Research

Objectives: To determine whether hospital patients identified as Indigenous are less likely than other inpatients to have a principal procedure recorded, and the extent to which any disparity in procedure use can be explained by differences in patient, episode and hospital characteristics. Design: Retrospective analysis of routinely collected administrative data from the National Hospital Morbidity Database (NHMD). Setting: Australian public and private hospitals. Patients: All patients included in the NHMD whose episode type was recorded as acute and whose separation occurred between 1 July 1997 and 30 June 1998. Patients admitted for routine dialysis treatment were excluded. Main outcome measure: Whether a principal procedure was recorded. Results: In public hospitals, patients identified as Indigenous were significantly less likely than other patients to have a principal procedure recorded, even after adjusting for patient, episode and hospital characteristics (adjusted odds ratio [OR], 0.67; 95% CI, 0.66–0.68). This disparity was apparent for most diseases and conditions. In private hospitals, no significant difference was observed (adjusted OR, 0.94; 95% CI, 0.83–1.06). Conclusions: The disparity in procedure use after adjustment for relevant factors indicates that in Australian public hospitals there may be systematic differences in the treatment of patients identified as Indigenous.

Joan Cunningham ScD (Epidemiol)

Objectives: To describe the management of early-pregnancy bleeding and miscarriage reported by general practitioners in Victoria.Design, setting, and participants: Self-administered, mailed survey of a stratified random sample of GPs in Victoria. Responses weighted by strata to reflect GP population.Main outcome measures: Reported management in referral; investigation (especially ultrasound); expectant versus interventional management; and prevention of rhesus iso-immunisationResults: 382 of 621 eligible GPs responded (response rate, 62%). GPs' reported referral was more likely if the patient had painful bleeding (55%) or if the pregnancy was not viable (77%). Ultrasound strongly influenced the assessment of bleeding. Two-thirds of doctors (262/369; 66%) would routinely order ultrasound for painless bleeding, and 328/369 (84%) for painful bleeding. Expectant management was recommended by 15/353 (4%) for incomplete miscarriage with light bleeding and by 6/351 (2%) when bleeding was heavy. Some GPs are uncertain of the indications for anti-D prophylaxis, including instrumentation of the uterus, for which 261/337 (77%) said they would routinely offer anti-D. There was less agreement about anti-D after threatened miscarriage, for which 213/353 (57%) said they offered the injection.Conclusions: GPs need a working knowledge of the management of early-pregnancy bleeding, and can probably encourage more rational management. There are significant areas where GPs are uncertain, often reflecting uncertainty elsewhere, and some areas where a minority of GPs are not aware of essential requirements.

Bruce McLaren DRANZCOG, FRACGP, MPH · Julia M Shelley MPH, PhD

The research enterprise The Menzies Centre for Population Health Research A unique and supportive local population was a vital ingredient in the Centre's success Terence Dwyer MJA 2001; 175: 617-620 Early days: the Tasmanian Infant Health Survey and Sudden Infant Death Syndrome - The post-SIDS era: taking stock - New directions - Genomics - The future - References - Authors' details - - More articles on Psychiatry I ACCEPTED THE CHAIR IN COMMUNITY HEALTH at the University of Tasmania in 1985 with the intention of setting up a research centre that focused on epidemiological research into preventable causes of disease. While I did not know how the centre would be funded, I was certain that Tasmania would be a very competitive site for such research. Already, valuable epidemiological studies on iodine deficiency, hydatid disease and asthma had been conducted in the absence of significant research infrastructure.1 The "Island State" provided a perfect source population for unbiased selection of cases and comparison samples or controls. Further, the land area and population size (around 500 000 people) made follow-up of cohorts relatively easy. Thus, Tasmania had important advantages for the two major strategies used to search for environmental and lifestyle causes of disease — case-control and cohort studies. Funding from the Menzies Foundation came about through the input of three people — Basil Hetzel, then Chief of the Commonwealth Scientific and Industrial Research Organisation Division of Human Nutrition in Adelaide, who had a close association with the Menzies Foundation; Professor Ian Lewis, Dean of the Medical School at the University of Tasmania and a member of the Menzies Foundation Board; and Eric Wigglesworth, the Director of the Foundation. To determine the likely success of such a centre, the Foundation Board held a three-day workshop attended by representatives of State and Federal health departments and the World Health Organization, notable Australians in the field of public health, and distinguished British epidemiologist Sir Richard Doll. History of the Menzies Centre 1987 Workshop ("Towards a Centre for Population Health Research") in Hobart, Tasmania. 1988 Official opening in January. Collection of Tasmanian Infant Health Survey (TIHS) data began (prospective study on Sudden Infant Death Syndrome [SIDS]). 1990 Designated as a World Health Organisation Collaborating Centre for the Prevention of Cardiovascular Diseases. 1991 Provides prospective evidence confirming importance of prone sleeping position as a cause of SIDS (Lancet 1991; 337: 1244-1247). 1992 Evidence that SIDS death rate was falling after a national campaign on infant sleeping position. 1993 Research helps explain how prone position interacts with other factors to increase risk (N Engl J Med 1993; 329: 377-382). 1995 First follow-up of TIHS cohort searching for early life influences on childhood diseases. Shows that the major decline in SIDS deaths from 1991 onwards is the result of changes in infant sleeping position (JAMA 1995; 273: 783-789) 1997 Contract signed with AMRAD pharmaceutical company. Provides funding for Genetic Epidemiology Unit. 1997-2000 Follow-up of the Tasmanian Infant Health Survey cohort into childhood provides important evidence about early life determinants of risk for osteoporosis, blood pressure and asthma (J Clin Endocrinol Metab 1998; 83: 4274-4279; J Bone Miner Res 1999; 14: 146-151; BMJ 1999; 319: 1325-1329; Thorax 1999; 54: 664-669). 2000 Named "Tasmanian Icon" by State Premier. Core funding doubles. Key events Major scientific achievements. Subsequently, the Menzies Foundation Board decided to support the establishment of an epidemiology research centre, to be named the Menzies Centre for Population Health Research. The Foundation then met with the Tasmanian Premier and Minister for Health, who matched the Foundation's initial contribution of $100 000 per year. Early days: the Tasmanian Infant Health Survey and Sudden Infant Death Syndrome Before my departure from Sydney University, I had been reviewing the data on disease distribution in Tasmania. Sudden Infant Death Syndrome (SIDS), with an annual rate in Tasmania twice the national average, stood out. The head of neonatology at the Royal Hobart Hospital, Neville Newman, convinced me that this should be the subject of a major research effort. The cause had not been clearly identified, and epidemiological research had been limited. With helpful input from Geoffrey Berry, Professor of Biostatistics at Sydney University, we planned the first prospective cohort study on this condition. Preliminary work began just before the decision of the Menzies Foundation to support the establishment of the Centre. The epidemiology research group within the Medical School at the University of Tasmania consisted of one epidemiologist, the research fellow Trevor Beard, and limited support staff. Even with the extra $200 000 that the establishment of the new Centre brought, it would not have been realistic to work on a broad front. It was decided that we would focus most of our effort on the new SIDS research program. The next step was to build an appropriately skilled team. We advertised for another epidemiologist and a biostatistician, but it proved very difficult to attract qualified applicants. It seemed that Australian academics were either not interested in living in Tasmania, or were not confident their careers would flourish there. This problem was compensated for by a stroke of good luck when a young Tasmanian medical graduate, Anne-Louise Ponsonby, became our first postgraduate student, working on SIDS. She put an incredible amount of intelligently directed energy into the SIDS program, and together, with financial help from the Australian Rotary Health Research Fund, we were able to develop momentum in the project. In 1988, we initiated the first full data collection for the cohort study — a huge endeavour that involved measurements each year in 1500 infants and their mothers on three occasions in the first three months after birth. That we could get this work under way was pleasing, but we needed to find well-qualified biostatisticians. Given the previous lack of success with advertising in Australia, I decided to use our international network. Sir Richard Doll referred Michael Jones, a young Master of Science graduate from Oxford, who was recruited to our ranks, and then Laura Gibbons, from the University of Massachusetts, joined us. This relatively small and young team of investigators coordinated the conduct, data management and analysis of the SIDS program. They also assisted with less well resourced but developing areas in cancer and cardiovascular disease. In late 1990 evidence was accumulating from case-control studies that prone sleeping position might be a major cause of SIDS, but the research was retrospective, creating concerns that recall bias might explain the findings. We had the only prospective data in the world and were able to show that the association was equally strong prospectively, ruling out recall bias.2 A number of countries, including Australia, launched campaigns to encourage parents not to place babies on their stomachs in the cot, with astonishing results — the death rate from SIDS in Australia fell from 507 in 1990 to 139 in 1998, with similar falls in a number of other countries.3 While our work was not the only important contribution to the understanding of this major cause of SIDS, it provided an important piece of evidence needed for solving the puzzle. Later, in 1993, our team explained why prone sleeping position seemed to exert a different effect in winter than summer and a different effect across countries.4 Then, in 1995, we provided evidence that showed clearly that the fall in deaths could only be attributed to the changes in prevalence of prone sleeping position.5 This success will undoubtedly rank as one of the major contributions of the Centre in the years to come. It also established the organisation as one which, in its special location, could have a significant impact on international medical science. It was the much-needed platform that would underpin future recognition and opportunities. The post-SIDS era: taking stock The death rate from SIDS fell so rapidly after the prone sleeping position campaign that, by late 1991, it was clear there would eventually be insufficient cases occurring annually in Tasmania for epidemiological research (when we started the SIDS program, there had been an average of 27 cases a year for an extended period, and by 1998 there were only three). While this outcome was tremendously gratifying, it was clear that the research money to support our staff of now approximately 20 would dry up unless we repositioned our research program. This was confirmed by the National Health and Medical Research Council (NHMRC) Regional Grants Interview Committee's decision not to recommend refunding of our cohort study for 1992. We went from triumph to a period of considerable adversity. One of our first responses was to tell the Tasmanian public that we needed its financial and moral support. They responded generously. With a major public fundraising appeal, helped greatly by our Board and new Chairman John Tomlinson, and a timely decision by the United States National Institutes of Health, we were able to continue the study long enough to thoroughly evaluate the impact of the prone sleeping intervention campaign. In 1992, I took some time to review where we were going as an organisation and to think about where our future research opportunities might lie. I visited people like Richard Doll in Oxford, who had provided very helpful mentoring since 1987. I also had discussions with Ken Rothman (author of Modern epidemiology6), and Dimitris Trichopoulos at Harvard. These visits confirmed that, if we were to continue to conduct work of global significance, we would have to search even more thoroughly for gaps in knowledge that might be filled by an epidemiological approach. I was also convinced that we would need to develop stronger working relationships with basic scientists if we wished to use epidemiology to understand aetiology. These strategies were challenging, but all our team had learned a great deal from the SIDS research experience. While overseas, I also upgraded my skills in organisation and management by attending a management course in Salzburg, led by Peter Drucker, one of the world's most prominent management theorists. New directions The perspectives gained during my overseas visit were incorporated into planning from 1992 onwards. The major new strategy we decided on was to follow the Tasmanian Infant Health Survey (TIHS) cohort, now numbering 11 000 infants and children. The focus would be to search for links between early life exposures and later disease, using our extensive database of infant measurements that provided information on more than 450 variables measured during the first three months of life. One disease we looked at was asthma, an important disease for which preventable causes had not yet been identified and for which there was a shortage of good epidemiological data. This investigation would be coordinated by Anne-Louise Ponsonby, with help from David Couper, a biostatistician who had joined us from Seattle. We also increased our activity in research on the development in childhood of risk factors for cardiovascular disease and diabetes. Fitting into this theme was the new program started by a recent recruit from the Garvan Institute in Sydney, Graeme Jones (we were finally starting to see interest in work opportunities from well-qualified Australians outside Tasmania). He had a strong track record in osteoporosis in the elderly, and he used that background to focus on the impact of early-life factors on bone density in childhood. Fortuitously, interest in the "Barker hypothesis", which concerns the impact of fetal development on later disease, was gaining momentum. We were well placed to make an important contribution in this field, and our capacity was greatly enhanced by the addition of Ruth Morley, from the Institute of Child Health in London. Supplemented by smaller research efforts in cancer and adult cardiovascular disease, by 1994 we were able to see evidence that the research program was growing again. Between 1994 and 2000, the team was able to attract 17 new NHMRC grants from 38 applications submitted. This overall level of success was built on the tremendous preparedness of the Tasmanian public to be involved in the research. Response rates for case-control studies in this period were about 90% for cases and 80% for controls sampled from the electoral rolls, with comparable figures for cohort follow-up. In addition to our growing research effort we took on an important role in ensuring that knowledge was transferred to countries with less developed research capacity. The World Health Organization designated our institution as a Collaborating Centre for the Prevention of Cardiovascular Disease (CVD) a decade ago. That role has expanded steadily to the point where the Centre is assisting in studies on CVD in countries including Vietnam, Fiji and Samoa, where CVD and diabetes are producing an unexpectedly high disease burden. Genomics For the first seven years (1988-1995) the Centre focused solely on the environmental and lifestyle causes of disease. Meanwhile, others had been using the deep family pedigrees available in Tasmania to search for genetic causes of diseases following a Mendelian pattern of inheritance. Novel genes or linkages were discovered for several conditions, including multiple endocrine neoplasia and Huntington's disease. These successes were based on special features of Tasmania that are replicated in few other locations, namely (i) a population descended largely from identifiable founder families; (ii) comprehensive genealogical records; (iii) a modern healthcare system capable of identifying disease outcomes; (iv) a demonstrated capacity to involve the population in studies; and (v) organisational structures to facilitate the research. In 1995, David Mackey, a Tasmanian medical graduate and ophthalmologist at the Victorian Eye and Ear Hospital, approached us. He was undertaking important work in Tasmania on the more complex genetics of glaucoma. He wanted a base in Tasmania, and the Walter and Eliza Hall Institute, in Melbourne, was seeking a Tasmanian institution to manage new research and development syndicate funds to support his research. We accepted the role and our interest in the use of epidemiology to find genes for human diseases increased. Then, in 1996, the Australian pharmaceutical company AMRAD approached us about increasing its involvement in gene discovery in Tasmania through the Menzies Centre. We agreed, on the condition that the funding would be for a genetic unit that would employ people who could provide intellectual input to the work from a Tasmanian base. AMRAD signed a contract in 1997 for a five-year grant of $2.5 million, and, in 1998, we attracted Tasmanian molecular geneticist Michele Sale to coordinate the work. With financial and other help we were able to very quickly get projects under way in multiple sclerosis and osteoarthritis, and have continued to develop activity with Cerylid, a spin-off from AMRAD formed to operate its discovery arm. The genomics development has also led to an increase in postgraduate student training at the Centre, with five PhD students currently enrolled. The future During the past 12 months there have been several important developments for the Centre, driven by our very committed Board, chaired by Jean Trethewey, and strongly supported by the Dean of the Faculty of Health Science at the University of Tasmania, Allan Carmichael. The Tasmanian Government introduced an "Icons Program", which supports the Tasmanian Symphony Orchestra and our State cricket team. To this list Premier Jim Bacon added the Menzies Centre, with a commitment to provide $500 000 a year to help our organisation develop its capabilities. This, together with a large donation in 2000 from the United States-based Atlantic Philanthropies Inc, has placed us in a previously unimagined position to recruit more staff and drive our research program. To enable us to undertake these future developments with vigour, the University of Tasmania Council has established the Menzies Centre as an independent company limited by guarantee, remaining within the university structure. In 2002, the Centre will become the "Menzies Research Institute". Already, the Centre has grown to support a staff of 60. The new institute is likely to start 2002 with a budget of approximately $5 000 000 that will see staff numbers increase to more than 100, working on both environmental and genetic causes of disease. A major NHMRC grant of $2 290 000 over the next five years will enable us to study a cohort of Australians first measured as schoolchildren in 1985. They will be followed up for the emergence of adult disease, and it is anticipated this will provide the first direct evidence available on the impact of childhood lifestyle and biology on diseases such as coronary heart disease. A collaboration with similar cohorts in the US and Finland has already been established. A new director of the Cohort Studies Unit, Alison Venn, who has a strong background in this research strategy, has been recruited from the Centre for the Study of Mothers' and Children's Health at La Trobe University to coordinate developments. A large adult cohort study in Tasmania, with a focus on exposures that occur closer to the time of disease development, will also commence in 2002. Both studies will benefit from the input of a now-strong biostatistics group of three staff headed by one of our own PhD graduates, Leigh Blizzard. The level of genetic research activity will expand greatly. Tasmania presents opportunities as good as any in the world for gene discovery, and we intend to take up these opportunities. An increasing number of epidemiological studies at the Centre are focused on finding novel genes or validating candidate genes identified through animal or cell studies, or bioinformatic "data mining". This growth in activity reflects the recognition by commercial and government sources of the opportunities here, as well as the developing capacity of our genetic unit. Increasingly, our "environmental" epidemiologists and biostatisticians are developing their interests and skills in genetic research. This has not only led to the more rapid development of a critical mass for projects on gene discovery and validation, but has also opened up the possibility for in-depth investigation of gene-environment interaction. Projects with this focus are already under way in multiple sclerosis. There is great scope for us to contribute in an internationally significant way to the understanding of gene-environment interactions using Tasmania's unique population and our skill base. In the coming decade Australia will be relying more and more on its medical research institutes to maintain its competitive advantage in a knowledge-based global economy. We are confident that the new Menzies Research Institute will be making its contribution. References King H, editor. Epidemiology in Tasmania. Canberra: Brolga Press, 1987. Dwyer T, Ponsonby AL, Newman NM, Gibbons LE. Prospective cohort study of prone sleeping position and sudden infant death syndrome. Lancet 1991; 337: 1244-1247. Australian Bureau of Statistics. Deaths, Australia, 1990, 1998. Canberra: ABS, 1998. (Catalogue no. 3302.0/3303.0.) Ponsonby AL, Dwyer T, Gibbons LE, et al. Factors potentiating the risk of SIDS associated with the prone position. N Engl J Medicine 1993; 329: 377-382. Dwyer T, Ponsonby AL, Blizzard CL, et al. The contribution of changes in the prevalence of prone sleeping position to the decline in SIDS in Tasmania. JAMA 1995; 273: 783-789. Rothman K. Modern epidemiology. Boston: Little John and Co., 1986. Authors' details Menzies Centre for Population Health Research Terence Dwyer, MD, FAFPHM, Director. Reprints will not be available from the author. Correspondence: Professor T Dwyer, Menzies Centre for Population Health Research, 17 Liverpool Street, Hobart, 7000 TAS. t.dwyerATutas.edu.au Make a comment

Terence Dwyer

Research Mega-dose vitamin C in treatment of the common cold: a randomised controlled trial Carmen Audera, Roger V Patulny, Beate H Sander and Robert M Douglas MJA 2001; 175: 359-362 Abstract - Methods - Results - Discussion - Acknowledgements - Competing interests - Reference - Authors' details - - - More articles on Infectious diseases and parasitology Abstract Objective: To determine the effect of large doses of vitamin C in the treatment of the common cold. Study design: Double-blind, randomised clinical trial with four intervention arms: vitamin C at daily doses of 0.03 g ("placebo"), 1 g, 3 g, or 3 g with additives ("Bio-C") taken at onset of a cold and for the following two days. Participants and setting: 400 healthy volunteers were recruited from staff and students of the Australian National University, Canberra, ACT, between May 1998 and November 1999. The trial continued for 18 months. Interventions: Participants were instructed to commence medication when they had experienced early symptoms of a cold for four hours, and to record daily their symptoms, severity, doctor visits and use of other medications. Main outcome measures: Duration of symptoms and cold episodes; cumulative symptom severity scores after 7, 14 and 28 days; doctor visits; and whether participants guessed which medication they were taking. Results: 149 participants returned records for 184 cold episodes. No significant differences were observed in any measure of cold duration or severity between the four medication groups. Although differences were not significant, the placebo group had the shortest duration of nasal, systemic and overall symptoms, and the lowest mean severity score at 14 days, and the second lowest at 7 and 28 days. Conclusions: Doses of vitamin C in excess of 1g daily taken shortly after onset of a cold did not reduce the duration or severity of cold symptoms in healthy adult volunteers when compared with a vitamin C dose less than the minimum recommended daily intake. A recent Cochrane systematic review of the effects of vitamin C on the common cold concluded that large maintenance doses of vitamin C do not lower the incidence of colds in well-nourished subjects in Western countries.1 Nevertheless, the meta-analysis of 17 trials found that prophylactic doses of at least 1g per day were associated with a statistically significant weighted mean reduction in symptom days of about 0.45 days per cold (9% of symptom days).1 However, the authors of the Cochrane review could not draw conclusions about the therapeutic effects of vitamin C (ie, effects when taken at onset of a cold).1 Findings of four well-conducted trials of the effects of treating colds with a loading dose of vitamin C were inconclusive2-5(Box 1). This prompted us to design a study to answer the question "Would vitamin C, when used exclusively as a therapeutic agent in doses that greatly exceed the required daily intake, reduce the duration or severity of symptoms of the common cold in healthy Australian adults?". Methods Our study was a double-blind, randomised trial comparing the effects of different doses and formulations of vitamin C. We chose as "placebo" a dose of 0.03g per day of vitamin C (about half the recommended minimum daily intake), recognising that all participants would have some nutritional vitamin C intake. Ethics approval was obtained from the Human Ethics Committee of the Australian National University, Canberra. Participants Staff and students of the Australian National University, Canberra, ACT, were recruited between May 1998 and November 1999 through personal letters and emails, announcements at student gatherings and direct approach in university common areas. Volunteers were eligible for the study if they were aged over 18 years, not pregnant or planning to become pregnant, in good general health, and did not take vitamin supplements regularly or take vitamin C, echinacea, zinc or Chinese herbal preparations regularly at the onset of a cold. Volunteers were clearly informed about the objectives of the study and signed an informed consent form. They also completed a questionnaire about their current health and medication status, including respiratory infections in the previous year. An information letter was provided for their general practitioners. Participants who returned information on one respiratory event were eligible to re-enrol in the study. Interventions Participants were randomised to receive one of four interventions: vitamin C in a daily dose of 0.03 g, 1 g or 3 g, or "Bio-C" (containing vitamin C [3 g daily] plus bioflavenoids [75 mg], rutin [150 mg], hisperidin [150 mg], rose hip extract [750 mg] and acerola [150 mg]). They were to take the medication at onset of cold symptoms and on the following two days. The medications were prepared by Blackmores Ltd (Sydney, NSW) as compressed tablets with identical appearance and packaging. Dosage was confirmed by chemical analysis of unused tablets at the end of the study. A random number table was constructed to order the medications sequentially so that each sequence of four numbers comprised all four types of medication. The medications were issued to investigators in 400 sequentially numbered sets of three bottles, each bottle containing the daily dose in three tablets. As volunteers joined the study they were given a set of three bottles and a correspondingly numbered "respiratory event card" to record outcome. The code was retained by the manufacturer until we were ready to analyse the results. Participants were instructed that they must have at least two of the following symptoms for a minimum of four hours before commencing medication: sore or scratchy throat, nasal congestion or discharge, headache or stinging eyes, muscle aches, fever, or "four hours of certainty that a cold is coming on". On the first day of illness, they were to take the contents of one bottle (three tablets) as soon as possible. For the next two days, they were to take three tablets a day at intervals of at least four hours. Outcome measures The respiratory event card was designed to be carried in a wallet or purse. When a cold began, participants were instructed to score symptoms daily, noting presence and severity (1, mild; 2, moderate; or 3, severe) of cough, nasal, throat, and systemic symptoms, including fever, headache, aches, feeling unwell and "other symptoms". Recording was to cease either when all symptoms disappeared or 28 days after onset of the cold. Participants were also instructed to record hours between onset of symptoms and first dose of medication, use of other medication and whether they sought medical attention. They were also invited to guess to which medication group they had been assigned. Duration of the cold was measured from day of symptom onset to the last day of any symptom. Cold severity scores were the sum of daily individual symptom scores throughout the duration of the cold. Symptom days and severity scores for cough, nasal, throat and systemic symptoms were considered separately, and cumulative scores were considered at 7, 14, and 28 days. For any one day of symptoms, the maximum severity score was 12. Participants who did not return a respiratory event card were sent reminder letters after nine months and 15 months. The initial 12-month study period was extended by six months in an effort to increase the response rate. Statistical analysis We aimed to study 75 individuals in each intervention arm, in the expectation that the study would have an 80% power to detect a 30% difference between groups in duration or severity, which we considered clinically significant. Desired sample size was calculated assuming a mean duration of seven days and a standard deviation of four days. Statistical comparisons were carried out using the software package SPSS.6 Distribution, mean and median of duration and severity scores for each symptom were compared between the four groups by t-tests, analysis of variance and box plots. Results Study population Four hundred sets of medication were distributed to 323 volunteers. By November 1999, when the study was terminated, 149 people had returned completed respiratory event cards for 184 cold episodes. These 149 were significantly older than those who did not return cards (45.1 versus 40.9 years; P < 0.05), but the two groups did not differ significantly in sex distribution or previous cold history. Personal characteristics and previous cold history of those who returned cards are shown in Box 2, along with time from symptom onset to beginning medication. Participants in the four medication groups were comparable in sex distribution and time to beginning medication, but those who took Bio C were significantly older and had fewer colds in the previous year than those in the other three groups (P < 0.05). Cold duration and severity Duration and severity of symptoms are compared between the four medication groups in Box 2. There were no significant differences between the groups in either mean duration of symptoms or mean severity scores at Days 7, 14 or 28, although the placebo group (30 mg vitamin C daily) had the shortest duration of nasal, systemic and overall symptoms, and the lowest mean severity score at 14 days, and the second lowest at 7 and 28 days. A box plot of cumulative severity scores at Day 28 (Box 3) revealed that the distribution of values was more dispersed in the 1 g and 3 g vitamin C groups, with the lowest median values occurring in the placebo and Bio C groups. A box plot of cold duration showed a similar pattern (Box 3). Only 31 participants (17%) recorded a guess about the dose of vitamin C they had taken, and 14 guessed correctly that they had taken a high dose. Seventeen gussed incorrectly that they had taken either a high or low dose. Actual power of the study Because the mean cold duration for the whole group was 9.8 days with a standard deviation of 6.6 days, the number of completed cold episodes returned per group provided 80% power to detect a 40% difference in cold duration with a 95% level of confidence. Similarly, given that the mean severity score at Day 28 was 32 with a standard deviation of 32.3, the number of completed cold episodes provided 80% power to detect a 50% difference in severity at the 95% level of confidence. Discussion Our study found no significant differences in severity or duration of cold symptoms between groups who took low-dose (placebo) and high-dose vitamin C as treatment for the common cold. The lack of benefit from high-dose therapeutic vitamin C is consistent with the findings of four other randomised controlled trials2-5 (Box 1). The Cochrane and other reviews of the published evidence on high-dose vitamin C and the common cold have drawn attention to the relatively consistent trend for those taking prophylactic doses in excess of 1 g daily to experience some reduction in duration or severity of colds.1,7-9 Although high-dose prophylactic vitamin C was also found not to reduce the incidence of colds in well-nourished adult populations,1,7 Hemila has proposed that it may have an effect in groups who are physically stressed or have low nutritional intake.8-10 The main weakness of our study is that it necessarily relied on study participants to decide when the criteria for commencing medication were met and to provide all outcome data. In such a study, double-blindness must be rigorously preserved, and allocation to intervention arms must avoid selection bias. We are confident that our study met these requirements and that the few participants who correctly guessed their medication dose did so by chance. The focus on the university community meant a potential bias in socioeconomic and educational status of participants. The observed spectrum of cold experience may not have been representative of the cold experience of the rest of the Canberra community. Many potential volunteers in our study were ruled ineligible because of their regular use of vitamin C and other, non-traditional approaches for cold therapy and prophylaxis. A recent US study found that 67% of patients seeking medical care for cold episodes believed that vitamin C reduces cold symptoms.11 Our target of 75 colds in each treatment group was not reached, despite extension of the study and repeated reminder letters to participants. Fewer than half those enrolled returned a completed respiratory event card. As we expected most to suffer at least one cold during the 18 months of the study, based on their previous history, we assume that many did not use the medication as instructed. Although those who completed a respiratory event card were older than those who did not, both groups had similar previous cold experience. The double-blind nature of the study makes it unlikely that greater compliance would have changed the result. Our study had medication groups of comparable size, and for each medication group colds were found to have occurred across the entire study period. The Bio-C group was slightly older than the other groups and, probably in consequence, experienced fewer colds in the previous year, as the incidence of colds tends to decrease with age. However, these differences were not associated with significant differences in outcomes. The average time between symptom onset and medication use was 13 hours, although we encouraged participants to begin medication as soon as four hours after they were certain that a cold was developing. However, the time to beginning medication did not differ significantly between groups. The power of our study to detect a possible significant difference in symptom severity and duration after high-dose vitamin C treatment was limited by the smaller than expected participation rate. However, the non-significant trend that was observed was the reverse: symptoms tended to be less severe and of shorter duration in the placebo group. The lack of observed benefit in this trial is fully consistent with the observations from the four previous randomised controlled trials that have sought to evaluate this issue.2-5 It is time to question again the wisdom and utility of the wide practice of well nourished adults taking mega-doses of vitamin C to treat the common cold, a practice which has become prevalent worldwide since the advocacy of Linus Pauling in the early 1970s.12,13 Acknowledgements The project was supported by a grant from Blackmores Ltd, who also provided the study medications. We thank all those who participated in the study for their patience and compliance. Competing interests Blackmores Ltd were not involved in conduct or analysis of the trial or preparation of this article. References Douglas RM, Chalker EB, Treacy B. Vitamin C for preventing and treating the common cold (Cochrane Review). In: The Cochrane Library, 3, 2001. Oxford: Update Software. Anderson TN, Suranyi B, Beaton GW. The effect on winter illness of large doses of vitamin C. Can Med Assoc J 1974; 111: 31-38. Karlowski TR, Chalmers TC, Frenkel LD, et al. Ascorbic acid for the common cold. A prophylactic and therapeutic trial. JAMA 1975; 231: 1038-1042. Elwood PC, Hughes SJ, St Leger AS. A randomized controlled trial of the therapeutic effect of vitamin C in the common cold. Practitioner 1977; 218: 133-137. Tyrrell DA, Craig JW, Meada TW, White T. A trial of ascorbic acid in the treatment of the common cold. Br J Prev Soc Med 1977; 31: 189-191. SPSS [computer program]. Version 10.0 for Windows. Chicago, Ill: SPSS Inc, 1999. Hemila H. Vitamin C and the common cold. Br J Nutr 1992; 31: 3-16. Hemila H. Vitamin C supplementation and the common cold: was Linus Pauling right or wrong? Int J Vitam Nutr Res 1997; 67: 329-325. Hemila H. Vitamin C and common cold incidence: A review of studies with subjects under heavy physical stress. Int J Sports Med 1996; 17: 379-383. Hemila H, Douglas RM. Vitamin C and acute respiratory infections. Int J Tuberc Lung Dis 1999; 3: 756-761. Braun BL, Fowles JB, Solberg L, et al. Patient beliefs about the characteristics, causes, and care of the common cold: an update. J Fam Pract 2000; 49: 153-156. Pauling L. The significance of the evidence about ascorbic acid and the common cold. Proc Natl Acad Sci USA 1971; 68: 2678-2681. Pauling L. Vitamin C, the common cold, and the flu. San Francisco: Freeman, 1976. Authors' details National Centre for Epidemiology and Population Health, Australian National University, Canberra, ACT. Carmen Audera, MD, MPH, Lecturer; Roger V Patulny, BEc, BA (Hons), Research Assistant; Beate H Sander, BAppSc (Nursing), MEcDev, Research Assistant; Robert M Douglas, MD, FRACP, FAFPHM, Visiting Fellow. Reprints will not be available from the authors. Correspondence: Emeritus Professor R M Douglas, National Centre for Epidemiology and Population Health, Australian National University, Canberra, ACT 0200. Bob. DouglasATanu.edu.au Make a comment 1: Previous randomised controlled trials of the therapeutic effect of high-dose vitamin C on cold symptoms Study Participants and setting Interventions Outcomes Anderson et al2 (1974) Toronto, Canada Hospital and business employees (>275 per arm) 4 arms: 2 placebo, 2 therapeutic (4g or 8g vitamin C taken on day of symptom onset) The two placebo arms unfortunately differed in outcome. Mean days of respiratory symptoms over 3 months: placebo, 5.4 and 4.16 days (combined placebo mean, 4.77 days); intervention, 4.82 days (4g dose) and 4.52 days (8g) Karlowski et al3 (1975) Bethesda, USA National Institutes of Health employees (46 placebo, 43 therapy) 3g vitamin C daily or placebo for the first 5 days of a cold Problem in blinding, as over half the participants correctly guessed their medication through taste. Although mean duration of colds was longer in the placebo than therapy group (7.1 v 6.5 days), difference was confined to those who guessed their medication ("unblinded"). Unblinded group: 8.6 (placebo) v 4.7 days (therapy); blinded group: 6.3 (placebo) versus 6.7 days (therapy). Elwood et al,4 (1977) South Wales, UK Community volunteers (119 placebo, 145 therapy) 3g vitamin C daily or placebo for 3 days Vitamin C significantly reduced duration of "simple" colds in men (5.7 days [placebo] v 3.97 days [therapy]), but had no benefit in women (4.97 days [placebo] v 6.05 [therapy]), or in "chest" colds in either sex. Tyrrell et al5 (1977) Salisbury, UK 482 volunteers 4g of vitamin C or an identical-tasting placebo daily for 2.5 days No evidence that vitamin C alleviated or shortened upper respiratory or general constitutional symptoms. Back to text 2: Characteristics of participants and outcomes of a study of the effect of therapeutic vitamin C on the common cold Vitamin C formulation (daily dose) 0.03g (n=42)* 1g (n=47)* 3g (n=50)* Participant characteristics Mean age in years (95% CI) 38.6 (34.2-43.0) 40.1 (35.8-44.4) 39.9 (36.2-43.6) Male sex (95% CI) 45% (30%-61%) 38% (26%-54%) 50% (36%-65%) Cold history in previous year Mean number of colds (95% CI) 2.2 (1.7-2.7) (n=40) 2.25 (1.9-2.6) (n=46) 2.2 (1.8-2.7) (n=49) Mean number of days unwell from colds (95% CI) 8.0 (3.4-12.5) (n=39) 7.7 (6.2-9.3) (n=46) 7.7 (6.5-9.2) (n=49) Mean hours from symptom onset to medication (95% CI) 13.3 (9.4-17.2) (n=39) 11.6 (8.7-14.7) (n=44) 10.2 (8.2-12.3) (n=48) Outcome measures Mean days of symptom (95% CI) 8.5 (6.6-10.5) 10.1 (8.1-12.1) 10.4 (8.5-12.2) Cough 5.3 (3.0-7.6) 6.4 (4.1-8.6) 6.3 (4.4-8.3) Nasal symptoms 7.3 (5.4-9.1) 8.4 (6.7-10.1) 9.2 (7.4-11.1) Throat symptoms 5.4 (3.6-7.2) 6.1 (4.3-7.9) 6.3 (4.6-7.9) Systemic symptoms 3.5 (2.1-4.9) 3.7 (2.3-5.2) 3.8 (2.7-4.8) Mean severity score‡ (95% CI) Day 7 20.2 (16.5-24.0) 22.1 (18.1-26.0) 23.0 (19.3-26.6) Day 14 25.6 (19.0-32.1) 31.1 (23.5-38.8) 30.8 (24.9-36.6) Day 28 29.0 (19.5-38.6) 35.4 (23.4-47.5) 34.3 (26.6-42.1) Doctor visit (95% CI) 7% (2%-20%) 19% (8%-31%) 4% (0.5%-14%) Other medication taken for symptoms (95% CI) 57% (41%-72%) 55% (40%-70%) 55% (39%-68%) Vitamin C formulation (daily dose) "Bio C" (3g plus additives) (n=45)* Total (n=184)* Participant characteristics Mean age in years (95% CI) 45.1 (40.6-49.5)† 40.9 (38.8-43.0) Male sex (95% CI) 51% (36%-66%) 46% (39%-54%) Cold history in previous year Mean number of colds (95% CI) 1.5 (1.3-1.8)† (n=44) 2.1 (1.9-2.3) (n=179) Mean number of days unwell from colds (95% CI) 6.8 (5.4-8.2) (n=43) 7.5 (6.4-8.9) (n=177) Mean hours from symptom onset to medication (95% CI) 18.6 (11.2-26) (n=44) 13.4 (11.1-15.8) (n=175) Outcome measures Mean days of symptom (95% CI) 9.9 (7.9-11.9) 9.8 (8.8-10.7) Cough 4.4 (2.2-6.5) 5.6 (4.6-6.7) Nasal symptoms 8.1 (6.1-10.1) 8.3 (7.4-9.2) Throat symptoms 5.4 (3.8-6.9) 5.8 (5.0-6.7) Systemic symptoms 4.4 (3.2-5.6) 3.9 (3.2-4.5) Mean severity score‡ (95% CI) Day 7 19.2 (15.4-23.0) 21.2 (19.3-23.0) Day 14 25.9 (19.1-32.6) 28.5 (25.2-32.8) Day 28 28.6 (20.0-37.3) 32.0 (27.3-36.7) Doctor visit (95% CI) 9% (3%-21%) 9% (6%-14%) Other medication taken for symptoms (95% CI) 53% (38%-68%) 55% (47%-62%) *Number of completed cold episodes; 35 participants were counted twice, as they reported two medicated colds. For variables with missing data, numbers of participants who provided information are shown in parentheses. †P Back to text 3: Box plots of cold severity and duration in groups taking different vitamin C formulations Back to text

Carmen Audera · Roger V Patulny · Beate H Sander · Robert M Douglas

Abstract Objective: To examine the relationship between compensation claims for psychiatric injury after motor vehicle accidents and physical injuries sustained. Design: Audit of Compulsory Third Party (CTP) insurance claims. Subjects and setting: 559 consecutive CTP claims referred by NRMA Insurance Limited to its sole provider of CTP legal services during a three-month period in 1994 after the claimant had engaged legal representation. Main outcome measures: Claim for psychiatric injury (any psychiatric disorder excluding traumatic brain injury) supported by a medicolegal report from a psychiatrist, other medical practitioner or psychologist; pre-existing psychiatric disorders; Injury Severity Score; initial treatment setting; hospital stay; percentage of accidents involving loss of consciousness or a death. Results: 522 claims were eligible for the study; 19.5% (102/522) included a claim for psychiatric injury. A pre-existing depression or anxiety disorder was documented in 11 claims (2.1% of all claims and 3.9% of those claiming psychiatric injury). Only very severe injuries, particularly those involving loss of consciousness, were associated with an increased rate of claims for psychiatric injury. Conclusions: No association was found between claims for psychiatric injury and severity of physical injuries, except among those most severely injured. More than 25 000 people are injured in motor vehicle accidents in New South Wales each year.1 Data from the NSW Motor Accidents Authority from 1998 show that about 60% of people who made a claim after a motor vehicle accident obtained legal representation,2 and that the percentage of claims with a psychiatric component rose from 2.2% to 8% between 1990 and 1998.3 NSW Motor Accidents Authority data also show that minor physical injuries result in 54% of all claims, but 77% of claims for psychiatric injury.3Motor vehicle accidents are associated with post-traumatic stress disorder (PTSD), other anxiety disorders and depression,4,5 although most recent publications refer to PTSD rather than other syndromes.4-9 However, the true incidence of PTSD and other disorders after these accidents is unclear; most studies have sampling bias and other methodological problems.4 The extent to which motor vehicle accidents cause the observed psychiatric disorders is also uncertain. Factors associated with PTSD include those not directly related to the accident, such as past psychiatric history6-9 and involvement in litigation,6,7 and others that are difficult to assess objectively, such as victims' recollections of fear of death6-10 and self-reported loss of consciousness.6,7 The relationship between PTSD and severity of physical injuries has also been investigated,6-13 with one study finding a positive correlation.7 In this study, I examined the relationship between compensation claims for psychiatric injury after motor vehicle accidents and physical injuries sustained. Methods Sample The sample comprised 559 consecutive claims on Compulsory Third Party (CTP) insurance that were referred by NRMA Insurance Limited to its sole provider of CTP legal services in a three-month period in 1994 after the claimant had engaged legal representation. All claims arose from motor vehicle accidents that occurred in New South Wales between 1989 and 1994. In this period, 39.5% of NRMA CTP claimants had legal representation. Of the 559 claims, 37 were excluded from the study as files were missing (25), the claim was from bereaved relatives (10) or the claimant had died since the claim (2). A further 50 files had data missing on one or more of the following: injuries, demographic details, or setting of medical care. As these files did not to appear to include claims for psychiatric injuries, they were included in denominators for proportions with psychiatric injury but were excluded from further analysis. Data collection and analysis I collected de-identified data from the files on systematic forms. The dependent variable was a claim for psychiatric injury (defined as any psychiatric disorder, excluding traumatic brain injury) supported by a report from a psychiatrist, other medical practitioner or psychologist. I recorded the presence of one or more of these injuries or traumatic brain injury attributed to the accident by the claimant's experts, as well as any pre-existing psychiatric disorder noted by claimants' or defendants' experts. Independent variables recorded were age and sex, type of motor vehicle accident and whether fatal (ie, any person killed), Injury Severity Score (ISS)14(calculated from information in the injury summary document in each file), setting of initial medical care (most medically intensive setting in the week after the accident), length of hospital stay, loss of consciousness during or after the accident (self-reported or corroborated), and self-reported neck or back pain (irrespective of physical or radiological signs). Data were analysed using the computer program SPSS.15 Logistic regression was used to determine the influence of independent variables on the presence of a claim for psychiatric injury. Results Most claimants (380; 73%) were passengers or drivers, and the remainder were motorcyclists, cyclists or pedestrians (92; 18%). The status of another 50 (10%) was not known because of incomplete files. Mean age was 34 years (range, 2-82 years), and 49% were male. Claims for psychiatric injury Claims for psychiatric or traumatic brain injury are shown in Box 1. One hundred and two people (19.5%) claimed at least one psychiatric injury related to the accident, combined with traumatic brain injury in six cases (another 11 people claimed traumatic brain injury alone). Thirty-six people claimed more than one psychiatric injury. Pre-existing psychiatric disorders are also shown in Box 1. These were documented in 25 people (4.8%), and comprised a depressive or anxiety disorder in 11 (2.1% of all claimants, and 3.9% of those claiming psychiatric injury). The reports supporting the psychiatric injury claims came from psychiatrists (65), psychologists (28) and other medical practitioners (9); mean time between the accidents and report dates was over two years. Experts disagreed on many claims, with treating practitioners and claimants' experts using the diagnoses of PTSD and depression (15% of claimants) more often than defendants' experts (2.5% of claimants), as described elsewhere.16 Variables associated with psychiatric injury claims The group that claimed psychiatric or traumatic brain injury had significantly longer hospital stay and higher mean ISS and proportion of accidents involving loss of consciousness than the group who claimed neither type of injury (Box 2). The group that claimed psychiatric injury but not traumatic brain injury also had significantly more accidents involving loss of consciousness and fatal accidents compared with those who claimed neither type of injury. There were no significant differences between the groups in proportions with self-reported neck or back pain. For the 102 who claimed psychiatric injury, the most medically intensive treatment in the week after the accident was provided by a local medical officer (33), in an emergency department (35), as a general inpatient (25) or in an intensive care unit (9). Psychiatric injury claims and injury characteristics are shown in Box 3 by initial treatment setting. The percentage of people who claimed psychiatric injury was significantly higher in those treated initially in intensive care than in those treated elsewhere (χ2 = 6.74; df = 1; P = 0.009). The percentage who claimed for PTSD and traumatic brain injury was also higher in the intensive care group (PTSD: χ2 = 8.99; df = 1; P = 0.003; and traumatic brain injury: χ2 = 103; df = 1; P <0.001). Mean ISS, hospital stay and percentage who reported loss of consciousness were also greater in those treated in more medically intensive settings (inpatient and intensive care), supporting the use of initial treatment setting as an indicator of injury severity. However, claims for neck or back pain were lower in claimants who received inpatient or intensive care treatment (Box 3). The percentage of people who claimed for a psychiatric injury did not increase with increasing ISS over the first nine deciles (mean, 20%). However, the percentage was significantly higher (33%) in people in the tenth ISS decile (ie, the most severely injured 10%) compared with those in the lower nine deciles (χ2= 4.34; df = 1; P = 0.04). A logistic regression analysis was performed using variables found to be significantly related to claims for psychiatric or traumatic brain injury by previous analyses (Box 4). Loss of consciousness and a fatal accident were significant predictors of claims for psychiatric injury. Discussion A claim for psychiatric injury was made in 19.5% of the legally represented CTP claims in this study, which is higher than the 4.6% estimated by the NSW Motor Accidents Authority for all victims of motor vehicle accidents during the same period.3 This confirms the previously reported association between psychiatric injury and legal representation.6,7 PTSD and depression were reported more often in this sample than in a recent survey of the Australian population.17 Conversely, pre-existing psychiatric disorders were documented much less often than the estimated prevalence of all psychiatric conditions in Australia (4.7% v. 17.7%17). This suggests that medicolegal assessments may under-report pre-existing psychiatric disorders and may sometimes wrongly identify a motor vehicle accident as the cause of a depressive or anxiety disorder that was actually pre-existing. In contrast to findings of the NSW Motor Accidents Authority,3 my study found that minor injuries were no more likely to be associated with a psychiatric injury than more severe injuries. However, a third of psychiatric injury claims (33/102) were made by people with physical injuries that were not severe enough for them to attend an emergency department or be admitted to hospital at the time of the accident. The study did find a positive relationship between the severity of physical injuries and claims for psychiatric injury in people who were very seriously injured. The psychological trauma of being severely injured may cause PTSD.5,7 Severe physical injuries may also cause psychiatric symptoms because of disability, pain or financial loss.5 However, in my study, the higher rate of claims for psychiatric injury in severely injured claimants was associated with loss of consciousness and involvement in a fatal accident rather than with other measures of injury severity. Reported loss of consciousness may be difficult to distinguish from amnesia resulting from emotional stress,18 which may predispose to psychiatric injury.19 The use of insurance claimants as the sample in this study led to selection bias and may have influenced the psychiatric injuries diagnosed by experts. The incidence and prognosis of whiplash injury are influenced by the system of assessing eligibility for compensation,20 and psychiatric injury may be similarly affected. More severely injured claimants may under-report their psychiatric symptoms because they are more concerned about their physical injuries and because the grounds for compensation for physical injuries have been clearly established. Claimants who are not seriously injured but are hurt or upset and have received less initial medical care may report more psychiatric symptoms. The opinions of expert witnesses may, in turn, be influenced by their role in the adversarial legal system.16 Under the current NSW system, some claimants may exaggerate their disability or genuinely become disabled because "significant disability" is a requirement for compensation;21 claimants who have a psychiatric injury but are less disabled are not compensated. Reform of the rules on expert evidence designed to reduce bias22,23 and a move to more detailed assessment of the cause of psychiatric symptoms after motor vehicle accidents may reduce the pressure for the NSW government to further limit psychiatric injury claims. Acknowledgements I would like to acknowledge NRMA Insurance Limited and Mr Victor Kelly of Abbott Tout Solicitors, Sydney, NSW, for making claimants' files available; Dr Timothy Heath (Concord Repatriation and General Hospital, Sydney, NSW) for his help with data analysis; and Dr Olav Nielssen (Psychiatrist, Sydney, NSW) for his assistance with the manuscript. The study was not funded. References Motor Accidents Authority and Roads and Traffic Authority of New South Wales. Road safety statistics, 2000. Available at <http//www.maa.nsw.gov.au/proftest/statistics/injury/report05.htm> (last sighted Jul 2001). Motor Accidents Authority and Road Traffic Authority of New South Wales. Compulsory third party statistics, 1999. Available at <http//www.maa.nsw.gov. au/professionals/statistics/CTP_stats_98.htm> Suhood S. Claims involving psychological disturbance, September 2000. Sydney: Motor Accidents Authority and Road Traffic Authority of NSW, 2000. Blaszczynski A, Gordon K, Silove D, et al. Psychiatric morbidity following motor vehicle accidents: a review of methodological issues. Compr Psychiatry 1998; 39: 111-121. Mayou R. The psychiatry of road traffic accidents. In: Mitchell M, editor. The aftermath of road traffic accidents. London: Routledge Press, 1997: 33-48. Ehlers A, Mayou RA, Bryant B. Psychological predictors of chronic posttraumatic stress disorder after motor vehicle accidents. J Abnorm Psychol 1998; 107: 508-519. Blanchard EB, Hickling EJ, Taylor AE, et al. Who develops PTSD from motor vehicle accidents? Behav Res Ther 1996; 34: 1-10. Mayou R, Bryant B, Duthie R. Psychiatric consequences of road traffic accidents. BMJ 1993; 307: 647-651. Ursano RJ, Fullerton CS, Epstein RS, et al. Acute and chronic posttraumatic stress disorder in motor vehicle accident victims. Am J Psychiatry 1999; 156: 589-595. Green MM, McFarlane AC, Hunter CE, Griggs WM. Undiagnosed post-traumatic stress disorder following motor vehicle accidents. Med J Australia 1993; 159: 529-534. Feinstein A, Dolan R. Predictors of post traumatic stress disorder following physical trauma: an examination of the stressor criterion. Psychol Med 1991; 21: 85-91. Bryant RA, Harvey AG. Initial posttraumatic stress responses following motor vehicle accidents. J Trauma Stress 1996; 9: 223-234. Blanchard EB, Hickling EJ, Taylor AE, Loos W. Psychiatric morbidity associated with motor vehicle accidents. J Nerv Ment Dis 1995; 183: 495-503. Baker SP, O'Neill B, Haddon W, Long WB. The Injury Severity Score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma 1974; 14: 187-196. SPSS for Windows. Release 9.0.1. Chicago: SPSS Inc, 1999. Large M, Nielssen O. An audit of medico-legal reports prepared for claims of psychiatric injury following motor vehicle accidents. Aust N Z J Psychiatry. In press. Henderson S, Andrews G, Hall W. Australia's mental health: an overview of the general population survey. Aust N Z J Psychiatry 2000; 34: 197-205. Kopelman MD. Fear can interrupt the continuum of memory. J Neurol Neurosurg Psychiatry 2000; 69: 431-432. Mayou RA, Black J, Bryant B. Unconsciousness, amnesia and psychiatric symptoms following road traffic accident injury. Br J Psychiatry 2000; 177: 540-545. Cassidy JD, Carroll LJ, Cote P, et al. Effect of eliminating compensation for pain and suffering on the outcome of insurance claims for whiplash injury. N Engl J Med 2000; 342: 1179-1186. Motor Accidents Compensation Act (NSW) 1999. Friston M. New rules for expert witnesses: The last shots of the medico-legal hired gun. BMJ 1999; 318: 1365-1366. Federal Court of Australia. Practice direction: guidelines for expert witnesses. Canberra: Federal Court of Australia, 1998. Available at <http://www. fedcourt.gov.au/pracproc/practice_direct.html> last sighted Jul 2001. (Received 31 Jul 2000, accepted 3 May 2001) Authors' details Department of Psychiatry, Royal Prince Alfred Hospital, Sydney, NSW. Matthew M Large, FRANZCP, Staff Specialist Psychiatrist. Reprints will not be available from the author. Correspondence: Dr M M Large, Department of Psychiatry, Royal Prince Alfred Hospital, Missenden Road, Camperdown, NSW 2050. mlargeATozemail.com.au Make a comment 1: Number of people claiming psychiatric or traumatic brain injuries and pre-existing psychiatric disorders among 522 insurance claimants Injury or disorder Injury claim Pre-existing disorder Traumatic brain injury (TBI) 17 (3.3%) 0 Post-traumatic stress disorder 48 (9.2%) 0 Depressive disorders 46 (8.8%) 8 (1.5%) Anxiety disorders 15 (2.9%) 3 (0.6%) Somatoform disorders 8 (1.5%) 3 (0.6%) Adjustment disorders 12 (2.3%) 0 Substance abuse 3 (0.6%) 4 (0.8%) Dementia or low IQ 0 3 (0.6%) Schizophrenia 0 2 (0.4%) Other 6 (1.1%) 2 (0.4%) Total* Psychiatric injury 102 (19.5%) 25 (4.7%) Psychiatric injury or TBI 113 (21.6%) 25 (4.7%) *36 people claimed more than one psychiatric injury. Back to text 2: Demographic, injury and accident characteristics among 472 insurance claimants* (95% CI) Variable No psychiatric or traumatic brain injury (n=359) Psychiatric injury (n=102) Psychiatric or traumatic brain injury (n=113) Age in years (95% CI) 33.4 (31.8-35.1) 36.4 (35.5-39.1) 35.7 (33.0-38.3) % Male 49% (44%-54%) 46% (36%-56%) 49% (39%-58%) Hospital stay in days 4 (3-5) 8 (4-12) 11 (6-16) Injury Severity Score† 11.7 (11.0-12.5) 14.0 (12.1-15.9) 16.1 (13.1-18.4) % With loss of consciousness 9% (6%-12%) 25% (16%-33%) 31% (23%-39%) % In fatal accident 1% (0-3%) 8% (3%-13%) 7% (2%-12%) % With neck or back pain 62% (57%-67%) 68% (59%-77%) 64% (55%-73%) Values in bold are significantly different from values for group with no psychiatric injury or TBI, as defined by non-overlapping 95% CIs. * 50 claimants were excluded from this analysis as no information was available on one or more of the following: nature of physical injuries, demographic details, or setting of medical care. † Maximum possible score, 75. Back to text 3: Psychiatric injury claims and injury characteristics among 472 insurance claimants,* according to initial treatment setting† (95% CI) Local medical officer (n=177) Emergency department (n=146) General inpatient (n=129) Intensive care unit (n=20) % With psychiatric injury 19% (13%-24%) 24% (17%-31%) 19% (12%-26%) 45% (26%-63%) % With post-traumatic stress disorder 10% (6%-14%) 10% (5%-15%) 8% (3%-12%) 30% (10%-50%) % With traumatic brain injury 0.6% (0-1.7%) 0.7% (0-2.0%) 5% (1%-8%) 45% (23%-67%) Mean Injury Severity Score 8.0 (7.5-8.5) 10.3 (9.5-11.0) 18.5 (17.0-20.0) 36.7 (31.3-42.0) Mean hospital stay (days) 0 1 13 (9-17) 40 (24-56) % With loss of consciousness 2% (0-4%) 12% (6%-18%) 27% (19%-35%) 60% (38%-82%) % With neck or back pain 81% (75%-87%) 71% (64%-78%) 33% (24%-41%) 20% (2%-38%) PTSD=Post-traumatic stress disorder. TBI=Traumatic brain injury. * 50 claimants were excluded from this analysis as no information was available on one or more of the following: nature of physical injuries, demographic details, or setting of medical care. † Most medically intensive treatment setting in the week after the accident. Back to text 4: Multivariate logistic regression analysis of variables potentially associated with psychiatric injury claims Odds ratio (95% CI) P Injury severity score* 0.98 (0.95-1.02) 0.29 ICU treatment 1.86 (0.80-4.35) 0.15 Hospital stay (days)* 1.00 (0.98-1.02) 0.45 Loss of consciousness 1.82 (1.20-2.88) 0.006 Fatal accident 3.47 (1.52-7.92) 0.003 ICU=Treatment in intensive care unit during first week. *Continuous variables. Back to text

Matthew M Large

MJA 2001; 175: 15-18 For editorial comment, see Hugh & Hugh Abstract - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - More articles on Surgery Abstract Objective: To measure and describe changes in the incidence of appendicectomy in the population of Western Australia (WA) for 1981-1997. Design: Population-based incidence study using hospital discharge data. Setting: All hospitals in WA (1981-1997). Patients: All patients who underwent an appendicectomy in WA hospitals. Main outcome measures: Changes in the incidence of appendicectomy procedures over time; age-standardised rates and age-sex profiles of four appendicectomy subgroups: (1) acute emergency admission, (2) other emergency admission, (3) incidental appendicectomy and (4) other appendicectomy. Results: From 1981 to 1997, there were 59 749 appendicectomies in WA hospitals. The age-standardised rate of appendicectomy declined by 63% in metropolitan females, by 44% in non-metropolitan females, by 41% in metropolitan males and by 21% in non-metropolitan males. The rate of decline was significantly greater in females and in metropolitan patients. From 1988 to 1997, acute emergency admission for appendicectomy was the most common admission status and was more common in males than females (122 v 103 per 100 000 person-years) and in non-metropolitan areas. The rate of incidental appendicectomy was higher among females than males (20 v 7 per 100 000 person-years). From 1988 to 1997, recorded diagnosis coding for appendicitis became more specific, with a marked reduction in the use of the "unspecified" appendicitis code. Conclusions: The overall incidence of appendicectomy has declined markedly in WA and includes a decline in the practice of incidental appendicectomy. The trend was greatest in the metropolitan hospitals. Appendicectomy is one of the most common surgical procedures in adults and children.1-3 Increases in the incidence of appendicitis were reported during the early part of the 20th century, but a decline has been reported since about 1930.4-6 Significant advances in diagnostic and surgical technology may have influenced treatment options for patients and surgical outcomes.3 Linked hospital discharge data from Oxford (UK), 1970-1986, reported by Primatesta and Goldacre, showed falls in acute appendicitis and the prophylactic and incidental use of appendicectomy, but no decline in conditions that mimic the disease.7 The authors raised the concern that appendicectomy without acute appendicitis was much more common in women than men, questioning the appropriateness of the use of the procedure.7 Our study used data from the Quality of Surgical Care Project8 stored in the WA Health Services Linked Database (WA Linked Database)9 to assess trends in appendicectomy in Western Australia (WA) for 1981-1997. Methods The WA Linked Database provided hospital morbidity data for all patients who underwent appendicectomy for 1981-1997. Hospital morbidity records with a separation date before 1988 were selected using the ICPM procedure code 5-470,10 while ICD-9-CM procedure codes 47.0 and 47.1 were used for patients separated in 1988-1997.11 Data for incidental appendicectomy were evaluated only for the period 1988-1997, as there was no specific incidental appendicectomy procedure code before 1988. To allow comparison with the Oxford study,7 patients who underwent appendicectomy were classified into four subgroups based on procedure and diagnosis codes in conjunction with admission status (Box 1). Western Australia occupies the western third of the Australian continent. It is sparsely populated, except for the southwest corner of the State and some coastal settlements to the north. Seventy-three per cent of the total population of 1.9 million reside in the capital city of Perth. We used postcode data to classify patients as residing in Perth (metropolitan) or non-metropolitan areas, following the Health Zone classification system of the Health Department of Western Australia. We estimated annual rates of appendicectomy procedures per 100 000 person-years (PY) by the direct method,12 age standardised to the WA population.13 Population estimates were obtained from the Australian Bureau of Statistics.14 Men and women were analysed separately. We analysed descriptive statistics with the statistical program SPSS,15 and time trends in rates of admission by Poisson regression models using the SAS procedure GENMOD.16 These models included terms for "locality" (metropolitan/non-metropolitan), "time", "age-group" and "sex", and associated rate ratios are reported. Depending on goodness of fit, "time" was modelled either as a single term for linear trend or categorically. In our modelling, we also assessed whether trend effects differed by sex and/or locality by using appropriate higher-order interaction terms. Results Trends in appendicectomy rates, 1981-1997 Of the 59 749 appendicectomies performed in WA in 1981-1997, 33 352 (55.8%) were performed on female patients and 26 397 (44.2%) on males. There was a marked decline in the rate of appendicectomy during the study period (Box 2). The age-standardised rate declined by 63% (from 386 to 144 per 100 000 PY) in metropolitan females, by 44% (from 393 to 221 per 100 000 PY) in non-metropolitan females, by 41% (from 240 to 142 per 100 000 PY) in metropolitan males and by 21% (from 258 to 204 per 100 000 PY) in non-metropolitan males. The decline was more marked in females than males and was also greater in the metropolitan area. The adjusted rate ratio (RR) in metropolitan females fell by 6.2% per year (RR, 0.938; 95% CI, 0.933-0.943), compared with 3.2% per year (RR, 0.968; 95% CI, 0.959-0.976) in non-metropolitan females. For metropolitan males, the adjusted rate ratio declined by 3.9% per year (RR, 0.961; 95%CI, 0.955-0.967), compared with the 1.6% per year decline (RR, 0.984; 95% CI, 0.976-0.993) in non-metropolitan males. Trends in admission classification, 1988-1997 Of the 30 934 appendicectomies performed in WA during 1988-1997, 18 961 (61.3%) were acute emergency admissions, 3820 (12.3%) were other emergency admissions, 2192 (7.1%) were incidental procedures and 5961 (19.3%) were recorded as other appendicectomy admissions. The age-sex profiles for each group are presented in Box 3. Acute emergency admission appendicectomy was more common in males than females (122.2 v 102.9 per 100 000 PY). The highest rates were in males aged 10-14 years (300 per 100 000 PY) and females aged 15-19 years (289 per 100 000 PY). There was an asymptotic decrease in rates of acute emergency appendicectomy after the 20-24-years age group in both sexes. Rates were higher in non-metropolitan areas for males (149 v 111 per 100 000 PY) and females (131 v 93 per 100 000 PY). The difference between metropolitan and non-metropolitan areas remained significant after adjustment for age, sex and year of separation (RR, 1.37; 95% CI, 1.30-1.45). There was a modest increase in the rate ratio of 1.5% per year over time (95% CI, 0.6%-2.4%) for patients in this group, with no difference between metropolitan and non-metropolitan areas in the rate of acute emergency admissions. Rates of other emergency appendicectomies were higher in females than males (31 v 15 per 100 000 PY). In females, the rates were highest in those aged 15-19 years (108 per 100 000 PY) and declined sharply after the 20-24-years age group. Rates were higher in non-metropolitan areas for both females (44 v 26 per 100 000 PY) and males (21 v 13 per 100 000 PY) and this effect remained after adjustment for age, sex and year of separation (RR, 1.66; 95% CI, 1.53-1.80). The age-sex profile of incidental appendicectomies showed a very different pattern. The rate of incidental appendicectomy was higher in females than males (20 v 7 per 100 000 PY). The age profiles were also different, with a sharp, bell-shaped pattern of increase and decrease in women between the ages of 15 and 49 years, with the highest rate occurring in women aged 35-39 years (37 per 100 000 PY). Rates were higher in non-metropolitan areas, with this difference considerably more pronounced in females (29 v 17 per 100 000 PY) than in males (8 v 6 per 100 000 PY). There was a marked decline in the rate of incidental appendicectomies over time among females (Box 4), with a significantly more pronounced trend in metropolitan than non-metropolitan areas (P < 0.001). The primary surgical procedures with which incidental appendicectomies were performed varied by sex. Incidental appendicectomies in females were most frequent during admissions for operations of the uterus (57%) and ovary (24%), and for operations on the intestines (52%), and hernia and abdomen (20%) in males. Rates of other appendicectomy were higher in females than males (50 v 22 per 100 000 PY). The highest rate occurred in females aged 15-19 years (139 per 100 000 PY). Rates in this group were higher in non-metropolitan areas for both females (55 v 48 per 100 000 PY) and males (26 v 20 per 100 000 PY). This locality effect was significant after adjustment for age, sex and year of separation (RR, 1.19; 95% CI, 1.09-1.29). There was a strong linear decrease in the incidence of other appendicectomies, with the rate ratio declining 14.4% per year (95% CI, 13.2%-15.5%). This rate of decline was significantly greater for males (17.2%) than females (13.1%; P = 0.002). Changes in recorded diagnosis, 1988-1997 Changes in the diagnostic profiles of appendicectomy records, excluding incidental appendicectomies, are shown in Box 5. There was a 10-fold reduction in the use of the unspecified appendicitis diagnosis code, with an increase in the use of acute appendicitis diagnosis codes. To assess whether the increased use of acute appendicitis codes was more likely to reflect changes in recording practices rather than in true disease incidence, trends in appendicectomy rates were examined in males aged 10-24 years, as this group predominantly reflected acute emergency admissions. From 1981 to 1997, age-specific rates of appendicectomy in young males declined by 42% (from 692 to 399 per 100 000 PY) in those aged 10-14 years, by 45% (from 629 to 346 per 100 000 PY) in those aged 15-19 years and by 33% (from 373 to 251 per 100 000 PY) in those aged 20-24 years. Discussion The incidence rate of appendicectomy in WA hospitals declined markedly from 1981 to 1997, consistent with trends reported from other industrialised countries.5,6 The age-sex profiles of the four different classifications of appendicectomy defined in our study were similar to those found in the Oxford Record Linkage Study.7 These profiles were unaffected by the different procedure classifications employed, namely ICD-9-CM in our study and the Office of Population Censuses and Surveys Operations Codes in the Oxford study. Improvements in diagnostic technology during the past decade have resulted in a much greater use of compression ultrasonography, laparoscopic examination and scoring systems to verify acute appendicitis in patients with abdominal pain.17,18 These technical improvements may have contributed to the decline in appendicectomy and an improvement in coding practice. Further research is warranted here given a recent finding of no significant benefits from ultrasonography compared with clinical diagnosis alone, other than reduced time to operation.19 Our study found changes in the specificity of coding of recorded diagnoses of appendicitis from 1988 to 1997. In 1988, most diagnoses of appendicitis were recorded using the non-specific code 541.x. By 1997, relatively few diagnoses of appendicitis were assigned this code. There was an increase in the number of diagnoses coded as acute appendicitis either with peritonitis (540.0 or 540.1) or without peritonitis (540.9). This change could be taken to indicate that the incidence of acute appendicitis increased in WA during 1988-1997. However, our data show a fall in the number of appendicectomies in WA since 1981 and a fall among males aged 10-24 years, the group most likely to be admitted with acute appendicitis. A more likely explanation is that there was an improvement over time in the accuracy of coding in WA hospitals. There is now concern about the continued practice of incidental appendicectomy.20 While the physiological role of the appendix is unclear, it may have surgical potential in reconstructive urology and the management of faecal incontinence. The frequency of emergency (acute and other) appendicectomy peaks in the 15-19-years age group, the frequency of incidental appendicectomy peaks in the 35-39-years age group in women and at around 70 years in men. A retrospective review and meta-analysis of incidental appendicectomy by Snyder and Selanders supported incidental removal of the appendix in young patients (< 35 years), suggested that the patient's clinical condition should determine incidental removal between 35-50 years, and could not justify incidental appendicectomy in patients older than 50 years.21 To address the concerns that incidental appendicectomy is unjustified, further comparison of the risk of appendicectomy and the risk of complications (especially adhesion formation) for different age groups is needed. The decline in incidental appendicectomy has also seen a convergence of appendicectomy trends for males and females, which most likely reflects a change in attitude by surgeons. The rate of incidental appendicectomy was about five times higher in females than males in 1988, but had reduced to twice the magnitude by 1997. There was no indication of a parallel decline in other abdominal procedures to account for the decline in appendicectomy rates, although the increased use of laparoscopic procedures may have contributed to the decline in incidental appendicectomy. The decline in the incidence of appendicectomy in WA from 1981 to 1997 is consistent with trends in other industrialised countries and most likely reflects a change in attitude to the use of the procedure, coupled with improvements in diagnostic technology. The trend was most notable in young women in the metropolitan area. There was a fivefold decline in incidental appendicectomy in women in both the metropolitan and non-metropolitan areas. Incidental appendicectomy was more common in women in non-metropolitan areas, which raises questions about differences in practice between the metropolitan and non-metropolitan areas. While the decline in the rates of incidental appendicectomy reflects a change in clinical practice, the question still remains whether incidental appendicectomy is justified to prevent future appendicitis, and does the risk of additional problems and complications outweigh the potential benefit. Acknowledgements We thank the National Health and Medical Research Council for the funds that supported this study, and Dr John Bass and the Extramural Unit of the Western Australian Health Services Research Linked Database Project for the linkage of patient records. Mr Neil Donnelly was on secondment from the Needs Assessment and Health Outcomes Unit, Central Sydney Area Health Service, Sydney, NSW, Australia. References Pearl RH, Hale DA, Molloy M, et al. Pediatric appendectomy. J Pediatric Surg 1995; 30: 173-181. Reid RI, Dobbs BR, Frizelle FA. Risk factors for post-appendectomy intra-abdominal abscess. Aust N Z J Surg 1999; 69: 373-374. Wilcox RT, Traverso LW. Have the evaluation and treatment of acute appendicitis changed with new technology? Surg Clin North Am 1997; 77: 1355-1369. Raguveer-Saran MK, Keddie NC. The falling incidence of appendicitis. Br J Surg 1980; 67: 681. Bisset AF. Appendicectomy in Scotland: a 20-year epidemiological comparison. J Public Health Med 1997; 19: 213-218. Blomqvist P, Ljung H, Nyren O, Ekbom A. Appendectomy in Sweden 1989-1993 assessed by the Inpatient Registry. J Clin Epidemiol 1998; 51: 859-865. Primatesta P, Goldacre MJ. Appendectomy for acute appendicitis and for other conditions: an epidemiological study. Int J Epidemiol 1994; 23: 155-160. Semmens JB, Lawrence-Brown MMD, Fletcher DR, et al. The Quality of Surgical Care Project: a model to evaluate surgical outcomes in Western Australia using population-based record linkage. Aust N Z J Surg 1998; 68: 397-403. Holman CDJ, Bass AJ, Rouse IL, Hobbs MST. Population-based linkage of health records in Western Australia: development of a health services research linked database. Aust N Z J Public Health 1999; 23: 453-459. International classification of procedures in medicine. Geneva: World Health Organization, 1978. The official NCC Australian version of ICD-9-CM. Tabular list (annotated) and index of procedures. Sydney: National Coding Centre, Faculty of Health Sciences, University of Sydney, 1995. Rothman KJ. Modern epidemiology. Boston/Toronto: Little, Brown and Company, 1986. Muir C, Waterhouse J, Mack T, et al. Cancer incidence in five continents, Vol. V. Lyon: IARC Scientific Publications, International Agency for Research on Cancer, 1987. Australian Bureau of Statistics. Estimated resident population by age and sex in statistical local areas, Western Australia (Catalogue no. 3203.5). Canberra: ABS, 1995. SPSS for Windows, release 5.0 [computer program]. Chicago, Ill: SPSS Inc., 1992. SAS version 6.12 [computer program]. Cary, NC: SAS Institute, 1997. Calder JDF, Gajraj H. Recent advances in the diagnosis and treatment of acute appendicitis. Br J Hosp Med 1995; 54: 129-133. Beasley SW. Can we improve the diagnosis of acute appendicitis? [editorial]. BMJ 2000; 321: 907-908. Douglas CD, McPherson NE, Davidson PM, Gani JS. Randomised controlled trial of ultrasonography in diagnosis of acute appendicitis, incorporating the Alvarado score. BMJ 2000; 321: 1-6. Wheeler RA, Malone PS. Use of appendix in reconstructive surgery: a case against incidental appendicectomy. Br J Surg 1991; 78: 1283-1285. Snyder TE, Selanders JR. Incidental appendicectomy — yes or no? A retrospective case study and review of the literature. Infec Dis Obstet Gynecol 1998; 6: 30-37. (Received 20 Sep 2000, accepted 20 Mar 2001) Authors' details Needs Assessment and Health Outcomes Unit, Central Sydney Area Health Service, Sydney, NSW. Neil J Donnelly, BSc (Hons), MPH, Statistician. Centre for Health Services Research, Department of Public Health, The University of Western Australia, Nedlands, WA. James B Semmens, MSc, PhD, Research Fellow, Quality of Surgical Care Project. C D'Arcy J Holman, MB BS, MPH, PhD, Director. University Department of Surgery, Fremantle Hospital, Fremantle, WA. David R Fletcher, MB BS, MD, FRACS, Professor. Reprints will not be available from the authors. Correspondence: Dr James B Semmens, Quality of Surgical Care Project, Centre for Health Services Research, Department of Public Health, The University of Western Australia, Nedlands, WA, 6907. Make a comment 1: Four appendicectomy subgroups Definitions based on ICD-9-CM diagnosis and procedure codes in conjunction with recorded admission type status: Acute emergency admission appendicectomy Diagnosis code for acute appendicitis with or without rupture (540.0, 540.1 or 540.9) + procedure code for appendicectomy (47.0) or Diagnosis code for unspecified appendicitis (541.0 or 541.9) + procedure code for appendicectomy (47.0) + emergency admission type status. Other emergency admission appendicectomy Patients who were clinically hard to define: patients treated with appendicectomy where the diagnosis did not include either acute or unspecified appendicitis (540.x or 541.x) but who were admitted as an emergency case (procedure code for appendicectomy (47.0) + emergency admission type status + any diagnosis codes not including 540.0, 540.1, 540.9, 541.0 or 541.9). Incidental appendicectomy Incidental or prophylactic excision of a normal appendix during abdominal operations (procedure code 47.1). Other appendicectomy All patients with a procedure code for appendicectomy (47.0) not included in subgroups 1 and 2. Back to text Age-standardised total annual incidence rates for appendicectomy in men and women in the metropolitan and non-metropolitan areas of Western Australia for the period 1981-1997. Back to text A: Acute emergency appendicectomy in males and females, Western Australia, 1988-1997. B: Other emergency appendicectomy in males and females, Western Australia, 1988-1997. C: Incidental appendicectomy in males and females, Western Australia, 1988-1997. D: Other appendicectomy in males and females, Western Australia, 1988-1997. Back to text Age-standardised total annual incidence rates for incidental appendicectomy in males and females in the metropolitan and non-metropolitan areas of Western Australia for the period 1988-1997. Back to text 5: Diagnostic profiles of appendicectomy records excluding incidental appendicectomy in Western Australia, 1988-1997 Acute rupture (540.0, 540.1) Acute non-rupture (540.9) Unspecified appendicitis (541.x) Other appendix (542.x, 543.x) Abdominal pain (789.x) Other 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 157 230 215 316 318 311 388 386 469 527 943 988 922 1084 1265 1478 1396 1364 1559 1563 1436 1138 1016 658 448 369 225 195 167 157 102 169 179 219 306 289 296 263 201 213 273 228 248 282 341 378 297 243 206 161 153 167 211 258 246 277 233 207 183 225 Coding numbers used in this table are from ICD-9-CM.11 Back to text

Neil J Donnelly · James B Semmens · David R Fletcher

MJA 2001; 174: 621-625 For editorial comment, see Barraclough Abstract - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - More articles on Administration and health services Abstract Objectives: To determine if an integrated clinical risk management program that detects adverse patient events in a hospital, analyses their risk and takes action can alter the rate of adverse events. Design: Longitudinal survey of adverse patient events over eight years of progressive implementation of the risk management program. Participants and setting: 49 834 inpatients (July 1991 to September 1999) and 20 050 emergency department patients (October 1997 to September 1999) at a rural base hospital in the Wimmera region of Victoria. Main outcome measures: Rates of adverse events detected by medical record review and clinical incident and general practitioner reporting. Results: The annual rate of inpatient adverse events decreased between the first and eighth years of the study from 1.35% of all patient discharges (69 events) to 0.74% (49 events) (P < 0.001). Absolute risk reduction was 0.61% (95% CI, 0.23%-0.99%), and relative risk reduction was 44.9% (95% CI, 16.9%-72.9%). The quarterly rate of emergency department adverse events decreased between the first and eighth quarters of monitoring from 3.26% of all attendances (84 events) to 0.48% (12 events) (P < 0.001). Absolute risk reduction was 2.78% (95% CI, 2.04%-3.52%), and relative risk reduction was 85.3% (95% CI, 62.7%-100%). Conclusions: Adverse patient events can be detected, and their frequency reduced, using multiple detection methods and clinical improvement strategies as part of an integrated clinical risk management program. Healthcare delivery in hospitals is associated with adverse patient events,1,2 and clinical risk management aims to reduce the probability of these events. One approach involves detecting adverse events, analysing their causes, estimating their likelihood and consequences and taking appropriate action to prevent the event recurring. Adverse events can be detected by medical record review3 and clinical incident reporting.4 However, after their detection, analysing the events and determining and taking appropriate action to reduce their rates are difficult tasks. Rates have been reduced in other complex industries, such as aviation, by analysing the systems of service delivery in which the events occurred and changing these systems to reduce their probability.5 This systems approach contrasts strongly with the blaming of individuals for errors in healthcare. In this study, we report the effect of an integrated clinical risk management program that used diverse methods to detect adverse patient events in a hospital and a systems approach to their analysis and action to reduce their rates. Methods Setting and patients The study was undertaken at Wimmera Base Hospital in Horsham, 300 km northwest of Melbourne, Victoria. The hospital provides services to 43 000 people in the Wimmera region, including 13 500 in Horsham. Eight specialists and 14 general practitioners live in the town. With the assistance of eight hospital medical officers, they treat about 6000 inpatients and 9000 emergency department patients annually. Another 14 specialists visit the town regularly to treat patients at the hospital. The components of the risk management program (Box 1) evolved over time. In 1989, the hospital medical staff chose four doctors to be medical reviewers and to form a surveillance committee. This committee was expanded to include a nurse in 1995 and a clinical risk manager in 1997. Detection of adverse events Inpatient medical record review: The medical records of all patients admitted to the hospital between July 1991 and September 1999 were reviewed shortly after discharge, using a process described previously.3 Briefly, each medical record was screened by medical records staff using eight general patient outcome criteria (Box 2). Records with at least one of these criteria were sent to one of the four nominated doctors to determine if an adverse event was present. This was defined as "an untoward patient event which, under optimal conditions, is not a consequence of the patient's disease or treatment".1 The reviewer independently completed an adverse event analysis form for discussion at bimonthly meetings of the surveillance committee. Recommendations for action relating to patient care were made by the committee and forwarded to medical and nursing staff groups in the hospital. When a clinical risk manager was appointed, the adverse event analysis forms were first forwarded to the manager to determine if immediate action was required. Emergency department medical record review: The medical records of all patients who attended the emergency department between October 1997 and September 1999 were reviewed. An administrative database of all inpatient admissions and emergency department attendances was screened for five general patient outcome criteria (Box 2), using software designed for the study. Attendances that screened positive were reviewed by the hospital's clinical risk manager and, if an adverse event was detected, by the director of medical services. If an event was confirmed, it was further analysed and recommendations to prevent its recurrence were made to relevant hospital staff, using the same committee review process as used for inpatient records. Clinical incident reporting: A clinical incident reporting system was developed in 1997 by the Australian Patient Safety Foundation, an independent organisation in Adelaide that promotes patient safety. Hospital staff in Horsham were educated about the system and encouraged to report clinical incidents and "near-misses". A clinical incident was defined as "any event that has caused harm, or has the potential to harm, a patient, visitor or staff member, or any event which involves malfunction, damage or loss of equipment or property, and any event which might lead to a complaint".5 Incident reporting forms developed by the Foundation were placed in all departments. These forms comprise two parts: the first provides details of actual or potential clinical incidents, while the second allows the incident to be reported anonymously to a national database. Staff members reporting incidents could identify themselves or remain anonymous. Completed forms for incidents reported between October 1997 and September 1999 were sent to the clinical risk manager for local analysis and were also reported to the national database. General practitioner reporting: As adverse events related to inpatient care may occur or be recognised after patient discharge from hospital, an adverse event reporting form was included in the inpatient summary routinely sent to each patient's general practitioner (GP). GPs were asked to attach the form to the patient's medical record for a month and to complete and return the form if they detected an adverse event. External sources: Some adverse events occur rarely in individual hospitals. Details about serious but infrequent adverse events at other hospitals were obtained from coronial and consultative committee reports, insurers, medical indemnity organisations, medical and nursing journals and the media. If the surveillance committee thought the event could occur locally, action was taken to reduce the risk. Patient satisfaction: Patient perspective on adverse events was sought through patient satisfaction surveys, focus groups and patient complaints. Satisfaction surveys were posted to every 10th patient who attended the emergency department or was admitted to the hospital. Event analysis and action When an adverse event was detected, its likelihood and consequences were estimated in accordance with the Australia/ New Zealand Risk Management Standard6 (Box 3). Events were ranked according to their risk severity (risk severity = consequence score x likelihood score) (Box 4). Events with high risk severity were given priority for analysis, and action was taken to reduce the risk, as described previously7 (Box 5). For adverse events with low risk priority, the surveillance committee decided whether to take action or to accept the risk and continue monitoring for that event. All data from the inpatient adverse event analysis forms were entered into a database program developed from the Clipper database compiler software package.8 Data from emergency department and GP reports were entered into access databases,9 and clinical incidents were entered into the Australian Patient Safety Foundation's database.4 Statistical analysis The χ2 test was used for categorical comparisons of data. A P value < 0.05 was considered to indicate statistical significance; all tests were two-tailed. Statistical analyses were performed using the statistical package GraphPad Instat.10 Confidence intervals were calculated using standard methods.11 Results \ Inpatient medical record review A total of 49 834 inpatients were discharged from the hospital between July 1991 and September 1999. The medical records of 4199 (8.43%) screened positive for one or more of eight general patient outcomes, and 386 (0.77%) contained an adverse event. These events were analysed and action was taken to reduce the probability of recurrence. The annual rate of adverse events decreased between the first and eighth years of the study from 1.35% of all patients discharged (69 events) to 0.74% (49 events) (χ2= 31.31; df = 7; P < 0.001). This trend was linear (χ2=11.52; df = 1; P < 0.001) (Box 6). The absolute risk reduction was 0.61% (95% CI, 0.23%-0.99%), and the relative risk reduction was 44.9% (95% CI, 16.9%-72.9%). Emergency department medical record review A total of 20 050 patients attended the emergency department between October 1997 and September 1999. The medical records of 544 screened positive for one or more of five general patient outcomes (2.71% of all patient attendances), and 250 (1.24%) contained an adverse event. Action was taken to reduce the probability of recurrence. The quarterly rate of adverse events decreased between the first and eighth quarters of monitoring from 3.26% of all attendances (84 events) to 0.48% (12 events) (χ2= 120.43; df = 7; P < 0.001). The trend was linear (χ2= 87.64; df = 1; P < 0.001) (Box 6). The absolute risk reduction was 2.78% (95% CI, 2.04%-3.52%), and the relative risk reduction was 85.3% (95% CI, 62.7%-100%). Clinical incident reporting Between October 1997 and September 1999, hospital staff completed 621 clinical incident forms, and 66 adverse events were found. The most common reported incidents were patient falls (280 incidents, 45% of all reported incidents) and medication errors (93; 15%). In response to the number of falls, a falls risk assessment tool was developed. Each patient over 65 years of age underwent a falls risk assessment on admission to hospital. Subsequently, the number of patient falls resulting in fractures while in hospital decreased. The 66 events detected by clinical incident reporting made up 16.3% of the total of 405 adverse events detected between October 1997 and September 1999 by clinical incident reporting and medical record review; 250 events (61.7%) were detected by emergency department medical record review and 89 (22.0%) by inpatient medical record review. Four adverse events were detected by more than one method; these were associated with failure of equipment in the operating room and an inpatient fall. For analysis, these four events were allocated to the first method by which they were detected. General practitioner reporting Between January and September 1999, 21 reports were made by general practitioners (0.25% of patients discharged). An adverse event was identified in 16 (76% of all reports). Events included discharge medication errors, postoperative wound infections and other surgical complications. Three patients required readmission. External sources Information about 12 adverse events at other hospitals in Victoria was obtained from the media and coronial reports. After analysis, preventive administrative and clinical changes were implemented in Horsham. These included the introduction of an organisation-wide policy on equipment service contracts, installation of thermostatic mixing valves in the hot water system to reduce burns, and development of an intercostal catheter insertion policy. Patient satisfaction Of the 69 formal complaints received by the hospital between October 1997 and September 1999, 11 (16%) related to clinical care. On review, four were associated with an adverse event. Complaints were analysed in the same way as adverse events from other sources. Discussion Our study has demonstrated that adverse events in hospitals can be detected using medical record review, clinical incident reporting and other methods. The rate of adverse events can then be reduced using a systems approach to event analysis, followed by appropriate action and continued monitoring for adverse events to evaluate the effectiveness of the action. As we found that few individual events were identified by more than one detection method, the use of multiple detection methods increased the total number of events identified. Scoring the risk associated with each event, using its likelihood and consequences, allowed events found by different methods to be ranked and prioritised for action to reduce risk. This allowed available resources to be directed to events with the greatest patient risk. To our knowledge, this is the first comparative study over time of a clinical risk management program that used diverse methods to detect adverse events and reduce their rate. Other studies have used a single detection method to measure adverse event rate at one point in time. For example, inpatient medical record review was used in the multihospital Harvard Medical Practice Study1 and the Quality in Australian Health Care Study.2 These studies detected adverse event rates of 3.7%1 and 16.6%2 of hospital admissions, respectively. Both used 18 screening criteria, including some that required clinical judgement, and neither measured the rate of adverse events over time after intervention. Our study found a lower rate of inpatient adverse events, but used only eight screening criteria, none requiring clinical judgement. We are not aware of any comparative studies that measured the adverse event rate in hospital emergency departments. The rate of adverse events found using clinical incident reporting depends on the rate of reporting. For example, an increase in reports of medication errors may reflect an increase in errors, the rate of reporting, or both. Reporting rates vary greatly between hospitals, making meaningful comparisons difficult. In a study in the United States comparing adverse events reported by resident medical staff and those found by medical record review, 30.8% of adverse events were found by both methods.12 This is a much greater proportion than the 0.9% of events found by both methods in our study. However, in our study, most events were reported by nursing staff, with medical staff reporting few events. Although we used multiple methods to detect adverse events, we did not find all adverse events that occurred in the hospital. Although we could have found more by using more screening criteria, our use of five to eight screening criteria for medical record review, rather than the 18 used in some other studies,1,2 meant that this would have required more resources. The study did not include a control hospital where adverse events were detected but no analysis was performed nor action taken. Also, adverse event rates were not adjusted for patient severity. Therefore, other factors may explain the reduction in adverse events. In our study, resource limitations meant that medical records identified by screening were reviewed by a single reviewer. In the Harvard Medical Practice Study and the Quality in Australian Health Care Study, each medical record identified by screening was reviewed by two reviewers and, if they disagreed, by a third. The resources for such intensive review of records are unlikely to be available in most hospitals. The strengths of our study included its eight-year duration and prospective nature. By keeping the method constant (eg, number and types of screening criteria and three of the four reviewers), the rate of adverse events could be meaningfully compared over a long period, and the effects of actions assessed. Further research is required to improve methods of detecting adverse events. A greater proportion of events might be detected with more effective and efficient screening criteria. A potentially effective criterion would be a code for external cause of injury13 to be assigned by medical records staff when coding diagnoses. Another potential screening device is clinical pathways that can detect adverse events by analysing deviations from the pathways. The development of electronic records may also assist in detecting adverse events. In addition, national databases of adverse events reported as clinical incidents or from coronial inquests would provide further information to reduce risk. Actions that are effective in changing clinical behaviour have been discussed in detail previously.7 Although some effective strategies are available, changing health delivery systems and clinical behaviour is frequently complex and difficult, and many strategies in use are not effective.14 More research is required to develop additional effective strategies. The components of the risk management program used in our study could be applied in hospitals of varying sizes. Hospitals or individual departments can decide which detection methods are appropriate for their services and available resources. For example, medical record screening does not need to use all outcome criteria, and further program components can be added over time. Finally, we believe that the risk management program has allowed our patients to receive better care with fewer adverse events and has been an effective use of resources. Acknowledgements We wish to thank the staff of the Wimmera Health Care Group for their enthusiastic participation in the program, especially Mrs Cathy Dooling, Manager Health Information Services, and staff, and previous members of the surveillance committee. This program was partly funded by a grant from the Victorian Department of Human Services. References Brennan TA, Leape LL, Laird NM, et al. Incidence of adverse events and negligence in hospitalised patients: results of the Harvard Medical Practice Study 1. N Engl J Med 1991; 324: 370-376. Wilson RM, Runciman WB, Gibberd RW, et al. The Quality in Australian Health Care Study. Med J Aust 1995; 163: 458-471. Wolff AM. Limited adverse event screening: using record review to reduce hospital adverse patient events. Med J Aust 1996; 164: 458-461. <eMJA full text> Australian Patient Safety Foundation. What is incident monitoring? The Australian Incident Monitoring Study. Adelaide: The Foundation, 1997. Leape LL. Error in medicine. JAMA 1994; 272: 1851-1857. Standards Australia. Australian/New Zealand Standard 43:60. Sydney: Standards Association of Australia, 1999. Wolff AM, Bourke J. Reducing medical errors: a practical guide. Med J Aust 2000; 173: 247-251. Clipper. Version 5.0. Los Angeles, Calif: Nantucket Corporation, 1990. Microsoft access. Relational database management system for windows. Version 7.0. Seattle, Wash: Microsoft Corporation, 1999. GraphPad Instat. Version 2.0. San Diego, Calif: GraphPad Software, 1992. Sackett DL, Haynes RB, Guyatt GH, Tugwell P. Clinical epidemiology: a basic science for clinical medicine. 2nd ed. Boston: Little Brown and Co, 1991: 218. O'Neil AC, Peterson LA, Cook EF, et al. Physician reporting compared with medical-record review to identify adverse medical events. Ann Intern Med 1993; 119: 370-376. O'Hara DA, Carson NJ. Reporting of adverse events in hospitals in Victoria, 1994-1995. Med J Aust 1997; 166: 460-463. Oxman AD, Thomson MA, Davis DA, Hayes RB. No magic bullets: a systematic review of 102 trials of interventions to improve professional practice. CMAJ 1995; 153: 1423-1431. (Received 23 Oct 2000, accepted 6 Feb 2001) Authors' details Clinical Risk Management Unit, Wimmera Health Care Group, Horsham, VIC. Alan M Wolff, FRACGP, MBA, Director of Medical Services, and Director of Accident and Emergency Department; Jo Bourke, RN, GradDipCM, Clinical Risk Manager; Ian A Campbell, FRACS, Visiting General Surgeon; David W Leembruggen, FRACGP, Visiting General Practitioner, and Director of Postgraduate Education. Reprints: Dr A M Wolff, Medical Administration, Wimmera Health Care Group, Baillie Street, Horsham, VIC 3400. whcgmedATnetconnect.com.au Make a comment Back to text 2: General outcome criteria used for screening medical records Inpatient criteria Death Return to operating theatre within 7 days Transfer from general ward to intensive care Unplanned readmission within 21 days of discharge Cardiac arrest Transfer to another acute care facility Length of stay greater than 21 days Booked for theatre and cancelled Emergency department criteria Death Unplanned re-presentation to department within 48 hours for same condition Length of stay greater than 6 hours Transfer to another acute care facility Presentation to department for same condition within 28 days of hospital inpatient discharge Back to text 3: Qualitative measures used to determine risk severity of adverse events (modified from Australian/New Zealand Standard 43:606) Measures of consequence or impact 1 Insignificant No injuries, low financial loss 2 Minor Minor treatment required, no increase in length of stay or readmission, minor financial loss 3 Moderate Major temporary injury, increased length of stay or readmission, medium financial loss 4 Major Major permanent injury, increased length of stay or readmission, major financial loss 5 Catastrophic Death, huge financial loss or threat to goodwill Measures of likelihood 1 Rare May occur only in exceptional circumstances 2 Unlikely Could occur at some time 3 Possible Might occur at some time 4 Likely Will probably occur in most circumstances 5 Almost certain Is expected to occur in most circumstances Back to text 4: Examples of risk severity scores Adverse event Source Consequence score (C)* Likelihood score (L)* Risk severity score (C x L) Missed diagnosis (abdominal pain, fracture, myocardial infarction) Emergency Department record review 3 3 9 Drug administration errors Clinical incident reporting 2 3 6 Postoperative wound infection Inpatient record review 3 2 6 Failure to admit when indicated Emergency Department record review 3 2 6 * Scores are defined in Box 2. Back to text 5: Actions taken to reduce the frequency of adverse events Changes to clinical and administrative protocols Focused audits to investigate specific adverse events Discussion with staff involved Education (including presentation of adverse events at postgraduate education meetings and clinical risk management presentations) Creation of worksheets containing details of clinical policy, space to write clinical notes and a patient management checklist Developing checklists for complex procedures Increasing the supervision of junior hospital medical officers Introduction of patient risk assessment tools to determine risk of falling, developing a pressure ulcer or thromboembolus and difficulty with discharge home Regular feedback to clinical staff about adverse events and the results of actions taken to reduce risk Back to text Back to text

Alan M Wolff · Jo Bourke · Ian A Campbell · David W Leembruggen

CLASS="LinkBox"> Research Prevalence of overweight and obesity in Australian children and adolescents: reassessment of 1985 and 1995 data against new standard international definitions Anthea M Magarey, Lynne A Daniels and T John C Boulton MJA 2001; 174: 561-564 For editorial comment, see Baur; see also Eckersley Abstract - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - More articles on Paediatrics Abstract Objective: To review the prevalence of overweight and ...

Anthea M Magarey · Lynne A Daniels

Henry Krum, Andrew M Tonkin, Robert Currie, Robert Djundjek and Colin I Johnston MJA 2001; 174: 439-444 For editorial comment, see Horowitz & Stewart; see also Krum Abstract - Methods - Results - Discussion - Acknowledgements - Reference - Authors' details - - More articles on Cardiology and cardiac surgery Abstract Objectives: To investigate the frequency and general practitioner awareness of patients with chronic heart failure (CHF), and to evaluate a cardiac algorithm and document cardiac investigations performed in establishing this diagnosis. Design and setting: Between March and August 1998, consecutive patients aged 60 years and older presenting to their GP were assessed. In patients previously diagnosed with CHF, aetiology and diagnostic assessments were documented. In patients with suspected CHF (by a standardised algorithm, based on World Health Organization guidelines), further investigations and GP diagnosis were recorded. Patients: 80 consecutive patients were assessed by each of 341 GPs throughout Australia, reflecting the Australian metropolitan/rural population mix of 1996. This provided a total of 22 060 evaluable patients. Main outcome measures: Estimated numbers of patients with CHF in general practice (previously and newly diagnosed); major aetiological factors; use of ancillary diagnostic tests; drugs prescribed. Results: CHF was diagnosed in 2905 of 22 060 patients (13.2%) (2485 previously diagnosed and 420 newly diagnosed). Major aetiological factors were ischaemic heart disease and hypertension. Echocardiography had been performed in 64% of previously diagnosed patients, but was performed in only 22% of possible CHF patients. Angiotensin-converting enzyme (ACE) inhibitors were prescribed in 58.1% of patients with CHF. Patients with evidence of left ventricular dysfunction were more likely to have received ACE inhibitors. Conclusions: CHF appears to be very common in the elderly, based on GP diagnosis of the condition. Of 100 patients aged 60 years and over presenting to their GP, two new cases of CHF will be detected using a simple clinical algorithm in conjunction with appropriate diagnostic tests. ACE inhibitors appear to be underutilised. Chronic heart failure (CHF) is a debilitating condition with high morbidity and mortality, and is a major public health burden. Its prevalence is increasing,1 despite a reduction in age-standardised mortality associated with cardiovascular diseases such as myocardial infarction and stroke.2 Factors implicated in this increased prevalence include the ageing of the population, decreased mortality rates following myocardial infarction, and more frequent diagnosis of CHF after investigations such as echocardiography.3The epidemiology of CHF in Australia has been assumed to be similar to that in the United States,4-6 the United Kingdom7-10 and Europe.11,12 However, substantive data have been lacking, and the approach taken to diagnosis of patients with suspected CHF in general practice in Australia is also unknown. Similarly, although international studies have suggested marked underutilisation of angiotensin-converting enzyme (ACE) inhibitors and β-blockers,4-11,13-17 no evaluation of use of drug therapies for CHF in Australia has been reported. Accordingly, the aims of the Cardiac Awareness Survey and Evaluation (CASE) Study were: to investigate the frequency, awareness, and aetiology of heart failure in general practice in Australia; to document cardiac investigations used by general practitioners in establishing the diagnosis of CHF; and to determine prescribing patterns in the treatment of CHF by Australian GPs. As the prevalence of heart failure increases steeply with age, the study focused on people aged 60 years and older. Methods Recruitment into the CASE study GPs were recruited solely on the basis of interest in participating in the study. Interest was first ascertained by the local pharmaceutical representative of the study sponsor (see Acknowledgements). The CASE steering committee then sent interested GPs a formal letter of invitation to participate in the study. GPs agreeing to participate attended a local education and information session. These sessions were spread across all Australian States and Territories with a mix of metropolitan, rural and remote regions in an effort to recruit a sample of GPs (and thus patients) representative of their distribution. Each GP was asked to assess 80 consecutive patients aged 60 years or older for the possibility of heart failure. GPs were recruited from March 1998, data were collected prospectively and the study was completed in August 1998. Assessment for CHF New patients: Patients not previously diagnosed as having CHF were assessed for that possibility using modified World Health Organization criteria (Box 1).18 Alternative conditions that may have contributed to these symptoms and signs were recorded. In patients suspected of having CHF based on the above criteria, further investigations (chest x-ray [CXR], electrocardiogram [ECG], and echocardiogram) were suggested (but not mandated). GPs were also asked to note any investigations that had been performed in the previous 12 months. For patients who had an echocardiogram, the GP was asked to indicate whether there was evidence of systolic or diastolic ventricular dysfunction (or both) from the echocardiogram report. At the conclusion of this process, the GPs assessed whether they thought the patient had CHF. Previously diagnosed patients: For patients who had previously been diagnosed as having CHF by their GP, records were retrospectively analysed for clinical and diagnostic criteria that contributed to that diagnosis. These included use of ECG, CXR and echocardiography, hospital admission for heart failure, and specialist referral for CHF. Pharmacotherapy For patients with previously diagnosed CHF, GPs were asked to document current drug therapy specifically prescribed for this condition (name of drug, daily dose, and frequency of administration). For patients with newly diagnosed CHF, GPs were asked whether they instituted pharmacotherapy immediately and what that pharmacotherapy comprised. A dosage equivalence table of commonly prescribed ACE inhibitors was compiled, and prescribing was divided into low, medium and high doses. To determine prescribing according to decade of life, prescribing was assessed in patients aged 60-69 years (n = 569), 70-79 years (n = 1360), and 80 years and older (n = 976). GP prescribing in patients with echocardiographic evidence of systolic or diastolic left ventricular dysfunction was specifically determined. Initial pharmacotherapy prescribed for patients diagnosed with CHF as part of the CASE study was evaluated. Because of the cross-sectional nature of the assessment in CASE, most of these patients did not have the opportunity to be up-titrated to target doses of drugs. For this reason, newly diagnosed patients were not included in the analysis of dose of ACE inhibitor prescribed. Results Of 523 GPs who originally expressed interest in participating, 341 completed the study. Their geographical distribution (78% metropolitan, 22% rural or remote) was similar to that observed for all Australian GPs (77% metropolitan, 23% rural or remote). In all, 23 845 patients were entered into the study. Of these, 1785 were excluded from analysis because of patient refusal or missing data, leaving 22 060 who made up the baseline population. Baseline demographics The distribution of the CASE patient population by area (capital city, 58.7%; metropolitan, 10.7%; rural, 21.4%; remote, 1.2%; 8% unclassified) was similar to the 1996 Australian population aged 60 years or older.19 However, there were fewer rural patients among the CASE cohort than in the census population. The total CASE study population comprised 45% men and 55% women; 8612 (39%) were aged 60-69 years (48% men), 9371 (42.5%) were aged 70-79 years (45% men) and 4077 (18.5%) were aged 80 years or older (39% men). Patients not previously diagnosed with CHF Box 2 shows the assessment process and results for the 22 060 patients. In the 4807 patients assessed as having possible CHF, at least one further investigation was performed in 2903 (60.4%). To determine whether the CASE audit itself may have prompted further investigation of these patients, the tests were divided into those ordered within the previous 12 months and those ordered subsequent to the CASE audit. Investigations ordered within the 12 months before the CASE audit were ECG in 1953 (40.6%), CXR in 1894 (39.4%), and echocardiogram in 366 patients (7.6%). Investigations subsequent to the CASE clinical assessment were performed in 488 patients who had ECGs (10.2%), 493 who had CXR (10.2%), and 466 who had echocardiograms (9.7%). The diagnosis of CHF was based on symptoms in 73%, signs in 66%, causative factors in 61%, and investigations in 49% (not mutually exclusive). Presence of at least one symptom and one causative factor had the highest sensitivity for detection of new heart failure (323 of the 420 [76.9%] new cases of heart failure). The sensitivity of the other diagnostic groupings for possible heart failure were > 2 symptoms, 68.6%; > 2 signs, 47.3%; > 1 symptom and > 1 sign, 68.6%; > 1 sign and > 1 causative factor, 66.9%. Possible false negative diagnoses: Of the 4807 patients who met diagnostic criteria for suspected CHF, 466 underwent echocardiography after the CASE audit (when systolic, diastolic or no dysfunction was specifically noted). Of these 466 patients, 108 had an echocardiographic report of left ventricular systolic or diastolic dysfunction. However, despite this objective evidence of ventricular dysfunction, GPs diagnosed CHF for only 77 of these 108 patients. Possible false positive diagnoses: Of the 420 patients newly diagnosed as having CHF, 162 underwent echocardiography. Nineteen of these 162 patients (11.7%) had no evidence of left ventricular systolic or diastolic dysfunction on this test, yet were still classified as having CHF by their GP. Patients previously diagnosed as having CHF CHF had been previously diagnosed in 2485 of the baseline population of 22 060 (11.3%). Both electrocardiograms and CXRs had been performed in 96% of the patients previously diagnosed as having heart failure, and 64% had had echocardiography performed. For these patients, in the previous 12 months: 1640 patients (66%) had been referred to a specialist; 1744 patients (70%) had either not been admitted to hospital with CHF or their hospitalisation status was unknown; and of the 741 patients admitted for CHF, 459 (62%) had one admission, 165 (22%) had two admissions, and 58 (8%) had three admissions. Six patients (1%) had been admitted 10 or more times for CHF. CHF patients in the CASE study At the end of the study, 2905 of the 22 060 baseline population (13.2%) were considered to have CHF: 2485 (11.2%) with a previous diagnosis, and 420 (1.9%) with a new diagnosis. The prevalence of CHF in these patients was closely related to age group (Box 3). The cardiovascular diagnoses that may be contributing to CHF in these patients are summarised in Box 4. Hypertension and ischaemic heart disease were major comorbidities and potential aetiological factors in both the new and previously diagnosed cohorts. Pharmacotherapy Specific CHF pharmacotherapy for the 2905 patients with CHF is summarised in Box 5. Most patients were receiving diuretics and ACE inhibitors. Alternatives to ACE inhibitors in patients who can not tolerate this medication include nitrates and hydralazine (prescribed for 0.6%) and/or angiotensin II receptor antagonists (prescribed for 4.3%). β-Blockers were prescribed for 12% of patients. Less than 50% of β-blocker prescribing was of the non-selective, vasodilating β-blocker carvedilol, approved for CHF in Australia. The dose of ACE inhibitor was determined in patients previously diagnosed as having CHF (Box 6). Based on our dosage equivalence table, prescribed doses were low in 60%, medium in 31%, and high in 9% of patients. Prescribing according to echocardiographic findings: Pharmacotherapy prescribed for CHF according to left ventricular (LV) systolic (found in 932 patients) or diastolic (found in 376 patients) dysfunction is summarised in Box 5. The presence of LV dysfunction on echocardiography resulted in higher prescribing of ACE inhibitors than in the overall patient cohort. However, the frequency of ACE inhibitor prescribing was not different between the systolic and diastolic LV dysfunction groups. Furthermore, there were very few differences among other agents in these groups according to systolic or diastolic LV dysfunction. Prescribing according to decade of life: Prescribing of drug therapy specifically for CHF according to decade of life is summarised in Box 5. ACE inhibitor prescribing was unaltered in the very elderly (≥ 80 years). Prescribing of β-blocker demonstrated an age-dependent decrease, whereas prescribing of digoxin and diuretics (thiazide and loop) demonstrated an age-dependent increase. Prescribing in patients with newly diagnosed CHF: Use of pharmacotherapy among newly diagnosed CHF patients was less than that observed in patients with previously diagnosed CHF. In particular, only 51% of newly diagnosed patients were prescribed ACE inhibitors, 37% diuretics, 17% received calcium-channel blockers, 10% digoxin and 8% β-blockers. Discussion We undertook a clinical algorithm approach to the possible diagnosis of CHF. By identifying possible CHF based on the grouping of symptoms, signs and causative factors, we demonstrated an 8.7% rate of identification of CHF (420 of 4807 patients). Although this approach resulted in a relatively low rate of successful diagnosis of CHF, the algorithm used was entirely clinical with a very simple screening process. Our results suggest that, of 100 patients aged 60 years or older presenting to a GP, two will have previously undetected CHF that can be simply diagnosed by attention to clinical symptoms and signs in conjunction with appropriate diagnostic tests. Prevalence of CHF We found somewhat higher rates of CHF patients (13.2%) than in earlier general practice based studies.4,7-9 Specifically, reported prevalence of CHF among patients aged 65 years or older in UK general practice ranged from 2.8%9 to 8.0%.7 The higher rate in the CASE study may reflect the differing methods by which the diagnosis of CHF was made (eg, patient file review,4,9 morbidity registry8), greater use of objective testing (ie, echocardiography in our patient cohort), or our study being a more representative sample of the true CHF population than previous geographically restricted studies performed in the US or UK. In contrast to the above assessments of CHF frequency, population studies where echocardiographic ventricular dysfunction was the main criterion for diagnosis detected fewer CHF patients within these age groups than in the CASE study.10,12 Numbers of CHF patients in the CASE Study increased dramatically with each decade of life — more than 20% of patients aged 80 years or older were diagnosed with CHF. Given the ageing of the population, these findings have important implications for resource allocation. Aetiology The aetiology of CHF was as expected, with a major contribution from ischaemic heart disease and previous myocardial infarction, as well as hypertension. Hypertension was a major contributor to CHF in the Framingham study,20,21 but less so in analysis of the Studies of Left Ventricular Dysfunction (SOLVD)22 and other, more recent data.23 The major contribution of hypertension in the CASE cohort may reflect the advanced age of the population studied, in which hypertension is a frequent comorbidity. Investigations Use of ECG and CXR was high, and echocardiography was used in more than half the patients. The lower use of echocardiography (despite recommendations by major organisations such as WHO)18 may reflect lack of full knowledge of the sensitivity and specificity of this diagnostic test, concerns regarding expense, and difficulty with access. Pharmacotherapy Use of ACE inhibitors: ACE inhibitor prescribing by Australian GPs ranges from 51%-71% of CHF patients, depending on the specific population studied. Prescribing of ACE inhibitor was more likely in patients in whom ventricular dysfunction had been objectively documented. These findings are consistent with international studies, in which prescribing of ACE inhibitors ranges from 10% to 60%.5,6,8-10,14-17 ACE inhibitors reduce morbidity and mortality across the entire spectrum of CHF severity, including in patients with asymptomatic systolic left ventricular dysfunction.24-26 Therefore, all patients with systolic left ventricular dysfunction should be receiving ACE inhibitors unless contraindicated or intolerant. Lack of compliance with these prescribing recommendations may relate to contraindications to ACE inhibitor therapy (ie, bilateral renal artery stenosis) or observed side effects such as hyperkalaemia or cough. Furthermore, ACE inhibitors are not of proven benefit in patients with diastolic CHF. Definitive diastolic dysfunction on echocardiography comprised only a small percentage in the CASE study, although the true percentage is undoubtedly considerably higher. Potential alternatives to ACE inhibitors in patients who are ACE intolerant or have contraindications include angiotensin II receptor antagonists and hydralazine (the latter as part of the hydralazine/nitrate combination). However, only 4.3% of patients were taking angiotensin II receptor antagonists and 0.6% were taking hydralazine. Our findings suggest that, despite the definitive data supporting the use of ACE inhibitors in CHF, these agents are still being underutilised. Dose of ACE inhibitors: Submaximal doses of ACE inhibitors were prescribed for those patients who are taking these drugs. The major clinical trials conducted in patients with CHF (SOLVD,24 CONSENSUS25) and LV dysfunction post-MI (SAVE27) used much higher doses (150 mg of captopril or 20-40 mg of enalapril) than the median and mean doses prescribed in this study. Doses prescribed by Australian GPs were generally lower than both the target and achieved ACE inhibitor doses used in these major trials. There are a number of reasons why the recommended target doses may not be achieved in general practice. First, patients may not tolerate the highest dosage because of hypotension, particularly if up-titration is rapid. Second, because these agents improve symptomatology, patients may become asymptomatic at lower doses of drug and the need to go to higher doses not be entertained in an asymptomatic patient. Finally, until recently, there has been no clear evidence that higher doses offer substantial clinical benefits over and above the use of ACE inhibitors at lower doses. The ATLAS study28 demonstrated a reduction in the combined endpoint of death/heart failure related hospitalisation with lisinopril 32.5-35 mg daily compared with lisinopril 2.5-5 mg daily. Mortality alone was reduced by 8% in the high-dose lisinopril subgroup. Although this reduction is modest, it does suggest that an attempt should be made to maximise ACE inhibitor dosage in every patient. β-Blockers: Prescribing of β-blockers was low, despite overwhelming evidence supporting the benefits of these agents in patients with New York Heart Association (NYHA) Class II-III symptoms.29-31 However, much of this evidence has only been published subsequent to the completion of the CASE study.30,31 Prescribing of the β-blocker vasodilator carvedilol occurred in fewer than half the patients receiving β-blockers. As carvedilol is the only β-blocker approved in Australia for CHF, this suggests that much of the prescribing of β-blockers for the CASE cohort was for indications other than CHF (eg, ischaemic heart disease and hypertension). As GPs in Australia are not permitted to prescribe carvedilol, our observed rate of use of β-blockers reflects specialist prescribing. Factors affecting prescribing: In patients with a definitive diagnosis of ventricular dysfunction by echocardiography, prescribing of ACE inhibitor therapy was higher than for the overall CASE CHF cohort. This may reflect more confidence with the cause of patient symptomatology as being related to CHF. Alternatively, it may be that a GP who is more likely to perform echocardiography to diagnose CHF is also more likely to prescribe best-practice pharmacotherapy. It was also noteworthy that the difference in overall prescribing for systolic versus diastolic dysfunction in these patients appeared similar, although many of the agents (eg, ACE inhibitors) are not of proven benefit in diastolic CHF. Furthermore, some drugs, such as non-dihydropyridine calcium-channel blockers, are relatively contraindicated in patients with systolic heart failure, yet prescribing rates were similar for the entire patient cohort. Conversely, these agents may be of particular benefit in diastolic heart failure, but again prescribing rates appeared similar to the entire patient cohort. There was no reduction in prescribing of ACE inhibitor with each decade of life, suggesting that GPs supported the use of these agents in CHF management in the very elderly (≥ 80 years). Very few data exist to support the use of ACE inhibitors in this group of patients, although studies are currently being conducted. In contrast, β-blocker use declined with each decade of life, suggesting less comfort in prescribing these agents for older patients. Increased digoxin and diuretic use with advanced age may reflect the need to increasingly prescribe these agents for comorbidities such as atrial fibrillation and oedema of other causes. Study limitations The CASE study had a number of potentially significant limitations. Selection of GPs was not random, but was based on interest in undertaking the study. This could introduce significant bias, and therefore we have not classified our evaluation as a prevalence or incidence study. Nevertheless, every effort was undertaken to ensure a representative distribution of general practices according to State, regional area and metropolitan versus rural practice. The study has also demonstrated the difficulty in making a clinical diagnosis of CHF. Diagnosis of CHF was left to the clinical judgement and decision of the GP. We noted a significant false positive and negative rate using documented left ventricular dysfunction on echocardiogram as the "gold standard" of CHF in conjunction with relevant signs, symptoms and causative factors. This is a limitation of many surveys of this type, in which the diagnosis is made based on subjective clinical criteria.4,9 A further limitation may have been the algorithm used to assist the GP in making the diagnosis. This clinical algorithm approach has not been used previously in the diagnosis of CHF. Therefore, the possibility exists of patients being wrongly assigned as having CHF using this approach. This is particularly true as echocardiography was not mandated for all patients. Nevertheless, this algorithm did yield an extra two new CHF patients for every 100 patients aged 60 years and older studied in this way. Acknowledgements The CASE study was supported by the National Heart Foundation of Australia and the Royal Australasian College of General Practitioners. The CASE Management Committee wish to thank all 341 GPs who participated in the CASE study and Servier Laboratories, Australia, who provided input into the study design as well as financial and logistical assistance in the conduct of the study. Servier Laboratories were not involved in the analysis of data. In addition, the Committee acknowledges the expert statistical assistance provided by Dr Chris Reid and Mr Stephen Lim (Baker Medical Research Institute, Prahran, VIC). References Bonneaux L, Barendregt J, Meeter K, et al. Estimating clinical morbidity due to ischemic heart disease and congestive heart failure: the future risk of heart failure. Am J Public Health 1994; 84: 20-28. Waters A-M, Bennett S. Mortality from cardiovascular disease in Australia. Cardiovascular Disease Series No. 3. Canberra: Australian Institute of Health and Welfare, 1995. McGovern PG, Pankow JS, Shahar E, et al. Recent trends in acute coronary heart disease. Mortality, morbidity, medical care and risk factors. N Engl J Med 1996; 334: 884-890. Ho KK, Pinsky JL, Kannel WB, Levy D. The epidemiology of heart failure: the Framingham Study. J Am Coll Cardiol 1993; 22: 6A-13A. Gardin JM, Siscovick D, AntonCulver H, et al. Sex, age and disease affect echocardiographic left ventricular mass and systolic function in the free-living elderly: the Cardiovascular Health Study. Circulation 1995; 91: 1739-1748. Lauer MS, Evans JC, Levy D. Prognostic implications of subclinical left ventricular dilatation and systolic dysfunction in men free of overt cardiovascular disease (the Framingham Heart Study). Am J Cardiol 1992; 70: 1180-1184. Mair FS, Crowley TS, Bundred P. Prevalence, aetiology and management of heart failure in general practice. Br J Gen Pract 1996; 46: 77-79. Morbidity statistics from general practice. 4th National Survey, 1991-92. Royal College of General Practitioners, Office of Population Census and Survey and Department of Health and Social Security. London: HMSO, 1995. Parameshwar J, Shackell MM, Richardson A, et al. Prevalence of heart failure in three general practices in west London. Br J Gen Pract 1992; 42: 287-289. McDonagh TA, Morrison CE, Lawrence A, et al. Symptomatic and asymptomatic left-ventricular systolic dysfunction in an urban population. Lancet 1997; 350: 829-833. Ambrosio GB, Riva LM, Casiglia E. Prevalence, clinical features and prognosis of congestive heart failure (CHF) in the elderly. A survey from a population in Veneto region. Acta Cardiologica 1994; 49: 324-325. Mosterd A, Bruijne de MC, Hoes AW, et al. Usefulness of echocardiography in detecting left ventricular systolic dysfunction in population based studies (The Rotterdam Study). Am J Cardiol 1997; 79: 103-104. Hillis GS, Trent RJ, Winton P, et al. Angiotensin-converting enzyme inhibitors in the management of congestive heart failure: are we ignoring the evidence? Q J M 1995; 89: 145-150. Mosterd A, Hoes AW, de Bruijne MC, et al. Prevalence of heart failure and (a) symptomatic left ventricular dysfunction in the general population. The Rotterdam Study. Eur Heart J 2001; in press. Bart BA, Gattis WA, Diem SJ, O'Connor CM. Reasons for underuse of angiotensin-converting enzyme inhibitors in patients with heart failure and left ventricular dysfunction. Am J Cardiol 1997; 79: 1118-1120. Newman J, Ahmed O, Hyngstrom T, et al. Heart failure treatment with angiotensin converting enzyme inhibitors in hospitalized Medicare patients in 10 large states. Arch Int Med 1997; 157: 1103-1108. Stafford RS, Saglam D, Blumenthal D. National patterns of angiotensin converting enzyme inhibitor use in congestive heart failure. Arch Intern Med 1997; 157: 2460-2464. World Health Organization/Council on Geriatric Cardiology Task Force on Heart Failure Education. Concise guide to the management of heart failure. Geneva: WHO/CGC, 1997; 6-9. Available at <http://www.who.int/ncd/cvd/concguid.pdf>. Rural, remote and metropolitan areas classification. 1991 Census Edition. Canberra: AGPS, 1994. McKee PA, Castelli WP, McNamara PM, Kannel WB. The natural history of congestive heart failure: the Framingham Study. N Engl J Med 1971; 285: 1441-1446. Kannel WB, Belanger JA. Epidemiology of heart failure. Am Heart J 1991; 121: 951-956. Bangdiwala SI, Weiner DH, Bourassa MG, et al. Studies of Left Ventricular Dysfunction (SOLVD) Registry: rationale, design, methods and description of baseline characteristics. Am J Cardiol 1992; 70: 347-353. Teerlink JR, Goldhaber SZ, Pfeffer MA. An overview of contemporary etiologies of congestive heart failure. Am Heart J 1991; 121: 1852-1853. The SOLVD Investigators. Effect of Enalapril on survival in patients with reduced left ventricular ejection fraction and congestive heart failure. N Engl J Med 1991; 325: 293-302. The CONSENSUS Trial Study Group. Effects of enalapril on mortality in severe congestive heart failure. Results of the Cooperative North Scandinavian Enalapril Survival Study (CONSENSUS). N Engl J Med 1987; 3161: 1429-1435. Nicklas JM, Pitt B, Timmis G, et al. Effect of enalapril on mortality and the development of heart failure in asymptomatic patients with reduced left ventricular ejection fractions. N Engl J Med 1992; 327: 685-691. Pfeffer MA, Braunwald E, Moye LA, et al. Effect of captopril on mortality and morbidity in patients with left ventricular dysfunction after myocardial infarction. Results of the Survival and Ventricular Enlargement Trial. N Engl J Med 1992; 327: 821-828. Packer M, Poole-Wilson PA, Armstrong PW, et al. Comparative effects of low and high doses of the angiotensin-converting enzyme inhibitor, lisinopril, on morbidity and mortality in chronic heart failure. ATLAS Study Group. Circulation 1999; 100: 2312-2318. Packer M, Bristow MR, Cohn JN, et al. The effect of carvedilol on morbidity and mortality in patients with chronic heart failure. N Engl J Med 1996; 334: 1349-1355. CIBIS II investigators and committees. The cardiac insufficiency bisoprolol study II (CIBIS II): a randomised trial. Lancet 1999; 353: 9-13. Effect of metoprolol CR/XL in chronic heart failure: Metoprolol CR/XL Randomised Intervention Trial in Congestive Heart Failure (MERIT-HF). Lancet 1999; 353: 2001-2007. (Received 24 Dec 1999, accepted 28 Oct 2000) Authors' details Alfred Hospital, Melbourne, VIC. Henry Krum, PhD, FRACP, Associate Professor, Clinical Pharmacology Unit, Department of Epidemiology and Preventive Medicine, and Department of Medicine, Monash University. National Heart Foundation of Australia, Melbourne, VIC. Andrew M Tonkin, MD, FRACP, Director, Health, Medical and Scientific Affairs. North East Valley Division of General Practice, Melbourne, VIC. Robert Currie, MB BS, FRACGP, Director. Servier Laboratories, Melbourne, VIC. Robert Djundjek, BSc, Manager, Scientific Projects. Austin and Repatriation Medical Centre, Melbourne, VIC. Colin I Johnston, MD, FRACP, Professor and Head, Department of Medicine, University of Melbourne. Reprints will not be available from the authors. Correspondence: Professor C I Johnston, Baker Medical Research Institute, Prahran, VIC 3181. Make a comment 1: Modified World Health Organization18 criteria for assessment of possible chronic heart failure Symptoms: Dyspnoea, chronic fatigue, oedema, and exercise intolerance. Signs: Third or fourth heart sounds, heart murmur, cardiomegaly, pulmonary crackles, raised jugular venous pressure, and dependent oedema. Causative factors: Angina, previous myocardial infarction, hypertension, valvular heart disease/rheumatic fever, and cardiomyopathy. Patients were considered to have possible CHF if they had: > 2 symptoms, > 2 signs, > 1 symptom and > 1 sign, or > 1 symptom and > 1 causative factor. Back to text Back to text Back to text 4: Cardiovascular comorbidities that may be contributing to chronic heart failure Percentage of patients New CHF Previous CHF Hypertension 69.1% 63.6% Angina 44.2% 53.4% Previous MI 28.1% 39.3% Valve disease 15.2% 23.0% Cardiomyopathy 5.2% 11.8% MI = myocardial infarction. Back to text 5: Percentage (95% CI) of patients with chronic heart failure prescribed each class of drug, by evidence of left ventricular dysfunction, and by age group Left ventricular dysfunction* All patients (n = 2905) Systolic (n = 932) Diastolic (n = 376) Diuretics (thiazide, loop) 63.3 (62.4-64.2) 60.3 (58.7-61.9) 62.2 (59.7-64.7) ACE inhibitor 58.1 (57.2-59.0) 70.7 (69.2-72.2) 70.5 (68.1-72.9) Digoxin 31.3 (30.4-32.2) 35.0 (33.4-36.6) 33.2 (30.8-35.6) beta-Blocker 11.8 (11.2-12.4) 13.9 (12.8-15.0) 17.8 (15.8-19.8) CCB-DHP 10.1 (9.5-10.7) 11.1 (10.1-12.1) 12.0 (10.3-13.7) CCB-NDHP 10.0 (9.4-10.6) 10.3 (9.3-11.3) 11.2 (9.6-12.8) Aspirin 10.3 (9.7-10.8) 9.3 (8.3-10.3) 9.6 (8.1-11.1) Warfarin 7.7 (7.2-8.2) 8.2 (7.3-9.1) 8.8 (7.3-10.3) Spironolactone 8.1 (7.6-8.6) 6.9 (6.1-7.7) 1.0 (0.5-1.5) Hydralazine 0.6 (0.5-0.7) 0.2 (0.1-0.3) 0.0 (0.0-0.0) AIIA 4.3 (3.9-4.7) 6.4 (5.6-7.2) 7.7 (6.3-9.1) Age group (years) 60-69 (n = 569) 70-79 (n = 1360) ≥80 (n = 976) Diuretics (thiazide, loop) 57.1 (55.0-59.2) 62.1 (60.8-63.4) 68.4 (66.9-69.9) ACE inhibitor 58.7 (56.6-60.8) 58.0 (56.7-59.3) 58.1 (56.5-59.7) Digoxin 24.8 (23.0-26.6) 29.9 (28.7-31.1) 37.2 (35.7-38.7) beta-Blocker 14.1 (12.6-15.6) 13.2 (12.3-14.1) 8.4 (7.5-9.3) CCB-DHP 9.5 (8.3-10.7) 11.0 (10.2-11.8) 9.0 (8.1-9.9) CCB-NDHP 10.2 (8.9-11.5) 11.0 (10.2-11.8) 11.2 (10.2-12.2) Aspirin 10.0 (8.7-11.3) 11.4 (10.5-12.3) 8.9 (8.0-9.8) Warfarin 9.5 (8.3-10.7) 8.8 (8.0-9.6) 5.0 (4.3-5.7) Spironolactone 6.7 (5.7-7.7) 8.8 (8.0-9.6) 8.0 (7.1-8.9) Hydralazine 1.1 (0.7-1.5) 0.6 (0.4-0.8) 0.3 (0.1-0.5) AIIA 3.5 (2.7-4.3) 5.4 (4.8-6.0) 3.4 (2.8-4.0) * According to echocardiography. ACE = angiotensin-converting enzyme. CCB-DHP = calcium-channel blocker - dihydropyridine. CCB-NDHP = calcium-channel blocker - non-dihydropyridine. AIIA = angiotensin II receptor antagonist. Back to text 6: Dosage equivalence table for angiotensin-converting enzyme (ACE) inhibitors, and median and mean doses prescribed for patients with diagnosed chronic heart failure Dosage equivalence (mg) Low Medium High Captopril ≤ 50 ≤ 100 ≤ 150 Enalapril ≤ 10 ≤ 20 ≤ 40 Perindopril ≤ 2 ≤ 4 ≤ 8 Lisinopril ≤ 10 ≤ 20 ≤ 40 Ramipril ≤ 5 ≤ 10 ≤ 20 Fosinopril ≤ 10 ≤ 20 ≤ 40 Trandolopril ≤ 2 ≤ 4 ≤ 8 Quinapril ≤ 10 ≤ 20 ≤ 40 Dose (mg) Media Mean SD n Captopril 50 67.5 44.1 383 Enalapril 10 15.6 12.0 308 Perindopril 4 3.6 2.2 223 Lisinopril 10 12.6 9.7 218 Ramipril 5 5.1 4.3 98 Fosinopril 10 14.3 7.1 81 Trandolopril 1 1.9 1.8 76 Quinapril 10 11.6 9.1 74 Back to text

Henry Krum · Andrew M Tonkin · Robert Currie · Robert Djundjek · Colin I Johnston

Research Iron deficiency in Australian-born children of Arabic background in central Sydney Margaret A Karr, Michael Mira, Garth Alperstein, Samia Labib Boyd H Webster, Ahti T Lammi and Patricia Beal MJA 2001; 174: 165-168 For editorial comment, see Couper and Simmer Abstract - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - More articles on Public and environmental health Abstract Objectives: To determine the prevalence of iron depletion and deficiency, and iron-deficiency anaemia, along with risk factors for iron depletion, in Australian-born children aged 12-36 months of Arabic-speaking background. Design: Community-based survey. Setting: Central Sydney Area Health Service (CSAHS), NSW, April to August, 1997. Participants: All children born at five Sydney hospitals between 1 May 1994 and 30 April 1996, whose mothers gave an Arabic-speaking country of birth and resided in the area served by the CSAHS. Main outcome measures: Full blood count (haemoglobin, mean corpuscular haemoglobin, mean corpuscular volume), plasma ferritin concentration, haemoglobin electrophoresis, potential risk factors for iron depletion. Results: Families of 641 of the 1161 eligible children were able to be contacted, and 403 agreed to testing (response rate, 62.9% among those contacted). Overall, 6% of children had iron-deficiency anaemia, another 9% were iron deficient without anaemia, and 23% were iron depleted. Multiple logistic regression analysis showed three significant independent risk factors for iron depletion: < 37 weeks' gestation (odds ratio [OR], 5.88, P = 0.001); mother resident in Australia for less than the median time of 8.5 years (OR, 1.96, P = 0.016); and daily intake of > 600 mL cows' milk (OR, 3.89, P = < 0.001). Conclusion: Impaired iron status is common among children of Arabic background, and targeted screening is recommended for this group. Numerous studies have documented the adverse health effects of iron deficiency in infants and preschool children, including growth retardation,1,2 gastrointestinal changes,3 impaired immune function,4 impaired behavioural and mental development5,6 and decline in psychomotor development.7,8 In 1992-1994, a study of Sydney children aged 9-62 months found that 1.1% had iron-deficiency anaemia, while 2.8% were iron deficient without anaemia and another 10.5% were iron depleted.9 The prevalence of iron-deficiency anaemia appeared to be higher among children of Arabic-speaking background, but the small number of these children prevented firm conclusions, and the reasons for any difference were not clear. The most important determinants of iron status in infants are growth rate relative to iron endowment at birth, dietary iron content and bioavailability and gastrointestinal blood loss.10 Risk factors for iron deficiency in infancy and childhood include prematurity, low birth weight,11,12 exclusive breastfeeding beyond six months of age,13 introduction of whole cows' milk before 12 months of age,14 and high intake of cows' milk.15 We examined the prevalence of impaired iron status in a large group of Australian-born children of Arabic-speaking background and evaluated their risk factors for iron depletion. Methods The study was a community-based survey undertaken between April and August 1997. Participants Children were identified from the medical records of five Sydney hospitals, which, according to the Midwives Data Base, account for 92% of deliveries to mothers born in an Arabic-speaking country and residing in the area served by the Central Sydney Area Health Service (CSAHS).16 Eligibility criteria were: birthdate between 1 May 1994 and 30 April 1996; mother gave an Arabic-speaking country of birth on admission; postcode of mother's place of residence was in the area served by the CSAHS. Hospitals were asked to exclude stillbirths and neonatal deaths. Contact details were obtained from the medical records. Survey Parents of all eligible children were sent a letter about the study in both Arabic and English. Five days later, they were telephoned to discuss queries and to invite their child's participation. If they agreed, an appointment was made at a convenient early childhood health centre, or a home visit was arranged. Parents were asked to bring the child's Personal Health Record for assessment of birth weight and gestation. Demographic data were obtained using a structured questionnaire administered by an Arabic-speaking research assistant (S L). Questions were also asked about the child's feeding habits since birth and whether the child had had a fever in the two weeks before the blood test, as fever can elevate plasma ferritin concentration.17 Investigations About 0.75 mL of blood was collected by fingerprick and tested at the Royal Alexandra Hospital for Children (RAHC), Sydney, NSW. Haematological investigations (using a Coulter S+IV, Fullerton, Cal, USA) included measurement of haemoglobin and red cell indices. Plasma ferritin concentration was measured by immunoradiometric assay (Biorad, Hercules, Cal, USA). Haemoglobin electrophoresis was performed on all samples to detect haemoglobinopathies. Definitions of impaired iron status are shown in Box 1. All parents were notified of their children's results. Children with poor iron status or haemoglobinopathy were referred to their general practitioners (GPs). A copy of the laboratory report was sent to the GP and to the parents, if they so requested. Statistical analyses Children found to have a haemoglobinopathy were excluded from the analyses, which were performed using Stata (version 5).21 Statistical tests were performed after adjustment for possible cluster effects both within hospitals and within families (as some families contributed more than one child to the study). Confidence intervals were similarly adjusted. Adjusted χ2 tests were used to examine relationships between iron depletion and demographic and risk factors. Variables found to be significantly associated with iron depletion were then entered into a multivariate logistic regression model. Prevalence of impaired iron status was compared with prevalence in children from the general population of central Sydney assessed in 1992-1994.9 Data from that study were re-examined for children aged 12-38 months, using a ferritin level < 10 µg/L to define iron depletion. To test the representativeness of our sample group, demographic characteristics of the mothers were compared with those of all women who in the 1996 census gave an Arabic-speaking country of birth, resided in the area served by the CSAHS, were aged 15-45 years and had children aged 12-38 months. These data were obtained from the Australian Bureau of Statistics. Ethical approval for all components of this study was obtained from the CSAHS Ethics Review Committee. All participating parents gave informed written consent. Results We were able to contact families of 641 of the 1161 eligible children and tested 403 of these children (63% response rate among those able to be contacted). Haematological testing identified a haemoglobinopathy in 21 children, who were therefore excluded from analysis, although two had other haematological parameters consistent with iron depletion. This left 382 children with a plasma ferritin result, and 315 with complete haematological results (blood volume was insufficient for a complete examination in the other 67). Median age of the 382 children at the time of data collection was 25 months (range, 12-38 months). Age distribution was 12-23 months (149 children), 24-35 months (204), and 36-38 months (29). Just over half the children (53%) were male. Prevalence of impaired iron status Prevalence of impaired iron status is shown in Box 2. Overall, 38% of children with an Arabic background had impaired iron status, comprising 6% with iron-deficiency anaemia, a further 9% with iron deficiency without anaemia and a further 23% with iron depletion. The Box also shows prevalences found in 1992-1994 among children the same age in the general population of central Sydney.9 The proportion of children with impaired iron status was substantially higher among Australian-born children of Arabic-speaking background in 1997 than among children of the same age in the general community in 1992-1994. Among the children of Arabic background, those who were reported as having a fever in the two weeks before the blood test were statistically less likely to fulfil the criteria for iron depletion (19/122 versus 68/260; F1,339 = 5.07; P = 0.025). However, they did not differ significantly in rates of iron deficiency (11/108 versus 16/207; F1,282 = 0.49; P = 0.48) or iron-deficiency anaemia (8/108 versus 12/207; F1,282 = 0.31; P = 0.58). The rate of iron depletion among the 260 children reported not to have had a fever in the two weeks before the blood test was 26% (95% CI, 21%-32%). There were no significant differences in iron status between the sexes or between age groups. Risk-factor analysis Potential risk factors among the 382 children are shown in Box 3. Univariate analysis revealed that prematurity, mother resident in Australia less than the median time of 8.5 years, mother born in a country other than Lebanon, and daily intake of more than 600 mL of cows' milk were significantly associated with iron depletion (Box 4). None of the other variables tested, including age of introduction of cows' milk, were significantly associated with iron depletion. Multivariate logistic regression analysis determined that prematurity, mother resident in Australia less than the median time of 8.5 years, and daily intake of more than 600 mL of cows' milk, but not mother born in a country other than Lebanon, were independently associated with iron depletion (Box 4). Children who had been born prematurely were almost six times more likely to be iron depleted, while those who drank more than 600 mL cows' milk per day were almost four times as likely and those whose mothers had been in Australia less than the median time (8.5 years) were almost twice as likely. Representativeness of sample We compared post-secondary education and time in Australia between the sample group and all women who in the 1996 census gave an Arabic-speaking country of birth, resided in the area served by the CSAHS, were aged 15-45 years and had children aged 12-38 months. In the sample group, 29% (116/403) had post-secondary qualifications (95% CI, 24%-33%), compared with 30.2% in the census group (421/1392). Similarly, 30% (121/401) of our sample had been in Australia for six to 10 years (95% CI, 26%-35%), while the corresponding figure for the census group was 26.4% (368/1392). The 238 parents who declined a blood test for their child were questioned by telephone about the age and sex of the child and the volume of cows' milk consumed daily; 180 parents (76%) responded. There were no significant differences between their children and those who had blood tests in age (P = 0.7), sex (P = 0.6) or reported volume of cows' milk consumed daily (P = 0.17). Among the 382 children tested for iron depletion, 67 had moved place of residence since birth and 315 had not moved. The proportion with iron depletion did not differ between these two groups (OR, 1.31; 95% CI, 0.73-2.36). Nor did it differ between the 67 children who had only ferritin level estimated and the 315 who gave sufficient blood for a full haematological examination (OR, 1.03; 95% CI, 0.53-2.01). Discussion These results indicate a public health problem in Australian-born children of Arabic-speaking background in central Sydney that could indicate a nationwide problem. More than a third of these children had impaired iron status, including 6% with iron-deficiency anaemia and another 9% with iron deficiency without anaemia. These prevalences are higher than those found in children the same age in the general population of central Sydney in 1992-1994.9There are several potential sources of bias in this study. The first was the use of retrospective records and consequent failure to contact about 45% of mothers. This is a common problem in such retrospective studies.22,23 However, the mothers of the children studied did not differ significantly from all women in the 1996 census who were aged 15-45 years with children in the target age range, gave an Arabic-speaking country of birth and resided in the CSAHS, while prevalence of iron depletion did not differ between children who had moved residence since birth and those who had not. It is unlikely that our sample differed substantially from the total study population. A second potential source of bias was non-response. However, children whose parents refused a blood test did not differ significantly from those who had a blood test in age, sex and proportion who drank more than 600 mL cows' milk daily. Finally, children whose blood samples were insufficient for full haematological assessment did not differ significantly in prevalence of iron depletion from those who had a full assessment. In the group of children reported to have had a fever in the two weeks before the blood test, a significantly lower proportion fulfilled the criteria for iron depletion. Therefore, the rate of iron depletion reported may be an underestimate. The definition of iron deficiency used in this study was particularly stringent, requiring abnormal values for three laboratory indicators of iron status. Criteria used by the United States Third National Health and Nutrition Examination Survey were less stringent: individuals were diagnosed as iron-deficient if they had abnormal values for two of three laboratory indicators (serum ferritin, free erythrocyte protoporphyrin or transferrin saturation).18 Nevertheless, that survey found rates of iron deficiency and iron-deficiency anaemia among children aged one to two years less than half the rates found in our study (3% versus 6% in our study). The US rate was similar to the rate found in children in the general population of central Sydney in 1992-1994. Risk of iron depletion in our study was greater in children whose mothers had been in Australia for less than the median time of 8.5 years. About 79% of these mothers spoke Arabic, or mainly Arabic, in the home. Early childhood health centres in central Sydney have specific days on which an Arabic interpreter is present, but anecdotal reports suggest that many mothers do not avail themselves of this service. Newly arrived mothers should be targeted in hospital, immediately postpartum, and given information as to which days an Arabic interpreter will be at their local centre and strongly encouraged to attend on a regular basis. Prematurity is well documented as a risk factor for iron deficiency, and this should be kept in mind by GPs and other healthcare providers. Of particular interest is the risk associated with the volume of cows' milk consumed daily. The National Health and Medical Research Council recommends that children aged under 12 months should not receive cows' milk as the main source of milk, while those aged over 12 months should not receive more than 600 mL per day.24 In the multiple logistic regression model, children who consumed more than 600 mL per day were almost four times as likely to have iron depletion, and targeted screening is strongly indicated based on this dietary history. Cows' milk is a poor source of iron, displaces foods with greater available iron and may also increase gastrointestinal occult blood loss. GPs should be aware of the importance of a dietary history for children of Arabic-speaking background and should enquire particularly about the volume of cows' milk consumed per day after 12 months of age. Acknowledgements We wish to thank the parents and children who participated in this study, the haematology laboratory staff at the Royal Alexandra Hospital for Children, Sydney, and the nurses of the participating Early Childhood Health Centres. The blood collection skills of Mrs Rhonda Dryden were invaluable to this study. The study was funded by the National Health and Medical Research Council Public Health Research Development Committee, Grant No: 97-417-7. References Aukett MA, Parks YA, Scott PH, Wharton BA. Treatment with iron increases weight gain and psychomotor development. Arch Dis Child 1986; 61: 849-857. Prasad AN, Prasad C. Iron deficiency; non-hematological manifestations. Prog Food Nutr Sci 1991; 15: 255-283. Berant M, Khourie M, Menzies IS. Effect of iron deficiency on small intestinal permeability in infants and young children. J Pediatr Gastroenterol Nutr 1992; 14: 17-20. Thibault H, Galtn P, Selz F, et al. The immune response in iron-deficient young children: effect of iron supplementation on cell-mediated immunity. Eur J Pediatr 1993; 152: 120-124. Oski FA, Honig AS, Helu B, Howanitz P. Effect of iron therapy on behavior performance in nonanemic, iron-deficient infants. Pediatrics 1983; 71: 877-880. Lozoff B, Jiminez E, Wolf AW. Long-term developmental outcome of infants with iron deficiency. N Engl J Med 1991; 325: 687-694. Williams J, Wolff A, Daly A, et al. Iron supplemented formula milk related to reduction in psychomotor decline in infants from inner city areas: randomised study. BMJ 1999; 318: 693-697. Walter T, De Andraca I, Chadud P, Perales CG. Iron deficiency anemia: adverse effects on infant psychomotor development. Pediatrics 1989; 84: 7-17. Karr M, Alperstein G, Causer J, et al. Iron status and anaemia in preschool children in Sydney. Aust N Z J Public Health 1996; 20: 618-622. Dallman PR, Siimes MA, Stekel A. Iron deficiency in infancy and childhood [review]. Am J Clin Nutr 1980; 33: 86-118. Gorten MK, Cross ER. Iron metabolism in premature infants: 2. Prevention of iron deficiency. J Pediatr 1964; 64: 509-520. Friel JK, Andrews WL, Matthew JD, et al. Iron status of very-low-birth-weight infants during the first 15 months of infancy. CMAJ 1990; 143: 733-737. Calvo EB, Galindo AC, Aspres NB. Iron status in exclusively breast-fed infants. Pediatrics 1992; 90: 375-379. Penrod JC, Anderson K, Acosta PB. Impact on iron status of introducing cow's milk in the second six months of life. J Pediatr Gastroenterol Nutr 1990; 10: 462-467. Mills AF. Surveillance for anaemia: risk factors in patterns of milk intake. Arch Dis Child 1990; 65: 428-431. NSW Department of Health, NSW Midwives Data Collection, 1994. Sydney: NSW Department of Health, 1995. Elin RJ, Wolff SM, Finch CA. Effect of induced fever on serum iron and ferritin concentrations in man. Blood 1977; 49: 147-153. Looker AC, Dallman PR, Carroll MD, et al. Prevalence of iron deficiency in the United States. JAMA 1997; 277: 973-976. Dallman PR, Siimes MA. Percentile curves for hemoglobin and red cell volume in infancy and childhood. J Pediatr 1979; 94: 26-31. Dallman PR, Looker AC, Johnson CL, Carroll M. Influence of age on laboratory criteria for the diagnosis of iron deficiency anaemia and iron deficiency in infants and children. In: Hallberg L, Asp N-G, editors. Iron nutrition in health and disease. London: J Libbey, 1996: 64-74. Stata statistical software release 5.0 [computer program]. College Station, Texas: Stata Corporation, 1997. McBride WG, Black BP, English BJ. Blood lead levels and behaviour of 400 preschool children. Med J Aust 1982; 2: 26-29. Ranmuthugala G, Karr M, Mira M, et al. Opportunistic sampling from early childhood centres: a substitute for random sampling to determine lead and iron status of pre-school children? Aust N Z J Public Health 1998; 22: 512-514. National Health and Medical Research Council. Dietary guidelines for children and adolescents. Canberra: AGPS, 1995. (Received 2 Mar, accepted 1 Sep, 2000) Authors' details Central Sydney Area Health Service, Sydney, NSW. Margaret A Karr, MPH, MSc(Med), Senior Research Officer, Division of General Practice; Michael Mira, MB BS, PhD, Clinical Professor, Department of General Practice, University of Sydney; Garth Alperstein, FRACP, Paediatrician and Clinical Senior Lecturer, University of Sydney, and Conjoint Senior Lecturer, University of New South Wales, Sydney, NSW; Samia Labib, BA, MEd(Health), Senior Interpreter, Health Interpreter Service. Department of Haematology, Royal Alexandra Hospital for Children, Sydney, NSW. Boyd H Webster, FRCPA, Senior Staff Specialist; Ahti T Lammi, FRACP, FRCPA, Senior Staff Specialist; Patricia Beal, MSc, Senior Hospital Scientist. Reprints will not be available from the authors. Correspondence: Professor M Mira, General Practice Casualty, Balmain Hospital, Booth Street, Balmain, NSW 2041. michaelmira_auATyahoo.co.uk 1: Definitions of impaired iron status used in the survey of children of Arabic background Iron depletion18 Plasma ferritin level Iron deficiency19 Iron depletion plus Mean corpuscular volume plus Mean corpuscular haemoglobin Iron-deficiency anaemia20 Iron deficiency plus Haemoglobin level Back to text 2: Prevalence of impaired iron status among children aged 12-38 months of Arabic background in central Sydney in 1997 Arabic background General population9 Iron status* Number % (95% CI) Number % (95% CI) Iron depletion Iron deficiency Iron-deficiency anaemia 87/382 27/315 20/315 23% (19%-27%) 9% (5%-12%) 6% (4%-9%) 36/381 14/329 5/329 9% (7%-12%) 4% (2%-7%) 2% (0-3%) *Definitions of iron status in children of Arabic background are shown in Box 1. The same definitions were used for children in the general population, except that iron deficiency was defined as iron depletion plus mean corpuscular volume < 70fL (age, 12-23 months) or < 73fL (age, 24-38 months), or red cell zinc protoporphyrin level > 80µmol/mol haem. Back to text 3: Potential risk factors for iron depletion among 382 children aged 12-38 months of Arabic background in central Sydney, 1997 Children with Children without Potential risk factors iron depletion iron depletion P* Born before 37 weeks' gestation 12/87 (14%) 10/295 (3%) 0.001 Birth weight 7/86 (8%) 8/293 (3%) 0.05 Breastfed initially 70/87 (81%) 245/295 (83%) 0.59 Breastfed at time of data collection 4/87 (5%) 9/295 (3%) 0.49 Cows' milk introduced before age of 12 months 32/87 (37%) 76/295 (26%) 0.06 Cows' milk introduced before age of 9 months 15/87 (17%) 36/295 (12%) 0.20 Consume >600mL cows' milk per day 54/83 (65%) 109/287 (38%) Consume ≥1L cows' milk per day 23/83 (28%) 30/287 (11%) Iron-fortified cereal as first solid 38/87 (44%) 149/295 (51%) 0.27 Consume meat 42/87 (48%) 122/295 (41%) 0.26 Receiving vitamin supplement 3/87 (3%) 11/295 (4%) 0.90 Receiving iron-containing supplement 1/87 (1%) 4/295 (1%) 0.88 Mother not born in Lebanon 26/87 (30%) 49/295 (17%) 0.01 Arabic or mainly Arabic spoken at home 58/87 (67%) 180/293 (61%) 0.39 Mother resident in Australia less than median time (8.5 years) 53/86 (62%) 137/294 (47%) 0.02 *By adjusted χ2 test. Back to text 4: Risk factors significantly associated with iron depletion among children aged 12-38 months of Arabic background in central Sydney, 1997 Univariate analysis Multivariate analysis Risk factor Odds ratio (95% CI) P Odds ratio (95% CI) P Gestation ≥37 weeks 1.00 1.00 4.55 (1.70-12.50) 0.003 5.88 (2.22-20.0) 0.001 Years mother in Australia ≥8.5 years 1.00 1.00 1.82 (1.10-3.03) 0.02 1.96 (1.36-3.33) 0.016 Cows' milk consumed daily ≤600mL 1.00 1.00 >600mL 3.04 (1.80-5.13) 3.89 (2.22-6.80) Country of birth Lebanon 1.00 Country other than Lebanon 2.14 (1.18-3.88) 0.01 NS NS=Not significant. Back to text

Margaret A Karr · Michael Mira · Garth Alperstein · Samia Labib · Boyd H Webster · Ahti T Lammi · Patricia Beal

Research Selective versus universal screening for gestational diabetes mellitus: an evaluation of predictive risk factors Richard X Davey and P Shane Hamblin MJA 2001; 174: 118-121 For editorial comment, see Wilson Abstract - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - More articles on Obstetrics & gynaecology and women's health Abstract Objective: To assess whether selective screening for gestational diabetes mellitus (GDM) on the basis of risk-factor assessment is a practicable alternative to universal screening. Design: Case-control study. Setting: A 212-bed regional specialist hospital in Melbourne, providing services in obstetrics and gynaecology, paediatrics, geriatrics and rehabilitation. Subjects: 6032 women who gave birth at the hospital, May 1996 to August 1997 and November 1997 to August 1998; all were screened for GDM, and 313 were diagnosed with the condition. Main outcome measures: Odds ratios (ORs) for risk factors (age, obesity, family history of diabetes mellitus and high-risk racial heritage) in women with GDM compared to those without GDM; proportion of women with GDM whose diagnosis would have been missed by selective screening. Results: ORs were 1.9 for age ≥25 years (95% CI, 1.3-2.7), 2.3 for body mass index ≥27 kg/m2 (95% CI, 1.6-3.3), 2.5 for high-risk racial heritage (95% CI, 2.0-3.2), and 7.1 for family history of diabetes mellitus (95% CI, 5.6-8.9). Other proposed criteria (previous GDM and glycosuria) added no further diagnostic power. Selective screening using the above four criteria would have missed two of 313 cases (0.6%) and could have saved screening up to 1025 women without GDM (17% of all women). Conclusions: Selective screening for GDM based on prior risk assessment can reduce the need for testing, with negligible loss of diagnostic efficiency. Gestational diabetes mellitus (GDM) is officially described as carbohydrate intolerance with onset or first recognition during pregnancy.1 It is associated with increased incidence of maternal hypertension, pre-eclampsia and obstetric intervention; a third of women with GDM develop diabetes mellitus in later life. Babies of mothers with GDM may be either macrosomic or small-for-gestational-age, and may suffer birth trauma, hypoglycaemia and other metabolic disturbances. Potential effects later in the child's life are still debated. In Australia, at least 5% of pregnancies are affected by GDM. However, there is no sharply defined maternal blood glucose level beyond which morbidity invariably ensues in either mother or baby,2 and there is disagreement about how to diagnose GDM and how aggressively to treat it.3In 1998, both the American Diabetes Association (ADA) and the WHO Consultation on diabetes published recommendations on diagnosis and classification of diabetes mellitus that included comments on GDM.4,5 The Australasian Diabetes in Pregnancy Society (ADIPS) has also published GDM guidelines.6 Both the ADA and ADIPS recommendations acknowledge that there are variable levels of risk for GDM and, consequently, that selective rather than universal screening can be considered. Selective screening both reduces costs and, for women deemed not to need screening, eliminates the minor physical inconvenience of the procedure and any anxiety raised by the possibility of suffering diabetes. Both ADA and ADIPS list risk factors for GDM (Box 1). The WHO Consultation's delineation of risk factors for GDM was less clear.5 We tested the hypothesis that selective screening for GDM is a practicable alternative to universal screening. We also investigated the effect of using the different age criteria of ADIPS and ADA as a basis for selective screening. Methods We undertook a case-control study to compare the likelihood of particular risk factors (defined in Box 2) among women with and without GDM. We also determined the proportion of women with GDM whose diagnosis would have been missed by selective screening, based on different sets of risk factors. Study population Sunshine Hospital is a 212-bed regional specialist hospital in Melbourne, Victoria, which provides service in obstetrics and gynaecology, paediatrics, geriatrics and rehabilitation. The study population comprised all 6032 women who gave birth at the hospital over the 26 months May 1996 to August 1997 and November 1997 to August 1998. Women who gave birth in September and October 1997 were excluded, as their laboratory data were incomplete. All women were screened with a 50 g glucose challenge test, according to the ADIPS protocol.6 Those with an abnormal result (defined as plasma glucose level after one hour of ≥7.8 mmol/L) proceeded to a 2 h 75 g oral glucose tolerance test (OGTT); GDM was diagnosed if the fasting plasma glucose level was ≥5.5 mmol/L, or the 2 h level was ≥8.0 mmol/L. Nearly all patients diagnosed with GDM were managed by an endocrinologist (P S H) in conjunction with one of the clinic obstetricians and were offered review and ongoing care from a dietitian and a diabetes nurse educator. Data retrieval Case group: We identified all post-delivery patient separations coded for GDM by computer search of the hospital medical information system, with cross-referencing to laboratory, dietitian and diabetes nurse educator records. Women who gave birth twice in the study period were included only once, using details from their first GDM-affected pregnancy. There were 313 women diagnosed with GDM. Information on risk factors for these women was obtained from medical records containing details of pregnancy management and delivery, endocrinologist's and dietitian's notes and laboratory records (by R X D). If racial heritage was unclear, patients were telephoned at home to obtain more details. Body mass index (BMI) was available for only 290 of the 313 women (93%), but other data were available for over 99%. Control group: For the 5719 women without GDM, information on age was also obtained from the hospital medical records. However, it was impracticable to investigate racial heritage as closely for this group as for the case group. Therefore, if country of birth was recorded as being in Europe, Asia or Central and South America, it was used for risk categorisation (45.5% of women). All other women were allocated to risk groups in the same proportions as found in the case group. While this biases the outcome in favour of the null hypothesis, it is more accurate than making no such allocation at all. BMI and family history were not available for the 5719 women without GDM. Therefore, the BMI comparison used BMI data obtained from 303 consecutive non-diabetic women presenting for a glucose challenge test at about 28 weeks' gestation as part of a 1995 study at Sunshine Hospital.8 As the patient catchment area was unchanged between 1995 and 1998, this group should represent an unbiased sample of women who presented between 1995 and 1998. For the family history comparison, a recent estimate of prevalence of diabetes mellitus in Australia9 was used for the non-GDM patients. Background risk was corrected for the bias caused by the tendency of patients with diabetes to visit their doctors twice as often as non-diabetic patients.10 It was also doubled to give a worst-case estimate, as each parent might pass on heritable risk independently. Statistical analyses and ethics approval Data were analysed using Stata statistical software.11 Odds ratios were calculated from the comparative prevalence in affected and control populations by Cornfield's method. Ethics approval for this study was not required by the Victorian Health Services Act 1988 and was not sought. Data were permanently de-identified after analysis. Results Risk-factor comparison Prevalence of risk factors among women with and without GDM is shown in Box 3, along with odds ratios. Women with GDM were almost twice as likely to be aged 25 years or over compared with those without GDM, more than twice as likely to have a BMI ≥27 kg/m2 or to have a high-risk racial heritage, and more than seven times as likely to have a family history of diabetes mellitus. To determine the value of racial heritage as a predictor of GDM in isolation from family risk, we determined the OR for high-risk racial heritage among women with no family history of diabetes mellitus. This OR was not statistically different from the earlier OR for high-risk racial heritage that included women with a family history. Furthermore, birth in Australia, New Zealand or North America (of non-Indigenous background) does not necessarily equate with low heritable GDM risk. Of the 313 women with GDM, 94 were born in these countries, but 19 of these had high-risk racial heritage. Finally, we also assessed whether glycosuria in pregnancy or previous GDM had any extra value as predictors of GDM. All women with these risk factors qualified for screening on other grounds. Effect of selective screening The numbers of women with GDM who would be screened on the basis of risk factors, using different age thresholds, are shown in Box 4. Only the 290 women with complete data for all risk factors are included. However, all 23 women with incomplete risk-factor data would have undergone screening under these selective screening policies, as all had at least one risk factor (10 had two factors and five had three). Selective screening on the basis of at least one risk factor, using the ADIPS age criterion (≥30 years), would have missed 12 women with GDM (95% CI, 6-19; 4%). Using the ADA age criterion (≥25 years), selective screening would have missed only two women with GDM (95% CI, 0-5; 0.6%). Furthermore, χ2 tests showed that the proportions of women with GDM who had risk factors other than age did not differ significantly between age groups (≥30 years, ≥25 years and all ages); all P values exceeded 0.67. Among the women without GDM, 83% were aged 25 years or over and 48% were aged 30 years or over. A selective screening policy could therefore have saved testing up to 17%-52% of women without GDM, depending on the age threshold used and the presence of risk factors other than age. Discussion We found that selective screening for GDM using the four criteria common to the ADA and ADIPS lists of risk factors -- older age, obesity, family history of diabetes and high-risk racial heritage -- would have missed few women with GDM in our study population. It is clear that the age threshold for screening proposed by ADA (25 years) is diagnostically safer than the ADIPS threshold of 30 years, missing only 0.6% versus 4% of women with GDM. However, our data also show that it is important to examine risk factors other than age even when the lower age threshold is used, as the other factors underlying susceptibility to GDM operate irrespective of age. Family history and heredity are immutable, and obesity may also be partly under genetic control. These observations are also consonant with the theory that pregnancy unmasks diabetes mellitus prematurely.12 We also found that previous GDM and glycosuria in pregnancy added nothing to the above four criteria for screening. However, this does not mean that GDM in a previous pregnancy should be ignored. It is often regarded as a criterion for a full OGTT, without a prior glucose challenge test, earlier than 28 weeks' gestation; the wisdom of this practice is not doubted. Not only are the ADA criteria for screening diagnostically safer than the ADIPS criteria, they are also more precise in their definitions. Australian women would benefit if ADIPS recommendations were brought into line with ADA recommendations. Our conclusions differ from those of Moses and colleagues, who found no benefit from selective screening in their study in the Illawarra region of New South Wales.13,14 Indeed, Moses has championed universal screening.15 However, the Illawarra protocol for GDM screening varied from contemporary practice, as it did not measure fasting glucose level or stringently control the time between the glucose load and blood sampling, making comparison difficult. Its outcomes have been questioned.16,17 In North America, a recent, albeit small, retrospective study of GDM screening in Michigan specifically assessed the ADA selective screening recommendations and concluded that they can be used as they miss "few" (4%) women with GDM.18 A larger study was reported by the Toronto Trihospital Investigators.19 They proposed a scheme that used among its criteria those later published by ADA to differentiate risk levels for GDM, sparing 35% of pregnant women the need for a glucose challenge test. This is at least twice as efficient as using the ADA criteria in our population, which potentially spared up to 17% of women a glucose challenge test (depending on the presence of risk factors other than age). However, an accompanying editorial concluded that the Trihospital criteria were "so hard to discern" that universal screening would continue as the only practicable alternative.20 For busy clinicians, simple systems are essential. Our study differed from the Toronto study in that only women with positive results on a glucose challenge test proceeded to an OGTT, while, in Toronto, all women had a full OGTT. We will have missed the small number of women who would have had positive results on an OGTT despite their negative results on a challenge test -- perhaps 3%, based on the Toronto data. Short of performing a full OGTT on all pregnant women, which is impracticable, this group will always escape detection. The Toronto ORs for risk factors among those with GDM generally accord with ours (1.6 for age ≥35 years [95% CI, 1.1-2.5], 3.2 for BMI ≥25.1 kg/m2 [95% CI, 2.1-4.8], 4.8 for Asian race [95% CI, 3.0-7.6]), but our results differ in two ways. Toronto race groupings, apart from "Asian", are difficult to interpret and, by using "white" and "black", ignore the extreme variation among "white" Europeans. Secondly, family history of diabetes mellitus in Toronto did not correlate significantly with higher risk of GDM, whereas our findings strongly support the inclusion of family history among criteria for a glucose challenge test. Our study, along with the Michigan and Toronto studies, indicates that a selective approach to GDM screening in pregnancy is justifiable. Our simplified algorithm for selective screening is shown in Box 5. In practice, the proportion of women spared a glucose challenge test by selective screening will vary between populations. Consequently, whether selective screening is locally practicable will be a decision for individual groups of obstetricians, endocrinologists and pathologists with local knowledge. It is clear that consideration of a patient's age, rigorous questioning about racial and family history and accurate measurement of height and weight can reduce the need for screening among suitable populations. Selective screening can reduce costs and maternal anxiety, with negligible loss in diagnostic power. Nevertheless, GDM poses still further challenges. Australia needs a better-directed, more organised, totally inclusive approach to follow-up of women who have had GDM. The third who will go on to develop diabetes mellitus need to be tracked, monitored, and managed prospectively into a healthier future. Acknowledgements We thank dietitians Candy d'Menzie-Bunshaw and Ruth Cuttler, specialist diabetes nurse consultant Elizabeth Borg, and health information manager Sianne Banks and her staff at Sunshine Hospital for their invaluable assistance with the study, and Lucy Inocencio for her excellent technical assistance. References Metzger BE, editor. Summary and recommendations of the Third International Workshop-Conference on Gestational Diabetes Mellitus. Diabetes 1991; 40 Suppl 2: 197-201. Sacks DA, Greenspoon JS, Abu-Fadil S, et al. Towards universal criteria for gestational diabetes: The 75-gram glucose tolerance test in pregnancy. Am J Obstet Gynecol 1995; 172: 607-614. Jovanovic L. A tincture of time does not turn the tide [editorial]. Diabetes Care 2000; 23: 1219-1220. The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Report of the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes Care 1998; 21 (Suppl 1): S5-S19. Alberti KGMM, Zimmet PZ for the WHO Consultation. Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1. Diagnosis and classification of diabetes mellitus. Provisional report of a WHO consultation. Diabet Med 1998; 15: 539-553. Hoffman L, Nolan C, Wilson JD, et al, for the Australasian Diabetes in Pregnancy Society. Gestational diabetes mellitus -- management guidelines. Med J Aust 1998; 169: 93-97. Beischer NA, Oats JN, Henry OA, et al. Incidence and severity of gestational diabetes mellitus according to country of birth in women living in Australia. Diabetes 1991; 40: 35-38. Davey R. The glucose challenge test: different drink dilutions. Diabet Med 1996; 13: 917-918. Welborn TA, Reid CM, Marriott G. Australian Diabetes Screening Study: impaired glucose tolerance and non-insulin-dependent diabetes mellitus. Metabolism 1997; 46 (12 Suppl 1): 35-39. Australian Bureau of Statistics. National health survey: diabetes, Australia, 1995. Canberra: AGPS, 1997. (Catalogue No. 4371.0.) StataCorp. 1999. Stata Statistical Software: Release 6. College Station. TX: Stata Corporation. Yue DK, Molyneaux LM, Ross GP, et al. Why does ethnicity affect prevalence of gestational diabetes? The underwater volcano theory. Diabet Med 1996; 13: 748-752. Moses R, Griffiths R, Davis W. Gestational diabetes: do all women need to be tested? Aust N Z J Obstet Gynaecol 1995; 35: 387-389. Moses RG, Moses J, Davis WS. Gestational diabetes: do lean young Caucasian women need to be tested? Diabetes Care 1998; 21: 1803-1806. Moses RG. Diabetes in pregnancy [editorial]. Med J Aust 1998; 169: 68-69. Davey R. Of gestational diabetes, finesse, and an antipodean snark [letter]. Diabetes Care 1999; 22: 873-874. Moses RG, Moses J, Davis WS. Response to Davey [letter]. Diabetes Care 1999; 22: 874. Williams CB, Iqbal S, Zawacki CM, et al. Effect of selective screening for gestational diabetes. Diabetes Care 1999; 22: 418-421. Naylor CD, Sermer M, Chen E, Farine D. Selective screening for gestational diabetes mellitus. N Engl J Med 1997; 337: 1591-1596. Greene MF. Screening for gestational diabetes mellitus [editorial]. N Engl J Med 1997; 337: 1625-1626. (Received 15 May, accepted 21 Sep, 2000) Authors' details Western Hospital, Melbourne, VIC. Richard X Davey, FRCPA, FACB, Clinical Pathologist; P Shane Hamblin, FRACP, Senior Endocrinologist. Reprints will not be available from the authors. Correspondence: Dr R X Davey, Western Hospital, Gordon Street, Footscray, VIC 3011. richard.daveyATwh.org.au Make a comment 1: Risk factors for gestational diabetes mellitus, listed by different sources Risk factor Australasian Diabetes in Pregnancy Society6 American Diabetes Association4 Age Obesity Family history of diabetes mellitus Previous GDM High risk "ethnic" group Glycosuria Previous adverse pregnancy outcome Yes (>30 years) Yes (not defined) Yes Yes Yes (examples given)† Yes Yes Yes (>25 years) Yes (BMI >27kg/m2) Yes (first-degree relative) Not mentioned* Yes (examples given)‡ Not mentioned Not mentioned* GDM=Gestational diabetes mellitus. BMI=Body mass index. *While the ADA did not consider previous GDM or adverse pregnancy outcome as sufficiently significant for women to be included in the high-risk GDM group, it did report them as criteria for diabetes testing in asymptomatic, undiagnosed individuals,4 thereby acknowledging them as markers of early, silent diabetes mellitus. †Including Australian Indigenous, Polynesian, Asian and Middle Eastern women. ‡Including Hispanic-American, Native American, Asian-American, African-American and Pacific Islander women. Back to text 2: Definitions of risk factors used in this study Age: Age was not further defined by ADA or ADIPS; we used age at estimated time of conception — the most conservative calculation. Obesity: As defined by the ADA — body mass index ≥27kg/m2, determined from pre-pregnancy mass and height. Family history of diabetes mellitus: As defined by the ADA — diabetes mellitus affecting a first-degree relative. Racial susceptibility: Termed "ethnic" risk by ADA and ADIPS. We classified a woman as having high-risk racial heritage if she or her parents were born in one of the countries around the Mediterranean (including the Levant, but not the rest of Europe), the Indian subcontinent or Asia, or belonged to the Indigenous populations of Australia, the Pacific or the Americas.7 ADA = American Diabetes Association. ADIPS=Australasian Diabetes in Pregnancy Society. Back to text 3: Prevalence and odds ratios of risk factors for gestational diabetes mellitus (GDM) Prevalence Risk factor Women with GDM (n=313) Women without GDM (variable n*) Odds ratio (95% CI) Age (years) >25 90.1% 82.9% 1.9 (1.3-2.7) >30 58.5% 47.8% 1.5 (1.2-1.9) Body mass index >27 kg/m2 36.2%† 19.8% 2.3 (1.6-3.3) Family history of diabetes mellitus 39.9% 8.6% 7.1 (5.6-8.9) High-risk racial heritage 68.7%‡ 46.4% 2.5 (2.0-3.2) Among women with no family history 71.7%§ 42.8%¶ 2.9 (2.1-4.0) *Sample size varied between risk factors: 5719 (age), 303 (body mass index), 50371 (family history) and 5719 (racial heritage). †Data were available for 290 of the 313 women. ‡High-risk racial heritage: peri-Mediterranean (56 women), Indian subcontinent (20), Asia (124), South America (12), Indigenous populations (3); low risk racial heritage: United Kingdom (61), other European countries (35) and other (2). §n=187. ¶n=5324. Back to text 4: Effect if selective screening were used among 290* women with gestational diabetes mellitus Number of women who would be screened (% of women with GDM) Risk factors† ADIPS age threshold (≥30 years) ADA age threshold (≥25 years) Age ≥ threshold Only risk factor Plus any one other factor Plus any two other factors Plus any three other factors Total 13 (4%) 78 (27%) 54 (19%) 20 (7%) 165 (57%) 23 (8%) 120 (41%) 90 (31%) 27 (9%) 260 (90%) Age ≤ threshold One risk factor Two risk factors Three risk factors Total 60 (21%) 42 (14%) 11 (4%) 113 (39%) 18 (6%) 6 (2%) 4 (1%) 28 (10%) Any risk factor 278 (96%) 288 (99%) *23 women with gestational diabetes mellitus but incomplete data on risk factors are not included. †Risk factors other than age were body mass index ≥27kg/m2, family history of diabetes mellitus and high-risk racial heritage. Back to text Back to text

Richard X Davey

A randomised controlled trial was planned to compare two different treatment strategies — structured problem solving and selective serotonin reuptake inhibitor (SSRI) medication — for patients with mild to moderate major depression. The trial was to be conducted in the primary care setting with all treatment given by general practitioners. When no patients had been recruited into the study after six months, we performed an audit of all patients with depressive symptoms attending the doctors' practices over three weeks. Exclusion criteria were changed to ease entry into the trial, but still no patients were recruited over the following six months. What went wrong? MJA 2001; 174: 144-146 Why did the trial fail? - Acknowledgements - References - Authors' details - - More articles on General practice and primary care The recent National Survey of Mental Health and Wellbeing found that depression was associated with significant disability and that 6.3% of the Australian population was estimated to have suffered a major depressive disorder in the previous 12 months.1 The survey also showed that general practices were the main points of contact for patients with a mental disorder, consistent with previous reports that only 5% of such patients are referred to psychiatrists.2 As most depressed patients will be treated in primary care, the evaluation of treatment in this setting is important. Structured problem solving is emerging as an effective treatment for clinical depression.3-6 This treatment aims to teach patients to use their own resources to deal with their problems and includes skills such as identifying and simplifying problems, "brainstorming" potential solutions, and implementing these solutions. The treatment is brief, has clearly identified stages, and is very suitable for delivery by primary healthcare professionals. However, little research in primary care settings has used general practitioners as the main treatment providers, and thus the degree to which the evidence for the effectiveness of structured problem solving can be generalised to general practice is questionable. For this reason we designed a randomised trial with all treatment conducted by GPs (Box 1). The study was, in part, a replication of an earlier United Kingdom project that showed that problem solving was better than placebo, but equivalent to amitriptyline.4 Six months after the trial commenced no patients had been recruited. Why did the trial fail? In practical terms, the trial should not have failed, as mild to moderate depression is a common presentation in primary care and the research protocol addressed many of the problems identified in previous primary care research failures.9,10 For example: We used GPs who had participated in a Masters program designed to improve the recognition and management of mental disorders in general practice; We involved the participating GPs in the development of the protocols; We made every effort to minimise tasks involved in the study for the GPs; and The chief investigator was responsible for the teaching on the Masters program, and had developed a close working relationship with the GPs. While our original intention was not to address the complexity of conducting randomised trials in clinical practice, the apparent unease of many of the GPs with randomisation raises some potential reasons for the failure (Box 2). Apparent ambivalence towards randomisation in medicine, despite strong support in the scientific literature, has been reported previously.11 Silverman argues that as medicine shifts from the traditional authoritarian stance of "doctor knows best" towards the use of clinical guidelines and standardised treatment protocols, there is an increased discomfort in any approach that might be seen to admit a lack of crucial knowledge on the part of the doctor.11 In other words, asking patients to consent to randomisation between two conditions admits an uncertainty that compromises the traditional doctor-patient relationship. However, it has not been difficult to engage doctors in trials involving randomisation of their patients, and there is evidence that many are prepared to follow simple randomisation protocols.12 For example, one of the first randomised controlled trials of giving aspirin to patients very early after myocardial infarction (to stop or reverse the thrombotic process) enrolled 2500 GPs. These GPs, who were blind to the treatment condition they were offering, agreed to give the allocated capsules (aspirin or placebo) to any patient who presented with chest pain or other symptoms likely to be caused by a myocardial infarct. Two thousand patients were recruited into the trial, suggesting that, in this example, randomisation was not a substantial concern. If it is not randomisation per se that causes reluctance to recruit, then other factors need consideration. In our study, the audit results indicated that one in six patients had been excluded because the GPs lacked confidence that structured problem solving would be of benefit. Perhaps the GPs lacked confidence in their ability to deliver the psychological treatment effectively. If this is the case, the failure of this trial may have been influenced by the inclusion of a psychological treatment. We have argued elsewhere that doctors remain cautious about using non-drug treatments, partly because of the lack of organised promotion of non-proprietary treatments, and partly because of the difficulty in ensuring quality control.13 Yet, these GPs had agreed problem solving was a useful treatment, had demonstrated competence as part of their training, and two GPs in the group had conducted their own research projects where they taught structured problem solving to other doctors. It is also possible that, given the obvious differences between the psychological and pharmacological approaches, the GPs had formed an opinion early in the recruitment process that one or other approach would better suit a particular patient. In this case, exclusion from the trial was based on the presumption that the doctor already knew what treatment was best.11,14 In a disturbing example of "doctor knows best", two-thirds of suitable patients were not enrolled in a randomised trial of antiarrhythmia drugs because their doctors were so convinced of the benefits of the drugs that they did not want their patients allocated to the placebo group.15 The study eventually showed that these drugs were capable of causing fatal arrhythmias, so those doctors were essentially withholding from their patients the 50% chance of being allocated to the safer placebo alternative. Other factors may cause doctors to hesitate when faced with recruiting patients, including the complexities of obtaining informed consent, the need to complete clinical ratings or ask patients to complete self-report questionnaires for measurement of outcome, or the need to work within a specific treatment protocol. Although tasks for the GPs were minimal, it is likely that the necessary role change from practitioner to scientist-practitioner was too great to facilitate a shift in treating behaviour. If this is the case, the inherent contradiction between research and delivery of care in the minds of many clinicians is a basic problem for clinical research.16 Nevertheless, treatments that are efficacious in research settings need to be evaluated under the conditions of routine care, and it may not be sufficient to rely on the use of qualitative methods to evaluate outcome in primary care.17 Perhaps an explicit use of the "uncertainty principle" in randomised controlled trials within routine clinical care will enhance recruitment rates.18 That is, if there is apparent certainty about what treatment is best, it is ethically untenable that patients should have their treatment chosen at random, so only patients for whom there is uncertainty about which treatment would be best should be recruited into a randomised trial. In this way the ethical dilemma for clinicians is solved, and the heterogeneity of the patient sample maintained, as clinicians will differ significantly in the types of patients they will be uncertain about. Financial incentives might increase the involvement of clinicians in research, but have caused public outcry in the United States;19,20 the resulting assertive recruiting can also erode informed consent.21,22 We argue instead that a more fundamental shift in ethos and knowledge of principles that underlie research in clinical practice is required. This argument has parallels in the recent Royal Australian College of General Practitioners report on an implementation strategy for evidence-based clinical practice guidelines.23 The report points to a lack of understanding of the principles underpinning the use of clinical practice guidelines in general practice, and the need to develop a culture in which such guidelines are used and valued. In regard to outcome research, until doctors accept a scientist-practitioner model of practice it is unlikely that they will feel comfortable using conventional research protocols. Clinical research conducted in routine care will help clinicians make informed decisions about what may be the best treatment for their patients based on scientifically derived knowledge. We hope that the questions raised in this article will stimulate debate and research that will directly address this important issue. Acknowledgements This research was supported by a grant from the School of Psychiatry, University of New South Wales. References Andrews G, Henderson S, Hall W. Prevalence, comorbidity, disability and service utilisation; and overview of the Australian national mental health survey. Br J Psychiatry 2001. In press. Gath D, Catalan J. The treatment of emotional disorders in general practice: psychological methods versus medication. J Psychosom Res 1986; 30: 381-386. Mynors-Wallis L, Davies I, Gray A, et al. A randomised controlled trial and cost analysis of problem-solving treatment for emotional disorders given by community nurses in primary care. Br J Psychiatry 1997; 170: 113-119. Mynors-Wallis LM, Gath DH, Lloyd-Thomas AR, Tomlinson D. Randomised controlled trial comparing problem-solving treatment with amitriptyline and placebo for major depression in primary care. BMJ 1995; 310: 441-445. Schulberg HC, Block MR, Madonia MJ, et al. Treating major depression in primary care practice. Eight-month clinical outcomes. Arch Gen Psychiatry 1996; 53: 913-919. Catalan J, Gath DH, Anastasiades P, et al. Evaluation of a brief psychological treatment for emotional disorders in primary care. Psychol Med 1991; 21: 1013-1018. Depression Guideline Panel. Depression in primary care. Vol. 2. Treatment of major depression. Clinical Practice Guideline No. 5. Rockville, MD: Department of Health and Human Services, Public Health Service, Agency for Health Care Policy and Research, April 1993. (AHCPR Publication No. 93-0551.) ICD-10 classification of mental and behavioural disorders. Geneva: World Health Organization, 1992. Foy R, Parry J, McAvoy B. Clinical trials in primary care. BMJ 1998; 317: 1168-1169. Peto V, Coulter A, Bond A. Factors affecting general practitioners' recruitment of patients into a prospective study. Family Practice 1993; 10: 207-211. Silverman W. Equitable distribution of the risks and benefits associated with medical innovations. In: Maynard A, Chalmers I, editors. Non-random reflections on health services research: on the 25th anniversary of Archie Cochrane's effectiveness and efficiency. London: BMJ Publishing Group, 1997: 184-193. Elwood P. Cochrane and the benefits of aspirin. In: Maynard A, Chalmers I, editors. Non-random reflections on health services research: on the 25th anniversary of Archie Cochrane's effectiveness and efficiency. London: BMJ Publishing Group, 1997: 107-121. Andrews G. On the promotion of non-drug treatments. BMJ 1984; 289: 994-995. Segelov E, Tattersall MHN, Coates AS. Redressing the balance -- the ethics of not entering an eligible patient on a randomised trial. Ann Oncol 1992; 3: 103-105. Moore TJ. Deadly medicine. New York: Simon and Schuster, 1995. Tognoni G, Alli F, Avanzini F, et al. Randomised clinical trials in general practice: lessons from a failure. BMJ 1991; 303: 969-971. Miller ML, Crabtree BF. Qualitative analysis: how to begin making sense. Family Practice Res J 1994; 14: 289-297. Collins R, Peto R, Gray R, Parish S. Large-scale randomised evidence: trials and overviews. In: Maynard A, Chalmers I, editors. Non-random reflections on health services research: on the 25th anniversary of Archie Cochrane's effectiveness and efficiency. London: BMJ Publishing Group, 1997: 197-230. Eichenwald K, Kolata G. Drug trials hide conflicts for doctors. New York Times, May 16, 1999; 1,28-29. Eichenwald K, Kolata G. A doctor's drug trials turn into fraud. New York Times, May 17, 1999. Ferguson C. Payment of financial incentives to GPs may invalidate informed consent process. BMJ 1998; 316: 75-76. Shalala D. Protecting research subjects -- what must be done. N Engl J Med 2000; 343: 808-810. Report on consultancy to develop an implementation strategy for CPGs. Sydney: Royal Australian College of General Practitioners, August 2, 2000. Authors' details School of Psychiatry, University of New South Wales, Sydney, NSW. Caroline J Hunt, MPsych, PhD, Lecturer (currently, Senior Lecturer, Department of Psychology, University of Sydney). Gavin Andrews, MD, Professor. Clinical Research Unit for Anxiety Disorders, St Vincent's Hospital, Sydney, NSW. Louise M Shepherd, BA(Hons), MPsych, Clinical Psychologist. Reprints will not be available from the authors. Correspondence: Dr C J Hunt, Department of Psychology (F12), University of Sydney, NSW, 2006. carolineATpsych.usyd.edu.au Make a comment 1: The planned trial Objective: To compare structured problem solving with SSRI medication and non-specific counselling for mild to moderate major depression. Trial development: The trial was initially designed with three treatment arms: structured problem solving, medication and placebo. In 1997, general practitioners in their second year of a two-year part-time Master of Psychological Medicine course were approached for feedback. The GPs indicated that, while they believed the study to be of value, they were uncomfortable with the use of a placebo control, so the placebo arm of the trial was abandoned. Detailed protocols for each treatment were developed, with medications which reflected best prescribing practice (United States Department of Health and Human Services Clinical practice guidelines 7), a treatment period of six months, and a follow-up period to track longer-term changes. The protocols were designed to be simple to use and, to minimise tasks for the GPs, telephone assessments by a clinical psychologist were planned to confirm diagnosis, assess severity, and evaluate outcome. Design: Patients assessed by their GPs as having mild to moderate major depression were to be randomly allocated to: structured problem solving alone; selective serotonin reuptake inhibitor (SSRI) medication and non-specific counselling; or SSRI medication and structured problem solving. Patients were required to meet International classification of diseases - 10th revision 8 criteria for a mild or moderate major depressive episode. Exclusion criteria included current or previous manic (or hypomanic) or psychotic symptoms, current suicidal intent, current drug or alcohol abuse, current pharmacological or psychological treatment for depression, and (to exclude severe depression) current somatic (or melancholic) features. Participating general practitioners: In 1998, the protocols were again reviewed by students in the Masters course, who agreed that they were consistent with current best primary care practice and could be delivered in this setting. Ten GPs agreed to participate - four were current second-year students in the Masters course, and six were graduates of the course. Three current second-year students did not participate because they were not working in general practice, and two because they did not wish to randomly allocate their patients to treatment. All participating doctors had been trained in assessing and managing depression (including structured problem solving), and clinical supervision was offered over the course of the trial. Back to text 2: Why were no patients recruited? Six months after commencement of the trial in 1998, despite frequent reminders and discussions about the trial protocol, no patients had been recruited. This prompted a clinical audit of all patients with depressive symptoms attending the participating doctors' practices over three weeks. Each general practitioner recorded patients presenting with depressive symptoms and the reasons why they considered them unsuitable for the trial on a form we designed for this purpose. Over the three weeks, 114 patients presented with depressive symptoms (12% of all presenting patients), but none were entered in the study. The results of the audit are shown in the Table. The most frequently cited reasons for exclusion were current pharmacological or psychological treatment for depression, or depression of insufficient severity to meet ICD-10 criteria. In an attempt to improve recruitment in the following six months, we dropped the exclusion criteria of current psychological treatment and insufficient severity (these features now to be assessed by the clinical psychologist), yet still no patients were recruited. Reasons for not entering trial No. (%) patients* Depression insufficiently severe to meet ICD-10 criteria 27 (23.6%) Severe depression or criteria met for "somatic syndrome" 13 (11.4%) Prior or current manic, hypomanic or psychotic episode 6 (5.3%) Serious suicidal intent 5 (4.4%) Current drug or alcohol abuse 6 (5.3%) Current pharmacological treatment for depression 38 (33.3%) Current psychological treatment for depression 34 (29.8%) Physical problems precluding use of an SSRI 1 (0.9%) Prior failure to respond to an SSRI 5 (4.4%) Patient refused randomisation 7 (6.1%) General practitioner not confident about using problem solving with this patient 20 (17.5%) Other (eg, dementia, communication difficulties, personality disorders) 22 (19.3%) *There were 114 patients, but general practitioners frequently nominated more than one reason for excluding patients. ICD-10=International classification of diseases - 10th revision.8 SSRI=selective serotonin reuptake inhibitor. Once the trial was formally abandoned, there were discussions with the four GPs still completing their Masters course. These GPs admitted unease with the process of randomisation, despite the demonstrated efficacy of both treatments for this population. Back to text

Caroline J Hunt · Louise M Shepherd · Gavin Andrews

Research The effects of Chinese medicinal herbs on postmenopausal vasomotor symptoms of Australian women A randomised controlled trial Susan R Davis, Esther M Briganti, Run Q Chen Fabien S Dalais, Michael Bailey and Henry G Burger MJA 2001; 174: 68-71 For editorial comment, see Eden Abstract - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - More articles on Obstetrics & gynaecology and women's health Abstract Objective: To evaluate the effects of a defined formula of Chinese medicinal herbs (CMH) on menopausal symptoms. Design: A double-blind randomised placebo-controlled trial. Methods: Between August 1998 and April 1999, 55 postmenopausal Australian women recruited from an urban population completed 12 weeks of intervention with either a defined formula of CMH (n = 28) or placebo (n = 27) taken twice daily as a beverage. Main outcome measures: The primary end-point was change in frequency of vasomotor events (hot flushes and night sweats). The secondary end-points were changes in score for the domains measured in the Menopause Specific Quality of Life (MENQOL) Questionnaire. Results: There was a reduction in average weekly frequency of vasomotor events with CMH (- 15%; 95% CI, - 31% to + 1%) and with placebo (- 31%; 95% CI, - 42% to - 21%). The difference between groups favoured the use of placebo; however, this was not significant (P = 0.09). Although significant reductions in scores for the various domains of the MENQOL Questionnaire were observed for both CMH and placebo, there were no significant differences between the two treatment groups for any domain. There was evidence for effect modification by previous use of natural therapies for the vasomotor, physical and sexual domains of the MENQOL Questionnaire: women with no prior use of natural therapies for their menopausal symptoms responded to therapy, whereas prior users did not. Conclusions: The defined formula of CMH was no more effective than placebo in reducing vasomotor episodes in Australian postmenopausal women, or in improving any of the four symptom domains in the MENQOL Questionnaire. Three of the MENQOL Questionnaire domains were modified by prior use of natural therapies. This finding has implications for future studies. In Australia, more than a third of postmenopausal women are troubled by vasomotor symptoms.1 Although hormone replacement therapy (HRT) eliminates 60% of flushes within three months,2 a significant number of postmenopausal women have absolute or relative contraindications to HRT, or are unwilling to use this therapy.3Chinese herbal medicine has been used for centuries in China for treating menopausal symptoms, and is still in current use. Clinical trials in China have shown significant effects of Chinese herbal medicine in alleviating menopausal symptoms in Chinese women,4,5 but these effects may not be generalisable. We therefore conducted a double-blind, randomised, placebo-controlled study of the effects of a modified traditional formula of Chinese medicinal herbs (CMH) on vasomotor symptoms in Australian postmenopausal women. Methods Study population Approval for the study was obtained from the Human Research and Ethics Committee of Monash Medical Centre. All patients gave written informed consent. The study was conducted between August 1998 and April 1999. Inclusion criteria: Patients were recruited through the Jean Hailes Foundation Newsletter, newspapers, radio station interviews and the Medical Unit of the Jean Hailes Foundation. Non-Asian women aged 45 to 70 years, who had lived in Australia for at least 10 years, were postmenopausal (> 12 months' amenorrhoea and follicle-stimulating hormone level > 25 IU/L), and reported at least 14 hot flushes or night sweats per week were eligible for the study. Exclusion criteria: Women were excluded if they had used HRT, CMH or other natural therapies (evening primrose oil, yam cream, progesterone cream or any other over-the-counter preparation for menopausal symptoms) during the eight weeks before baseline, or if they had preexisting gastrointestinal, renal or liver disease, diabetes mellitus requiring treatment, uncontrolled hypertension, undiagnosed vaginal bleeding, systemic glucocorticosteroid use, or were undergoing cancer therapy. Women who had consumed a high phytoestrogen diet (according to a food frequency questionnaire6) for the four weeks before baseline were also excluded. Randomisation: Subjects were randomised to CMH or placebo using a randomisation chart constructed by randomising numbers 1 to 88 into two groups using Microsoft Excel.7 Study intervention The defined formula of CMH for the active preparation used in the trial is listed in Box 1. Placebo was cornstarch with a bitter taste enhancer. All herbs are listed with the Australian Therapeutic Goods Administration, and were administered within standard dosage levels. All were screened for heavy metal contamination by Ningbo Daekang Herbs Co Ltd (Ningbo, China) and the National Analytical Laboratories in Melbourne. The CMH and placebo were produced by Ningbo Daekang Herbs Co Ltd as granules soluble in warm water, and each dose was prepackaged in identical aluminium foil sachets. Patients were instructed to drink one sachet of granules dissolved in 200 mL of warm water, twice a day. Both solutions had similar unusual tastes. A four-week supply was dispensed at each treatment visit. The design did not accommodate the diagnostic and therapeutic principles of Chinese medicine, but was for analysing the therapeutic efficacy of the herbs. Outcome measures Demographics, body mass index (kg/m2), medical history, gynaecological history, and use of previous HRT or natural therapies for menopausal symptoms were documented at baseline. The primary end-point of the study was the effect of treatment on the frequency of vasomotor symptoms. Each woman completed a daily diary of the frequency of hot flushes and night sweats for four weeks before the commencement of treatment and for the entire 12 weeks of the study period. The secondary end-points of the study were the effect of treatment on each of the four domains of the Menopause Specific Quality of Life (MENQOL) Questionnaire and on urinary phytoestrogen excretion. The MENQOL Questionnaire is a validated instrument that tests physical, vasomotor, psychosexual and sexual domains of quality of life.8 It can differentiate between women according to quality of life, as well as measure changes in quality of life. A minimum score of zero corresponds to no symptoms and a maximum score of seven corresponds to extremely bothersome symptoms. The smallest clinically relevant change is a difference in one point within the domains, representing a 15% change. As dietary history is a poor guide to phytoestrogen ingestion, a potential confounding factor, we also measured urinary phytoestrogen excretion in women participating in the trial. Total urinary daidzein and genistein excretion was measured in 24-hour urine samples at baseline and at Week 12.9 Subjects were asked not to modify their dietary pattern for the study period. Sample size and statistical analysis The sample size was calculated based on the primary end-point of change in hot flushes and night sweats. A clinically relevant effect of treatment is considered to be at least a 40% reduction in vasomotor events.10-12 Anticipating a 30% placebo response, for power of 80% and a significance level of 5%, a sample size of 28 subjects in each treatment group was required. This sample size was also adequate to determine a clinically relevant change of score of one point in the MENQOL domains.8Statistical analysis was performed using Statview.13 For each participant the percentage change from baseline in the frequency of hot flushes and night sweats for each of the 12 weeks of the study period, and the absolute difference in scores between Week 12 and baseline for each domain of the MENQOL Questionnaire, were calculated. Repeated analysis of variance was used to analyse the effects of treatment within and between groups over the study period for these outcomes. The Wilcoxon two-sample test was used to analyse the effect of treatment on the excretion of the phytoestrogen metabolites daidzein and genistein. Additional analysis was undertaken to determine the effect of baseline characteristics (age, BMI, duration of amenorrhoea, and previous use of HRT or natural therapies for relief of vasomotor symptoms) on the average percentage change in vasomotor symptoms and on the difference in scores for each domain of the MENQOL Questionnaire. This was performed using analysis of covariance. Continuous variables were categorised into two groups based on the median value (age: < 55 years and ≥ 55 years; BMI: ≤ 25 kg/m2 and > 25 kg/m2; duration of amenorrhoea: < 4 years and ≥4 years). Results Study population Of the 78 subjects who were randomised, 28 in the active group and 27 in the placebo group completed the study (Box 2). Baseline characteristics of those who withdrew and those who completed the study were similar, except for the previous use of natural therapies for menopausal symptoms, which was more frequent in those who withdrew. Most withdrawals were owing to taste intolerance and occurred within the first week after randomisation. There were no significant differences in baseline characteristics between the placebo and CMH treatment groups (Box 3). Effect of intervention on vasomotor symptoms The frequency of vasomotor symptoms was reduced in both CMH and placebo groups (CMH: - 15.0%; 95% CI, - 31.1% to + 1.2%; placebo: - 31.4%; 95% CI, - 41.5% to - 21.2%). The difference between the two groups was not significant (+16.4%; 95% CI, + 35.2% to - 2.4%; P = 0.09). A progressive decline in the frequency of vasomotor symptoms with treatment duration was seen in both groups (CMH, P = 0.001; placebo, P = 0.006). The difference between the two groups was not significant (P = 0.26). Effect of intervention on MENQOL Questionnaire scores A reduction in score was seen in all four domains of the questionnaire with both CMH and placebo. This was significant only for the physical (- 1.14; 95% CI, - 1.78 to - 0.50), vasomotor (- 0.57; 95% CI, - 0.89 to - 0.24) and sexual (- 0.69; 95% CI, - 1.10 to - 0.28) domains in the CMH group, and the physical (- 0.74; 95% CI, - 1.41 to - 0.07) domain in the placebo group. However, the difference between the two treatment groups was not significant for any of the four domains. Effect of baseline characteristics on treatment effect The effects of baseline characteristics are shown in Box 4. Women with more than four years of amenorrhoea had a significantly greater response to placebo than to CMH (Box 4). A significantly greater reduction in score was seen with CMH compared with placebo in the vasomotor domain of the MENQOL Questionnaire for the baseline characteristics of age < 55 years, BMI ≤ 25 kg/m2, and previous non-users of natural therapies (Box 4). There was a significant difference in treatment effect between those who had and those who had not previously used natural therapies for the physical, vasomotor and sexual domain scores. A significant difference in treatment effect was also seen for the two age categories and for the two BMI categories for the vasomotor domain score. Phytoestrogen measurements No significant change was seen in daidzein or genistein excretion with either CMH or placebo, and there was no difference between the two treatment groups. Adverse events The frequency of reported adverse events did not differ between the two groups. Abdominal bloating was reported by three women treated with placebo and one with CMH; two women with CMH reported lower abdominal pain and loose stools. Fifteen women (placebo, 9; CMH, 6) reported headache, joint pain or dizziness. Discussion In our study of extracts of CMH administered as granules reconstituted to a beverage, there was no significant or clinically relevant difference in the frequency of vasomotor symptoms between placebo and CMH therapy. Furthermore, there was no significant or clinically relevant difference in the scores for the four domains of the MENQOL Questionnaire. This study was adequately powered to distinguish at least a 40% reduction in the frequency of vasomotor symptoms, as well as a clinically meaningful reduction in the MENQOL domain scores with treatment. The effect of CMH compared with placebo on the frequency of vasomotor symptoms was not modified by age, BMI, duration of amenorrhoea, previous use of HRT or natural therapies. However, the scores for the physical, vasomotor and sexual domains were modified by previous use of natural therapies: women who had no prior use of natural therapies for their menopausal symptoms responded to therapy, whereas prior users did not. The vasomotor domain was also modified by age and BMI. It is of interest that prior users of natural therapies showed a clinically relevant greater response to placebo than CMH for the physical, vasomotor and sexual domains. That baseline characteristics, particularly prior natural therapy use, significantly modified the response to treatment in this study is an important observation of relevance to future studies of natural therapies. That our findings differ from reports of studies conducted in China4,5 is most likely owing to differences in study design. The Chinese studies have not been placebo-controlled, have employed raw herbs, and have allowed for modification of herbal constituents during the studies according to individual responses.4,5 The preparation of medicinal tea from raw herbs is very time consuming, and was deemed a major obstacle to compliance in a non-Asian study population. The granules used in this study were supplied by a company that routinely prepares herbs in this manner for medicinal purposes in China, and there is no evidence to indicate that the granules do not retain the therapeutic properties of the original herbs. In summary, our study of the effects of CMH versus placebo on vasomotor events and menopausal symptoms in non-Asian Australian women found no overall benefit of the CMH. The modifying effects of prior use of natural therapies is an important positive finding that deserves further investigation and confirmation. Acknowledgements The study was supported by a research grant from the Australasian Menopause Society. Cathay Herbal, Sydney, kindly donated the study preparations. We thank Dr Tikky Wattanapenpaiboon of Monash University for assistance in the evaluation of the food frequency questionnaire, and Associate Professor Flavia Cicuttini of Monash University for valuable input to this study. Dr James C G Doery of Monash Medical Centre offered guidance in testing the study herbal preparations for heavy metal contamination. We also thank all the participants. References Dennerstein L, Smith AM, Morse C, et al. Menopausal symptoms in Australian women. Med J Aust 1993; 159: 232-236. Abraham S, Perz J, Clarkson R, Llewellyn-Jones D. Australian women's perception of hormone replacement therapy over 10 years. Maturitas 1995; 21: 91-95. Waldman TN. Menopause: when hormone replacement therapy is not an option. Part 1 [review]. Women's Health 1998; 7: 559-565. Yao SA. Review on the research and development of Chinese medicine in menopausal syndrome (in Chinese). J Tradit Chin Med 1994; 35: 112-114. Li CJ. Menopausal symptoms. In: Dai DY, editor. Current application and research of Chinese medicine and pharmacology: gynecology. Shanghai: Shanghai University of Traditional Chinese Medicine Publishing House, 1995; 174-182. Wahlqvist M, Kouris-Blazos A, Hsu-Hage B, et al. Food habits and health status of Anglo-Celtic Australians. A questionnaire on food frequency. Melbourne: Monash University. Microsoft Excel 95 [computer program]. Redmond: Microsoft Corporation, 1995. Hilditch JR, Lewis J, Peter A, et al. A menopause specific quality of life questionnaire: development and psychometric properties. Maturitas 1996: 24; 161-175. Dalais FS, Rice GE, Wahlqvist ML, et al. Effects of dietary phytoestrogens in postmenopausal women. Climacteric 1998; 1: 124-129. Murkies AL, Lombard C, Strauss BJ, et al. Dietary flour supplementation decreases postmenopausal hot flushes: effect of soy and wheat. Maturitas 1995; 21: 189-195. Poller L, Thomson JM, Coope J. A double-blind cross-over study of piperazine oestrone sulphate and placebo with coagulation studies. Br J Obstet Gynaecol 1980; 87: 718-725. Hailes JD, Nelson BJ, Schneider M, et al. Conjugated equine oestrogen versus placebo in the management of menopausal symptoms. Med J Aust 1981; 2: 341-342. Statview [computer program]. Abacus Concepts Inc, Berkeley, CA, 1995. (Received 28 Aug, accepted 1 Oct, 2000) Authors' details The Jean Hailes Foundation, Melbourne, VIC. Susan R Davis, FRACP, PhD, Associate Professor and Director of Research; Run Q Chen, MA, Master's Student; Henry G Burger, AO, FRACP, FAA, Consultant Endocrinologist. Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, VIC. Esther M Briganti, MB BS, FRACP, Senior Lecturer; Michael Bailey, MSc, Statistical Consultant. International Health and Development Unit, Monash University, Clayton, VIC. Fabien S Dalais, PhD, Senior Research Officer. Reprints will not be available from the authors. Correspondence: Associate Professor S R Davis, The Jean Hailes Foundation, 173 Carinish Road, Clayton, VIC 3168. suedavisATnetlink.com.au Make a comment 1: The defined formula of the Chinese medicinal herbs Pharmaceutical name Chinese name Dose* Rehmannia glutinosa Cornus officinalis Dioscorea opposita Alisma orientalis Paeonia suffruticosa Poria cocos Citrus reticulata Lycium chinensis Albizzia julibrissin Zizyphus jujuba Eclipta prostrata Ligustrum lucidum Shu Di Huang Shan Zhu Yu Shan Yao Ze Xie Dan Pi Fu Shen Chen Pi Di Gu Pi He Huan Pi Suan Zao Ren Han Lian Cao Nu Zhen Zi 15 10 12 8 8 12 5 20 15 10 15 10 *Dose in grams of dried herb per day. Back to text 2: Flow diagram of participation in the study CMH = Chinese medical herbs. HRT = hormone replacement therapy. Back to text 3: Baseline characteristics of study participants Baseline characteristics Placebo Chinese medicinal herbs P Number Age (years)* Body mass index (kg/m2)* Duration of amenorrhoea (years)* Previous use of hormone replacement therapy Previous use of natural therapies Frequency of hot flushes or night sweats, per week* MENQOL domains Physical domain* Vasomotor domain* Psychosexual domain* Sexual domain* 27 54.1 (52.6, 55.5) 26.1 (24.3, 27.9) 4.6 (3.0, 6.2) 44.4% 37.0% 46.6 (35.4, 57.8) 5.6 (4.9, 6.2) 4.0 (3.3, 4.8) 3.9 (3.3, 4.6) 3.4 (2.5, 4.3) 28 56.3 (54.3, 58.3) 25.7 (23.9, 27.5) 5.8 (3.9, 7.7) 53.6% 35.7% 46.2 (38.75, 53.7) 5.5 (5.2, 6.5) 3.8 (3.1, 4.5) 3.6 (3.0, 4.2) 3.3 (2.4, 4.3) 0.07 0.75 0.34 0.50 0.92 0.94 0.57 0.67 0.45 0.95 *Values are mean (95% confidence limits). Back to text 4: Effect of Chinese medicinal herbs (CMH) compared with placebo on frequency of vasomotor symptoms and MENQOL domain scores, by baseline characteristics* Mean reduction in MENQOL domain score (95% confidence limits) Mean reduction in vasomotor symptoms (95% confidence limits) Physical Vasomotor Psychosexual Sexual Age +8.1% (-16.5%, +32.8%) -0.74 (-1.91, +0.42) -0.94 (-1.74, -0.14) -0.22 (-1.00, +0.57) -0.72 (-2.82, +1.38) ≥55 years +17.9% (-10.7%, +46.5%) -0.36 (-1.76, +1.03) +0.51 (-0.56, +1.58) +0.10 (-0.77, +0.96) -0.14 (-1.30, +1.03) Body mass index ≤25 kg/m2 +7.1% (-21.4%, +35.5%) -1.16 (-2.57, +0.25) -0.85 (-1.61, -0.08) -0.17 (-1.01, +0.68) -1.00 (-2.66, +0.65) >25 kg/m2 +26.2% (-0.48%, +52.9%) +0.28 (-0.92, +1.48) +0.42 (-0.73, +1.58) +0.13 (-0.68, +0.95) +0.18 (-1.46, +1.82) Amenorrhoea < 4 years +4.8% (-24.6%, +34.2%) -0.28 (-1.72, +1.17) -0.31 (-1.67, +1.06) +0.18 (-0.81, +1.16) +0.52 (-1.59, +2.62) ≥4 years +26.8% (+3.8%, +49.9%) -0.49 (-1.70, +0.72) -0.09 (-0.61, +0.42) -0.08 (-0.70, +0.54) -0.96 (-2.08, +0.16) Previous use of hormone replacement therapy No +12.1% (-14.4%, +38.6%) -0.69 (-2.02, +0.65) -0.24 (-1.47, +1.00) +0.03 (-0.74, +0.80) +0.17 (-1.74, +1.40) Yes +22.4% (-6.3%, +51.2%) -0.05 (-1.37, +1.27) -0.16 (-0.75, +0.43) +0.02 (-0.90, +0.95) -0.59 (-2.43, +1.26) Previous use of natural therapies for symptoms of menopause No +11.1% (-11.2%, +33.3%) -1.16 (-2.35, +0.02) -1.09 (-1.68, -0.49) -0.33 (-1.07, +0.41) -1.19 (-2.75, +0.37) Yes +26.1% (-10.0%, +62.3%) +0.93 (-0.38, +2.24) +1.30 (-0.06, +2.67) +0.63 (-0.29, +1.56) +1.15 (-0.33, +2.64) *Values are the percentage or point difference between CMH and placebo (CMH effect minus placebo effect). As vasomotor symptoms and all domain scores improved for both treatment groups, a negative value indicates a greater effect from CMH and a positive value indicates a greater effect from placebo. For shaded values, 95% confidence limits do not include 0. For boxed values, there is a significant difference in treatment effect between the two categories (P Back to text

Susan R Davis · Esther M Briganti · Run Q Chen · Fabien S Dalais · Michael Bailey · Henry G Burger

Research Use of fake tanning lotions in the South Australian population Kerri R Beckmann, Barbara A Kirke, Kieran A McCaul and David M Roder MJA 2001; 174: 75-78 Abstract - Methods - Results - Discussion - Conclusions - References - Authors' Details - - More articles on Public and environmental health Abstract Objective: To explore the relationship between the use of fake tanning lotions and repeated sunburn among South Australian adults, with a view to informing the Anti-Cancer Foundation of South Australia's (ACFSA) policy on fake tanning products. Study design: Population survey. Participants: 2005 South Australians aged 18 years or older, selected randomly from the electronic White Pages. Main outcome measures: Self-reported use of fake tanning lotions in the past 12 months; frequency of sunburn over summer; and various sun-protective behaviours. Results: 2005 of the 2536 eligible participants (79%) were surveyed by telephone. Fake tan use was most prevalent among women (15.9%), people aged 18-24 years (15.4%), and people with household incomes above $40 000 per year (11.9%). Fake tan users were more likely than non-users to use sunscreens (81.3% v 56.5%; P < 0.001), but less likely to take other precautions such as wearing hats (40.9% v 51.0%; P = 0.04) and protective clothing (22.3% v 34.1%; P = 0.005). They were also more likely to report having been burnt more than once over summer (26.2% v 16.5%; P = 0.025). Multivariate analysis indicates a statistically significant association between fake tan use and repeated sunburn (odds ratio, 2.07; 95% confidence interval, 1.17-3.69), which was independent of age, sex, skin type and sun-protection practices. Conclusion: Users of fake tanning products may be at greater risk of repeated sunburn. The ACFSA sees no justification at this stage for altering its present policy position of not actively promoting the use of fake tanning lotions as a means of reducing sunburn. Anticancer organisations in Australia have been conducting programs aimed at reducing Australia's high rate of skin cancer for over two decades.1 The main objective of these programs is to encourage people to reduce their exposure to solar ultraviolet radiation, the major contributing factor to the development of skin cancer.2In Australia public awareness about the dangers of overexposure to the sun is generally high. In spite of this, a suntan is still desired by some sectors of the community -- in particular, young, fashion-conscious people.3,4 Skin cancer prevention programs have attempted to change attitudes that value tanned skin as attractive and healthy with such messages as "there is no such thing as a safe tan" and "a tan is a sign of skin damage". Last year, Chapman challenged anticancer organisations to consider the role that fake tanning lotions might play in reducing sun exposure, suggesting that they should be assessed as a potential harm-reduction strategy.5 The Anti-Cancer Foundation of South Australia (ACFSA) has, for a number of years, provided information on fake tanning lotions. While not actively encouraging their use, the information suggests that, for those desiring a tan, using fake tanning lotions is preferable to exposure to artificial or solar ultraviolet radiation. In October 1999, the ACFSA included a question on the use of fake tanning lotions in a Health Monitor Survey along with questions on skin type, experience of sunburn and frequency of wearing hats, cover-up clothing, applying sunscreen and seeking shade. This article reports the findings of that survey and discusses them in relation to the position taken by the ACFSA regarding fake tanning lotions. Methods Questions relating to sun exposure and ultraviolet radiation protective behaviours, including one relating to the use of fake tanning lotions, were asked of a random sample of South Australians, by computer-assisted telephone interviewing. These questions (Box 1) were part of a larger health-related survey organised and conducted by the Department of Human Services, South Australia, in October 1999. Except for the question relating to fake tan use, these questions have been used routinely in monitoring sun-protection behaviours in South Australia and were originally validated as written questions in a national Secondary School Children's Survey conducted triennially since 1990.6 The question on fake tan use is a slightly modified version of a question asked in Victorian surveys in 1993 and 1995.7,8Ethical approval for the survey that incorporated the questions used in this study was obtained through the Department of Human Services, with legal authorisation under section 64d of the South Australian Health Commission Act (1976). A sample of 3400 residences from rural and metropolitan areas within South Australia was drawn from the electronic White Pages. One adult from each household (the person whose birthday was the most recent) was invited to participate. Two thousand and five interviews were conducted from the 2536 households that could be contacted after six callback attempts, giving a participation rate of 79.1%. All data were weighted by age, sex and region, and on the probability of selection within the household. The population profile for weighting was obtained from the Australian Bureau of Statistics' estimated population for South Australia, 1997. Geographical region was defined as either metropolitan or country region. Both descriptive analysis of the survey data and logistic regression modelling were undertaken using STATA version 6,9 as this software allows calculation of robust estimates of standard error using methods devised by Huber10 and White.11 Consequently, variance estimates are adjusted for the data weighting. The relationship between fake tan use and reported sunburn over summer was examined using logistic regression modelling, allowing adjustment for age, sex, skin type and sun-protective behaviours. We first constructed a model containing known risk factors for sunburn and then added fake tan use to this model to establish if this improved the fit of the model. "Having been burnt two or more times during the previous summer" was the dependent (outcome) variable. For each of the sun protective behaviour questions, respondents were coded as regular users if they indicated that they "usually, almost always, or always" took such precautions when out in the sun for an hour or more, and were coded as irregular users if they indicated that they "never, rarely or sometimes" took these measures. Results Based on results from this survey, the estimated prevalence of fake tanning lotion use during the past 12 months among South Australians aged 18 years or more was 8.7% (95% confidence interval [CI], 7.3%-10.5%). The prevalence of fake tan use among various subgroups of the population is shown in Box 2. The use of fake tanning lotions is most common among younger people, particularly women, with the peak prevalence being 28% among young women aged 18-24 years. Fake tan use is also more common among those who report that their skin burns before tanning, compared with those whose skin just tans or just burns. Fake tan use also appears to be related to household income, with those with relatively high household incomes (above $40 000 per year) more likely to use fake tanning lotions. Individuals who had used fake tanning lotions in the past year were more likely to report regularly using sunscreen with a sun protection factor (SPF) of 15+ or higher when in the sun than non-users (81% v 57%; P < 0.001). However, they were less likely to report regularly wearing hats (41% v 51%; P = 0.04) or protective clothing (22% v 34%; P = 0.005). They reported seeking shade at levels similar to those who had not used fake tanning lotions (80% v 76%, P = 0.4). Those who had used fake tanning lotions were also more likely to report having been burnt two or more times during the previous summer (26% v 17%; P = 0.025) (Box 3). Factors such as age, skin type, sex and regular sun-protective behaviours are likely to confound the association between fake tan use and risk of burning. Results of logistic regression modelling, which takes into account the effects of these potential confounders, indicate an increased risk among fake tan users of having been sunburnt more than once (odds ratio [OR], 2.07; 95% CI, 1.17-3.69), as shown in Box 4. As the use of fake tanning lotions was much more prevalent among women than men, we also undertook regression analyses for women and men separately. No association between fake tan use and sunburn was found among men (OR, 0.90; 95% CI, 0.14-5.97). This was most probably owing to the fact that only 12 men reported using fake tanning lotions. There was, however, a strong association between using fake tanning lotions and repeated sunburn among women (OR, 2.47; 95% CI, 1.38-4.42). Discussion While the overall prevalence of fake tan use among adult South Australians is low (8.7%; 95% CI, 7.3%-10.5%), the use of fake tanning lotions is fairly common in younger women, with more than one in four women aged 18-24 years reporting having used fake tanning lotions in the past year. These findings are consistent with the reported prevalence of use in Victoria.7,8Respondents who reported using fake tanning lotions were more likely to report regularly using SPF 15+ or higher sunscreen when out in the sun during summer, but were less likely to report wearing hats or protective clothing. Fake tan users were more likely to report being sunburnt two or more times over the past summer. When other known risk factors were taken into account, fake tan users had twice the risk of repeated sunburn over summer compared with non-users. The only previously reported findings in relation to the association between fake tan use and sunburn are from two surveys conducted in Victoria, one in 19937 and one in 1995.8 The first of these surveys found a higher prevalence of sunburn among fake tan users (66% v 46%), while the latter survey found no difference (39% v 40%). The inconsistency of these two reports may have been owing to the relatively small sample size of each survey (n < 700). Owing to the limited nature of the questions in this survey, we were unable to determine whether fake tanning lotions were used just at the start of the season to give a tanned look before a sun-induced tan could be achieved, or throughout the summer as a substitute for sunbathing. Given the timing of the survey (ie, spring), there may be some inaccuracy in people's recall of sunburn in the previous summer. However, it seems unlikely to us that one group would have been more or less likely to under-report having been sunburnt, so any recall effect would have been equivalent in both groups. Another limitation in relation to the timing and cross-sectional nature of the survey is the inability to establish a temporal relationship. In some cases, sunburn may have preceded the use of fake tanning lotions. We can not conclude that fake tan use contributes directly to an increased risk of sunburn. We can only suggest that the behaviours of fake tan use and sun exposure may be linked. A further limitation of this study is that we did not ask about the reason for or frequency of use. We do not know whether there are differences in the risk of sunburn among those who use fake tanning lotions only on special occasions (eg, theatrical performances) compared with those who use such products regularly to maintain a tanned appearance. This lack of detail does not negate the finding that, as a whole, those who use fake tanning lotions are at greater risk of sunburn. Regardless of when and why people use fake tanning lotions, the results of this survey do not offer any evidence that use of fake tanning lotions, as currently practised, protects against sunburn. However, since this is an observational study, we can not rule out the possibility that the use of fake tanning lotions may actually offer some protection. It is conceivable that, had users not been applying fake tanning lotions, sunburn levels could have been even higher in this group. Further clarification of this issue would require a longitudinal (experimental) study design. Our results suggest that, rather than reducing their sun exposure, fake tan users are more likely to be exposing their skin to damaging levels of ultraviolet radiation than non-users. The evidence also suggests that, in general, fake tan users take fewer precautions to protect their skin from the sun. While fake tan users are more likely to report using sunscreens, they appear to rely on sunscreens alone for sun protection rather than using multiple strategies, as recommended by the Anti-Cancer Foundation. Some fake tan users may believe that the tanned effect provided by fake tanning lotions offers protection against the sun. Further confusion may arise in cases where their chosen brand of tanning lotion contains sunscreen. An inspection of fake tanning lotions currently available in South Australia showed that most brands do not include a sunscreen, and many state on the label that the fake tan does not provide protection against solar ultraviolet radiation. Some brands also include the advice to use a regular SPF 30+ sunscreen when going into the sun. However, a few brands of fake tanning lotions do contain sunscreen, varying in their sun protection factor from 4 to 15. One commonly available product with an SPF 4 rating states on its label, "UV Protection: Protects you in the sun, providing 4 times your natural sunburn protection". While technically correct, the protection would apply only to the period immediately after application and not for the time that the tan remains visible on the skin. Such claims are obviously very misleading. Anticancer organisations advise that sunscreens need to be reapplied regularly, ideally two-hourly, to maintain adequate protection. If such advice was followed when using a fake tanning lotion containing a sunscreen, the colour of the tan would deepen with each application. Also, it may take up to four hours for the tan colour to fully develop. Most products recommend removal of dry or flaky skin before applying the fake tanning lotion. The need to do this and the effect of repeated applications on tan colour are likely to preclude regular reapplication of fake tanning lotions, thereby increasing the likelihood of users risking sunburn if they rely on the protection offered by one application. In our view, including a sunscreen in a fake tanning lotion offers no obvious benefit. On the contrary, it has the potential to generate a false sense of protection which may lead to sunburn in fake tan users. Conclusions In response to the discussion posed by Chapman5 in relation to the use of fake tanning lotions as a "harm minimisation" approach, the ACFSA was prompted to review its policy in relation to the promotion and sale of fake tanning lotions. The results of this study do not point to a reduced risk of harmful sun exposure among fake tan users. Rather, they suggest an elevated risk of sunburn. In the light of these findings, the ACFSA sees no justification for altering its current position on the use of fake tanning lotions. The use of fake tanners is not actively promoted by the ACFSA. However, where there is a strong desire for a tan, people are advised that the use of fake tanning lotions is a better alternative than sunbathing or using a solarium. They are also advised that a fake tan does not provide protection against the sun and are warned about the limited protection offered by products that contain a sunscreen. More in-depth investigation of why and when fake tanning lotions are used, and the extent to which fake tan users believe they are protected from the harmful effects of the sun while using such products, is needed to inform education strategies and guide any policy change by organisations such as the ACFSA. The potential of the labelling of fake tan products containing sunscreens to be misleading needs to be brought to the attention of the relevant authorities. References Gray N. Report of the Chairman of the Education Committee. Australian Cancer Society Annual Report, 1979. Sydney: ACS, 1979: 9. Armstrong BK. Stratospheric ozone and health. Int J Epidemiol 1994; 23: 873-885. Arthey S, Clarke V. Suntanning and Sun protection: A review of the psychological literature. Soc Sci Med 1995; 40 (2): 265-274. Clarke V, Williams T, Arthey S. Skin type and optimistic bias in relation to the sun protection and suntanning behaviors of young adults. J Behav Med 1997; 20: 207-222. Chapman S. Faking it: should cancer control agencies promote fake tanning lotions? Med J Aust 1999; 170: 603-604. Broadstock M, Borland R, Hill D. Knowledge, attitudes and reported behaviours relevant to sun protection and suntanning in adolescents. Psychol Health 1996; 11: 527-539. Purchase M, Borland R. Public reaction to the 1992/93 SunSmart campaign: results from a representative survey of Victorians. SunSmart Evaluation Studies 3. Melbourne: Anti-Cancer Council of Victoria, 1994: 93. Dixon H, Cappiello M, Borland R. Reaction to the 1994/95 SunSmart campaign: results from a representative household survey of Victorians. SunSmart Evaluation Studies 5. Melbourne: Anti-Cancer Council of Victoria, 1997: 64. Stata version 6 [computer program]. College Station, TX: Stata Corporation, 1999. Huber PJ. The behavior of maximum likelihood estimates under non-standard conditions. Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability 1967; 1: 221-233. White H. A heteroskedasticity-consistent covariance matrix estimator and direct test for heteroskedasticity. Econometrica 1980; 48: 817-830. (Received 21 Mar, accepted 31 Jul, 2000) Authors' Details Anti-Cancer Foundation of South Australia, Adelaide, SA. Kerri R Beckmann, BSc(Hons), MPH, Program Evaluation Officer; Barbara A Kirke, DipN, MPHC, Skin Cancer Prevention Project Officer. Collaborative Research Centre for Asthma, University Department of Medicine, Sir Charles Gairdner Hospital, Perth, WA. Kieran A McCaul, BSc, MPH, Biostatistician. Epidemiology Branch, South Australian Department of Human Services, Adelaide, SA. David M Roder, AM, MPH, DDSc, Director. Reprints will not be available from the authors. Correspondence: Ms K R Beckmann, Anti-Cancer Foundation of South Australia, PO Box 929, Unley, SA 5061. kbeckmannATcancersa.org.au Make a comment Back to text Back to text Back to text 4: Factors associated with being burnt more than once over summer All respondents Women only Adjusted Adjusted odds ratio 95% CI P odds ratio 95% CI P Fake tan use Non-users Users 1.00 2.07 - 1.17-3.69 0.013 2.47 1.38-4.42 0.002 Sex Female Male 1.00 2.99 - 2.09-4.29 - - - Age (group years) 65+ 18-24 25-34 35-44 45-54 55-64 1.00 13.70 5.83 4.72 2.46 1.27 - 7.00-26.78 3.19-10.65 2.66-8.38 1.34-4.53 0.61-2.66 0.004 0.529 15.46 4.40 5.12 3.56 0.40 5.34-44.77 1.71-11.34 2.04-12.86 1.37-9.23 0.09-1.65 0.002 0.001 0.009 0.203 Skin type Just tan Burn then tan Just burn 1.00 2.07 2.47 - 1.20-3.56 1.47-4.16 0.009 0.001 1.86 2.12 0.69-5.01 0.83-5.39 0.222 0.117 Sunscreen use Irregular Regular 1.00 1.03 - 0.72-1.48 0.876 0.82 0.44-1.49 0.509 Hat wearing Irregular Regular 1.00 0.69 0.48-0.98 0.038 0.91 0.52-1.59 0.746 Protective clothing Irregular Regular 1.00 0.63 - | 0.44-0.91 0.014 0.95 0.56-1.61 0.856 Shade seeking Irregular Regular 1.00 0.91 - 0.62-1.32 0.604 0.55 0.30-1.01 0.056 *Logistic regression modelling using forced entry of all variables. Separate models for all respondents and women only. Back to text

Kerri R Beckmann · Barbara A Kirke · Kieran A McCaul · David M Roder

The Research Enterprise Australian medical patents granted in the United States in 1984-1999 Eugen Mattes and Michael C Stacey MJA 2001; 174: 83-87 Abstract - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - More articles on Law Abstract Objective: To describe all medical patents granted in the United States to Australian-resident inventors between 1984 and 1999. Data sources: All patent data originated from the US Patent and Trademark Office. Data for 1984-1994 were compiled by CHI Research Inc, and data for 1995-1999 were obtained from the Community of Science website. Main outcome measures: Number of medical patents granted in the US to Australian-resident inventors; assignees (owners) of these medical patents; proportion of these medical patents related to biotechnology. Results: From 1984 to 1999, 7835 utility patents were granted in the US to Australian-resident inventors. Of these, 1308 patents (17%) were identified as medical patents; 489 (37%) of these were biotechnology patents. Medical patents account for an increasing proportion of all US patents granted to Australian inventors, increasing from 10% in 1984 to 25% in 1999. Biotechnology accounted for an increasing proportion of medical patents, rising from 10% to 55% between 1984 and 1999. More than half the medical patents are owned by commercial interests, and 33% by only 14 organisations, six of which are universities and their affiliated institutions. Conclusion: Only a few organisations account for most of the patenting of medical technology. The inventors and their organisations listed on medical patents could be canvassed when developing government policy and targeted for support in commercialising their medical technology. Interest in harnessing the economic value of medical technology conceived and created in Australia has been growing.1-3 This is reflected in the creation of Cooperative Research Centres (CRCs). Since 1991, 67 CRCs have been created, of which 10 have a medical focus.4 Despite these initiatives, medical inventors, unlike sports stars, are virtually unrecognised by the Australian public and scientific community. Where do our new technologies in medicine come from -- industry, universities, or the lone inventor? We chose to study all patents for inventions (utility patents) granted in the United States to Australian residents. These patents are economically more significant than patents granted in Australia, as the increased cost and effort of patenting in another country is thought to filter out trivial inventions.5 Because of the importance of the US market, the US is also the first country where multinational corporations submit their patent applications outside the patent's country of origin.6 Thus, Australian patents in the US are arguably the most important subset of Australian patents in other countries.7 Our aim was to describe Australian medical patents and to compare them with non-medical patents. Methods This is a descriptive study of patents granted in the US from 1984 to 1999 to inventors resident in Australia. Utility patents were examined, with design and plant patents excluded. Sources of data All patent data originated from the US Patent and Trademark Office. US patents listing one or more Australian-resident inventors for 1984-1994 were compiled by Computer Horizons Incorporated (CHI) Research Inc8 of the US and, for 1995-1999, were updated from the Community of Science website.9 Our electronic patent database contained the following information for each patent: year of patent being granted; US patent registration number; title of patent; all listed inventors and assignees and their country of residence; and number of citations of scientific literature. When required for classifying patents, the patent abstracts or full patents were examined on the Internet (using the US patent registration number) at either the US Patent and Trademark Office Web Patent Database Centre,10 or the Delphion (formerly IBM) Intellectual Property Network.11 Categorisation of medical patents As there are no published guidelines for selecting medical patents, we defined a medical patent as any technology used for: managing patients and their illnesses, such as drugs, diagnostic tests, surgical instruments, and rehabilitation devices (dental technologies were excluded); preventing illness, such as sunscreen lotion; or medical research, such as laboratory instruments. Generic technologies used in other fields, such as information technology, were excluded. Patents were classified as medical, possibly medical or non-medical after reading the title and the name of the assignees (owners). For all patents labelled as possibly medical, we read the abstract, and if necessary the complete patent, to properly classify the patent. Medical patents related to biotechnology12 were identified separately, and included devices, processes, DNA sequences, transgenic animals and manufacturing processes in the medical industry. Biotechnology patents related to other industries, such as agriculture, mining and food processing, were excluded. Describing inventors and assignees The inventors and assignees on patents were sorted alphabetically in Microsoft Excel,13 and any errors or differences in spelling were corrected. The median number of inventors and assignees per patent was calculated as a measure of collaboration. The assignees on each patent were categorised as a business, university, government, research institute, CRC, non-government organisation, technology transfer office, or individual. Categorising was usually straightforward using the name of the assignee, but, if there was uncertainty, a search was made on the Internet using the search engine Dogpile.14 Data analysis The data were stored, tabulated and graphed in Microsoft Excel,13 and statistical analysis was conducted using SPSS for Windows.15 The χ2 test or Fisher's exact test was used to compare independent proportions. The Mann-Whitney U test was used to compare medians. Ethical issues All the information in this study, including the names of inventors and companies, is publicly available on numerous patent bibliographic databases. Results From 1984 to 1999, 7835 utility patents were granted in the US to Australian-resident inventors. From examination of the title and assignee names of these patents, 9% (673/7835) were classified as medical and 35% (2767/7835) as possibly medical. The abstracts of all patents classified as possibly medical were examined, and 11% (869/7835) required the full patent to be read. In total, 1308 (17%) Australian patents in the US were classified as medical. Of these, 489 (37%) were biotechnology patents. Trends in medical patenting The annual number of patents granted in the US to Australian-resident inventors in 1984-1999 more than doubled, rising from 310 to 800 (Box 1). During this 16-year period, the proportion of medical patents rose from 10% (30/310) to 25% (202/800). Biotechnology accounted for an increasing proportion of medical patents, rising from 10% (3/30) to 55% (112/202) over the same period. Comparison of medical and non-medical patents In terms of inventors, medical patents were: more likely to have multiple inventors listed, with a median of two inventors per medical patent versus one per non-medical patent (P < 0.001, Box 2); and twice as likely to be part of an international collaboration, with co-inventors who are residents of other countries in 21% (275/1308) of medical and 10% (649/6527) of non-medical patents (P < 0.001, Box 3). Most US patents listing Australian inventors have either an Australian inventor or assignee owning the patent (68% for medical and 81% for non-medical patents) (Box 3). The technology most likely to arise from another country is that owned by an assignee in another country and listing an inventor from another country. Thus, 15% of medical and 7% of non-medical patents in our study may have originated outside Australia (Box 3). Patents are either assigned, usually to an organisation, or unassigned (thus owned by the inventor). We found 82% of medical patents were assigned, compared with 69% of non-medical patents (P < 0.001, Box 2). For assigned patents, both medical and non-medical patents usually have one assignee (Box 2). For these assigned patents, there were three large differences, with medical patents being (Box 4): less likely to be owned by a business; four times more likely to be owned by a university; and 40 times more likely to be owned by a research institute. Both medical and non-medical patents are increasingly owned by business and universities, with fewer being unassigned. From 1984 to 1999, patents assigned to business increased from 49% to 63%; patents assigned to universities increased from 2% to 7%; and unassigned patents decreased from 36% to 20% of all patents. Medical patents were three times as likely as non-medical patents to quote from published scientific articles. Sixty per cent (785/1308) of medical patents cited one or more scientific publications, compared with only 23% (1475/6527) of non-medical patents (P < 0.001). Characteristics of medical patents Most of the 1785 medical inventors are not prolific, with 67% (1200/1785) listed only once in 1984-1999. About 18% (318/1785) of medical inventors are listed on three or more medical patents. However, the 17 most prolific medical inventors (Box 5) were responsible for 13% (169/1308) of Australian medical patents. Eleven of these prolific inventors are clustered around four different technologies: electromedical devices (cardiac pacemakers and cochlear ear implants), biosensors, ribozymes, and the relaxin gene. Just 14 organisations own 33% (438/1308) of medical patents; six of these organisations are Australian universities and their affiliated institutions (Box 6). Surprisingly, 20% (264/1308) of medical patents were owned by just five organisations: the University of Melbourne, Telectronics, the Commonwealth Scientific and Industrial Research Organisation (CSIRO), Biotech Australia, and the University of New South Wales. The three most common types of medical technologies are cardiac pacemakers (7%), syringes or parenteral drug delivery technology (4%), and cochlear ear implants (3%). Discussion Principal findings We believe that this is the first published report that has examined Australian medical patents in detail. Australia, like all countries of the Organisation for Economic Co-operation and Development (OECD), has shown a strong rise in the number of US patents granted per capita since the 1960s.5,16 However, Australia is still ranked a low 16 of 20 OECD countries,7 with only a modest increase in the proportion of utility patents granted to Australian inventors in the US during 1984-1999 (from 0.46% [309/67200] to 0.52% [800/153492]16). As expected, Australia's comparative technological advantage is found mainly in mining and agriculture. This is similar to other resource-abundant OECD countries like Canada, Finland and Norway.5,7 However, our study, along with others,7,17 indicates a shift in Australia to patenting in higher technologies. Our study suggests that medical technology, especially medical biotechnology, is an increasingly important part of Australia's intellectual property portfolio. This may explain why Australia appears to be developing a technological advantage in biotechnology and pharmaceuticals.17 To place this trend in a global perspective, our study would need to be repeated for other, especially OECD, countries. There is a dip in the total number of US patents granted to Australian inventors from 1990 to 1993 (Box 1), possibly reflecting the economic recession at the time. Interestingly, the trend in medical and medical biotechnology patents did not show this decrease, suggesting that development of such technology may be more resistant to downturns in the economy. Another feature was the 51% jump in the number of US patents granted to Australian inventors from 1997 to 1998 -- possibly a flow-on from the 54% increase in the number of patent applications filed by Australians in the US between 1994 and 199818 (noting that it usually takes two years from lodging a patent application until it is granted19). This trend coincides with increased research and development spending in Australia, particularly by business (which unfortunately declined in 1996-97).20 However, these changes may also reflect increased processing of patent applications by the US Patent and Trademark Office7 -- the overall number of utility patents issued jumping 32% from 111 983 in 1997 to 147 520 in 1998.16 Our study lends support to recent findings of the importance of university-based research in underpinning high-technology patents and industries.1,21 Universities and their affiliated institutions: make up more than a third of the most prolific patenting organisations; own an increasing proportion of US patents granted to Australian inventors; and are the source of 97% of the scientific articles cited in Australian medical patents.17 These findings could be the result of Australian governments actively encouraging universities to fund and commercialise research and develop links to industry. The CRCs were part of such initiatives, but the fact that only eight patents are owned by CRCs suggests that they are not very productive in commercialising research. However, this is difficult to judge, as the patents may be assigned to commercial or university partners. Our data also support the emerging ideas on the importance of clusters of co-located industries and universities where collaboration and competition act as constant spurs to innovation, such as is seen in Silicon Valley in California.21 In Australia, such clusters appear to be growing in Melbourne and Sydney for industries in biotechnology and electromedical devices. This is demonstrated by examining the prolific inventors and their assignees, indicating varied links between industry and publicly funded institutions. Strengths and weaknesses of the study Possible weaknesses in our study relate to three areas of potential misclassification in our patent data. Classification of country of origin: We classified patents as Australian if any inventor was an Australian resident. This may result in the inclusion of technology originating in another country but which had an Australian inventor working on it (estimated to be about 15% for medical and 7% for non-medical patents). Definition of medical patents: Given the absence of any published guidelines, it could be debated whether certain technologies are really "medical" (such as those related to optometry and sunscreen lotions) and whether dental technologies should have been excluded. Classification of assignees as "business": Assignees with a business-type suffix (ie, "Pty Ltd", "Ltd", "Corp", "Inc", "NV", "AG" and "GMBH"), unless detected through searches on the Internet as belonging to another category such as a technology transfer organisation, would have been misclassified as a business. It is difficult to predict whether the first two potential biases could alter our conclusions. The third may lead to an overestimation of the number of patents owned by business. When identifying the country of origin of a patent, the main convention is to use the residency of the inventor16,17 rather than assignees, partly because a large proportion of patents are unassigned (29% in our study). Like other technology or innovation indicators, patent statistics have advantages and disadvantages.22,23 Our study treats all patents as being of equal importance; however, a patent's commercial value can vary enormously.23,24 Furthermore, patent data do not capture all new technology, as some may not be patentable, and patenting can vary with economic conditions and with the strategic concerns of companies.23,24 For example, patenting as a means of protecting intellectual property is very important for the pharmaceutical industry but of little relevance to the rubber industry.24 In addition, the difficulty when describing patents owned by business is that many are granted under the names of subsidiaries and divisions that are different from the names of parent companies. Some companies even actively hide emerging technologies under different company names, so-called "submarine" patents.5,25 Possible mechanisms and implications for policymakers Australia has a substantial and growing trade deficit in high-technology goods,21 making it more imperative to capture more of the economic value of Australian medical patents. But how? First, the more prolific medical inventors and their organisations could be canvassed when developing government policy which may impact on the commercialisation of medical technology. Such surveys could also identify emerging technologies, which may be the basis of new industries, enabling government to take an anticipatory stance on industry policy. Second, the medical inventors and assignees could be actively targeted with assistance in developing their medical technology. To foster the growing culture of enterprise and innovation within academia,4 it may be worth considering a reward for the prolific inventors and assignees. Such rewards may encourage other inventors and promote inventors as role models for other scientists. Acknowledgements We are grateful to Associate Professors Sam Garrett-Jones and Tim Turpin from the Centre for Research Policy at the University of Wollongong for providing the database of Australian patents in the US for 1984-1994, and to Ms Christine Porter, Manager of the European and Commonwealth Office, Community of Science, for providing free access to the US patents on their website. We would like to thank Dr Dora Marinova from ISTP at Murdoch University and Professor Jane Marceau, Pro Vice Chancellor (Research) at the University of Western Sydney Macarthur, for their valuable advice. Eugen Mattes was the recipient of an Eva K A Nelson Medical Research Scholarship from the University of Western Australia from 1995 to 1998 and an advanced academic registrar funded by the Royal Australian College of General Practitioners in 1999 and the Registrar Scholarship and Research Fund of the College in 2000. We also want to acknowledge the support of Professor Max Kamien and the Department of General Practice, University of Western Australia. We would also like to thank the reviewers for their helpful comments. References Wills PJ (Chairman). Health and Medical Research Strategic Review. The virtuous cycle: working together for health and medical research. Canberra: Department of Health and Aged Care, 1999. The National Innovation Summit. <http://www.isr. gov.au/industry/summit/index.html> Accessed 24 February 2000. Biotechnology Australia. Developing Australia's biotechnology future. Discussion Paper. Canberra: Commonwealth of Australia, 1999. Mercer D, Stocker J (Steering Committee). Review of greater commercialisation and self funding in the Cooperative Research Centres Programme. Canberra: Department of Industry, Science and Tourism, 1998. Patel P, Pavitt K. Australia's technological capabilities: an analysis using US patenting statistics. Brighton: Science Policy Research Unit, University of Sussex, 1995. Bertin G, Wyatt S. Multinationals and industrial property: the control of the world's technology. Hemel Hempstead, Hertfordshire, England: Harvester-Wheatsheaf, 1988. Department of Industry, Science and Technology (DIST). Australian business innovation: a strategic analysis. Report No. 5. Canberra: AGPS, 1996. CHI Research Inc. <http://www.chiresearch.com/> Accessed 7 February 2000. Community of Science. US patents. <http://patents.cos.com/> Accessed 10 July 2000. US Patent and Trademark Office. USPTO Web Patent Database. <http://www.uspto.gov/patft/index.html> Accessed 7 February 2000. Intellectual Property Network. <http://www. delphion.com/home> Accessed 7 December 2000. National Science and Technology Council. Biotechnology for the 21st Century: New Horizons. Washington: USGPO, 1995. Available at <http://www.nal.usda.gov/bic/bio21>. Microsoft Excel 97 SR-2 [computer program]. Cambridge, Massachusetts: Microsoft, 1997. Dogpile. <http://www.dogpile.com/> Accessed 7 February 2000. SPSS for Windows [computer program]. Version 8.0. Chicago, Illinois: SPSS Inc, 1997. Patent counts by country/state and year. Utility patents. January 1, 1963 - December 31, 1999. Technology Assessment and Forecast (TAF) Program, Office for Patent and Trademark Information, US Patent and Trademark Office. <http://www.uspto.gov/web/offices/ac/ido/oeip/taf/ cst_utl.pdf> Accessed 10 July 2000. Narin F, Albert M, Kroll P, Hicks D. Inventing our future: the link between Australian patenting and basic science. <http://www.arc.gov.au/ publications/arc_pubs/00_02.pdf> Accessed 30 October 2000. Number of utility patent applications filed in the United States, by country of origin, calendar years 1965 to present. Technology Assessment and Forecast (TAF) Program, Office for Patent and Trademark Information, US Patent and Trademark Office. <http://www.uspto.gov/web/offices/ac/ido/oeip/taf/ appl_yr.pdf> Accessed 30 October 2000. Trilateral Statistical Report 1997. US Patent and Trademark Office. <http://www.uspto.gov/web/offices/ dcom/olia/trilat/tsr97/> Accessed 10 July 2000. Science and Technology Policy Branch of the Department of Industry, Science and Resources. Australian science and technology at a glance 2000. <http://www.isr.gov.au/science/analysis/glance2000/> Accessed 30 October 2000. Marceau J, Manley K, Sicklen D. The high road or the low road? Alternatives for Australia's future. Sydney: Australian Business Foundation; 1997. Patel K, Pavitt K. Paterns of technological activity: their measurement and interpretation. In: Stoneman P, editor. Handbook of the economics of innovations and technical change. Oxford: Blackwell, 1995; 14-51. Industry Analysis Branch of the Department of Industry, Science and Resources. Measuring the knowledge-based economy. How does Australia compare? Canberra: Commonwealth of Australia, 1999. Geroski P. Markets for technology: knowledge, innovation and appropriability. In: Stoneman P, editor. Handbook of the economics of innovation and technical change. Oxford: Blackwell, 1995; 90-131. Garrett-Jones S, Aylward D. Measuring linkages between basic scientific research and Australian industrial technologies using patent data. Wollongong: Centre for Research Policy, University of Wollongong, 1995. (Received 14 Jul, accepted 2 Nov, 2000) Authors' details University of Western Australia, Perth, WA. Eugen Mattes, MB BS, MPH, Advanced Academic Registrar, Department of General Practice, and PhD Scholar, Department of Surgery, Fremantle Hospital; Michael C Stacey, DS, FRACS, Associate Professor, Department of Surgery. Reprints will not be available from the authors. Correspondence: Dr E Mattes, Visiting Research Fellow, TVW Institute for Child Health Research, Division of Population Sciences, 100 Roberts Road, Subiaco, WA 6008. emattesATcyllene.uwa.edu.au Make a comment 1: Number of utility patents granted in the United States to Australian-resident inventors from 1984 to 1999 Back to text 2: The number of medical and non-medical patents granted in the United States to Australian-resident inventors between 1984 and 1999, and numbers of inventors and assignees (owners) Medical Non-medical Total Patents 1308 6527 7835 Inventors Number listed Number of individuals Inventors/patent (median)* 3270 1785 2 11127 7092§ 1 14397 8744¶ Assigned patents† 1068 (82%) 4533 (69%) 5601 Assignees Number listed Number of individual assignees‡ Assignees/assigned patent (median)* 1192 448 1 4909 2309 1 6101 2701¶ * The medians were significantly different (Mann-Whitney U test; P < 0.003). For assignees this is unlikely to be of practical significance. † Medical and non-medical groups are significantly different (Pearson χ2, P < 0.001). ‡After correcting errors or differences in spelling (16% of medical and 9% of non-medical assignees were either misspelt or spelt differently). § Calculated using ratio from medical patents: there were 1785 individual medical inventors after correcting the spelling of the initial list of 1878 individuals. So, for non-medical inventors, it was estimated that there were 7092=7461 x (1785/1878) individuals, assuming a similar 5% difference in spelling of the 7461 non-medical inventors listed initially. ¶Not equal to the sum of inventors or assignees on medical and non-medical patents, as 133 inventors and 56 assignees are on both types of patents. Back to text 3: The location of inventors and assignees for United States patents listing Australian inventors for 1984-1999 Location of inventors Location of assignees Medical Non-medical Australia Unassigned patents* Australia Australia and other countries Other countries* Subtotal 226 (17%) 660 (50%) 8 (1%) 139 (11%) 1035 (79%) 1942 (30%) 3357 (51%) 74 (1%) 505 (8%) 5878 (90%) Australia and other countries (international collaboration) Unassigned patents Australia* Australia and other countries* Other countries* Subtotal 14 (1%) 48 (4%) 22 (2%) 191 (15%) 275 (21%) 52 (1%) 114 (2%) 40 (1%) 443 (7%) 649 (10%) Total 1308 (100%) 6527 (100%) Global Pearson χ2 is significant (P <0.001.). * Significant difference between medical and non-medical patents (P <0.001). Back to text 4: Classifications of assignees listed on the 1068 medical and 4533 non-medical assigned US patents listing Australian-resident inventors for 1984-1999 Assignee Medical patents Non-medical patents Business* University* Research institute* Government* CSIRO Individual* Technology transfer organisation Non government organisation† Cooperative research centre Total 700 (59%) 206 (17%) 97 (8%) 70 (6%) 60 (5%) 30 (2.5%) 22 (1.8%) 6 (0.5%) 1 (0.1%) 1192 (100%) 3913 (80%) 181 (4%) 8 (0.2%) 195 (4%) 272 (6%) 257 (5%) 69 (1%) 7 (0.1%) 7 (0.1%) 4909 (100%) Global Pearson χ2 is significant (P <0.001). Pairwise comparisons were done using Pearson χ2, except where Fisher's exact test was needed when an expected count was less than 5. *Significant difference between medical and non-medical patents (P <0.004). †Significant difference between medical and non-medical patents (P=0.027). CSIRO=Commonwealth Scientific and Industrial Research Organisation. Back to text 5: The 17 most prolific Australian medical inventors listed on 10 or more Australian medical patents in the US for 1984-1999, grouped by technology Main technology Inventor Number of patents Main assignees Method for constructing proteins and other molecules Simpson R 18 Ludwig Institute for Cancer Research, US Cardiac pacemakers and cochlear ear implants Money D Kuzma J Daly C Milijasevic Z 15 12 10 10 Telectronics Pty Ltd, NSW, and Cochlear Pty Ltd, NSW Biosensors Cornell B Braach-Maksvytis V Raguse B 14 13 10 Australian Membrane and Biotechnology Research Institute, NSW Ribozymes - gene shears Jennings P Cameron F 14 (1) 11 Gene Shears Pty Ltd, NSW and ACT Vitamin B12 as carrier for oral drugs Russell-Jones G 14 (1) Biotech Australia Pty Ltd, NSW Intraocular lenses Barrett G 13 Alcon Laboratories Inc, and Chiron, US, and Oversby Pty Ltd, WA Relaxin gene Tregear G Niall H 13 (1) 10 (1) Howard Florey Institute, VIC Syringe or drug infusion devices Whisson M 12 Eastland Technology Australia Pty Ltd, WA Matrix metalloprotease inhibitors Grobelny D 10 Glycomed Inc, US, and Narhex Ltd, Hong Kong Contact lenses Meijs G 10 (6) CIBA Vision Group, US Back to text 6: The most prolific assignees listed on 15 or more Australian medical patents in the United States for 1984-1999 Assignee Medical Non-medical Total University of Melbourne and affiliated institutions 76 28 104 Telectronics NV or Telectronics Pty Ltd or Telectronics Pacing Systems Inc 75 0 75 CSIRO 60 272 332 Biotech Australia Pty Ltd 28 3 31 University of New South Wales and affiliated institutions 25 51 76 Australian National University and Anutech Pty Ltd 23 30 53 University of Sydney 22 26 48 Monash University and affiliated institutions 21 8 29 AMRAD Corp Ltd 21 3 24 Ludwig Institute for Cancer Research* 21 0 21 University of Queensland and Queensland Institute of Medical Research (QIMR) Council 20 18 38 Cochlear Pty Ltd 16 0 16 Commonwealth of Australia 15 87 102 Gene Shears Pty Ltd 15 0 15 Total 438 526 964 *16 of the 21 patents originated from the Melbourne branch (Dr C Thumwood, Scientific Administrator, Ludwig Institute for Cancer Research, personal communication). Back to text

Eugen Mattes · Michael C Stacey

The Research Enterprise General practice research in Australia, 1980-1999 Alison M Ward, Derrick G Lopez and Max Kamien MJA 2000; 173: 608-611 For editorial comment, see Van Der Weyden Abstract - Methods - Results - Summary and conclusions - Acknowledgements - References - Authors' details - - More articles on General practice and primary care Abstract There has been a nearly fivefold increase in the amount of Australian general practice research published in 1990-1999 compared with the previous decade. The university departments of general practice and other university departments have been responsible for most of the research. GPs were involved in at least 60% of all of the research reviewed. Half of the research was found to be clinically pertinent to the front-line GP. The National Health Priority areas, introduced in 1994, were poorly represented, but it is probably too soon for this research to be published. There has also been little research on rural general practice. This review provides a starting point for classifying general practice and primary healthcare research in the future. In Australia, the general practice research effort, begun by a few enthusiastic general practitioners (GPs) in the early 1970s, now extends to academic programs involving GPs and social scientists with an interest in general practice. The inception of academic departments of general practice resulted in a gradual professionalisation and an increase in research.1In 1991-92 the research effort received an extra boost through the General Practice Evaluation Program (GPEP) of the Commonwealth Department of Health, which aimed to develop data on the dynamics of general practice in Australia and to improve standards and quality assurance. In April this year, the Department announced a new funding initiative -- a national strategy for Primary Health Care Research, Evaluation and Development.2 In the light of these developments, we reviewed and analysed Australian general practice research published in the past 20 years to provide a baseline for future research initiatives. Methods Data sources We conducted literature searches in March and April 2000 to identify all Australian articles relating to general practice research published since 1980. Letters and editorials were excluded. The search strategies, search terms and outcomes are shown in Box 1. Study selection We restricted the review to the two decades 1980-1989 and 1990-1999; and We used a similar definition of general practice research to that used by Starfield for primary care research; that is, even if the problem being examined is a problem seen in general practice, it is not necessarily general practice research unless the context of the research is general practice.3We identified general practice research published in the past two decades by scanning the abstracts of the 2242 articles retrieved by our search (those articles without abstracts [about 28%] were viewed in full). Scanning identified 663 general practice research articles. The remainder were discussion papers, or not general practice research. Classification categories We classified each article and the research reported according to nine variables: Country of publication; Journal; Affiliation of authors: Determining affiliation was complicated by the name changes of departments of general practice. We compiled lists of department names and of academic GPs to help identify who was doing the research. Type of research: Searches of the Australian and overseas literature failed to find a comprehensive classification system for type of general practice research.3,4-6 We extended the classification categories used by the National Information Service (NIS) for the GPEP grants, giving 13 categories of general practice research. Research methodology: The research methods used were difficult to classify.7,8 The usual method categories for public health research do not include some methods used in general practice. We used the public health categories (see Box 6) as a starting point and expanded on these. International Classification of Primary Care (ICPC);9 National Health Priority areas; 10,11 Rural health; Other areas of concern (eg, Aboriginal or Torres Strait Islanders; aged or adolescent populations). Data analysis One of us (A M W) coded all articles and trained a research officer to check the coding. A test-retest reliability check on 41 articles showed an average reliability across the classification variables of 92% (98% for affiliation, 83% for type of research, 83% for methodology, 95% for ICPC category, 98% for National Health Priority areas, 98% for areas of concern, 90% for rural research). The classified data were analysed using SPSS.12 Results Number and location of publications (Boxes 2 and 3) Publications of Australian general practice research increased markedly in the second decade, with 82% published since 1990. As expected, most were published in Australia, but the proportion published overseas has increased from 17.1% in 1980-1989 to 31.7% in 1990-1999, with a marked increase in publications in England in the second decade. Over half of the research has been published in two journals -- Australian Family Physician (31%) and The Medical Journal of Australia (21%). The proportion published in Australian Family Physician almost halved in the second decade, with publications increasing across a range of other journals. Authors and affiliation (Box 4) The States with two or more universities with departments of general practice produced most research, with the amount from New South Wales (37%) being more than double that from any other State. Victoria was second with 18%, followed by Queensland (12%) and South Australia (12%). Universities were responsible for 58% of the research, with hospitals accounting for 12% and the Royal Australian College of General Practitioners (RACGP) 7%. The "other" category included health departments, and the Divisions of General Practice (the latter published less than 1%). The number of articles without a recorded affiliation decreased over the decades (the "institution" field was added to MEDLINE in 1988), but, as many of the early articles had no abstracts and were viewed in full, we were often able to identify the first author's affiliation. The contribution made by GPs to research not conducted in general practice academic departments was assessed by searching for GPs among the co-authors; 39% of the research in other academic departments had GPs as co-authors. We also found that 36% of the research conducted in hospitals and other organisations (eg, health departments) had GPs as co-authors. Overall, GPs were involved as authors in at least 64% of all research in general practice. This has implications for discussions about research conducted by GPs as opposed to research conducted on GPs. Type of research (Box 5) Twenty-seven per cent of research was in the category "GP behaviour, views and opinions". This includes screening, health prevention and promotion, prescribing, counselling and many other general practice activities. The number of GP behaviour studies increased in the second decade. Next most common was studies into education and training -- both undergraduate and postgraduate training. The "encounter" category included studies such as the Australian Morbidity and Treatment Survey (AMTS), clinical epidemiology studies, clinical presentation studies and Health Insurance Commission (HIC) data analyses. Very little research was found in workforce, finance or evidence-based medicine. The "other" category included studies evaluating national health promotion campaigns, and studies that could not be classified elsewhere. Methodologies used (Box 6) Two-thirds of the research was observational, nearly a quarter involved some form of intervention, and little qualitative research was published. There was an increase in both interventional and observational studies over the decades. Forty-one per cent of the studies were purely descriptive, many of these being surveys of GPs' views. Only 5% were randomised controlled trials. Many of the evaluation studies were evaluations of educational or training programs for GPs. Research topics (Box 7) ICPC categories: The psychological category was the most frequently studied (13%); over a quarter of these studies were on smoking behaviour. Most of the "female genital" studies were cervical or breast screening studies and 62% of the "respiratory" studies were on asthma. In over half of the studies, no specific ICPC category applied; 22% of these were education and training studies. However, at least half the research topics focused on a clinical area, indicating their pertinence for practising GPs. The main differences between the two decades were increases in cervical, breast and skin cancer studies, and in respiratory studies. National Health Priority (NHP) areas: Nearly two-thirds of the research did not focus on NHP areas. Research into cancer control accounted for 15%, and the proportion doubled over the two decades, followed by mental health (10%). Studies on asthma, cardiovascular disease and diabetes combined accounted for less than 10% of all the research. Other areas of concern: Only 5% of studies dealt with Aboriginal and Torres Strait Islanders, and aged or adolescent populations. Rural research: We found very little research on rural populations. The amount of rural research has remained constant (16%) over these two decades. The bulk is conducted by academic departments (56%), followed by hospitals (14%) and the RACGP (6%). Summary and conclusions This is the first rigorous review of general practice research in Australia. It has demonstrated a marked increase in the amount of research conducted in the past decade. The university departments of general practice have been responsible for most of the research, and GPs were involved in at least 60% of all of the publications reviewed. Half of all the research was found to be clinically pertinent to the front-line GP. The National Health Priority areas were poorly covered in the reviewed articles, but, as these were only formulated in 1994, there has been insufficient time to conduct the research and get it published. There has also been little research in rural general practice; most was conducted by university departments, although not always the general practice departments. The paucity of research on rural populations may be due partly to the fact that many rural studies were not focused specifically on general practice and not listed under "Family practice" in MEDLINE. In addition, the Australian Journal of Rural Health was not indexed in MEDLINE until 1995, which may account for the small number of general practice publications identified in this journal. The limitations of our review are that only one database was searched (MEDLINE) and several journals relevant to general practice research (eg, Education for General Practice) are not indexed in MEDLINE. By checking the curricula vitae of three prominent researchers in Australian general practice, we know that we missed some publications. This is because of inconsistencies in department names in the institution field and omission of the terms "family practice" or "physicians, family" in the medical subject headings (MeSH) field. We encountered difficulties in developing a comprehensive classification system for type of general practice research. We wanted the codes to be useful to both policy makers and providers, but the final categories were still broad. We will continue to refine this coding system in the future. This review did not examine quality of the research, an area which has presented problems for previous researchers.8 However, the proportion of randomised controlled trials was small, mirroring the findings of a review of general practice research in the United Kingdom in the mid-1990s.5 The 11 university departments of general practice in Australia are relatively small, with an average of three full-time equivalent core academic positions each. Most staff have large teaching loads and little time to pursue research activities. The marked increase in research being conducted by these departments over the last decade shows a commitment to research, despite the problems of finding time and adequately trained staff. To further encourage a research culture in general practice in Australia, the General Practice Strategy Review made several recommendations in 1998, including strengthening the research environment in academic departments of general practice, developing and supporting multidisciplinary career pathways in general practice research, and involving consumers and Divisions of General Practice.13 The success of the recently announced national strategy for Primary Health Care Research, Evaluation and Development1 in implementing the recommendations of the General Practice Strategy Review will, in part, be demonstrated by an increase in the quality, quantity and relevance of general practice research in Australia over the next decade. This overview of general practice research takes stock of where we are now and provides a starting point for classifying general practice and primary healthcare research in the future. Acknowledgements We would like to thank Sandra Pullman, of the University of Western Australia medical library, for performing the literature searches. References Lawson KA, Chew M, Van Der Weyden MB. The rise and rise of academic general practice in Australia. Med J Aust 1999; 171: 643-648. General Practice Branch, Department of Health and Aged Care. General practice in Australia: 2000. Canberra: The Department, May 2000: p 377. Starfield B. A framework for primary care research. J Fam Pract 1996; 42: 181-185. Silagy CA, Schattner P, Baxter RG. Current status of general practice research in Australia. Med J Aust 1992; 157: 108-113. Thomas T, Fahey T, Somerset M. The content and methodology of research papers published in three United Kingdom primary care journals. Br J Gen Pract 1998; 48: 1229-1232. Jones R. Primary care research: ends and means. Fam Pract 2000; 17: 1-4. Dawson-Saunders B, Trapp RG. Basic and clinical biostatistics. Connecticut: Appleton & Lange, 1990. Meijman FJ, de Melker RA. The extent of inter- and intrareviewer agreement on the classification and assessment of designs of single-practice research. Fam Pract 1995; 12: 93-97. Lamberts H, Wood M, editors. International classification of primary care. New York: Oxford University Press; 1989. National Health Priorities. In: Wood T, editor. Australia's health 1998. The sixth biennial health report of the Australian Institute of Health and Welfare. Canberra: Australian Institute of Health and Welfare, 1998: 75. Better Health Outcomes for Australians. Canberra: AGPS; 1994. SPSS [computer program], version 8.0. Chicago, Ill: SPSS Inc, 1997. General Practice Strategy Review Group. General practice, changing the future through partnerships. Department of Health and Family Services. Canberra: The Department, 1998. (Received 27 Sep, accepted 9 Nov, 2000) Authors' details Department of General Practice, University of Western Australia, Perth, WA. Alison M Ward, BPsych, PhD, Senior Research Fellow. Derrick G Lopez, BSc, MMedSci, Research Officer. Max Kamien, MD, FRACGP, FRACP, Professor and Head of Department. Reprints: Dr A M Ward, Department of General Practice, University of Western Australia, 328 Stirling Highway, Claremont, WA 6010. alison.wardATuwa.edu.au Make a comment 1: Australian general practice research -- main search strategies, search terms and outcomes MEDLINE (via Ovid Technologies) 1. General practice research by Australian researchers The terms (physicians, family or family practice) in the medical subject headings (MeSH) field combined with the terms (australia$) or (tasmania or queensland or victoria$ or ACT or northern territory or new south wales or NSW) in the institution field; and 2. Research conducted by GPs regardless of the topic The terms (australia$) or (tasmania or queensland or victoria$ or ACT or northern territory or new south wales or NSW) in the institution field combined with the names of all the academic departments of general practice ((family or general) and (physician$ or practi$)) in the institution field. Outcome 881 articles 3. General practice research published in Australia whether or not by GPs The terms (physicians, family or family practice) in the text field combined with (australia$) in the country-of-publication field. Outcome 834 articles 4. Research from departments possibly missed because of name changes The terms (australia$ or tasmania or queensland or victoria$ or ACT or northern territory or new south wales or NSW) in the institution field combined with ((community or primary) and (health or medicine or practice)) in the institution field. Outcome 480 articles 5. Additional searches, including using additional department names in the institution field; and the names of all heads of departments of general practice, and all professors of general practice in the author field. Outcome 47 articles Australasian Medical Index (AMI) (publications not catalogued in MEDLINE) Using the terms (family practice or physicians, family) in the MeSH field, we identified a further 991 possible articles. Many of these were in non-refereed journals, were duplicates of the MEDLINE articles or were only marginally relevant to general practice research. Because their short truncated titles and abstracts made it impossible to classify the type of research being reported they were not included. We estimate this meant possibly another 100-200 references were missed. Total 2242 articles $ = wild card (ie, search finds all words beginning with the letters before the $ sign). Back to text 2: Country of publication -- number (%) of publications Country 1980-1989 (n = 117) 1990-1999 ( n = 546) Australia England United States Europe (excluding England) Canada Unknown 97 (82.9%) 16 (13.7%) 3 (2.6%) 1 (0.9%) 0 0 373 (68.3%) 126 (23.1%) 35 (6.4%) 9 (1.6%) 1(1.6%) 2 (0.4%) Back to text 3: Top five journals publishing Australian general practice research -- number (%) of publications Journal 1980-1989 Journal 1990-1999 Aust Fam Physician Med J Aust Fam Pract Aust N Z J Psychiatry Community Health Stud* 59 (50.4%) 30 (25.6%) 9 (7.7%) 4 (3.4%) 2 (1.7%) Aust Fam Physician Med J Aust Aust N Z J Public Health* Fam Pract Aust J Rural Health † 145 (26.6%) 112 (20.5%) 44 (8.1%) 36 (6.6%) 21 (3.8%) * Community Health Stud became Aust N Z J Public Health in 1996. †Not indexed in MEDLINE until 1995. Back to text 4: Affiliation of first author -- number (%) of publications Afiliation 1980-1989 (n = 117) 1990-1999 (n = 546) University department of general practice Other university departments Other (eg, health department) Hospital RACGP Unknown 35 (29.9%) 17 (14.5%) 19 (16.2%) 12 (10.3%) 6 (5.1%) 28 (23.9%) 205 (37.5%) 131 (24.0%) 89 (16.3%) 65 (11.9%) 38 (7.0%) 18 (3.3%) Back to text 5: Type of research -- number (%) of publications Type of research 1980-1989 (n = 117) 1990-1999 (n = 546) Encounter and clinical epidemiology studies GP behaviour, views, opinions Education and training Patient behaviour, views, opinions, compliance GP and patient behaviour Workforce Organisation of general practice Health services interface Research methodology Finance Ethical/legal/professional Evidence-based medicine Other 28 (23.9%) 27 (23.1%) 15 (12.8%) 15 (12.8%) 10 (8.5%) 7 (6.0%) 5 (4.3%) 4 (3.4%) 2 (1.7%) 1 (0.9%) 0 0 2 (2.6%) 66 (12.1%) 153 (2.80%) 81 (14.8%) 34 (6.2%) 62 (11.4%) 14 (2.6%) 31 (5.7%) 51 (9.3%) 20 (3.7%) 7 (1.3%) 5 (0.9%) 3 (0.5%) 19 (3.5%) Back to text 6: Methodology used -- number (%) of publications Methodology 1980-1989 (n = 117) 1990-1999 (n = 546) Intervention Evaluation of a service Randomised controlled trial Other controlled trial 15 (12.8%) 3 (2.6%) 1 (0.9%) 102 (18.7%) 28 (5.1%) 6 (1.1%) Observation Descriptive without analytical component Cross-sectional quantitative or qualitative Observational over time Cohort 46 (39.3%) 19 (16.2%) 1 (0.9%) 1 (0.9%) 224 (41.0%) 112 (20.5%) 26 (4.8%) 9 (1.6%) Reviews Systematic review Meta-analysis alone 2 (1.7%) 0 11 (0.2%) 1 (0.2%) Other None of the above 29 (24.8%) 27 (4.9%) Back to text 7: International Classification of Primary Care categories -- number (%) of publications* ICPC category 1980-1989 (n = 117) 1990-1999 (n = 546) Psychological Female genital (pap smears, mammogram) Respiratory Pregnancy, child bearing, family planning Skin Circulatory Endocrine, metabolic, nutritional Digestive General and unspecified Male gential (eg, prostate screening) No ICPC category applied 20 (17.1%) 3 (2.6%) 1 (0.9%) 5 (4.3%) 1 (0.9%) 5 (4.3%) 4 (3.4%) 2 (1.7%) 3 (2.6%) 1 (0.9%) 69 (59.0%) 65 (11.9%) 37 (6.8%) 36 (6.6%) 20 (3.7%) 22 (4.0%) 17 (3.1%) 16 (2.9%) 12 (2.2%) 8 (1.5%) 8 (1.5%) 282 (51.6%) * Categories with fewer than 1% of articles (musculoskeletal; neurological; social problems; blood-forming organs; immune mechanisms; eye; urological; ear; hearing) are not listed. Back to text

Alison M Ward · Derrick G Lopez · Max Kamien

The Research Enterprise Exploring the unknown: the challenges of a career in biomedical research Gordon L Ada Gordon Ada reminisces on his career as a researcher and a facilitator MJA 2000; 173: 612-615 Getting started - Walter and Eliza Hall Institute, Melbourne - John Curtain School of Medical Research, Canberra - World Health Organization, Geneva - Retirement projects - The take-home message? - References - Authors' details - - More articles on Immunology and allergy The attractions of a career in research are many, but first among these is the opportunity to be involved in important discoveries, either personally or though close association with other researchers. Experimenting first with viruses and then in immunology at the Walter and Eliza Hall Institute (1948-1968) was a great start. Later, by becoming head of a world-class microbiology department at the John Curtin School of Medical Research at the Australian National University, I was able to establish an environment that spawned important discoveries in medical science. Getting started I had a happy childhood. I was the fourth in a family of six children -- three boys and three girls. My father studied electrical engineering at Sydney University and later became a senior executive with the New South Wales Railways, but my mother had to leave school early when her mother died. When I entered Sydney University in 1940, my aim was to study biochemistry, having received a fascinating book the previous Christmas -- The science of life, by H G Wells, Julian S Huxley and G P Wells. My years at university might not have been so enjoyable if Jack Still had not returned to Sydney University in 1941 from Gowland Hopkins' Biochemistry Department at Cambridge. He enthused us with stories about the exciting research being done there. My first research position at the Commonwealth Serum Laboratories (CSL) (1944-1946), studying ways of stabilising human serum and avoiding denaturation, convinced me of the need for new techniques to isolate and study individual proteins. I applied for leave from CSL to work at the National Institute of Medical Research in London, where moving-boundary electrophoresis and ultracentrifugation were being used for this purpose. This request, although supported by the Director of CSL, Frederick G Morgan, was refused at a higher level, so I resigned, travelled to England, and worked unpaid at the Institute with Arthur S McFarlane, Head of the Biophysics Department. After a few months, McFarlane recommended my paid appointment to the research staff. Walter and Eliza Hall Institute, Melbourne In the early 1940s, Macfarlane Burnet, Director of the Walter and Eliza Hall Institute (WEHI), on a visit to Harvard University to give the Dunham Lectures, saw the need for his institute to gain these new techniques for studying proteins. He obtained a government grant of £20 000 to establish the technologies (including, later, electron microscopy) at the WEHI. As there was no expertise in Australia, Burnet invited me to join the staff at the WEHI and, with the senior biochemist, Henry Holden, to set up moving-boundary electrophoresis and ultracentrifugation. I arrived back in Australia in August 1948. Fortunately, Holden did much of the establishing and I was able to spend most of my time on research. I became a virologist, working mainly with influenza and later Murray Valley encephalitis viruses, studying their composition and biological properties. I crystallised the Vibrio cholerae neuraminidase. In 1957, after publication of his clonal selection theory,1 Burnet decided to phase out virology in favour of immunology at the Institute. In 1962, I decided to make the switch and, after much reading, began studying immune responses, in particular the fate of tiny amounts of antigen (using the highly immunogenic Salmonella flagella and flagellin labelled with radioactive iodine) to establish the nature and location of cells which bind antigen. Because of my general ignorance of this field, I asked Gus Nossal, then the Deputy Director (Immunology) at the Institute, to help me get started. He kindly agreed, but, when the first autoradiographs showing localisation of antigen over rat primary lymphoid follicles (Box 1) were so striking, Gus decided to collaborate full-time. When presented at a meeting in the United States, our findings ranked a column in the New York Times. The next six years studying the fate and role of antigen during primary and secondary immune responses were like a taste of researcher's heaven, and Gus was a great colleague. We studied the role of antibody in antigen localisation and demonstrated the absence of antigen in antibody-forming cells. Burnet later wrote: What I can be certain about however, is the immense importance of the work on the cellular localisation of antigen led by Ada and Nossal in the 1962-5 period.2 All these findings, together with studies on the influence of antigen structure on immunogenicity with a new PhD student, Chris Parish, were published individually and then finally woven into a monograph.3 John Curtain School of Medical Research, Canberra Despite the attractions of working at WEHI, an invitation to head a department with an international reputation in virology was too exciting to refuse, and in 1968 I succeeded Frank Fenner as Head of the Department of Microbiology at the John Curtin School of Medical Research. Although much was known about the humoral response to viral infections, knowledge about cell-mediated immune responses was almost non-existent. I reasoned that research projects combining both virological and immunological approaches, supported by basic research in both these fields, would surely lead to some exciting findings. This turned out to be the case. For example: In 1972 Chris Parish was the first to show the inverse relationship between antibody and cell-mediated immune responses, which led others to describe two classes of helper T cells. Bruce Stillman's studies on adenovirus in 1978 started him on the road to becoming Director of the renowned Cold Spring Harbor Laboratories in New York State. Robert Blanden was to lay the foundations for a major finding. Our department was acknowledged as a world leader in poxvirus research,4 and Blanden was studying the immune response to ectromelia, a poxvirus pathogenic for mice. In the next few years, using technology for assaying the newly discovered cytotoxic T cells gleaned from overseas meetings, Blanden, in late 1972, became the first to show that cytotoxic T cells would kill cells infected with ectromelia virus. But how did these cells recognise virus-infected cells? Early indications were that major histocompatibility (MHC) antigens were involved in some way, so some inbred mouse strains (members of the same strain having identical MHC antigen specificities) were imported to facilitate further studies. Peter Doherty came to the department as a postdoctoral fellow in 1972, and started work on lymphocytic choriomeningitis (LCM) virus infections of mice. In early 1973, Rolf Zinkernagel, a Swiss medical graduate, worked for a while with Blanden to learn about assaying cytotoxic T cell activity. I then asked Doherty and Zinkernagel to share the same laboratory, as they clearly had similar research interests. In some very elegant experiments, they found that cytotoxic T cells formed during an LCM viral infection would only lyse infected target cells if effector and target cells shared at least some MHC antigen specificities (ie, the T cell lytic activity was "MHC restricted"). They suggested that the cytotoxic T cell receptor recognised at the infected cell surface some virus-induced alteration of the MHC molecule, possibly caused by complexing with a viral antigen.5 They proposed the fundamental concept -- that a central function of MHC antigens on cells was to signal changes in "self", to what they now called "altered self", to the immune system.6 Once identified, such a cell would be lysed. This finding stimulated much research both in the department and elsewhere, and studies investigating the details of MHC restriction of T cell responses became a leading immunological topic internationally. Needless to say, Zinkernagel was awarded a PhD scholarship and graduated in record time. Both he and Doherty left to work overseas in the mid-1970s. Subsequently, analysis of crystals of MHC molecules, isolated from the surface of infected cells by US researchers, showed a viral peptide occupying a cleft in the MHC molecule so that parts of each were recognised by the cytotoxic lymphocyte receptor (Box 2). The award of the 1996 Nobel Prize in Physiology or Medicine to Rolf Zinkernagel and Peter Doherty (Box 3) recognised the importance of their original discovery, as it was the first description of the molecular mechanism used by vertebrates for the control and clearance of most intracellular infectious agents, especially viruses. World Health Organization, Geneva For 20 years, from 1971, I became associated with different World Health Organization programs, concerned mostly with the development and use of vaccines (Box 4). I was the first Chairman of the Programme for Vaccine Development (1984-1989), which is now a much larger WHO program with Gus Nossal as Chairman. These experiences focused my own research towards defining the roles of different components of the immune response to viral infections. Retirement projects As I approached retirement (December 1987), I was invited to do a six-month consultancy at WHO (to plan for a major review of studies on developing a vaccine to control pregnancy in women), to spend my retirement at Johns Hopkins School of Hygiene and Public Health in Baltimore, and to give the plenary lecture on The prospects for HIV vaccines at the Fourth International AIDS Congress in Stockholm in May 1988. I had never worked with HIV, but the Swedes apparently wanted an "independent" opinion. Stockholm By 1986, HIV RNA had been largely sequenced, and there was great optimism that a vaccine could be developed quickly. However, in the next two years, several disturbing findings were made, especially the very great sequence variation of the envelope antigen in different HIV isolates. There are three desirable properties of an infectious agent which can facilitate vaccine development (Box 5). At the Stockholm lecture,7 I listed seven reasons why it would be very difficult to develop an HIV vaccine based primarily on strong infectivity-neutralising antibody formation (Box 5). I then drew on recent research by two of my colleagues, David Boyle and Ian Ramshaw, at the John Curtin School of Medical Research. They had shown that DNA, coding for antigens of other infectious agents and of cytokines, could be inserted into the DNA of a poxvirus, such as vaccinia virus. Vaccination with this "chimeric" virus could protect against infection by the agent which was the source of the inserted DNA. I therefore suggested that, because the internal antigens of HIV (which are the source of many T cell epitopes) showed considerably less variation, a vaccine might be developed based on vaccinia virus containing the genes coding for the internal HIV antigens, gag and pol, as well as for the cytokine, interferon gamma.7 In mice, such a construct generated a strong cytotoxic T cell response; in man, this might be sufficient to better control, if not clear, an HIV infection. The 8000-strong audience was largely stunned by my assessment of the situation, although none subsequently disputed it. However, major vaccine manufacturers ignored it, determined to make an antibody-inducing subunit vaccine based on the HIV envelope antigen, a strategy driven by the success of the hepatitis B viral vaccine which contains the surface antigen of that virus. Baltimore and Washington On arrival in Baltimore in July 1988, I was warmly welcomed, made Associate Director of a new Center for AIDS Research and later became Director. In Washington, I was asked to participate in meetings and activities of the Division of AIDS (DAIDS) of the National Institute of Allergy and Infectious Disease. In 1991, after three years in the United States, my wife and I decided to return to Australia, but I was invited to continue the relationship with DAIDS and to join a new HIV Vaccine Working Group. The crunch came in 1995, when the Director of the US National Institute of Allergy and Infectious Disease refused to support a Phase III clinical trial of the then leading HIV candidate vaccine, based on the envelope antigen. Many reasons were given, but two critical ones were: Antibody from volunteers immunised with this candidate vaccine did not prevent infection by newly isolated HIV field strains; and The vaccine did not induce cytotoxic T cell formation in the volunteers. This was a major turning point in international HIV vaccine research. The National Institute of Allergy and Infectious Disease completely revamped its HIV vaccine development program, and my hectic travel schedule to and from the United States came to an end. My last task for the Working Group was to review the evidence supporting a role for cytotoxic T cells in controlling HIV infections.8 Return to Canberra In 1991, I was appointed Visiting Fellow in the (now) Division of Immunology and Cell Biology at the John Curtin School and Chairman of the HIV Vaccine Working Group, one of the committees of the National Centre in HIV Epidemiology and Clinical Research in Sydney. Ian Ramshaw had recently shown that a vaccination schedule involving priming with plasmids containing DNA coding for selected antigens, followed by boosting with chimeric fowlpox virus coding for the same antigens, gave a greatly enhanced immune response in mice. Stephen Kent (now at the University of Melbourne) and Ramshaw and their colleagues showed that Macaca nemestrina monkeys immunised in this way developed a strong cytotoxic T cell response and rapidly cleared a subsequent HIV infection.9 Any antibody induced was irrelevant. Supporting findings for this approach were later reported from the United States. Now Australia was set to develop an HIV vaccine initiative based on this vaccination technology. At a meeting of the HIV Vaccine Working Group, David Cooper, Head of the National Centre in HIV Epidemiology and Clinical Research, was elected to head an Australian HIV Vaccine Consortium. In June this year, out of 20 international applications received, the National Institute of Allergy and Infectious Disease awarded four contracts, three to US groups and the fourth to the Australian consortium ($27 million over five years) to carry out clinical trials of their vaccine formulation. It is anticipated that a strong immune capability based on cytotoxic T lymphocyte activity will greatly reduce viral titres. Thus, those infected by HIV will live longer and be much less likely to infect others. If this vaccination technology can be shown to generate strong cytotoxic lymphocyte responses in humans, it heralds a new approach to controlling other difficult infectious diseases, such as malaria, trachoma and pelvic inflammatory disease, and even pandemic influenza. The take-home message? From a career path in biochemistry, I switched to virology, then to immunology and became an enthusiastic supporter for the application of immunisation technology, not only for the more difficult infectious diseases but also for non-communicable diseases. Young researchers should jump at the chance to switch fields when exciting opportunities arise. Acknowledgement: I wish to acknowledge with gratitude the great support of my wife, Jean Ada, during my career. References Burnet FH. A modification of Jerne's theory of antibody production using the concept of clonal selection. Aust J Sci 1957: 20; 67-69. Macfarlane Burnet I. Walter and Eliza Hall Institute, 1915-65. Melbourne: Melbourne University Press, 1971. Nossal GJV, Ada GL. Antigens, lymphoid cells and the immune response. New York: Academic Press, 1971. Fenner F. Nature, nuture and my experience with smallpox eradication. Med J Aust 1999; 171: 638-641. Zinkernagel RM, Doherty PC. Restriction of in vitro cell-mediated cytotoxicity in lymphocytic choriomeningitis within a syngeneic or semi-allogeneic system. Nature 1974; 248: 701-702. Doherty PC, Zinkernagel RM. A biological role for the major histocompatibility antigens. Lancet 1975; 1: 1406-1409. Ada GL. Prospects for HIV vaccines. J Acquir Immune Defic Syndr 1988; 1: 295-303. Ada GL, McElrath MJ. Perspectives. HIV type-1 vaccine-induced cytotoxic T cell responses: potential role in vaccine efficacy. AIDS Res Hum Retoviruses 1997; 13: 243-248. Kent SJ, Zhao A, Best SJ, et al. Enhanced T-cell immunogenicity and protective efficacy of a human immunodeficiency virus type 1 vaccine regimen consisting of consecutive priming with DNA and boosting with recombinant fowlpox virus. J Virol 1998; 72: 10180-10188. Authors' details John Curtin School of Medical Research, Australian National University, Canberra, ACT. Gordon L Ada, AO, DSc, FAA, Emeritus Professor, and Visiting Fellow in the Division of Immunology and Cell Biology. Correspondence: Professor G L Ada, John Curtin School of Medical Research, P O Box 334, Canberra, ACT 2601. Make a comment 1: Antigen in the immune response Autoradiograph showing localisation of antigen over primary lymphoid follicles of rat popliteal lymph nodes, after footpad injection of Salmonella flagellin labelled with radioactive iodine. Back to text 2: The function of major histocompatibility antigens Schematic diagram of the cytotoxic T lymphocyte receptor recognition of the complex between the major histocompatibility antigen molecule and a nonapeptide derived from an infectious agent protein expressed on the surface of the infected cell. Back to text 3: At the 1996 Nobel Prize awards Evening banquet after the awarding of Nobel Prizes, Stockholm, December 1996. From left to right: Gordon Ada, Peter Doherty and Frank Fenner at the display of Nobel Prize medals and citations (photograph courtesy of Peter Pockley). Back to text 4: Involvement with World Health Organization programs 1971-1973 Member, Fellowship Selection Committee 1973-1976 Member, then Chairman (1975-1976), Scientific Council, International Agency for Research on Cancer, Lyons, France 1978-1984 Member, Scientific and Technical Advisory Committee, Tropical Diseases Research 1981-1984 Member, Global Advisory Committee on Medical (Health) Research 1984-1989 Chairman, Scientific Advisory Group of Experts, Programme for Vaccine Development. Member and later Consultant (1988), Vaccination Committee, Human Reproduction Programme 1985-1988 Member, Regional (Western Pacific) Advisory Committee on Health Research 1987-1989 Member, Research and Development Group, Expanded Programme on Immunization Back to text 5: Factors for and against the development of an effective vaccine Factors favouring the development of an effective vaccine Only one or a few strains of the infective agent exist; little or preferably no antigenic variation within a strain. Infective agent causes an acute infection; host completely recovers from a sublethal dose of the agent; agent does not persist. Agent is moderately (rather than highly) infectious. Factors militating against development of an effective vaccine (all these factors apply to HIV) Great antigenic variation; antigenic drift. Integration of viral DNA/cDNA into the host cell genome. Infection may be transmitted by cells which are latently infected. Immune enhancement: antibody can enhance infection of macrophages/monocytes if these cells are susceptible to infection. Agent infects cells in immunoprivileged sites in the host. Crucial cells of the immune system are infected, and either destroyed or their function is impaired. Failure to produce protective antibody and/or persisting cell-mediated immunity responses. Back to text

Gordon L Ada

The Research Enterprise John Kerr and apoptosis Michael G E O'Rourke and Kay A O Ellem MJA 2000; 173: 616-617 On 14 March 2000, John Foxton Ross Kerr, Emeritus Professor of Pathology at the University of Queensland, received the Paul Ehrlich and Ludwig Darmstaedter Prize for his description of apoptosis, a form of cell death. The prize, which he shared with Boston biologist Robert Horvitz, is considered to be one of the most prestigious European awards in science, second only to the Nobel Prize. John Kerr's discovery, initially called "shrinkage necrosis" but which he later renamed "apoptosis", came about in the late 1960s, when his attention was caught by a curious form of liver cell death during his studies of acute liver injury in rats. The findings of this seminal study were first published in 1965.1 Subsequently, Kerr and his co-authors (including Jeffrey Searle) described the unique morphological changes of this type of cell death, compared with those of necrosis, in a series of articles published during the 1970s and 1980s.2-4 These studies extended the range of pathological and physiological states in which apoptosis is known to occur. Further studies with other collaborators (who later included Alastair Currie and Andrew Wyllie)5 led to an increase in the understanding of the role of apoptosis in embryogenesis, spermatogenesis, cancer growth, and tissue remodelling during healing or functional regression. At first thought to be somewhat arcane as a topic, the literature on apoptosis initially grew slowly. However, recognition of the significance of this "protected" form of cell death on immune function and regulation was followed by an explosion of related publications in immunology. Biochemists now recite a mantra of enzymes involved in apoptosis, and chant a list of factors capable of modulating or regulating its expression. It has become de rigueur to adorn seminars and lectures with charts of the increasingly complex interactions which occur between signalling pathways as they are traversed by the informing reactions (which either trigger or defuse the suicidal steps leading to apoptotic cell death). The astonishing total number of publications on apoptosis is now over 35 539, including some of the world's leading scientific journals, such as Nature6-8 and Science.9-11 Apoptosis is now a growth industry, the clinical implications of which can be applied to chemotherapy, the endocrine treatment of cancer, autoimmune disease and neurodegenerative disease. This body of evidence is a tribute to the catalytic influence that John Kerr's insights have had in so many disparate disciplines and areas of biological study. These insights revealed the importance of this process as a universal microphenomenon in the macroevents of tissue and organismal function, and in disease. References Kerr JF. A histochemical study of hypertrophy and ischaemic injury of rat liver with special reference to changes in lysosomes. J Path Bact 1965; 90: 419-435. Kerr JF, Cooksley WG, Searle J, et al. The nature of piecemeal necrosis in chronic active hepatitis. Lancet 1979; 20: 827-828. Searle J, Lawson TA, Abbott PJ, et al. An electron-microscope study of the mode of cell death induced by cancer-chemotherapeutic agents in populations of proliferating normal and neoplastic cells. J Pathol 1975; 116: 129-138. Weedon D, Searle J, Kerr JF. Apoptosis. Its nature and implications for dermatopathology. Am J Dermatopathol 1979; 2: 133-144. Kerr JF, Wyllie AH, Currie AR. Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br J Cancer 1972; 26: 239-257. Wallach D. Apoptosis: Placing death under control. Nature 1977; 388: 123-126. Hengartner MO. Apoptosis: death cycle and Swiss army knives. Nature 1998; 391: 441-442. Martinou JC. Apoptosis: key to the mitochondrial gate. Nature 1999: 399; 411-412. Barinaga M. Apoptosis: forging a path to cell death. Science 1996; 273: 735-737. Barinaga M. Apoptosis: death by dozens of cuts. Science 1998; 280: 32-34. Brenner C, Kroener G. Apoptosis: Mitochondria -- the death cell integrators. Science 2000; 289: 1150-1151. Cotran R, Kumar V, Collins T, editors. Pathological basis of disease. 6th ed. Philadelphia: W. B Saunders, 1997: 18-24. Kerr JF, Winterford CM, Harmon BV. Apoptosis. Its significance in cancer and cancer therapy. Cancer 1994; 73: 2013-2026. Walker NI, Harmon BV, Gobe GC, et al. Patterns of cell death. Methods Achiev Exp Pathol 1988; 13: 18-32. Sandford N, Searle JW, Kerr JF. Sucessive waves of apoptosis in the rat prostate after repeated withdrawal of testosterone stimulation. Pathology 1984; 16: 406-410. Soubrane C, Mouawad R, Antoine EC, et al. A comparative study of Fas and Fas-ligand expression during melanoma progression. Br J Dermatol 2000; 143: 307-312. Andrane F, Casciola-Rosen L, Rosen A. Apoptosis in systemic lupus erthymatosus. Clinical implications. Rheum Dis Clin North Am 2000; 26: 215-227. Make a comment John Kerr (right) with Roland Koch, Prime Minister of Hesse, Germany, and Honorary Chairman of the Board of Trustees of the Paul Ehrlich Foundation, at a dinner for award recipients. 1: Other awards and lectures recognising John Kerr's achievements Keynote opening addresses at the Cold Spring Harbor, New York, Symposia, 1990. Opening lecture, The cell and molecular biology of apoptosis, at the Queensland Institute of Medical Research Cell and Molecular Biology Symposium, 1992. Opening lecture at the conference Apoptosis in AIDS, Paris, 1993. 12th Mildford D. Schultz Lecture, Harvard Medical School, Boston, 1993. Bancroft Medal (Queensland AMA), 1993. John Earnshaw Memorial Lecture, International Melanoma Conference, Brisbane, 1994. Fred W Stewart Award, Memorial Sloane-Kettering Cancer Center, New York, for contribution to cancer research, 1995. Doctor of Science honoris causa, University of Queensland, 1998. Fellowship of the Australian Academy of Science, 1998. Back to text 2: Apoptosis: programmed cell death The term "apoptosis" is derived from the Greek for "falling off" and describes a distinct form of cell death whereby cells die in a tightly regulated and morphologically uniform fashion.12 Morphologically, there is condensation of the nucleus and cytoplasm, membrane blebbing, and the formation of discrete, packaged apoptotic bodies which are phagocytosed by nearby cells, without provoking an inflammatory reaction.13 In contrast to necrosis, a degenerative process in which cells swell and lyse after irreversible tissue injury, apoptosis appears to be an active process14 which is subject to genetic regulation. Apoptosis can be triggered either from within the cell, or from outside the cell (mediated by binding of surface membrane receptors to "death activators" such as Fas-ligand and tumour necrosis factor). Electron micrograph showing apoptosis occurring spontaneously in cell culture. Note the discrete, membrane-enclosed nuclear fragments, with characteristic segregation of uniformly compacted chromatin, the crowding of well preserved cytoplasmic organelles and the marked convolution of the cellular surface, which is a prelude to conversion of the cell into a number of membrane-bound fragments or apoptotic bodies. N = nucleus, O = organelles. Apoptosis may occur in several different physiological, adaptive and pathological settings. For example, apoptosis acts as a homoeostatic mechanism for controlling cell populations, and in hormone-dependent tissue involution such as endometrial breakdown during menstruation, prostatic atrophy after castration,15 and cessation of lactation after weaning.12 Localised apoptosis plays a role in embryonic development, such as formation of interdigital clefts and involution of phylogenetic vestiges.5 In malignant tumours, apoptosis may occur spontaneously, or may increase in response to cytotoxic chemotherapy or irradiation.13 Impaired regulation of apoptosis is known to be associated with the development of various types of cancer,16 and with the pathogenesis of some autoimmune diseases such as systemic lupus erythematosus.17 Back to text

Research Recent trends in the use of antidepressant drugs in Australia, 1990-1998 Peter McManus, Andrea Mant, Philip B Mitchell William S Montgomery, John Marley and Merran E Auland MJA 2000; 173: 458-461 For editorial comment, see Parker Abstract - Introduction - Methods - Results - Discussion - Acknowlegdements - References - Authors' details - - - More articles on General practice and primary care Abstract Objective: To determine the pattern of use of antidepressant drugs in the Australian community, 1990-1998, and to compare this with those of other developed countries. Design: Retrospective analyses of prescription and sales data, together with information about patient encounters for depression (from an ongoing survey of service provision by general practitioners) and population-based prevalence estimates for affective disorders (from community health surveys). Main outcome measures: National and international consumption of antidepressants, expressed in defined daily doses (DDDs) per 1000 population per day. Changes in both the frequency of general practice patient encounters for depression and population-based prevalence estimates for affective disorders. Results: Dispensing of antidepressant prescriptions through community pharmacies in Australia increased from an estimated 12.4 DDDs/1000 population per day in 1990 (5.1 million prescriptions) to 35.7 DDDs/1000 population/day in 1998 (8.2 million prescriptions). There has been a rapid market uptake of the selective serotonin reuptake inhibitors (SSRIs), accompanied by a decrease of only 25% in the use of tricyclic antidepressants (TCAs). In 1998, the level of antidepressant use in Australia was similar to that of the United States, while the rate of increase in use between 1993 and 1998 was second only to that of Sweden. In Australia, depression has risen from the tenth most common problem managed in general practice in 1990-91 to the fourth in 1998-99, and the number of people reporting depression in the National Health Surveys (1995 v 1989-90) has almost doubled. Of the prescriptions dispensed in 1998 for antidepressant drugs subsidised by the Pharmaceutical Benefits Scheme, 85% were written by general practitioners, and 11.2% by psychiatrists. Conclusions: As in most developed countries, antidepressant use increased between 1990 and 1998. The rapid market uptake of the new antidepressants, particularly SSRIs, is likely to have been driven by increased awareness of depression, together with availability and promotion of new therapies. Introduction The World Health Organization report on the global burden of disease placed major depression fourth among the leading causes of disease burden in the developing world in 1990, and predicted that it would rise to second by the year 2020.1 In parallel with the increasing awareness of depression as an important health issue, the past decade has seen an increase in the pharmacotherapy options for managing depression with the arrival of several new classes of antidepressants. To review trends in antidepressant use in Australia, the Drug Utilisation Sub-Committee (DUSC) of the Pharmaceutical Benefits Advisory Committee, Department of Health and Aged Care, convened a working group in 1998. The working group, which comprised representatives from the DUSC and from the Australian Pharmaceutical Manufacturers Association (APMA), reviewed Australian and international data on antidepressant sales and dispensing. The aim was to determine patterns of antidepressant use in Australia between 1990 and 1998 and to compare Australian patterns with those in similar developed countries. To assist in interpretation of Australian drug use trends, the group reviewed changes in both the frequency of general practice patient encounters for depression and in population-based prevalence estimates for affective disorders. Methods Prescription and sales data Prescription dispensing data were obtained from the database maintained by the DUSC that monitors the dispensing of prescription medicines through community pharmacies in Australia.2 No data on public hospital use are included in this database. The measurement units used are either prescription volumes or the number of defined daily doses (DDDs) per 1000 population per day. The DDD is based on the assumed average daily dose of the drug when used for its main indication by adults. It is the unit approved by the World Health Organization (WHO) for drug use studies, and allows for comparisons independent of differences in price, preparation and quantity per prescription.3 Within the data on dispensing of antidepressant drugs subsidised by the Pharmaceutical Benefits Scheme (PBS), it is also possible to determine the major specialty of the prescribing doctor. Data on total sales of antidepressants from wholesalers to retail and hospital pharmacies for all countries, except Sweden, were obtained from IMS Health Incorporated. IMS Health is the leading international provider of information on drug usage to the pharmaceutical and healthcare industries.4 Data were retrieved as kilograms of active ingredient and then converted to DDDs per 1000 population per day. Excluded were the use of lithium, Hypericum (St John's wort) or tryptophan, and combinations involving these drugs or their active constituents. Utilisation data for Sweden, where separate local arrangements apply, were supplied by the Swedish Association of the Pharmaceutical Industry (LIF). The 1999 WHO defined daily doses (DDDs) were used in calculations. Drugs unique to particular markets that did not have DDDs available were provisionally assigned values using standard references and information provided by drug information centres in the countries involved.5 Prescriber surveys Information related to general practice patient encounters for depression was obtained from the General Practice Statistics and Classification Unit of the Family Medicine Research Centre (FMRC), University of Sydney, which is conducting an ongoing survey of service provision by general practitioners (GPs).6 This involves 1000 randomly selected, active, recognised GPs per year, each recording details of 100 consecutive consultations on structured encounter forms. Rolling recruitment ensures that the recording weeks are distributed evenly over the year and that there is constant change in participants. These data can be compared with the findings of an earlier FMRC study of morbidity and treatment in general practice that used simpler but compatible methods.7 Information on prescribing by specialists is not included in these GP surveys. Community health surveys The 1995 National Health Survey was a household survey conducted by the Australian Bureau of Statistics to obtain national benchmark information on a range of health-related issues and to enable the monitoring of trends in health over time.8 A previous health survey, collecting broadly comparable data, was conducted in 1989-90.9The 1997 National Survey of Mental Health and Wellbeing of Adults was also conducted by the Australian Bureau of Statistics and used a representative sample of people aged 18 years or over living in private dwellings.10 The survey was interview-based with a diagnostic component administered through a modified version of the WHO Composite International Diagnostic Interview (CIDI). The CIDI translates the criteria of the Diagnostic and statistical manual of mental disorders, 4th edition (DSM-IV),11 and the International classification of diseases, 10th edition (ICD-10),12 into sets of questions that can be readily answered by the general adult population. Specific combinations of symptoms may indicate a specific mental disorder. Results Antidepressant use in Australia The dispensing of prescriptions for antidepressants through community pharmacies in Australia increased from an estimated 12.4 DDDs/1000 population per day in 1990 (5.1 million prescriptions) to 35.7 DDDs/1000 population per day in 1998 (8.2 million prescriptions). Trends in the use of the selective serotonin reuptake inhibitors (SSRIs), tricyclic antidepressants (TCAs), moclobemide, venlafaxine and nefazodone between 1990 and 1998 are shown in Box 1. The market uptake of the SSRIs has been rapid and accompanied by a decrease of only 25% in the use of the TCAs. Other new agents included moclobemide (a reversible monoamine oxidase type A inhibitor), nefazodone (a 5-HT2 antagonist) and venlafaxine (a serotonin-noradrenaline reuptake inhibitor). The 10 most commonly dispensed antidepressants in Australia in 1998 were, in descending order, sertraline, dothiepin, paroxetine, amitriptyline, fluoxetine, doxepin, moclobemide, imipramine, venlafaxine and citalopram. Of these, only the four tricyclic antidepressants were on the market in 1990, with dothiepin alone maintaining or improving its position over this period. Of the PBS-subsidised prescriptions dispensed for antidepressants in 1998, 85% were written by GPs, while 11.2% were written by psychiatrists. International comparisons We compared retail and hospital sales of antidepressants in Australia and seven major developed countries for the years 1993 and 1998 (Box 2). In 1998, sales of antidepressants in Australia (34.2 DDDs/1000 population per day) were similar to those of the United States (34.2 DDDs/1000 population per day), less than in Sweden (37.1 DDDs/1000 population per day) and France (36 DDDs/1000 population per day) and higher than in Canada (30.8 DDDs/1000 population per day) and the United Kingdom (30.4 DDDs/1000 population per day). Germany and Italy had considerably lower usage levels (12 and 9.9 DDDs/1000 population per day, respectively). The rate of increase in Australia between 1993 and 1998 was second only to that of Sweden. For these same countries in 1998, Box 3 shows the percentage split (based on DDDs/1000 population per day) of the antidepressant market by drug class. There was considerable variability in the percentage that TCAs represented of overall antidepressant use, from a low level of 11% in Sweden through to a high of 67% in Germany. Australia, Canada and France had a similar profile, with TCAs representing about 20% of antidepressant use. Venlafaxine was marketed in all eight of the countries surveyed and ranged between 1.5% and 5.2% of the total use. Mianserin had a low level of use in most countries, except for France and Italy, where it represented about 4% of antidepressant use. It was not available in North America. Similarly, moclobemide had a low level of use in most countries, except in Australia, where it represented 12% of the antidepressant market. Prescriber surveys Surveys conducted in 1990-91 and 1998-99 by the Family Medicine Research Centre have shown the increasing prominence of depression as a problem managed in general practice.6,7 In 1998-99, depression ranked as the fourth most common general practice problem, compared with the tenth in 1990-91. The rate of patient encounters involving depression per 100 encounters has increased from 2.1 in 1990-91 to 3.5 in 1998-99. In 1998-99, compared with 1990-91, antidepressants were more likely to be prescribed per every 100 encounters for depression (58.4 prescriptions [95% CI, 56.1-60.8] v 52.3 prescriptions [95% CI, 49.2-55.5]). Comparisons with age and sex demographics for total general practice encounters (women, 58.7%) suggest that female patients were over-represented at encounters for depression. The most frequent patient age group in encounters at which a tricyclic antidepressant was prescribed was 45-64 years (38%), whereas for encounters at which SSRIs were prescribed it was 25-44 years (43%). Sex distribution was similar for both drug groups, with about a third of the patients being men. Depression was the most common problem for which TCAs and SSRIs were prescribed in 1998-99, although the proportion of TCAs prescribed for depression (48.8% [95% CI, 44.3%-53.3%]) was lower than that of SSRIs (81.9% [95% CI, 79.7%- 84.1%]). Other specific problems managed with TCAs were sleep disturbance (7%), anxiety (5%) and back complaints (4.5%). For the SSRIs, these were anxiety (5.8%) and phobia/compulsive disorder (1.7%). When used for depressive disorders, TCAs had a prescribed daily dose consistently lower than the WHO DDD. The prescribed daily doses and DDDs for the most commonly dispensed TCAs were amitriptyline (mean, 59 mg; median, 50 mg; DDD, 75 mg), doxepin (mean, 61 mg; median, 50 mg; DDD, 100 mg) and dothiepin (mean, 85 mg; median, 75 mg; DDD, 150 mg). The prescribed daily doses for the most commonly dispensed SSRIs were much closer to the DDD: fluoxetine (mean, 24 mg; median, 20 mg; DDD, 20 mg), paroxetine (mean, 23 mg; median, 20 mg; DDD, 20 mg) and sertraline (mean, 72 mg; median, 50 mg; DDD, 50 mg). Community health surveys The 1997 National Mental Health and Wellbeing Profile of Adults identified a 5.8% prevalence of affective disorders (depression, 5.1%; dysthymia, 1.1%) during the 12 months before the survey among people aged 18 years or over.10 Women were more likely than men to have experienced affective disorders (7.4% compared with 4.2%). Although based on self-reports, household surveys conducted by the Australian Bureau of Statistics in 1989-90 and 1995 identified marked changes in the number of people reporting current or previous depression. In the 1995 National Health Survey, 8.1 persons per 1000 population reported depression as a long term condition, compared with 2.8 persons per 1000 in the 1989-90 survey. For depression as a recent illness, 11.4 per 1000 population reported this in 1995, compared with 5.8 per 1000 in 1989-90.8,9 Discussion The past decade has seen a remarkable change in the number of people recognised with and managed for depression, in the range of drug therapy options available, and in the volume of antidepressants prescribed. Previously, depression had been reported as under-recognised and undertreated.13-15Prominent among the likely reasons for this change are increased community awareness of depression as an important health issue, and attempts, most notably through government and community campaigns, to reduce the stigma of mental illness and the gaps in professional expertise inhibiting adequate recognition and treatment of depression.16,17 Coincident with these campaigns, important treatment recommendations were released in the United Kingdom in 1992 (the Royal College of General Practitioners and the Royal College of Psychiatrists) and, in the United States, in 1993 (Agency for Health Care Policy and Research).13,18 In Australia, the Psychotropic drug guidelines19 are the endorsed national standard, and the National Health and Medical Research Council has published clinical practice guidelines for managing depression in young people.20,21 The 1995 Australian National Health Survey showed that the number of people reporting depression as a recent and/or long term condition had nearly doubled compared with the earlier survey conducted in 1989-90. Such a change in the true underlying prevalence of disease is unlikely over a relatively short period of time, and the increase is far more likely to reflect a greater awareness of depression, with patients being more comfortable about coming forward for help and doctors, particularly in general practice, being more willing to provide it. This increased awareness of depression by doctors and patients, together with the availability and promotion of new drug therapy options (between 1990 and 1998, five SSRIs have been approved for PBS subsidy together with moclobemide, venlafaxine and nefazodone), accounts for the rise from the tenth to the fourth most common problem managed in general practice between 1990-91 and 1998-99. In 1998-99, encounters for depression were also more likely to generate a prescription for an antidepressant. This change is reflected in drug utilisation statistics. The market uptake of the SSRIs has been rapid and, remarkably, accompanied by only a relatively small decrease in the use of the TCAs. As a result, the overall antidepressant market has expanded greatly, with utilisation (as defined by DDDs/1000 population per day) being nearly three times greater in 1998 than in 1990. Prescription rates, however, have risen only 60% over that time, as the newer antidepressants are more likely to be dosed closer to the DDD than the older tricyclic antidepressants. TCAs are prescribed for sleep disturbance in a small proportion (7%) of patients, which is not the case for SSRIs. Most developed countries have seen similar trends, with sales in Australia consistent with US sales and slightly higher than those in the UK. The percentage that the SSRIs represented of total antidepressant use in Australia in 1998 was similar to that in the United Kingdom. The considerably lower levels of antidepressant use in Germany are probably related to Germany's strong tradition of use of complementary medicines (substantial use of Hypericum preparations [St John's wort] were not included in the comparisons); and the lower levels in Italy may be because, in 1994-98, SSRIs were not reimbursed by the national health system in Italy, but were fully paid for by the patient (Dr Alberto Vaccheri, Associate Professor, Department of Pharmacology, University of Bologna, personal communication, June 1999). Although there are interesting differences between countries, the rapid uptake of the new antidepressants is likely to have been driven by increased awareness, together with the availability and promotion of new therapies. The drug utilisation patterns, supported by evidence from population and general practice surveys, showed that there has been growth in the actual market rather than just redistribution within the market. Public health benefits of this major change in drug use (eg, reductions in suicide rates) are anticipated in the long term, but measuring population-level outcomes from changes will not be easy. Acknowledgements Other members of the Antidepressants Working Group who helped prepare these data were the Australian Pharmaceutical Manufacturers Association and the pharmaceutical industry (Susan Alexander, Mark Bradley, Michelle Burke, Liz Campbell, Victoria Croker, Marnie Firipis, Deborah Monk, Michael Ortiz, Ruth Stokes, Nick Williams). Drug Utilisation Sub-Committee secretariat (John Dudley). General Practice Statistics and Classification Unit, Family Medicine Research Centre, University of Sydney (Helena Britt and Geoff Sayer, who conducted the analyses of the depression data from BEACH). Disclosure: Philip B Mitchell has been a member of scientific advisory boards for Eli Lilly, SmithKline Beecham and Wyeth. References Murray CJ, Lopez AD. The global burden of disease: summary. Cambridge, Mass: Harvard School of Public Health, Harvard University Press (on behalf of the World Health Organization and the World Bank), 1996. Edmonds DJ, Dumbrell DM, Primrose JG, et al. Development of an Australian drug utilisation database: a report from the Drug Utilization Sub-Committee of the Pharmaceutical Benefits Advisory Committee. PharmacoEconom 1993; 3: 427-432. World Health Organization Collaborating Centre for Drug Statistics Methodology. Guidelines for ATC classification and DDD assignment. 2nd edition. Oslo, Norway: WHO, 1998. Hurley SF, McNeil JJ, Berbatis CG. Sources of Australian pharmacoepidemiology data. Commun Health Stud 1988; 12(1): 82-96. World Health Organization Collaborating Centre for Drug Statistics Methodology. ATC Index with DDDs, 1999. Oslo, Norway: WHO, 1998. Britt H, Sayer GP, Miller GC, et al. BEACH (Bettering the Evaluation And Care of Health): a study of general practice activity, six-month interim report. AIHW Catalogue No. GEP 1. Canberra: Australian Institute of Health and Welfare (General Practice series no.1). Bridges-Webb C, Britt H, Miles D, et al. Morbidity and treatment in general practice in Australia 1990-1991 [Errata in Med J Aust 1993; 158: 72, 652]. Med J Aust 1992; 157 (Suppl Oct 19): S1-S56. Australian Bureau of Statistics. 1995 National Health Survey: use of medications, Australia. Canberra: ABS, 1995. (Catalogue No. 4377.0.) Australian Bureau of Statistics. 1989-90 National Health Survey: summary of results, Australia. Canberra: ABS, 1991. (Catalogue No. 4364.0.) Australian Bureau of Statistics. 1997 Mental Health and Wellbeing: profile of adults. Canberra: ABS, 1997. (Catalogue No. 4326.0.) American Psychiatric Association. Diagnostic and statistical manual of mental disorders, 4th edition (DSM-IV). Washington, DC: APA, 1994. World Health Organization. International classification of diseases, 10th edition (ICD-10). Geneva: World Health Organization, 1993. Paykel ES, Priest RG. Recognition and management of depression in general practice: a consensus statement. BMJ 1992; 305: 1198-1202. Hirschfeld RMA, Keller MB, Pamico S, et al. The National Depressive and Manic-Depressive Association consensus statement on the undertreatment of depression. JAMA 1997; 277: 333-340. Kendrick T. Prescribing antidepressants in general practice: watchful waiting for minor depression, full dose treatment for major depression. BMJ 1996; 313: 829-830. Paykel ES, Tylee A, Wright A, et al. The Defeat Depression Campaign: psychiatry in the public arena. Am J Psychiatry 1997; 154 (6 Suppl): 59-65. The National Mental Health Strategy. Community Awareness Program: a review. Canberra: Commonwealth Department of Health and Aged Care, November 1998. US Department of Health and Human Services. Agency for Health Care Policy and Research (AHCPR). Depression in primary care: Vol 11. Treatment of major depression. Rockville, Md: AHCPR, 1993. Psychotropic drug guidelines. 4th edition. Melbourne: Therapeutic Guidelines, 2000. National Health and Medical Research Council. Depression in young people. A guide for general practitioners. Canberra: NHMRC, 1997. National Health and Medical Research Council. Depression in young people. A guide for mental health professionals. Canberra: NHMRC, 1997. (Received 5 May, accepted 31 Aug, 2000) Authors' details Drug Utilisation Sub-Committee, Department of Health and Aged Care, Canberra, ACT. Peter McManus, MMedSc, BPharm, Secretary. South Eastern Sydney Area Health Service, Sydney, NSW. Andrea Mant, MD, FRACGP, MA, Area Adviser, Quality Use of Medicines; and Associate Professor, School of Community Medicine, University of New South Wales, Sydney, NSW. School of Psychiatry, University of New South Wales, NSW. Philip B Mitchell, MD, FRANZCP, FRCPsych, Professor; and Administrative Director, Mood Disorders Unit, Prince of Wales Hospital, Sydney, NSW. Health Economics and Outcomes Research, Eli Lilly Australia Pty Ltd, Sydney, NSW. William S Montgomery, BPharm, DipHospPharm, GradCertHealthEcon, Health Outcomes Research Manager. Department of General Practice, University of Adelaide, Adelaide, SA. John Marley, MD, MB ChB, Professor. Health Economics and Pricing Department, SmithKline Beecham (Australia) Pty Ltd, Melbourne, VIC. Merran E Auland, PhD, Health Economist. No reprints will be avaliable from the authors. Correspondence: Mr P McManus, Secretary, Drug Utilisation Sub-Committee, Mail Drop Point 83, Department of Health and Aged Care, GPO Box 9848, Canberra, ACT 2601. peter.mcmanusAThealth.gov.au Make a comment Back to text Back to text Percentage split of antideprssant sales (based on defined daily doses per 1000 population per day) by drug class in 1998 (data for all countries, except Sweden, from IMS Health; Swedish data from the Swedish Association of the Pharmaceutical Industry). SSRI = selective serotonin reuptake inhibitor. TCA = tricyclic antidepressant. Back to text

Peter McManus · Andrea Mant · Philip B Mitchell · William S Montgomery · John Marley · Merran E Auland

Research Diabetes-related lower-limb amputations in Australia Craig B Payne MJA 2000; 173: 352-354 For editorial comment, see Colman & Beischer; see also Campbell et al. Abstract - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - More articles on Endocrinology Abstract Objective: To identify the prevalence of diabetes-related lower-limb amputations and its regional variations in Australia. Design and setting: Cross-sectional analysis of a hospital morbidity dataset in Australia. Methods: Analysis of the National Hospital Morbidity Database of all hospital separations for the ICD codes 84.10-84.19 (lower-limb amputations) and 250.0-250.9 (diabetes and its complications) for the financial years 1995-96 to 1997-98. Main outcome measure: Number of lower-limb amputations in people with diabetes mellitus in Australia, and in each State and Territory. Results: 7887 diabetes-related lower-limb amputations were reported during the study period, with a mean ± SD of 2629 ± 47 per year. The prevalence in Australia was 13.97 per 100 000 total population, and varied from 11.34 per 100 000 in the Australian Capital Territory to 20.68 per 100 000 in South Australia. Conclusion: Diabetes-related lower-limb amputation poses a substantial personal and public health cost in Australia. The loss of a limb is a frequent complication of diabetes mellitus, most commonly the result of diabetic foot problems such as ulcers and infection. The risk of amputation of the lower limb is increased up to 15-fold in people with diabetes. Contributory factors include the loss of sensation from the sensory neuropathy; deformity and gait abnormalities from the motor neuropathy; abnormal blood flow regulation from the autonomic neuropathy; ischaemia from the macrovascular disease; limited joint mobility from the increased glycolation of collagen; poor glycaemic control; and increased risk of infection. It is usually some trigger or traumatic event superimposed on these risk factors that causes a lesion such as ulceration or infection which starts a pathway leading to amputation.2,3 Inadequate and inappropriate self-care is also a major factor. The National Diabetic Foot Disease Management Program, as part of the National Diabetes Strategy and Implementation Plan,4 has called for a 50% reduction in lower-limb amputations by the year 2005. Data on diabetes-related lower-limb amputations in Australia are lacking.4 The aim of my study was to identify the prevalence of diabetes-related lower-limb amputations in Australia, as well as variations among States and Territories. Methods Approval for the study was given by the Faculty of Health Sciences Human Ethics Committee at La Trobe University (Victoria). The dataset for my analysis was obtained from the Australian Institute of Health and Welfare (AIHW) for the financial years 1995-96, 1996-97 and 1997-98. The AIHW obtained permission from the relevant State and Territory agencies to release the information, which did not include any personal identifying data. Information was obtained from the National Hospital Morbidity Database (compiled by the AIHW) on all separations from public and private hospitals in Australia for the International Classification of Disease (ICD)5 procedure codes 84.10 to 84.19 (amputations of the lower extremity) and diagnosis codes 250.0 to 250.9 (indicating diabetes and its complications) as the principal or secondary diagnoses. Information was also obtained on sex, age, ethnicity, duration of hospital stay, and State or Territory of residence of each patient who had an amputation. A spreadsheet was used to determine the number of amputations in each region and the duration of hospital stay. The data for each State and Territory were age- and sex-standardised6 to the estimated Australian population as at 30 June 1998.7 This information was then used to determine the rate for each State and Territory. Results A total of 7887 diabetes-related lower-limb amputations (68.2% in men) were recorded as occurring in the three-year period, with an annual mean of 2629 ± 47 (SD) (Box 1). Most occurred in the 65-79 years age groups (Box 2). The age- and sex-standardised prevalence of lower-limb amputation varied among the States and Territories (Box 3), from 11.34 per 100 000 total population in the Australian Capital Territory to 20.68 per 100 000 total population in South Australia. The duration of hospital stay (Box 3) also varied among the States and Territories. The shortest mean hospital stay was 20.0 (95% CI, 17.4-22.6) days in South Australia and the longest was 40.2 (95% CI, 23.1-57.3) in the Northern Territory. It was not possible to analyse the ethnicity data, as two States/Territories would not agree to the release of this information. Discussion The 2629 diabetes-related lower-limb amputations in Australia per year represent a significant personal burden on people with diabetes and on the healthcare system. The loss of a limb is a personal tragedy for those with diabetes,8 and is associated with a deterioration of functional status and residential status,9 with a significant number requiring long term care.10 People with diabetes who have a lower-limb amputation have a higher mortality rate,1,11 especially perioperative mortality.12 Half the people with an amputation will require an amputation of the remaining limb within five years.13,14 This morbidity results in high medical and rehabilitation costs: about 10% of diabetes-related healthcare costs are associated with lower-limb amputations.15The sex differences in lower-limb amputation rates of about 2:1 for men to women reported here are consistent with previous reports,20 and may be related to the levels of adherence to advice, the amount of social support, psychological factors such as denial, or a higher prevalence of the physiological risk factors for amputation such as macrovascular disease.21 Ethnicity is a well-recognised risk factor for lower-limb amputation,22,23 but was not analysed in this project as two of the States/Territories would not release this information. The duration of hospital stay has been identified as one of the main determinants of cost associated with a lower-limb amputation.17 The mean number of bed-days reported here (24.7 days) is less than the mean in the Netherlands15 (42 days) and more than that in the United States16 (15.9 days). There was a large variation among the Australian States and Territories in the mean hospital stay; South Australia has the highest prevalence of lower-limb amputation, but the shortest mean stay. Regional variations have been reported previously in New Zealand for hospital admissions for diabetic foot complications.18 Such regional variations are most likely to be due to variations in clinical practice and access to services.19 A number of shortcomings are inherent in the type of dataset analysed here. Of primary concern is the accuracy of the recording of data. Diabetes has been reported as being under-recorded on discharge records,24,25 so the numbers reported here are most likely an underestimate. There is also concern that the dataset does not distinguish the number of multiple amputations in the same individual; this will bias the population towards the characteristics of these individuals. A number of modifiable risk factors for diabetes-related lower-limb amputation have been identified,26-28 including the lowering of blood pressure, improving glycaemic control and reducing or eliminating smoking. With proper foot care, patient education and provision of appropriate services, such as regular podiatric care, a reduction in the number of amputations can be achieved.4 A number of studies have shown the value of multidisciplinary teams in reducing amputations by up to 50%.29-32 A reduction of this magnitude has the potential to save up to $24 million (based on the assumption that the direct cost of diabetes-related lower-limb amputations in Australia is $48 million per year4). However, a significant proportion of this potential saving will need to be directed to programs to prevent the amputations. Acknowledgements Funding for this project was provided by the Australasian Podiatric and Education Foundation. References Nelson R, Gohdes DM, Everhart JE, et al. Lower extremity amputations in NIDDM: 12 year follow up study in Pima Indians. Diabetes Care 1998; 11: 8-16. Payne CB, Scott RS, Moir C. Trigger events for acute admission to hospital for diabetic foot disease. Australas J Podiatric Med 1998; 32: 57-64. Pecoraro RE, Reiber GE, Burgess EM. Pathways to diabetic amputation -- basis for prevention. Diabetes Care 1990; 13: 513-521. Colagiuri S, Colagiuri R, Ward J. National Diabetes Strategy and Implementation Plan. Canberra: Diabetes Australia, 1998. The International classification of diseases. 9th Revision. Clinical modification. Commission on Professional and Hospital Activities. Michigan, 1990. Beaglehole R, Bonita R, Kjellstrom T. Basic epidemiology. Geneva: World Health Organization, 1993. Australian Bureau of Statistics. Australian demographic statistics. Canberra: ABS, 1999. (Catalogue no. 3101.0.) Fitzpatrick MC. The psychologic assessment and psychosocial recovery of the patient with an amputation. Clin Orthop 1999; 381: 98-107. Frykberg RG, Arora S, Pomposelli FB, LoGerfo F. Functional outcome in elderly following lower extremity amputation. J Foot Ankle Surg 1998; 37: 181-185. Lavery LA, van Houtum WH, Armstrong DG. Institutionalisation following diabetes related lower extremity amputation. Am J Med 1997; 103: 383-388. Faris I, Duncan H, Young C. Factors affecting the outcome of diabetic patients with foot ulcers or gangrene. J Cardiovasc Surg 1988; 29: 736-740. Ebskov LB. Relative mortality in lower limb amputees with diabetes mellitus. Prosthet Orthot Int 1996; 20: 147-152. Silbert S. Amputation of the lower extremity in diabetes mellitus. Diabetes 1952; 1: 297-299. Ebskov LB. Diabetic amputation and long-term survival. Int J Rehab Res 1998; 21: 403-408 Van Houtum WH, Lavery LA, Harkless LB. The costs of diabetes-related lower extremity amputations in the Netherlands. Diabetic Med 1995; 12: 777-781. Ashry HR, Lavery LA, Armstrong DG, et al. Cost of diabetes related amputations in minorities. J Foot Ankle Surg 1998; 37: 186-190. Solomon C, van Rij A, Barnett R, et al. Amputations in the surgical budget. N Z Med J 1994; 107: 78-80. Payne CB, Scott RS, Moir C. Hospital discharges for diabetic foot disease in New Zealand 1980-1993. Diabetes Res Clin Pract 1998; 39: 69-74. Sanders D, Coulter A, McPherson K. Variations in hospital admission rates: a review of the literature. London: King Edward's Hospital Fund, 1989. Armstrong DG, Lavery LA, van Houtum WH, Harkless LB. The impact of gender on amputation. J Foot Ankle Surg 1997; 36: 66-69. Vogt MT, Wolfson SK, Kuller LH. Lower extremity arterial disease and the aging process -- a review. J Clin Epidemiol 1992; 45: 529-542. Lavery LA, Ashry HR, van Houtum W, et al. Variation in the incidence and proportion of diabetes related amputations in minorities. Diabetes Care 1996; 19: 48-51. Simmons D, Scott D, Kenealy T, Scragg R. Foot care among diabetic patients in South Auckland. N Z Med J 1995; 108: 106-108. Williams DRR, Fuller JH, Stevens LK. Validity of routinely collected hospital admissions data on diabetes. Diabetic Med 1998; 6: 320-324. Phillips DE, Mann JI. Diabetes -- inpatient utilisation, costs and data validity. Dunedin 1985-9. N Z Med J 1992; 105: 313-315. Moss SE, Klein R, Klein BEK. The prevalence and incidence of lower extremity amputation in a diabetic population. Arch Intern Med 1992; 152: 610-616. Lehto S, Ronnemaa T, Pyorala K, Laakso M. Risk factors predicting lower extremity amputations in patients with NIDDM. Diabetes Care 1996; 19: 607-611. Hamalainen H, Ronnemaa T, Halonen JP, Toikka T. Factors predicting lower extremity amputations in patients with type 1 or type 2 diabetes mellitus: a population based 7 year follow-up study. J Intern Med 1999; 246: 97-103. Edmonds ME, Blundell MP, Morris ME, et al. Improved survival of the diabetic foot -- the role of a specialised foot clinic. QJM 1986; 60: 763-771. Malone LM, Snyder M, Anderson G, et al. Prevention of amputation. Am J Surg 1989; 158: 520-523. Ebskov LB. Epidemiology of lower extremity amputation in Denmark. Int Orthop 1991; 15: 285-288. Larson J, Apelqvist J, Agardh CD, Stenstrom A. Decreasing incidence of major amputation in diabetic patients -- a consequence of a multidisciplinary foot care team approach. Diabetic Med 1995; 12: 770-777. (Received 25 Nov 1999, accepted 20 Jul 2000) Authors' details Faculty of Health Sciences, La Trobe University, Melbourne, VIC. Craig B Payne, DipPod(NZ), MPH, Lecturer, Department of Podiatry. Reprints: Dr C B Payne, Department of Podiatry, School of Human Biosciences, Faculty of Health Sciences, La Trobe University, Bundoora, VIC 3083. c.payneATlatrobe.edu.au Make a comment 1: Number of diabetes-related lower-limb amputations in Australia Men Women Total1995-96 1996-97 1997-98 1729 1849 1804 851 824 830 2580 2673 2634 Mean ±SD 1795 ±61 834 ±14 2629 ±47 Total 5382 (68%) 2505 (32%) 7887 Back to text Click in box for larger versionBack to text 3: Age- and sex-standardised prevalence and duration of hospital stay for lower-limb amputations in Australia for 1995-1998 Mean ±SD lower extremity amputations per year Rate (95% CI) per 100000 total population New South Wales Victoria Queensland South Australia Western Australia Tasmania Northern Territory Australian Capital Territory Australia 801 ±13 695 ±12 468 ±8 308 ±6 219 ±4 67 ±2 36 ±1 35 ±1 2629 ±47 12.59 (9.54-15.78) 14.87 (11.6-18.17) 13.48 (10.56-16.45) 20.68 (17.18-24.18) 11.89 (8.90-14.88) 14.21 (12.63-16.17) 18.86 (15.53-22.19) 11.34 (8.34-13.56) 13.97 (11.98-15.87) Duration of hospital stay Mean (95% CI) bed days Median (range) bed days New South Wales Victoria Queensland South Australia Western Australia Tasmania Northern Territory Australian Capital Territory Australia 24 (23-26) 22 (21-23) 30 (28-33) 20 (17-23) 26 (22-29) 27 (19-35) 40 (23-57) 33 (18-47) 25 (24-26) 18 (1-210) 16 (1-183) 21 (1-283) 13.5 (1-176) 18 (1-183) 21 (1-197) 24 (1-224) 24 (1-223) 17 (1-283) Back to text

Craig B Payne

Research Rates of in-hospital arrests, deaths and intensive care admissions: the effect of a medical emergency team Peter J Bristow, Ken M Hillman, Tien Chey, Kathy Daffurn, Theresa C Jacques, Sandra L Norman, Gillian F Bishop and E Grant Simmons MJA 2000; 173: 236-240 Abstract - Methods - Results - Discussion - Conclusion - Acknowledgements Authors' details - - More articles on Emergency medicine Abstract Objectives: To evaluate the effectiveness of a medical emergency team (MET) in reducing the rates of selected adverse events. Design: Cohort comparison study after casemix adjustment. Patients and setting: All adult (≥ 14 years) patients admitted to three Australian public hospitals from 8 July to 31 December 1996. Intervention studied: At Hospital 1, a medical emergency team (MET) could be called for abnormal physiological parameters or staff concern. Hospitals 2 and 3 had conventional cardiac arrest teams. Main outcome measures: Casemix-adjusted rates of cardiac arrest, unanticipated admission to intensive care unit (ICU), death, and the subgroup of deaths where there was no pre-existing "do not resuscitate" (DNR) order documented. Results: There were 1510 adverse events identified among 50 942 admissions. The rate of unanticipated ICU admissions was less at the intervention hospital in total (casemix-adjusted odds ratios: Hospital 1, 1.00; Hospital 2, 1.59 [95% CI, 1.24-2.04]; Hospital 3, 1.73 [95% CI, 1.37-2.16]). There was no significant difference in the rates of cardiac arrest or total deaths between the three hospitals. However, one of the hospitals with a conventional cardiac arrest team had a higher death rate among patients without a DNR order. Conclusions: The MET hospital had fewer unanticipated ICU/HDU admissions, with no increase in in-hospital arrest rate or total death rate. The non-DNR deaths were lower compared with one of the other hospitals; however, we did not adjust for DNR practices. We suggest that the MET concept is worthy of further study. Certain in-hospital deaths may be preventable.1-3 Nearly 85% of hospital inpatients who suffer a cardiorespiratory arrest have documented observations of deterioration in the eight hours before the arrest.4,5 Recent studies have demonstrated suboptimal care of hospitalised patients before their admission to the intensive care unit (ICU), and that these patients have a higher mortality.6-8 The authors of these studies urge earlier intervention. One approach to providing an early response to at-risk inpatients throughout the hospital is the medical emergency team (MET),9,10 which replaces the conventional cardiac arrest team. The MET responds to specific clinical criteria (such as bradycardia, tachycardia, hypotension and threatened airway) in order to prevent further deterioration. Others have advocated similar approaches to reduce unexpected hospital deaths and morbidity.11,12 In our study, selected outcomes in a hospital with a MET were compared with outcomes in two hospitals with conventional cardiac arrest teams. These outcomes were rates of cardiorespiratory arrest, unanticipated admission to the ICU or high dependency unit (HDU), death, and deaths where there was no prexisting "do not resuscitate" (DNR) order. Methods This study was a prospective cohort comparison of three hospitals, testing whether an early intervention team, the MET, was associated with fewer adverse events among inpatients, after adjusting for casemix differences. The study was approved by the ethics committees of each participating hospital and the University of New South Wales. Setting The hospitals were similarly sized Australian public hospitals, with bed capacities in the range 380-530. At Hospital 1, the cardiac arrest team was replaced by a MET, which any staff member could call for immediate assistance. Staff could summon the MET if concerned about a patient's condition or if the patient's vital signs exceeded certain levels (Box 1).10 An education program explained the MET's role to all new staff. However, calling the MET when criteria were met was not compulsory. The MET consisted of the ICU registrar and senior nurse, and medical registrar. At Hospitals 2 and 3, the arrest team was paged by nursing or medical staff for cardiorespiratory arrest. The arrest team consisted of the ICU registrar, medical registrar, and ICU or coronary care nurse. Data collection We identified all cardiorespiratory arrest calls, deaths, and ICU/HDU admissions at the three hospitals among patients 14 years and over in hospital during the period from 8 July to 31 December 1996. These were designated "events". Soon after an event, the patient's medical record was reviewed for demographic information. In addition, for cardiac arrests and deaths, documentation of a DNR order before arrest or death was recorded. Each ICU/HDU admission was classified as to whether the patient was admitted to ICU/HDU for the same reason he or she was admitted to hospital. If not, the ICU/HDU admission was defined as unanticipated. For example, a patient admitted to ICU with respiratory distress after a cholecystectomy would be unanticipated. Data were collected by three critical care nurses (one at each hospital) trained in the use of a specifically designed form, which was piloted for two weeks before data collection. The nurses were familiar with the medical record at the three hospitals. Where the information contained in the history was unclear, the attending staff were asked for clarification. Data were entered into a database.13 Data cleaning was performed and any anomalies checked by reference to the datasheet or medical record. Data were then exported to SAS14 for analysis. Outcome measures The primary endpoints were the casemix-adjusted rates of ICU/HDU unanticipated admission, cardiac arrest, death, and deaths without a prior DNR order. These were called "total event rates". For any patient, one event could result in additional events (eg, cardiac arrest followed by unanticipated ICU admission and then death). The "index event" was defined as the event the data collectors considered the first in a series of events. The casemix-adjusted rates of index events were compared between the three hospitals as secondary endpoints. To ensure that any decline in the rate of unanticipated admissions was not caused by excess anticipated admissions, the casemix-adjusted rate of all ICU/HDU admissions was calculated as a control measure. Casemix adjustment Demographic and diagnostic data on the patient population (aged ≥ 14 years) admitted to the hospitals for the period were obtained. The study data were merged by medical record number and date of admission with the complete inpatient statistical data to create a dataset with 50 942 records. This enabled us to identify the admissions for which an event occurred and analyse the data at the patient level. Using simple and multiple logistic regression, we modelled the probability of an event occurring during hospitalisation, adjusted for patient demographics and diagnostic characteristics. Models were derived independently for each total event and for each index event. Parameters were added to the model in a stepwise fashion. To prevent overparameterising the models (where minor, non-significant differences cumulatively hide true differences), when C (equivalent to the area under the receiver operator characteristics curve) reached 0.85 no further parameters were added to the model. This always occurred with fewer than six parameters used. Demographic and casemix independent variables that were tested for use in the models are detailed in Box 2. The models developed used groups of diagnostic categories based on ICD-9-CM codes using the principal diagnosis and the stay diagnosis only.15 The ICD-9-CM code groupings used are available from the principal author (PJB). The performance of the models was assessed by Hosmer-Lemeshow goodness-of-fit tests.16 The risk of an event occurring in a hospital compared with the MET hospital was presented as an adjusted odds ratio with 95% confidence intervals. A level of significance of 5% was used in all statistical tests. Results Hospital demographics Characteristics of all patients (aged ≥ 14 years) admitted to the three hospitals during the study period are shown in Box 3. Hospital 2 had fewer admissions than the other hospitals. Hospital 1 had a higher proportion of male patients admitted, and a lower proportion of admissions from the emergency department (ED). This hospital also had a younger patient population, which is reflected in differences in casemix: Hospital 1 had lower proportions of patients with stroke, severe acute heart disease, gastrointestinal disease, and musculoskeletal and connective tissue diseases, but higher proportions with severe trauma and follow-up care without acute diagnosis (eg, dialysis). The rates of DNR orders in dying patients were 77% in Hospital 1, and 64% and 70% in Hospitals 2 and 3, respectively (P = 0.006). Prevalence and characteristics of events A total of 1510 adverse events (unanticipated ICU/HDU admissions, arrest calls, and deaths) were recorded during the study period for the three hospitals. There were 1100 index events. The prevalence and characteristics of events are summarised in Box 4 for total event rates and Box 5 for index rates. There was a significantly reduced rate of unanticipated ICU/HDU admissions at the MET intervention hospital after casemix adjustment (for both the total event rate and the index rate). After adjustment, Hospital 2 had 49 (95% CI, 20-87) more unanticipated ICU/HDU admissions over a six-month period, and Hospital 3 had 92 (95% CI, 47-146) more, compared with Hospital 1. The rate of all ICU/HDU admissions was lower at Hospital 1 than at one control hospital, and trended to lower than at the other. There was no statistically significant difference in cardiac arrest rate or death rate after casemix adjustment. The casemix-adjusted death rate in patients where there was no documentation found of a DNR order was significantly higher at Hospital 2, translating to 27 (95% CI, 7-53) extra non-DNR deaths. Model performance Box 6 presents an example of the odds ratios after addition of the most significant variables in the multiple logistic regression models derived from the data for the total arrest data. It shows the C statistic as each variable was added to the cardiac arrest model. In the cardiac arrest models, the variables that were adjusted for were emergency admissions, age over 74, heart disease, lung disease and infectious disease as diagnoses. In the total death models, the terms adjusted for were emergency admissions, age over 74, single day stay emergency admissions, and cancer and infectious disease. The same demographic variables were used in the index death model, although in the total non-DNR death model both age ≥ 75 and age 65-74 were used in the model. In the unanticipated ICU/HDU models, the variables adjusted for were single day admission, emergency admission, and cancer and gastrointestinal disease diagnoses. The total unanticipated ICU/HDU model also included the variables stroke and infectious disease as diagnoses. All models satisfied the Hosmer- Lemeshow test.16 Discussion Rationale for our methods In this study, we attempted to determine if the MET system was associated with a reduced rate of adverse events among inpatients. To do this, we compared the rates of adverse events between three hospitals after casemix adjustment.17-19 This method was chosen as we decided that randomisation at the patient level was impractical. A random pattern of response to calls would probably have dissuaded staff caring for patients from calling the MET. Randomisation by ward would have risked contamination bias and engendered problems of casemix, as wards differ in the nature of their patients. Historical comparison at Hospital 1 between a period before and a period after introduction of the MET team was impractical, as the team had been trialled and evolved for six years before the study. The models we used appear to adequately fit the data, according to the Hosmer-Lemeshow goodness-of-fit tests, and with good model performance measured by C statistics. However, multiple methods of casemix adjustment are possible, and these may give divergent results. This is a limitation of casemix adjustment methodology.20 To avoid concealing real differences by excessive modelling, parameters were added stepwise by multivariate analysis until the models reasonably represented the data. The terms which appeared in the final models were usually those that could be expected to influence the outcomes. Thus, advanced age and emergency admissions were factors in the death and cardiac arrest models. The cardiac arrest models also included the terms for heart disease, lung and infectious disease. Infectious disease was an unexpected variable and was also significant in the total death model. Other differences (such as levels of hospital funding, ICU/HDU capacity, the number and seniority of medical and nursing staff, and the level of out-of-hours cover) may also have contributed to the results. However, to adjust for these would have been more difficult than for the variables studied, which relate directly to the patients at risk and are easily and reliably obtained. Explanation of findings After casemix adjustment, we found reduced rates of both total and index unanticipated ICU/HDU admissions at the MET intervention hospital. There were no differences in the rates of cardiac arrests or deaths. However, at one hospital without the MET, there was a higher rate of non-DNR deaths, the subset of deaths most likely potentially preventable by a MET. The reduction in unanticipated ICU/HDU admissions that was seen in the MET intervention hospital could result from many factors. One possible explanation is that the MET was effective and able to intervene on the wards and prevent further deterioration. Another possible reason may relate to differences in referral practices: perhaps the presence of MET backup engendered a feeling that ICU/HDU referral was not needed. Misclassification of ICU/HDU admissions as anticipated rather than unanticipated was excluded as an explanation of the difference by the finding that the rate of all ICU/HDU admissions was lower at the intervention hospital than one control hospital and trended to lower at the other. The lack of efficacy of the MET to prevent cardiorespiratory arrest and modify death rate may be related to lack of sensitivity of calling criteria, or because pathophysiological processes (eg, shock) become irreversible. Another possible explanation for the lack of effect of the MET on event rates is underutilisation. Based on a previous study,21 up to 706 MET calls could have been expected, and yet only 150 were made. Frequent education is probably also required to ensure the appropriate calling of a MET.22 No special efforts regarding staff education in the study period were made. The clinical staff of the hospital were unaware of the study, to negate any possible Hawthorne effect.23 Finally, organisational changes such as introduction of a MET are difficult to implement in hospitals.24,25 Our results probably reflect the effectiveness of the implementation of the MET system as much as the concept of early intervention. Future directions Our study cannot answer definitively if the MET was the cause of the benefit we observed; it does show that the MET concept is worthy of further study. The study could be likened to a Phase II trial of a drug comparing three hospitals at one point in time. Further studies of the MET system's efficacy are needed, such as a before-after comparison in several hospitals, or a comparison of a larger sample size of intervention and control hospitals. Such studies should be repeated some time after the intervention. Is the benefit observed useful and worth pursuing? If it is possible to reduce unanticipated ICU admissions without increased mortality this may result in cost saving. It has been estimated that the US spends about 1% of its gross national product on intensive care facilities.26 However, any savings in intensive care would be offset by the cost of establishing and maintaining a MET. The MET may also have unexpected costs and benefits on processes such as staff satisfaction with care provided. Again, these need to be quantified. Conclusion In this study, we found that fewer patients were unexpectedly admitted to ICU or HDU at a hospital with the MET system, and this hospital had fewer non-DNR deaths than one of the other hospitals. There was no significant change in the casemix-adjusted rate of arrests or total deaths. This may be an advantage of an early response team, which could have important implications for patient care in hospitals. We believe that the MET concept should be studied further in a larger sample of institutions. Acknowledgements Funding for the study was provided by a Commonwealth Department of Health and Family Services Research and Development Grant (HS338). Associate Professor Robert Gibberd assisted with the statistical analysis. References Brennan TA, Leape LL, Laird N, et al. Incidence of adverse events and negligence in hospitalised patients: results of the Harvard Medical Practice Study I. N Engl J Med 1991; 324: 370-376. Leape LL, Brennan TA, Laird N, et al. Nature of adverse events in hospitalised patients: results of the Harvard Medical Practice Study II. N Engl J Med 1991; 324: 377-384. Wilson R McL, Runciman WB, Gibberd RW, et al. The Quality in Australian Health Care Study. Med J Aust 1995; 163: 458-471. Schein RMH, Hazday N, Pena M, et al. Clinical antecedents to inhospital cardiopulmonary arrest. Chest 1990; 98: 1388-1392. Franklin C, Mathew J. Developing strategies to prevent inhospital cardiac arrest: analyzing responses of physicians and nurses in the hours before the event. Crit Care Med 1994; 22: 246-247. Lundberg JS, Perl TM, Wiblen T, et al. Septic shock: an analysis of outcomes for patients with onset on hospital wards versus intensive care units. Crit Care Med 1998; 26: 1020-1024. Goldhill DR, Sumner A. Outcome of intensive care patients in a group of British intensive care units. Crit Care Med 1998; 26: 1337-1345. McQuillan P, Pilkington S, Allan A, et al. Confidential inquiry into quality of care before admission to intensive care. BMJ 1998; 316: 1853-1858. Lee A, Bishop G, Hillman KM, Daffurn K. The medical emergency team. Anaesth Intensive Care 1995; 23: 183-186. Hourihan F, Bishop G, Hillman KM, et al. The medical emergency team: a new strategy to identify and intervene in high risk patients. Clin Intensive Care 1995; 6: 269-272. Frank ED. A shock team in a general hospital. Anesth Analg 1967; 46: 740-745. Goldhill DR. Introducing the postoperative care team [editorial]. BMJ 1997; 314: 389. Microsoft Access [computer program]. Version 2.0. Redmond, Wa: Microsoft, 1994. SAS for Windows [computer program]. Version 6.12. Cary, NC: SAS Institute Inc, 1997. Stremple JF, Bross DS, Davis CL, McDonald GO. Comparison of postoperative mortality and morbidity in VA and nonfederal hospitals. J Surg Res 1994; 56: 405-416. Hosmer DW, Lemeshow S. Applied logistic regression. New York: John Wiley and Sons, 1989. Iezzoni LI. The risks of risk adjustment. JAMA 1997; 278: 1600-1607. Dubois RW, Rogers WH, Moxley JH, et al. Hospital inpatient mortality. Is it a predictor of quality? N Engl J Med 1987; 317: 1674-1680. Green J, Passman LJ, Wintfield N. Analyzing hospital mortality. The consequences of diversity in patient mix. JAMA 1991; 265: 1849-1853. Iezzoni LI, Shwartz M, Ash A, et al. Severity measurement methods and judging hospital death rates for pneumonia. Med Care 1996; 34: 11-28. Hillman KM, Bishop G, Lee A, et al. Identifying the general ward patient at high risk of cardiac arrest. Clin Int Care 1996; 7: 242-243. Daffurn KD, Lee A, Hillman KM, et al. Do nurses know when to summon emergency assistance? Intensive Crit Care Nurs 1994; 10: 115-120. Grufferman S. Complexity and the Hawthorne effect in community trials [editorial]. Epidemiology 1999; 10: 209-210. Garside P. Organisational context for quality: lessons from the fields of organisational development and change management. Qual Health Care 1998; 7 Suppl: S8-15. Koeck C. Time for organisational development in healthcare organisations [editorial]. BMJ 1998; 317: 1267-1268. Cerra FB. Healthcare reform: the role of coordinated critical care. Crit Care Med 1993; 21: 457-464. (Received 15 Feb, accepted 10 Jul, 2000) Authors' details Liverpool Hospital, Sydney, NSW. Peter J Bristow, MB BS, FRACP, Staff Specialist, Department of Intensive Care; Ken M Hillman, MB BS, FFICANZCA, Professor, University of New South Wales Clinical School; Kathy Daffurn, RN, MAppSc, Co-Director, Division of Critical Care; Sandra L Norman, MN, BAppSc, Clinical Nurse Specialist, Department of Intensive Care; Gillian F Bishop, MB ChB, FFICANZCA, Director, Department of Intensive Care; Tien Chey, BSc, MAppStat, Statistician, Epidemiology Unit. Department of Intensive Care, St George Hospital, Sydney, NSW. Theresa C Jacques, MB BS, FFICANZCA, Director. Department of Intensive Care, Illawarra Regional Hospital, Wollongong, NSW. E Grant Simmons, MB BS, FFICANZCA, Director. Reprints will not be available from the authors. Correspondence: Dr P J Bristow, Intensive Care Offices, Alfred Hospital, Commercial Road, Prahran, VIC 3181. p.bristowATalfred.org.au Make a comment 1: Criteria for calling the medical emergency team10 Cardiorespiratory arrest Threatened airway Respiratory rate ≤5 breaths per minute ≥36 breaths per minute Pulse rate ≤40 beats per minute ≥140 beats per minute Systolic blood pressure ≤90mmHg Repeated or prolonged seizures Fall in Glasgow Coma Score >2 points Concern about patient status not detailed above Back to text 2: Variables available for calculation of the various models Sex (binary) Seven age categories (14-24, 25-34, 35-44, 45-54, 55-64, 65-74, ≥75) Same-day admission (binary) (ie, admission and discharge occurred on the same calendar day) Referral from emergency department (binary) Australian born (binary) Casemix categories (16 indicator variables, available from author) Hospital (three indicator variables) Back to text 3: Characteristics of admissions at the three study hospitals from 8 July to 31 December 1996 Hospital* Characteristic 1 2 3 Test of Independence† Number of admissions 18338 13059 19545 Male admissions 44.9% 42.9% 42.8% χ2=21.06‡ (2 df) Same-day admissions 47.7% 47.0% 46.7% χ2=4.35 (2 df) Admission via emergency department 29.6% 36.0% 35.1% χ2=186.53‡ (2 df) Australian born Country of birth not stated 49.3% 6.8% 67.2% 0.5% 50.2% 23.2% Not tested Age distribution 14-24 25-34 35-44 45-54 55-64 65-74 ≥75 9.7% 14.9% 14.3% 12.4% 18.1% 20.5% 10.0% 8.6% 15.2% 9.6% 9.8% 18.5% 22.2% 16.0% 7.8% 13.1% 11.1% 10.4% 14.4% 22.1% 21.1% χ2=1146‡ (12 df) Diagnostic category 1. Cancer 2. Stroke 3. Severe acute heart disease 4. Metabolic and electrolyte disorders 5. Pulmonary disease 6. Ophthalmologic disease 7. Low risk heart disease 8. Gastrointestinal disease 9. Urologic disease 10. Musculoskeletal, connective tissue disease 11. Infectious diseases 12. Symptoms and ill-defined conditions 13. Severe trauma 14. Follow-up care without acute diagnosis 15. Pregnancy, childbirth, puerperium 16. Others 4.4% 1.4% 2.6% 1.3% 3.3% 2.0% 3.5% 6.4% 1.9% 1.8% 1.0% 3.2% 2.9% 34.0% 10.8% 19.5% 4.1% 1.8% 3.0% 1.6% 2.9% 1.2% 2.9% 8.9% 1.7% 3.4% 0.7% 3.0% 2.1% 30.1% 14.1% 18.5% 5.3% 1.6% 3.2% 1.1% 4.2% 0.5% 4.6% 10.2% 1.9% 3.2% 1.0% 6.7% 1.8% 23.4% 11.0% 20.5% χ2=1562‡ (50 df) *Hospital 1 had the medical emergency team. †Test for any difference between the three hospitals. ‡P Back to text 4: Comparisons of total event rates by hospitals Event n Crude rates/10000 Unadjusted ORs Adjusted ORs* Cardiac arrest Hospital 1 Hospital 2 Hospital 3 69 66 99 38 51 51 1.00 1.34 (0.96-1.89) 1.35 (0.99-1.83) 1.00 1.14 (0.81-1.61) 1.00 (0.73-1.37) Death Hospital 1 Hospital 2 Hospital 3 243 240 295 133 184 151 1.00 1.39 (1.16-1.67) 1.14 (0.96-1.35) 1.00 1.08 (0.89-1.30) 0.83 (0.70-1.00) Non-DNR death Hospital 1 Hospital 2 Hospital 3 55 86 88 30 66 45 1.00 2.20 (1.57-3.09) 1.50 (1.07-2.11) 1.00 1.68 (1.19-2.36) 0.94 (0.67-1.33) Unanticipated ICU/HDU admission Hospital 1 Hospital 2 Hospital 3 118 146 234 64 112 120 1.00 1.73 (1.36-2.21) 1.86 (1.49-2.32) 1.00 1.59 (1.24-2.04) 1.73 (1.37-2.16) *Odds ratios (ORs) adjusted for patient characteristics and diagnostic categories. Hospital 1 (which has the medical emergency team) is the reference for the ORs. For shaded ORs, 95% CIs do not cross 1.0. DNR="do not resuscitate" order documented. ICU=intensive care unit. HDU=high dependency unit. Back to text 5: Comparisons of index event rates by hospitals Event n Crude rates/10000 Unadjusted ORs Adjusted ORs* Cardiac arrest Hospital 1 Hospital 2 Hospital 3 60 63 84 33 48 43 1.00 1.48 (1.04-2.10) 1.31 (0.94-1.83) 1.00 1.24 (0.87-1.78) 0.96 (0.69-1.35) Death Hospital 1 Hospital 2 Hospital 3 119 139 191 65 106 98 1.00 1.65 (1.29-2.11) 1.51 (1.20-1.90) 1.00 1.24 (0.97-1.60) 1.05 (0.82-1.33) Unanticipated ICU/HDU admission Hospital 1 Hospital 2 Hospital 3 82 140 222 45 107 114 1.00 2.41 (1.83-3.17) 2.56 (1.98-3.30) 1.00 2.17 (1.65-2.87) 2.35 (1.82-3.04) *Odds ratios (ORs) adjusted for patient characteristics and diagnostic categories. Hospital 1 (which has the medical emergency team) is the reference for the ORs. For shaded ORs, 95% CIs do not cross 1.0. ICU=intensive care unit. HDU=high dependency unit. Back to text 6: An example of how the odds ratios and C statistic changed as variables were added stepwise to the model for total cardiac arrests Odds ratio* (95% CI) Hospital 2 Hospital 3 C statistic Crude odds ratio Admission via emergency department Age ≥75 years Severe acute heart disease Low risk heart disease Infectious disease Pulmonary disease 1.34 (0.96-1.89) 1.16 (0.83-1.63) 1.06 (0.75-1.49) 1.06 (0.75-1.49) 1.07 (0.76-1.50) 1.09 (0.77-1.53) 1.14 (0.81-1.61) 1.35 (0.99-1.83) 1.19 (0.87-1.62) 0.98 (0.72-1.34) 0.99 (0.72-1.35) 0.99 (0.72-1.35) 1.00 (0.73-1.37) 1.00 (0.73-1.37) 0.533 0.755 0.798 0.809 0.826 0.833 0.850 *Hospital 1 is the reference for the odds ratios. Back to text

Peter J Bristow · Ken M Hillman · Tien Chey · Kathy Daffurn · Theresa C Jacques · Sandra L Norman · Gillian F Bishop

Research Characteristics and outcomes of older patients presenting to the emergency department after a fall: a retrospective analysis Anthony J Bell, Janet K Talbot-Stern and Annemarie Hennessy MJA 2000; 173: 179-182 For editorial comment, see Close & Glucksman Abstract - Methods - Results - Discussion - References - Authors' details - - More articles on Emergency medicine Abstract Objectives: To study older patients presenting to the emergency department after a fall -- factors associated with the fall, injuries sustained and outcome. Design: A retrospective analysis using the Emergency Department Information System (EDIS), the Trauma Registry and the patient information database (CCIS), in addition to the patient's emergency and inpatient medical records. Setting: Emergency department of a major inner city teaching hospital, 1 June - 30 November 1997. Patients: All patients over 65 years presenting to the emergency department (ED) after a fall, for whom complete medical records were available. Results: Of 803 patients over 65 years presenting to the ED after a fall, complete records were available for 733 (91.3%) (283 men and 450 women). Extrinsic (accidental) causes were implicated in more than a third of falls (313 patients [42.7%]). A high proportion of the patients were living at home (520; 70.9%) and walking unaided (389; 53.1%). Although absolute numbers of women increased with age, men were as likely as women to present after a fall. Many patients had fallen before -- 39% of the men (111/283) and 24% of the women (110/450). In 78 patients (10.6%), alcohol misuse may have been a direct cause of the fall. The overall injury rate was 70.5% (517/733 patients), the most common injury being an isolated fracture (269/517 patients; 52.0%). In all, 419 patients (57.2%) were admitted to hospital, 48% (200/419) with a fracture and 52% (219/419) for investigation of the medical cause of the fall. The median length of hospital stay was 6 days (mean, 10.4 days; range, 1-129 days); 35% (146/419) of patients were in hospital for more than 10 days. Conclusion: Older patients presenting to the ED after a fall had high injury rates, high admission rates and often prolonged hospitalisation. About a third had fallen before. Patients at risk can be identified in the ED and referred to falls prevention programs. Census data for 1996 show that 12.1% of Australians are aged 65 years or over.1 This proportion is expected to double in the next 40 years,2 with major implications for healthcare costs. Alone, the cost of falls in patients over 70 years in Australia was estimated to be $398 million in 1989.3In the United States, trauma causes a considerable proportion of presentations (and subsequent hospital admissions) of older patients. Falls account for most of these presentations.4 The annual incidence of all falls increases from 25% at age 70 years to 35% after the age of 75; the risk increases with age and is higher among those living in long-stay institutions.5 Up to 10%-15% of falls result in serious injury, of which at least half are fractures. Even falls not resulting in injury may have serious psychological consequences.5,6 The "postfall anxiety syndrome"7 and fear of falling leads to decreased activity,8 and ultimately an increased risk of future falls.9 Patients have reported continued disability two months after a fall.10 No Australian report has been published specifically about patients in this age group presenting to the emergency department (ED) after falls, although previous studies have looked at older people presenting to the ED.11,12 Our aim was therefore to focus on patients over 65 years who presented to our ED as a result of a fall. Several features were of interest: why the patients fell; what, if any, injuries were sustained; what proportion of patients required admission to hospital; and what morbidity and mortality resulted from the fall. Methods Royal Prince Alfred Hospital is a 700-bed tertiary referral centre with 60 000 admissions and 45 000 ED attendances per year. A retrospective review of attendances for the six-month period June - November 1997 was undertaken. All older patients who had fallen were eligible for the study. Patient data Data were obtained from the sources below and thereafter patients remained anonymous. EDIS: Patients eligible for the study were identified by a search of the EDIS (Emergency Department Information System) for "falls" in the age group chosen. EDIS is a computerised database in the ED with demographic information, presenting complaint, diagnosis and disposition for each patient. Medical records: A predetermined dataset was recorded from the medical record for each patient presenting to the ED. This included medical record number, age, sex, type of residence (home, hostel or nursing home), prefall mobility, nature of fall, alcohol misuse, recurrent fall, referral status, triage category, injury score, specific area of the body injured, fracture, admission, specialty, length of stay, mortality, and discharge disposition. Prefall mobility was further defined as unaided versus aided (use of a stick, frame, crutches, assistance by another person) versus unknown. Trauma Registry: Additional data were obtained from the hospital's Trauma Registry. An Injury Severity Score (ISS) is calculated for patients requiring admission after trauma. ISS is the sum of the squares of the highest Abbreviated Injury Scores (an anatomical system classifying injuries by body region on a scale of 1 [minor] to 6 [serious]) for the three most seriously injured body regions. ISS ranges from 1 (minor injury) to 75 (severe injury).13 CCIS: For patients transferred to an affiliated geriatric and rehabilitation hospital, the patient information database (CCIS [Central Sydney Area Health Service Clinical Information System]) was accessed for the length of stay. None of the patients in our study were transferred to non-affiliated geriatric and rehabilitation hospitals. Population data: The Australian Bureau of Statistics supplied population data for the hospital's catchment area.14 Definitions Fall: "Inadvertently coming to rest on the ground or other lower level with or without loss of consciousness."15 Extrinsic (accidental) causes: Environmental factors (eg, rugs, steps, uneven floors). Falls as a result of external trauma, such as motor vehicle accidents and violence, were excluded. Intrinsic (non-accidental) causes: Syncope, dizziness or vertigo, postural drop, central nervous system lesion (haemorrhage or infarct), drop attack, and balance or gait disturbance. Alcohol misuse: A history of alcohol misuse related temporally to the event, a record of alcohol on the breath, or a statement in the ED record about the patient's being intoxicated. Statistical analysis We used Minitab Statistical Software16 for statistical analysis and performed χ2 tests. Analysis was based on age group or sex and compared with a number of variables: presentation as a result of a fall, nature of the fall, outcome of a fracture, and admission status. A multivariate analysis was performed on four aspects of the falls considered to be related to place of residence or mobility: extrinsic cause, recurrent falls, fracture/no fracture and admission. Odds ratios (95% CI) were calculated for each of these groups. Multivariate analysis was also used to calculate odds ratios (95% CI) for whether alcohol use contributed to selected outcomes: admission (yes/no), extrinsic or recurrent falls versus other falls, and age under or over 80 years. Results Patient characteristics Of a total of 22 782 patients presenting to the ED during the six-month study period, 4489 (19.7%) were patients older than 65 years and 803 (17.8%) of these patients presented as a direct consequence of a fall. Of these patients, 733 (91.3%) had medical records available for review at the time of analysis and complete for the purposes of the dataset. Age and sex: The average age was 78.6 years (range, 65-101 years) and the median age was 79 years: 263 patients were aged 65-74 years, 279 were 75-84 years and 191 were 85 years or older. Increasing age of the patients was associated with presenting to the ED as a result of a fall (χ2 test for trend, P < 0.001) (Box 1). There were 283 men and 450 women. However, the number of men and women presenting to the ED after a fall reflected the age and sex distribution within the catchment population (Box 2). Thus, men were as likely as women to present as a result of a fall. Residence: At the time of the fall, 83% (211/253) of the 65-74 year olds, 74% (200/269) of the 75-84 year olds and 57% (109/191) of those over 85 years were living in their own homes. Thus, the proportion of those living in either a hostel or a nursing home increased with advancing age. In 20 patients residence could not be classified. Previous falls: 39% of the men (111/283) and 24% of the women (110/450) had fallen before. Mobility: Patients were classified according to mobility: walking aided or unaided. As expected, as the patients aged the use of a walking aid increased. Cause of fall Extrinsic or intrinsic: Overall, extrinsic causes for the fall accounted for 42.7% of patients presenting to the ED. In the age group 65-74 years extrinsic causes accounted for 49.4% of falls, which is more than expected when compared with the proportion in the older age groups (39.0% and 38.7%, respectively). Intrinsic causes were more likely with advancing age (χ2 test; P = 0.018) and accounted for 50.5% (95% CI, 45%-57%), 60.9% (95% CI, 55%-67%) and 64.2% (95% CI, 54%-68%) of falls in the respective age groups. The breakdown of all causes for falls presenting to the ED is shown in Box 3. Despite extensive review of the medical records we were unable to classify 23% of falls as either extrinsic or intrinsic. Alcohol misuse: This was documented in 78 patients (10.6%): 18% of the 65-74 year olds, 10% of the 75-84 year olds, and was not a factor in those over 85 years (χ2 test; P = 0.001). Sixty-five (83%) of these patients were living in their own homes. Multivariate analysis for alcohol misuse at the time of fall showed it to be significantly associated with an increased risk of both accidental and recurrent falls (Box 4). Outcomes Injury: 517 (70.5%) patients sustained an injury as a result of the fall: 73.3% (379/517) had an ISS of 4 or less (a score of 9 correlated with a femoral fracture); 13 patients had scores between 15 and 25, with all of these patients (except one with spinal cord compression) sustaining intracranial injury. The most common injuries were fractures (36.7%), soft tissue injuries (16%), lacerations and skin tears (14.5%). Fracture: 269 patients (36.7%) sustained a fracture: 36% (98/269) of which were neck-of-femur fractures, 16% fractured wrists, 12% fractured humeral neck and 5% pelvic fractures. The breakdown of fractures in each group is shown in Box 5. Women sustained both neck-of-femur and all fractures more frequently than men (χ2 test; P < 0.001): 64% (63/98) of femoral-neck fractures and 73% (125/171) of all other fractures (95% CI, 66%-80%). Interestingly, in women, the proportion of fractured neck of femur to all fractures was 33.5% (63/188) (95% CI, 27%-40%), whereas in men it was 43% (35/81) (95% CI, 32%-54%). Fracture rate overall was not found to be related to advancing age in either sex. Admission: The total number of patients admitted to hospital was 419, or 57.2% of all older patients with falls (representing 38% of all older patients admitted during the study period). Sixty-three per cent of those 85 years or older were admitted, compared with 60% of the 75-84 year olds and 50% of the 65-74 year olds (χ2 test; P = 0.009). Of the 269 patients with fractures, 200 (74%) were admitted. There was no statistically significant difference in the fracture admission rate across the age groups (χ2 test; P = 0.53). Of the 200 patients admitted, in 49% the cause of the fracture was intrinsic. Patients admitted to hospital after a fall had a mean length of stay of 10.4 days (95% CI, 10.2-10.6) and a median stay of 6 days (range, 1-129 days). Hospitalisation for more than 10 days was necessary in 35% (146/419) of patients. Deaths: Thirty-two patients died in hospital, representing 4.4% of all patients presenting to the ED after a fall: half of those who died were over 85 years of age and half were from nursing homes. In those who died, the cause of the fall was intrinsic rather than extrinsic (27/32), and the most common injury was a fracture of the neck of the femur (10/32). Data analysis: Multivariate analysis of place of residence or mobility and extrinsic cause, recurrent falls, fracture/no fracture and admission showed no significant interaction. Discussion We found that older patients presenting to the ED after a fall had a high injury rate (71%), high admission rates (57%) and often prolonged hospitalisation (> 10 days in about a third of those admitted). Our study complements others performed in Australia and elsewhere on older patients who fall, particularly those who present to an ED.11,12Some studies have found that women in the community fall more frequently than men,17 and others, as we did, found no difference.18 Institutionalised patients have been reported to have higher fall rates than patients living at home,17,19 but most of our patients lived at home and walked unaided. Falls may be caused by an environmental hazard alone or a simple syncopal event, or there may be a complex interaction of environment, physical illness, and type of activity. Changes in vision, vestibular function and proprioception affect physical stability, and musculoskeletal changes affect gait. Postural hypotension from dehydration, drug effects or autonomic dysfunction may be involved. Additionally, acute illness such as respiratory tract infection, arrhythmias, carotid sinus hypersensitivity,20 cardiac failure and neurological problems (eg, Parkinson's disease) may increase the risk of falling. All these intrinsic factors may be compounded by environmental hazards.5,6,17 We found gait disturbance, syncope, central nervous system lesion, postural hypotension and dizziness to be the most common intrinsic causes, and these were statistically more likely to be the underlying reason for a fall as age increased. The proportion of patients with falls in association with alcohol misuse contrasts with the findings of Adams et al.21 They surveyed older patients over an eight-week period for alcohol use, and found a negative relationship between alcohol use and falls. A high proportion of our patients with alcohol misuse lived at home, with perhaps easy access to alcohol. These patients had a greater risk of extrinsic and recurrent falls, a potential relationship that warrants further study. A UK study found that most falls in the community do not result in serious injury.17 We found that patients presenting to the ED after a fall have a high rate of injuries, consistent with previous reports,17,22 but the rate was significantly higher than that found by Tinetti et al.23 We found women to be statistically more likely to suffer a fracture than men. Grisso et al,10 in an older inner-city population in the United States, found that women generally had higher rates of fall injury than men. In addition, they found that injury rates increased with advancing age, a finding that we could not confirm. There were fewer hip fractures in older men than older women in our study, confirming previous findings.24 This is probably related to the higher prevalence of osteoporosis in women. Previous reports have shown that older men with hip fracture have higher mortality rates than age-matched women.23 The high admission rate in our study, which increased in older patients, is only slightly higher than that found by Richardson,11 but this was in patients over 75 years, in whom a higher admission rate is expected. A UK study found admission was needed in only 34% of patients.22 Admission rates for patients with a fracture did not vary significantly across our three age groups, nor were they different according to place of residence or prefall mobility. Length of hospital stay similarly did not depend on place of residence or prefall mobility, differing from the Richardson study, in which a significant relationship was found between accommodation status and outcome at 90 days.11 US studies report that 75% of deaths after a fall occur in patients over 65 years.6 We found that the single most important factor associated with death was hip fracture, a finding similar to that in previous studies.7,11 Modification of the environment and dealing with intrinsic problems such as drug side effects and gait dysfunction can reduce falls,25-27 prevent hospitalisation26 and shorten length of stay.15 If 95% of problems can be identified from the history and physical examination alone, as suggested by Rubenstein et al,15 the emergency physician is well able to identify those patients at risk of further falls. Intrinsic causes can be treated and the patient's general practitioner or specific falls prevention programs can then proceed to modify the risk of recurrence. References Australian Bureau of Statistics. Australia in brief (Census data, 1996). Canberra: ABS, 1998. <www.abs.gov.au> Davis JA. Older Australia: a positive view of ageing. Sydney: Harcourt Brace, 1994. Smith RD, Widiatmoko D. The cost-effectiveness of home assessment and modification to reduce falls in the elderly. Aust N Z J Public Health 1998; 22: 436-440. Spaite DW, Criss EA, Valenzuela TD, et al. Geriatric injury: an analysis of prehospital demographics, mechanisms and patterns. Ann Emerg Med 1990; 19: 1418-1421. Tinetti ME, Speechley M. Prevention of falls among the elderly. N Engl J Med 1989; 320: 1055-1059. Nelson RC, Murlidhar AA. Falls in the elderly. Emerg Med Clin North Am 1990; 8: 309-324. Rubenstein LZ, Josephson KR, Robbins AS. Falls in the nursing home. Ann Intern Med 1994; 121: 442-451. Nevitt MC, Cummings SR, Kidd S, Black D. Risk factors for recurrent nonsyncopal falls: a prospective study. JAMA 1989; 261: 2663-2668. Gostynski M, Ajdacic-Gross V, Gutzwiler F, Michel JP. Epidemiological analysis of accidental falls by the elderly in Zurich and Geneva. Schweiz Med Wochenschr 1999; 129: 270-275. Grisso JA, Schwarz DF, Wishner AR, et al. Injuries in an elderly inner city population. J Am Geriatr Soc 1990; 38: 1326-1331. Richardson DB. Elderly patients in the emergency department: a prospective study of characteristics and outcome. Med J Aust 1992; 157: 234-239. Stathers GM, Delpech V, Raftos JR. Factors influencing the presentation and care of elderly people in the Emergency Department. Med J Aust 1992; 156: 197-200. Baker SP, O'Neill B, Haddon W. The Injury Severity Score. J Trauma 1974; 14: 187. Needs Assessment and Health Outcomes Unit. A demographic profile of the Central Sydney Area Health Service from the 1996 Census. Sydney: Central Sydney Area Health Service, March 1998. Rubenstein LZ, Robbins AS, Josephson KR, Schulman BL. The value of assessing falls in an elderly population: a randomised clinical trial. Ann Intern Med 1990, 113: 308-316. Minitab Statistical Software [computer program], version 12. State College, Pa: Minitab Inc, 1998. Blake AJ. Falls in the elderly. Br J Hosp Med 1992; 47: 268-272. Campbell AJ, Borrie MJ, Spears GF, et al. Circumstances and consequences of falls experienced by a community population 70 years and over in a prospective trial. Age Ageing 1990; 19: 136-141. Cummings SR, Nevitt MC. Falls [editorial]. N Engl J Med 1993; 331: 872-873. Ward CR, McIntosh S, Kenny RA. Carotid sinus hyersensitivity -- a modifiable risk factor for fractured neck of femur. Age Ageing 1999; 28: 127-133. Adams WL, Magruder-Habib K, Trued S, Broome HL. Alcohol abuse in elderly Emergency Department patients. J Am Geriatr Soc 1992; 40: 1236-1240. Davies AJ, Kenny RA. Falls presenting to the Accident and Emergency Department: types of presentation and risk factor profile. Age Ageing 1996; 25: 362-366. Tinetti ME, Speechley M, Ginter SF. Risk factors for falls among elderly persons living in the community. N Engl J Med 1988; 319: 1701-1707. Diamond TH, Thornley SW, Sekel R, Smerdely P. Hip fracture in elderly men: prognostic factors and outcomes. Med J Aust 1997; 167: 412-414. Province MA, Hadley EC, Hornbrook MC, Lipsitz LA. The effects of exercise on falls in elderly patients: a preplanned meta-analysis of the FICSIT trials. JAMA 1995; 273: 1341-1347. Close J, Ellis M, Hooper R, Glucksman E. Prevention of falls in the elderly trial (PROFET): a randomised controlled trial. Lancet 1999; 353: 93-97. Tinetti ME, Baker DI, McAvay G, Claus EB. A multifactorial intervention to reduce the risk of falling among elderly people living in the community. N Engl J Med 1994; 331: 821-827. (Received 10 Aug 1999, accepted 29 May 2000) Authors' details Department of Emergency Medicine, Royal Prince Alfred Hospital, Sydney, NSW. Anthony J Bell, MB BS, Emergency Medicine Registrar. Janet K Talbot-Stern, MD, FACEM, FACEP, Director, Emergency Department; and Clinical Senior Lecturer, Department of Surgery, University of Sydney. Department of Medicine, University of Sydney, Sydney, NSW. Annemarie Hennessy, MB BS, PhD, Senior Lecturer. Reprints will not be available from the authors. Correspondence: Dr A J Bell, Department of Emergency Medicine, Royal Prince Alfred Hospital, Missenden Road, Camperdown, NSW 2050. Make a comment 1: Patients presenting to the emergency department, by age group, June - November, 1997 65-74 years (n=2060) 75-84 years (n=1672) ≥85 years (n=757) Total (n=4489) Presentation after a fall Other presentations 295 (14.3%) 1765 317 (19.0%) 1355 191 (25.2%) 566 803 (17.9%) 3686 χ2 test for age trend (P<0.001). Back to text 2: Age and sex distribution of patients presenting to the emergency department after a fall compared with the catchment population 65-74 years 75-84 years ≥85 years Men Presentation after a fall Proportion of catchment population 124/263 (47%) 15415/32185 (47.9%) 107/279 (38%) 7251/18448 (39.3%) 52/191 (27%) 1650/6038 (27.3%) Women Presentation after a fall Proportion of catchment population 139/263 (53%) 15770/32185 (49.0%) 172/279 (62%) 11187/18448 (60.6%) 139/191 (73%) 4388/6038 (72.7%) Back to text Back to text 4: Multivariate analysis (logistic regression) of alcohol misuse and selected variables in older patients presenting to the emergency department after a fall Variable Alcohol misuse odds ratio (95% CI) Age at presentation 5.5 (2.8-10.6) Extrinsic cause of fall 1.72 (1.05-2.83) Recurrent falls 2.24 (1.35-3.72) Back to text 5: Fractures in older patients presenting to the emergency department after a fall, by age group (years) Fracture 65-74 (n=263) 74-85 (n=279) >85 (n=191) Total (n=733) Neck of femur 28 (11%) 37 (13%) 33 (17%) 98 (13.4%) Other 74 (28%) 59 (21%) 38 (20%) 171 (23.3%) No fracture 161 (61%) 183 (66%) 120 (63%) 464 (63.3%) Back to text

Anthony J Bell · Janet K Talbot-Stern · Annemarie Hennessy

Research Moderate alcohol intake is associated with survival in the elderly: the Dubbo Study Leon A Simons, John McCallum, Yechiel Friedlander, Michael Ortiz and Judith Simons MJA 2000; 173: 121-124 For editorial comment, see Stockwell Abstract - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - More articles on Cardiology and cardiac surgery Abstract Objective: To examine the relationship between alcohol intake and survival in elderly people. Design and setting: A prospective study over 116 months of non-institutionalised subjects living in Dubbo, a rural town (population, 34 000) in New South Wales. Participants: 1235 men and 1570 women aged 60 years and over who were first examined in 1988-89. Main outcome measures: All-causes mortality; gross cost of alcohol per life-year gained. Results: Death occurred in 450 men and 392 women. Intake of alcohol was generally moderate (ie, less than 14 drinks/week). Any intake of alcohol was associated with reduced mortality in men up to 75 years and in women over 64 years. In a proportional hazards model, the hazard ratio for mortality in men taking any alcohol was 0.63 (95% CI, 0.47-0.84) and in women was 0.75 (95% CI, 0.60-0.94). Cardiovascular deaths in men were reduced from 20/100 (95% CI, 14-26) to 11/100 (95% CI, 9-13) and in women from 16/100 (95% CI, 13-19) to 8/100 (95% CI, 6-10). The reduction in mortality occurred in men and women taking only 1-7 drinks/week -- hazard ratios, 0.68 (95% CI, 0.49-0.94) and 0.78 (95% CI, 0.61-0.99), respectively, with a similar protective effect from intake of beer or other forms of alcohol. After almost 10 years' follow-up, men taking any alcohol lived on average 7.6 months longer, and women on average 2.7 months longer, compared with non-drinkers. The gross cost for alcohol per life-year gained if consuming 1-7 drinks/week was $5700 in men, and $19 000 in women. Conclusions: Moderate alcohol intake in the elderly appears to be associated with significantly longer survival in men 60-74 years and in all elderly women. The consumption of moderate amounts of alcohol, compared with abstention or with heavy alcohol intake, appears to be associated with reduced all-causes mortality in middle-aged subjects.1-3 This effect may be partially mediated through a reduced risk of coronary heart disease (CHD)4 and stroke.5 Some studies attribute the protection to a specific effect of wine;6,7 other studies attribute it to any type of alcohol.8In elderly people, some of this benefit from moderate alcohol intake may be negated by mortality from other causes.9 In a prospective study of men and women aged 65 years and over in the United States (Established Populations for Epidemiologic Studies of the Elderly), alcohol intake under 21 drinks/week was associated with a 30%-40% lower all-causes mortality in two cohorts, but with no influence in a third cohort.10 In Australian men and women aged 60 years and over with 77 months' follow-up, the intake of 1-7 drinks/week was associated with a 22%-25% reduction in all-causes mortality, although this reduction did not achieve statistical significance.11 We have examined the relationship between alcohol intake and survival in this Australian cohort during a more extended follow-up of 116 months. We present the results and include an economic analysis of moderate alcohol intake. Methods Dubbo Study The Dubbo Study is an ongoing prospective examination of cardiovascular and other diseases in an elderly Australian cohort first examined in 1988-89. All non-institutionalised residents of Dubbo, New South Wales, born before 1930 were eligible; participation rate was 73% (1235 men and 1570 women). Methods and measures have already been described in detail.12,13 The baseline examinations comprised demographic, psychosocial and standard cardiovascular risk assessments, including examination of fasting blood samples. Alcohol usage Questions on alcohol usage were those asked in the National Heart Foundation Risk Factor Prevalence Study,14 and yielded an approximation of usual alcohol intake coded as zero, 1-7, 8-14, 15-28 and more than 28 drinks/week (referring to a standard drink containing 10 g of alcohol). Specific intakes of beer, wine and spirits were not separately sought, but subjects were asked whether they normally drank beer or not, allowing a separation of drinking behaviour into beer and "other". Survival analysis Outcomes from August 1988 to 31 December 1998 were included in the analysis, a median 116 months' follow-up. Hospitalisation and death records were monitored continuously, and postal surveys were conducted every two years to confirm vital status. The survey in 1997 successfully traced more than 98% of surviving participants. Records were coded according to the International classification of diseases, ninth revision, clinical modification (ICD-9-CM). The independent contribution of any risk factor to mortality was examined in a Cox proportional hazards model. Point estimates and 95% CIs for the relative hazard of death were calculated from the regression coefficients (presented as hazard ratio, a measure of relative risk). The models included categories of alcohol intake as described above, and, where relevant, a categorical term describing whether a subject normally drank beer or "other" (ie, wine/spirits). The proportional hazards model assumes constant relative hazard over the length of follow-up. This assumption was confirmed for each model by a plot of log-minus-log survival, demonstrating parallel curves over all categories of alcohol intake. Statistical analyses were conducted using SPSS for Windows NT.15 Economic analysis A weighted alcohol acquisition cost per week using the midpoint of the intake ranges was calculated for each alcohol intake stratum and for each sex using 1999 Dubbo club prices. Individuals were assumed to remain in their initial consumption strata over the whole time period. Survival curve data from the Cox model were used to calculate total expenditure on alcohol, as well as survival benefit for each stratum of intake. An incremental analysis (difference in cost/difference in benefit) was conducted using the no-alcohol-consumption stratum as the reference. This yielded an estimate of the gross cost per life-year gained. Although the study collects hospitalisation records, hospitalisation costs were not available and a net cost per life-year gained could not be estimated. Ethical approval The study was approved by the institutional ethics committees at St Vincent's Hospital, Sydney, the University of New South Wales and the Australian National University. All participants gave informed, written consent. Results Pattern of alcohol intake The pattern of alcohol intake and its clinical associations have been fully documented in an earlier report.11 The pattern of alcohol intake, by quantity and type, is shown in Box 1. On a day when alcohol was consumed, 40% of all men and 45% of all women took one or two drinks, 23% and 7% took three or four drinks, and 15% and 1%, respectively, took five or more drinks. All-causes mortality Death occurred in 450 men (36%) and 392 women (25%). Consumption of more than 14 drinks/week was uncommon. Hence, where relevant, the use of alcohol has been grouped into zero use and any use. Alcohol intake: Age-specific all-causes mortality by alcohol intake is presented in Box 2. Alcohol use in men appeared to be associated with reduced mortality up to age 74 years, but not beyond. In women, its use was associated with reduced mortality in all groups older than 64 years. Predictors of all-causes mortality: The independent contribution of alcohol to all-causes mortality was explored in men 60-74 years and in all women in proportional hazards models which adjusted for the presence of major demographic, psychosocial and cardiovascular variables at study entry. The significant predictors of all-causes mortality are summarised in Box 3. (Alcohol intake was not a significant predictor of mortality in men aged more than 74 years.) Any alcohol intake was significantly associated with reduced all-causes mortality in both sexes. Quantity or type of alcohol intake: The relationship between quantity or type of alcohol intake and all-causes mortality in the proportional hazards model is presented in Box 3. The risk of mortality was significantly reduced at all levels of alcohol intake, except in women taking 15-28 drinks/week (representing only 3% of women in the study). A similar degree of reduction in all-causes mortality was observed at all levels of alcohol intake. The protection observed was broadly similar in those using beer versus wine/spirits, although this only reached statistical significance for beer consumption. Pattern of alcohol intake: In a subsequent model, the quantity of alcohol consumed per week was replaced by a variable denoting the usual number of drinks taken on a given day, a measure of the pattern of drinking. Using zero intake as the reference group, the hazard ratio in men with a consumption of one or two drinks on a given day was 0.64 (95% CI, 0.46-0.89), with three or four drinks 0.68 (95% CI, 0.47-0.98), and with five or more drinks 0.69 (95% CI, 0.45-1.06). The corresponding hazard ratios in women were 0.74 (95% CI, 0.59-0.93), 0.68 (95% CI, 0.38-1.22) and 1.38 (95% CI, 0.58-3.29) (there were only 17 women in this group). Hazard curves: The hazard curves calculated from the proportional hazards models in men and women are presented in Box 4. By the end of almost 10 years' follow-up, men taking any alcohol lived on average 7.6 months longer, and women on average 2.7 months longer, than their counterparts taking no alcohol. Specific causes of death: Cardiovascular death (ie, CHD and stroke) was reduced from 20/100 (95% CI, 14-26) in non-drinkers to 11/100 (95% CI, 9-13) in men taking any alcohol. The corresponding reduction in cardiovascular death in women was from 16/100 (95% CI, 13-19) to 8/100 (95% CI, 6-10). Deaths attributed to any cancer were unchanged in men (6/100 [95% CI, 3-9] versus 7/100 [95% CI, 5-9]) and in women (4/100 [95% CI, 2-6] versus 5/100 [95% CI, 3-7]). Gross cost for alcohol per life-year gained From the number of life-years added to survival for each quantity of intake, we have estimated the gross cost for alcohol per life-year gained. In men 60-74 years and in women 60 years and over taking 1-28 drinks/week, the respective costs were $13 000 and $31 000 per life-year gained. Since much of the benefit from alcohol intake was observed at a moderate intake of only 1-7 drinks/week, the respective costs at this intake were $5700 and $19 000 per life-year gained. Discussion In this well-defined, community-based sample of rural elderly Australians, alcohol intake could be described as moderate rather than heavy.11 Our results confirm that any intake of alcohol is associated with significantly reduced all-causes mortality in men 60-74 years and in all elderly women, consistent with our previous report at 77 months' follow-up.11 Statistical significance has now been reached due to the greater number of deaths and increased statistical power. Absolute death rates in men were substantially higher than in women, especially in the "young old". This would account for the greater average survival advantage shown in Box 4 (7.6 months versus 2.7 months). Equivalent reductions in mortality occurred at 1-7 drinks/week and at higher intakes. In men there was no evidence of a differential effect between 1-2 drinks on a given day and an intake of five or more drinks on a given day. Not all studies have documented an association between all-causes mortality and alcohol intake.16,17 The reduction in CHD and stroke mortality associated with alcohol use we observed is consistent with findings of previous reports,8,18,19 as well as those of recent reports in large middle-aged cohorts from the United States.4,5,20 These data are gradually causing health authorities to reconsider public policy on moderate alcohol intake, say 1-7 drinks/week, in the prevention of future morbidity and mortality.21 Studies in younger populations indicate a "U"- or "J"-shaped relationship between alcohol intake and all-causes mortality.2 This has not been a general finding in the elderly, possibly because these cohorts contain an excess of "healthy survivors".10 It is surprising that we could find no relationship between alcohol intake and mortality in men aged over 74 years, but it is plausible that very elderly men lose the benefit of alcohol intake because they become subject to competing causes of mortality.9 There is a potential for misclassification of alcohol intake between zero and low intake because of under-reporting. This would diminish any apparent relative benefit of alcohol intake on all-causes mortality. Hence, our statistically significant findings may represent a minimum estimate of the benefit of moderate alcohol intake. We have demonstrated essentially a "threshold effect" between alcohol intake and all-causes mortality in either sex. Protection does not improve greatly as alcohol intake increases further (Box 3). Thus, under-reporting would then have less impact on our findings. This threshold effect is important, particularly if we were to move to a public health position of suggesting that abstainers should begin to imbibe! The economic findings take no account of any changes in healthcare costs arising downstream through the benefits or otherwise of alcohol intake. What we regard as an acceptable cost per life-year gained is arbitrary, but it is informative to compare our calculated costs with, for example, recently published gross costs for the use of simvastatin in patients with established CHD, with survival increased by around 20%.22 Assuming such patients receive lifetime therapy with simvastatin from their mid-50s, the gross cost per life-year gained in the UK population would be £5100 (or about $13 000). Although one is comparing unrelated "therapy", moderate alcohol intake, at least in older men, may turn out to be a popular and cost-effective means of improving survival which does not require government subsidy! The relative merits of wine versus other forms of alcohol consumption remain controversial. Data from France6 and Denmark7 highlight specific benefits of wine, and this has generated a new research effort to identify which components of wine, apart from alcohol, may be the most beneficial. Antioxidants are among the most prominent suggestions.23 Others consider that alcohol in any form gives protection against cardiovascular disease,8 largely through its effect in raising high density lipoprotein (HDL) cholesterol levels.11 In Dubbo, the quantity of alcohol intake was highly correlated with HDL cholesterol (r = 0.32, P < 0.001 and r = 0.23, P < 0.001 in men and women, respectively). Other suggested mechanisms for cardiovascular protection include favourable effects of alcohol on thrombotic and fibrinolytic pathways, reduced insulin resistance and improved endothelial function through increased nitric oxide production.5 Alcohol may also influence survival in ways which currently defy measurement: regular alcohol intake may reflect a special lifestyle integrated with less tangible factors.21 Although excess alcohol intake is undoubtedly toxic to the central nervous system, recent studies suggest that a moderate intake may reduce the risk of dementia. In a study of elderly French people 65 years and over, appropriately from Bordeaux, the rate of hospitalisation for dementia in non-drinkers over three years was 4.9/100, but only 3.9/100 in those taking any alcohol (with a more striking effect on the risk of Alzheimer's disease).24 During 116 months' follow-up in the Dubbo population, the respective rates of hospitalisation for dementia were 4.3/100 and 2.5/100 (P < 0.01). It is premature to promote the use of alcohol for prevention of dementia, but the 21st century may witness a completely new role for alcohol in health. Acknowledgements The Dubbo Study is supported in part by grants from the National Health and Medical Research Council of Australia, Astra Pharmaceuticals Pty Ltd, Amrad Pharmaceuticals Pty Ltd, Bristol-Myers Squibb Australia Pty Ltd, Merck Sharp & Dohme Australia Pty Ltd, Parke Davis Pty Ltd and Pfizer Pty Ltd. We acknowledge the dedication of the Dubbo Nurse-Manager Kerrie Pearson, the assistance of Gina Brinsmead in economic analysis and Helen Adams in preparation of the manuscript. References Klatsky AL, Armstrong MA, Friedman DD. Alcohol and mortality. Ann Intern Med 1992; 117: 646-654. Holman CDJ, English DR, Milne E, Winter MG. Meta-analysis of alcohol and all-cause mortality: a validation of NHMRC recommendations. Med J Aust 1996; 164: 141-145. Thun MJ, Peto R, Lopez AD, et al. Alcohol consumption and mortality among middle-aged and elderly US adults. N Engl J Med 1997; 337: 1705-1714. Albert CM, Manson JE, Cook NR, et al. Moderate alcohol consumption and the risk of sudden cardiac death among US male physicians. Circulation 1999; 100: 944-950. Berger K, Ajani UA, Kase CS, et al. Light-to-moderate alcohol consumption and the risk of stroke among US male physicians. N Engl J Med 1999; 341: 1557-1564. Renaud S, Geuguen R, Siest G, Salamon R. Wine, beer, and mortality in middle-aged men from Eastern France. Arch Intern Med 1999; 159: 1865-1870. Gronbaek M, Deis A, Sorensen TIA, et al. Mortality associated with moderate intakes of wine, beer or spirits. BMJ 1995; 310: 1165-1169. Rimm EB, Klatsky A, Grobbee D, Stampfer MJ. Review of moderate alcohol consumption and reduced risk of coronary heart disease: is the effect due to beer, wine or spirits? BMJ 1996; 312: 731-735. Van de Water HA, Boshuizen HC. The impact of substitute morbidity and mortality on public health policies. Leiden: TNO Prevention and Health, Division of Public Health and Prevention, 1995. Scherr PA, LaCroix AZ, Wallace RB, et al. Light to moderate alcohol consumption and mortality in the elderly. J Am Geriatr Soc 1992; 40: 651-657. Simons LA, Friedlander Y, McCallum J, Simons J. Alcohol intake and survival in the elderly: a 77 month follow-up in the Dubbo Study. Aust N Z J Med 1996; 26: 662-670. Simons LA, McCallum J, Friedlander Y, et al. Dubbo Study of the elderly: sociological and cardiovascular risk factors at entry. Aust N Z J Med 1991; 21: 701-709. Simons LA, McCallum J, Friedlander Y, Simons J. Predictors of mortality in the prospective Dubbo Study of Australian elderly. Aust N Z J Med 1996; 26: 40-48. Risk Factor Prevalence Study Management Committee. Risk Factor Prevalence Study: Survey No 3 1989. Canberra: National Heart Foundation Australia and Australian Institute of Health, 1990. SPSS for Windows NT [computer program], version 9.0. Chicago, Ill: SPSS Inc, 1999. Leino EV, Romelsjo A, Shoemaker C, et al. Alcohol consumption and mortality. II. Studies of male populations. Addiction 1998; 93: 205-218. Hart CL, Smith GD, Hole DJ, Hawthorne VM. Alcohol consumption and mortality from all causes, coronary heart disease, and stroke: results from a prospective cohort study of Scottish men with 21 years follow up. BMJ 1999; 318: 1725-1729. Hennekens CH, Willett W, Rosner B, et al. Effects of beer, wine and liquor in coronary deaths. JAMA 1979; 242: 1973-1974. Stampfer MJ, Colditz GA, Willett WC, et al. A prospective study of moderate alcohol consumption and the risk of coronary disease and stroke in women. N Engl J Med 1988; 319: 267-273. Sacco RL, Elkind M, Boden-Albala B, et al. The protective effect of moderate alcohol consumption on ischemic stroke. JAMA 1999; 281: 53-60. Hommel M, Jaillard A. Alcohol for stroke prevention? N Engl J Med 1999; 341: 1605-1606. Pickin DM, McCabe CJ, Ramsay LE, et al. Cost effectiveness of HMG-CoA reductase inhibitor (statin) treatment related to the risk of coronary heart disease and cost of treatment. Heart 1999; 82: 325-332. Frankel EN, Kanner J, Germann JB, et al. Inhibition of oxidation of human low-density lipoprotein by phenolic substances in red wine. Lancet 1993; 341: 454-457. Orgogozo J-M, Dartigues J-F, Lafont S, et al. Wine consumption and dementia in the elderly: a prospective community study in the Bordeaux area. Rev Neurol (Paris) 1997; 153: 185-192. (Received 23 Dec 1999, accepted 10 Apr 2000) Authors' details University of New South Wales Lipid Research Department, St Vincent's Hospital, Sydney, NSW. Leon A Simons, MD, FRACP, Associate Professor of Medicine; Judith Simons, MACS, Analyst-Programmer. Faculty of Health, University of Western Sydney MacArthur, Sydney, NSW. John McCallum, DPhil, Professor and Dean. Department of Social Medicine, Hebrew University - Hadassah Hospital, Jerusalem, Israel. Yechiel Friedlander, PhD, Associate Professor in Epidemiology. Pfizer Pty Ltd, Sydney, NSW. Michael Ortiz, PhD, Health Outcomes Manager. Reprints will not be available from the authors. Correspondence: Professor L A Simons, Lipid Research Department, St Vincent's Hospital, Darlinghurst, NSW 2010. l.simonsATnotes.med.unsw.edu.au Make a comment Back to text 2: Age-specific all-causes mortality rate and alcohol intake during 116 months' follow-up of subjects 60 years and over. Data are mortality rates per 100 subjects (95% CI and number of subjects in each group in parentheses) Men Women Zero alcohol Any alcohol Zero alcohol Any alcohol 60-64 y 65-69 y 70-74 y 75-79 y 80+ y All ages 26 (17-35) (88) 37 (24-50) (54) 57 (45-69) (65) 51 (35-67) (37) 83 (68-98) (23) 44 (38-50) (267) 17 (13-21) (326) 26 (21-31) (266) 41 (34-48) (184) 64 (55-73)(120) 83 (74-92) (63) 34 (31-37) (959) 9 (5-13) (193) 22 (16-28) (172) 31 (24-38) (154) 45 (36-54) (121) 72 (63-81) (104) 31 (28-34) (744) 10 (7-13) (288) 14 (9-19) (191) 11 (6-16) (166) 36 (27-45) (107) 54 (42-66) (68) 20 (17-23) (820) Back to text 3: Proportional hazards model of all-causes mortality and alcohol intake in elderly subjects Hazard ratio (95% CI) Men 60-74 years Women 60+ years Significant predictors of all-causes mortality Any alcohol intake Age (/year) Current smoker Ex-smoker Prior stroke Blood pressure medication Diabetes Atrial fibrillation Poor expiratory flow 0.63 (0.47-0.84) 1.07 (1.04-1.11) 2.81 (1.90-4.16) 1.76 (1.26-2.46) 1.58 (1.03-2.42) 1.54 (1.16-2.06) 1.99 (1.38-2.87) -- 1.53 (1.09-2.17) 0.75 (0.60-0.94) 1.08 (1.06-1.10) 1.74 (1.22-2.49) -- -- -- 2.06 (1.46-2.92) 2.77 (1.65-4.64) 2.00 (1.43-2.79) Relationship of quantity or type of alcohol to all-causes mortality Drinks/week Nil 1-7 8-14 15-28 >28 1.00 0.68 (0.49-0.94) 0.58 (0.39-0.85) 0.62 (0.40-0.95) 0.56 (0.33-0.96) 1.00 0.78(0.61-0.99) 0.66 (0.45-0.97) 0.67 (0.29-1.55) Alcohol type Nil Beer Wine/spirits 1.00 0.62 (0.46-0.84) 0.70 (0.44-1.12) 1.00 0.64 (0.47-0.86) 0.85 (0.66-1.11) The reference category for alcohol usage was zero intake. Other variables in the models were body mass index, family history of coronary heart disease (CHD), prevalent CHD, blood pressure, lipid levels, self-rated health, and physical disability. Poor expiratory flow refers to peak expiratory flow tertile I. Back to text Back to text

Leon A Simons · John McCallum · Yechiel Friedlander · Michael Ortiz · Judith Simons

Research A randomised crossover trial of chemotherapy in the home: patient preferences and cost analysis Danny Rischin, Michelle A White, Jane P Matthews, Guy C Toner, Kathryn Watty, Anthony J Sulkowski, Jan L Clarke and Lois Buchanan MJA 2000; 173: 125-127 Abstract - Methods - Results - Discussion - References - Authors' details - - More articles on Oncology Abstract Objectives: To determine patient preferences and cost differences between home-based and hospital-based chemotherapy. Design: Randomised crossover trial. Setting: A tertiary cancer hospital in Melbourne, Victoria. Participants: 20 patients who required chemotherapy suitable for administration at home. Interventions: Patients were assigned at random to receive their first chemotherapy treatment in either the home or the hospital and the second treatment in the alternative setting. Main outcome measures: Patient preference, costs. Results: There was universal agreement by the 20 patients in the randomised trial that home-based chemotherapy was the preferred option (P < 0.0001). No problems were nominated by the patients as being associated with home-based chemotherapy. Home-based treatment was estimated to result in an increased cost of $83 (P = 0.0002) for each chemotherapy treatment compared with hospital-based treatment. Reported advantages for chemotherapy in the home included the elimination of travel, reduction in treatment-associated anxiety, reduction in the burden on carers and family, and the ability to continue other duties. There were no significant complications associated with administration of chemotherapy in the home. Conclusions: Patients prefer home-based chemotherapy to hospital-based treatment. The future of chemotherapy-in-the-home programs in Australia will depend on whether patient preferences are deemed to offset any potential increase in costs. Patients with cancer who require treatment with chemotherapy will experience major changes in lifestyle and overall well-being. Some intravenous chemotherapy regimens require frequent visits to hospital to receive treatment. This may be time-consuming and inconvenient for a patient, and may also disrupt the lives of other family members and carers. The concept of home-based therapy is not new,1-3 but there have been few reports on chemotherapy-in-the-home programs, and these have had a different emphasis from our study (eg, costs [in a US paediatric population];4 costs and safety [in a retrospective review of an Australian adult population]5). We performed a randomised crossover trial, the aim of which was to compare (i) patient preference for hospital-based versus home-based chemotherapy; and (ii) the cost of therapy administered in hospital versus that in the home. Methods Patient eligibility Patients were considered eligible if they met the following criteria: they were to receive chemotherapy that was suitable to be given at home; their first two treatments were planned to be identical; they had not received chemotherapy in the preceding 12 months; they lived in an area that was geographically suitable for treatment at home; and they were aged 18 years or over. Patients gave written informed consent and the study was approved by the ethics committee of the Peter MacCallum Cancer Institute. Study design At enrolment, patients were randomly assigned to receive their first chemotherapy treatment in hospital and the second at home, or their first treatment at home and the second in hospital. They were assigned according to a computer-generated randomisation chart, using an allocation scheme based on a biased coin design.6Chemotherapy treatment refers to the first two administrations of chemotherapy. Following completion of the first two treatments, patients filled out a questionnaire regarding the two different locations of therapy. Questions focused on (i) patients' preference for where to receive their remaining chemotherapy after completing their two study treatments; and (ii) any perceived difficulties or advantages of treatment in hospital or in the home. Chemotherapy nurse specialists who also worked in the chemotherapy day ward at the hospital administered all home chemotherapy treatments. Patients were reviewed by a doctor before each chemotherapy cycle. Cost assessment Cost comparisons for hospital-based versus home-based therapy were made specifically from the perspective of the treating hospital, not the patient or society in general. Costs were estimated using Transition software (Eclipsys Transition Systems, Boston, MA), which distributes direct and indirect costs for an entire financial year between patient episodes on the basis of the services received. It was decided to compare only those components of the cost for which there could be a genuine difference to the hospital attributable to the site of delivery of the chemotherapy. Thus, costs related to patient records, allied health, medical staff and pharmacy were excluded. All overheads associated with the chemotherapy-in-the-home program, including vehicle costs and travelling time, were apportioned by Transition to nursing costs on the basis of time spent with each patient. Similarly, hospital overheads were apportioned on the basis of nursing times. The cost of providing a single meal was included in the hospital costs. Statistical methods A target sample size of 20 eligible patients with identical chemotherapy for their first two treatments was chosen, to provide 84% power to test the null hypothesis that no one setting is preferred versus the alternative hypothesis that at least 85% of patients prefer one setting over the other, using a two-sided test of significance at a significance level of 0.05. To determine if significantly more patients preferred treatment at home rather than in hospital, or vice versa, the proportion of patients preferring to have their third treatment in the same location as their first treatment was compared between the two randomisation arms using Fisher's exact test for 2 x 2 contingency tables. This test is valid even if there are "period" effects -- that is, if patients tolerate their second chemotherapy treatment better than their first, or vice versa.7 Standard methods for a 2 x 2 crossover trial7 were used to compare costs of chemotherapy in hospital with costs in the home, and costs between the first and second chemotherapy given ("period" effects), after ensuring there were no significant carryover effects. (Carryover effects were tested by comparing the sum of the costs in the home and hospital for patients in the "hospital first" arm with patients in the "home first" arm.) Statistical significance and 95% confidence intervals (CIs) were estimated from the means and standard errors assuming a Student's t-distribution. Two-sided P-values have been given throughout. All statistical tests were carried out using Stat Xact 4 (CYTEL Software Corporation, Cambridge, MA, 1998) and Microsoft Excel (Microsoft Corporation, Redmond, WA, 1996) software. Results Patient selection and profile The trial accrued the target 20 patients, out of a total of 64 registered on the chemotherapy-in-the-home program, between February 1996 and March 1997 (see Box 1). Patient demographics are shown in Box 2. Patient preferences When asked where they would have preferred to receive their first two treatments if they had had their time again, 70% of patients expressed a preference for having both treatments at home, while none said they would have preferred to have both treatments in hospital (Box 3). Patients were then asked to nominate their preferred site for the remaining treatments (the primary endpoint of the trial). All 20 patients (100%; 95% CI, 83%-100%) preferred to have their remaining therapy given at home (P < 0.0001). None of the patients in the trial reported concerns with chemotherapy being given in their home; however, four (20%) reported concerns with treatment in hospital, relating to transport difficulties and waiting times. Eighteen (90%) of the patients felt there were advantages with treatment in the home. The reasons given included convenience; avoidance of travel and parking problems (particularly not having to travel while feeling unwell); reduction in treatment-associated anxiety; not burdening their carers and family; and being able to continue other duties, such as caring for their dependants. Only one patient felt there were specific advantages to chemotherapy in the hospital. This patient felt it was good to see other people who were worse off. No major complications of chemotherapy administration (eg, hypersensitivity reactions or extravasation) were reported. Costing Overall, chemotherapy in the home was associated with an estimated average increased cost of $83 (95% CI, $46-$120; P = 0.0002) relative to the cost of chemotherapy in the hospital. The average cost of the first treatment was estimated to be $57 more than the cost of the second (95% CI, $20-$94; P = 0.0044). There was no carryover effect (P = 0.16). Discussion This study has demonstrated that patients have an overwhelming preference for home-based therapy. Clearly, home-based therapy is not possible, or indeed appropriate, for all patients. Patients living outside designated geographical areas or having special needs that can be met in the hospital setting (eg, need for an interpreter) would be more easily treated at the hospital.8 Complex or prolonged chemotherapy regimens or those associated with a risk of an immediate serious complication are more appropriately administered in the hospital day ward setting. Nevertheless, many commonly administered chemotherapy regimens are suitable for administration in the home, and this study clearly demonstrates that, given the choice, patients prefer to have such treatments at home. While in our study the cost of home-based treatment was on average $83 higher than the cost of hospital-based therapy, this estimate did not include costs to the patient (such as travelling costs, lost time for the patient or carers, and childcare costs). These could all have made the hospital episode more costly relative to the home episode. Furthermore, given that the cost per visit for any chemotherapy-in-the-home program is dependent on the throughput of patients and the geographical spread of the patients, an increase in the frequency of home visits or a more limited geographical spread of patients might further reduce the difference between home and hospital costs. Unlike some other hospital-in-the-home programs, chemotherapy in the home does not necessarily result in cost savings to the administering hospital, as treatment in the hospital does not require overnight admission. The future of chemotherapy-in-the-home programs in Australia will depend on how governments, hospital administrators, oncologists and nurses balance the overwhelming patient preference for treatment at home with any potential increase in costs. References Grayson ML, Silvers J, Turnidge J. Home intravenous antibiotic therapy. A safe and cost effective alternative to inpatient care. Med J Aust 1995; 162: 249-253. Koopman MMW, Prandoni P, Piovella F, et al. Treatment of venous thrombosis with intravenous unfractionated heparin administered in the hospital as compared with subcutaneous low molecular weight heparin administered at home. N Engl J Med 1996; 334: 682-687. Bielory L, Long GC. Home health care costs: intravenous immunoglobulin home infusion therapy. Ann Allergy Asthma Immunol 1995; 74(3): 265-268. Close P, Burkey E, Kazak A, et al. A prospective, controlled evaluation of home chemotherapy for children with cancer. Pediatrics 1995; 95: 896-900. Lowenthal RM, Piaszczyk A, Arthur GE, et al. Home chemotherapy for cancer patients: cost analysis and safety. Med J Aust 1996; 165: 184-187. Wei LJ, Lachin JM. Properties of the urn randomization in clinical trials. Control Clin Trials 1988; 9: 345-364. Jones B, Kenward MG. Design and analysis of cross-over trials. London: Chapman and Hall Ltd, 1989. Zalcberg JR, Siderov J, Petty M. Outpatient chemotherapy: there's no place like home - sometimes. Med J Aust 1996; 165: 182. (Received 11 Oct 1999, accepted 24 May, 2000) Authors' details Peter MacCallum Cancer Institute, Melbourne, VIC. Danny Rischin, MB BS(Hons), FRACP, Consultant Medical Oncologist, Division of Haematology and Medical Oncology; Michelle A White, MB BS(Hons), FRACP, Clinical Fellow, Division of Haematology and Medical Oncology; Jane P Matthews, BSc(Hons), PhD, AStat, Director, Statistical Centre; Guy C Toner, MD, BS, FRACP, Head of Medical Oncology, Division of Haematology and Medical Oncology; Kathryn Watty, RN, RM, Clinical Nurse Consultant, Division of Nursing; Anthony J Sulkowski, RN, BEd, Clinical Nurse Consultant, Division of Nursing; Jan L Clarke, RN, Clinical Nurse Consultant, Division of Nursing; Lois Buchanan, RN, RM, Clinical Nurse Consultant, Division of Nursing. Reprints will not be available from the authors. Correspondence: Dr D Rischin, Division of Hematology and Medical Oncology, Peter MacCallum Cancer Institute, Locked Bag 1, A'Beckett Street, Melbourne, VIC 8006. drischinATpetermac.unimelb.edu.au Make a comment Click in box for larger version Back to text 2: Patient demographics, by randomisation arm (first chemotherapy treatment given at hospital vs first treatment at home). Values are number of patients unless otherwise stated Hospital first Home first Total (%) Sex Male Female 1 8 4 7 25% 75% Age Median (years) Range (years) 40-49 50-59 60-69 70 61 47-71 0 2 1 5 1 59 26-69 1 4 1 5 0 -- -- 5% 30% 10% 50% 5% Diagnosis Breast cancer Colon cancer Non-Hodgkin's lymphoma Pancreatic cancer 5 3 0 1 5 5 1 0 50% 40% 5% 5% Chemotherapy CMF(P)* 5-FU† ± folinic acid or levamisole CHOP‡ 5 4 0 5 5 1 50% 45% 5% Support at home Spouse/parent Parent Son/daughter Other Not specified 4 1 3 0 1 7 0 1 3 0 55% 5% 20% 15% 5% * Cyclophosphamide, methotrexate, 5-fluorouracil ±prednisolone. † 5-Fluorouracil. ‡Cyclophosphamide, doxorubicin, vincristine and prednisolone. Back to text 3: Preferred location for chemotherapy treatments, by randomisation arm (first chemotherapy treatment given at hospital vs first treatment at home). Values are number of patients Hospital first Home first Total (%) For first 2 treatments Both at home First at home, second at hospital No preference First at hospital, second at home Both at hospital 7 0 1 1 0 7 2 1 1 0 14 (70%) 2 (10%) 2 (10%) 2 (10%) 0 For subsequent treatments Home Hospital 9 0 11 0 20 (100%) 0 Back to text

Danny Rischin · Michelle A White · Jane P Matthews · Guy C Toner · Kathryn Watty · Anthony J Sulkowski · Jan L Clarke · Lois Buchanan

Research Early discharge and postnatal depression: a prospective cohort study Jane F Thompson, Christine L Roberts, Marian J Currie and David A Ellwood MJA 2000; 172: 532-536 For editorial comment see Lumley Abstract - Methods - Results - Discussion - Acknowledgements - References - Authors' Details - - More articles on Psychiatry Abstract Objectives: To determine whether women discharged from hospital ≤ 72 hours after childbirth (early discharge) were at greater risk of developing symptoms of postnatal depression during the following six months than those discharged later (late discharge), their reasons for early discharge and their level of postnatal support. Design and setting: Population-based, prospective cohort study with questionnaires at Day 4, and at 8, 16 and 24 weeks postpartum, conducted at all birth sites in the Australian Capital Territory (ACT). Participants: Women resident in the ACT giving birth to a live baby from March to October 1997. Main outcome measure: A score > 12 on the Edinburgh Postnatal Depression Scale (EPDS). Results: 1295 (70%) women consented to participate; 1193 (92%) were retained in the study to 24 weeks and, of these, 1182 returned all four questionnaires. Of the 1266 women for whom length-of-stay data were available, 467 (37%) were discharged early and 799 (63%) were discharged late. There were no significant differences between the proportion of women discharged early who ever scored > 12 on the EPDS during the six postpartum months and those discharged late (17% v. 20%), even after controlling for other risk factors (adjusted OR, 0.67; 95% CI, 0.44-1.01). Of women discharged early, 93% had at least one postnatal visit at home from a midwife and 81% were "very satisfied" with the care provided. Most women (96%) reported they had someone to help in practical ways. Conclusions: Women discharged early after childbirth do not have an increased risk of developing symptoms of postnatal depression during the following six months. In Australia, and internationally, the length of time spent in hospital following childbirth has been steadily decreasing since the early 1980s. This has prompted concern about the consequences of early discharge for both mothers and babies.1,2 Postnatal depression (PND) is a common disorder with long-term consequences for both mother and infant.3,4 It has been associated with psychosocial and obstetric factors3,5-7 and possibly dissatisfaction with length of stay.8,9 The association with early discharge is not clear. A small randomised controlled trial of early discharge from Sweden10 reported no difference in depression, while in a similar Canadian study women discharged early were less likely to be depressed.11 In observational studies, women discharged early have been reported to be either equally8,9,12,13 or less likely14,15 to be depressed. A recent Australian study reported an increased risk of PND in women discharged early.16 However, this study made no detailed assessment of support available to women, a factor that may be critical to emotional wellbeing of new mothers.17Here, we aimed to investigate whether early discharge following childbirth was associated with an increased risk of PND, why women elect early discharge, and to examine the social support available to women after discharge from hospital. Methods Participants This population-based, prospective cohort study included women resident in the ACT, planning to reside there for at least six months, aged ≥ 16 years, who gave birth to a live baby between March and October 1997 in any of the ACT's two public hospitals (one included a birth centre), two private hospitals, or at home. Women were excluded if their baby was admitted to the neonatal intensive care unit or adopted, if critically ill themselves, unable to give informed consent or complete the questionnaires for other reasons, or participating in another study. Participants were compared with all women who gave birth in the ACT during 1997 using data from the ACT Maternal and Perinatal Data Collection.18 Procedure The study was approved by the ACT Department of Health & Community Care Research Ethics Committee, and the ethics committees of participating hospitals. Postnatal ward or domiciliary midwives gave information sheets to women in the first few days after giving birth. Participants gave written informed consent when completing the first questionnaire as close to Day 4 as possible. They were then mailed questionnaires at eight, 16, and 24 weeks postpartum. Questionnaires The first questionnaire covered sociodemographic characteristics of mother and partner, a nine-item personality scale identifying "vulnerable" and "resilient" personality dimensions;16 a maternity "blues" questionnaire;19 a subset of four items from the Medical Outcomes Study Social Support Scale;20 questions about availability of and satisfaction with practical support, emotional support from partner, and a single summary question assessing global satisfaction with partner scored on a five-point Likert scale. Three questions asked about the nature of women's past relationship with their parents in relation to warmth/care, overprotection/controlling and independent decision making, with responses on a four-point scale. Questions were also included about the mother's history of depression (at any time as well as during or after a pregnancy) and whether the infant was breastfed. In the second questionnaire, women were asked to indicate whether any of a list of 30 possible reasons for their actual length of stay applied to them and whether they thought their length of stay was too long, about right or too short. The third questionnaire included questions about the number of and satisfaction with domiciliary visits. Satisfaction was measured by quality of care, accessibility and convenience,21 and a single summary question assessing overall satisfaction with care. Postnatal depression Postnatal depression was assessed at eight, 16 and 24 weeks using the 10-item Edinburgh Postnatal Depression Scale (EPDS), a self-report measure of depression developed for use in the postpartum period.22-24 Women scoring above 12 are likely to be suffering from a depressive illness. Length of stay For comparability with previous Australian research,16 and in keeping with ACT definitions, early discharge was defined as discharge up to 72 hours after giving birth, and late discharge as more than 72 hours after giving birth. Power of study A sample size of 944 is sufficient to detect with 95% confidence and 80% power an increase in prevalence of PND in the early discharge group at eight, 16 or 24 weeks from 7% to 14%,16 assuming 30% are discharged early and an overall attrition rate of 25%. To allow for variations in these assumptions, we set a target sample size of 1200 women. Statistical analyses The prevalences of EPDS scores greater than 12 were compared between women discharged early and late by means of contingency tables and unconditional logistic regression. We used logistic regression to assess the effect of previously identified risk factors on the association between early discharge and high EPDS scores. Six separate models were fitted for women ever scoring > 12 during the six postpartum months; for those who scored > 12 at eight weeks, 16 weeks or 24 weeks; and for women who scored > 12 on either two occasions or all three occasions. Results are expressed as crude and adjusted odds ratios (OR) with 95% confidence intervals. Results Study population Of 1961 ACT residents asked to participate in the study, 105 were ineligible and 1295 (70%) of the remainder agreed to participate. After 24 weeks, 1193 (92%) remained in the study. Of the 1295 who agreed to participate, 869 (67%) gave birth in a public hospital, 411 (32%) in a private hospital, and 15 (1%) at home, of whom six were transferred to hospital. Compared with all women who gave birth in 1997, participants were slightly older, more likely to be married or in a defacto relationship, to have given birth in a private hospital, and to have been discharged late (Box 1). The 102 (8%) who were lost to follow-up differed from those who remained in that they were significantly (P ≤ 0.001) more likely to be aged < 25 years (32% v. 12%), unmarried (14% v. 4%), born in a non-English-speaking country (19% v. 9%), and public patients (74% v. 56%). They were not significantly more likely to be in the early discharge group (46% v. 36%), but were significantly (P ≤ 0.001) less likely to be in paid employment in the previous 12 months (56% v. 74%), to have a paid position to resume after maternity leave (43% v. 63%) and to have been educated beyond Year 11 (60% v. 82%). They did not differ with respect to the following factors known to be associated with PND: vulnerable personality, level of social support, past history of depression, dissatisfaction with relationship with partner, or dissatisfaction with past relationship with mother. Length of stay After excluding the nine women who gave birth at home and were not transferred to hospital, and the 20 with missing data, there were length-of-stay data for 1266 (98%) women, 467 (37%) with early and 799 (63%) with late discharge. The early discharge group was significantly (P ≤ 0.001) more likely to be aged < 25 years (21% v. 10%), public patients (80% v. 44%) or to have delivered in a public hospital (94% v. 52%), multiparous (61% v. 54%), and to have given birth at > 39 weeks' gestation (85% v. 75%). Women discharged early were also significantly (P ≤ 0.001) more likely to have had a spontaneous onset of labour (74% v. 55%), an unassisted vaginal birth (90% v. 56%) and to formula-feed their infant from birth (10% v. 5%). They were significantly (P ≤ 0.001) less likely to have been in paid employment in the past 12 months (68% v. 76%), to have a paid position to resume (55% v. 66%) and to have been educated beyond Year 11 (74% v. 84%). They were significantly (P ≤ 0.001) less likely to rate their length of stay as "about right" than women discharged late (72% v. 82%), and more likely to rate their length of stay as "too short" (23% v. 8%). There were no statistically significant differences in vulnerable personality (17% v. 16%), level of social support (median score, 7 for both groups), past history of depression (29% v. 29%), maternity blues score (median score, 4 v. 5), dissatisfaction with partner (8% v. 6%), and past relationship with mother (not warm/caring, 6% v. 5%; overprotective/controlling, 52% v. 52%; did not encourage independent decision-making, 20% v. 21%). Postnatal depression Of the 1252 (97%) women with complete data at eight weeks postpartum, 129 (10%) scored > 12 on the EPDS. At 16 weeks, 91 of 1219 (8%) and at 24 weeks 90 of 1187 (8%) scored > 12. The cumulative incidence of an EPDS score > 12 over the six months of follow-up was 224/1295 (17%). Among women with complete data, 37/1172 (3%) had EPDS scores > 12 on two occasions and 23/1172 (2%) on all three occasions. Length of stay and postnatal depression Women discharged early were not more likely to ever score > 12 on the EPDS during the six months of follow-up than women discharged late: 72/429 (17%) compared with 150/751 (20%). The association between length of postnatal stay and PND symptoms remained statistically non-significant after adjusting for other risk factors (Box 2). This finding was robust for other outcome measures (Box 3). For all outcomes there was a consistent trend towards a reduced risk of PND symptoms for women discharged early. Reasons for choosing early discharge Reasons were given by 447 women (96%). The most common reason was a preference to be at home with their partner or family (74%). Other reasons were feeling confident with their baby and preferring to be at home (73%); so the father could be more involved in baby care (47%); being unhappy in hospital and unable to sleep or rest (41%); greater privacy (41%); to rest or recover after the birth (40%); not liking hospitals (34%); to establish breastfeeding (33%); and to have time to focus on the baby (32%). Of the women discharged early, 15% said they did not feel they had a choice about length of stay, and 14% felt under pressure from midwives to leave early. Women discharged early who felt they had no choice about length of stay were not significantly more likely to ever score > 12 on the EPDS (17% v. 17%), and neither were those who felt pressured to leave early (21% v. 16%; P = 0.4). Early discharge and postnatal support All women discharged within three days were eligible for home visits from midwives. Data were available on the number and nature of these visits for 432 women discharged early (93%). Of these, 404 (94%) had at least one visit, and 176 (41%) were visited up to Day 7. The maximum number of visits in the first seven days was 12, and 107 women (27%) were visited at least once after seven days. Eighty-eight per cent of the women thought the number of visits was "just right", while 8% thought it "not enough" and 4% "too many". Overall, 81% of these women said they were "very satisfied" with the care provided at home, 15% "satisfied in some ways but not in others" and 4% "very dissatisfied". Some women discharged after 72 hours were also eligible for and received home visits. The participating public hospitals' policy was for home visiting to be available up to Day 3, but sometimes this extended beyond 72 hours depending on the time of birth. Home visits were also available in cases of special need, and community-based maternal and child health nurses and midwives also offer some home visiting. Of the women discharged late, 364 (47%) received at least one visit at home from a midwife. In addition to support provided by health professionals, 96% of women discharged early reported that they had someone to help in practical ways in the first eight weeks postpartum. For 92% this was a partner, and for 47% their mother. Other family members (22%), friends (20%) and mothers-in-law (19%) were the next most common sources of help. The help received was satisfactory for 90% of the respondents; however, 23% said that they would like to know more people who could be asked for help. There were no statistically significant differences between women discharged early or late in availability of and satisfaction with practical help. Discussion The postnatal stay has become shorter without being properly evaluated in randomised controlled trials (RCTs). As shorter stays have become standard practice, the window of opportunity for conducting an appropriate RCT may have been lost.25 The best alternatives are prospective studies with heterogeneous samples of sufficient size to detect clinically significant effects of short postnatal stays. We found that women resident in the ACT electing short postnatal stays were not at increased risk for developing symptoms of PND. Our finding is consistent with those from two Victorian surveys,8,13 but differs from that of a study in Sydney which found a significantly increased risk of PND during the first six months in mothers discharged within three days.16 There are several possible explanations for the discrepant results between this study and our own, including differences in the outcome measures, population characteristics, postnatal support and reasons for early discharge. The same assessment tool to identify possible cases was used in the Victorian,8,13 Sydney16 and ACT studies, but in the Sydney study a psychiatric interview was added to confirm the diagnosis of PND. The use of a psychiatric examination is unlikely to explain differences in the results, as high scores on the EPDS coincide closely with diagnoses of PND. A validation study in Australian women found the EPDS to be highly sensitive (100%) and specific (96%), with a positive predictive value of 70%.22 Differences in the characteristics of women discharged early may also contribute to the different outcomes observed. In both the Sydney and the ACT studies, women discharged early were more likely to be multiparous, to have a lower level of education and to formula feed their infants in the first week than women discharged late. However, the women discharged early in the Sydney study16 were more likely than those in our study to report a poor relationship with parents and to have a history of depression; these associations were controlled for in the statistical analyses in the Sydney study and so cannot fully explain their different findings. Low levels of social support have been identified as a risk factor for PND,5 and another possible explanation for the differences in findings is that the two populations differed in the extent, nature of and satisfaction with postnatal support provided to women through early discharge programs and non-professional contacts. The authors of the Sydney study reported that only half of the women who elected early discharge participated in an early discharge program with domiciliary midwifery care.26 In our study, 94% of the women discharged early had had at least one visit from a midwife at home and the level of satisfaction with this care was high. Also, only 4% of women discharged early said they had no one to provide practical support at home. An overwhelming majority reported that their partners assisted them and most were satisfied with the level of help. Early discharge was introduced partly to offer more choices in care in a climate of increasing consumer participation in decisions. However, economic imperatives to increase patient throughput may lead to increasing pressure on women to leave hospital earlier than they would otherwise choose. The Sydney study26 did not report reasons for early discharge. Reasons given by women for leaving ACT hospitals early were generally positive, although 15% felt they did not have a choice and 14% felt pressured to leave (not mutually exclusive reasons). Several things should be considered when interpreting the results of our study. Firstly, as the Sydney study found double the rate of PND after early discharge,16 we formed our null hypothesis (that early discharge makes no difference) in the expectation that it would be disproved. Instead, we found no evidence of a significant difference in the rate of PND; in fact, early discharge tended to a protective effect (but this was not statistically significant). Our study does not prove that there is not an increased rate of PND for early discharge. However, in this population, it is unlikely that the true risk for early discharge was double that for late discharge. Secondly, although only 70% of women approached participated in the study, the characteristics of women who did participate were similar in important respects to those of the source population. However, our findings may not be generalisable to populations with differing PND risk factors or early discharge programs. Lastly, the women in our study selected their length of postnatal stay. Although we have controlled for known determinants, confounding by unknown determinants for PND cannot be excluded. We found that women who selected early discharge from hospital after childbirth and received midwifery support at home, and who were well supported by other family members, were not more likely to experience depressive symptoms in the first six months after childbirth. Very few women in this study were discharged early without home support, so it was not possible to determine whether early discharge without support was associated with an increased risk of PND symptoms. It may be important to ensure that all women discharged early after childbirth, in particular those lacking other sources of support, receive additional help from the healthcare system. What constitutes adequate postnatal support could be examined by RCTs comparing different patterns of home visiting. Acknowledgements This work was supported by a project grant from The Canberra Hospital Private Practice Fund. Additional funding was provided by The Canberra Hospital Auxiliary, the Nurses' Board of the ACT, and the ACT Department of Health & Community Care. The assistance of the midwives in recruiting women for this study is gratefully acknowledged. Robyn Attewell provided statistical advice. We are especially grateful to the women of the ACT who so generously gave their time to complete this study. References Buist AE. Counting the costs of early discharge after childbirth [editorial]. Med J Aust 1997; 167: 236-237. Braveman P, Egerter S, Pearl M, et al. Problems associated with early discharge of newborn infants. Early discharge of newborns and mothers: a critical review of the literature [review]. Pediatrics 1995; 96: 716-726. Boyce PM, Stubbs JM. The importance of postnatal depression. Med J Aust 1994; 161: 471-472. Murray L, Cooper P. Effects of postnatal depression on infant development. Arch Dis Childhood 1997; 77: 99-101. O'Hara MW, Swain AM. Rates and risk 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. Warner R, Appleby L, Whitton A, Faragher B. Demographic and obstetric risk factors for postnatal psychiatric morbidity. Br J Psychiatry 1996; 168: 607-611. 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. Dowswell T, Piercy J, Hirst J, et al. Short postnatal hospital stay: implications for women and service providers. J Public Health Med 1997; 19: 132-136. Waldenström U. Early and late discharge after hospital birth: fatigue and emotional reactions in the postpartum period. J Psychosomatic Obstet Gynaecol 1988; 8: 127-135. Carty EM, Bradley CF. A randomized, controlled evaluation of early postpartum hospital discharge. Birth 1990; 17: 199-204. Beck CT, Reynolds MA, Rutowski P. Maternity blues and postpartum depression. J Obstet Gynecol Neonatal Nurs 1992; 21: 287-293. Brown S, Lumley J, Small R. Early obstetric discharge: does it make a difference to health outcomes? Paediatr Perinat Epidemiol 1998; 12: 49-71. Burnell J, McCarthy M, Chamberlain GVP, et al. Patient preference and postnatal hospital stay. J Obstet Gynaecol 1982; 3: 43-47. James ML, Hudson CN, Gebski VJ, et al. An evaluation of planned early postnatal transfer home with nursing support. Med J Aust 1987; 147: 434-438. Hickey AR, Boyce PM, Ellwood D, Morris-Yates AD. Early discharge and risk for postnatal depression. Med J Aust 1997; 167: 244-247. Barclay KM, Chamberlain ME, Homer CS, Barclay LM. Early discharge and risk for postnatal depression [letter]. Med J Aust 1998; 168: 419-420. Bourne M. Maternal and perinatal status, ACT, 1997 tables. Canberra: Clinical Epidemiology and Health Outcomes Centre, ACT Department of Health & Community Care, 1999. Kennerley H, Gath D. Maternity blues. 1. Detection and measurement by questionnaire. Br J Psychiatry 1989; 155: 356-362. Sherbourne CD, Stewart AL. The MOS social support survey. Soc Sci Med 1991; 32: 705-714. Kenney P, Cameron S, King M, et al. Evaluation of obstetric early discharge: client satisfaction. Centre for Health Economics Research & Evaluation. Discussion Paper Series No 10. 1992. Boyce P, Stubbs J, Todd A. The Edinburgh postnatal depression scale: validation for an Australian sample. Aust N Z J Psychiatry 1993; 27: 472-476. 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. Murray L, Carothers AD. The validation of the Edinburgh Postnatal Depression Scale on a community sample. Br J Psychiatry 1990; 157: 288-290. Thompson JF, Roberts CL, Ellwood DA. Early discharge after childbirth: Too late for a randomised trial? Birth 1999; 26: 192-195. Hickey AR, Boyce PM, Morris-Yates AD, Ellwood DA. Early discharge and risk for postnatal depression [letter]. Med J Aust 1998; 168: 420. (Received 7 Oct 1999, accepted 3 Apr 2000) Authors' Details The Canberra Hospital, Garran, ACT. Jane F Thompson, MSc, PhD, Senior Research Officer, Women's & Children's Health. Marian J Currie, BapplSc, GDPH, Midwife, Maternity and Gynaecology Outpatients and Fetal Medicine Unit. David A Ellwood, FRANZCOG, Dphil, Professor of Obstetrics and Gynaecology, The Canberra Clinical School. New South Wales Centre for Perinatal Health Services Research, Departments of Obstetrics and Gynaecology and Public Health and Community Medicine, School of Population Health Services Research, University of Sydney, NSW. Christine L Roberts, MB BS, MHP, Senior Lecturer. Reprints: Dr J F Thompson, Women's and Childrens Health, The Canberra Hospital, PO Box 11, Woden, ACT 2606. jane.thompsonATact.gov.au ©MJA 2000

Jane F Thompson · Christine L Roberts · Marian J Currie · David A Ellwood