. MJA 1997; 167: 244-247 Introduction Over the past 20 years, technological advances have contributed to improved maternal and infant outcomes. In parallel, women have become more involved in their obstetric care: wanting to be more informed, be able to make choices and have childbirth demedicalised and "normalised". One initiative aimed at normalising the process has been early discharge programs, enabling women to return home earlier after childbirth than had been traditional. Women's desire for early discharge has fitted in well with the financial constraints put upon health care systems; reduced length of stay is seen as desirable because it reduces the cost of obstetric care. Postpartum studies have reported early-discharge programs to be a safe, cost-effective and satisfying alternative to the standard length of hospital stay among low-risk mothers and their infants.1-5 While these studies have shown that early discharge has no adverse physical effects, the only psychological effects examined have been women's satisfaction about making a choice or taking control of discharge rather than psychological morbidity.5,6 As postnatal depression is a common disorder which may have a long-term impact on the mother and her infant,7 the psychological outcome after childbirth is increasingly being recognised as important.8 Postnatal depression is associated with psychosocial factors, such as an excess of adverse life events, lack of social support, dysfunctional personality style and a past history of depression.8,9 Obstetric factors, such as emergency caesarean section, have also been implicated,10,11 but the impact of early discharge has not been extensively examined. The aim of our study was to investigate whether early discharge following childbirth had an effect on the development of postnatal depression in a heterogeneous sample of women. Methods Subjects All women who delivered at the Nepean Hospital, NSW, over a three-and-a-half-month period in 1993 were invited to participate in the study. Women were excluded from the study if their baby was stillborn, born with congenital abnormalities, adopted out, transferred to another hospital, or if the mother was diagnosed as suffering from puerperal psychosis. Early discharge was defined as discharge within 72 hours (the definition used at Nepean Hospital). Survey instruments Questionnaire: A self-report questionnaire was used to assess social and demographic circumstances. Socioeconomic status was determined according to a status-ranking list of occupations in Australia,12 using the women's partner's occupation (or her parents' occupation when this was not appropriate). A nine-item personality scale was included that identified different personality traits, scored on a 5-point Likert scale. Factor analysis of this personality scale identified "vulnerable" and "resilient" personality dimensions. The vulnerable personality dimension included being a worrier, nervy, timid and unassertive, sensitive, liable to emotional outbursts and needing to do things right. This scale is reliable (Cronbach α= 0.75), and is associated with increased risk for postnatal depression (Boyce P M, Hickey A R, Greenstock J A and Talley N J, Nepean Hospital, unpublished data). Interview: A semi-structured interview (conducted by A R H) was used to assess obstetric history, satisfaction with the labour, whether the delivery was congruent with expectations, the quality of the woman's social support, her personal and family psychiatric history, and the quality of her intimate relationships. We also obtained obstetric details from the labour ward computerised database. Edinburgh Postnatal Depression Scale (EPDS):13 The EPDS is a specifically designed measure of postnatal depression that has shown reliability and validity for the postpartum period.14-16 It consists of a 10-item self-report scale, with each item scored 0-3. A score of 13 or above is indicative of postnatal depression. It is not contaminated by symptoms (e.g., fatigue, weight loss, broken sleep) which may be reported by healthy nursing mothers. Tennant and Andrews life events scale:17,18 A self-report scale comprising 10 relevant life events (e.g., death of a family member or close friend, moving house, marriage); subjects were asked whether each event had happened to her over the past 12 months. A modified version of this scale was used in our study. Structured Clinical Interview for DSM-III-R (SCID):19 The SCID was administered to confirm postnatal depression "caseness". Postnatal depression was defined as meeting the criteria for major depression according to the Diagnostic and statistical manual of mental disorders, third edition, revised (DSM-III-R).20 Procedure The women were recruited immediately after delivery and asked if they would participate in a longitudinal study of postnatal depression. Women who agreed were given an initial semi-structured interview on the second or third day post partum, a self-report questionnaire and a baseline EPDS. The women were assessed a further four times -- at six, 12, 18 and 24 weeks post partum. At each interval, the women were posted the EPDS, a self-report questionnaire (consisting of questions on the baby's condition and feeding patterns as well as the quality of the women's relationship) and a stamped self-addressed envelope for returning the forms. Women who scored > 13 on the EPDS on two occasions (excluding their baseline EPDS score) were further interviewed using the SCID to determine postnatal depression "caseness". An equal number of women who scored < 13 on the EPDS were randomly selected and interviewed using the SCID. Statistical analysis Statistical analysis was performed using the SPSS computer program.21 Women were categorised as "cases" or "non-cases" of postnatal depression and into early and standard length of stay. Categorical data were analysed using odds ratios (OR), and chi-squared and t tests were used for comparing groups on dimensional data. A logistic regression analysis was used to determine whether length of stay remained a risk factor when controlling for other risk factors. Results Sample Of the 749 women who delivered during the three-and-a-half month period, 522 women (69.7%) agreed to participate in the study. Full data were available for 425 women (81.4%). The mean age of the sample was 26.9 years (standard deviation [SD], 5.0; range, 15-43), 367 women (86.4%) were married or in a de facto relationship, 58 women (13.6%) were single or separated, and 184 women (43.3%) were primiparous. The average period of formal education was 11.4 years (SD, 1.9; range, 9-15). Twenty-six women did not meet the inclusion criteria, 122 refused to participate in the study and 79 were unable to be contacted. There were no statistically significant differences in demographic and obstetric details between women who consented to participate and women who did not consent. The 97 women who withdrew from the study or had insufficient data were more likely to be younger (t = 5.91; P < 0.01), unemployed (chi-squared = 33.37; P < 0.01), single or separated (chi-squared = 59.46; P < 0.01) and have fewer years of formal education (t = 3.47; P < 0.01) compared with women who had full data. Forty-two women (9.9%) were categorised as having major depression, based upon scoring > 13 on the EPDS on two occasions and meeting DSM-III-R criteria for major depression. All of these women met "caseness" criteria for major depression on the SCID. None of the sample of women who screened negative (< 13) on the EPDS and were interviewed using the SCID met "caseness" criteria for major depression. Length of stay The average length of stay in hospital following delivery was 90.9 hours (SD, 43.1 hours; range, 16-307). Of the 425 participants in the study, 153 women (36%) were discharged within 72 hours (early-discharge group) and 272 women (64%) were discharged after 72 hours (standard-length-of-stay group). We compared sociodemographic and risk factors for the early-discharge and standard-length-of-stay groups of women. The early-discharge group were more likely to be multiparous (71.2% v. 48.5%; OR, 2.63; 95% confidence interval [CI], 1.69-4.10), have fewer years of formal education (11.01 years v. 11.5 years; t = 3.09; P < 0.01), to bottle feed in the first week post partum (20.9% v. 12.1%; OR, 1.91; 95% CI, 1.06-3.46), and report a poor relationship with parents (52.9% v. 40.1%; OR, 1.68; 95% CI, 1.11-2.56). They were also more likely to have a history of depression or postnatal depression, but this was not statistically significant (9.1% v. 4.4%; OR, 2.18; 95% CI, 0.92-5.19). The two groups did not differ with respect to age, marital or socioeconomic status, number of life events in the past year, or global satisfaction with their partner. Women in the early discharge group did not have a more vulnerable personality style and did not differ in baseline EPDS scores (5.1 v. 5.9; t = 0.24). Postnatal assessment The proportion of women who developed postnatal depression (as defined above) was higher among those discharged early than those discharged after three days. Three of the 13 women (23.1%) discharged within 24 hours, nine of the 58 women (15.5%) discharged between 24 and 48 hours, and 10 of the 82 women (12.2%) discharged between 48 to 72 hours developed postnatal depression. Of the 153 women discharged early (within 72 hours), 22 (14.4%) developed postnatal depression compared with 20 of the 272 women (7.4%) who had standard length of stay (OR, 2.12; 95% CI, 1.07-4.21). Of the 42 women who developed postnatal depression, 22 (52.4%) were discharged early. The possibility that the increased risk associated with early discharge could have arisen as a result of other, antecedent, risk factors for postnatal depression was examined using a hierarchical logistic regression,22 with postnatal depression as the dependent variable. Sociodemographic, historical, obstetric, and psychosocial factors (see below) were entered in the first step, baseline EPDS score in the second step and early discharge in the final step to see whether it would remain a risk factor after controlling for other risk factors. Sociodemographic factors: Age (5-yearly intervals), level of education (less or more than 10 years), social class (low or high socioeconomic status) and marital status (single, married or de facto relationship). Obstetric factors: Parity (multiparous or primiparous) and delivery type (spontaneous vaginal delivery, instrumental delivery or caesarean section). Psychosocial risk factors: Personality style (high or low vulnerability [extracted from the personality scale, with women scoring one standard deviation above the mean on the vulnerability component of the scale considered to be vulnerable]); quality of relationship with parents and in-laws (poor or good); satisfaction with relationship with partner (dissatisfied or satisfied); history of depression or postnatal depression; experience of none, one or two, or three or more life events over the previous 12 months; and whether they had had the desired-sex baby. When all sociodemographic, obstetric and psychosocial factors were entered, the model was highly significant (Box), with significantly increased risk for postnatal depression associated with dissatisfaction with partner relationship(s), reporting three or more life events in the past 12 months, and having a vulnerable personality. The model significantly improved (had better predictive power; chi-squared = 41.5; P < 0.0001) when baseline EPDS score was entered, and again when early discharge was entered. Early discharge increased the risk threefold (OR, 3.06; 95% CI, 1.22-7.69), after controlling for other risk factors. These results also support the more specific finding of early discharge, in which there was a 2.1-fold increased risk (95% CI, 1.07-4.21) of developing postnatal depression. Discussion Early discharge following childbirth is becoming more routine. It has been shown that neither maternal nor infant health is compromised by this practice; however, there is a paucity of research evaluating the psychological functioning of the mother following early discharge. We are aware of only one prospective study of the psychological morbidity associated with early discharge compared with customary length of stay in hospital. Beck et al.23 assessed depressive symptoms among 49 women at six and 12 weeks post partum using the Beck Depression Inventory (BDI).24 They found no significant difference between the early- versus standard-discharge groups on BDI scores. However, this study was confined to privately insured, primiparous women who had had uncomplicated pregnancy and labour. Therefore, their results have limited generalisability because most studies have identified that women opting for early discharge are more likely to be multiparous, older, have fewer years of formal education and lower socioeconomic status.25,26 While there have been no prospective studies on the psychological impact of early discharge in heterogeneous samples, one retrospective study found that women who reported that their postnatal hospital stay was too short were significantly more likely to have postnatal depression than those who reported the right length of stay.27 Our study shows that women discharged within three days of childbirth had an increased risk of developing postnatal depression. This increased risk is not the result of other well recognised risk factors for postnatal depression, such as experiencing life stresses, having a dysfunctional personality style, unsatisfactory interpersonal relationships28 or a history of depression. When these risk factors were controlled in a logistic regression analysis, early discharge was still a significant risk for postnatal depression, with an odds ratio of 3.06. Therefore, early discharge exerted an independent effect on the development of postnatal depression and was not simply a proxy for other, pre-existing, risk factors. We also found that one-third of the women who stayed in hospital for more than five days also experienced postnatal depression. However, they had all experienced complicated labours and some infants had been in neonatal intensive care for a prolonged period, which may have contributed to the higher rates of postnatal depression. Women discharged within three days of giving birth may be at an increased risk of postnatal depression because they have had insufficient time to recover from and work through the experience of childbirth. They may still be tired from the physical demands of childbirth and have to return to the stresses of everyday life, which includes looking after a household, their new baby and, in many cases, other children. This additional stress so soon after childbirth may increase the risk of depression. The "blues" are minor, transient depressive symptoms which generally arise on the third to fifth day post partum. Therefore, women on an early-discharge program leave hospital before the onset of the "blues". If they return to an environment where there is insufficient support, experiencing the "blues" may lead to depressive symptoms being perpetuated and contribute to the development of postnatal depression. Finally, women discharged early go home without having properly established breastfeeding, and therefore without the supervision and round-the-clock support of midwives to help them with the inevitable difficulties (e.g., breast engorgement, change in milk flow, unable to latch the baby correctly, leading to cracked, sore nipples) that arise after the milk "comes in" (generally on the third day). These difficulties can lead to an irritable baby and a tired, depressed mother. The association we found between early discharge and postnatal depression merely suggests causality. A prospective controlled study with women randomly assigned to early- or standard-discharge programs (controlling for parity, social support and past history) is needed to identify any causal link between early discharge and postnatal depression. The association between early discharge and postnatal depression has major public health implications. While health services are having to cut costs, early discharge may result in short-term cost savings. However, the consequences of postnatal depression could lead to escalating health care costs in the long term. Women entering an early-discharge program need a careful assessment, particularly those at risk for postnatal depression (e.g., poor family support or a history of depression). The possibility of postnatal depression should be kept in mind for all women who have been discharged early when they return for their postnatal check-ups. The EPDS could be used as a routine screen for such women. Finally, women with a history of postnatal depression should have counselling if they are considering early discharge and be advised that a longer hospital stay may reduce their risk of recurrent depression. References Yanover MJ, Jones D, Miller MD. Perinatal care of low-risk mothers and infants: Early discharge with home care. N Engl J Med 1976; 294: 702-705. Avery MD, Fournier LC, Jones PL, Sipovic CP. An early postpartum hospital discharge program: implementation and evaluation. J Obstet Gynecol Neonatal Nurs 1982; 11: 233-235. Burnell I, McCarthy M, Chamberlain GVP, et al. Patient preference and postnatal hospital stay. J Obstet Gynaecol 1982; 3: 43-47. Lemmer CM. Early discharge: outcomes of primiparas and their infants. J Obstet Gynecol Neonatal Nurs 1987; 16: 230-236. Hall WA, Carty EM. Managing the early discharge experience: taking control. J Adv Nurs 1993; 18: 574-582. Kenny P, King MT, Cameron S, Shiell A. Satisfaction with postnatal care -- the choice of home or hospital. Midwifery 1993; 9: 146-153. Murray L, Cooper PJ, Stein A. Postnatal depression and infant development. BMJ 1991; 302: 978-979. Boyce PM, Stubbs JM. The importance of postnatal depression [editorial]. Med J Aust 1994; 161: 471-472. O'Hara MW, Swain AM. Rates and risks of postpartum depression -- a meta-analysis. Int Rev Psychiatry 1996; 8: 37-54. Boyce PM, Todd AL. Increased risk of postnatal depression after emergency caesarean section. Med J Aust 1992; 157: 172-174. Murray L, Cartwright W. The role of obstetric factors in postpartum depression. J Reprod Infant Psychol 1993; 11: 215-219. Congalton AA, Cheshire FW. Status and prestige in Australia. Melbourne: Chesire Publishing Pty Ltd, 1969. Cox JL, Holden JM, Sagovsky R. Detection of postnatal depression: development of the 10-item Edinburgh Postnatal Depression Scale. Br J Psychiatry 1987; 150: 782-786. Harris B, Huckle P, Thomas R, et al. The use of rating scales to identify post-natal depression. Br J Psychiatry 1989; 154: 813-817. Murray L, Carothers AD. The validation of the Edinburgh Postnatal Depression Scale on a community sample. Br J Psychiatry 1990; 157: 288-290. Boyce PM, Stubbs JM, Todd AL. The Edinburgh Postnatal Depression Scale: validation for an Australian sample. Aust N Z J Psychiatry 1993; 27: 472-476. Tennant C, Andrews G. A scale to measure the cause of life events. Aust N Z J Psychiatry 1977; 11: 163-167. Brugha T, Bebbington P, Tennant C, Hurry J. The list of threatening experiences: a subset of 12 life event categories with considerable long-term contextual threat. Psychol Med 1985; 15: 189-194. Spitzer RL, Williams JBW, Gibbon M, First MB. Structured Clinical Interview for DSM-III-R: SCID user's guide for the structured clinical interview for DSM-III-R. Washington: American Psychiatric Press Inc, 1990. American Psychiatric Association. Diagnostic and statisitical manual of mental disorders, third edition revised (DSM-III-R). Washington DC: American Psychiatric Association, 1987. SPSS: Statistical package for the social sciences [computer program], release 6.1. Chicago, Ill: SPSS Inc, 1994. Fleiss JL, Williams JBW, Dubro AF. The logistic regression analysis of psychiatric data. J Psychiatr Res 1986; 20: 195-209. Beck CT, Reynolds MA, Rutowski P. Maternity blues and postpartum depression. J Obstet Gynecol Neonatal Nurs 1992; 21: 287-293. Beck AT, Ward CH, Mendelson M, et al. An inventory for measuring depression. Arch Gen Psychiatry 1961; 4: 561-571. Scott A. A cost analysis of early discharge and domiciliary visits versus standard hospital care for low-risk obstetric clients. Aust J Public Health 1994; 18: 96-100. Mitchell SD, Counsell AM, Geddis DC. Planned early discharge from New Zealand maternity hospitals. N Z Med J 1993; 106: 152-154. Astbury J, Brown S, Lumley J, Small R. Birth events, birth experiences and social differences in postnatal depression. Aust J Public Health 1994; 18: 176-184. O'Hara MW, Zekoski EM. Postpartum depression: a comprehensive review. In: Kumar R, Brockington IF, editors. Motherhood and mental illness 2: causes and consequences. London: Wright, 1988: 17-63. (Received 17 Feb, accepted 5 June 1997) Authors' details Department of Psychological Medicine, Nepean Hospital, Penrith, NSW. Anthea R Hickey, Psychologist, BPsych, MMedSc; Philip M Boyce, MD, FRANZCP, Professor of Psychiatry; Allen D Morris-Yates, BA(Hons), Senior Research Officer. Department of Obstetrics and Gynaecology, Canberra Hospital, ACT. David Ellwood, FRACOG, DPhil(Oxon), Professor of Obstetrics and Gynaecology. Reprints: Prof P M Boyce, Department of Psychological Medicine, Nepean Hospital, PO Box 63, Penrith, NSW 2751. E-mail: pboyceATmail.usyd.edu.au ©MJA 1997 <URL: http://www.mja.com.au/> © 1997 Medical Journal of Australia.

Anthea R Hickey · Philip M Boyce · David Ellwood · Allen D Morris-Yates