Volume 192 Issue 6

The challenge of convincing governments of the value of a nationally comprehensive data collection Congenital anomalies are worth bothering about — they affect around one in 20 births in Australia.1-3 They are the second most common cause of perinatal and infant mortality and the fourth commonest cause of mortality in 1–14-year-olds in Australia.4,5 They are major contributors to hospital admissions6 and often result in lifetime disability. They are costly to our health system, including the considerable expense of providing programs to screen for, diagnose and terminate pregnancies affected by major congenital anomalies (Down syndrome and neural tube defects in particular). Importantly, some anomalies are preventable, including neural tube defects (70% preventable with adequate periconceptional folic acid7) and anomalies resulting from exposure to teratogens (eg, by avoiding alcohol during pregnancy). For many congenital anomalies, early identification allows interventions to decrease the risk of secondary disabilities. We need good data to monitor trends in congenital anomalies, to identify clusters of cases that may require investigation for possible environmental causes, and to evaluate the effectiveness of interventions for screening, treatment and prevention. Nationally, the Australian Institute of Health and Welfare (AIHW) National Perinatal Statistics Unit (NPSU) collates information on congenital anomalies that is supplied voluntarily by health departments in the states and territories. However, there remains considerable variability between jurisdictions in the scope and quality of the data collected. This variability — which includes the sources of case ascertainment, the upper age limit for inclusion, definitions and classifications used, methods of operation, and resources available for collecting, updating, validating and using the information — limits the utility of the collection. National data published by the NPSU can only be as complete as the data provided by individual states and territories. Lack of completeness is evident in two recent AIHW reports. The first, Congenital anomalies in Australia 2002–2003,8 does not include data from the Northern Territory, because data were not available. This may change, as the NT is reviewing its perinatal data needs. In addition, data were only available from four states on terminations of pregnancy at less than 20 weeks’ gestation for congenital anomalies.8 The absence of information on early terminations is also evident in the second report, Neural tube defects in Australia.9 The prevalence of neural tube defects at birth for the period 1998–2005 was similar for all states included in the report, at around 5 per 10 000 births, but the total prevalence (including early terminations of pregnancy from the four states collecting such information) was more than twice as high, at 10.1 per 10 000 pregnancies. Furthermore, of the four states collecting data on early terminations, the total prevalence in 2005 in South Australia, Victoria and Western Australia was 13.3 per 10 000, double that in New South Wales (6.2 per 10 000), suggesting incomplete ascertainment of terminations in NSW. One of the purposes of the report on neural tube defects9 was to provide baseline data against which to monitor the effect of the introduction of mandatory fortification of bread-making wheat flour with folic acid, in place nationally by 13 September 2009. Because such a high proportion of neural tube defects are diagnosed prenatally and affected pregnancies terminated, post-intervention monitoring in Australia will be restricted to the three states where there is complete ascertainment of such terminations. The inclusion of terminations of pregnancy is essential for a national data collection on congenital anomalies — not only for evaluating interventions such as folic acid fortification, but also for evaluating and monitoring the safety and quality of prenatal screening programs and diagnostic tests, and the associated health and psychosocial impacts. In response to the limitations in national data collection, a program was commenced in 2007 to develop a national minimum dataset on congenital anomalies. Members of a committee representative of the states and territories reached consensus on collecting good data on a limited number of conditions, particularly those with important clinical, social or health care impacts; and on the use of internationally agreed definitions for congenital anomalies. However, because of existing data limitations in some jurisdictions, commitment to a national minimum dataset is not currently possible. In addition, the scope of the proposed national collection has been limited to the perinatal period, which means terminations of pregnancy for congenital anomalies before 20 weeks’ gestation will not be included. These decisions are very disappointing and suggest that Australian policymakers and governments are still to be convinced of the value of monitoring congenital anomalies, which, despite their magnitude and importance as a cause of mortality and morbidity, are clearly not seen as a public health priority. Historically, data collection for congenital anomalies has been unfunded or under-resourced in Australia. Apart from mandatory folic acid fortification, there has been no national policy on the surveillance, prevention and management of congenital anomalies. There has also been no consumer involvement in deliberations on the societal impact of these anomalies, the need to collect national data, and the ways in which these data should be collected and used. The challenge remains to convince governments of the value of a nationally comprehensive collection that can be used to monitor trends, identify clusters that may require investigation, evaluate the effectiveness of screening and interventions for treatment and prevention, and allow research into the prevention of congenital anomalies.

Carol I Bower MB BS, PhD, FAFPHM · David Lester-Smith BM BS, FRACP, MPH · Elizabeth J Elliott MD, MPhil, FRACP

Comparative effectiveness research may stimulate heated debate, but ultimately, those who question its findings need to provide high-quality data to support their arguments One consequence of the continuing rise in the cost of health care has been the emergence of comparative effectiveness research.1 This variant of evidence-based medicine is defined as: . . . the generation and synthesis of evidence that compares the benefits and harms of alternative methods to prevent, diagnose, treat and monitor a clinical condition, or to improve the delivery of care.1 Furthermore, the purpose of comparative effectiveness research is: . . . to assist consumers, clinicians, purchasers, and policy makers to make informed decisions that will improve health care at both the individual and population levels.1 When confronted with health care consuming an ever-increasing percentage of the gross domestic product, politicians and policymakers have enthusiastically embraced comparative effectiveness research,2 and Prime Minister Rudd is no exception. In a recent speech, Mr Rudd proclaimed that medical research needed to play a greater role in reducing burgeoning health budgets. “Patients need treatments, technologies and procedures for which there is evidence from research that these are safe and effective.”3 He cited a recent article in the New England Journal of Medicine (NEJM), in which research by an Australian team “found a commonly available treatment for fractures of the bones of the spinal cord was in fact no better than doing nothing at all.”3 He was referring to the treatment of osteoporotic vertebral fractures with vertebroplasty — that is, the percutaneous injection of medical cement into the fractured vertebral body. Such procedures are performed in some 100 000 patients per year in the United States4 and about 700 patients per year in Australia.5 Late last year, this area of practice received a seismic shock when the NEJM simultaneously published two randomised controlled trials (RCTs) — one conducted in Australia6 and one in the US, the United Kingdom and Australia.7 These trials were conducted independently of each other, and both showed that the outcomes of vertebroplasty in patients with osteoporotic vertebral fractures were no different than for a placebo procedure. It was doubtlessly anticipated that publication of these two RCTs would inflame debate, arousing passionate defence of vertebroplasty.4 This is to be expected whenever evidence-based medicine clashes with the collective wisdom of clinical experience. For more than a decade, it had been argued that vertebroplasty was so successful that RCTs were unnecessary or even unethical!4 Not surprisingly, the two RCTs turned the practice of vertebroplasty on its head. In view of the seminal importance of these studies and seeking to inform the broad readership of the Journal, I duly sought an editorial from the lead authors of the NEJM studies, Professor Rachelle Buchbinder from the Monash Department of Clinical Epidemiology at Cabrini Hospital in Melbourne and Professor David Kallmes from Mayo Clinic in Rochester in the US. Then strange things began to happen. Just before the editorial was published, I received an email critical of its content. Then, subsequent to its appearance in the 2 November 2009 issue of the Journal, further emails arrived advising, among other things, that the editorial be retracted. Medical science has always thrived on debate in an open forum, wherein discussion and interpretation of the evidence is to be encouraged. Yet, I was the recipient of closed communications pointing out the weaknesses of the RCTs, as well as suggesting that the reputations of the NEJM and the Medical Journal of Australia had been diminished by the original publication of the RCTs and our subsequent editorial. More sinister, perhaps, is the fact that Professor Buchbinder was subjected to a far more vitriolic campaign, necessitating the threat of legal action (Rachelle Buchbinder, personal communication). In this issue of the Journal, we publish the views of Clark and colleagues, a group of Australian vertebroplasty experts,8 and the rejoinder by Buchbinder and colleagues.9 Clark et al point, among other things, to problems with patient selection and recruitment as a reason for the negative findings of the RCTs, while Buchbinder et al robustly defend the findings and their subsequent interpretation. It is up to the readers of the Journal to decide for themselves whether the two trials and the editorial that sought to interpret them represent the best available evidence on the effectiveness of vertebroplasty. Where do we go from here? It is easy to be critical of study methods, findings and interpretations, but I strongly believe that, when considering important clinical issues, criticisms must not be ad hominen but be supported by new data. It may well be argued that vertebroplasty should no longer be performed except in the context of a study aimed at resolving unresolved questions.4 At the very least, vertebroplasty practitioners should now relate the outcomes of the NEJM trials in their discussions with patients before proceeding to gaining their informed consent.10

Martin B Van Der Weyden MD, FRACP, FRCPA

This short report summarises the proceedings of the international symposium “Precincts, people and places — forging new partnerships”, held in Brisbane on 24 and 25 July 2009 The international symposium “Precincts, people and places — forging new partnerships” was organised by the University of Queensland (UQ) and the Pharmacy Australia Centre of Excellence (PACE) precinct initiative. The PACE precinct is a 1.7 hectare site adjoining the Princess Alexandra Hospital (one of Australia’s largest tertiary referral hospitals) and only 10 minutes walk from the St Lucia campus of UQ. The precinct will encompass the UQ School of Pharmacy and adjacent health-related commercial developments, the Translational Research Institute and BioPharmaceuticals Australia. The UQ School of Pharmacy will open in 2010, and other developments on the site will be phased in over the next 5 years. The location of the precinct offers an opportunity to explore innovative solutions to bioscience developments around translational research, health promotion and diseases prevention and the various forms of e-health platforms. A physical location does not, itself, ensure that those working around that precinct will interact — something else is required. While the PACE precinct offers a significant opportunity for health research innovation in that it links the universities, hospitals and medical research institutes with health industry developments, it is that “something else” we wished to debate in the “Precincts, people and places” symposium. Participants in the symposium were drawn from universities, research agencies and institutes, industry, government, and health service providers, and included architects and designers. Discussion sessions and panels were interspersed with provocations from speakers from the United States, United Kingdom and India. Transcripts of these contributions, videos of special presentations, and copies of selected presentation materials will be available on the PACE website.1 Norman Swan, well known health journalist and commentator, moderated the symposium proceedings with energy and critical probing. Presentations from those working in health “precincts” in Australia, the UK, the US and India were interspersed with small group sessions and discussion. We are grateful for the table facilitators who helped crystallise the outcomes of the discussions. The symposium tackled the role of special precincts as productive hotspots and “living laboratories” for innovation, and examined the particular importance of precincts in the health system as a springboard and case study for discussion. Four broad challenges were identified. How do we move beyond mere collocation to multi-party engagement and deep collaboration? How do we better promote real innovation outcomes, and mark out a place for innovative precincts within the wider innovation system?2 How do we shape precincts that are “fit for purpose” in the 21st century and, in particular, how do we optimise across the physical and the virtual? What, for example, can really only be achieved by a face-to-face interaction and what can be done from a “virtual” perspective? Within the health system, could we envisage a national effort similar in ambition to the one on which the UK has embarked?3,4 The discussion at the symposium was organised around four basic questions. Do precincts matter? What has been our experience: what has worked and not worked? What are the challenges and issues we need to address to get better outcomes? What is an action agenda for moving forward? Why precincts matterThe main conclusions from the group are summarised below. A consensus emerged that precincts matter because they: Provide the scale to attract good people, funding, facilities and global partnerships; Support first-class facilities to support first-class research and learning; Provide proximity and access to people and capabilities that are task-critical, as well as to the end-users of outcomes (thus nesting problem-solving within an engaged community); Facilitate synergies, serendipity, focus and, most importantly, multidisciplinary and transdisciplinary capabilities (allowing unstructured conversations across disciplines and entities); Provide a catalyst for intergovernmental coordination and for economic development strategies (combining new facility and infrastructure development, the coordination of information technology platforms and hubs, and the expression of new business models); Address the inbuilt human need for face-to-face social contact (fostering social capital); Unlock latent intellectual capital and underused assets; Secure better public (socioeconomic) outcomes for the investment of time, effort and money; and Facilitate a more creative, culturally attuned innovation ecosystem. In contemplating the question of the interface between the physical and the virtual, it was concluded that special attention needed to be given to this interface and to focusing on what can only be best done within a physical precinct. The purpose of precinctsThe answers to the basic question about the particular function of precincts were summarised in seven propositions. These were that precincts should: achieve transformative and sustained excellence in science, industry and economic outcomes; provide “living laboratories” and arenas for innovation through embedded practice; set a challenge that hasn’t been set before; shape emerging value chains (accelerating the change to new models) and support innovations across these value chains; shift emphasis from the physicality of place to a sense of social belonging and community; connect and cross-fertilise ideas drawn from deep expertise from different areas of knowledge to resolve complex challenges needing significant economies of scale and scope; and enable smart people to be in close proximity and develop an ecology that keeps them engaged. What can we learn from our experience?A panel of people with experience of a cross-section of precincts from around Australia opened up this discussion. It would be fair to say that, essentially, the overall verdict was that most precincts are currently collocations of convenience, often prompted by the desire for access to, or to support, large-scale facilities or infrastructure. Such collocation typically translates into little else. The implication is that “more of the same” is not going to change things or unleash the potential of significant investments. This led into a productive discussion of how we might describe the critical specifications for a successful precinct. The resulting summary checklist of the views of participants shown in Box 1 is in no particular order of priority. It was generally agreed that there is no “one size fits all” model, but rather a spectrum of models and approaches, shaped around different priorities and emphases. This was illustrated by one discussion group as shown in Box 2. Precinct issues and challengesEight particular challenges were identified as demanding close attention and further work. The process for engaging stakeholders and the community. Calibrating meaningful success factors (including benefits to the community), involving: Performance indicators and milestones; A focus on outcome measures; and Assuring adequate and appropriate contributions by participants. Articulating the governance options on a “horses for courses” basis. Delivering the “curatorial” role in precinct organisation and evolution. How to embed “openness” and inclusiveness in all their dimensions (and noting the tension with focus and differentiation). Avoiding the risk of precincts becoming silos in themselves, and thus inward looking. Articulating the characteristics of a good “precinct leader”. Assuring recurrent as well as capital investment, in order to avoid the brass plaque syndrome which leaves institutions with inadequate ongoing resources. Barriers to progressIn elaborating on these issues, the symposium participants identified and highlighted a number of barriers to progress. These barriers included: risk aversion, inflexibility, lack of ambition, and conflicting paradigms; asymmetric power structures and relationships; stakeholder fear of brand dilution within partnerships and collaborations; use of traditional and inappropriate metrics to represent outcomes; lack of scale and critical mass; the unintended consequences of government regulatory and planning constraints; and the negative impact of perverse incentives. In addressing these issues various participants stressed the need to: appreciate the evolutionary options and pathways for progressive precinct development; recognise the different motivations in play and to accommodate them within inclusive incentive structures; develop new metrics, such as the use of social network analysis to map emerging outcomes and relationship opportunities;5 think big and not be under-ambitious; and provide for transient access and engagement. It was noted that the expertise from business schools, which routinely deal with governance structures and the building of innovative teams, was an underused resource in the discussion of precincts. One table group summarised the enabling and constraining factors as shown in Box 3. Other important issues that were raised included the need to consider precincts in the wider context of the overall ecosystem in which they interact, and the importance of managing the accelerating pace of change so important to their development. We need to build flexibility and adaptability into our systems, to ensure we have appropriate ways of measuring the success of collaborations and to avoid creating silos within precincts. Essentially, we need to be in the business of building new models of collaboration and embedded practice, all within an end-user environment. These need to be unashamedly focused on community outcomes and with community engagement at all levels of the process. At the conclusion of the symposium, distinguished IBM Fellow and Alumnus Nick Donofrio, outlined some challenges which we ignore at our peril. These are listed in Box 4. A number of exciting new bioscience hubs will be developed in Australia over the next few decades. Innovation precincts are emerging around areas such as next-generation manufacturing, marine resources and tropical industries, and it is to be hoped that the discussions reported here will provide some guidance as to how to get the best out of each particular precinct. 1 Summary of critical specifications for a successful precinct as developed by participants at the “Precincts, people and places — forging new partnerships” symposium A clearly articulated and understood mission, with short-term, medium-term, and long-term goals. This needs to be a shared vision, and one that is “owned” by all the interested parties. The tension between what is achievable in the short term and stretch targets was noted. The vision needs “fit for purpose” governance arrangements to sustain it. Shared cultural values around excellence, openness, adaptability, risk-taking and risk-sharing, tolerance for failure, and generosity. Strong and self-assured leadership with “fit for purpose” management and enabling skills. Clear “cross-fertiliser” roles, possibly involving: a curatorial function; funding compacts and incentives; and mobilising functions for program leaders. Clear terms of engagement around resources and roles. An engagement strategy for stakeholders and the community. Pulling power through being on the global radar (and developed through brand, marketing, and communications). A high concentration of appropriate expertise and standing. The capacity to deal with complexity and to ensure diversity of contributions from the major partners in the precinct. An embedded educational role (and learning capacity). Industry connections, connectivity and relevance. Porous, permeable boundaries which promote a clear focus on market and outcome domains (not institutional domains). Incentives geared to outcomes (not inputs). “3D” use of land and space (through tiered rather than linear occupancy categories, with more than one thing in the one space). Shared and creative “play spaces”. Time-based and sequenced development priorities, enabling evolutionary pathways. Sufficient scale to be able to accommodate risk. 2 The spectrum of available precinct models as illustrated by one discussion group at the “Precincts, people and places — forging new partnerships” symposium In this model, the precinct model depends on the intended goal to be achieved. Simple collocations may be appropriate when only critical mass is required, whereas integration of entities may be needed to address issues of coordinated focus. Between these are precincts which either share services to drive limited resources further, or share branding to create momentum around a common vision. 3 Enabling and constraining factors in building and sustaining precincts 4 Challenges discussed at the “Precincts, people and places — forging new partnerships” symposium by IBM Fellow and Alumnus Nick Donofrio Leadership is what it is all about — and that leadership needs to be real and hopeful. Look to developing 21st century models — look for how and where the physical adds value over the virtual to create new hybrid models. Remember that money is not everything — too much may be as bad as too little — and scarcity can help to drive innovation. Precincts need to be dynamic — activities should be allowed to expand and contract. For the right task, at the right time, and within the right environment, the talent will come. A collection of things is important, but to make precincts really work we need a system that works as well.

Terry Cutler PhD, FTSE, FAIM · Michael Still MBA · James B Moody PhD, BEng(Elec), BInfoTech · Peter M Brooks MD, FRACP, FAFPHM

Objectives: To examine trends and predictors of prescription medications on discharge after first (index) hospitalisation for heart failure (HF), and the effect on all-cause mortality of evidence-based therapy.Design: A retrospective multicentre cohort study, with medical record review.Setting: Three tertiary-care hospitals in Perth, Western Australia.Patients: WA Hospital Morbidity Data were used to identify a random sample of 1006 patients with an index admission to hospital for HF between 1996 and 2006.Main outcome measures: Proportion of patients prescribed evidence-based therapy for HF on discharge from hospital; and 1-year all-cause mortality.Results: Among 944 patients surviving to hospital discharge, the prescription rate of angiotensin-converting enzyme (ACE) inhibitors/angiotensin receptor blockers (ARBs) (74.3%) and loop diuretics (85.5%) remained high over the study period, whereas that of β-blockers and spironolactone increased (10.5% to 51.3% and 1.4% to 23.3%, respectively), and digoxin prescription decreased (38.1% to 20.7%). The temporal trends in use of β-blockers, spironolactone and digoxin were in line with clinical trial evidence. Age ≥ 75 years was a significant, negative predictor of β-blocker and spironolactone prescription. In-hospital echocardiography, performed in 53% of patients, was associated with a significantly greater likelihood of treatment with ACE inhibitors/ARBs, β-blockers and spironolactone. Both ACE inhibitors/ARBs and β-blockers prescribed on discharge were associated with a lower adjusted hazard ratio (HR) for mortality at 1-year (HR, 0.71; P = 0.003; and HR, 0.68; P = 0.002, respectively).Conclusion: ACE inhibitors/ARBs and β-blockers, prescribed during initial hospitalisation for HF, are associated with improved long-term survival. Therapy became more evidence based over the study period, but echocardiography, an important predictor of evidence-based therapy, was underutilised.

Tiew-Hwa Katherine Teng MPH · Joseph Hung MB BS(Hons), FRACP, FACC · Judith Finn RN, PhD, FRCNA

Aim: To determine the prevalence and causes of vision loss in Indigenous Australians.Design, setting and participants: A national, stratified, random cluster sample was drawn from 30 communities across Australia that each included about 300 Indigenous people of all ages. A sample of non-Indigenous adults aged ≥ 40 years was also tested at several remote sites for comparison. Participants were examined using a standardised protocol that included a questionnaire (self-administered or completed with the help of field staff), visual acuity (VA) testing on presentation and after correction, visual field testing, trachoma grading, and fundus and lens photography. The data were collected in 2008.Main outcome measures: VA; prevalence of low vision and blindness; causes of vision loss; rates of vision loss in Indigenous compared with non-Indigenous adults.Results: 1694 Indigenous children and 1189 Indigenous adults were examined, representing recruitment rates of 84% for children aged 5–15 years and 72% for adults aged ≥ 40 years. Rates of low vision (VA < 6/12 to ≥ 6/60) were 1.5% (95% CI, 0.9%–2.1%) in children and 9.4% (95% CI, 7.8%–11.1%) in adults. Rates of blindness (VA < 6/60) were 0.2% (95% CI, 0.04%–0.5%) in children and 1.9% (95% CI, 1.1%–2.6%) in adults. The principal cause of low vision in both adults and children was refractive error. The principal causes of blindness in adults were cataract, refractive error and optic atrophy. Relative risks (RRs) of vision loss and blindness in Indigenous adults compared with adults in the mainstream Australian population were 2.8 and 6.2, respectively. By contrast, RRs of vision loss and blindness in Indigenous children compared with mainstream children were 0.2 and 0.6, respectively.Conclusion: Many causes of vision loss in our sample were readily avoidable. Better allocation of services and resources is required to give all Australians equal access to eye health services.

Hugh R Taylor AC,MD, FRANZCO · Jing Xie PhD · Sarah Fox BA · Ross A Dunn BAppSc(AppChem), GradDipBIT · Anna-Lena Arnold BSc · Jill E Keeffe OAM, PhD

Objective: To test the effectiveness of using a full-time project nurse to assist residential aged care facilities in using evidence-based approaches to falls injury prevention.Design, setting and participants: Cluster randomised controlled trial involving 5391 residents in 88 aged care facilities in the Hunter and Lower Mid North Coast areas of New South Wales. Residents were followed for 545 days or until death or discharge. Data were collected from July 2005 to June 2007.Intervention: Employment of a project nurse to encourage best-practice falls injury prevention strategies during the 17-month intervention period.Main outcome measures: Monthly data about falls, falls injury and falls injury prevention programs; audit of hospitalisation for fractured neck of femur.Results: Despite significant increases in the provision of hip protectors and use of vitamin D supplementation in both intervention and control facilities, there was no difference in the number of falls or falls injuries between the intervention and control groups, nor a reduction in falls overall. There was also no difference between the 7-month pre-intervention period and the intervention period in the number of falls or falls injuries. Factors related to residents having an increased risk of falls with fractured neck of femur included being ambulant, having dementia, increasing age, and having a high falls risk assessment score.Conclusion: It is difficult to change falls risk among high-risk populations, including people with dementia. The use of important strategies such as hip protectors and vitamin D and calcium supplementation increased during the study, probably with contamination of control facilities. Longer follow-up may be required to measure the impact on falls outcomes of the strategy of using a facilitating nurse.Trial registration: Australian New Zealand Clinical Trials Registry ACTRN12605000540617.

John A Ward MB BS, FRACP MSc(CommHealth) · Mandy Harden BA, GradDipEd · Richard E Gibson BSc, DipEd, DipMedStats · Julie E Byles BMed, PhD

Objective: To describe the management and outcomes of a population-based cohort of patients with pancreatic cancer treated with chemotherapy or radiotherapy in Victoria, Australia.Design, setting and patients: Questionnaire-based study of patients diagnosed with pancreatic cancer during 2002–2003 in Victoria who were retrospectively identified from the Victorian Cancer Registry and followed up for a minimum of 5 years.Main outcome measures: Reported treatment, referral patterns and survival rates.Results: 1044 patients with pancreatic cancer were identified, of whom 927 were eligible for the study. Completed questionnaires were obtained for 831 eligible patients (response rate, 89.6%) and data for 66 patients with tumours of the ampulla of Vater and neuroendocrine tumours were excluded. Of the remaining 765 patients, 6.5% were managed in multimodality clinics. Chemotherapy was considered for 413 patients and radiotherapy was considered for 162. One-third of the cohort (275 patients) received chemotherapy, most commonly as palliative treatment (185). Single-agent gemcitabine was the most common palliative treatment (154), and was associated with a median overall survival of 6.6 months. Radiotherapy was used in 119 patients (15.6% of the cohort) — it was used alone or with chemotherapy, as postoperative adjuvant treatment, as potentially curative radical treatment, or as palliative treatment. For 45 patients with locally advanced disease who were treated with chemoradiation as radical treatment, median overall survival was 13.1 months.Conclusions: There appears to be under-referral of patients to medical and radiation oncologists. Median survival of patients treated with radical chemoradiation or palliative chemotherapy is consistent with clinical trial data, but outcomes for patients in our cohort were generally poor. Development and implementation of treatment guidelines may result in improved outcomes.

Michael Jefford MB BS, PhD, FRACP · Vicky Thursfield BSc, GradDipApplStats · Yvonne Torn-Broers BA(Hons) · Trevor Leong MB BS, MD, FRANZCR · Mario Guerrieri MB BS, FRANZCR · Tony Speer BE, MB BS, FRACP

Objective: To conduct a descriptive study investigating the effect of access to motor vehicle accident (MVA) compensation on recovery outcomes at 24 months after injury.Design and setting: Longitudinal cohort study conducted in two Level 1 trauma hospitals in Victoria, Australia. Participants were 391 randomly selected injury patients with moderate-to-severe injuries. Compensable and non-compensable patients were compared at 24 months after injury on a number of health outcomes.Main outcome measures: Health outcomes at 24 months, including anxiety and depression severity, quality of life and disability.Results: Medical records identified two groups of compensation patients: MVA-compensable and non-compensable patients. After controlling for baseline variables, the MVA-compensable patients, at 24 months, had higher levels of post-traumatic stress disorder, anxiety and depression, and were less likely to have returned to their pre-injury number of work hours. However, some patients in the non-compensable group had accessed other forms of compensation (eg, private health care or compensation for victims of crime). When these were removed from the non-compensable group, the differences between MVA-compensable and non-compensable groups all but disappeared.Conclusion: Our findings do not support previous research showing that access to compensation is associated with poor recovery outcomes. The relationship between access to compensation and health outcomes is complex, and more high-level research is required.

Meaghan L O’Donnell BSc(Hons), MA(Clin), PhD · Mark C Creamer BA(Hons), MA(Clin), PhD · Alexander C McFarlane MMed(Psych), FRANZCP · Derrick Silove MMed(Psych), FRANZCP · Richard A Bryant BA(Hons), MA(Clin), PhD

We use vertebroplasty for patients with the most severe pain caused by osteoporotic vertebral fractures less than 6 weeks old, and have observed dramatic pain relief in this acute setting. A recent editorial in the Journal, written by the authors of two recent vertebroplasty trials, suggested that vertebroplasty is not an effective therapy for acute osteoporotic vertebral fractures. The trials described in the editorial sampled a very different patient cohort to the one that we treat with vertebroplasty. Our clinical experience and most of the published literature relating to the benefits of vertebroplasty are in striking contrast to the opinions presented in that editorial.

William A Clark MB BS, FRANZCR · Terrence H Diamond MB BS, MB BCh, FRACP · H Patrick McNeil MB BS, FRACP, PhD · Peter N Gonski MB BS, BMedSci, FRACP · Glen P Schlaphoff MB Bch, FCRad(SA), FRANZCR · John C Rouse MB ChB, FRANZCR

Our recent editorial in the Journal presents an accurate summary of our two randomised trials of vertebroplasty, which found no benefit of vertebroplasty over placebo. Participants in both trials are representative of patients seen in clinical practice and who would qualify for government-subsidised funding of vertebroplasty in Australia. Clinical experience and previous published literature are likely to have overestimated the treatment benefit of vertebroplasty for many reasons. This is why randomised placebo-controlled trials are required to determine the efficacy of treatment interventions, particularly when the condition being treated is self-limiting and the primary end point is improvement of symptoms. Based on the best evidence currently available, the routine use of vertebroplasty outside of the research setting for painful osteoporotic vertebral fractures appears unjustified.

Rachelle Buchbinder MB BS(Hons), PhD, FRACP · Richard H Osborne BSc, PhD · David Kallmes MD

Assessment of medical fitness to drive can be a sensitive and difficult task, particularly when it involves a condition such as epilepsy, where impairment is intermittent. The patient, their doctor and the driver licensing authority (DLA) each have responsibilities, both to the patient and to the wider community of road users. DLAs in Australia have shifted most of the responsibility for determining fitness to drive to the treating doctor. This creates a conflict of interest and may lead to unsafe decisions, damage to the doctor–patient relationship, interference with medical management and legal vulnerability for the doctor. Australian neurologists have argued for a system in which the treating doctor provides objective information about the patient’s condition, rather than an opinion on fitness to drive, and the DLA uses that information to determine fitness. This must be supported by an expert review process. Although drivers are legally obliged to notify the DLA when they become unfit, most people are unaware of this law. However, passing this responsibility to doctors in the form of mandatory reporting is counterproductive to road safety.

Ernest R Somerville MB BS, FRACP, FRCP · Andrew B Black BMedSc, MB BS, FRACP · John W Dunne MB BS, FRACP

Clinical record Patient 1 A 90-year-old man with bilateral neovascular age-related macular degeneration (AMD) presented with sudden onset of painful visual loss in his left eye. Visual acuity in his right eye was 1/60 due to scarring associated with AMD, and visual acuity in his left eye had deteriorated to hand movements; it had previously been 6/120. He had been taking warfarin for about 2 years for atrial fibrillation and a transient ischaemic attack. The international normalised ratio (INR) had previously been measured almost weekly, but as it had been stable at 2.5–2.9 (target INR, 2.0–3.0), it had not been measured for 6 weeks. B-mode ultrasonography and dilated fundus examination revealed massive suprachoroidal haemorrhage (Figures A and B). The INR was 6.0, which was reversed with intravenous administration of two units of fresh frozen plasma and 1 mg of vitamin K. Visual acuity in the patient’s left eye deteriorated to light perception. As there were no signs of spontaneous improvement and this had been his better eye, surgery was performed to drain the haemorrhage. After surgery, the haemorrhage decreased in size but his vision did not improve. Patient 2 An 82-year-old man presented with a 3-week history of a shadow in his left eye. He had been taking low-dose aspirin for ischaemic heart disease and stroke, but had no known ocular history. On examination, visual acuity was 6/5 and hand movements in his right and left eye, respectively, and massive subretinal haemorrhage was noted in the left eye. Non-neovascular AMD was present in his right eye, and the haemorrhage in his left eye was presumed to be due to a combination of neovascular AMD and the effects of aspirin. He was offered surgery, but declined as there was little chance of improvement in central vision. Subsequently, he gradually lost all vision (including light perception) in his left eye. Aspirin was continued because of his significant cardiovascular history. Over the following year, neovascular AMD developed in his right eye and visual acuity dropped from 6/5 to 6/36 (Figure C). He was treated with regular intravitreal ranibizumab injections, and his vision stabilised. However, several months later, vision in his right eye suddenly deteriorated to vague perception of light. On examination of his right eye, there was no view of the fundus due to a dense cataract and vitreous haemorrhage, but B-mode ultrasonography once again revealed massive choroidal, subretinal and vitreous haemorrhage. The patient was still taking low-dose aspirin therapy, which was subsequently ceased. In an attempt to preserve vision in his only seeing eye, he underwent vitrectomy and silicone oil insertion, but there was no improvement and he subsequently lost light perception in his right eye. A: B-mode ultrasonogram of Patient 1’s left eye, showing massive suprachoroidal haemorrhage. B: Fundus photograph of Patient 1’s left eye, showing massive suprachoroidal haemorrhage bulging forward (therefore largely out of focus). C: Fundus photograph of Patient 2’s right eye, showing right neovascular AMD with subretinal haemorrhage, prior to development of massive suprachoroidal, subretinal and vitreous haemorrhage. Age-related macular degeneration affects about one-third of people aged over 75 years.1 Of those with AMD, 10%–15% develop the neovascular (“wet”) form,2 characterised by abnormal new blood vessel formation in the choroid, under the retina. These abnormal vascular membranes are prone to rupture, leading to subretinal bleeding, fibrous scar formation and severe visual loss. Many older patients who have AMD also take medications that can exacerbate or promote bleeding, such as anticoagulants or antiplatelet agents. In rare cases, intraocular bleeding — in the form of subretinal, suprachoroidal, or vitreous haemorrhage — can be catastrophic and blinding. Previous reports link systemic anticoagulation therapy to intraocular haemorrhage and blindness in AMD patients,3-7 including a recent report in this Journal.8 In two of these reports, patients taking warfarin had very high INRs (4.1 in one case;3 6.3 in the other4). Other reports link systemic anticoagulation therapy to spontaneous suprachoroidal haemorrhage, even in the absence of neovascular AMD.9-11 Additionally, patients with neovascular AMD can develop massive submacular haemorrhage, even if they are not taking antiplatelet or anticoagulant agents. Unfortunately, as many patients with AMD have one eye with poor visual acuity due to macular scarring, it is all the more catastrophic when a massive haemorrhage leads to blindness in their “good” eye. We suggest that the risk of catastrophic bleeding may be stratified, with the greatest risk being associated with anticoagulants (eg, warfarin), followed by antiplatelet agents (eg, clopidogrel and ticlopidine, with aspirin conferring a lower risk). In addition, we would expect a greater risk with combination therapy comprising simultaneous use of multiple antiplatelet and/or anticoagulant agents. In the cases reported here, we suggest that haemorrhage is likely to have been more severe and blinding than it would have been in a patient who was not taking those medications. Although these medications should not be withheld in cases where they are warranted to reduce cardiovascular morbidity and mortality, they are not without risk, and a careful risk–benefit analysis should be performed for each patient before they are prescribed. For example, the risk of stroke is different among patients with simple atrial fibrillation compared with aortic valve replacement. It has been previously noted that if a patient has only one functioning eye, the patient’s general practitioner or cardiologist should seek an ophthalmologist’s opinion to assess the risk of neovascular AMD in the seeing eye before, or soon after, commencing warfarin.12 Further, ophthalmologists should ask their patients whether they take warfarin, and should communicate to the treating doctor whether a patient has, or is at high risk of developing, neovascular AMD.12 Patients taking warfarin who develop neovascular AMD should be advised to maintain an INR at the lower end of the recommended range.3 In addition, we recommend that it would be prudent for INR monitoring to be at the more frequent end of the spectrum. Lessons from practice Patients taking warfarin who develop neovascular age-related macular degeneration (AMD) should maintain an international normalised ratio (INR) at the lower end of the recommended range. INR monitoring should be more frequent for patients with neovascular AMD. If a patient has only one functioning eye, an ophthalmologist should assess the risk of developing neovascular AMD in the seeing eye before, or soon after, commencing warfarin or an antiplatelet agent, including aspirin. Ophthalmologists should ask all patients whether they take warfarin, and should communicate to the treating doctor whether this patient has, or is at high risk of developing, neovascular AMD.

Helen M Garrott FRANZCO,MB BS(Hons), BMedSc · Richard J Haynes MB BCh, FRCOphth, MD

Haemophilus ducreyi is a well recognised causative agent of genital ulcers and chancroid. We report two unusual cases of non-sexually transmitted H. ducreyi infection leading to chronic lower limb ulcers. Both patients were Australian expatriates visiting Australia from the Pacific Islands — one from Papua New Guinea and the other from Vanuatu. Clinical recordsPatient 1A 51-year-old Australian man presented with a 2-month history of non-healing ulcers on the lateral aspect of his lower left leg. The ulcer developed subsequent to a dog bite received in Papua New Guinea. The wound was treated initially in Papua New Guinea with debridement and oral doxycycline and the wound resolved. Subsequently, the patient developed another ulcer adjacent to the original dog bite wound. The ulcer became increasingly painful and erythematous, and was associated with a purulent discharge. On return to Australia, he was referred to a plastic surgery unit. The patient underwent debridement of the ulcer and split skin grafting to cover the defect. He was prescribed flucloxacillin (500 mg orally every 6 hours) after surgery, but had further ulceration of the skin graft. Tissue samples were cultured and examined histologically. Subsequently, a gram-negative coccobacillus was isolated on chocolate horse blood agar incubated at 35°C in CO2. The cultured organism was sent to a reference laboratory for further identification, and a presumptive diagnosis of Haemophilus ducreyi infection was made from the characteristic “schools of fish” morphology on staining (Box 1); this was confirmed with 16S rDNA (ribosomal DNA) sequencing. The patient was referred to an infectious diseases service. On review, the wound was 2 × 1.5 cm, with a sloughy base, rolling edges and surrounding erythema (Box 2). He had no risk factors for sexually transmissible infections and had no evidence of genital lesions or inguinal lymphadenopathy. He did not undergo testing for sexually transmitted infections. He was prescribed oral azithromycin 1000 mg, with significant improvement in the ulcer appearance. The patient returned to Papua New Guinea and was subsequently lost to follow-up. Patient 2A 60-year-old Australian man who resided in Vanuatu presented to an infectious diseases unit with a 6-week history of a painful ulcer on the back of his lower leg. The ulcer developed following a visit to a jungle area of Vanuatu; he did not recall any prior injury. Before presentation, he had received dicloxacillin and metronidazole and had undergone regular iodine dressings for several weeks without clinical improvement. At review at the infectious diseases unit, the most striking feature was the severe degree of pain associated with the lesion. On clinical examination, the ulcer was 3–4 cm in diameter, with heaped-up necrotic edges and a sloughy base. The patient was not systemically unwell and had no risk factors for sexually transmissible infections. A Gram stain of tissue obtained from biopsy and swabs of the ulcer showed numerous neutrophils and a large number of gram-negative coccobacilli; an organism was isolated from the swab and from the biopsy specimen on chocolate agar incubated at 35°C in CO2. The reference laboratory identified the organism as H. ducreyi using 16S rDNA sequencing. The patient was treated with intravenous ceftriaxone 1 g daily for 1 week, followed by amoxycillin 1 g orally thrice daily for 2 weeks, and the size and associated pain of the ulcer decreased rapidly. He did not report any genital lesions and was not tested for sexually transmissible infections. The patient returned to Vanuatu but reported complete resolution of the ulcer to the treating physician. DiscussionChronic skin ulcers secondary to H. ducreyi infection in the absence of genital lesions, as in our two cases, have not been reported until recently. H. ducreyi is a fastidious gram-negative coccobacillus. The association between this organism and chancroid, a sexually transmissible genital ulcer disease, is well established.1,2 Extragenital skin ulcers have been described resulting from autoinoculation among patients with genital ulcers.3,4 Human volunteers have developed mildly painful skin lesions following inoculation with H. ducreyi.5 Humans are the only known host for H. ducreyi; there is no recognised environmental or zoonotic reservoir.1,6 Apart from our two reported cases, to our knowledge, only two other publications have described chronic skin ulcers secondary to H. ducreyi infection.6,7 Box 3 summarises these cases and the cases we have described here. In all cases, the patients had chronic skin ulcers on lower limbs without evidence of genital ulcers or inguinal lymphadenopathy. All of these patients acquired the infection in the Pacific Islands.6,7 In most cases, the organism was cultured on selective media (eg, chocolate agar) and identified using 16S rDNA sequencing techniques.6,7 H. ducreyi is difficult to isolate from culture owing to several factors. H. ducreyi grows best between 33°C and 35°C in 5% CO2 and a humid atmosphere, and will not grow at 37°C. Enriched media, such as chocolate horse blood agar, are required for the organism’s growth. Microbiological specimens must be inoculated as quickly as possible to ensure organism survival.1 Upon Gram staining, H. ducreyi aligns in a “railroad”, “chaining” or “schools of fish” (Box 1) arrangement. Definitive identification of H. ducreyi requires specific molecular tests, such as 16S rDNA sequencing.1 Recommended treatment for H. ducreyi infection includes oral or intravenous azithromycin, ceftriaxone or ciprofloxacin.8 Plasmid-mediated β-lactamase production by this organism is well recognised; in a Californian report, 88% of H. ducreyi isolates produced plasmid-mediated β-lactamase.9 However, antimicrobial susceptibilities vary geographically. This may account for the response to penicillin or amoxicillin in three of the six cases reported thus far. In summary, H. ducreyi may be an emerging aetiological agent in chronic skin ulcers in patients from the Pacific Island region. Given its fastidious nature, a high degree of suspicion is required to enable specific diagnosis. This organism should be considered in patients from this region who have non-healing ulcers without a causative organism isolated by normal culture methods. Consideration should be given to culture swabs and tissue on enriched media, such as chocolate agar, and for the use of other diagnostic tools, such as 16S rDNA sequencing. 1 Gram stain appearance of ulcer showing characteristic “schools of fish” appearance 2 Appearance of lower limb ulcer showing wound ulceration at edge of split skin graft 3 Six cases of chronic skin lesions due to Haemophilus ducreyi infection Report Patient age (years), sex Country of acquisition Presentation Culture results Confirmatory diagnosis Treatment Outcome Ussher et al6 9, female Samoa Three chronic skin ulcers on lower legs Gram-negative coccobacilli on chocolate agar 16S rDNA Penicillin Not reported 6, female Samoa Single painful ulcer on lower leg Gram-negative coccobacilli on supplemented GC agar base 16S rDNA Oral azithromycin Complete resolution 5, female Samoa Three painful ulcers on lower leg Gram-negative coccobacilli on chocolate agar 16S rDNA Oral azithromycin Complete resolution McBride et al7 23, female Vanuatu Single non-healing ulcer on lower leg No organism cultured 16S rDNA Intramuscular benzathine penicillin Complete resolution Current report 51, male Papua New Guinea Single non-healing ulcer on lower leg Gram-negative coccobacilli on chocolate horse blood agar 16S rDNA Oral azithromycin Clinical improvement 60, male Vanuatu Single painful non-healing ulcer on lower leg Gram-negative coccobacilli on chocolate agar 16S rDNA Intravenous ceftriaxone then oral amoxycillin Clinical improvement GC = gonococcal. rDNA = ribosomal DNA.

Trisha N Peel MB BS, FRACP · Deepak Bhatti MB BS · Jim C De Boer BSc, MPH, FAIMS · Ivan Stratov MB BS, FRACP, PhD · Denis W Spelman FRACP, FRCPA

To the Editor: In their article, Couzos and Delaney Thiele raise many good points with respect to Medicare Australia’s Practice Incentives Program (PIP) Indigenous health incentive.1 The funds in question are part of the “closing the gap” spending by the Rudd government. However, the central question is whether spending many millions of dollars of this allocation to add to the income of general practitioners (through a patient enrolment program and the generation of “care plans”) will in fact convert into greater access to, and greater utilisation of, health services by the Aboriginal and Torres Strait Islander population (hereafter referred to as the “Aboriginal” population). The authors correctly point out that most practices in Australia do not treat Aboriginal patients, and that the uptake of targeted, extended-primary-care items by Aboriginal people is much lower than in the general population.1 In most areas, apart from the most remote, the problem is not a lack of services, but the red tape that blocks Aboriginal patients from taking advantage of health services. Adding further layers of red tape, such as patient enrolment, care plans, and health assessments, will only make it harder for Aboriginal patients to access extra health care. For example, an allied health service, normally accessed by walking in off the street, requires a care plan plus a team care plan. Such red tape only adds to the burden of compliance — the problem that lies at the heart of the reason why so many Aboriginal patients fail to meet basic health outcomes. From an economic viewpoint, one needs to ask what could be done with the money that will go to the health provider for administration, rather than for actual clinical care. For example, paying a GP $500 to enrol a patient could instead pay for a significant amount of dental work, speech pathology, diabetes education or physiotherapy. The Department of Veterans’ Affairs (DVA) Gold Card offers an efficient, simple and highly efficacious model that would serve the Aboriginal population a lot better.2 Under the DVA model, doctors are paid a modestly higher rebate for treating veterans (or their families). However, the real benefit for DVA Gold Card holders lies in their ability to access an expanded pharmaceuticals scheme, a comprehensive range of allied health and medical equipment, free patient transport, and private hospital care. Under the DVA, such benefits are accessed with minimal paperwork for the referring doctors and patients. Therefore, the DVA model, in contrast to the PIP Indigenous health incentive model, better targets funds towards service delivery. It is time for politicians and the Department of Health and Ageing to adopt a DVA-style model for the Aboriginal population.

Aniello Iannuzzi

To the Editor: We read with interest Walters and Dawson’s call for a clinical revolution to tackle access block1 and are heartened by the interest shown by general physicians in a problem that primarily affects the emergency department (ED). The efficient management of admitted medical patients is paramount to patient flow within the hospital, and “buy-in” from general physicians is essential. When considering any new model of care, it is important to note that longer patient assessments in ED by emergency doctors has a relatively small effect on access block; the claim that the length of assessments is a significant factor in access block has been established as a “myth” by investigators who have mapped process times.2 Therefore, it is unlikely that substituting one workforce of acute physicians for another would make any difference to overall patient flow through the ED. On the contrary, it is likely to be associated with increased costs3 and adverse effects on the emergency medicine labour supply.4 In addition, the ability and willingness of the general physician workforce to implement and sustain the newer role of “acute physician” is unknown. The root cause of access block lies in ward-bed shortages, ward processes and community capacity, which should be solved by improved flow processes across the continuum of care. Access block will not be solved by a second tier of acute physicians duplicating the role of emergency physicians. However, there are many aspects of Walters and Dawson’s model of change that would improve patient flow, in particular: improved rostering of medical staff; improved access to pathology and radiology services; and, perhaps, specific retraining of medical staff in the efficient discharge of inpatients. These aspects should be rigorously explored as we strive together to tackle access block.

Biswadev Mitra · Peter A Cameron · Pieter De Villiers Smit

To the Editor: Walters and Dawson1 correctly highlight access block (hospital overcrowding) as a whole-of-system problem. Acute medical assessment and admission units (AMAAUs), or other similar incarnations in Australia and New Zealand, are part of the solution, although the evidence presented is low-level non-Australasian data. The reduced length of stay achieved by these units and reported in papers cited by Walters and Dawson would, if replicated in Australasia, produce additional capacity, improving bed availability and patient flow. As has been repeatedly stated: it’s all about available beds!2 However, Walters and Dawson’s article stretches well beyond the evidence in its approach to emergency department (ED) roles and the interactions between AMAAUs and EDs. None of the cited studies suggested that AMAAUs provide better environments than EDs for sick undifferentiated patients. None studied effects that AMAAUs have on ED treatment and none proposed interventions specifically designed to alter ED management. They essentially examined improved patient journeys for front-loaded AMAAU versus standard (slower) general medical inpatient care. In addition, Walters and Dawson imply that these changes related to introduction of the United Kingdom’s 4-hour rule. However, many references were either non-UK or not specific to the 4-hour rule. Rigorous research in an Australasian context would be required before adopting models from a different system. The authors promote a view that undifferentiated acutely sick patients bypass the ED to be managed by “new” acute-care specialists. No evidence is presented to support this change, and it is difficult to see how this would be a sensible policy for Australia and NZ, which have mature ED systems. Emergency physicians are specialists specifically trained and skilled in early diagnosis, management and disposition of the undifferentiated, unwell patient. What is required is a system that builds on the excellent start made by the ED, removes the blocks to patient care caused by waiting for beds in the ED and then continues to emphasise rapid diagnosis, early management, disposition and flow. This is what AMAAUs can deliver and why they should be effective. Australasian EDs already provide an exemplary service in a difficult, access-blocked environment. What patients need is sufficient hospital capacity — hospitals that provide enough appropriate beds. We look forward to seeing AMAAU staff meet this need in partnership with their emergency physician colleagues. In summary, it’s all about available beds, about having enough overall capacity and optimising patient flow to maximise bed availability.2 The only revolution required is for governments to recognise this fundamental precept.

David Mountain · Daniel M Fatovich · Drew B Richardson · Sally M McCarthy

To the Editor: Although written from a United Kingdom perspective, the recent article by Walters and Dawson1 suggests a change in the clinical culture within hospitals, so that patient care and throughput can be improved. A critical factor in achieving change is the creation of an acute medical assessment and admission unit (AMAAU) within each district hospital. Characteristics of the AMAAU will “depend on local circumstances”: there is no one size that fits all.1 Because of Australia’s unique demography, and the number of communities beyond the reach of tertiary centres, many primary-care physicians (general practitioners and “rural generalists”) provide the continuum of care required by patients, both within the community and within their local hospitals (the acute admission, ongoing inpatient care and discharge planning). Twenty-first-century GPs deal daily with patients needing management of multiple comorbidities and the consequences of polypharmacy (the “sick general” and “complex elderly” clinical streams1). GP training prepares doctors for these responsibilities and could easily be expanded to include an AMAAU role for interested GPs, especially those in outer urban and major rural areas. The advent of AMAAUs is an opportunity to change the mindset in medicine: after 8 to 9 years of primarily hospital-based training, some GPs suddenly have no hospital access! This would seem to be a callous waste of talent and resources.

Frank R Jones

To the Editor: Walters and Dawson1 highlight growing interest in new models of care aimed at ameliorating hospital-bed pressures and access block. They advocate acute medical assessment and admission units (AMAAUs) as a potential solution, and claim, principally based on the United Kingdom’s experience, that these units can significantly improve clinical care and patient outcomes. A recent systematic review confirms that these units (which have attracted several different synonyms) have promise, although controlled trials have yet to be performed, and publication bias remains a potential confounder.2 Experience with such units in Australia and New Zealand is growing, with more than 30 units in operation, and up to another 15 due to open over the next few years. Several national workshops conducted during the past 12 months have allowed staff of the units to share lessons and insights, and to debate how to balance service needs with resource availability. Operating standards for AMAAUs have been developed by the Internal Medicine Society of Australia and New Zealand (IMSANZ),3 which represents consultant general physicians. A recent survey shows the operations of Australasian units concord, in the most part, with these standards.4 We caution against Walters and Dawson’s suggested separation of AMAAU physicians into two streams — acute physicians working shifts, and ward-based general physicians responsible for patients requiring transfer from the AMAAU. Given that at least half of AMAAU patients will require transfer to inpatient wards, and many may warrant ongoing outpatient care even if discharged from the AMAAU, the need for continuity of care is paramount at the interface between the AMAAU and ward or clinic. To minimise the number of handovers and their attendant hazards and inefficiency, the medical team assessing and managing the patient in the AMAAU should ideally be the same team that provides ongoing inpatient (and indeed subsequent outpatient) care. This practice also eliminates any confusion around who is ultimately responsible for decisions about individual patient care, particularly for patients who remain in the AMAAU for any length of time. General physicians can acquire and maintain skills in acute medicine by making use of professional development programs sponsored by the IMSANZ. Clinical directors are needed in AMAAUs to oversee unit operations, develop policies and procedures, and provide capacity for rapid consultant response if on-call consultants are temporarily unavailable. The real challenge, to which Walters and Dawson refer, is the need for health care professionals to recognise that whole-of-hospital redesign solutions — which include AMAAUs — are needed, if access block in emergency departments is to be successfully overcome.

Ian A Scott · John W Henley

To the Editor: The Journal took a significant step forward in publishing the three articles on access block in the 6 April 2009 issue.1-4 Walters and Dawson’s viewpoint article,4 in a later issue, touches on some ideas that will be useful in finding solutions to access block — ideas that some hospitals are implementing. However, I am not sure a microsolution aimed purely at acute medical patients can be called a whole-of-hospital revolution. The acute medical assessment and admission unit (AMAAU) is potentially a good idea. Fortunately, many hospitals all over Australia already have units that are highly efficient at the role that is proposed for it — they are called emergency departments (EDs). Most acute medical patients can be identified as needing admission after a few seconds in the ED by experienced emergency physicians. The remaining patients need some basic pathology or imaging service before a decision can be made, which should take an hour at the most. Having secondary inpatient units providing this role to the community via direct general practitioner referrals, as well as having some patients bypassing the ED by being cherry-picked by inpatient teams, may generate inefficient duplications of service. The AMAAU has merit, streaming patients to the right specialty and the right inpatient bed early in their presentation. Emergency physicians have largely known this for over a decade and these kinds of units have already been introduced in hospitals all over the country. Nepean Hospital, in western Sydney, has the PECC (Psychiatric Emergency Care Centre), AGS (Acute Gynaecological Service), MAU (Medical Assessment Unit), EDMAU (ED Medical Assessment Unit), EMU (Emergency Medical Unit) and ASU (Acute Surgical Unit), to name just a few acronyms. Unfortunately, this does not deal with the 20 patients in the ED, already admitted and sorted, waiting for an inpatient bed at 8 am on a Monday. Increased inpatient bed numbers to cope with the predicted acute ED admissions and the planned elective surgical workload must be the number-one priority. Once we have bed numbers to cope with demand, then we can plan how to use them. I propose my own revolution. We need to provide a true 7-day-a-week service to our hospital inpatients. Ward rounds should be conducted 7 days a week. All inpatient consults, including those of allied health practitioners, should be completed on the same day, including weekends. All complex imaging should be completed on the day it is ordered, not the next working day, with formal reports available the same day. Once we acknowledge that acute hospital medicine does not fit in with the 38-hour working week, then we can truly start acting as patient advocates.

James L Mallows

To the Editor: The recent article by Chang and colleagues concluded that “the clinical course and outcomes of pandemic (H1N1) 2009 influenza virus are comparable to those of the current circulating seasonal influenza”, and that: “The high number of hospital admissions reflects a high incidence of disease in the community rather than an enhanced virulence of the novel pandemic influenza virus”.1 We are concerned that these assertions underemphasise the true severity of the influenza pandemic, and have led to inappropriate reporting in the media.2 The single-centre series reported by Chang et al had a small sample size and was almost certainly underpowered to detect important differences. It is also likely that some of the five untypeable patients who were categorised in the seasonal influenza group had had false-negative test results for pandemic (H1N1) 2009 influenza. Despite the small numbers, this study suggested that patients with pandemic (H1N1) 2009 influenza were younger and less immunocompromised than those with seasonal influenza; both of these characteristics of the patients affected may indicate that the pandemic virus is a more virulent strain. However, regardless of whether its virulence was greater, our significant concern is that our community will underestimate the real burden of the pandemic, which was substantial in Australia and New Zealand during the recent winter. In recent publications, we described more than 700 people admitted to intensive care units (ICUs) throughout these two countries with pandemic (H1N1) 2009 influenza.3,4 These were often young and previously healthy people, and many were pregnant women.3 Two-thirds needed mechanical ventilation for influenza-induced respiratory failure,3 and a smaller but substantial number developed rapidly progressive acute respiratory distress syndrome and required extracorporeal membrane oxygenation (ECMO),4 the most extreme life support available. This is not the normal pattern of influenza in Australasia. In comparison to a normal winter, ICU admissions for viral pneumonitis increased 15-fold,3 and the use of ECMO for acute lung injury increased 17-fold.4 Patients infected with pandemic (H1N1) 2009 influenza required prolonged stays in both the ICU and hospital and, despite optimal care, more than 100 died. ICU bed occupancy by patients with pandemic (H1N1) 2009 influenza ran as high as 19%3 in a system that normally runs close to maximal occupancy. There is a real risk that the pandemic will affect Australia again next winter or earlier, and we feel the Australasian medical community should not be misled into believing that the pandemic (H1N1) 2009 influenza virus is not virulent and has not been responsible for significant mortality and morbidity in a population not normally affected.

Andrew R Davies · Steven A Webb · Ian M Seppelt · Rinaldo Bellomo

In reply: Our study found that both subtypes of influenza caused substantial morbidity and mortality, but that there was no difference in outcomes between patients infected with seasonal and pandemic (H1N1) 2009 influenza.1 The large numbers of cases of pandemic (H1N1) 2009 infection, and especially those requiring intensive care unit (ICU) management, is not disputed — 37 537 cases, 655 ICU admissions and 191 deaths in Australia had been reported as of 18 December 2009.2 However, this is not a function of enhanced virulence of the pandemic strain, but rather a function of vast numbers of infected individuals in a naïve population. The article by the Australian and New Zealand Intensive Care study investigators supports our conclusions, as they found no difference in outcome between pandemic (H1N1) 2009 and seasonal influenza.3 As specimens were tested in “real time”, RNA degradation is unlikely to have resulted in incorrect categorisation. The real impact of influenza has probably been under-reported before this pandemic, as virological testing of respiratory specimens has not been routine. Influenza per se is a significant disease, and thus appropriate planning, including resource allocation for ICU management and routine virological testing of respiratory specimens should be undertaken.

Iain B Gosbell · Sebastiaan J van Hal · Peter M Spencer · Ya-Shu Chang · Peter W Collett

To the Editor: Chang and colleagues suggest that the clinical course and outcomes of patients infected with pandemic (H1N1) 2009 influenza virus are comparable to those with seasonal influenza infection, and that increased hospital and intensive care unit (ICU) admissions with influenza reflected a higher incidence of disease in the community rather than enhanced virulence of the pandemic influenza virus.1 However, this conclusion was based on an analysis of data from a single hospital, which did not examine whether the community incidence of influenza was increased or whether infection with pandemic (H1N1) 2009 influenza increased the risk of admission to a hospital or ICU compared to infection with seasonal influenza. Seasonal influenza co-circulated with the pandemic influenza strain during the recent influenza epidemic in New South Wales.2 We used the results of influenza tests (excluding rapid antigen tests) performed at eight major NSW public laboratory services to estimate the risk of admission to an ICU with pandemic (H1N1) 2009 influenza compared with seasonal influenza A infection. These public laboratories confirmed 95% of people admitted to hospital with pandemic (H1N1) 2009 influenza, and 84% of all laboratory-confirmed pandemic influenza cases in NSW. For the weeks ending 5 June to 2 October 2009, 38 060 specimens were tested for respiratory viruses, of which 7602 were positive on polymerase chain reaction for influenza A. Of these, 4172 (55%) were identified as pandemic (H1N1) 2009 influenza. During the same period, we collected data on all patients admitted to an ICU with influenza A infection.3 In total, 274 patients with influenza A were admitted to NSW ICUs — 229 with pandemic (H1N1) 2009 influenza, 38 with seasonal influenza A and seven with unsubtyped influenza A infection. Excluding the seven unsubtyped influenza A cases, we calculated a relative risk of 4.9 (95% CI, 3.5–7.0) for admission to ICU with pandemic (H1N1) 2009 influenza compared with seasonal influenza A infection. A sensitivity analysis assigning the unsubtyped influenza A cases admitted to ICUs to either pandemic (H1N1) 2009 influenza or seasonal influenza A produced no significant change in our findings, with a relative risk of admission to ICU with pandemic influenza compared with seasonal influenza A infection of 5.1 (95% CI, 3.6–7.1) and 4.2 (95% CI, 3.1–5.7), respectively. Although we attempted to collect data on all patients admitted to an ICU with influenza A infection, repeated testing for pandemic (H1N1) 2009 influenza in ICU patients may have resulted in identification of pandemic (H1N1) 2009 influenza in a greater proportion of ICU patients than community patients. However, our results suggest there was a significantly increased risk of admission to an ICU with pandemic (H1N1) 2009 influenza infection compared with other seasonal influenza A strains circulating in NSW. Given the potential for a subsequent pandemic wave, this provides support for maximising rates of community vaccination with a pandemic-specific vaccine.

Craig B Dalton · Michelle A Cretikos · David N Durrheim · Ian M Seppelt · William D Rawlinson · Dominic E Dwyer

To the Editor: Kelly showed that the pandemic (H1N1) 2009 influenza epidemic experience in Victoria was similar, overall, to a moderately severe influenza season.1 As with seasonal influenza, 0–4 year olds had the highest hospital admission rates,2,3 with those in the first year of life most likely to be admitted to intensive care units (ICUs). “Swine flu” did have differences to seasonal flu, with “younger” people having disproportionally more serious illness.2,3 However, these younger groups were not necessarily what most would regard as young. Eighty per cent or more of deaths associated with pandemic (H1N1) 2009 influenza were among those aged over 35 years, and most had identifiable risk factors.2,3 The peak 5-year age group for ICU admissions was 50–54 years.2 Those aged over 65 years seemed relatively protected — presumably because most had pre-existing immunity resulting from infection with H1N1-type viruses circulating since 1918. Pre-existing immunity was also not uncommon in others. In 18–65-year-olds, 30% had protective antibody levels to pandemic (H1N1) 2009 influenza before vaccination.4 Using data from New South Wales, we can ascertain the severity of pandemic (H1N1) 2009 influenza in different age groups.2 In NSW last winter, 1214 people were hospitalised, 225 were admitted to ICUs and 48 died with pandemic (H1N1) 2009 influenza (17.2, 3.2 and 0.7 per 100 000 population, respectively). Pregnant women had a 10 times higher risk of death (1.4 per 100 000) than other women of their age. For the overall population aged under 40, the death rate was 0.4 per 100 000, with most having identifiable risk factors. Thus, the death rate for those aged under 40 years, but with no known risk factors, was about one per million people. Most of the mortality predictions for this epidemic have been consistently wrong and exaggerated. Some were suggesting 10 000 deaths in NSW alone.5 These, and other predictions of second and third killer waves, have generated needless fear and inappropriate responses.6 Many experts and even health departments were postulating that more than 20% of the population would become infected, with an associated case fatality rate of 1% or more. That translates to a population mortality rate of 200 per 100 000 people, which is 300 times higher than what actually occurred. We need to learn from our recent experience so we can better plan and act in the future. Analysis such as that performed by Kelly is essential if we do not want to repeat the mistakes we made during this pandemic, caused by a virus of relatively low virulence.

Peter J Collignon

To the Editor: The debate raised by Sitas and colleagues about whether smoking status should be recorded on death certificates1 represents data acquisition at the last point. What needs to be considered is not the merits of the proposal, but the systematic failure to record this crucial data in patients’ clinical records. The argument I present here concerns cancer, but may be extrapolated to other smoking-related diseases. The absence of smoking history from the clinical records of patients with cancer up until now is understandable. Unlike the situation with infectious diseases, causative agents have had no relevance to the prognosis or the treatment of malignancy. Even if this pregenomic outlook persisted, the evidence we now have, showing that smoking affects response to therapy (eg, in prostate cancer),2 would warrant documentation of smoking history for every cancer patient. Among developed countries, the risk attributable to smoking in lung cancer is 70%–90%. In single-gene terms, polycyclic aromatic hydrocarbons and tobacco-specific nitrosamines such as 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone mediate malignant transformation by point mutations in onocogenes and tumour suppressor genes, typified by G-to-T transversions in codons 157, 158, 245, 248 and 273 of TP53; mutations in the KRAS gene and in the gene for epidermal growth-factor receptor (EGFR) are likewise relevant.3 Enough distinction can thus be made between smoking-related and non-smoking-related lung cancer to identify two entities.3 Patients with non-small-cell lung cancer, with exon 19 and 21 mutations in the EGFR gene are typically non-smokers, and their tumours have an 80% response rate to erlotinib.4 Indeed, these tumours respond better to chemotherapy than in patients lacking such mutations.4 As the single-gene era closes, smoking history will be required for all relevant biospecimens to avoid making genomic-wide data on pathways consequent on tobacco-related etiology — representing hundreds of tumours — inaccessible.5 Tobacco smoking causes cancer of the oral cavity, oropharynx, nasopharynx, and hypopharynx, oesophagus (adenocarcinoma and squamous-cell carcinoma), stomach, colorectum, liver, pancreas, nasal cavity and paranasal sinuses, larynx, lung, uterine cervix, ovary (mucinous), urinary bladder, kidney (body and pelvis), ureter and bone marrow (myeloid leukaemia), and possibly causes female breast cancer.6 The prospect of genomic-wide data for these tumours not being accompanied by data on smoking is daunting. The case for smoking data in relation to clinical trials has been made previously.7 Beyond this, the immediate goals for Australia are that smoking history be recorded for tobacco-related malignancies in hospital-based tumour-specific clinical cancer registries and for biospecimens. The need is for the adoption of uniform national vocabulary and coding methods. This will require leadership by an authority such as the National Health and Medical Research Council (NHMRC), the Cancer Council Australia, the Clinical Oncological Society of Australia, or Cancer Australia.

Bernard W Stewart

To the Editor: As health professionals, we urge Australian politicians (and the public) to recognise the overlap in the underlying cause of two great health threats that our population now faces: the rise of obesity and its life-threatening disease consequences, and the great threats to health from global climate change.1-3 Big health gains have been made since the onset of industrialisation. However, we are now seeing the emergence of health risks caused by excesses in market-driven consumerism (including the consumption of energy-dense processed foods), energy-subsidised exertion-free living, an over-arching preoccupation with gross domestic product and (particularly relevant to climate change) population growth.4,5 In light of our profession’s long experience with the smoking–disease debate, we recognise the serious threat posed to population health by the well financed, doubt-fostering opposition of vested interests. In the case of smoking, it took 50 years and several million deaths to progress from acknowledgement of the health hazard to application of effective interventions. With climate change, we lack the luxury of time — and the stakes are much higher.6 Hence, a prudent, precautionary strategy to tackle the big issues is essential. We propose two initiatives: Convene a high-level, ongoing forum to discuss post-growth alternatives to unsustainable, consumption-based growth as the economic norm. This should involve health and social scientists, among others, and economists prepared to objectively take into account the real environmental costs of our actions; and Convene a multidisciplinary taskforce to develop an Australian population policy. Population is a key driver of energy use, greenhouse pollution and resultant ill-health.7 Despite the apparently widespread bias against public discussion of population issues, the topic is now attracting renewed attention and the debate should be facilitated. The irony of adopting a “no change” approach to these two issues is that, if this happens, involuntary change to everyone’s lifestyle will become unavoidable. We must seek a sustainable economic system and stable population size that ensures prosperity without endangering both health and environmental quality.8 For the moment, our actions are seriously damaging our planet and our children’s health. One in three children born today will become obese and/or diabetic in their lifetime, and the population at large will face increasing health risks from climate change.

Garry J Egger · Boyd Swinburn · Fiona Stanley · Kerryn Phelps

Childhood obesity and lifespan Obese children are far less likely to reach old age, according to US researchers. In a cohort of almost 5000 American Indians, they studied the effect of cardiovascular risk factors that were present in childhood on lifespan. A history of childhood obesity more than doubled the risk of dying from causes other than accident, suicide or homicide before the age of 55 years (incidence rate ratio 2.3). Although obesity conferred the strongest association with premature death, there were also significant associations between premature death and both impaired glucose tolerance and hypertension. Interestingly, no association with cholesterol levels was observed. The study was also of note because of its length — participants were followed up for a median of 24 years. N Engl J Med 2010; 362: 485-493 Ageing hearts Should octogenarians with acute coronary events undergo per-cutaneous coronary intervention (PCI)? The elderly are generally underrepresented in trials, but a recent Dutch study provides some insights. Of all patients with ST elevation myocardial infarction (STEMI) admitted to one institution for PCI, the proportion of octogenarians treated more than doubled over the 11-year study period, indicating that this is likely to be an ongoing dilemma. Unfortunately, mortality was high in this age group and did not improve (mortality at 30 days was 21.4% compared with 9.4% among 60-79-year-olds). As well as having more comorbidities, the octogenarians were less likely to have successful restoration of coronary perfusion — although this did improve over the study period, possibly reflecting better technology. Heart Online 4 Dec 2009 Haiti quake In two dispatches from Haiti, medical staff describe their experiences working in a collapsed infrastructure. While trying to restore a semblance of order, staff were faced with a lack of fuel, electricity and communication as well as threats of violence, with thousands of prisoners having escaped from the nearby national penitentiary. Hospitals across the city had varying capacities to provide services, with an urgent call for a centralised system informing people where to go for appropriate care. Meanwhile, another hospital about 60 km from the epicentre remained intact, allowing surgical teams to respond while operating rooms in Port-au-Prince were incapacitated. The acute management of the enormous number of injuries inflicted by this disaster is expected to evolve into wound care, fitting of prostheses and rehabilitation. N Engl J Med 2010; 362: 575-777 N Engl J Med 2010; 362: e19 (1)-e19 (2) The long and the short of it Are short babies really more likely to be shorter and weigh less when they grow up? A study involving almost 600 Israeli army recruits indicates that final adult height (at age 17 years) can be predicted by length at birth. In this retrospective study, those babies born short for gestational age (less than 48 cm at birth) were significantly more likely to be shorter and weigh less at their conscription medical than those with a normal birth length. This finding was apparent for both sexes. Arch Dis Child 2009; 94: 959-961 Pregnancy and pandemic flu Advanced pregnancy appears to be associated with an increased risk of severe complications of H1N1 influenza, according to analysis of the outcomes in 43 pregnant women with confirmed infection admitted to Melbourne hospitals during the 2009 pandemic. Of these, 65% were in the third trimester (of whom about half were at term), with 30% in their second trimester and only 5% in the first. Although about half had at least one associated comorbidity (including asthma, diabetes and obesity), half had no other risk factors. Antiviral therapy was prescribed in 77% and 8 women required admission to ICU. Median length of stay was 2 days, and 8 were hospitalised for more than 7 days.1 Although the mechanism by which pregnancy increases flu severity is unclear, a related study may shed some light. After finding a low level of IgG2 in one pregnant woman with severe H1N1 infection, levels were assessed in 39 other patients with confirmed infection (9 of whom were pregnant). Pregnant patients with H1N1 were found to have significantly lower levels compared with healthy pregnant controls, suggesting a possible role of IgG2 deficiency in the pathogenesis of H1N1 infection in pregnancy.2 1. Clin Infect Dis 2010; 50: 686-690 2. Clin Infect Dis 2010; 50: 672-678

Alison Williams