Statistics

Objective: To describe the frequency, characteristics and outcomes of reports of possible harms related to medical devices submitted to the Australian Therapeutic Goods Administration (TGA) using data made publicly available on the TGA website.Design and setting: A retrospective analysis, conducted in January 2012, of data made publicly available on the TGA website from January 2000 to December 2011.Main outcome measures: The number and nature of reports of medical device incidents, recalls and alerts.Results: Up to December 2011, 6812 incidents involving medical devices were reported to the TGA, although there were several periods where data were unavailable. Incidents were reported more frequently in later years, most often by device sponsors, and were often attributed to mechanical problems. 295 deaths and 2357 serious injuries have been related to incidents, with serious injury (597) highest in 2009. Most incidents involving medical devices were not investigated (47.5%), or, after investigation, no further action was taken (25.0%). During the same time period, there were 35 medical device recalls and 34 medical device alerts issued by the TGA, with no consistent increase over time.Conclusions: Despite TGA reform proposals, greater transparency is still needed. Issues that have not been addressed include patchy and conflicting data in the public domain and lack of explanations for the large proportion of uninvestigated reports. To maintain public confidence in the national regulatory system these problems need to be resolved.

Richard G McGee PGDipSurgAnat, MMed(ClinEpi) · Angela C Webster MMed(ClinEpi), PhD · Thomas E Rogerson BSc · Jonathan C Craig MMed(ClinEpi), PhD

In Australia, regionalised maternity care aims to ensure provision of the appropriate level of care to all women, through assessment of risk, referral and transfer. An important element of this care is providing high-quality, safe and accessible services to women and babies in rural and remote areas. For women who experience pregnancy complications, this means their needs may not necessarily be met by their local hospital. They may be admitted ...

Tim Badgery-Parker BSc(Hons), MBiostat · Jane B Ford BA(Hons), PhD · Mary G Jenkins RN, CM, MScM · Jonathan M Morris MB ChB, FRANZCOG, PhD · Christine L Roberts MB BS, FAFPHM, DrPH

Objective: To examine the differential in cancer survival between Indigenous and non-Indigenous people in Queensland in relation to time after diagnosis, remoteness and area-socioeconomic disadvantage.Design, setting and participants: Descriptive study of population-based data on all 150 059 Queensland residents of known Indigenous status aged 15 years and over who were diagnosed with a primary invasive cancer during 1997–2006.Main outcome measures: Hazard ratios for the categories of area-socioeconomic disadvantage, remoteness and Indigenous status, as well as conditional 5-year survival estimates.Results: Five-year survival was lower for Indigenous people diagnosed with cancer (50.3%; 95% CI, 47.8%–52.8%) compared with non-Indigenous people (61.9%; 95% CI, 61.7%–62.2%). There was no evidence that this differential varied by remoteness (P = 0.780) or area-socioeconomic disadvantage (P = 0.845). However, it did vary by time after diagnosis. In a time-varying survival model stratified by age, sex and cancer type, the 50% excess mortality in the first year (adjusted HR, 1.50; 95% CI, 1.38–1.63) reduced to near unity at 2 years after diagnosis (HR, 1.03; 95% CI, 0.78–1.35).Conclusions: After a wide disparity in cancer survival in the first 2 years after diagnosis, Indigenous patients with cancer who survive these 2 years have a similar outlook to non-Indigenous patients. Access to services and socioeconomic factors are unlikely to be the main causes of the early lower Indigenous survival, as patterns were similar across remoteness and area-socioeconomic disadvantage. There is an urgent need to identify the factors leading to poor outcomes early after diagnosis among Indigenous people with cancer.

Susanna M Cramb BAppSc(Med Sci), Grad Cert(Sc), MPH · Gail Garvey BEd, MEd · Patricia C Valery PhD, MD, MPH · John D Williamson BA(Hons), GradCert(PublicHlth), MEpi(Clin Epi) · Peter D Baade BSc, MMedSc, PhD

Objectives: To document temporal trends in maternal overweight and obesity in Australian women and to examine associations with pregnancy outcomes....

H David McIntyre MB BS, FRACP, MD · Kristen S Gibbons BMaths(Hons), BInfoTech · Vicki J Flenady MMedSc(ClinEpid) · Leonie K Callaway MB BS, PhD, FRACP

Plotting the trajectory of some common artery conditions.

on behalf of the REACH Registry Investigators

When “should not” becomes “must not” — how mandatory compliance with guidelines can threaten professional independence There is a proliferation of rules and red tape at all levels of society. In New South Wales, for example, the Department of Health has been generating around 70–100 policy directives a year. Among other things, such directives tell obstetricians when and how to deliver babies — compliance is mandatory, under threat of disciplinary action and loss of indemnity cover. One of the most prominent recent directives, PD2010_045 (Maternity — towards normal birth in NSW), requires a reduction in caesarean section rates to 20% by 2015,1 a target that is illusory and possibly dangerous. Others are worse. Policy directive PD2007_024 (Maternity — timing of elective or pre-labour caesarean section) states: The risk of respiratory morbidity is increased in babies born by caesarean section before labour, but this risk decreases after 39 completed weeks. Therefore elective or pre-labour caesarean section must not routinely be carried out before 39 completed weeks. These findings are supported by recent studies.2 This text is taken almost verbatim from a guideline of the National Institute for Health and Clinical Excellence (NICE) in the United Kingdom3 and, at the time, accorded with the relevant Royal Australian and New Zealand College of Obstetricians and Gynaecologists (RANZCOG) guideline on the timing of elective caesarean section, the wording of which has since been softened slightly.4 However, there is one very substantial difference: the “should not” of NICE and RANZCOG became a “must not”. The NICE guideline was produced by a committee of 16 members, two of whom were obstetricians. Nicholas Fisk, then chairman of the Royal College of Obstetricians and Gynaecologists (RCOG) Scientific Advisory Committee, strongly criticised the outcome and stated that the committee “selectively interpreted” the facts to suit its case.5 Timing of elective caesarean delivery is a complex issue. There are factors (mainly related to pulmonary maturity) suggesting delay until 39 weeks’ gestation, and others (mainly related to unexplained stillbirth6,7) supporting the traditional timing between 38 weeks and 38 weeks and 6 days. A large randomised controlled trial (RCT) or a carefully conducted observational study employing propensity scoring would be required to provide a solid evidence base. For the moment it remains a matter of opinion, and RCOG, NICE and RANZCOG can, and should, express expert opinion and review the available evidence. Guideline committees and their products make an important contribution to improving clinical practice, and the voluntary work of countless individuals in this regard should not be denigrated. However, the situation may be altered completely once bureaucrats become involved. Changing the “should” of a College guideline to a “must” in a NSW Health policy directive has major implications. PD2007_024 effectively forbids doctors in NSW public hospitals to schedule routine elective caesarean section before 39 weeks. Anyone doing so risks disciplinary action and may forfeit their indemnity cover. PD2007_024 is not currently supported by high-level evidence, and even if there were solid data from a large RCT or a meta-analysis, it would be inappropriate to stipulate “mandatory compliance”. Evidence-based medicine has great potential to improve patient care, but it is a tool to inform clinicians, not an excuse to switch off one’s brain. As recently argued in the case of mesh use in pelvic reconstructive surgery,8 guidelines derived from a large RCT (or a valid meta-analysis of appropriate trials) are of limited use as a guide to the care of an individual patient if that patient is far off the trial sample mean in predictors of outcome. The Term Breech Trial9 is an excellent example. If there is a very high likelihood of an uncomplicated birth for an individual woman (eg, fast progress, small baby, previous normal births), a clinician may want to ignore the findings of the Term Breech Trial when deciding how to advise her. The process of starting with some initial information (probabilities of events of interest) and combining it with data related to an individual case or circumstance has been mathematically formalised by statisticians as “Bayesian updating”.10-13 Information from a clinical guideline, prevalence study or meta-analysis provides the “prior probability”, the data related to the individual case at hand are embodied in the “likelihood”, and the application of Bayes’ theorem13,14 gives the updated “posterior probability” on which a decisionmaker can act. This implies that even the best evidence in literature and guidelines can provide only prior probabilities. It is our job to adapt these to the clinical management of the individual patient, generating posterior probabilities that have been updated by the relevant particulars of the patient and processed by clinical intuition and common sense. However, in a trend that is encouraged by guidelines and policy directives, clinicians increasingly omit this updating process. Once a document — a seminal paper, a guideline, a policy directive — is published, it is likely to be interpreted cautiously (ie, as widely as possible) because of the litigious, risk-averse nature of our society. We do not particularise to the individual patient because our work environment strongly discourages doing so. Finally, even if a guideline is produced after an optimally diligent process and published so rapidly that it is up to date at its launch, it still freezes current best practice. Inevitably, today’s best practice will be obsolete tomorrow. Practice guidelines therefore have the potential to slow down progress, and unthinking adherence to such documents is unlikely to deliver the best outcomes to the greatest number of individuals. The expert system of our brains, employed in the assessment and treatment of the individual patient — and using relevant prior information from evidence-based sources — is a potentially superior tool when compared with the replication of a course of action outlined in a document issued by a government agency or professional body. It is time to reassert our professional independence.

Hans Peter Dietz MD, PhD, FRANZCOG · Barrie J Stokes BSc, MMath

To the Editor: Participation by patients in cancer clinical trials is low. Doctors’ reluctance to participate in clinical trials has been reported as a key barrier to recruitment.1 The Physician Orientation Profile (POP) examines doctors’ attitudes to and behaviour regarding randomised clinical trials.2 In the context of a 2010 cluster-randomised trial to assess the impact of a consumer-friendly cancer clinical trials website (http://www.australiancancertrials .gov.au), we asked medical oncologists to complete the POP and identify barriers they faced when recruiting patients to trials. The option of providing a comment on barriers identified was included. The sociodemographic characteristics of the 28 medical oncologists who participated in the trial and the barriers to recruitment reported are shown in the Box. Twenty-one doctors reported that they faced barriers when recruiting patients to clinical trials. The most common barriers identified included poor organisational infrastructure, insufficient time, slow and bureaucratic ethics approval processes, and lack of staff (medical oncologists and clinical trials coordinators). This was associated with the feeling that it is too hard to do trials. However, in contrast to studies reported during the 1990s in the United States and the United Kingdom,1,2 we found that doctors were strongly supportive of clinical trials. A high proportion gave answers consistent with the attitudes of “researchers” (practice guided by published data) rather than “clinicians” (practice guided by clinical experience). For example, if there was uncertainty about treatment, most saw this as an opportunity to do a randomised trial (26 of 28), and most believed that clinical trials were central to a medical oncologist’s practice (24 of 28). We acknowledge our small sample size and the likelihood of selection bias. However, the doctors in our study were experienced in recruiting patients to clinical trials and had firsthand knowledge of barriers to recruitment in the Australian medical system. Our findings suggest that, rather than negative attitudes of doctors, issues regarding organisational infrastructure, staffing, time and ethics approval processes are now significant barriers to recruitment of patients to cancer clinical trials. The participating doctors’ comments about the challenges of ethics and governance processes are consistent with recent publications which highlight that multisite studies are under threat because of onerous ethics review and regulatory requirements.3,4 These issues are also addressed in the first rec-ommendation of the Australian Government’s Clinical Trials Action Group report, which is “to improve the timeliness of ethics and research governance review”.5 Our results support the importance of this recommendation. Australian medical oncologists’ sociodemographic characteristics and barriers to recruiting patients to trials (n = 28), 2010 Number* Sociodemographic characteristics Mean age (years) 48 Men 20 Location New South Wales 17 Victoria 11 Practice type Mostly salaried 10 Mostly fee-for-service 9 About equal 9 Barriers to recruitment One or more barriers 21 Poor organisational infrastructure (eg, “Lack of support to employ clinical trials staff”) 11 Lack of time 9 Lack of access to clinical trials due to geographic isolation 8 Difficulty identifying eligible patients 4 Trials competing for the same patients 2 Lack of awareness about ongoing trials 2 Preference for a particular treatment arm 0 Other† 11 No barriers 7 * Data are numbers unless otherwise indicated. † Examples of comments provided for other barriers: “Hopeless, slow ethics in NSW — central ethics has not helped. Average time to get trials going exceeds average working life of trials coordinators!! It is TOO HARD to do trials. The logistics and lack of infrastructure is a major barrier” “Bureaucratic ethics and governance requirements” “Profound delays in ethics approval” “Lack of resources and man power to put more clinical trials in our treatment centre. Sometimes forgetting about ongoing trials if I am not the principal investigator and sometimes not aware of ongoing trials outside our centre” “Lack of additional medical oncologists” “Insufficient clinical trials coordinators — constant overload of staff means we have to keep halving study accrual”

Rachel F Dear · Alexandra L Barratt · Martin H N Tattersall

Patients with cardiovascular disease may be especially prone to the adverse effects of complementary medicines Cardiovascular disease (CVD) is the leading cause of morbidity and mortality in Australia.1 Given the chronic nature of CVD, medical management is the most common mode of care, and medications are prescribed to most patients with CVD.2 In general practice, treatment of CVD accounts for almost one in five encounters, with hypertension the most commonly treated risk factor, followed by lipid disorders.2 Although effective management of CVD depends on a range of factors, one that is often neglected is the use of complementary medicine (CM). While CM can potentially benefit patients with CVD, it can have an adverse impact on the effectiveness of conventional CVD therapies in two main ways: through drug interaction or through reduced adherence to conventional therapies. In Australia, CM is defined as the use of medicines containing herbs, vitamins, minerals, nutritional supplements and homoeopathic medicines.3 The popularity of CM has increased in recent years,4 and in people with CVD, rates of use have been estimated at up to 42%, with up to 21% of these patients taking herbal supplements.5 The most common types of CM taken by people with CVD are multivitamins and minerals (34% of surveyed cardiac patients), calcium (22%), vitamin E (20%) and vitamin C (14%). The most common herbal agents used are mint or lemon balm (11%), nettle (8%), green tea (7%), echinacea (6%) and garlic (6%). The use of coenzyme Q10 (3%), fish oil or omega-3 fatty acids (2%) and hawthorn (1%) are surprisingly low.5 Although CM is often considered safe by patients, harmful effects associated with its use do exist. The use of CM alongside prescription medications has the potential of reducing the effectiveness or increasing the potency of therapy. These effects may be compounded in people with CVD for several reasons. First, CM users tend to consume more than one CM product concurrently over periods of months to years.5 Second, the likelihood of interference with prescribed therapy is increased since certain medications, such as digoxin and warfarin, have a narrow therapeutic window. Patients with CVD also tend to be elderly, are likely to suffer from multiple comorbidities and therefore are likely to be using multiple long-term medications.5 Several studies in people with CVD indicate that not only do a significant proportion use CM concurrently with prescribed therapy, but about half do not inform treating doctors about their use.5 The relatively widespread use of CM in people with CVD means it is vital that doctors are aware of CM use. Some commonly observed interactions between CM and conventional CVD medications and possible mechanisms are summarised in the Box. Apart from the possibility of drug interactions, the bioactive components of CM products can also directly cause adverse effects. Some examples of CM that are used in the treatment of CVD (and their adverse effects) include Aloe vera (diarrhoea and potassium depletion), fenugreek (diarrhoea and hypoglycaemia), garlic (inhibition of platelet function), ginseng (insomnia), ephedra (stroke and myocardial infarction), Ginkgo biloba (bleeding) and red yeast rice (myopathy and rhabdomyolysis).7 More generally, there are safety issues associated with the quality of CMs. These relate to the lack of standard dosing, which in part is due to variations in potency, depending on where plants are grown; the possibility of contamination by pesticides and heavy metals during cultivation; and the possibility of contamination by bacteria in the storage, transport or manufacturing processes.7 The second main mechanism in which CM use may impact on effective management of CVD is through reduced adherence to prescription medications. However, a limited number of studies investigating the impact of CM on cardiovascular medication adherence show inconsistent results. Some studies report a lower adherence to prescription medications among cardiovascular patients taking CM, while others found the opposite.5 Importantly, we found in our review that CM use was not disclosed by cardiac patients up to 65% of the time. The main reasons were fear of clinician disapproval and because clinicians had not asked about CM use.5,9 This strongly suggests that clinicians need to take a less judgemental and more proactive approach to encourage patient discussion of CM use, regardless of adherence to conventional treatment. What, then, are the identifying characteristics of a CM user? In the general population, use of CM is associated with higher levels of education and income10 but poorer self-reported health status. Although few differences have been found between cardiac patients who are CM users compared with non-users,5 it appears that people with chronic disease are more likely to use CM, especially if they have multiple comorbidities.10 As well as being aware of the potential impact of CM on prescribed therapy, clinicians should also be aware of the reasons patients use CM. About half of CM users with CVD believe that CM is of greater benefit than conventional treatment, or perceive that the CM used has proven benefits for their condition.5 Many CM users also believe that there are fewer side effects associated with “natural” therapies.5 CM therefore meets a real or perceived need for health care, and serves an important psychological function in enabling patients to manage their own health.11 Given the increasing prevalence of both CM use and CVD, it is likely that clinicians will see more cases of patients who are using CM either specifically for their cardiovascular condition or towards improving general health. Regardless of personal viewpoint, CM is now an integral part of the therapeutic armamentarium used by the Australian population. Thus, there is a need to better inform doctors about CM use and the associated potential for adverse reactions and herb–drug interactions. There are significant gaps in clinician knowledge of the risks and benefits associated with CM use, and clinicians report feeling ill-equipped to respond to patient enquiries about CM.12 Medical schools overseas, such as those in Asia, France, Germany, Switzerland and some in the United States, have incorporated CM into their medical curricula. Awareness of CM needs to be integrated into routine medical practice, such as by asking patients about CM as part of a medication review. CM modules should be included in continuing professional education programs for general practitioners and specialists alike. The key issues and complexities of CM that should be understood include the differing definitions of CM in different countries, such as the American and Australian classifications, and the processes underlying its regulation. In Australia, for example, the Therapeutic Goods Administration regulates CM products, and the difference between “listed” medicines (which are not evaluated for efficacy) versus “registered” medicines is important.3 There is growing research on the efficacy of select CMs for various conditions. Results from robust studies show that omega-3 fatty acids may assist in the treatment of hypertriglyceridaemia,13 reduce the incidence of thrombotic stroke14and coronary heart disease,15 and may prevent sudden cardiac death in patients with prior myocardial infarction.16 Coenzyme Q10 supplements appear to reduce the symptom of statin-induced myalgia17 and may have favourable effects in those with heart failure.18 Physical therapies such as qi gong also appear helpful for hypertension19 and are likely to be beneficial in some patient groups as alternatives to higher intensity physical activity. Furthermore, considerable efforts are being made to investigate the efficacy of CM in CVD prevention and risk factor reduction. Research bodies such as the National Institutes of Health in the US and the National Institute of Complementary Medicine in Australia are supporting studies investigating the effects of hawthorn leaf in milder forms of heart failure, the use of disodium edetate (EDTA) chelation therapy to treat coronary artery disease, vitamin D3 for prevention of CVD and melatonin for lowering hypertension. Further research is required to strengthen the evidence base for these and other CM therapies; however, certain CMs may prove to be efficacious for people with CVD, and critical awareness of this work is necessary for effective clinical practice. The use of CM is common among patients with cardiovascular conditions and a high proportion of patients using CM believe they have remedial benefits. Many CM users also believe that CM is as safe as or safer than their prescribed treatments, and are often unwilling to inform their doctors of their use of CM products. Commonly used CM products have the potential to interfere with the intended action of prescription medications although clinician knowledge of these interactions may be limited. We have restricted our discussion to patients with CVD, although the importance of considering CM in medical management applies equally to other patient groups, especially those with chronic disease. We emphasise the need for education about CM, so clinicians become more aware of patients’ CM use, become more knowledgeable about CM, and more capable of advising their patients about this issue. Letter p 660 Potential interactions of commonly used CM with cardiovascular medications6-8 CM (prevalence of use*) Possible herb–drug interactions and mechanism of action Coenzyme Q10 (3%) May partially antagonise antiplatelet effect of clopidogrel and decrease response to warfarin Vitamin E (20%) Contains curbicin, which antagonises the effect of vitamin K on coagulation, which, with clopidogrel, may increase the risk of bleeding and potentiate warfarin effects Garlic (6%) Additive antiplatelet effects may occur with clopidogrel and other anticoagulants Ginkgo biloba (4%) Ginkgo inhibits platelet aggregation and increases the effect of warfarin Fish oil (2%) Increases the effect of antiplatelet agents and vitamin K-dependent coagulation Concomitant use of warfarin and fish oils may increase risk of bleeding Vitamin D (4%) Improves calcium absorption, which increases the toxic effect of digoxin St John’s wort (2%) Reduces the effectiveness of statins, warfarin and digoxin through induction of hepatic enzymes and P-glycoprotein Hawthorn (1%) Increases the inotropic effects of digoxin Ginseng (4%) Has an additive effect with antiplatelet agents Decreases the effect of anticoagulants Glucosamine/chondroitin (4%) Increases the effect of anticoagulants Capsicum (5%) Increases the effect of antiplatelet agents Bilberry (1%) Increases the effect of anticoagulants Aloe vera (4%) Increases the effect of antiplatelet agents Produces hypokalaemia, leading to increased toxic effect of digoxin; enhances the effect of digoxin Liquorice (prevalence of use unknown) Mineralocorticoid effects promote potassium excretion. Causes hypokalaemia when used with some antihypertensive agents CM = complementary medicine. * Prevalence of use is based on figures from a systematic review of the literature combining data from multiple studies.5

Hosen Kiat MB BS, FRACP, FACC · Yu Sun Bin BSc(Hons) · Suzanne Grant BApplSc(TCM), MPS, PhD · Dennis Hsu-Tung Chang MB BS, MSc, PhD

Objective: To examine the effects of paternal depression during children’s first year on their wellbeing at 4–5 years of age using a large, representative sample of Australian families.Design, setting and participants: Prospective study of Australian families from 2004 to 2008. Two-biological-parent families (n = 2620) from the Longitudinal Study of Australian Children were included if depression measures were available for both parents in 2004, behavioural and developmental measures were available for children in 2008, and the families had not separated by 2008.Main outcome measure: Child scores on the Strengths and Difficulties Questionnaire and on a set of Derived Outcome Indices, measured when the child was 4–5 years old. Logistic regression modelling was used to investigate how early paternal depression in 2004 influenced child outcomes 4 years later.Results: Early paternal depression was a significant predictor of a range of poorer child outcomes (odds ratio [OR] for behavioural difficulties, 3.34 [95% CI, 3.06–3.65]; OR for a low development and wellbeing score, 2.70 [95% CI, 2.44–2.98]). These effects remained significant after controlling for both early maternal depression and later paternal depression (adjusted OR for behavioural difficulties, 1.93 [95% CI, 1.75–2.14]; OR for a low development and wellbeing score, 1.65 [95% CI, 1.48–1.85]).Conclusions: Depression in fathers during the first year of a child’s life can have a detrimental impact on their child’s behaviour, and social and emotional development at the point of school entry, in addition to and uniquely compared with depression in mothers. Early intervention to identify and address the mental health needs of fathers is required for the benefit of fathers, children and families.

Richard J Fletcher PhD · Emily Feeman PhD · Craig Garfield MD, MAPP · Graham Vimpani FRACP, FAFPHM, PhD

Objective: To measure alcohol-related harms to the health of young people presenting to emergency departments (EDs) of Gold Coast public hospitals before and after the increase in the federal government “alcopops” tax in 2008.Design, setting and participants: Interrupted time series analysis over 5 years (28 April 2005 to 27 April 2010) of 15–29-year-olds presenting to EDs with alcohol-related harms compared with presentations of selected control groups.Main outcome measures: Proportion of 15–29-year-olds presenting to EDs with alcohol-related harms compared with (i) 30–49-year-olds with alcohol-related harms, (ii)15–29-year-olds with asthma or appendicitis, and (iii) 15–29-year-olds with any non-alcohol and non-injury related ED presentation.Results: Over a third of 15–29-year-olds presented to ED with alcohol-related conditions, as opposed to around a quarter for all other age groups. There was no significant decrease in alcohol-related ED presentations of 15–29-year-olds compared with any of the control groups after the increase in the tax. We found similar results for males and females, narrow and broad definitions of alcohol-related harms, under-19s, and visitors to and residents of the Gold Coast.Conclusions: The increase in the tax on alcopops was not associated with any reduction in alcohol-related harms in this population in a unique tourist and holiday region. A more comprehensive approach to reducing alcohol harms in young people is needed.

Steve R Kisely MD, PhD, FAFPHM · Joanne Pais MSc · Angela White MCP, PhD, MAPS(Clin) · Jason Connor PhD, MAPS · Lake-Hui Quek PhD · Julia L Crilly BNurs, MN(Hons), PhD · David Lawrence PhD

To the Editor: In 2006, the Cancer Institute NSW established a single ethics committee, in order to improve the efficiency of ethics reviews for multicentre cancer clinical trials. This predates the National Health and Medical Research Council (NHMRC) Harmonisation of Multi-centre Ethical Review (HoMER) but exemplifies what HoMER aims to encourage nationally. Previously, such trials were submitted to each institution’s ethics committee, resulting in replication of effort and cost and prolonged review times, potentially making sites uncompetitive in attracting clinical trials.1 Under the Cancer Institute’s model, with the agreement of individual institutions, multicentre projects are submitted directly to a single ethics committee. Governance issues, such as the capability of an institution to provide appropriate support, and insurance issues have remained local health unit responsibilities.2 Data were prospectively collected on applications received from 1 July 2007 to 30 June 2009, and their processing times. The Australian Research Ethics Database was used to track and manage multicentre research projects. The aim was to achieve a 60-calendar-day time period from submission (1 month before the ethics committee meeting) to approval. Submissions were made on the NHMRC National Ethics Application Form, along with a protocol and patient information form. We evaluated 89 studies from the 2-year period. Fourteen trials were reviewed from single sites before other sites were engaged or whose ethics committees lacked cancer expertise. The median time range for review was 61–70 days. Forty-four per cent of applications were reviewed within 60 days and 70% within 80 days, ranging from eight studies taking 31–40 days to the extreme of seven studies requiring 131–160 days. The median time for assessment improved as the committee streamlined its processes: from 89 days in 2007 to 68 days in 2008 and 59 days in 2009. There were 15 studies, predominantly in the first year, that had long assessment times because of multiple issues concerning research merit and inadequate information for participants. Of the studies evaluated, one was approved without revision, 48 required minor revisions that were approved between meetings, and 40 required review by the full committee. Eighteen studies were reviewed twice, and 22 were reviewed more than twice. Efficiencies introduced included disseminating all documentation electronically, empowering the chair and deputy chair to approve minor amendments or responses between meetings, and the website carrying standard wording for use in sections of the patient information forms that were proving to be recurrently problematic. To resolve difficult issues that correspondence had not resolved, investigators were invited to meet with the committee. We conclude that a single ethics review for multicentre trials may be able to deliver a rapid response and be efficient without compromising ethical rigour. However, to make a difference to researchers, the governance review process will need to become correspondingly efficient. Our experience parallels that internationally, where such ethics committees can demonstrate cost savings with faster response times, yet provide ethical reviews of similar quality.3,4,5

Ian N Olver · Sharon J P Falleiro · Marion L Marson · James F Bishop

Objective: To assess mortality trends among people with cystic fibrosis (CF) in Australia.Design and setting: We augmented Australian summary data for deaths from CF registered during 1979–2005 with information from Australian transplant centres on lung transplantation among CF patients for 1989–2005 to allow us to follow trends in all “mortality events” (death or lung transplantation).Main outcome measure: Age at death or lung transplantation.Results: Between 1979 and 2005, the mean age at death increased from 12.2 years to 27.9 years for males and from 14.8 years to 25.3 years for females. Overall, female deaths in childhood (0–14 years) occurred at an age-standardised rate of 0.40 per 100 000 (95% CI, 0.34–0.45) during 1979–2005, which exceeded the corresponding rate for males of 0.24 (95% CI, 0.20–0.28) per 100 000. Among 0–14-year-old boys, event rates declined markedly after 1989, but they declined later and more gradually for girls, with the result that the age-standardised rate for girls was 2.38 times that of boys during 1989–2005 (95% CI, 1.69–3.36).Conclusions: The pattern of CF mortality in Australia has changed substantially. Mortality rates continue to be higher for girls than for boys, but death in childhood has become uncommon. Survival has increased since 1979, but females continue to have reduced length of life.

David W Reid BSc, MB ChB · C Leigh Blizzard PhD · Dace M Shugg RN · Ceri Flowers BSc · Catherine Cash BSc · Hugh M Greville MD

To the Editor: Scott and Guyatt are absolutely correct that as much transparency as possible is required in the process by which any clinical guidelines are formulated.1 I also agree that this must include careful management of conflicts of interest, particularly in the selection of the members of the expert panels required to formulate guidelines and in the functioning of those panels.1 However, transparency does not necessarily require adopting the GRADE (Grading of Recommendations Assessment, Development and Evaluation) system or any other hierarchical system for formulating guidelines, nor does it necessarily require seeking formal feedback from external stakeholders. The GRADE system, like most hierarchical systems, assumes there is at least some evidence in the literature that addresses the relevant clinical problem — thus, it gives its lowest grade to observational studies with a very uncertain estimate of effect. But, often, practitioners most want help with the complex and multifaceted clinical problems on which there are not even formal observational studies. The GRADE system would presumably put these problems in the category of “You’re on your own, chaps, and don’t expect us to help you”. Recommendations on the problems that are most vexing to practitioners will be mainly (and inevitably) based on clinical experience and circumstantial evidence — in other words, expert opinion. Most practitioners looking for guidance will be well aware that such recommendations are likely to be based on expert opinion rather than excellent formal evidence. When deciding whether or not to trust the recommendations, they will be much more concerned that any conflicts of interest for the experts making the recommendations have been recognised and appropriately redressed, rather than that the level of evidence has been rigorously graded. Seeking formal feedback from stakeholders is also a separate issue from transparency. Potential users should be well represented on any writing panel, and their feedback is obviously essential — there is no point in producing guidelines that don’t address the problems that the potential users are facing. However, many stakeholders — for example, the pharmaceutical industry — will have much more serious conflicts of interest than do the members of the expert panel, and these can be impossible to deal with. Of course, feedback should be sought where appropriate, but to make it an absolute requirement under the guise of transparency seems nonsensical. It is clearly essential that users of clinical guidelines must be able to have confidence that they have been developed as rigorously as possible, but an excessively doctrinaire approach to their development will almost certainly lead to a decline in usefulness without a counterbalancing increase in trustworthiness.

Robert F W Moulds

To the Editor: In his thought-provoking editorial in the May issue of the Journal,1 Tait made reference to an apparent recent improvement in the life expectancies of Indigenous Australians by citing a 2010 Australian Bureau of Statistics (ABS) report entitled The health and welfare of Australia’s Aboriginal and Torres Strait Islander peoples, Oct 2010.2 In this report, the life expectancy for Indigenous Australians was quoted as 67.2 years for males and 72.9 years for females, compared with 78.7 and 82.6 years for non-Indigenous males and females, leaving a “gap” of 11.5 years and 9.7 years, respectively. These figures are from 2005–2007 (which includes the 2006 Census year), and are quoted again in this year’s update report from the Australian Institute of Health and Welfare.3 At first glance, they appear to be a startling improvement on the figures from 1996–2001, which quote (as late as 2005) Indigenous life expectancies of 59.4 and 64.8 years for males and females respectively, representing a “gap” of about 17 years for both.4 Unfortunately, the apparent improvement represents not a miraculous leap forward in Indigenous health care and outcomes, but rather a change in the methodology used to calculate life expectancies around the time of the 2006 Census. The essence of the change was from an indirect to a direct demographic method of compiling life-expectancy estimates, which entailed correcting Indigenous death registration data before calculating death rates. The ABS anticipated the potential for confusion (not to mention premature celebration), and so included warnings that comparisons should not be made between published estimates of Indigenous life expectancies on their website and in subsequent reports, as well as producing a discussion paper outlining and justifying the changes.5 The October 2010 ABS report explicitly stated that: “Differences should not be interpreted as measuring changes in Aboriginal and Torres Strait Islander life expectancy over time”.2 As with any statistical analysis, the underlying issue is the quality of the data. As we continue to work to narrow the “true” Indigenous life-expectancy gap, we need to be mindful of the importance of accurate record-keeping, including the identification of Indigenous status, if future analysis of mortality statistics is to stand up to scrutiny.

Lachlan J McIver

To the Editor: Contrast induced nephropathy (CIN) is one of the most important and frequent complications of invasive coronary procedures.1 We have previously reported a multicentre randomised trial comparing use of iso-osmolar and low osmolar contrast agents (iopromide and iodixanol, respectively) in patients with pre-existing renal impairment.2 The overall proportion of patients developing CIN by Day 7 was around 25%, and there was no statistical difference between the different contrast media. There have been few prospective randomised controlled trials to determine the late effects on renal function and outcomes in relation to dialysis and mortality in patients who developed CIN after invasive coronary procedures. We report here on the long-term follow-up of patients enrolled in our initial study. Of the original cohort of 191 patients, 21 were excluded because of lack of follow-up information. We divided patients into two groups based on whether or not they had initially developed CIN, defined as an absolute increase in the serum creatinine concentration of at least 44 μmol/L or by a relative increase of at least 25% from the baseline value on Day 2 or 7 after exposure to the contrast media. The primary end point was persistent renal impairment, which we defined by these same criteria for serum creatinine, and alternatively, by an absolute reduction in estimated glomerular filtration rate (eGFR) of at least 10 mL/min/1.73 m2 (accounting for the coefficient of variation of creatinine and also the age-related decline in GFR3-5). The secondary end point was a composite of death and need for dialysis. Median length of follow-up was 43 months (interquartile range, 31–48 months). Latest serum creatinine results were available from physicians, hospital records or private laboratories for 157 patients. Significantly higher proportions of patients who had CIN at baseline showed evidence of persistent renal impairment compared with patients who did not have CIN at baseline, based on both serum creatinine results (20/40 v 31/117; P = 0.006) and eGFR (22/40 v 28/117; P < 0.001). Mortality was determined for all 170 patients by direct contact or from the national death registry. A significantly higher proportion of patients who had CIN at baseline (2/41) compared with those who did not have CIN (3/129) needed dialysis (P = 0.60). Twelve patients who had CIN at baseline had outcomes of death, dialysis or both, compared with 32 in the other group (P = 0.57). Multivariate analysis showed CIN at Day 2 or 7 was an independent predictor of persistent renal impairment (odds ratio, 3.31 [95% CI, 1.39–7.86]; P = 0.007). Age, diabetes mellitus, sex, hypertension, body mass index, contrast type and baseline eGFR were not predictive. This long-term follow-up showed that CIN after invasive coronary procedures is associated with increased risk of persistent renal dysfunction in patients with pre-existing renal dysfunction. Physicians should be alert to this complication.

Akash Dhawan · Devang Parikh · Ibrahim Shugman · John French · Hisham Hallani · Clyne Fernandes · Craig P Juergens

Objective: To describe how Medical Benefits Schedule (MBS) chronic disease (CD) item claims vary by sociodemographic and health characteristics in people with heart disease, asthma or diabetes.Design, setting and participants: A cross-sectional analysis of linked unit-level MBS and survey data from the first 102 934 participants enrolled in the 45 and Up Study, a large-scale cohort study in New South Wales, who completed the baseline survey between January 2006 and July 2008.Main outcome measure: Claim for any general practitioner CD item within 18 months before enrolment, ascertained from MBS records.Results: The proportion of individuals making claims for MBS CD items was 18.5% for asthma, 22.3% for heart disease, and 44.9% for diabetes. Associations between participant characteristics and a claim for a CD item showed similar patterns across the three diseases. For heart disease and asthma, people most likely to claim a CD item were women, older, of low income and education levels, with multiple chronic conditions, fair or poor self-rated health, obesity and low physical activity levels. The pattern of claims was slightly different for participants with diabetes in that there was no significant association with number of chronic conditions, smoking or physical activity.Conclusions: Many individuals with self-reported CD do not claim CD items. People with diabetes and individuals with greatest need based on health, socioeconomic and lifestyle risk factors are the most likely to claim CD items.

Kirsty A Douglas DipRACOG, MD, FRACGP · Laurann E Yen BSc, MPsych · Rosemary J Korda BAppSci, MAppSci, PhD · Marjan Kljakovic MB ChB, FRNZCGP, PhD · Nicholas J Glasgow BHB, MB ChB, MD

Objective: To examine the association of adolescent depression and anxiety symptoms with alcohol abuse or dependence in young adulthood.Design, setting and participants: Cohort study of the health and wellbeing of adolescents and young adults in Victoria, assessed at 8 waves (periods) of data collection, from age 14 to 24 years, between 1992 and 2003. Young people who participated in the cohort study at least once during the six adolescent assessment points (conducted 6 months apart, from age 14 to 17 years), at least once during young adulthood and who were alive at Wave 8 (n = 1758).Main outcome measure: Alcohol abuse or dependence assessed using the alcohol and substance abuse modules of the Composite International Diagnostic Interview at age 24 years.Results: Adolescents with moderate to high levels of depression and anxiety symptoms (measured by the revised Clinical Interview Schedule) had an increased risk of alcohol abuse or dependence in young adulthood, compared with young adults with low levels of adolescent depression and anxiety symptoms, after adjusting for potential confounding factors. Risk was higher for those with symptoms at more than two adolescent assessment points (odds ratio [OR] 1.9; 95% CI, 1.7–2.0) and for those with symptoms at one or two assessment points (OR 1.3; 95% CI, 1.2–1.4), compared with those with no above-threshold symptoms in adolescence.Conclusions: Adolescents with depression and anxiety symptoms are at increased risk for alcohol use disorders into young adulthood. They warrant vigilance from primary care providers in relation to alcohol use well into adulthood.

Maria McKenzie BBSc(Hons) · Anthony F Jorm PhD, DSc · Helena Romaniuk BSc, MSc, PhD · Craig A Olsson PhD · George C Patton MB BS, MD

Objective: To assess the relationship between dietary sodium intake, as measured by urinary electrolyte excretion, and blood pressure within a population of older Australian adults.Design, setting and participants: A cross-sectional study of adults enrolled in the Melbourne Collaborative Cohort Study, stratified by sex, country of birth (Italy, Greece, Australia/New Zealand) and age (50–59 and 60–75 years). Blood pressure measurements were taken in 2003–2007 and 24-hour urine collections in 2007–2008.Main outcome measures: 24-hour urinary excretion of sodium and potassium, urinary sodium-to-potassium ratio, and clinic blood pressure measurement.Results: The mean ± SD age of 783 participants was 64.0 ± 6.3 years. Mean ± SD urinary sodium was 155.1 ± 63.1 mmol/day (8.9 ± 3.6 g salt/day), urinary potassium was 82.3 ± 27.9 mmol/day, and urinary sodium-to-potassium ratio was 1.99 ± 0.83. In the 587 participants with blood pressure measurements, urinary sodium and the sodium-to-potassium ratio were both associated with systolic blood pressure in all adjusted and unadjusted models (mmHg change per 100 mmol/day increase in sodium: regression coefficient, 2.3, 95% CI, 0.1–4.6; P = 0.049, adjusted for age, sex, body mass index, country of birth and antihypertensive medication use).Conclusion: This study has demonstrated, for the first time within an Australian population sample of older adults, that sodium intake is positively associated with blood pressure. These results suggest that a population-wide reduction in sodium intake could be effective in reducing blood pressure in adults in Australia.

Catherine E Huggins BSc(Hons), PhD · Sharleen O’Reilly BSc(Hons), PhD, CertHEd · Maree Brinkman BSc(Nutrition), MND · Allison Hodge BSc, GradDipDiet, PhD · Graham G Giles BSc, MSc, PhD · Dallas R English BSC, PhD · Caryl A Nowson PhD, DipNutDiet

To the Editor: Recently, Professor Hoy argued for the full inclusion of Aboriginal and Torres Strait Islander people in the Australian Health Survey (AHS), including the measurement of clinical variables and the proposed sample repository.1 Although much of the argument is plausible, several points were overlooked that make it untenable overall. First, the current study design arose with input from at least five Indigenous representative bodies, including the National Aboriginal Community Controlled Health Organisation.2 They identified social and cultural issues as priority areas to be addressed — correctly so, as the underlying causes of health disparity are located in these domains, not primarily in the clinical and biomedical aspects of the AHS. The input from these major national bodies cannot be ignored. Second, yes — there are concerns that “the stored samples and their results might be somehow misused”.1 These concerns are legitimate and well founded in historical and contemporary experiences of Indigenous people. The argument for applying “current scientific and epidemiological knowledge, methods and safeguards”1 to the use of information held in the AHS is correct as far as it goes, but ignores equally important Indigenous knowledge and methodologies, Indigenous intellectual property issues, the principles of “ownership, control, access and possession” of Indigenous information,3 and certain aspects of the United Nations Declaration on the Rights of Indigenous Peoples. This position is therefore inconsistent with the National Health and Medical Research Council guidelines on values and ethics in Aboriginal and Torres Strait Islander health research, particularly as they relate to “survival and protection”.4 Third, denying Indigenous people control over how their health information is used by mainstream research institutions prevents accountability of researchers to communities. Using and publishing this information requires review by relevant experts, in this case Aboriginal and Torres Strait Islander community representatives. Biomedical expertise alone is insufficient to enable effective peer review and, at worst, it risks promoting destructive policies that ignore social, cultural and political realities for Aboriginal people and Torres Strait Islanders. Aboriginal people and Torres Strait Islanders rightly feel that they have been one of the most researched groups in history. And yet, even with this background of decades of being constantly studied, researched and examined, it seems that there is still not enough information being collected. Wellbeing is “grounded in the respect given to people, and the control afforded to them, in their daily lives”.5 Sometimes it’s up to Aboriginal and Torres Strait Islander people to identify what is important in Aboriginal and Torres Strait Islander health: it’s our health!

Kevin G Rowley · Alister H Thorpe

In reply: I thank Dr Rowley and Mr Thorpe for their response.1 It is hard to justify exclusion of any Australian from opportunities to participate fully in important initiatives on the recommendation of bodies whose membership sometimes has no direct link to the persons affected. There is no other population group in Australia to whom this applies. Medical and clinical approaches should complement initiatives to address critical social and cultural issues; they are not in competition nor mutually exclusive. The inclusion of health measures in the adult (but not youth) components of the Australian Health Survey (AHS) acknowledges that there is much to be learned and remediated clinically. Any interpretation of the deliberate exclusion of Indigenous youth from the “measures” elements of the survey is unsettling. There is more, not less, to be learned from this group. Their exclusion deprives policymakers of robust evidence that could improve health status. It condemns enquiry to the current sidestream method of short-term research projects on small pockets of people. These sometimes yield results of dubious generalisability and cause ongoing competition for the impossibly stretched research dollar. Alternatively, is it implied that Indigenous parents are less able to make sound decisions on their child’s participation or that the minors are less likely to cooperate? I suggest that the matter of participation in the AHS be aired through general media channels, as well as those with an Indigenous focus, such as “Living Black” (SBS television) and Imparja television, and through local Indigenous radio stations and community networks. With a developed sampling frame for Indigenous people, dialogue about elements of the examination should at least be conducted with the specific individual tribal groups or communities, if not with the targeted individuals (the preferred option). Subsequently, the whole issue of representation to policymakers in Indigenous health matters might be re-examined on a national basis.

Wendy E Hoy

Objectives: To determine issues that affect newly resettled refugees in accessing an emergency department (ED).Design, setting and participants: We conducted a descriptive community survey using a semistructured questionnaire. Newly resettled refugees from the Middle East and Africa were interviewed, statistical analysis was performed, and standard content analysis methods were applied to free-text responses.Main outcome measures: Emergency health-seeking behaviour, sociocultural barriers and beliefs about Australia’s emergency health services.Results: Half the African refugees (53/106) (50%), compared with only 15/49 (31%) of the Middle Eastern refugees, preferred an ED service over other forms of care for an urgent medical condition (P = 0.024). Qualitative data revealed that most newly resettled refugees understand how to use the emergency health services. However, while most indicated that they were able to make a call for emergency medical help, a substantial number of our respondents revealed that they were afraid to make such a call for fear of security implications, on the basis of experiences from their home countries.Conclusion: Reasons for differences in preferences of health care access, and determining how best to educate the community on the use of ED services, warrant further investigation. From a policy perspective, the increasing health care needs of refugees need re-examination when planning health care provision to refugees.

Mohamud Sheikh DrPH, MHSc, MIPH · Peter I Nugus PhD · Zhanhai Gao PhD · Anna Holdgate MB BS, MMed, FACEM · Alison E Short PhD · Ayman Al Haboub BPharm, MIPH · C Raina MacIntyre FRACP, MAppEpid, PhD

CorrectionMean number of days in hospital incorrectly expressed: In “Hospital and emergency department use in the last year of life: a baseline for future modifications to end-of-life care” in the 6 June 2011 issue of the Journal (Med J Aust 2011; 194: 570-573), the mean number of days in hospital was given per hospital admission for decedents with and without cancer. This should have read: “Decedents with cancer had a mean number of hospital admissions of 7.6 (SD, 10.2; median, 5) with the mean number of days in hospital per decedent of 41.3 (SD, 35.2; median, 34). Decedents with non-cancer diagnoses had a mean number of hospital admissions of 8.1 (SD, 23.1; median, 3) with the mean number of days in hospital per decedent of 49.6 (SD, 60.0; median, 30).” The html and pdf versions of this article are corrected

Lorna K Rosenwax · Beverley A McNamara · Kevin Murray · Rebecca J McCabe · Samar M Aoun · David C Currow

Objective: To assess the influence of area-level socioeconomic status (SES) on incidence and case-fatality rates for stroke.Design, setting and participants: Analysis of pooled data for 3077 patients with incident stroke from three population-based studies in Perth, Melbourne, and Auckland between 1995 and 2003.Main outcome measures: Incidence and 12-month case-fatality rates for stroke.Results: Annual age-standardised stroke incidence rates ranged from 77 per 100 000 person-years (95% CI, 72–83) in the least deprived areas to 131 per 100 000 person-years (95% CI, 120–141) in the most deprived areas (rate ratio, 1.70; 95% CI, 1.47–1.95; P < 0.001). The population attributable risk of stroke was 19% (95% CI, 12%–27%) for those living in the most deprived areas compared with the least deprived areas. Compared with people in the least deprived areas, those in the most deprived areas tended to be younger (mean age, 68 v 77 years; P < 0.001), had more comorbidities such as hypertension (58% v 51%; P < 0.001) and diabetes (22% v 12%; P < 0.001), and were more likely to smoke (23% v 8%; P < 0.001). After adjustment for age, area-level SES was not associated with 12-month case-fatality rate.Conclusions: Our analysis provides evidence that people living in areas that are relatively more deprived in socioeconomic terms experience higher rates of stroke. This may be explained by a higher prevalence of risk factors among these populations, such as hypertension, diabetes and cigarette smoking. Effective preventive measures in the more deprived areas of the community could substantially reduce rates of stroke.

Emma L Heeley MSc, PhD · Jade W Wei BPharm · Kristie Carter PhD · Md Shaheenul Islam MB BS, MPH, MSc · Amanda G Thrift PhD · Graeme J Hankey MD · Alan Cass PhD · Craig S Anderson MD, PhD

Australian Health Ministers have endorsed the hospital standardised mortality ratio (HSMR) as a key indicator of quality and safety, and efforts are currently underway towards its national implementation. In the United Kingdom, Canada, the Netherlands and the United States, the HSMR has been used for several years within organisations to monitor performance and response to various quality and safety programs. In the UK and Canada, the HSMR is also publicly reported and used to compare performance between hospitals. The validity and reliability of the HSMR as a screening tool for distinguishing low-quality from high-quality hospitals remain in doubt, and it has not yet been proven that HSMR reporting necessarily leads to worthwhile improvement in quality of care and patient outcomes. Institutions may respond to an unfavourable HSMR by “gaming” administrative data and risk-adjustment models or implementing inappropriate changes to care. Despite its apparent low cost and ease of measurement, the HSMR is currently not “fit for purpose” as a screening tool for detecting low-quality hospitals and should not be used in making interhospital comparisons. It may be better suited to monitoring changes in outcomes over time within individual institutions.

Ian A Scott MHA, MEd, FRACP · Caroline A Brand BA, MPH, FRACP · Grant E Phelps MBA, FRACP, AFRACMA · Anna L Barker BPhty, MPhty, PhD · Peter A Cameron MB BS, MD, FACEM

Objectives: To describe hospital and emergency department use in the last year of life by people for whom death from cancer or one of another nine conditions was an expected outcome.Design, participants and setting: Retrospective cross-sectional study based on death registrations and morbidity data for 1071 Western Australians who died between 1 August 2005 and 30 June 2006. Decedents had an informal primary carer, did not live in residential aged care and died of a condition amenable to palliative care.Main outcome measures: Total number of hospital admissions; emergency presentations (with and without hospital admission); days spent in hospital by age group at death, sex, metropolitan or rural place of residence and cancer versus non-cancer diagnosis; proportion in hospital on any day in the last 365 days of life; time points of change in the last 365 days of life at which there was an increasing proportion of hospital admissions for those with cancer and non-cancer conditions.Results: All but 4% of the decedents spent time in hospital with a marked increase in hospitalisations in the last 108 days of life for people who died of cancer and the last 83 days of life for people who died of non-cancer conditions. Those with cancer spent less time in hospital than those with other diagnoses. Seventy per cent of the cohort had at least one emergency presentation. On the last day of life, 61.5% of people were in hospital and 4.0% had been seen in emergency departments.Conclusions: Western Australian hospitals currently provide extensive and progressively greater care at the end of life. Identifying patterns of emergency and inpatient use for various disease trajectories will assist in the planning of appropriate services for people where death is an expected outcome.

Lorna K Rosenwax BAppSc(OT), MSc, PhD · Beverley A McNamara BA(Hons), Dip(Teach), PhD · Kevin Murray BSc(Hons), MSc · Rebecca J McCabe BBus, GradCertBus, MCom(Bus Admin) · Samar M Aoun BSc(Hons), PhD, MPH · David C Currow BMed, FRACP, MPH