Letters

To the Editor: Dugdale proposed recently in the Journal that people who have a skin reaction to mosquito bites are less likely to be infected by Ross River virus than those who do not.1 As he quotes Kumar, who made a similar comment about malaria infection,2 one could presumably extend his idea to other conditions transmitted by mosquitoes. This accords with my own personal experience of dengue fever acquired in Fiji. While serving there, I had two separate proven infections with dengue virus. As I react very little to mosquito bites, I could not identify the time of infection. Indeed, on the first occasion, I had just returned from a three-month stay in Adelaide and could not recall being bitten by a mosquito at all. In contrast, my wife, who developed large weals whenever bitten, went through at least three epidemics of dengue without being infected. My advice to travellers who consult me is that there is an advantage to reacting badly to mosquitoes, as one is then more likely to take anti-mosquito precautions, whereas the non-reactor is more likely to disregard them. However, Dugdale's suggestion that a local inflammatory reaction may be a factor in defence against infection is intriguing and worth following up. It should be simple to enquire retrospectively about reactions to mosquito bites in those who have had a mosquito-borne disease, as Dugdale has done for Ross River virus infection. This would provide evidence on which to base pathological and immunological studies.

Michael Sorokin

To the Editor: Dugdale recently suggested that people who develop skin reactions to mosquito bites may be protected against Ross River virus (RRV) disease.1 He noted that seven patients with a history of RRV disease all reported no skin reaction to mosquito bites, and 18 patients with no past history of RRV disease reported reacting to such bites. We argue that this correlation is entirely to be expected, as the lack of reaction to mosquito bites illustrates that the individuals have been previously exposed to many bites. At least two studies have shown a clear inverse correlation between mosquito exposure and bite reactions.2,3 Clearly, exposure to a large number of mosquito bites increases the risk of infection.4 Thus, a reaction to mosquito bites probably does not protect against RRV disease, but is simply a marker for low exposure to mosquito bites and therefore low risk of RRV infection. The rationale behind the association of itching bites and protection against RRV infection is also tenuous. Virus is likely to reach the circulation within seconds of introduction by the mosquito, whereas allergic reactions take minutes to develop. It is unlikely that a local reaction will affect viral replication at distant sites. In Dugdale's study, only people with a history of symptomatic RRV disease had undergone RRV serological testing. As about 30% of Queenslanders are seropositive,5 and 60%–75% of RRV infections are thought to be asymptomatic,6 some of the 18 people with no past history of RRV disease might be expected to have had asymptomatic RRV infection. This raises the question, does reaction to mosquito bites correlate with asymptomatic RRV infection?

Peter A Ryan · Jillann F Farmer · Brian H Kay · Andreas Suhrbier

To the Editor: I support Tattersall and colleagues1 in their attempt to use a wee bit of science and a dose of common sense to improve doctors' letters. While I would love to receive letters based on their proposed "prompt sheet", and should really try much harder myself to follow it, I'd be only too pleased — in the interim, while we await enlightenment — to receive any letter from some of my colleagues with whom I share what is supposed to be "multidisciplinary care" of patients. There is a research project in this for the brave to find out why, too often, communication is not merely inadequate but non-existent.

Alan Rodger

In reply: The support for my study1 is pleasing, especially the letter from Martin. I agree with Hockey and Pitt that including Queensland data would have enhanced my study, as Queensland is the only other Australian State with a substantial injury database like South Australia's and Victoria's. However, when I analysed the Queensland data only around 1% of baby-walker injuries could be classified as "proximity", compared with 20%–25% in Adelaide and Melbourne. Surveillance collection can very easily miss critical details. The SA questionnaire asks "What was the victim doing at the time of the injury?", then "What went wrong?" and, finally, "How exactly was the injury caused?". As an example, "a child in a baby walker accesses the fireplace and burns her hand". If one or two of the above questions are left out, or the coder does not capture all the detail, this narrative easily becomes "child burns hand on fireplace" and the detail that the baby walker facilitated the child's access to the fireplace is lost. In 1995, after the Victorian Injury Surveillance System moved to a "minimum" dataset, their "proximity" component dropped to just 1%, the same as for Queensland. Analysts at the Victorian Injury Surveillance System recommended that these post-1995 data not be used for my study. In my opinion the Queensland system has the same systematic problem, and discussions with Hockey suggest that this is a possibility (Richard Hockey, Senior Data Analyst, Queensland Injury Surveillance Unit, personal communication). My assertion — that the data I used are largely representative of baby walker injury events in metropolitan Australia — is justified, as any variations in "proximity" rates (even if they do exist) caused by different architectural styles in Queensland would represent only a very small proportion of the events Australia-wide. The presence of additional steps and higher steps would only explain a higher proportion of these types of injuries, and possibly increased severity, but not the almost total absence of "proximity" injuries, as there is no association between them. The recent finding by the US Consumer Product Safety Commission2 that steps and stairs injuries still occur with new-style walkers confirms the need to discontinue the proposed mandatory regulations and instigate an immediate ban. The letter from Beard is disturbing in its attitude. Inadequate supervision is another name for victim blaming, and an unsafe environment is a perfect description of a baby walker. Moreover, to suggest that the ban would include any form of purpose-designed apparatus, including a commercial baby walker, to assist disabled children is nonsense.

Peter G Thompson

To the Editor: The article on the risks of baby walkers by Thompson is welcome.1 However, he used a broad definition of "serious" (ie, those children who required admission to hospital). Readers should be aware that some of these accidents cause major disfiguring injuries with life-long consequences. Burns sustained as a result of being in a baby walker are more extensive and deeper than those of the average patient admitted to our Burns Unit. In the 5 years to July 2000, 24 patients were admitted to the Burns Unit at The Children's Hospital at Westmead (the NSW State Paediatric Burn Referral Unit) with burns as a result of being in a baby walker. As baby walkers are used by children before they can walk, all were 13 months old or younger (range, 6–13 months), whereas only 15% of all admissions are under one year. In the same 5-year period, 178 patients were under 12 months, with the number rising to 246 if those under 13 months are also included. Only one of the group with burns from using a baby walker was older than 12 months, and only by a day. (Ironically, he was burnt while playing in a baby walker that he had been given for his birthday the day before.) The baby-walker burns were extensive — the largest burn covered 48% of the body surface, with the average being 17%. Deep burns were more common than average, 19 of the 24 children needed grafting (80%) versus only 50% of all admissions. Grafting means that the burns are deep enough to cause permanent scarring. Eight of the 24 required admission to the intensive care unit. Of the 24 children, 21 had scald burns, mostly due to the child being able to reach a cord, tablecloth or container; two had contact burns because the child could reach out to a hot stove; and one sustained a flame burn, because the child moved close to a parent starting a fire with flammable fluid. Thus, the danger results from the child being able to reach higher and move faster than expected. No improvement has been evident in the last 2 years. From July 2000 to June 2002, seven more children were admitted with burns caused by baby walkers: six spill scalds (four water and two oil) and one contact burn (iron pulled down). This helps confirm Thompson's contention that regulations (introduced in September 2000) will not stop baby-walker injuries. Apart from the danger of severe injury, baby walkers impede normal balanced muscle development and so are deleterious to every child who uses one. Apart from being a child-minding device, they have only potential for harm. Now that Thompson has shown that the "safety standards" do not protect children from injury, it is time for them to be banned altogether.

Hugh C O Martin

To the Editor: Thompson's report highlighting the danger of baby walkers and the applicability of the proposed Australian mandatory standard is welcome.1 However, we have some concerns with the statement that the injury data used in the analysis are "largely representative of such events in the major proportion of metropolitan Australia". The only data sources used were the South Australian Department of Human Services Injury Surveillance System and the original Victorian Injury Surveillance System. Equivalent data from Queensland, although readily available, were not used. An analysis of 177 baby-walker injuries recorded in the Queensland Injury Surveillance Information System (ISIS) for the period 1989–1994 shows that only 9% of injuries were not stability related, and that 73% were associated with steps and stairs. Among the cases admitted to hospital, only one was not related to stability or falls down steps or stairs. Examination of more recent Queensland injury surveillance data, although not directly comparable, reveals a similar pattern. These results, which appear to contradict the findings of Thompson, are not altogether surprising considering the differences in architectural styles between Queensland, with its high-set houses with verandahs and wooden steps, which present quite a different injury-prevention challenge to the style of housing more common in Adelaide and Melbourne. The introduction of a baby-walker standard to address the problem of stability and falls down steps and stairs would thus appear worthwhile in this setting. However, we accept that, in other parts of Australia, baby-walker-related injuries may be less amenable to prevention through this intervention. It is also of concern that the new standard is being promoted by the Commonwealth as able to prevent stair-related baby-walker injuries. In reality, this is not the case: a recent study by the US Consumer Product Safety Commission found that 15% of stair-related injuries involved the new-style walkers.2 Furthermore, introduction of the standard, and its ensuing publicity, may result in an increase in use of baby walkers, along with a perception that they are now safe. This has the potential to increase baby-walker injuries. It is our belief that, in this particular case, it may be better to keep publicity very low key. Finally, use of baby walkers has also been associated with a delay in normal physical development.3 For all these reasons we would continue to advocate that the use of baby walkers be actively discouraged.

Richard L Hockey · Rob Pitt

To the Editor: The article by Thompson recommended that baby walkers be banned.1 He described various injuries caused by baby walkers: An 8-month-old baby in a walker pulled the cord of a deep fryer and was splashed by oil, sustaining full-thickness burns to chest, abdomen and upper arms. A 10-month-old baby in a walker sustained severe finger laceration when a fly-screen door slammed shut. It is suggested that baby walkers allow babies to reach hazards, such as heaters, ashtrays, hot drinks, etc. However, all of these injuries result from lack of supervision. In any case, within a couple of months these babies will be standing and then walking and, without supervision, exactly the same injuries may occur. Most injuries in infants are related to an unsafe environment or inadequate supervision. One aspect that Thompson does not mention is walkers for disabled children. Does he recommend that these be banned? For infants and young children with severe disabilities, a walker is their only means of locomotion.

Donald D Beard

To the Editor: Colorectal cancer is, after skin cancer, the most common cancer in Australia, with 11 245 new cases diagnosed in 1997, and over 4600 deaths.1 In clinical trials, screening programs using faecal occult blood testing (FOBT) have been shown to reduce mortality. The Commonwealth Department of Health and Ageing estimates that implementation of effective FOBT screening programs would save around 400 lives per year.1 However, such screening programs have not been widely implemented because of perceived difficulties with patient acceptance, funding, and the complexity of support structures. General practitioners are in the front line of healthcare, and well placed to institute FOBT screening. Thus, we established an opportunistic screening program whereby patients over the age of 50 years attending surgery are asked by reception staff to complete a short questionnaire while in the waiting room. This questionnaire, developed locally to quickly establish whether a patient has symptoms or a family history of bowel cancer, is given to the GP by the patient during the consultation. If the questionnaire indicates colorectal symptoms, appropriate clinical assessment is undertaken. If a family history of colorectal cancer is elicited, the GP further defines the patient's risk by using the established National Health and Medical Research Council guidelines.2 If there are neither symptoms nor a family history, the patient is offered annual FOBT screening. From 17 June to 30 September 2002, 731 patients under the care of 29 GPs completed the questionnaire. Our findings are summarised in the Box. GP-based opportunistic screening can reach significant numbers of people. Moreover, unlike other strategies (eg, distribution of test kits by pharmacies), review by GPs of patients' questionnaires ensures that cases unsuitable for FOBT screening (such as those with previously undeclared symptoms or family history) are appropriately assessed. Data reported so far on patients who completed general practice questionnaires for eliciting family history or symptoms of bowel cancer FOBT = faecal occult blood testing.

Susan J Harnett · SK Cyril Wong · Gavin W Lackey

To the Editor: To achieve greater dissemination of mental health education to general practitioners, the Adelaide Central and Eastern Division of General Practice developed a program to be taken to GPs, based on individual need and using a medical expert/facilitator. The "GP Meets the Psychiatrist Project" is an initiative of the Division in collaboration with the Eastern Mental Health Service, and supported by the Lundbeck Institute — a Danish foundation with a special interest in psychiatric education and pharmaceuticals (www.luinst.org). The specific objective of the project is to facilitate access by GPs to psychiatrist support, in the form of education and advice. This takes the form of a psychiatrist visiting a practice for a one-hour "open tutorial", on a topic preselected by the GPs in the practice. Lundbeck funded the psychiatrist's time. GPs were not funded, and no Continuing Medical Education points were sought for these sessions. The project began in February 2001, and during the year 75 GPs from 22 practices were involved. Following the tutorial, both the GPs and the psychiatrist completed an evaluation questionnaire that covered issues such as the topics chosen by GPs, discussion of medications, referral for psychiatrist support, the need for further sessions and how they rated the sessions.1 The main topics raised by GPs were depression, medication issues, difficulties with access to psychiatry services, psychosis, and management of acute situations or angry patients. GPs felt the sessions were very useful, and 87% were interested in having meetings with other specialists. From October 2001, 23 GPs rated the usefulness of the tutorial using a Likert scale of 1 ("no use") to 5 ("very useful"). The value of the tutorials was clearly demonstrated by the mean rating of 4.6. Three psychiatrists participated and all found the experience of attending general practices and running the sessions very rewarding. They found their assumptions about the nature of family medicine were often wrong; for example, they were interested to find that practices were often focused towards particular areas of health. The project has demonstrated the usefulness of tailoring education packages to the specific needs of GPs and has shown that this would be a suitable avenue to improve links between GPs and specialists. With specialist support, a similar session involving an endocrinologist discussing diabetes is now being run through the Division, with no funding, which suggests this type of program may be sustainable in the long term.

Greg A Lovell · Phillipa J Hay

To the Editor: Your enthusiastic comments accompanying the publication of a "physicians' charter" prepared by the Medical Professionalism Project1 do not acknowledge strikingly enhanced approaches to medical professionalism in Australia dating from the landmark Doherty Report in 1988.2 One of the most noticeable changes since that time is the emphasis now placed on professional development throughout medical school curricula3 and in the preregistration year.4,5 In addition, our medical colleges are poised to do more in this area in their postgraduate training and continuing professional development programs.6 You fail to point out that the "physicians' charter" contains nothing new, as an examination of the Code of Ethics of the Australian Medical Association will reveal.7 We are probably fortunate that an equivalent process of developing such a charter has not taken place in Australia. The document repeatedly speaks of a "contract with society", but it is an oddly one-sided contract, prepared without consulting members of the communities the authors purport to represent. The charter seems to be a response to frustrations and challenges caused by changes to healthcare systems, especially in the United States and Canada, and carries a tone of living in the past. There are effective means of engaging with our community to ensure that essential aspects of medical professionalism are valued and maintained.8 Many of these are already being used in Australia. I refer to such developments as community membership of medical boards, community input into selection of medical students, establishment of independent health complaints commissions and widespread engagement with the health consumer bodies by most sections of the profession. Additional initiatives that will assist the community to trust and value medical professionalism in Australia include the increased expectation that all doctors will engage in continuing medical education and the establishment by medical boards of pathways to identify and assist poorly performing doctors. Should the organised profession ever develop a similar charter for Australia, it is to be hoped that the authors will ensure that the concerns and needs of our broader community are taken into account and that we do not engage in the self-pity evident in the Medical Professionalism Project.

Kerry J Breen

To the Editor: You recently lent your support to the "physicians' charter" produced by the members of the Medical Professionalism Project.1 According to the document, the charter is part of the process of reforming healthcare systems. If the authors are hoping to use this document to change the way medicine is delivered, they should be prepared for criticism. The simplest criticism is that they present no evidence. Sweeping statements are made, without supporting documentation, about the potential for dishonest behaviour by physicians. Even broader generalisations are made about the appropriate way to run healthcare, again without evidence. The authors state that "physicians should never exploit patients for . . . personal financial gain or other private purpose". While the word "exploit" has a pejorative flavour, it is reasonable for doctors to be paid for what they do. The "sackcloth and ashes" version of medical practice appeals to a small minority. They then move deeper into socialist territory: "Medical professionalism demands that the objective of all healthcare systems be the availability of a uniform and adequate standard of care." But the standard of healthcare can never be uniform, and vigorous attempts to make it so could conceivably undermine its quality. "Political correctness" proceeds apace. "A commitment to equity entails the promotion of public health and preventive medicine . . .". Why, for goodness' sake? Preventive medicine is not always cost effective. The authors then propose their own version of health economics. In their view "the provision of unnecessary services not only exposes patients to avoidable harm and expense but also diminishes the resources available for others". This is not necessarily so. Some services will turn out to be superfluous. There are other areas of extravagance in our society (eg, fast cars, big houses, and all restaurants), but they still make an important contribution to the economy. Growth in healthcare expenditure expands the whole economy, and I am not aware of any evidence that spending in one area of healthcare necessarily deprives others. It could well be the reverse. Politically correct attitudes are widespread in the medical community,2 but should not be imposed on others, as the Editor was perhaps suggesting.

Keith V Woollard

To the Editor: Young doctors and medical students are acculturated into healthcare systems in which governments, corporations, consumer groups and other stakeholders have an increasing presence. These forces have so significantly altered medical practice in the United States, for example, that many doctors have questioned the applicability of traditional patient-centred values. Whether or not Australian doctors have faced such adversity, medical practice is becoming increasingly complex, and the Medical Professionalism Project's charter1 encourages us to consider how we might best serve the future needs of patients, families and the community. To this end, the charter upholds traditional patient-centred values while embracing less familiar civic responsibilities. It challenges us to think outside the doctor–patient dyad about what it means to be a doctor in contemporary Australian society. The inevitable rationing of finite health budgets, inequalities in access to and quality of care, and growing recognition of social determinants of health are all compelling reasons for engagement with social processes. The Australian medical profession has served us well through its health policy leadership. We have no reason for complacency, however, given the appalling state of Indigenous health, the challenges posed by rural and refugee populations and the progressive infiltration of for-profit interests into the Australian healthcare system. Upholding the principle of social justice will require skills not often called upon in medical practice. If we are to remain respected public advocates we will need heightened awareness of population health issues and the ability to effectively collaborate with other players. We will need ways of translating time-honoured virtues such as altruism and compassion into public arenas. Partnerships may open new opportunities in a publicly responsive and outward-looking profession, while retaining the inherent worth of patient care. By its social orientation, the charter does more than exalt old values in the face of modern healthcare challenges. It helps to articulate the aspirations of a more broadly engaged profession. The charter will be valuable to those of us near the start of our careers, for whom professionalism will be as much characterised by its engagement with the future as it is by its links with the past.

Russell L Gruen

In reply: I welcome the comments on the physicians' charter. The purpose in publishing the charter was to promote a dialogue on medical professionalism, as, in contrast to the continuing discussion and debate on professionalism in North America1-3 and the United Kingdom,4 interest in the subject in Australia is virtually absent. Enter the terms "medical professionalism" and "Australia" in PubMed and the search yields eight publications in the Australian literature over the past 30 years. Despite this meagre tally, Breen informs us that all is well, through the efforts of bodies such as the Australian Medical Council (AMC). While the AMC's efforts are commendable, its recommendations for instilling professional attitudes in medical students and interns must compete with other priorities in an already crowded curriculum and with the pressures of the intern year. Under the daily stresses of real life, abstract concepts of professionalism are likely to be given low priority. Breen also suggests that the medical colleges are "poised to do more in this area". This may be so, but a search of college websites reveals that most are silent on matters of professionalism. Finally, Breen laments that the physicians' charter contains nothing new — but no one really expects revolutionary concepts in precepts of professionalism, which have evolved over hundreds of years. Besides, what harm is there in being exposed to the views of our colleagues in the northern hemisphere? As the comments of Gruen suggest, the charter is one of the more broad-ranging contemporary expositions of medical professionalism. Woollard wants evidence for the principles of the charter and its purpose. He also suggests that it is tinged with socialism and "political correctness". But the medical ethics that underpin our professionalism are dependent not on evidence but on humanistic principles. He hints that making the objective of all healthcare systems "the availability of a uniform and adequate standard of care" smells of socialism. Perhaps so, but I prefer to think of it as embodying the Australian ethos of a "fair go". Finally, "political correctness" is a subjective rather than an objective concept. I thank my colleagues for their insightful comments and for starting the debate on new professionalism. Long may it continue.

Martin B Van Der Weyden

To the Editor: Optimising the management of cancer patients requires objective decisions about "best treatment" strategies, based on high quality data collected systematically from all treated patients (or at least a representative sample of them). Relating treatment and stage at diagnosis to individual outcome can allow monitoring of whether treatment is consistent with best practice, and can provide a systematic foundation for evidence-based care. Clinical cancer data collection also allows treatment services to be evaluated, as institutions can monitor throughput and endpoints. However, institution-based data collections may not be representative of all cancer patients, and aggregation of data from several institutions is needed to obtain a comprehensive picture. Population-based cancer registries, which operate in all Australian States and Territories, include data on the site and morphology of cancers. Notification of cases to the registries is mandatory for hospitals and pathology providers, and survival of patients is assessed by linkage to mortality data. The registers do not routinely record stage or treatment data. Until recently, there have been no nationally agreed data items or standard data definitions to facilitate the collation of clinical cancer data across institutions. In 1999, the National Cancer Control Initiative (NCCI) commissioned a nationwide consultation process to seek expert advice on developing a core clinical cancer data set. Representatives from the State and Territory population-based cancer registries, the Australian Institute of Health and Welfare and many large cancer treatment centres were consulted.1 A workshop was held in Melbourne in July 2000 to identify key items for inclusion in the data set, and a group was established to work on data definitions. These are now available on the NCCI's website (<http://www.ncci.org.au/projects/data/dat01.htm>). The data set is designed to be compatible with, and expand on, data currently collected by State cancer registries. Definitions are consistent with the New South Wales clinical cancer data set,2 and we acknowledge the input from this source. Items would be collected by treatment centres. Some institutions would need to standardise information already collected for ongoing patient management, while others would need to establish and maintain new collections. The Faculty of Radiation Oncology of the Royal Australian and New Zealand College of Radiologists has recommended incorporation of the NCCI data set into its proposed quality assurance program. Collation of data across institutions requires careful attention to patient identification issues in order to protect privacy and avoid duplication of data from multiple sources. Use of the data set by clinicians and health planners and evaluators at a national level is the ultimate aim. This would require funding and commitment, and attention to issues of privacy, confidentiality, and data ownership. At present, adoption of the data set on a voluntary basis by treatment centres is the best way forward.

Margaret P Staples · J Mark Elwood · Alan S Coates · Lizbeth M Kenny

To the Editor: We applaud the efforts of Wasiak et al to apply the principles of evidence-based medicine to answer clinical questions.1 However, it is important to understand the historical context of clinical trials reported in the literature, and, when necessary (eg, when conducting a meta-analysis or when the results of trials appear to conflict), to seek additional information from the authors. One of us (H K G) designed the randomised-controlled trial (RCT) reported by Corry et al.2 It was a pilot study and not a definitive clinical trial. The primary outcome measure was resonant frequency, an objective measure of muscle stiffness. This trial was conducted before the introduction of validated outcome measures for assessing upper limb function in children with cerebral palsy, and it was not possible to perform any sample size calculation for functional outcomes. At 12 weeks in the group receiving injections of botulinum toxin A, there was a significant difference in grasp and release but not in the ability to pick up coins. It is not surprising therefore that this study found significant decreases in muscle stiffness, but the functional results were inconclusive. The other RCT identified by Wasiak et al also involved one of us (D F).3 It was designed specifically to investigate functional outcomes, a sample size calculation was performed from pilot work, and a specific functional outcome measure (QUEST) was used. This study reported significant functional improvements after the use of botulinum toxin combined with occupational therapy. These two studies, when understood in their historical sequence, should therefore be considered complementary and not contradictory. It is important to assess the quality of randomised clinical trials as well as their conclusions (eg, using the Physiotherapy Evidence database PEDRO scale <http://ptwww.fhs.usyd.edu.au/pedro>).4,5 The smaller study by Corry et al2 had insufficient power and inadequate methodology to investigate functional outcomes. On the other hand, the conclusions of the study by Fehlings et al3 should be taken as the current level of evidence. We therefore submit that the conclusion drawn by Wasiak et al is incorrect. We support further research to evaluate and strengthen the evidence relating to botulinum toxin A and upper-extremity function.6

H Kerr Graham · Roslyn N Boyd · Darcy Fehlings

In reply: We thank Graham et al for their response to our article.1 It is important to understand that the clinician who posed the question regarding botulinum toxin A injection wished to find the "best available medical evidence". We were not asked to take account of the historical context of previously published articles, nor were we asked to exclude specific types of RCTs. If we were to exclude specific RCTs based on the preference of an author, then the strong methodological principles that surround the evidence-based practice movement would be open to extreme forms of bias. We also disagree that our conclusions were incorrect. Our reading of the article by Corry et al2 differed from that of Graham et al. We do not consider that their study showed that botulinum toxin injection significantly improved the function of the hemiplegic upper limb. Together with the results of the study by Fehlings et al,3 indicating a significant improvement in weight-bearing at four weeks (part of the QUEST assessment), our conclusion — that we could not support or refute the efficacy of botulinum toxin injections for improving upper-limb function in cerebral palsy because of differing opinions — remains unchanged.

Jason Wasiak · Brian J Hoare

To the Editor: I read with interest and alarm your comments on generalists and gerontology in the 16 September 2002 issue of the Journal.1 I believe the role of the general internal medicine physician in managing patients with complex multisystem disease is still important and will expand in the future. General physicians remain passionate about general medicine.2 General medicine has been threatened over the past 20 years by the emergence of the medical subspecialties. However, there is increasing recognition of the need for general medicine specialists, especially in North America and Europe. In Australia, general medical units are being established in the major teaching hospitals in capital cities. Sydney remains a unique exception — all of its general medical units were closed prior to the 2000 Olympics (for reasons that are unclear), and only one has subsequently been re-established (at Royal North Shore Hospital). The Internal Medicine Society of Australia and New Zealand (IMSANZ) has been active in promoting the role of the general physician. IMSANZ provides support for the professional profile and culture of general physicians throughout Australasia. We would welcome enquiries from trainees and physicians (<http://www.racp.edu.au/imsanz>). The Royal Australasian College of Physicians has recognised the necessity for a strong general physician workforce and will be holding a General Medicine Forum in March 2003 to examine the key issues confronting general medicine in Australia and New Zealand. By 2010 the majority of "baby boomers" will be in their sixties, or older, and will require the services of well trained general physicians to provide integrated, cost-effective, whole-of-patient specialist healthcare.

Leslie E Bolitho

To the Editor: Survey data are difficult to publish when response rates fall below 60%. Reminder letters are the most reliable method of improving response rates in postal surveys,1,2 and three to four reminders are needed to achieve the 60% benchmark. The additional benefit of sending five or more reminders appears negligible, suggesting that four is the optimum number to maximise response rates using this approach.1 Our recent experiences suggest that human research ethics committees (HRECs) are seeking to limit the number of reminder letters on the grounds that they could be construed as harassment of research subjects. We recently submitted applications to two HRECs, seeking permission to conduct a multicentre postal survey in our local area. One committee approved the use of a single reminder letter, while the second approved the use of two. We appealed both decisions, citing evidence from earlier studies.1,2 The first committee reversed its decision and subsequently approved use of three reminders; the second committee upheld its decision to allow only two reminders. Thus, our study was effectively restricted to two reminder letters. These experiences highlight two limitations of the ethics review process in Australia. First, the process is highly fragmented and lacks standardisation. HRECs exist as autonomous and independent entities, with varying interpretations of ethical practice. The decisions of one committee frequently undermine those of another. This is frustrating for researchers, who often have to seek permission from two or more committees during the course of a study. Second, the process confers higher status on the rights of study participants than on the methodological demands of science. Whenever there is any actual or perceived conflict between the two, HRECs consistently rule in favour of individual rights. Researchers are reluctant to challenge the decisions of their local committee, as there is no independent review process, and appeals are usually referred back to the original committee. The result is a directive and autocratic process which fails to consider the demands of scientifically valid research. Under current guidelines of the National Health and Medical Research Council, HRECs may endorse procedures that are potentially invasive or intrusive, provided certain conditions are met: the procedures are scientifically justified; there is no acceptable alternative; and reasonable steps are taken to protect individual rights. This certainly provides scope for sanctioning four reminder letters in postal surveys.

Stuart C Howell · Susan Quine · Nicholas J Talley

Comment: Howell and his coauthors invite human research ethics committees (HRECs) to standardise their judgements about key aspects of methods, such as response-aiding strategies.1 Their balanced and thoughtful analysis of their experience in securing approval from two HRECs for reminders to enhance response rates to a postal community survey adds to previous concerns about decision-making by HRECs in Australia.2,3 In an era of evidence-based healthcare, there are two main reasons to insist that HRECs only approve protocols for surveys that propose scientifically based procedures to increase response rates. First, applicants are applying empirical insights from previous research in their own practice. Hence, methods are evidence based. Second, applicants are doing their best to ensure the validity of their future data. As eloquently quantified elsewhere, surveys with low response rates are plagued by response bias.4 Indeed, it was recently asserted that, for mailed surveys, "you need an 80–85% response rate to make it epidemiologically significant".5 Fortunately, there is the most rigorous evidence (Level 1) for specific response-aiding strategies for surveys of medical practitioners in Australia.6 Howell and colleagues cite two recent studies of response-aiding strategies in lay surveys. Yet such compelling evidence appears to have been inadequate to secure identical responses from at least two HRECs. That both required Howell and colleagues to apply "homoeopathic" measures to their reminders (sending two rather than the proven four reminder letters) suggests that these HRECs may have been poorly apprised of the relevant scientific literature, unconvinced of its generalisability, or concerned about the acceptability of proposed procedures to research participants. As current National Health and Medical Research Council (NHMRC) guidelines provide a framework for approving research procedures that are "scientifically justified", it seems HRECs would benefit from regular and independent updates about key methodological advances, such as response-aiding strategies, their benefits, risks and harms. I am mindful that such methodological guidance must not add unnecessarily to the copious reading that is already typically demanded of members of HRECs. Perhaps the NHMRC Australian Health Ethics Committee could consider this issue in its next triennium, to commence 2003.

Jeanette E Ward

To the Editor: The recent information that methylphenobarbitone (60 mg tablets) will be unavailable after 1 January 2003 has caused anxiety in patients with epilepsy previously treated satisfactorily with this drug. The suggested substitution of phenobarbitone, primidone or a newer antiepileptic agent seems appropriate. However, patients and perhaps practitioners may assume, as my patients have, that phenobarbitone and primidone are equivalent to methylphenobarbitone on a milligram-for-milligram or a tablet-for-tablet basis. This may not be so. The equivalence is close to 30 mg of phenobarbitone for 60 mg methylphenobarbitone, and probably around 200 mg of primidone for 60 mg of methylphenobarbitone.1 Plasma phenobarbitone concentrations should be checked before and after any changeover. In recent years, several other old, but therapeutically satisfactory, neurological drugs have also been withdrawn from the Australian market (oral neostigmine, several anticholinergic antiparkinsonian agents, ethosuximide and some phenytoin preparations, and the only ergotamine preparation not also containing caffeine). Subject to safety issues, ethics committees usually will not approve a clinical trial of a new drug unless patients who benefit from it are guaranteed supplies until the drug is marketed. Surely similar considerations should apply for patients who have had completely satisfactory long-term responses to marketed drugs. If such drugs must be withdrawn, except for safety reasons, there should be extensive prior consultation with prescribers and patient groups, prescribers should know the situation before their patients discover it from other sources, and there should be a sufficient lead time for everyone receiving the drug to return for another prescription (and for advice) before the drug becomes unavailable (a minimum lead time of six months in the case of drugs subsidised under the Pharmaceutical Benefits Scheme). The withdrawal of useful neurological drugs in Australia has reached the stage where therapeutic options are becoming limited. In the case of drugs required for long-term use, to protect the interests of new patients it has become necessary to consider whether the drug will continue to be available for the expected duration of the patient's therapy. In this regard, the prescriber's only guide may be the track record of the firm which markets the drug otherwise chosen.

Mervyn J Eadie

To the Editor: The recent article by McLean and Higginbotham1 and the accompanying editorial by Melding2 highlight the problems faced by elderly people in aged-care facilities. It is likely that many elderly people living alone in the community are suffering equal, if not worse, pain. At a recent strategic planning meeting, the Australian Pain Society identified this group of people as a high priority for the development of pain management treatment strategies. These strategies are now well into the development process. While it is appropriate for the Journal to focus on medical practitioners' care of these patients, it must be remembered that most direct care for people in aged-care facilities is delivered by nurses and nurse assistants/carers. The Australian Pain Society will be focusing its strategies on non-drug techniques that can be used by this group of healthcare workers. Assessment and documentation of pain-related behaviour, particularly in people with cognitive impairment, is critical if progress is to be made. It is also important to appreciate the contribution in this area from other allied health professionals, such as physiotherapists, psychologists and occupational therapists. These practitioners have much to offer this patient group and have been important contributors to the Australian Pain Society's management strategies. It is hoped that State and federal funding can be made available for nurse educators to deliver these low-tech, non-drug management strategies within aged-care facilities. The two articles quite rightly focus on the regular use of simple oral analgesics, such as paracetamol, and low-dose opioids. Oral analgesics, together with more widespread use of non-drug treatments (eg, exercise, transcutaneous electrical nerve stimulation, hot and cold topical applications, relaxation, distraction, mental stimulation, lifestyle modification) and increased awareness among aged-care workers of the problems and solutions, should lead to an enhanced quality of life for this growing sector of our community.

C Roger Goucke

To the Editor: The articles by Melding1 and McClean and Higginbotham2 highlight the important problem of chronic pain in residential care. We have conducted two studies to investigate factors related to depression in residential care. In our first study, in 1994, we approached a random one-in-two sample of the non-nursing-home population of a Sydney retirement village (n = 1466). We excluded residents who were under 65 years, had severe dementia, were away from the village at the time of the survey, or were too deaf or ill to participate. Of 610 eligible residents, 513 participated (response rate, 84%). Of these, 42.1% lived in hostels and 57.9% in independent living units. In a second, similar study, in 2000–2001, we surveyed residents of three Sydney aged-care hostels (n = 205). Of 159 eligible residents, 148 (93%) participated. In both studies, residents were asked how often over the previous six months they had experienced recurring pain and asked to rate the severity of pain at its worst (see Box). Using the Geriatric Depression Scale (GDS),3 Study 1 found that residents reporting frequent/constant pain were significantly more likely to be depressed (ie, to have a GDS score ≥ 11) than people reporting rare/occasional pain; similarly, people who felt rare/occasional pain were more likely to be depressed than those with no pain (odds ratio, 1.44; 95% CI, 1.13–1.83). In Study 2, there was a non-significant association between frequent pain and depression (odds ratio, 1.47; 95% CI, 0.96–2.24). We implemented pain management programs at each facility. In Study 1, the program was part of a multifaceted intervention for depression,4 but residents could attend whether depressed or not. Based on general practitioner referral, the program provided interdisciplinary assessments by a visiting rehabilitation specialist together with a physiotherapist, occupational therapist and registered nurse from the facility. Consultative psychiatric input was also available. Neuropathic and musculoskeletal pain were the most common reasons for referral. Recommended interventions included drug treatment, exercise and preventive measures. Our impression was that they were well received by residents and GPs. In Study 2, residents with chronic pain were referred to a physiotherapist specialising in pain management and reported that this was beneficial. A clinical psychologist also offered to assist, but residents were reluctant to accept this form of help. Our impression was that psychological assistance would have been better received as part of an interdisciplinary pain management program. Older people in residential care may find it difficult to travel to hospital-based pain management programs. Our experience indicates that it is feasible to conduct pain management programs in residential care. However, improving pain management is not only a matter of pharmacological interventions. If we are serious about achieving adequate standards of pain management in residential-care facilities in Australia, resources should also be devoted to providing accessible interdisciplinary pain management programs and to changing the attitude that pain is an inevitable part of old age. Self-reported pain frequency and severity among residents of aged-care facilities Study 1 (1994) (n = 513) Study 2 (2000–2001) (n = 148) Pain frequency Not at all 230 (44.8%) 54 (36.5%) Rarely/occasionally 115 (22.4%) 52 (35.1%) Frequently/constantly 168 (32.8%) 42 (28.4%) Pain severity* Minimal/mild 58 (20.5%) 27 (28.7%) Moderate 98 (34.6%) 36 (38.3%) Severe/bad as could be 127 (44.9%) 31 (33.0%) *Severity rated only for residents experiencing pain.

Robert H Llewellyn-Jones · Karen A Baikie · Heather E Smithers · Philip D Funnell

To the Editor: In their editorial,1 Hammett and Harris have overlooked one of the most important contributory factors to the increased use of diagnostic tests by community-based practitioners (ie, GPs) — patient demand. It is not unusual for a GP to be faced with a request by a patient to be "tested for everything", or for a specific test that may be quite inappropriate ("I just want my hormones checked"). It takes far longer to explain to the patient that the tests are inappropriate than to give in and sign the appropriate pathology form. And then, if the patient a year later does come down with some obscure syndrome, he or she can come back in the courts and say, "If only the doctor had listened to my request for tests I would be okay now". The frontline GP is in a lose–lose situation, stuck between the Health Insurance Commission and its demands for reasonable levels of testing, the expectations of patients that everything can be detected by a blood test, and the excessively perfectionist ideals of the legal system.

Warwick Carter

In reply: As Murray notes, appropriate test ordering belongs firmly in the domain of quality care and clinical accountability. In his seminal article on clinical leadership of healthcare system improvement, Berwick lists appropriate use of testing and therapy as the first challenge facing people who wish to improve healthcare systems in the developed world.1 Murray also identifies the paucity of structured educational programs aimed at providing junior medical staff with the skills to exercise "knowledge and judgement" in test ordering. The Royal Australasian College of Pathologists is seeking to address this deficiency through the development of educational modules on test ordering. It is unlikely, however, that education alone will curb the increase in test ordering in Australia. Educational programs for junior medical staff are notoriously resource-intensive and difficult to sustain. In addition, they rarely provide the point-of-care guidance that seems to be more effective in sustainably modifying behaviour. Such guidance may require test-ordering software that provides guidelines for ordering and feedback of individual performance, or the use of structured test-stratification programs such as that described by Stuart et al.2 Improved education, supervision and point-of-care guidance will prove ineffective if fears of litigation continue to drive the behaviour of clinicians. As Carter points out, concerns about litigation must be considered in any program aimed at improving practice. However, although litigation related to missed diagnosis is a recurring theme, this may relate more to time pressures rather than a failure to perform investigations. Indeed, many malpractice suits result from failure to adequately check and act upon the results of the barrage of tests ordered. Attempts to reduce medicolegal risk by ordering all conceivable tests may increase practitioners' risk unless they have extremely well-designed follow-up systems. It is important that clinicians not sacrifice high-quality, evidence-based investigation and treatment in an attempt to minimise perceived litigation risks. By testing inappropriately, clinicians may in fact expose themselves to greater risks of litigation, as their patients are exposed to the risks of the tests themselves, the chance of false-positive results and inappropriate treatment, and the failure to follow up on investigation results. It is unfortunate, and an indictment of our current reimbursement system, that the financial realities of community practice make it difficult for clinicians to take sufficient time to communicate with patients about the appropriateness of an investigation or treatment. If we continue to allow this to become the way we practise, we will continue to see a diminution of our professional role as we become merely booking agents for tests. As Berwick says, "Efforts to reform the health system from the outside can help motivate and set the stage for improvement. Yet, if clinicians do not wish to make specific changes in their own work to better meet society's need for better outcomes and lower cost, no-one outside the health system can be clever enough or powerful enough to make them do it."1

Rohan J H Hammett · Roger D Harris

To the Editor: While agreeing that sensible utilisation of pathology tests in emergency departments (EDs) is important, we are concerned that the article by Stuart et al1 might be misinterpreted to justify wholesale reductions in important diagnostic microbiological tests, particularly blood cultures. Stuart and colleagues imply they could safely reduce the number of blood cultures by 80%.1 Other local data have suggested a minority of blood cultures in the ED influence patient management.2 Confirmation of aetiology will be denied for patients by "rationalisation" of blood cultures in EDs. Although most pathogens are susceptible to broad-spectrum anti-microbial agents, widespread empiric prescribing of such agents in an era of increasing antimicrobial resistance is unwise. A recent Australian study evaluating blood cultures found that a third of patients with positive blood culture results were not clinically suspected to be bacteraemic.3 Furthermore, the Journal recently reported the emergence of community-acquired methicillin-resistant Staphylococcus aureus (MRSA),4 and increasing resistance in Streptococcus pneumoniae.5 Missing MRSA or multidrug-resistant pneumococcal bacteraemia will result in adverse patient outcomes. What about missed cases of meningococcal disease, or typhoid fever, with their associated public health costs? Paradoxically, amid emerging antimicrobial resistance, we may become less aware of the problem. Furthermore, what about the infection control costs required to control the resultant outbreaks of multidrug-resistant organisms? Empiric broad-spectrum antibiotic prescribing, driven by failure to undertake important microbiological investigations, is bad medicine: It teaches everyone to guess the microbiological diagnosis, and, if you do not test, who can prove you wrong? Perhaps only when the patient presents to the tertiary referral hospital with therapeutic failure and evolving multisystem organ failure. It logically extrapolates to all patients getting vancomycin plus meropenem to ensure covering MRSA and resistant gram-negative bacilli. It inevitably drives resistance, which is increasing rapidly. It has never been subject to rigorous scientific scrutiny with cost-effectiveness studies. Moreover, the study by Stuart et al1 provides no data on readmission rates, lengths of stay, adverse events and rates of missed or incorrect diagnoses; the ED setting studied has limited generalisability; and United States guidelines, which may be inappropriate in the Australian healthcare context, were used to develop the diagnostic testing protocol. Might not reducing the ordering of some microbiological tests cause "spiralling therapeutic empiricism"? Might not the overall healthcare budget growth accelerate because of increased prescribing of expensive broad-spectrum antimicrobials?

Iain B Gosbell · Peter J Collignon · John D Turnidge · Christopher H Heath · Joan L Faoagali