Letters

In reply: Black has argued that the role of evidence in policy development is to create concern and set agendas.1 This was the aim of our article. Although there has been a national, organised approach to preventing cervical cancer since 1991, our study found that screening rates for Indigenous women still lag well behind those of non-Indigenous women. Workers in Indigenous health might have suspected as much, but valid data have not previously been available for a wide geographical area. As Murray and Lopez point out, the lack of good data on a health issue is often taken to mean that the problem is not important.2 We agree that the development of effective programs should be done in consultation and partnership with Indigenous communities. However, there is an additional need to influence decision-makers and budget-holders at national, State and regional levels. An evidence base that identified effective interventions would facilitate this process. Such evidence need not come from randomised controlled trials. On the other hand, case reports of successful interventions in a single community that cannot be sustained when key personnel leave are not very persuasive. It is also useful to demonstrate that the situation is not hopeless. The encouraging finding from our study was that the participation rate in cervical cancer screening was more than 50% in three of the 13 communities studied.

Michael D Coory · Patricia S Fagan · Jennifer M Muller · Nathan AM Dunn

To the Editor: The beyondblue guidelines for treating depression in primary care by Ellis and Smith1 are intended to assist both healthcare professionals and consumers. While they provide several helpful indications, they also include some misleading suggestions. The authors state that drug treatment of depression should continue for at least one year for a first episode of depression, and at least two years for repeated episodes or when there are other risk factors for relapse. However, no background literature is cited in support of this statement, and indeed would be difficult to find. Maintenance pharmacotherapy has been advocated as an effective tool for reducing relapses and recurrences in major depression.2 A number of studies have shown the superiority of antidepressant drugs (mostly tricyclics) compared with placebo in protecting the patient from relapse. Duration of drug treatment, however, did not seem to affect long-term prognosis once treatment with the drug was discontinued. In clinical terms this means that, whether you treat a depressed patient for three months or three years, it does not matter when you stop therapy with the drug. In fact, after recovery from an index episode of major depression, risk of postdiscontinuation relapse was nearly significantly greater after longer treatment (ρ = 0.37; P = 0.052).3 Further, Ellis and Smith1 fail to mention a vexing clinical problem in maintenance antidepressant treatment: the return of depressive symptoms.3 Dose increase is likely to entail only a temporary solution to the problem, which may occur in up to 57% of patients. However, there is a promising alternative. Treatment of depression by pharmacological means is likely to leave substantial residual symptoms.4 Residual symptoms hinder lasting recovery and are one of the strongest risk factors for relapse. In randomised controlled trials, cognitive behavioural treatment of residual symptoms was found to significantly improve long-term outcome of recurrent depression and to allow discontinuation of drug therapy.4 Preventing recurrence in major depression cannot simply be based on prolonging ongoing pharmacological treatment. The belief that a longer course of treatment will result in a more favourable outcome after discontinuation of antidepressant drug therapy is not supported by research evidence.5 Active collaboration with the patient (in choosing a treatment option, in lifestyle modification, in seeking treatment again when needed) is a crucial, and yet neglected, variable. It can lead to a more rational use of antidepressant drugs and to therapeutic efforts of more enduring quality than those prevailing today.

Giovanni A Fava · Chiara Ruini · Eliana Tossani

In reply: Fava observes that duration of treatment with an antidepressant does not affect the subsequent rate of relapse. Indeed, it would be unexpected if it did; medication only works while it is being taken. He then states that the duration of antidepressant treatment is immaterial. However, Figure 1 of the reference he quotes1 indicates that, for at least 54 months after the index episode, continuing medication provides greater protection against relapse than early discontinuation. This argues strongly for the beneficial effects of continuing antidepressant therapy for a significant period after recovery. The duration of this period depends on the number of previous episodes of depression and is a compromise between the benefits of effective prophylaxis and the burden of treatment. He also refers to a study of "tolerance" to 20 mg fluoxetine.2 Of patients who responded to treatment, 31.4% relapsed while on maintenance treatment. Of these, 57% responded to an increase to 40 mg fluoxetine and remained well for six months. He appears to have interpreted these data differently. I agree that treating depression involves more than prescribing. The beyondblue guidelines promote active collaboration with the depressed person in addressing key issues in their lives and a focus on relapse prevention.3 Cognitive behavioural therapy is one strategy for achieving this.

Pete M Ellis

To the Editor: Van Der Weyden has commented on the World Health Organization's Utopian definition of "health", first promulgated in 1948.1 In 1973, I addressed this matter in a speech to the All Nations Club in Sydney, and again in 1980 when presenting a paper to The Hope Foundation (a large charitable health organisation in the United States). For what it is worth, my suggested definition of "health" was: Health is a high level of physical, mental and social comfort appropriate to the age of the person concerned and attainable within the economic constraints of the particular environment. It seems we have some agreement, and I would like to see the WHO again consider definition!

Keith S Jones

In reply: I thank Sir Keith for his pragmatic proposal. The concept of health has individual and societal connotations. At its most basic level, it is the avoidance of pain and suffering. More broadly, it is a basic human resource for the pursuit of life's goals.1 But, as argued by Lewis and Leeder, "health is a good to be pursued, but not an absolute one" for the "functioning of social institutions does not require perfectly healthy citizenry, nor does the individual have to be perfectly healthy to take part in social life."2 As such the perfect definition of health espoused by the WHO is Utopian and removed from reality. It is, as poignantly captured by René Dubos, the "mirage of health", as "complete freedom . . . from disease is but a dream remembered from imaginings of a Garden of Eden."3

Martin B Van Der Weyden

To the Editor: The editorial by Cameron and Campbell on access block is an excellent summary of the causes and potential solutions to access block.1 The effects of overcrowding in the emergency department (ED) have been previously reported, including risks to patient safety, prolonged pain and suffering, and decreased clinical productivity and effectiveness.2 Access to emergency care is impaired. The fundamental problem is that while demand has increased, the capacity of the system has decreased.3 This is best exemplified by the significant reduction in hospital bed numbers. Healthcare in Western societies has been through a period of severe economic rationalisation, resulting in closure of thousands of beds in the acute hospital system. For example, in the United States, the number of medical and surgical beds declined by 18% in the period 1994–1999. From 1990 to 1999, attendances at EDs increased 15%. There have also been many aged care beds closed or changed to community-based facilities.4 It has been suggested that the problem is "a badly flawed approach to financing health care that values profits over patients."5 Spare bed capacity is essential for the effective management of emergency admissions. At least one study has found that if hospital bed occupancy rates exceed 85%, then bed crises occur.6 In fact, the Guinness Book of World Records now has a category for longest wait on a hospital trolley!7 The official record currently stands at 77 hours 30 minutes, although anecdotes report longer times. It is paradoxical that other departments within a hospital cannot exceed 100% occupancy, and yet the ED, which may contain some of the most seriously ill or injured, is allowed to exceed the safe level of 100% occupancy. The ED has always been available to help if all else fails in the healthcare system. That basic tenet is now being challenged, and the general public may no longer be able to rely on EDs for quality and timely emergency care, placing the safety of people at risk.8 In addition, Derlet has stated that should there be a major infectious disease epidemic or national catastrophe, EDs and hospitals could not accommodate the demand, undoubtedly leading to increased suffering and excess mortality.3 For all patients, increasing the capacity of the hospitals across the system (viz beds) would really make a difference.

Daniel M Fatovich

To the Editor: The series of articles concerning access block1 indicates that access block is a major health issue in this country. It is remarkable that, despite all these efforts to avoid admissions and reduce inpatient length of stay, only a few occasions of brief success at reducing ambulance diversions were described, and only one case of reducing access block (Royal Melbourne Hospital). We contend that it is now time to increase available beds. In several cases, the association between worsening access block and closure of beds was noted (Australian Capital Territory, Queen Elizabeth Hospital, Royal Perth Hospital). Available data show a steady reduction in hospital beds per thousand population in Australia, from 3.3 in 1995–96 to 2.8 in 1999–2000,2 a decrease of more than 11%. Keeping inpatient occupancy below a threshold percentage is important to controlling access block.3,4 Reducing occupancy by increasing available beds is the logical recommendation.

Peter A Roberts · Paul A Cunningham

To the Editor: I was interested in the letter by Marks et al,1 indicating that the efforts of over-worked medical staff in emergency departments to introduce patients to local general practitioners had been largely unsuccessful. A few years ago I noted the success with which this problem was handled by the emergency department management at Huddinge University Hospital in Stockholm. All patients were charged 60 krone at triage. Those who sat in the waiting room were confronted by two large electronic signs. The first listed the waiting time for the 10 most common GP-type ailments. The second listed 10 local GPs, where the consultation fee was then 50 krone, with the offer to refund their initial payment if they chose to take their business elsewhere. I was told that this was the very successful first of eight "barriers" between the emergency department door and the intensive care unit. Since the middle of last century, Sweden has been held up as a model provider of an egalitarian and "free" healthcare service. Perhaps our country could benefit from the revisions and improvements that the Swedes have made over recent decades.

Peter J Burke

To the Editor: An 8-month-old girl, born in Perth, Western Australia, who received only breast milk feeds for her first six months of life, was referred urgently to Princess Margaret Hospital for Children (the tertiary paediatric hospital in Western Australia) for investigations of seizures. Her parents described the seizures as episodic, involving all limbs, lasting less than five minutes and occurring over the 10 days before presentation. At presentation, the infant had carpopedal spasm. She was afebrile, and a septic screen gave negative results. Venous blood gas analysis showed a low ionised serum calcium level of 0.71 mmol/L (reference range [RR], 1.13–1.32 mmol/L). An x-ray film of her wrists showed signs of rickets, with cupping and fraying of the distal metaphyses of both radius and ulna. Other blood tests revealed elevated parathyroid hormone levels of 8.6 pmol/L (RR, 0.80–8.00 pmol/L), an elevated alkaline phosphatase level of 523 U/L (RR, 100–350 U/L), and an extremely low level of serum 25-hydroxyvitamin D3 (25OHD3) of 5 nmol/L (RR, 30–150 nmol/L). The infant responded to treatment with an infusion of 0.5 mg/kg of calcium gluconate and a 4-month oral course of 0.2 μg calcitriol daily. After treatment commenced, she did not have any further seizures. Radiogaphy performed 4 months later showed increasing calcium deposition at the distal metaphyses of the radius and ulna. Seizures are described as a presenting feature of hypocalcaemia in vitamin D deficiency rickets.1,2 Ultraviolet radiation and/or dietary vitamin D are required to prevent rickets in children. Perth, Western Australia, located at latitude 32° South has an average daily sunshine duration of at least five hours per day. However, even with this amount of sunshine, vitamin D deficiency can still occur in people who, for various reasons, receive little or no sun exposure, especially if their skin is darkly pigmented, and who have an inadequate dietary intake of vitamin D. The girl's mother, who wore a veil and clothing which protected her body from exposure for religious reasons, had a serum 25OHD3 level of 7 nmol/L. Breast milk is a poor source of dietary vitamin D, especially when the lactating woman is vitamin D deficient.3 In Australia, vitamin D deficiency rickets in infants of parents who have migrated from Mediterranean, African, Middle Eastern and southern Asian regions has been reported since the 1960s.4,5 Education of healthcare providers and their patients about the requirement for sunlight exposure or dietary supplementation to prevent vitamin D deficiency rickets needs to continue.

Graeme H Johnson · Francis Willis

Comment: Nutritional rickets is highly prevalent in countries such as Mongolia, Tibet and China1 where winter sunlight is reduced and there is no universal vitamin D supplementation. Paradoxically, rickets is also prevalent in developing countries in the tropics and subtropics where sunlight is unlikely to be a limiting factor. Low calcium intakes (including vegetarian diets), prolonged breast feeding, and covering of the skin may all contribute.2,3 Published reports and clinical experience in Sydney suggest an increase in prevalence of rickets, especially in infants and mothers in immigrant populations2,4-7 Other Western countries have reported similar findings. The vitamin D deficiency described by Johnson and Willis in an infant in Perth highlights a high-risk group — infants of mothers who are veiled. Treatment of associated nutritional deficiencies, especially iron deficiency,4,5 and giving a minimum of 300 000 IU of vitamin D over 6–8 weeks, should resolve the rickets. Data on the epidemiology of vitamin D deficiency in these high-risk groups in Australia and other Western societies are lacking and should be the subject of future research. The major source of vitamin D and its circulating form, 25-hydroxyvitamin D3 (25OHD3), in children and adults is the skin. It is estimated that exposure to sunlight for 15 minutes three times per week normalises 25OHD3 levels.8 Dark skin, increasing age, sun protection agents, and the angle of the sun in winter will attenuate this increase in 25OHD3.8 Neonates acquire their vitamin D3 stores from their mothers via the placenta, with only a small amount transferred in breast milk.9 By screening high-risk pregnant women, specifically veiled women and those with dark skin,10,11 prevention of most cases of infant rickets is possible. Levels of 25OHD3 should be measured and, if low, the mother should receive 4000 IU of vitamin D daily until 25OHD3 levels are normal. There is currently no recommendation for routine supplementation of vitamin D in infants, and most cereals and foods are not fortified with vitamin D. However, infant formulas are supplemented with 200 IU of vitamin D per litre. It is estimated that sufficient vitamin D levels to prevent rickets could be achieved if 400 IU of vitamin D were provided daily as part of a multivitamin supplement to high-risk infants.

Christopher T Cowell

To the Editor: It has been well established that improved compliance in hand washing significantly reduces hospital-acquired infections and cross transmission of methicillin-resistant Staphylococcus aureus.1 Although there have been no reports of infections resulting from cross-contamination via stethoscopes, studies have demonstrated that 80%–100% of these appliances are colonised by bacteria.2-5 However, most of the organisms isolated are considered non-pathogenic. The most common potentially pathogenic organism isolated from stethoscopes is S. aureus, with a prevalence of 4.2%–27.5%.2-5 Cleaning with either 70% isopropyl alcohol or benzalkonium chloride wipes can reduce the bacterial count on stethoscopes by 94%–100%.3,5 We undertook a study to assess the prevalence of S. aureus carriage on stethoscopes and hands of staff of the Canberra Hospital, and to measure the effectiveness of cleaning and hand washing in reducing colonisation. A convenience sample of healthcare workers from various areas of the hospital was obtained over a period of 6 months. The diaphragm of each participant's stethoscope was directly impressed on to mannitol salt agar, before and after being cleaned with a 70% isopropyl alcohol wipe. The dominant hand of each participant was also tested before and after washing with triclosan (1%) antimicrobial handwash and water. S. aureus was identified by standard laboratory methods. There were 134 participants: 69 doctors, 50 nurses, 10 medical and nursing students and five physiotherapists. Most doctors and physiotherapists used their own stethoscopes, whereas most nurses and students used ward stethoscopes. S. aureus was isolated from five stethoscopes before cleaning (4%), but from none after cleaning. The organism was also isolated from the hands of 11 people before hand washing (8%) and of one after hand washing (0.7%). Two people had S. aureus isolated from both sites. There was no statistically significant difference between the prevalence of S. aureus on hands and stethoscopes (P = 0.15, McNemar's test). Hand washing is the best recognised means of preventing cross-contamination in hospitals. However, the simple intervention of cleaning stethoscopes with an alcohol wipe was highly effective, and we believe that this practice should be more widely promoted.

Karina J Kennedy · Dianne E Dreimanis · Wendy D Beckingham · Francis J Bowden

To the Editor: At the Australasian Schizophrenia Conference in Sydney in October 2002, Professor Patrick McGorry of the Orygen Research Centre, University of Melbourne, presented draft guidelines on the management of schizophrenia and early psychoses.1 One of the recommendations was that atypical antipsychotic drugs should be used as the first-line pharmacological treatment in preference to typical antipsychotics and depot antipsychotics. With a shift in management of schizophrenia to community-based care, the number of patients with schizophrenia managed by general practitioners has increased over the past decade (from 36 per 10 000 encounters in 1990–91 to 45 per 10 000 in 2000–02).2 With the pending introduction of the guidelines, a baseline measure of GP prescribing rates of antipsychotics, both typical and atypical, will allow future measurement of the impact of the guidelines. We analysed the 1998–2002 data from the Bettering the Evaluation and Care of Health (BEACH) program, a continuous national cross-sectional survey of general practice.3 About 1000 GPs participate in this program every year, each providing details (on structured forms) about 100 consecutive patient encounters. Data collected include GP and patient characteristics, problems managed and treatment provided. We examined 401 300 encounters from 4013 GPs, with 431 537 medications recorded. Prescription rates were calculated and regression analyses performed using SAS software4 to adjust for the cluster effect of the study design. There were 1988 schizophrenia or psychosis problems managed in the four years of data collection (a rate of 49.5 per 10 000 encounters); 1883 medications were prescribed (94.7 per 100 contacts), of which 926 (49.2%) were typical antipsychotic drugs and 484 (25.7%) were atypical antipsychotic drugs. In 1998–99, the prescription rate of atypical antipsychotics was 15.7 per 100 contacts with patients with schizophrenia or psychosis (95% CI, 11.8–19.7). This rate increased to 31.1 per 100 contacts (95% CI, 26.3–36.0) in 2001–02. In the same period, the prescription rate for typical antipsychotics fell from 51.3 per 100 contacts (95% CI, 45.6–56.9) in 1998–99 to 40.6 per 100 contacts (95% CI, 35.4–45.8) in 2001–02. Linear regression showed that the prescription rate of atypical antipsychotics had increased by an average of 5.1 per 100 contacts per year over the four years (P < 0.0001), while the prescription rate of typical antipsychotics had decreased by an average of 3.8 per 100 per year (P < 0.005). Over the four-year period, there was no significant increase in the rate of overall prescriptions for people with schizophrenia (92.0 prescriptions per 100 contacts [95% CI, 83.4–100.7] in 1998–99 v 96.3 [95% CI, 89.4–103.3] in 2001–02). Between 1998 and 2002, the relative prescribing rate of atypical antipsychotics for schizophrenia and other psychoses in general practice nearly doubled. These results show that, even before the introduction of the guidelines, there has been a shift towards prescribing atypical antipsychotics in preference to typical antipsychotics. This change may reflect a change in specialist behaviour, as specialists have a direct effect on GP prescribing.5 The BEACH study will be able to assess the effect of the guidelines on the prescription rate of atypical antipsychotics by GPs.

Christopher M Harrison · Helena C Britt

To the Editor: Figures from the Australian Bureau of Statistics show that Australians make about 3.3 million overseas departures each year.1 Few data are published on the extent to which Australian travellers seek pre-travel health advice, what vaccinations they receive, and what risks they are exposed to during travel. A series of surveys of travellers examining these questions has been conducted under the auspices of the Travel Health Advisory Group, a coalition of Australian travel and medical organisations. Surveys were conducted in 1996, 1997, 2000, 2001 and 2002. On each occasion, a market research company telephoned people from mainland capitals using numbers randomly selected from the telephone directory. This process continued until 500 people aged 18 or over who had travelled overseas in the previous two years had been interviewed. In the 2002 survey, about 10 000 calls were made to complete the interviews. The questions on vaccinations focused on hepatitis A and B, two of the most common vaccine-preventable diseases associated with travel.2 Results from the 2002 survey are shown in the Box. A minority of people (31%) reported seeing a doctor or travel clinic for pre-travel health advice. Of those who saw a doctor or travel clinic, 31% did so two weeks or less before departure. Over the series of surveys, there has been an increase in travel to destinations with high- or intermediate-risk for hepatitis A infection, from 40% of travellers in 1996 to 58% in 2002. Despite this increase, only a minority of travellers could recall ever being vaccinated for this illness. Travellers were informed about how they might be exposed to hepatitis A and B and asked if they believed that they could have been at risk during their most recent overseas trip. Substantial numbers recalled a risk (34% and 16% for hepatitis A and B, respectively) and some of these could not recall being vaccinated (18% and 7%, respectively) (Box). These surveys have methodological limitations, including lack of information on the consent rate and potential recall bias. However, the results suggest that substantial numbers of travellers do not seek pre-travel health advice and are at risk of vaccine-preventable diseases during travel. Not only does this have implications for individual travellers, it also creates public health risks, as travellers can introduce hepatitis A and B into their home communities. These results suggest more public education is needed about the importance of pre-travel health advice and appropriate vaccination. Results of 2002 survey of Australian travellers overseas Variable No. of travellers (n = 500) Age (years) 18–29 156 (31%) 30–49 188 (38%) ≥ 50 153 (31%) Not stated 3 (0.6%) Male sex 205 (41%) Hepatitis A risk in country visited* High 226 (45%) Intermediate 62 (12%) Low 212 (42%) Sought pre-travel health advice from doctor or travel clinic Doctor 135 (27%) Travel clinic 20 (4%) No professional advice 345 (69%) Believed could have been at risk of hepatitis on most recent trip Hepatitis A 168 (34%) Hepatitis B 79 (16%) Vaccinated against hepatitis Hepatitis A 195 (39%) Hepatitis B 197 (39%) Believed at risk of hepatitis on most recent trip and not vaccinated or unsure Hepatitis A 91 (18%) Hepatitis B 36 (7%) Hepatitis A risk in country visited* among those not vaccinated for hepatitis A or unsure High 123 (25%) Intermediate 40 (8%) Low 142 (28%) * As defined by the United States Centers for Disease Control and Prevention, 2000.3 Destinations for the cohort of 500 were Asia (45%), northern Europe (28%), southern Europe (14%), North America (15%), Oceania (14%), eastern Europe (3%), Africa (3%), Middle East (2%), South America (1%) and Central America (1%), with some having more than one destination.

Nicholas A Zwar

To the Editor: I would like to comment on a recent article by Young and Dobson on the decline of bulk-billing and the increase in out-of-pocket expenses.1 It is true, as the authors state, that "Australia has no legislation restricting how much a general practitioner can charge for a consultation". It is a shame that they did not take equal time to point out that there is no binding requirement on governments to ensure that Medicare rebates remain within striking distance of the real cost of service provision. Policy change is required — with as much political and moral urgency as Young and Dobson advocate for patient access reform — to enable doctors to provide affordable healthcare under a fee-for-service system without being penalised for accepting a substantial number of elderly or socially disadvantaged patients. Primary healthcare policy needs to be adjusted to maintain rebate justice for low-income patients by linking patients' rebates to their doctors' real-life market costs (or even to relative value studies), not budget "bottom lines". Is the patient to be out of pocket, or the doctor? Surely, both positions are equally unfair, and equally unlikely to bring about an equitable system.

Warwick H Ruse

To the Editor: Young and Dobson1 have confirmed what has been long suspected by many rural doctors and patients — that women (and patients in general) in rural areas are paying higher out-of-pocket costs for general practice consultations than those in urban areas. I wonder whether the authors have considered analysing the data by State and Territory, as the structure of a State healthcare system often has a significant impact on healthcare costs to individuals. I am also interested to know whether total costs of healthcare (including on-costs, referrals, specialist fees, hospital and investigative care) were analysed. There is a long held view that rural doctors have a more holistic approach to patient care than their urban colleagues, who are more inclined to recommend unnecessary investigations and specialist referrals. It would be fascinating to know whether urban women would actually be more "out-of-pocket" than rural women if total contact with the healthcare system (not just general practice consultations) was taken into account.

Chris A Harrison

To the Editor: Do the data in Young and Dobson's study of general practice consultation fees1 support their conclusion that women in rural and remote areas lack access to affordable healthcare services? The declining prevalence of bulk-billing by general practitioners suggests that healthcare may be becoming increasingly unaffordable for people on lower incomes. But affordability is not only income-related — it depends also on choices regarding discretionary expenditure. Considering median levels of disposable weekly income, together with the prices of basic daily commodities such as milk or bread (not to mention a $10 pack of cigarettes!), how "unaffordable" is an occasional $5–$10 out-of-pocket fee for a GP consultation? As Young and Dobson concede, a further complicating factor in their study was that "the consenters . . . tended to have higher socioeconomic status and so may be less likely to be bulk-billed". It would have been helpful if the authors had defined affordability of GP care in relation to family income (perhaps analogous to advice that rental or mortgage repayments should not exceed a third of household disposable income) and identified just how much their frequent attenders' GP costs exceeded a specified limit of affordability. Services free at point of delivery are overused, both by patients and doctors, as evidenced by the Commonwealth's implementation of the Professional Services Review Scheme and the States' legislating to curb the costs of workers compensation and third-party motor vehicle insurance claims. There is probably an optimal price range that would facilitate affordable access without penalising the less affluent or encouraging "inappropriate practice". Closer attention to the affordability of GP services for individual households would help target resources to people who truly need the services but cannot afford them. This may be a better alternative to the authors' suggestion of simply making policy changes (taxpayer-funded?) to lower the price of GP services for all women in rural and remote Australia.

Peter C Arnold

In reply: We thank the authors of these letters for raising many of the complex issues that underlie the current geographical inequities in costs of general practice consultations. Ruse is concerned that we did not point out the inadequacy of the current Medicare rebates to practitioners. An appraisal of the adequacy of Medicare Benefits Schedule fees was beyond the scope of our study. We presented data on the out-of-pocket costs of general practice consultations, according to demographic and health-related characteristics of consumers, for consideration by all interested parties — practitioners and patients. As suggested by Harrison, we could also look at differences in bulk-billing and costs by State and Territory. However, as Medicare rebates are a Commonwealth issue, looking at national data seemed a sensible first step. We cannot examine total costs of healthcare, as the Medicare data do not include all costs related to care. Arnold argues that "affordability" should be better defined by us and questions, "how unaffordable is an occasional $5–$10 out-of-pocket fee for a GP consultation?". We have three responses. Firstly, many medical practices require an "up-front" payment. As written by one older respondent living in a rural area, "Small country town medical clinics do not give bulk-billing to aged pensioners and insist on cash payment on the day of the visit . . . many pensioners would not seek medical help when needed if at the time no cash was available". Our second response is that, regardless of how "affordability" is defined, the issue is one of equity. Is it reasonable that a major factor identified in our study as influencing access to bulk-billing is whether you consult a practitioner in an urban area or a rural area? Finally, "affordability" can only be assessed in relation to income and other expenditure and commitments. In our surveys we ask women how satisfied they are with the costs of GP care and, while the responses are subjective, they are likely to take into account these contextual issues. These data have not yet been fully analysed.

Anne F Young · Annette J Dobson

To the Editor: I note with interest the case study of experimental Ancylostoma caninum infection in a 22-year-old student.1 In his accompanying editorial, Van Der Weyden highlights the courage of these researchers, as well as some of the risks and discomforts associated with their participation,2 including in two studies in which he was a co-author. However, it is also worth highlighting some of the ethical issues associated with such experimentation. Larry Altman, who provided many of the quoted examples of self-experimentation, also reflects on Walter Reed's experiments with yellow fever vectors in Cuba. Although later credited with the use of written consent forms, on an earlier occasion Altman alleges that Reed withdrew at the last moment from inoculation experiments in which one of his colleagues died.3 Although not mentioned in the published work, the A. caninum experiment was initiated and undertaken by Landmann under the supervision of Prociv, who has himself self-experimented with both A. caninum and Necator americanus (human hookworm) in previous work (Juergen Landmann, Student; Paul Prociv, Senior Lecturer, Department of Microbiology and Parasitology, University of Queensland, personal communication). In this case, the study was wholly initiated by the student so consent was not an issue, but other such studies raise the potential problem of consent in situations of an unequal power relationship. Students under supervision constitute a "vulnerable" group in that consent may be given under a form of duress.4 Just as special protection is needed for populations for whom research is combined with care, protection is required for students who may feel obliged to participate in such research. There have been recent calls for a fuller discussion of ethical issues in published experimental studies,5 where ethical justification should be accorded the same weight as statistical considerations. The unusual study by Landmann and Prociv highlights the need for such discussion in potentially controversial research protocols.

Allen C Cheng MB BS, FRACP

To the Editor: The most recent newsletter of the New South Wales Medical Board1 opens with a plea by the president for better understanding of the Board's initiatives, inspired by the "public interest". We hope that the board recognises the gulf between legitimate public interest and unrealistic, illegitimate public expectation. The newsletter goes on to justify the Board's policy against doctors self-prescribing, with six examples (hardly significant from a register of 25 000). The first, Dr A, aged 70 (the only one whose age was given) was referred to the Board by colleagues for mental impairment, and was not self-prescribing. He was prescribed warfarin by his cardiologist, but also took aspirin. The other five were all involved with drugs of addiction, earning whatever sanctions the Board applied. The newsletter then states that these were "ordinary doctors providing ordinary services in the community" (p. 3).1 They were certainly not, and it is this patronising assessment of the behaviour of ordinary doctors that is objectionable. We are next told that "these dramatic examples represent the tip of an alarming iceberg" (p. 3).1 The Board has enough to do without plumbing the depths for imagined "icebergs", and should reconsider their policy on the alleged dangers of self-prescribing, a policy both unwarranted and unwanted. The mocking adage that "doctors who treat themselves have a fool for a physician" is not made correct by repetitive quotation. There are not many fools in our profession, and sanctions applied to them should not affect the Board's assessment of the whole profession. Although not mentioned in this newsletter, similar motives are apparent in the Board's disapproval of self-referral, where we are not credited with sufficient wit to discover for ourselves appropriate specialists for clinical referral. All medical boards are currently seized with similar agenda. The Victorian Medical Board ponders the "problem" of retired doctors, with the comment in their newsletter on doctors affected by "increasing age and commensurate reduction in cognitive ability", another gratuitous observation without supporting evidence. They could as correctly, and more kindly, have referred to the accretion of clinical wisdom commensurate with age, but respect for seniors seems to have declining value in current professional ethics.

G Douglas Tracy

In reply: Tracy's letter seems to focus on age. The Board's policy about treating family members is not about age, but about the wisdom or otherwise of this practice for medical practitioners at any stage of their career. The policy is not mandatory, but reflects what the Board considers to be prudent practice. The policy does not prohibit writing referrals or repeat prescriptions, but emphasises the importance of having an independent treating practitioner responsible for initiation of treatment and ongoing management. The New South Wales Medical Board is not alone in having such a policy, with similar views being expressed by UK's General Medical Council, the Medical Council of New Zealand, and Canadian, American and other Australian medical boards. The Australian Medical Association position statement on the "Health of medical practitioners" emphasises the importance of medical practitioners and their families having their own general practitioners. The case studies were published following a request from members of the profession for the Board to provide examples of problems arising through treating themselves or family members. Sadly, there are many more instances than the five referred to in the article. The doctors were certainly not ordinary once their attempts to treat family members went astray, but the point is that they had been ordinary doctors who got into difficulties because they crossed the professional–personal boundary.

Brian C McCaughan

To the Editor: This is the first report of the use of magnesium sulfate to treat Irukandji syndrome. A previously well 26-year-old commercial diver was stung on the neck by a jellyfish while collecting sea cucumbers in Barrier Reef waters off Townsville in February 2003. As is typical for an Irukandji syndrome, he was asymptomatic for about 30 minutes, after which he developed back and abdominal pain, nausea and headache. He was retrieved from the scene by helicopter and arrived in the emergency department (ED) two hours after the onset of symptoms. On retrieval he had a blood pressure of 150/90 mmHg, agitation, marked diaphoresis, piloerection and some dyspnoea. A typical carybdeid jellyfish sting mark was present on the neck. The cardiac troponin I level was elevated from the time of admission. En route and in the ED he was treated with intravenous morphine and diazepam. Skin scrapings were taken for nematocyst identification. After he had received 27.5 mg of morphine and 15 mg of diazepam, his pain settled somewhat, but abdominal discomfort, agitation and profuse diaphoresis persisted. Despite the dyspnoea, he showed no other overt clinical signs of cardiac failure. Concern with the patient's increasing hypertension (170/100 mmHg five hours after envenomation) led to his being transferred to the high dependency unit (HDU) six hours after envenomation. It was decided to try a therapeutic trial of magnesium sulfate for this patient in an attempt to control the hypertension. This decision was taken on the basis of: the unsatisfactory results of measures taken thus far, the postulated hyperadrenergic basis of hypertension in Irukandji syndrome, the known 20%–30% fall in systemic vascular resistance associated with magnesium administration in hyperadrenergic states,1 and considerable experience within the HDU with managing severe pre-eclampsia. Intravenous magnesium sulfate was administered as a loading dose of 10 mmol followed by an infusion of 5 mmol per hour. Sympathetic features and agitation resolved, and pain nearly completely resolved towards the end of the loading dose. Of note, an early reduction in the rate of magnesium sulfate infusion resulted in recrudescence of hypertension, back pain and piloerection. The infusion was uneventfully reduced to 3 mmol per hour at 11 hours after envenomation, and discontinued at 20 hours after envenomation. No adverse effects related to the magnesium infusion were noted. The patient subsequently remained well. The troponin I level rose to a peak of 6.4 μg/L, and an echocardiogram was normal at 20 hours after envenomation. Irukandji syndrome is produced by carybdeid jellyfish envenomation2 and has been shown (in animals) to be associated with dramatically elevated serum noradrenaline levels.3 Severe hypertension in Irukandji syndrome can be difficult to treat and has been associated with two deaths from intracranial haemorrhage. The origin of the extensive, severe pain associated with the syndrome is unknown. Postulated mechanisms include ischaemia from widespread small vessel vasoconstriction resulting from a hyperadrenergic state, and sodium channel opening in afferent pain fibres. Other mechanisms are equally likely. Induced catecholamine release or direct toxicity have been proposed as the cause of myocardial injury. This may produce overt, and occasionally severe, cardiac failure. Magnesium decreases both catecholamine release and sympathetic terminal receptivity to catecholamines1 via multiple sites of action, including most calcium channel subtypes (both at the cell membrane and intracellularly), as well as modifying other cation fluxes. It reduces catecholamine-induced myocardial necrosis in phaeochromocytoma (Professor M James, Department of Anaesthesia, University of Cape Town, personal communication) and is widely used in other hyperadrenergic states, such as phaeochromocytoma and pre-eclampsia.1 The apparent efficacy of intravenous magnesium in our patient suggests the need to further investigate this therapy. A larger case series, a multicentre randomised trial of magnesium sulfate administration in Irukandji syndrome and a dose-finding study are under way.

Michael A Corkeron

To the Editor: We wish to comment on the article by Whiting and colleagues1 on the proposed causal relationship between herbal remedies and severe acute hepatitis. Investigators have found that reported adverse effects of herbal medicines are not, in fact, caused by herbs alleged to be in the product, but result from substitution or contamination of the declared ingredients, intentionally or by accident, with a more toxic herb, a poisonous metal or even a pharmaceutical compound.2,3 In reports of adverse effects, there is often no effort to establish a positive identification of the herb involved or any adulterants. The attribution of toxicity to the wrong plant leads to inaccurate information being provided to patients, practitioners and regulators.4 A significant problem is the use of common names. As mentioned by Whiting and colleagues,1 Cimicifuga racemosa (black cohosh) has at least 20 different common names, which can be very confusing. The most tragic example of such confusion is the fatal substitution of Stephania tetrandra by the toxic herb Aristolochia fangchi owing to the similarity of the common names of the two herbs.5 In recording and responding to adverse events involving herbs, certain key questions need to be asked by those reporting the event and, more crucially, by those subsequently citing the report. No details regarding verification of the herbal products taken by the individual patients were supplied by Whiting and colleagues.1 Because of this failure to authenticate the plant compounds in the preparations, one cannot establish that the herbs were the cause of the hepatotoxicity. No information about plant parts used, solvent, concentration, manufacturing process or chemical analysis was supplied. Although Whiting et al exclude other causes for hepatitis, external factors may have contributed to the reported liver reactions. Hepatitis for which no cause can be identified is not uncommon.6 In addition, absence of hepatitis B surface antigen does not exclude the possibility of hepatitis B virus infection.7 The correlation of the liver diseases with preparations of Cimicifuga racemosa is speculative, as viral causes were not definitively ruled out. Without further pathophysiological or biochemical investigation, no conclusion can be made as to the exact mechanism. In a review of eight human studies on the effectiveness of an extract of black cohosh for alleviating menopausal symptoms, the authors concluded that black cohosh appears to be a safe, effective alternative to oestrogen replacement therapy for patients in whom oestrogen replacement therapy is refused or contraindicated.8

Luis Vitetta · Michael Thomsen · Avni Sali

To the Editor: We urge caution about the recently published findings of Durna and colleagues on use of hormone replacement therapy (HRT) by breast cancer patients.1 They found no increase in risk of breast cancer recurrence or reduction in life expectancy in women using HRT after treatment for primary breast cancer, leading them to conclude that HRT seems a safe treatment for women with a history of breast cancer. We believe that their study has many limitations and consequently their conclusions are unfounded Durna's group conducted a retrospective, observational study using unmatched patient groups. The HRT group was statistically younger, with smaller tumours and fewer positive nodes, and thus had better prognosis. The authors discussed regression analyses for age of diagnosis and stage to reduce potential bias, but did not include these data. The omission of tumour grade and receptor status in the analysis is an obvious flaw. The mean duration of HRT use was short compared with many previous studies (mean, 1 year). This is particularly important, as any risk is likely to be cumulative.2 There were also statistically more women in the HRT group who had received HRT before diagnosis, and for a longer duration (mean, 6.5 years v 3 years). This too may bias results, as breast cancer that develops after HRT has been suggested to carry a better prognosis.3 There was also no subgroup analysis of those who took concomitant tamoxifen. Durna and colleagues also do not address the concerns raised by the Women's Health Initiative study, which showed an excess of cardiovascular and thromboembolic events in "healthy" women taking HRT.4 Many of these events occurred in the first years of HRT use, the time assessed in Durna's study. The risks of short-term HRT in breast cancer patients may be not only tumour-related, but may also include cardiac and thrombotic events, which were not specifically considered by Durna and colleagues. Of additional concern is the extensive portrayal in some sections of the media that this study is conclusive regarding the risks of HRT use by women with breast cancer. The importance of correct media interpretation of studies is discussed in the accompaning editorial by Patel and colleagues.5 Despite this editorial and the inadequacies of Durna's study, the results of this study, released through the media, inaccurately suggested safety and even a possible benefit of HRT.6 This simply adds to the "HRT furore". All studies addressing this issue to date have been small and non-randomised, often with no control group or unmatched groups.2 Because of these limitations, no clear recommendation can be made about the safety of HRT in women with breast cancer. Until results are available from rigorous randomised controlled trials, such as the ongoing HABITS (Hormone Replacement Therapy After Breast Cancer — Is It Safe?) study (IBCSG 17-98), we should not be "advocating" HRT to breast cancer patients. For Durna's group to suggest otherwise is very premature. In addition, there are well studied, effective alternatives to HRT for treating menopausal symptoms in breast cancer patients (eg, venlafaxine),7 and these would seem the current safer option.

Theresa M Hayes · Craig R Underhill · Kerrie Clarke · Martin HN Tattersall

In reply: We feel that Hayes and colleagues repeat the editorial comments of Dixon1 that accompanied our article.2 Our investigation, being retrospective, had biases and limitations normally associated with that type of study. In the third and fourth paragraphs of the discussion section of our article, we examined further the sources of bias and their impact. Hayes and colleagues state that the omission of tumour grade and receptor status from our analysis is a flaw. We agree that including these variables would have made our study more complete, even if we doubt that this would have altered the conclusions. So many of these data were missing that statistical analysis would have been unreliable. As was stated in the methods section, the analysis was adjusted for the covariate HRT use before diagnosis. Tamoxifen use was very similar among HRT users and non-users (58% and 60%, respectively), and any bias introduced by this difference would be insignificant. The scope of our article, as suggested by its title, was limited to the issue of cancer recurrence and mortality. We did not specifically investigate the impact of HRT on cardiovascular and thromboembolic events. Nonetheless, if car-diovascular disease had increased mortality, this would have been apparent as an increase in all-cause mortality. We agree with Hayes and colleagues that it is vital that the media interpret studies correctly for the general public. We are disappointed that Hayes and colleagues have misinterpreted our conclusion. Our article does not advocate HRT use by women with breast cancer. We explicitly stated that "These results need to be confirmed in a randomised trial before HRT can be advocated for all women who have had breast cancer." Furthermore, ". . . the observed association between HRT use and [reduced] risks of breast cancer recurrence and death cannot be inferred to be causal". We believe that we are justified in our conclusion that HRT use in women diagnosed with breast cancer is not associated with an increased risk of recurrence of breast cancer or a shortened life span. It is our clinical practice to use HRT only when alternative therapy fails.

Eva M Durna · Barry G Wren · Gillian Z Heller · Leo R Leader · Peter Sjoblom · John A Eden

To the Editor: McBride reports that there were four deaths from cancer before the age of 40 years in a group of 480 thalidomide-affected people in the United Kingdom, an approximate cumulative mortality rate of 0.83%.1 He compares this with the annual death rate in the under-40-years age group, and concludes that the rate is increased almost 100 times in those affected by thalidomide. The correct comparison is with the cumulative mortality rate in the general population from birth to age 40. From 1999 UK statistics,2 this is about 0.31% — that is, we would expect 1.48 deaths among 480 people. While the observed number of four is greater than this, it is only slightly greater, and the difference is not statistically significant (the mortality ratio is 2.7, with exact 95% confidence limits of 0.7 to 6.9, based on a Poisson distribution). McBride's conclusion is based on an inappropriate comparison. A full analysis would use population death rates over the 40-year period and take account of censoring, but that is unlikely to affect the result substantially.

J Mark Elwood