Volume 187 Issue 5

Significant similarities between the challenges of lung transplantation in patients of all ages should lead to better access to this life-saving surgery for children and adolescents Lung transplantation (LTx) is firmly established as a therapy for end-stage lung and pulmonary vascular diseases in patients aged over 18 years and into the seventh decade of life.1,2 However, for those under the age of 18, be they child or adolescent, the role of LTx is less clear.3,4 In Australia, this has contributed to a perception that the risk of undertaking LTx in children and adolescents does not warrant the reward. Indeed, presently in this country, there is no major paediatric hospital offering a lung transplant program, likely recognising the complexity of treating such patients coupled with the potential risk of achieving poor results with a low case load — the reality is that the projected case numbers will only be of the order of four to eight per year across Australia and New Zealand. Thus, by focusing on successful LTx outcomes for an adolescent population, the article by Morton and colleagues in this issue of the Journal5 highlights a number of the key issues regarding the efficacy and utility of LTx for younger Australians (→ Successful lung transplantation for adolescents at a hospital for adults). Although adolescence refers to a transitional state from childhood to adulthood, patients 15 years and younger are generally excluded from adult hospitals and those 18 years and above excluded from paediatric hospitals. Two-thirds of the patients in the study by Morton et al could have been “routinely” treated in adult hospitals. Notwithstanding this limitation, the report gives important insight into the issues, experience and successful outcomes that can be achieved in younger lung transplant recipients. From this article, it is apparent that in Australia, a well developed, large adult LTx unit is able to use its highly specialised services to overcome some of the problems and deficiencies that can limit a stand-alone service for such a small population as children and adolescents requiring LTx. However, the age of any potential Australian lung transplant recipient is critically important — at this time, this technology is not being routinely offered to younger children. Indeed, at present, Australia’s youngest ever lung transplant recipient was aged 9 years at the time of LTx.6 The improved outcomes for LTx now described in adolescents5 should provide an impetus to provide access for younger potential LTx recipients. In looking to achieve this advance, we need to keep in mind that the transplant recipient’s age can matter in several different ways. Fortunately, severe lung disease warranting consideration of LTx in children and adolescents is relatively rare, although interestingly, it does have a bimodal distribution. The International Society for Heart and Lung Transplant (ISHLT) Registry 2005 paediatric report notes about 65 procedures performed worldwide each year.7 In older paediatric patients, typically over 12 years of age, about 70% will have cystic fibrosis as the primary indication for LTx, whereas in infants aged less than 3 years, the indication in about 60% is congenital heart disease or pulmonary hypertension. Despite the perception that transplant recipients fare worse if they are younger, the recent ISHLT Registry reports a half-life of around 5 years after LTx, and no significant survival difference between adults, adolescents and the very young.7 Rates of early graft dysfunction and late graft dysfunction (ie, bronchiolitis obliterans syndrome [BOS]) are also similar. However, causes of death are quite different, with adults and adolescents dying from respiratory failure related to BOS, and younger children dying from infection. The functional status of survivors is excellent, with over 80% reporting no activity limitations at 5 years,7 although morbidity related to the obligatory immunosuppressant drugs is very common across all age groups. Further, there are some specific issues (medical, psychosocial and legal) associated with LTx in adolescents and children compared with adults. Post-transplant lymphoproliferative disorders, growth retardation, respiratory tract infections and medical non-adherence appear much more commonly in children.8 As discussed by Morton and colleagues, facilitating compliance with therapies and medication are particularly challenging areas when working with adolescents.5 As an example, immunosuppressive protocols need to reflect potential concerns about physical appearance. Also, a particular “at risk” period arises when paediatric LTx recipients transition from paediatric to adult care.9 Performing major surgery with substantial short-term and long-term mortality risks in a patient unable to give consent presents ethical and legal dilemmas. For paediatric patients with severe lung disease, recent technological advances provide the potential to build on the excellent results of LTx in adolescents presented by Morton et al.5 Minimal waiting list mortality is a critical component of any assessment of the efficacy and utility of organ transplantation. Thus, the management of severe lung disease by experienced teams, with appropriate use of newer therapies such as bi-level positive airway pressure (BiPAP), dornase alfa and azithromycin in patients with cystic fibrosis, may lead to a successful “bridge to transplant”. Similarly, intravenous epoprostenol, oral bosentan and sildenafil may provide a bridge to transplant for patients of all ages with severe pulmonary hypertension. The study by Morton et al included several terminally ill individuals transplanted after support with mechanical ventilation or extra-corporeal membrane oxygenation.5 Morton and colleagues are to be commended for their successful endeavour, but we contend that further detailed discussion about excessive early mortality10 and resource use is needed before bridging in this fashion is routine in any age group. Such bridging has become increasingly used in the United States (11% of all LTx in 200611) and we believe that many, including ourselves, would argue that Australia does not have the intensive care facilities and staff to routinely bridge in this manner. There are also other developments that should increase transplant opportunities and access to LTx for children and adolescents, hopefully shortening waiting times, thereby further decreasing waiting list mortality, and potentially allowing at least the possibility of retransplantation in the event of late graft dysfunction. One possibility is that large-volume LTx transplant centres (typically not small-volume paediatric-only centres, as yet) might increase organ availability by using extended donor lungs (eg, where there are secretions or an abnormal chest x-ray, etc),12 or cadaveric or living-related lobar transplants (eg, so-called “cut-down lungs”).13 The use of cut-down lungs typically involves transplanting one lobe from each of two adults to make a bilobar transplant for a child or smaller adolescent. Although this resource-intensive and challenging operation is possible, some question the philosophy of undertaking the only known procedure to have a “potential 300% mortality”.13 Donation-after-cardiac-death (DCD) retrieval of lungs for transplantation (as distinct from the usual donation-after-brain-death retrieval) is also now a viable prospect being used to acquire adult lungs for LTx,14 and will soon be extended to paediatric DCD lung donation.15 Thus, evidently, expanding the complexity and extent of LTx offered to children and adolescents might consume significant resources, so LTx results must be carefully considered and evaluated to ensure continued successful outcomes. In this regard, we note with great interest the recent institution of a complex mathematical lung allocation score model by the American United Network for Organ Sharing (UNOS).16 This model uses disease-relevant clinical and physiological variables to predict who will get the most significant improvement in survival with LTx and, therefore, who should be preferentially transplanted. Although historically based, the model will evolve with ongoing clinical experience and should be able to provide new evidence to guide future practice. Interestingly, because of differences in diagnostic categories and post-LTx outcomes in younger lung transplant recipients, the UNOS lung allocation score is only to be applied to those aged over 12 years.16 So, although there are important differences to consider when evaluating the efficacy and utility of LTx across the wide age-spectrum of disease and physiology in the very young, adolescents and adults with terminal lung disease, there is also significant overlap. Medical and allied health experts in paediatric and adolescent medicine have much to offer adult LTx programs venturing into adolescent transplantation; their involvement should be routine. Similarly, units experienced in adult LTx bring knowledge and technology to paediatric and adolescent LTx that can only benefit the small number of critically ill young Australians previously without local access to LTx expertise.

Gregory I Snell MB BS, FRACP, MD · Glen P Westall MB BS, FRACP · Trevor J Williams MB BS, FRACP, MD

Australia is well placed to again lead the world in health technology assessment Australia led the world in 1993 when it introduced the so-called “fourth hurdle” of economic evaluation into the approvals process for drugs (in addition to the usual regulatory “hurdles” of quality, safety, and efficacy).1 We are among the dozen or so developed countries that had invested in health technology assessment (HTA) since the early 1980s, but it was the requirement of a favourable economic evaluation that attracted international attention to HTA in Australia.2 While economic evaluation had always been considered a component of HTA, a policy requiring evidence of cost-effectiveness was groundbreaking.3 In 1998, the federal Minister for Health created a parallel HTA process for new medical services. Evidence of sufficient safety, effectiveness and cost-effectiveness to be included in the Medicare-subsidised benefits package now forms the basis for coverage recommendations to the Minister by the Pharmaceutical Benefits Advisory Committee (PBAC) for drugs, and the Medical Services Advisory Committee (MSAC) for medical services and technologies (Box).5 In contrast to most other countries, HTA in Australia has been woven into the fabric of health services funding, giving it greater impact on the introduction of new treatments. Our approach is similar to that of the United Kingdom’s National Institute for Health and Clinical Excellence6 but differs from Canada’s more “hands off” implementation approach,7 both described in this issue of the Journal ("Health technology assessment in England: assessment and appraisal" and "Health technology assessment in Canada: diversity and evolution", respectively). Most other countries have structured their HTA processes to be “advisory” to doctors and health care services. It is unclear whether this separation of advice from funding is more effective than the direct application of HTA to coverage decisions seen in Australia and the UK, but a common lament from academics and policy-makers in such systems is that HTA findings are not “taken up” by health care providers.8,9 Separating HTA from coverage decision making may lead to less contention with professional groups and the biotechnology industries, but perhaps also reduces the impact of the HTA effort. Because of their direct impact on government coverage decisions, and the still novel requirement for an acceptable incremental cost-effectiveness ratio, both the PBAC and MSAC have been subject to industry and political scrutiny. The most comprehensive inquiry was a 2005 Productivity Commission report on advances in medical technology in Australia.10 The PBAC was a major focus of negotiation leading to the Australia–United States Free Trade Agreement,11 and the MSAC has conducted an internal review and consultation process,12 in part as a response to industry criticism of delays in the assessment process.4 A third article in this issue of the Journal by Petherick and colleagues (→ An evaluation of methods used in health technology assessments produced for the Medical Services Advisory Committee) examines the evolution and shortcomings of systematic reviews in published MSAC assessment reports since 1998.13 Although the Australian system is apparently fragmented (medicines versus services, federal versus states, public versus private systems), differing characteristics of each may justify separate approaches. It is clear that the longer history of drug safety regulation makes pharmaceutical evaluation more straightforward than evaluation of medical services.10,14 Surgical interventions provide their own unique challenges to evaluation methods, and the Australian Government has funded ASERNIP-S (Australian Safety and Efficacy Register of New Interventional Procedures — Surgical) to conduct HTAs under the sponsorship of the Royal Australasian College of Surgeons.15 Funders at each level of the system (federal and state) have differing responsibilities and interests, probably best served by dedicated evaluation efforts, but there has been considerable synergy in the development of HTA among these stakeholders. The Productivity Commission report highlighted a number of recent developments that fill gaps and reduce friction between the needs of different HTA users.10 Through the Australian Health Ministers’ Advisory Committee (AHMAC), the states pool funds to sponsor HealthPACT (Health Policy Advisory Committee on Technology) which performs “horizon scanning” for state health departments,16 and shares secretariat and other functions with MSAC. This mechanism alerts the funders of public hospitals to emerging medical technologies with potential to influence their health care systems. The states together determine HealthPACT’s budget and work program. In addition, AHMAC has delegated to MSAC a role in advising on Nationally Funded Centres (NFC). These are services where the volume of relevant cases is not sufficient to justify more than one or two units in the country — historically, these have been transplant units. The NFC designation ensures that all states contribute to funding of such units, thus guaranteeing access for residents of all states. State health departments are developing their own HTA capabilities. For example, the Victorian Policy Advisory Committee on Clinical Practice and Technology17 was set up in 2004 to undertake a variety of HTA activities, including horizon scanning, assessment and monitoring for the Victorian Department of Human Services. State-based committees commonly consider applications for high-price and/or high-volume drugs, devices and procedures, and create a mechanism to approve funding for novel or statewide specialty services outside normal hospital funding arrangements. Hospitals and regional health services in Queensland, Western Australia, South Australia and Victoria have established internal HTA committees to oversee the introduction of new drugs and medical procedures, with examples from Bayside Health and Southern Health in Victoria cited by the Productivity Commission in its report.10 Public hospitals, with their role in medical education and research, may need to focus on different technologies at different stages of the product development cycle than do private hospitals and health insurers. In contrast to Australia, the HTA efforts of Canada and the UK have a unified approach to drugs and other technologies. The UK National Institute for Health and Clinical Excellence has an advantage over MSAC and PBAC in setting its own agenda, the so-called “needs-led” prioritisation of HTA topics. Canadian HTA organisations seem to balance the needs of the system as a whole against those of particular interests, including those of funders,7 but probably come closer to HealthPACT’s user-led prioritisation. All jurisdictions grapple with the politically charged problem of “disinvestment” — that is, ceasing to support therapies whose effectiveness (and/or cost-effectiveness) cannot be demonstrated. Both the UK and Canada have successfully pioneered “rapid response” methods for HTA users requiring timely answers to tightly framed clinical questions, an approach not yet common in Australia. Clinical evidence for HTA is derived from systematic reviews, and these in turn rely on randomised controlled trials (RCTs). Evidence-based medicine has refined the tools available for evaluating evidence of clinical benefit; basic physiological evidence of efficacy can come from trials in any country. However, evidence of real-world effectiveness is dependent on the medical culture, workforce and referral patterns of a particular health system, and economic evaluation is even more dependent on the organisational forms of health care, including the skills mix and relative wages of different professional groups. Generally economic assessments use decision–analytic models, with key outcomes costed locally to determine the cost-effectiveness of an intervention in each health care system. None of the national HTA processes described has the capacity to commission new clinical research, and there is little articulation with existing medical research priority-setting processes. This often results in rejecting new technologies because there is no RCT evidence of their efficacy or effectiveness, rather than evidence that they are ineffective.4 Both the UK and the US are trialling “coverage with evidence” approaches to funding new medical technologies as a way of bridging current gaps in evidence.18 These allow introduction of new services or biotechnologies on the condition that patients are entered into rigorous clinical trials, and with the understanding that continued funding will depend on the evidence from these trials. The coming of molecular medicine with its individualised and gene-based therapies will exacerbate the lack of clinical evidence from RCTs.19 The “demise of the blockbuster” drug will demand a new research-intensive paradigm for evaluation and regulation of therapies in developed countries.20 Monitoring drug safety21 and funding the collection of randomised evidence22 are possible in Australia, even with current evaluation tools. However, we will need focused effort to develop more “fine-grained” clinical epidemiology techniques to identify patient characteristics that mediate response to treatment and define which subgroups are likely to derive how much benefit from new treatments at what cost. The phrase “rapid learning health system” has recently been coined to characterise the ways in which computerised medical information can be used to inform health care decision making from the bedside to national HTA efforts.23 With Australia’s large health information technology investment, well established disease registries and systematic metadata specifications, we are in a position to pioneer rapid learning strategies that can be used earlier in the evaluation process, and at an acceptable research cost. The challenges then are to manage the inevitable tensions that arise when HTA directly influences funding decisions, to tailor HTA methods to the needs of different stakeholders with differing timelines, to focus HTA strategically to meet national needs including the capacity to disinvest from ineffective treatments, and to supplement RCT efficacy evidence with real-world evidence of effectiveness and cost-effectiveness. Australia could once again claim leadership of international HTA by creating the evaluative and funding mechanisms to rise to these challenges. Requirements for Medicare subsidy of drugs and medical technologies in the Australian health care system Drugs Applicants are required to prepare detailed evidence-based submissions to the Pharmaceutical Benefits Advisory Committee (PBAC), once drug safety has been assessed by the Therapeutic Goods Administration (TGA). Applications are rigorously assessed by health technology assessment (HTA) organisations contracted to PBAC, which then provide confidential reports to PBAC. All documentation for PBAC recommendations is considered “commercial in confidence”, and only brief reports on decisions and deferrals are published. Medical services and technologies The Medical Services Advisory Committee (MSAC) undertakes its own assessments, also by contracted HTA organisations. MSAC reports are published once the Minister has made a determination about listing on the Medical Benefits Schedule. About a third of the work program of MSAC comes as referrals from the federal Department of Health and Ageing.4

Terri J Jackson PhD

Objective: To ascertain the beliefs of young people and their parents about the role of alcohol, tobacco and marijuana in the prevention and treatment of mental disorders.Design, setting and participants: Between May and August 2006, a national computer-assisted telephone survey was conducted on a representative sample of Australian youths aged 12–25 years. 3746 young people and 2005 of their parents were presented with a case vignette portraying psychosis, depression, depression with alcohol misuse, or social phobia in a young person.Main outcome measures: Participants’ beliefs regarding the role of substance use in preventing or dealing with mental disorders in young people.Results: Over 85% of participants agreed that alcohol, tobacco and marijuana were harmful for the young people in the vignettes, and over 80% of youths agreed that not using marijuana or drinking alcohol in excess would reduce the risk of developing a similar problem.Conclusion: Young people and their parents are fully aware of the negative impact of substance use on mental disorders. Translating this knowledge into behavioural change will be a major challenge for future public health campaigns.

Dan I Lubman PhD, FRANZCP, FAChAM · Leanne Hides BBehSc(Hons), PhD(Clin) · Anthony F Jorm PhD, DSc

Objective: To determine the live birth rate following surgical reversal of sterilisation in women aged 40 years and older.Design: Retrospective cohort study of pregnancy outcome following day surgery microsurgical reversal of sterilisation performed by two reproductive microsurgeons in the private sector.Setting and patients: 47 patients (aged 40 years or older) who had reversal of sterilisation performed between 1997 and 2005 in Adelaide, South Australia (n = 35), or the Infertility Centre of St Louis, Missouri, USA (n = 12).Main outcome measures: Independently audited live birth surviving the neonatal period.Results: Of the 47 patients on whom follow-up was obtainable from the two centres, 19 (40%) had a live birth, 7 had had only a first trimester miscarriage at the time of follow-up, and 21 (44%) had failed to conceive. Age at conception ranged between 40 and 47 years. Two women had two live births following surgery. The total direct costs (Australian dollars, adjusted to 2005) in Australia were $4850 per treatment, and $11 317 per live birth. The corresponding direct cost of a single cycle of in-vitro fertilisation (IVF) in Australia has been estimated at $6940, with a cost per live birth of $97 884 for women aged 40–42 years and $182 794 for older women.Conclusion: Previously sterilised women wanting further pregnancy should be offered tubal surgery as an alternative to IVF, as it offers them the opportunity to have an entirely natural pregnancy. In settings where IVF is financially supported by government agencies or insurance, tubal reversal is a highly cost-effective strategy for the previously fertile woman.

Oswald M Petrucco FRCOG, FRACOG, CREI · Sherman J Silber MD · Sarah L Chamberlain · Graham M Warnes PhD, BAgricSci(Hons) · Michael Davies BA(Hons), MPH, PhD

Objective: To evaluate the outcomes 10 years after a flexible sigmoidoscopy colorectal cancer (CRC) screening program in asymptomatic average-risk individuals.Design, setting and patients: In 1995, a program of flexible sigmoidoscopy-based screening of asymptomatic average-risk individuals aged 55–64 years was established at Fremantle Hospital, Western Australia. Insertion depths, pathological findings and subject-rated pain scores have been prospectively recorded. A follow-up flexible sigmoidoscopy examination was offered to attendees 5 years after the initial screening. Post-screening malignancies were determined by linkage with the Western Australian Cancer Registry in September 2006.Main outcome measures: Yield of neoplasia at initial and follow-up sigmoidoscopy, and the incidence of CRC detected after screening.Results: Between 1995 and 2005, 3402 people underwent an initial flexible sigmoidoscopy screening examination (mean age, 60 years; women, 41%) and 1025 had a 5-year recall examination. Mean insertion depth was greater in men than women (60 cm v 52 cm, P < 0.001). The insertion depth in women was more likely to be < 40 cm (17% v 6%, P < 0.001). Mean pain score was 2.9 for men and 4.0 for women (P < 0.001). Fourteen per cent of initial screenings detected at least one adenoma. Over a mean follow-up time of 8 years, invasive CRC was detected by flexible sigmoidoscopy screening in 0.4% of participants; 0.7% of those with a normal result of screening later developed CRC, with 75% of these found proximal to the splenic flexure.Conclusions: Flexible sigmoidoscopy is a viable screening method, with well defined utility and limitations, for CRC screening of asymptomatic people with average risk.

Charlie H Viiala MB BS, FRACP · John K Olynyk MB BS, FRACP, MD

David Jones combined a love of medicine and mathematics with an incredible humanitarian spirit. As the first medical statistician in the New South Wales State Health Department, he made a significant and passionate contribution to the advancement of Aboriginal health in the 1970s, collaborating with Fred Hollows on the National Trachoma Project and with Max Kamien in studying Aboriginal health issues in Bourke, north-western NSW. Compassionate and non-judgemental with people, irrespective of their situation, David was incredulous at needing to redefine the statistical categories of housing levels for Aboriginal people so that he and his colleagues could conduct their work (all were below the lowest standards for white housing). He was also shocked that only one doctor in the divided town of Bourke in the 1970s would treat Aboriginal people. David was born on 2 February 1926 in Adelaide and attended King’s College (now Pembroke School). He graduated in medicine in 1950 and, after doing his residency at Royal Perth Hospital, worked as a general practitioner in Busselton, south-western Western Australia, from 1954 to 1959. He later worked in the Department of Quarantine service at Port Adelaide (1960–1961). In this capacity he visited incoming ships, often in the middle of the night, climbing rope ladders from rocking dinghies. He also worked for the Commonwealth Health Poliomyelitis Services (1961–1962). On completing a Bachelor of Science degree at the University of Adelaide (majoring in mathematics and statistics) in 1963, David entered the new field of medical statistics as a research fellow in the Lung Cancer Registry of the University of Sydney’s Department of Medicine, at a time when links between lung cancer and smoking were being established. In the NSW State Health Department, he also set in place sampling protocols for assessing height and weight of children, and produced a landmark report comparing urban, rural and immigrant children. He obtained a Master of Applied Statistics degree from Macquarie University, Sydney, in 1985. David was an intellectual with remarkably broad interests, ranging from the sciences to music, classical literature and languages. He came from a musical Welsh family and had a great love of classical music, especially early music. In retirement, he learnt to play the organ, sang in a choir and learnt Welsh. He was a staunch life-long churchgoer. David was a reclusive, self-effacing, generous man whose life centred on his family. He expressed a pragmatic humour about ageing and death. He died in his sleep of congestive heart failure while recovering from pneumonia in hospital. His condition remained undiagnosed at the time of his sudden and unexpected death on 18 December 2006 at The Entrance, NSW. He is sadly missed by his wife Ena and daughters Ceridwen, Euronwy and Menna.

Menna Jones

Objective: To describe the results of lung transplantation (LTx) in adolescents at a hospital for adults.Design and setting: Prospective cohort study set in an LTx unit at an adult tertiary referral hospital from 1991 to 2006.Patients: 37 consecutive adolescent lung transplant recipients including 13 males and 24 females (mean age, 16.7 ± 2.0 [SD] years; range 12–19 years) who received heart–lung (six patients) or bilateral LTx (31 patients) for cystic fibrosis (29), congenital heart disease (four), acute respiratory failure (two), or another disorder (two). Two patients were transplanted after invasive ventilation, five after non-invasive ventilation and two after extracorporeal membrane oxygenation.Main outcome measures: Overall survival compared with an adult cohort; survival free of bronchiolitis obliterans syndrome (BOS); overall and BOS-free survival in those transplanted before and after January 2000.Results: Mean waiting time was 273 days (range, 5–964 days; median, 163 days), mean donor age was 28 years (range, 9–53 years). Median inpatient stay was 11 days (range, 7–94 days). Mean follow-up was 1540 ± 1357 days (range, 35–5163 days). The 5-year survival rate for the 16 patients transplanted before January 2000 was 38%, versus 74% for the 21 transplanted since January 2000 (P = 0.05; Mantel–Cox). Overall, 18 of 35 evaluable patients developed BOS. Only BOS was associated with an increased mortality risk (P < 0.01).Conclusion: LTx may be performed successfully in adolescents at a hospital for adults.

Judith M Morton MB BS, FRACP · Monique A Malouf MB BS, FRACP · Marshall L Plit MB, FRACP, PhD · Phillip M Spratt MB BS, FRACS · Allan R Glanville MB BS, FRACP, MD

The Health Technology Assessment (HTA) Programme in England is a government-funded but independent research program. It is “needs-led”, identifying technologies of most importance to the National Health Service and commissioning research to provide answers on these technologies useful to policymakers, clinicians and patients. It is “science-added”, refining problems to researchable questions and working with researchers to ensure that the question is addressed, and disseminating the findings to key audiences. There is a clear distinction in England between assessment (a scientific process and the role of the HTA Programme) and appraisal (the role of policymakers, like the National Institute for Health and Clinical Excellence). There are many features common to HTA in Australia and England, but also differences, as HTA in each country has to adapt to its own environment.

Tom Walley MD, FRCP, FRCPI

Canada has health technology assessment programs at national, provincial and local levels. The programs have been complementary in providing advice to decisionmakers in health care. A national strategy for the management of health technologies is expected to strengthen communication with policy areas.

David M Hailey MSc, PhD

Objective: To examine the methods used in health technology assessments (HTAs) produced for the Medical Services Advisory Committee (MSAC) reviewing the effectiveness of a technology or procedure.Design and setting: Data were extracted from the effectiveness section of HTA application assessment reports published between 1 January 1998 and 17 July 2006 and available on the MSAC website. Only HTAs of effectiveness interventions were examined, as the methods used to undertake such reviews are well established.Main outcome measures: Variables reflecting methods used in the HTAs to evaluate the effectiveness of health technologies or procedures.Results: Of 56 MSAC HTA reports available, 31 met the inclusion criteria. Considerable variability was shown to exist between the various indicators of quality and the methodology used within the HTAs. Reports did not describe potential conflicts of interest of participants. The majority of reports (19/31) did not formally state the research question that the assessment was attempting to answer. Just over half of the reports (18/31) provided details of validity assessment of the included studies.Conclusions: Minimum and consistent standards of methodology and reporting are required in Australian HTAs, using international recommendations of best practice to increase the transparency and applicability of these reports.

Emily S Petherick BSc(PhysEd), MPH · Elmer V Villanueva MD, ScM · Jo Dumville MSc, PhD · Emma J Bryan BSc, PhD · Shyamali Dharmage MD, PhD

In Australia, many deaths and significant cardiac disability result from delayed response to symptoms of heart attack. Although delays due to transport and initiation of reperfusion therapy in hospital may contribute to late treatment, the major component of delay is the time patients take in deciding to seek help. A critical examination of campaigns to shorten patient delay concludes that they were based on a factual, short-term, non-targeted approach that included education and mass media strategies. They achieved equivocal results. One randomised controlled trial has been conducted. Although this showed an improved understanding of heart attack symptoms, it did not shorten pre-hospital delays. The implications of these findings are that future campaigns to shorten patient delay are likely to be more effective if they address the psychosocial and behavioural blocks to action, are ongoing rather than short term, and focus on people at highest risk, including those with known or high risk of coronary heart disease, those in rural locations, and Indigenous Australians. The National Heart Foundation of Australia proposes a comprehensive strategy to incorporate this approach into its future campaigns to reduce patient delay for suspected heart attack.

on behalf of the National Heart Foundation of Australia Chest Pain Every Minute Counts Working Group

Postural syncope is a transient loss of consciousness secondary to a reduction in cerebral blood flow and is typically precipitated by standing. It is the commonest cause of recurrent transient loss of consciousness. Recurrent unexplained postural syncope is most often due to one of the five disorders of circulatory control: vasovagal syncope, postural tachycardia syndrome, chronic autonomic failure, initial orthostatic hypotension, or persistently low supine systolic blood pressure. Failure to identify the underlying cause of postural syncope can result in ongoing morbidity, impaired quality of life and high health care costs. With a detailed history, examination, blood pressure assessment and electrocardiography, most disorders of circulatory control can be diagnosed. In difficult cases, analysis of sympathetic nervous system and circulatory responses during head-up tilting can aid diagnosis. Treatment is challenging and compounded by a lack of evidence. Most patients can be managed in an outpatient setting, and hospital admission or emergency department assessment is rarely warranted.

Gautam Vaddadi MB BS, BMedSci, FRACP · Elisabeth Lambert BSc, PhD · Susan J Corcoran MB BS, FRACP · Murray D Esler MB BS, FRACP, PhD

A 57-year-old man had a colonoscopy which revealed a partially obstructing polyp in the rectum. Biopsy revealed an adenocarcinoma. A computed tomography scan of the abdomen revealed multiple low-attenuation lesions diffusely throughout the liver (Box). Differential considerations included multiple small hepatic cysts, Caroli’s disease, microabscesses and metastases. The patient had a low anterior resection for his rectal tumour and wedge biopsy of the liver. Macroscopically, there were multiple 2–3 mm white nodular lesions throughout the liver, with the macroscopic appearance of diffuse liver metastases. Pathology of the liver biopsy revealed von Meyenberg complexes. Von Meyenberg complexes are benign liver malformations that play an important part in the differential diagnosis of liver lesions. The clinical significance is that they are easily misdiagnosed as multiple liver metastases on imaging, or even on gross examination.1 Because the presence of liver metastasis is crucial to the correct therapeutic approach, ruling out other possible diagnoses, such as von Meyenberg complexes, is of great importance. Von Meyenberg complexes in the liver A: Non-contrast computed tomography (CT) scan of the abdomen showing multiple diffuse, low-attenuation lesions throughout the liver. B: CT scan of the abdomen with contrast showing multiple low-attenuation, non-enhancing lesions throughout the liver.

Bassem M Chehab MD · Shaker R Dakhil MD, FACP

We report a successful kidney transplant (A1 donor to an O recipient), with antibody removal pre- and post-transplant, and pre-transplant administration of anti-CD20 monoclonal antibody (rituximab), intravenous immunoglobulin, and conventional transplant immunosuppression. The transplant, which was performed without splenectomy, is the first such transplant in Australia. At 20 months, the patient’s creatinine level was 110–130 μmol/L, with no evidence of rejection and no complications. ABO-incompatible transplantation should increase “live donor” kidney transplantation, reduce waiting times, and improve patient outcomes. Clinical recordA 24-year-old white man with an antineutrophil cytoplasmic autoantibody (ANCA)-positive crescentic glomerulonephritis remained dialysis-dependent despite several months of immunosuppression. He was placed on the deceased-donor transplant waiting list, and potential live donors were evaluated. His mother (ABO blood group-compatible) was medically unsuitable. His father was blood group A1, while the patient was blood group O. The patient’s anti-A antibody titre, although moderately high at 1 : 256 (measured by conventional tube agglutination testing), was considered potentially amenable to lowering by systematic antibody removal to a preoperative target of between 1 : 8 and 1 : 16. A protocol for an ABO-incompatible transplant was established, reviewed and approved by the institutional ethics committee, and carefully discussed with the patient and his family. A month before surgery, the patient received the anti-B-cell (anti-CD20) monoclonal antibody, rituximab (375 mg/m2). Before infusion, his anti-A antibody titre was 1 : 1024 (the rise in titre was attributed to cessation of cyclophosphamide 6 months earlier). Two weeks before surgery, the antiproliferative immunosuppressant mycophenolate mofetil (MMF; 1000 mg orally, twice a day) was commenced, as was antibody removal using immunoadsorption treatments with Glycosorb ABO Columns (Glycorex Transplantation, Lund, Sweden). The antibody titre reduced to 1 : 64, but rebounded, and antibody removal was continued using plasma exchange. The patient eventually underwent transplantation at a stable antibody titre of 1 : 32 (5 weeks after commencing rituximab, and after 14 antibody removal treatments [five immunoadsorption; nine plasma exchange]). Intravenous immunoglobulin, 0.5 g/kg (12 hours before surgery), and daclizumab, an interleukin-2 receptor blocker (immediately before surgery) were administered. After surgery, tacrolimus, an oral calcineurin inhibitor (target trough blood levels, 8–12 ng/mL), and oral prednisolone (25 mg/24 h) were commenced. Splenectomy was not performed. The transplanted kidney functioned immediately, and the serum creatinine level fell from > 700 μmol/L to 110 μmol/L (reference range [RR], < 110 μmol/L) within 72 hours. Three postoperative antibody removal treatments were performed (one immunoadsorption, two plasma exchange) on postoperative Day 2, 4 and 6. By 1 month, immunosuppression consisted of 15 mg prednisolone, MMF, 750 mg twice daily, and tacrolimus (dosed to trough whole blood levels, 5–8 ng/mL) The serum creatinine level was 110 μmol/L (estimated glomerular filtration rate [eGFR], 75 mL/min [RR > 60 mL/min]). The anti-A antibody titre remained between 1 : 16 and 1 : 32. At Week 6, the creatinine level rose to 140 μmol/L. A transplant biopsy showed no rejection, no recurrent disease and no drug toxicity. At 3 months, the patient resumed work (not having worked for the previous year). At 20 months, the creatinine level ranged between 110 and 130 μmol/L; there was no evidence of rejection (on protocol biopsy), no opportunistic infections, and the patient had had no unscheduled admissions to hospital. Maintenance immunosuppression at this time was MMF 500 mg twice daily, tacrolimus (trough levels, 3–6 ng/mL), and prednisolone 5 mg/24 h. DiscussionRenal transplant recipients have an 80% lower mortality rate compared with those remaining on the transplant waiting list, largely due to the increased cardiovascular mortality rate in dialysis patients.1 In the 20–39-year age group, kidney transplant recipients are estimated to gain, on average, more than 17 years of life.2 The increase in deceased-donor transplant waiting times (which adversely affect the patient and transplant survival) has encouraged more transplantation from living donors (Australian average in 2004, > 37%3). However, around 30% of potential live donors are thwarted by blood group incompatibility, where there is a high risk of immediate, rapid graft loss due to (hyperacute) antibody-mediated rejection. As early as the 1950s, transplantation across the ABO barrier resulted in rapid loss of most kidneys due to hyperacute rejection.4 Sporadic attempts at blood group incompatible transplantation have occurred with limited success, employing plasma exchange for antibody removal. A small uncontrolled series in the mid 1980s reported improved results, and concluded that splenectomy was essential for transplant success.5 Japanese centres (without deceased-donor transplant programs) performed over 400 blood group incompatible kidney transplants between 1989 and 2001, all patients undergoing splenectomy, plasma exchange and intense immunosuppression.6 While the Japanese cohort had inferior early graft survival, the 9-year transplant survival (around 60%) was comparable with that of the concurrent blood group compatible transplant population in Japan (and in Australia) over that period. These results, as well as greater attention to measuring and monitoring of anti-blood group antibody titres, and the concurrent development of diagnostic tools and therapies for antibody-mediated rejection, revived interest in ABO-incompatible transplantation. Small series with excellent results have been reported from the United States and Sweden using MMF-based immunosuppression, pre-transplant antibody removal and, in some cases, anti-T-cell antibody (thymoglobulin) therapy, and splenectomy and/or administration of rituximab.7-10 The vast majority of the 300 ABO-incompatible transplants performed globally over the past 4 years have been performed without splenectomy; the collective 1-year graft survival is over 95% (First International Workshop on ABO-incompatible Kidney Transplantation, Stockholm, March 2007). Ten years ago, under the heading “ABO incompatible renal transplantation: a chance to re-examine?”, Mackie and Tiller reported an inadvertent ABO-incompatible transplant performed in Australia,11 with a fortuitously good outcome attributable to a very low antibody titre (1 : 8), and an A2 donor kidney (A2 is associated with lower antigen expression than A1, but is found in only 20% of the Australian population). The importance of antibody titres as a predictor of risk in ABO-incompatible transplantation has caused some centres to avoid transplanting patients with pre-treatment titres exceeding 1 : 128.8 Gloor and colleagues reported a cohort of 18 patients where the risk of antibody-mediated rejection and graft loss correlated strongly with pre-treatment antibody titres regardless of whether the kidneys were from A1 or A2 donors.9 In this, the first intentional ABO-incompatible kidney transplant undertaken in Australia (performed without splenectomy or anti-T cell antibody), the pretreatment titre was 1 : 1024, and yet no antibody-mediated rejection occurred. An important factor in the avoidance of rejection may have been the “incorporation” of post-transplant antibody removal into the protocol, a practice adopted by most centres currently undertaking ABO-incompatible transplantation.10-13 Our patient received rituximab, but its significance in ABO-incompatible regimens remains unclear.12 Although rituximab effectively eliminates B cells, it does not target the antibody-producing plasma cells (these express negligible amounts of CD20, the target antigen of rituximab). Segev et al have reported successful ABO-incompatible transplantation without splenectomy or rituximab,12 even in the presence of relatively high-titre anti-ABO blood group antibody, and again identified post-transplant plasma exchanges as a key factor. Both plasma exchange and immunoadsorbent columns are effective in removing antibodies. The latter are a safer alternative, as they specifically remove only the relevant anti-blood-group antibodies. Plasma exchange removes all antibodies, and other proteins; this includes clotting factors, which creates difficulties in the perioperative period and if diagnostic renal biopsies are required. Additional problems associated with plasma exchange are reactions to the replacement fluid and exposure to blood products. No complications have been reported to date with the use of the immunoadsorbent columns (but they are expensive). The essential components of protocols for ABO-incompatible transplantation are yet to be determined, and significant questions remain. The use of tacrolimus and MMF is common to almost all centres, while significant numbers of patients have been transplanted without splenectomy or rituximab. What is the highest anti-ABO titre that can be successfully overcome, and the highest titre that is acceptable at the time of transplantation? Regardless of these unknowns, blood group-incompatible transplantation has become an acceptable procedure in selected individuals, and selected centres, and provides a transplant option where sometimes none existed previously. While the 9-year data from Japan is reassuring, longer-term outcomes in patients receiving transplants under present protocols are awaited. An increase in transplantation from live donors by using ABO-incompatible donors should significantly reduce transplant waiting times, increase survival of patients with end-stage kidney disease, and improve the quality of the lives of patients and their families. The direct economic benefit of transplantation compared with maintenance dialysis is estimated to be between $40 000 and $60 000 per patient per year.14 Significant indirect benefits also arise from greater participation in the workforce and reduced reliance on welfare or social services. Safe, successful ABO-incompatible transplantation represents an important advance in the management of end-stage kidney disease in Australia.

Shlomo J Cohney PhD, MRCP, FRACP · Rowan G Walker MB BS, MD, FRACP · Michael N Haeusler BSc, FAIMS · David M Francis MS, MD, FRACS · Chris J Hogan MB BS, FRCPA

Clinical records Patient 1 A 55-year-old man presented with tenesmus, rectal bleeding and discharge of 3 weeks’ duration. He had a past history of treated syphilis and anal warts. High-resolution anoscopy, performed by a sexual health physician, revealed an extensive anterior ulcer distal to the dentate line, suggestive of anal carcinoma (Figure A). However, histological examination of repeated rectal biopsies revealed non-specific ulceration, with chronic inflammation in the adjacent rectal glandular mucosa. Patient 2 A 54-year-old man with previously treated syphilis presented with a 2-week history of per-rectal bleeding, pain and associated fevers. Colonoscopy, performed by a gastroenterologist, revealed extensive rectal ulceration. Histological examination of a biopsy from the ulcer revealed ulceration with mixed acute and chronic inflammatory cells and a lymphoid infiltrate, suggestive of a lymphoma (Figure B). A sigmoidoscopy was performed 7 days later, when the patient re-presented with worsening rectal pain and bleeding. Repeat rectal biopsies confirmed non-specific inflammation with atypical Epstein–Barr virus-associated lymphoid proliferation rather than lymphoma. Patient 3 A 43-year-old man with previous Kaposi’s sarcoma presented with a 2-month history of per-rectal bleeding and diarrhoea. Colonoscopy, performed by a surgeon, showed multiple rectal ulcers suggestive of Crohn’s disease (Figure C), but biopsies showed non-specific acute and chronic inflammatory infiltrates only. All three patients were HIV-infected men who have sex with men. They were receiving highly active antiretroviral therapy, with good virological control and CD4+ T cell counts. All were investigated for a broad range of diagnoses, including inflammatory bowel disease, colorectal malignancy, lymphoma and sexually transmitted infections; all underwent one or more invasive procedures by various specialists before being referred to our infectious diseases outpatient clinic between 2005 and 2006. After assessment at the clinic, the aetiology of the anorectal condition was determined in all cases by a simple rectal swab. The swab was initially tested for Chlamydia trachomatis using a routine nucleic acid amplification method. As these tests were initially positive, the samples were further analysed by sequencing of the outer membrane protein gene. This confirmed, in all three cases, that the strains were of the lymphogranuloma venereum (LGV) 2b serovar, which is commonly associated with LGV. Screening for other sexually transmitted infections was negative. All patients made a full clinical recovery after treatment with doxycycline 100 mg twice a day for 3 weeks. Patient 2 underwent a repeat colonoscopy 2 months later, which revealed complete resolution of the ulcer. Chlamydia trachomatis is a human pathogen and a common cause of sexually transmitted infections, including lymphogranuloma venereum (LGV).1 LGV was previously confined to endemic areas in tropical regions — principally Africa, India and northern South America. However, since 2003, LGV has emerged as an increasingly important infection worldwide, with outbreaks occurring in communities of men who have sex with men (MSM) in The Netherlands, Belgium, France, Germany, Sweden, the United Kingdom and North America.2-6 Risk factors identified in these outbreaks include HIV seropositivity, previously diagnosed sexually transmitted infections, concurrent ulcerative disease, and unprotected receptive anal sex with casual partners. In Australia, there have only been two previous reported cases of LGV. Both patients were MSM. One patient presented with inguinal lymphadenopathy acquired in Melbourne,7 while the other had anorectal LGV acquired after sexual exposure in Europe.8 Unlike other chlamydial infections, which are generally restricted to epithelial surfaces, LGV is invasive and causes severe inflammation, often with systemic symptoms and with a preference for lymphatic tissue.1 The manifestations of LGV infection vary depending on the site of inoculation, presenting either as a painful unilateral inguinal syndrome or an anorectal syndrome. Lessons from practice Lymphogranuloma venereum (LGV) is an invasive inflammatory disease of the urogenital tract caused by infection with Chlamydia trachomatis. LGV is an important cause of anorectal disease in men who have sex with men. Anorectal LGV may masquerade as inflammatory bowel disease, colorectal malignancy, lymphoma or other ulcerative rectal sexually transmitted infections. Diagnosis requires a high index of suspicion. It is important to take a detailed sexual history and conduct specific microbiological testing for C. trachomatis. Screening for coinfection, contact tracing, general education and health promotion are important public health components of managing LGV. LGV infection is characterised by three stages. In the first stage, the primary lesion is usually an asymptomatic small genital ulcer that heals spontaneously. This is followed by a painful inguinal lymphadenopathy associated with systemic features. Lymph node inflammation may progress to involve the surrounding subcutaneous tissue, causing an inflammatory mass (bubo) and/or abscess. Complications occur in 30% of cases as a result of bubo rupture and/or sinus tract or fistula formation. In anorectal disease, acute haemorrhagic inflammation of the colon and rectum is associated with involvement of perirectal lymphatic tissue.1,9 The third stage is characterised by chronic granulomatous inflammation leading to lymphatic obstruction, fibrosis and stricture formation.1 Clinical proctitis is a common problem in MSM, and C. trachomatis is one of the most frequent infectious agents found in this population. When suspected, C. trachomatis infections can be quickly identified and treated. However (as was the case with the patients described here), infected people may present to non-sexual-health practitioners (eg, gastroenterologists or colorectal surgeons) for persisting symptoms.9 Endoscopic features are non-specific, with a wide range of differential diagnoses including Crohn’s disease, lymphoma, anorectal carcinoma and other sexually transmitted ulcerative infections (eg, syphilis, herpes).3,4,9 Biopsies typically show only non-specific inflammatory features. C. trachomatis is divided into 15 serovars, labelled A, B, Ba, C–K and L1–L3, based on analysis of the major outer membrane protein. The various serovars are associated with specific disease manifestations: serovars A, B, Ba and C cause trachoma; serovars D–K are associated with urogenital infection; and serovars L1–L3 cause LGV.10 The L2 serovar can be further separated into L2, L2', L2a or L2b according to minor differences in their component amino acids.9 The commonly used commercial diagnostic tests for C. trachomatis are nucleic acid tests that can be performed on urine and on cervical and urethral swabs. Although not currently licensed for use on rectal swabs, the test may be used “off licence” and is the investigation of choice. Alternative tests, such as culture, are slow to perform and less sensitive, while serological tests cross-react with other Chlamydia species and can not distinguish between previous infection and current infection on single specimens. As commercial nucleic acid tests can not distinguish between uncomplicated rectal chlamydial infections (serovars D–K) and LGV (serovars L1–L3), directed testing is required (eg, outer-membrane protein sequencing). Therefore, to make a diagnosis of LGV, it is important at the outset to discuss the optimum specimen collection technique with a microbiologist at the laboratory to which the specimen will be sent. Two dry swabs should be sent with a specific request for LGV testing. This is a reference laboratory test and is offered by only a few public laboratories. The charge for the test is under the Medicare nucleic acid rebate, with no gap amount charged to the patient. The laboratory turnaround time for such a specimen is typically about 2 weeks from receipt. The correct diagnosis is essential, as treatment regimens recommended for LGV infection are much more prolonged than those for uncomplicated genital chlamydial infections — for example, a patient may receive 3 weeks’ treatment with doxcycline 100 mg twice a day or azithromycin 1 g weekly. However, good clinical trial data are lacking, and inadequate therapy may be associated with progressive disease and tissue destruction. Careful follow-up of the index patient is therefore essential. As with other sexually acquired infections, rigorous contact tracing is important to prevent further spread within the community. Counselling about healthy sexual practices is an important public health component of management, and patients should be alerted to the increased risk of HIV transmission associated with genital ulcer disease. A. Patient 1 — high-resolution anoscopy showing extensive ulceration distal to the dentate line. B. Patient 2 — biopsy from a rectal ulcer showing a lymphoid aggregate with a germinal centre (arrow) (haematoxylin and eosin stain; original magnification 3100). C. Patient 3 — colonoscopy showing multiple ulcerative lesions in the anorectal area.

Sebastiaan J van Hal MB ChB · Richard Hillman MD, FRCP · Damien J Stark BSc, PhD · Jock L Harkness FRCPA · Debbie Marriott FRACP, FRCPA

Re: “Chocoholic” in In Other Journals in the 4 June issue of the Journal (Med J Aust 2007; 186: 604). The item referred to “an intake of between 46 and 100 mg of polyphenol-containing dark chocolate per day”. The intake amount should have been stated as “between 46 and 100 g”. The html and pdf versions of the article have been updated.

Tanya Grassi

To the Editor: The recent article by Venn et al reported that childhood overweight carries through into adult overweight and obesity, but that most obese young adults in their study were “healthy” weight as children in 1985.1 As demonstrated by National Health Surveys, age is one of the strongest predictors of overweight,2 with body mass index (BMI) increasing as we grow older. However, there are two additional time-related components influencing obesity. Since 1985 (when Venn et al reported the prevalence of overweight and obesity in children was less than 10%), the environment appears to have become more obesogenic — a 2004 survey in New South Wales showed that 26% of children were overweight or obese.3 It is not only children who are vulnerable — the percentage of overweight adult Australians increased for almost all age groups from 1990 through 2001, and the mean BMI at which Australians enter adulthood has increased with each subsequent survey. For example, for women aged 20–24 years, mean BMI increased from 22.1 kg/m2 (1990) to 22.5 kg/m2 (1995) to 23.2 kg/m2 (2001) to 23.3 kg/m2 (2004). As the heights and weights were self-reported in these surveys, true BMI values may be even higher. We recently reported that year of birth (birth cohort) also predicts prevalence of overweight and obesity, independent of age and survey period; the prevalence of overweight and obesity in adults increased progressively with birth cohorts born since 1960.4 This birth span includes the cohort in the study by Venn et al.1 While obesity begins in childhood for only a small proportion of adults, the so-called healthy weight children now have a higher mean BMI, giving little margin for the seemingly inevitable increases in weight with ageing, before the population mean BMI reaches the cutpoint for overweight and later obesity. The 2004–2005 National Health Survey showed that men reached the overweight cutpoint at 25–29 years and women reached it at 30–34 years.5 Given increasing child and adult obesity, the need for allocation of public health resources to improve dietary and physical activity habits is undisputed. However, these data1,4 indicate that efforts should be directed to the hard-to-reach group, young adults, to prevent weight gain at this point. This will pose considerable challenges, because this group has minimal contact with health services, and perceives the threat of chronic illness as irrelevant. However, swift intervention is required, not only for their own health and that of their children as they become parents, but also because they will become overconsumers of health care for chronic diseases within a generation.

Margaret A Allman-Farinelli · Lesley King · Adrian E Bauman

In reply: Allman-Farinelli et al make an important point about the influence of age, survey period and cohort effects on the prevalence of overweight and obesity. While age and cohort effects could not be clearly separated in the 1985 Australian Schools Health and Fitness Survey, the prevalence of overweight and obesity increased with age in 7–15-year-olds.1 Our data collected from 4571 of the individuals in that survey at follow-up about 20 years later also showed an increase in the prevalence of overweight and obesity with age, although these findings were not presented in our report.2 In the Box, we show the distribution of body mass index (BMI) values for men and women in three age groups (24–27 years, 28–30 years and 31–34 years). Mean BMI values across the age groups were 25.2 kg/m2, 25.6 kg/m2, and 26.5 kg/m2 in men and 23.5 kg/m2, 24.2 kg/m2, and 24.6 kg/m2 in women. The prevalence of obesity (BMI ≥ 30 kg/m2) increased with increasing age as follows: age 24–27 years — men 12.2%, women 9.9%; age 28–30 years — men 12.3%, women 12.0%; and age 31–34 years — men 15.6%, women 14.6%. Distribution of body mass index values for men and women in three different age groups* * 24–27 years, 757 men and 854 women; 28–30 years, 767 men and 807 women; and 31–34 years, 673 men and 691 women in the 20-year follow-up of the 1985 Australian Schools Health and Fitness Survey.

Alison J Venn · Russell J Thomson · Michael D Schmidt · Verity J Cleland · Beverley A Curry · Hanni C Gennat · Terence Dwyer

To the Editor: O’Leary and colleagues1 rightly point to the need for better evidence on whether low to moderate maternal alcohol intake affects the fetus. Evidence to date is weak and inconsistent, largely because alcohol consumption in pregnancy is generally poorly documented,2 and few studies have recorded data on factors that could potentially modify fetal exposure. Evidence could come from large pregnancy cohort studies, usually designed to address other issues, but there are no published guidelines on how to collect information relevant to fetal alcohol exposure during gestation. There is, therefore, a great need to define a core dataset for research studies, as well as one that is sufficiently simple to use in routine pregnancy care settings. We have identified a number of issues that need to be addressed. To stimulate discussion, the Box shows our suggested core dataset. Maternal alcohol intake: Alcohol questionnaires have largely been designed to identify women who are heavy drinkers, misuse alcohol or are alcohol-dependent.3 Questionnaires are needed that capture information on alcohol intake across the range, from minimal to heavy drinking, as well as information on drinking patterns and alcohol intake at different periods of gestation. Factors that may modify the relationship between maternal intake and fetal alcohol exposure: For a given maternal intake over a given period, maternal blood alcohol level and hence fetal alcohol exposure may vary according to maternal size and body composition. Other factors can affect maternal alcohol absorption and elimination, such as whether alcohol is taken with food,4 and possibly maternal genotype.5 This information is rarely reported in pregnancy studies. Factors that may modify effects of alcohol on the fetus: There is animal evidence that maternal micronutrient supplementation may protect the fetus against some of the adverse effects of gestational alcohol exposure.6 We need to consider recording supplement use and measures of maternal nutritional intake or status. In well resourced studies, fetal genotype could also be considered.6 Researchers with expertise in the field need to reach consensus and provide guidelines on the best way to assess fetal alcohol exposure, so that pregnancy researchers and clinicians with little experience of alcohol research do not need to create their own. Better data should provide better evidence on which to base advice to women who are (or may be) pregnant. Good studies may also provide explanations for the apparently variable link between maternal alcohol consumption and adverse sequelae in the offspring. Suggested core dataset for studies of maternal alcohol intake during gestation and outcome of offspring As a basis for discussion, we suggest the folllowing dataset: Baseline Weeks of gestation at pregnancy recognition, or expected date of delivery and date that pregnancy was recognised (to allow calculation) Height For specific periods of gestation (eg, from date of start of last menstrual period until pregnancy recognition; from pregnancy recognition to 12 weeks’ gestation; from 13 to 28 weeks; and from 28 weeks to term) Body weight (eg, at 12 and 28 weeks) Alcohol intake — we suggest: Average number of standard alcoholic drinks per week; Average number of days per week on which alcoholic drinks are taken; and Maximum number of drinks on one occasion. Whether alcohol is taken with or soon after food (never, sometimes, usually or always) Dietary intake of fruit and vegetables Use of nutritional supplements

Ruth Morley · Jane L Halliday · Susan M Donath

To the Editor: I would like to put forward a consumer’s perspective in response to the recent article by O’Leary and colleagues.1 I am puzzled that they concluded that the National Health and Medical Research Council (NHMRC) guideline is in step with policies of the United Kingdom and Canada. My research of Canadian policy suggests that it is in direct contrast to current NHMRC guidelines. Health Canada states very clearly, “Whether you are trying to get pregnant or are pregnant already, stop drinking alcohol”,2 and “No amount or type of alcohol during pregnancy is considered safe”.3,4 The potential harm to the birth mother that results from an abstinence-based message was also raised by O’Leary and colleagues. Elizabeth Russell, the birth mother of two sons affected by prenatal exposure to alcohol, offers an alternative viewpoint: By not discussing alcohol and pregnancy through misplaced compassion, we are hurting one person for the sake of another. Very few mothers would want that but that is exactly what is happening — children are being sacrificed to ensure that the anxiety level of a mother is kept within acceptable limits — neither mother nor child will benefit from this methodology.5 Many pregnant women give up eating shellfish and processed meats and drinking coffee but still continue to consume alcohol, thinking it is safe because they have not been told otherwise. I would like to pose the following questions. What is the potential harm of not having an abstinence message? Might this prevent women who are alcohol-dependent from seeking help to alter their drinking behaviour when pregnant because they are not aware of the risks? Should further research to “elucidate the true association between low to moderate alcohol consumption and fetal harm” really be a priority? What are the benefits to the unborn child of trying to ascertain a safe level of consumption of a teratogen and neurotoxin that is known to disrupt fetal development, particularly the fetal brain, throughout the three trimesters of pregnancy?6 While I acknowledge the importance of scientifically sound research on the risks of prenatal exposure to alcohol, I am concerned that this argument is diverting attention and dollars away from the urgent need for diagnosis and management of fetal alcohol spectrum disorder (FASD) in Australia. The reality is that children, adolescents and adults with FASD are seldom recognised, seldom treated effectively, and seldom connected to service dollars. Addressing this situation needs to be the priority.

Sue Miers

In reply: Morley and colleagues and Miers raise a number of interesting discussion points. As we reported in our policy review, the Canadian, United Kingdom and Australian guidelines have similar intent but differ in emphasis.1 Health Canada’s policy position is that, although abstinence is the prudent choice, fetal risk is relative to the amount of alcohol consumed and is minimal with low levels of maternal alcohol intake. Australian policy addresses the same issues, with less emphasis on abstinence and more on avoiding intoxication and ensuring low-level alcohol consumption. Since our review was published, UK guidelines have been reframed to emphasise abstinence, but their message has not changed — they now place more weight on avoiding alcohol during pregnancy.2 We strongly agree with Morley and colleagues that screening for alcohol should be routine in all pregnant women, and that standardised items should be included in a core dataset.3 If this were implemented, Australia would be in a unique position to make a valuable contribution to alcohol and pregnancy research. Miers comments that research into the impact of low to moderate alcohol exposure during pregnancy should not be a priority, because it may direct “attention and dollars away from the urgent need for diagnosis and management of fetal alcohol spectrum disorder”. Although we agree that specialised clinical services are important — and lacking — in Australia, it is short-sighted to suggest that there is no need for further research to establish the true risks from low to moderate alcohol consumption. Rates of alcohol consumption in Australia are high: about 80% of women report alcohol consumption in the 3 months before pregnancy, 14% report binge drinking, while 47% report that pregnancy was unplanned.4 Many fetuses may thus be exposed to alcohol before women are aware they are pregnant. Unfortunately, many women are unable to stop drinking and may expose their babies to high alcohol levels. Health professionals need to have a true estimate of risk to the fetus and know what additional factors (eg, genetics and nutrition) may alter risk, and women deserve to be well informed. The need for research is well articulated by Morley et al. Research evidence in humans does not clearly indicate a risk to the fetus from low levels of alcohol consumption, and this has led to inconsistent policy.5,6 As we point out, the potential for harm from an abstinence message should be considered when Australian alcohol guidelines are reframed. Whatever the policy, it needs to be widely disseminated, in a “digestible” format, to health professionals and the community. Research being conducted at the Telethon Institute for Child Health Research is evaluating educational materials for health professionals about alcohol and pregnancy.

Colleen M O’Leary · Louise M Heuzenroeder · Elizabeth J Elliott · Carol I Bower

To the Editor: Kawasaki disease (KD) is an acute vasculitis of unknown aetiology occurring mostly in infants and young children. KD is characterised by fever of more than 4 days’ duration; conjunctivitis; rash; cervical lymphadenopathy; erythema of the lips, oral mucosa, palms and soles; and oedema of the hands and feet.1 Coronary artery aneurysms develop in 15%–25% of untreated children,2 with attendant risk of ischaemic heart disease, myocardial infarction and sudden death.3,4 Treatment with intravenous immunoglobulin (IVIG) within the first 10 days reduces the incidence of aneurysm to less than 5%.4 A KD diagnosis is clinical, based on the recognition of a characteristic set of signs and symptoms.4 The 10%–45% of children who meet only some of the classical criteria are said to have “atypical” or “incomplete” KD. These children have a higher risk of coronary artery aneurysm than children with typical KD.4 Abdominal symptoms, including acute appendicitis and appendicular vasculitis, can occur before the development of classical features of KD.4,5 A 50% coronary artery aneurysm rate has been reported in children with KD and acute abdomen. It is still unclear whether this reflects a delay in diagnosis and treatment or is a marker of a more severe vasculitis involving the intestinal tract. A 3-year-old boy presented with a 2-week history of remittent, high-spiking fever (37.5–39.0° C; 2–3 spikes/day), right lower quadrant abdominal pain, and McBurney’s sign with rebound tenderness. Abdominal ultrasonography suggested a diagnosis of acute appendicitis with peritonitis. The postoperative diagnosis was appendicular vasculitis with peritoneal inflammation and serous secretion. Fever persisted despite treatment with intravenous cephalosporin. Several days later, the boy developed conjunctivitis, cracked lips, a raised erythrocyte sedimentation rate and C-reactive protein level, and thrombocytosis (715 × 109 platelets/L). KD was suspected, and an echocardiogram revealed two sacciform coronary artery aneurysms (diameters, 3.1 mm and 2.9 mm) in the proximal part of the common trunk. The child was given IVIG (2 g/kg) and oral acetylsalicylic acid (100 mg/kg per day in four divided doses). As the fever failed to resolve, the patient was given a second dose of IVIG,4 this time leading to a dramatic clinical improvement. Five days later, he developed oedema of the hands and periungual peeling of the fingers. His aspirin dose was reduced to 5 mg/kg/day. Follow-up echocardiograms at 3 months and 6 months demonstrated persistent coronary artery dilatation. Persistent fever with conjunctivitis and cracked lips should alert clinicians to the possibility of KD. At our patient’s age, acute appendicitis is rare, and other causes of abdominal pain must be excluded. In this case, the unusual postoperative course, with persistent fever even after antibiotic treatment, was another clue to establishing the correct diagnosis.

Maria Cristina Maggio · Andrea Liotta · Salvino M Vitaliti · Giovanni Corsello

To the Editor: Gerber’s article1 about the case Royal Women’s Hospital v Medical Practitioners Board of Victoria raises important issues but contains significant errors. The Medical Practice Act 1994 (Vic) states that the main purposes of the Act are “to protect the public by providing for . . . investigations into the professional conduct . . . of registered medical practitioners” — Section 1(a). Section 22(1) makes it clear that any person may notify the Board if they believe a person may have engaged in unprofessional conduct. Section 25(1) states that the Board must investigate notifications unless they meet certain criteria, which the Board did not believe were met in this case. Gerber is not correct in stating that the Board made no attempt to question the medical practitioners involved. In the absence of the patient’s consent, all but one of the doctors refused to provide any information in response to the complaint. Gerber’s statement that the subcommittee of the Board that conducted the preliminary investigation concluded that the complaint was “frivolous and vexatious” is wrong. The subcommittee recommended that the matter be closed. Later, the full Board chose not to accept this recommendation, as the investigation had been hampered by lack of information, including access to the original hospital records. While formal hearing panels have the power to subpoena documents or persons, this power does not extend to the Board’s preliminary investigations. The Board does have the power to apply to a magistrate for the issue of a search warrant. Gerber refers to the powers of the Board under Section 48 and Section 49 of the Act. These powers only come into play if the Board has determined that it will conduct a formal hearing. He is also incorrect when he states that the Board can compel medical practitioners to appear before the Board. This power is not available during preliminary investigations. The considerable delay in the matter being finalised was due to the legal appeals mounted by the Royal Women’s Hospital against the decision of the Magistrate to allow the Board access to the hospital records. Another issue raised by Gerber requires clarification: the Board does not currently have the power to conciliate disputes or conduct mediations. The Board did not ultimately dismiss the matter as frivolous and vexatious. When the subcommittee, having been provided with the records, reported to the Board that it did not find evidence of unprofessional conduct, the Board closed the investigation. I trust that this information will correct the public record on this important matter.

Joanna M Flynn

In reply: Flynn points to some minor technical differences in my historical recount of the handling, by the Medical Practitioners Board of Victoria, of the complaint against the medical specialists involved in a late-term abortion.1 None requires a reply, save for Flynn’s assertion that “the Board does not currently have the power to conciliate disputes or conduct mediations”. The Board does not require statutory power to approach a hospital in a conciliatory manner so as to explore whether an impasse, involving confidentiality, can be resolved without recourse to litigation. Was the Board’s only remedy to raid the hospital, trawling for evidence to decide whether there were grounds for the possible suspension or cancellation of registration of the doctors involved in the complaint? We both agree that the relevant legislation mandated the Board, on the material before it, to investigate the charge of serious professional misconduct. Where we disagree is that, having overruled its own subcommittee’s recommendation that the matter be closed, the Board failed (I maintain) in its statutory duty to promptly institute a formal hearing. Had the various specialists been subpoenaed, this would have cleared them of professional misconduct, thereby preventing the considerable and unnecessary stress to these witnesses over a period of 5 years.

Paul Gerber

To the Editor: Breen’s timely call for a reality check on medical professionalism noted major global changes that affect contemporary doctor–patient relationships: new technology, changing market forces, evidence-based treatment protocols, and resource-driven health services and policies.1 The call by our colleagues in the United States and United Kingdom to restore “trust that the public used to have in the profession” was urgent. Breen’s thesis posits that lost trust is due to an “altered balance” of ethical issues faced by doctors because of a generational shift from the ethical principle of “beneficence” to “autonomy” to “justice” and “distributive justice”. This view contrasts with Green and Bloch’s analysis of the mental health care system’s ethical concerns arising from an adherence to “efficiency-driven” policies that started in many countries during the 1980s.2 They suggest the system itself is flawed. Green and Bloch suggested that the legacy of efficiency-driven policies created two current moral compromises for our profession: first, a threat to the “ethic of agency”; second, the constraints those policies imposed on ethical principles precisely because they were not based on justice, instead being created to meet wider socioeconomic and political considerations. Their views point to the heart of the matter, beyond Breen’s suggested remedy to be found in “stronger leadership”, which may be necessary, but is not sufficient without an urgent update on personal medical ethics. At the individual doctor’s experience, Green and Bloch locate conflict arising from competing interest when doctors’ “principle of fidelity is juxtaposing their financial interests alongside patients’ needs”. In the US, a study found 28% of physicians receive direct payment for consulting, lectures or enrolling patients in trials, and 94% report “some type of relationship with the pharmaceutical industry”.3 To avoid the conundrum posed by the ethics of conflict of interests we face when confronted by these physician–industry relationships, or by efficiency-driven policies, risks perpetuating the very loss of trust that we need to restore. Breen’s analysis, an important step in the needed debate on medical professionalism, should account for not merely shifts in the ethical balance, but also the incremental erosion of trust. As it stands, he expressed our very Australian attitude “she’ll be right”. Our overseas colleagues, as well as locals, have suggested that “she won’t be right, mate” when it comes to managerialism eroding the ethical foundations of medicine.4

George Halasz

In reply: My recent article was submitted under the category of “For Debate”, so it is pleasing that Halasz has joined the debate. I am disappointed that he interprets my view as “she’ll be right”. My point is that revising or repackaging existing ethical codes will not, on its own, fix any of the perceived problems of “managerialism eroding the ethical foundations of medicine”. Working constructively, consistent with existing ethical codes, within our health care system, as is also suggested by Green and Bloch,1 is more likely to achieve better outcomes for our community. As I stated and as Green and Bloch imply, this will not always be a simple matter. I am in fierce agreement with Halasz over steps to reduce erosion of trust,2 but that was not the focus of my article.

Kerry J Breen

To the Editor: Clark et al have claimed to map the distribution of services for people with chronic heart failure (CHF) against the distribution of these people.1 An examination will show that they have mapped the distribution of people likely to have CHF, using age and Aboriginality as surrogate markers. The stated mapping of the services shows the services probably available to these people. A map is drawn to show us what the cartographer wants us to see.2 The authors note that high prevalence in remote regions has been shown on the maps, but they have not considered different mapping methods to provide a better representation of their results.3 This has led to an anomaly so that, when calculating numbers of people with CHF, the maps show remote areas with giant households containing between 24 and 300 people. CHF programs were located by a snowball sampling technique, which by its nature will miss isolated examples. Isolation is a feature of rural and remote practice, so this method is biased to finding metropolitan examples. In my own rural practice in Griffith, I found a local program that had been operating in 2004 and a distance program run by a health fund, neither of which had been identified by Clark et al. On the medical front, the authors asserted that access to a CHF management program is a mark of equity in health services. This is based on a metro-centric model looking at admission with CHF to a major teaching hospital, and showing that rates of readmission and death were reduced where a nurse and a pharmacist made a single visit to someone with CHF in their home after hospital discharge. The relevance of this activity to a rural person with a single community pharmacist who may make a home visit, a single point of contact with medical services in their general practitioner, and the possible availability of a community nurse to visit them regularly is unproven. Perhaps the city folk were copying the principles of the services we already had?

Elizabeth A Dodd

In reply: We thank Dodd for her commentary on our article.1 We concur with many of the highlighted issues relating to our suboptimal response to the burden and management of chronic heart failure (CHF) in rural and remote Australia. These include the lack of rigorous epidemiological data and lack of specialist services “beyond city limits”. Unfortunately, we have limited space to respond fully. However, we re-emphasise that, although our previous estimates2 complement that of the Canberra Study,3 neither can replace an Australia-wide study of CHF that samples metropolitan, regional, rural and Indigenous communities. We also stand by (within the context of the stated limitations) the accuracy of our mapping of the CHF programs and the location of general practice services in Australia for the study period. Our geo-mapping approach and data have been well validated by the National Centre for Social Applications of Geographic Information Systems (GISCA).4 For example, Jenks’ (natural breaks) classification is used for all sociodemographic thematic mapping at GISCA. Overall, we identified only four CHF programs which were located in regional areas. Other rural programs were excluded as they did not meet our prespecified definition of a CHF program. In summary, we acknowledge the need for better data to describe the burden of CHF throughout Australia. We also explicitly acknowledge the need for a less “metro-centric” approach to CHF service: perhaps by using remote monitoring techniques.5

Robyn A Clark · Andrea Driscoll · Justin Nottage · Skye McLennan · David M Coombe · Errol J Bamford · David Wilkinson · Simon Stewart

To the Editor: With the ski season in Australia drawing to a close for another year, it is a good time to reflect on the injury prevention benefits of wearing helmets when skiing or snowboarding. Skiing falls can be fatal. Two people have died from head injuries on Australian skifields in recent years: a skier died after colliding with a tree branch on an intermediate run at Mt Buller, Victoria, in 2003; and in 2006, a novice snowboarder died after falls sustained while snowboarding at Thredbo, New South Wales.1 Neither person was wearing a helmet. In Australia in 2002–03, 3.5% of skiing-related hospital admissions and 6.2% of snowboarding-related admissions were due to intracranial injuries.2 During the 2004 and 2005 ski seasons we collected data on the use of helmets in snowboarders presenting to the Mt Buller Medical Centre. Of 494 snowboarders, 17.6% had been wearing helmets, and none had sustained a head injury. Of the nine patients with head injuries, none had been wearing helmets. These figures are similar to those reported in overseas studies, which have shown that wearing a helmet can reduce the snow-sport head injury rate by up to 60%.3,4 The use of helmets for snow sports makes intuitive and biological sense, as it does for cyclists, but, unfortunately, Australia is yet to issue a snow-sport helmet performance standard, as it does for bicycle helmets. Helmet use should be strongly recommended for all snowboarders and skiers. In particular, helmets should be made compulsory for children, who are more susceptible to head injury4 and who are often present at ski resorts in large organised school groups that could readily be made to comply. At present in Australia, helmet use is not compulsory for children attending skiing or snowboarding lessons, as it is in North America. This is in spite of the fact that helmet use is compulsory in Australia for school skiing and snowboarding competition events. Some skiers and snowboarders are gradually getting the message about helmets, and a recent informal survey at Mt Buller (Buller Ski Lifts personnel, personal communication) estimated the rate of helmet use to be 20% among adults and 68% among children — but this still leaves over 30% of children vulnerable. Snow-sports helmets now come in many colours, shapes and sizes, and are increasingly acceptable to young people. A helmet is probably the cheapest individual item of clothing for a ski holiday. And it may save your life.

Graham M Slaney · Judith Finn · Angus Cook · Philip Weinstein

To the Editor: It was interesting to read of the trial conducted by Manley et al1 using yoghurt containing Lactobacillus rhamnosus to clear vancomycin-resistant enterococci. I would like to add an historical note. The use of yoghurt in restoring bowel flora was practised by Dr J H Kellogg (of Corn Flakes fame) around the end of the 19th century. Kellogg was the chief physician at the Battle Creek Sanitarium in Michigan and was an advocate of high colonic irrigation, for he believed the colon was a sewer of toxic materials that were the causal factor in many diseases. Following this procedure, the patient was given a pint of yoghurt — half to be taken orally, and the remainder given by enema.2,3 By these measures, Kellogg claimed to have cured many conditions, from cancer of the stomach to psychiatric problems.

H Reginald Magee

Suffering of child soldiers An estimated 250 000 child soldiers currently suffer from abuse around the world. Some children exposed to traumatic events as child soldiers are more likely to experience feelings of revenge and are less open to the idea of reconciliation, according to German researchers. A total of 169 Ugandan and Congolese former child soldiers were interviewed in African rehabilitation centres, and symptoms of post-traumatic stress disorder (PTSD) were assessed. The mean age of the participants was 15.3 years and the mean age of recruitment into the armed forces was 12.1 years. Reported traumatic experiences included being a witness to shooting or wounding, and having been beaten or forced into sexual contact. Over half reported having killed someone. Children who showed more PTSD symptoms were significantly less open to reconciliation and experienced significantly more feelings of revenge. The authors comment that post-traumatic stress may hamper the process of overcoming feelings of hate and revenge in former child soldiers, and that such children may see retaliation as an appropriate method of regaining their integrity. JAMA 2007; 298: 555-559 Cannabis and mental health Regular users of cannabis have a higher than average risk of developing psychosis beyond the transient intoxication stage, suggests a recent systematic review. The evidence for an association between cannabis and affective disorders including depression and anxiety was less strong but still present. Despite the possibility of confounding factors or bias affecting the result, the researchers conclude that the evidence is consistent enough to provide a clear association between use of the drug and psychotic symptoms, including severe psychotic disorders. They comment that young people should be warned of the increased risk of developing a psychotic illness if they use cannabis. Lancet 2007; 370: 319-328 Progesterone prolongs pregnancy Administration of progesterone appears to reduce the rate of spontaneous early preterm delivery in women with a short cervix, according to a large British randomised trial. Over 24 000 pregnant women were screened to determine cervical length, with 413 found to fulfil the criteria for a short cervix (15 mm or less). From 24 to 34 weeks’ gestation, these women were randomly assigned to receive either 200 mg per night of vaginal progesterone, or a placebo. Spontaneous delivery before 34 weeks’ gestation was significantly less frequent in the progesterone group. The authors advocate routine ultrasonographic screening for cervical length in pregnant women, with administration of prophylactic progesterone to those found to have a short cervix. N Engl J Med 2007; 357: 462-469 Cocaine kitty Sydney veterinarians treating an anxious moggy found more than they bargained for when they performed toxicological studies on the fractious feline. The 8-month-old cat was presented with a history of agitation and on examination was tachycardic with dilated pupils. The animal paced incessantly, climbing the bars of its cage and reacting nervously to any approach. A urinary toxicology screen revealed cocaine metabolites and benzodiazepines. Astute veterinary clinicians quizzed the cat’s owners and learned of a recent dinner party where cocaine was available “on plates” and at which it seemed the cat had been present. Treatment with a cardioselective b-antagonist was commenced and the cat’s condition improved substantially. The authors comment that although there are few documented reports of cocaine intoxication in companion animals, urinary drug screening is a valuable adjunct in the investigation of veterinary patients with unexplained neurological and cardiovascular signs. J Feline Med Surg 2007; 9: 265-270 Doctors’ work hours affect mortality A US study suggests that changes in work-hour regulations for hospital medical residents appear to be associated with reduced patient mortality. Researchers analysed the hospital stay data of over 1.5 million people over a 4-year period, before and after the introduction of restricted work hours for junior hospital doctors. Non-teaching hospital patients were used as a control group. After restriction of work hours to 75 hours per week in 2003, there was a 0.25% reduction in absolute mortality and a 3.7% reduction in relative risk for death. The changes applied only to medical patients, with death rates in surgical patients remaining unchanged. The improvement in mortality rates was more pronounced in older patients, in people admitted with infectious diseases, and in those with congestive heart failure and gastrointestinal bleeding. Despite admitting a number of limitations, including the possibility of confounding variables affecting mortality rates, the researchers suggest that the work-hour regulations may have shifted care to more experienced clinicians, resulting in better outcomes for patients. Ann Intern Med 2007; 147: 73-80

Tanya Grassi