Letters

To the Editor: I would like to add my perspective, after nearly 50 years of experience in Australian medicine, to the very welcome articles by Moynihan1 and Webber.2 A few months after Medibank (Mark I) was launched in 1975, a colleague and I wrote a satirical article called “How to rob Medibank blind”.3 We deliberately wrote this article anonymously, so that it could be judged on its merits, and not by its authors. It was picked up by the Sydney Sunday Telegraph and featured across its middle pages. The responses of the then Minister for Health, Ralph Hunt, and Australian Medical Association (AMA) president, Lionel Wilson, were, to paraphrase their words, “doctors are not like that!” This view was shared by Medibank architect, economist John Deeble, when I asked, 20 years later, why he and his co-architect Richard Scotton had designed “a mechanism with an accelerator pedal but no brakes”. They had, he said, not thought it necessary. Around that time, on behalf of the AMA, I advised on the restructuring of the Professional Services Review (PSR). I had been fortunate, meanwhile, to have chaired Professional Standards Committees of the New South Wales Medical Board, and to have sat on the bench of the NSW District Court Medical Tribunal, looking at alleged poor professional conduct. This opened my eyes to the small number of doctors whose standards were poor overall, and led to the Board’s performance assessment program. Many had also been investigated by Medibank or Medicare. Instead of having the PSR chasing illusionary and indefinable “over-servicing” by general practitioners (curiously, specialists were almost never investigated), I suggested peer review, whereby trustworthy, practising GPs or relevant specialists could advise on whether or not a doctor’s pattern of practice was inappropriate. This could be judged on an overall view of their work, much as was being done in the NSW Board’s Performance Program. Another decade later, as Chairman of the Doctors Health Fund (then the AMA Health Fund), I attended a small dinner in Melbourne with the chairmen of a few restricted membership funds and the then Minister for Health, Tony Abbott, who asked me to sit next to him. I asked pointedly how he could justify commercial corporations’ continuing to make profits for their shareholders out of Medicare benefits which were supposed to reflect the value of a doctor’s professional services. How could there be so much fat in the system that these corporations could cream off their substantial profits? His disappointing response was that it was too difficult to do anything about it. Webber is correct — both sides of politics are to blame.

Peter C Arnold

To the Editor: An ethical concern that was raised by Conway in his article supporting the freedom to practise in accordance with conscience1 related to the problems that arise when patients — specifically children — are unable to express a considered opinion on any conflict between their clinicians and their legal guardians about their best interests. Parents know their child best, are responsible for continuing care of the patient (including during any adverse effects of treatments) as well as of any other dependants. However, they may also be guided more by strongly held beliefs than by the views of clinicians on the best interests of their child. The law gives parents considerable rights and obligations to make decisions in the best interests of their children, but the law also recognises that children have rights, independent of the parents. When children are able to form and express opinions, these are considered; when they can’t, they are dependent on others to promote their interests. Within this legal framework of competing rights, the solution can only be a legal one, as occurred with the patient in Conway’s example.1 Clinicians have to be prepared to go to court to put their case for the best interests of the child, and then accept the legal decision with good grace, however personally distressing. This is just another example in medicine of the need for all to work with less than desirable outcomes. And one must always remember that not complying with a court order is an option — with consequences — that may seem ethically justifiable to a clinician or parent.

William R Adam

To the Editor: In a recent Opposing Views article, Conway claims that doctors should be able to practise according to their personal values,1 but the argument he uses is not persuasive. The ethical dilemma in the case is not characterised correctly. Autonomous choices can be made while treatment options, based on scientific knowledge, are available. If there are no medical options to improve the outcome — if we are at the very limits of medicine — there is no duty to provide anything other than palliation. In other words, the “do no harm” principle takes precedence over the duty to treat the patient according to his or her needs. Under these circumstances, refusing to “treat” is to practise with medical knowledge and the comfort of the patient (in Conway’s example, a child) in mind, regardless of the parents’ wishes. There is no ethical dilemma in that sense. In fact, the dilemma doctors have is not clinical, but a professional one created by the court order compelling doctors to continue futile treatment, causing unnecessary discomfort that obviously conflicts with the “do no harm” principle. Hence, Conway’s case does not support his argument. In addition, his argument may lead to the violation of patient wellbeing and of professional values. The moral problem related to defining a space for personal values in medical care is that they may conflict with professional values, legitimising discrimination. Then it would be nearly impossible to criticise doctors and institutions that refuse abortion, do not examine patients of the opposite sex on religious grounds, and refuse to operate on HIV-positive patients or to treat people with different political affiliations. It would be hard to call that kind of environment a “healthy diversity”. As well as defending the right to health, we should protect the professional value of non-discrimination to preserve trust between medicine and society. In fact, in order to “safeguard against abuses of power, error and exploitation in medicine”,1 we have already established universal professional codes, instead of leaving the stage to the relativity of personal values.

Murat Civaner

To the Editor: In his Opposing Views article, Travis claims that fee- for-service (FFS) “provides the best transparency, accountability and incentive for everyone”.1 However, FFS models reward volume and intensity, rather than quality of outcomes. Evidence suggests that FFS results in increased numbers of patient visits, investigations and procedures,2 which contribute to inflation in the cost of health care.3 Further, an FFS model is “transparent” only on a most superficial level, because doctors are paid for the number of items they deliver. What matters most is transparency about whether the care being delivered to patients is in those patients’ best interests. Patients do not have the same grasp of technical information as their doctors. They assume, with justification, that most doctors perform their clinical practice not primarily to serve their own incomes, but patients’ best interests. Regardless, it is undeniable that the reward structure associated with an FFS system provides powerful incentives for delivering more services, with clinicians subsequently inducing increased demand for medical services.4 On the other hand, the problem with purely salaried providers is that they are more likely to underservice their patients, with an increase in the length and a decrease in the number of consultations.2 Clearly, neither of these payment systems is optimal for driving quality or efficiency in resource use. The ideal outcome from health care reform would be one in which the focus is on delivering high-quality, efficient care to patients, rather than high-quantity care. In designing the optimal payment structure, the aim should be to align the incentive with the quality and appropriateness of the care that is delivered — to reward doctors for their efforts to provide good health outcomes for patients.3 Our current health system provides much lower rewards for the doctors who practise much of our preventive care (general practitioners and physicians) than it does for procedural specialists. This may not be the most cost-effective strategy, and part of the debate about how doctors are paid should be directed at this imbalance. There has been experimentation with “capitation combined with pay-for-performance” or “shared savings” models, which are promising attempts to find the best balance of incentives for cost control and quality improvement.5

Matthew H R Anstey · Stephen P Gildfind

To the Editor: In 2009, we calculated that an estimated 50 405 prison inmate separations occurred in Australia in the 2007–08 financial year,1 and argued that the significant mortality risks associated with release from custody necessitated accurate and routine dissemination of such information. We believe this is still required, and that reporting the most current estimate of inmate separations may help facilitate better through-care and post-release health service provision. Our previous estimates have been used to contextualise discussions of prisoner health,2 estimate the number of post-release deaths,3 and advocate for needle and syringe programs in Australian prisons.4 Consistent with the approach detailed elsewhere,5 we estimated the number of inmate separation episodes occurring in Australia for the financial year 2008–09, using benchmark data from public documents on the website of each state and territory government department responsible for prisons. We attempted to obtain the total (ie, sentenced and on remand) number of inmates released from prison in 2008–09. This figure was available for Victoria (5458), South Australia (4489) and the Northern Territory (3078). For New South Wales, data could only be obtained for separations of sentenced prisoners (8941). To adjust for separations of prisoners on remand in NSW, we assumed that separations of sentenced prisoners comprised 38% of all separations, reflecting the ratios of separations in other jurisdictions. Hence, the total estimated number of separations in NSW was 21 468. Adding this to the other three figures gave a total of 34 493 separations in four jurisdictions that collectively hold 63% of the national prisoner population. A multiplier of 1.59 (1/0.63) was applied to this figure to produce a national estimate of 54 751 prison separations in 2008–09, each of which is associated with a significant increase in risk of death. Our estimate represents an 8.6% increase on the estimate from the previous financial year. This substantial increase suggests that routine reporting of actual incidents of separation, as well as numbers of unique individuals released from prison, remains vital to the provision of adequately scoped and resourced post-release health and welfare services in Australia.

Kristy A Martire · Sarah Larney

To the Editor: Significant mortality and morbidity occur among pedestrians involved in road traffic incidents.1 The National Road Safety Strategy 2001–20102 aimed to achieve a 40% decrease in pedestrian fatalities in Australia through road safety measures, including urban speed limit reductions. To assess the impact of such measures, we examined the long-term mortality trend in pedestrians presenting to an inner Sydney major trauma centre and determined whether incident location was a predictor of major trauma (defined as in-hospital mortality, intensive care unit admission and/or Injury Severity Score > 15). We identified 1944 trauma registry records of adult patients admitted to Royal Prince Alfred Hospital between 1992 and 2010 as the result of a pedestrian incident within the hospital’s catchment area, excluding transfers from other facilities (Box 1). Incident locations noted on clinical and ambulance records were classified as city or suburban according to current City of Sydney local government area boundaries. Population data were obtained from the Australian Bureau of Statistics. From 1992–1994 to 2007–2009 (population data for 2010 were unavailable at time of analysis), the age-adjusted mortality rate fell from 6.25 to 2.31 per 100 000 population (Box 2), a relative decrease of 63% (incidence rate ratio [IRR], 0.94; 95% CI, 0.91–0.98; P = 0.003). Negative binomial regression analysis showed that the reduction in mortality was statistically significant for incidents in the city (IRR, 0.91; 95% CI, 0.86–0.96; P = 0.01) but not suburban areas (IRR, 0.96; 95% CI, 0.92–1.00; P = 0.09). Multivariable logistic regression showed the odds of major trauma were 70% higher for suburban incidents than city incidents after adjusting for age, sex and time of presentation (after hours and weekends) (adjusted odds ratio, 1.7; 95% CI, 1.1–2.6; P = 0.02). The decrease in pedestrian mortality appeared most prominent after 2003, around the time the urban speed limit was reduced from 60 km/h to 50 km/h, and to 30–40 km/h in areas with high pedestrian activity.3 The 10 km/h speed limit reduction alone is estimated to have resulted in a 6% decrease in pedestrian fatalities.4 However, the trend appears to be limited to city incidents, with an annual reduction of 10% in the city versus 3% in suburban areas (Box 1). The observed improvement in outcomes, particularly in the City of Sydney, is likely a result of speed-limit reductions, drink-driving laws and improvements in clinical care. However, more needs to be done to protect pedestrians in suburban areas. 1 Baseline characteristics and outcomes of pedestrian incidents City (n = 883) Suburban (n = 1061) P Male 561 (47%) 625 (53%) 0.04 Median age (IQR) 34 (24–53) 46 (29–67) < 0.001 Median ISS (IQR) 4 (2–9) 5 (4–17) < 0.001 ICU admission 124 (11%) 241 (19%) < 0.001 Median hospital length of stay, days (IQR) 3 (1–9) 5 (1–13) < 0.001 Major trauma* 175 (20%) 336 (32%) < 0.001 Deaths 32 (4%) 76 (7%) 0.001 Crude mortality rate (95% CI)† 1992–1994 2.1 (− 0.4, 4.7) 1.5 (− 0.03, 3.0) 1995–1997 1.2 (− 0.6, 3.0) 2.3 (0.4, 4.4) 1998–2000 0.9 (− 0.6, 2.4) 2.0 (0.2, 3.8) 2001–2003 1.6 (− 0.3, 3.6) 1.5 (− 0.03, 2.9) 2004–2006 0.3 (− 0.5, 1.2) 1.3 (− 0.09, 2.7) 2007–2009 0.3 (− 0.5, 1.2) 0.9 (− 0.2, 2.0) Annual change over 18 years‡ − 10% − 3% IQR = interquartile range. ISS = Injury Severity Score. ICU = intensive care unit. * ISS > 15, requiring ICU admission, and/or death. † Rate per 100 000 population among adult pedestrians averaged over 3-year periods. ‡ Calculated using the formula: change (%) = (P1/P01/n − 1) × 100 where n is number of years observed − 1, P1 is the final average rate and P0 is the initial average rate. 2 Age-adjusted mortality rate per 100 000 population for adult pedestrians, 1992–2009* * Age standardised to Australian Bureau of Statistics 2006 census data (bars represent 95% confidence intervals).

Matthew Oliver · Michael M Dinh · Susan Roncal · Soufiane Boufous · Bernardino Branco · Christopher M Byrne

To the Editor: Tenofovir disoproxil fumarate is a nucleotide analogue reverse transcriptase inhibitor that is used in Australia as first-line antiviral treatment for HIV infection.1,2 Tenofovir may be nephrotoxic, particularly affecting proximal tubular function.3 We report a case of tenofovir-associated Fanconi syndrome, which demonstrates the need for vigilance in patients taking tenofovir. A 17-year-old rurally residing boy with perinatally acquired HIV infection presented to a tertiary referral hospital with renal impairment. His serum creatinine level had been normal 4 years previously. He had had severe leg pain and weakness for over a year, and was unable to mobilise without a walking frame. Because of his illness, he had stopped studying. He had commenced antiretroviral therapy in 1996, when he was 3 years old, and at presentation was taking tenofovir plus emtricitabine and lopinavir plus ritonavir. Follow-up with his local clinician was intermittent. The patient’s HIV viral load was undetectable, and he had a CD4 count of 0.62 × 109 cells/L. His serum creatinine level was 150 μmol/L (reference interval, 60–110 μmol/L) and his estimated glomerular filtration rate (eGFR) was 52 mL/min/1.73 m2 (normal value, > 90 mL/min/1.73 m2). He had all the features of Fanconi syndrome — hypophosphataemia with renal phosphate wasting, glycosuria, aminoaciduria, a reduced serum uric acid level and proteinuria (1.93 g/day). He also had hypokalaemia and acidosis, and his urine tested positive for β2-microglobulin. Histological analysis of a renal biopsy specimen showed proximal tubular abnormalities (Box). In addition, he had severely reduced bone mineral density, elevated bone turnover markers and vitamin D deficiency. The patient’s tenofovir therapy was ceased, he was started on an alternative antiretroviral regimen, and he was given phosphate, potassium, bicarbonate and calcium supplementation plus vitamin D and calcitriol therapy. Two months later, his serum creatinine level was 130 μmol/L, his eGFR was > 60 mL/min/1.73 m2, and his proteinuria, glycosuria and aminoaciduria had resolved. His leg pain and weakness resolved within another month. He has since been able to recommence full-time study. The persistent biochemical and renal abnormalities in this patient may have been a result of delayed recognition of Fanconi syndrome leading to prolonged illness. Risk factors for tenofovir-associated nephrotoxicity include a GFR of less than 90 mL/min/1.73 m2, use of nephrotoxic medications, comorbidities (eg, diabetes and hypertension) and use of some protease inhibitors. Patients on tenofovir should be screened at least 6-monthly for eGFR, serum phosphate levels, proteinuria and glycosuria.4 Three-monthly testing is also suggested in the initial year, due to the occurrence of tenofovir-associated nephrotoxicity without risk factors.5 Urinary protein to creatinine ratio is usually increased in tenofovir-associated nephrotoxicity, and tests for some urinary proteins may be useful in subtle cases of the condition.3 Testing for albuminuria (a marker of glomerular disease) and sole reliance on eGFR are insufficient for detecting tenofovir-associated tubular dysfunction.3 Due to potential renal toxicity, optimal outcomes for patients on tenofovir require careful monitoring of patients and close liaison between treating practitioners. Renal biopsy specimens showing proximal tubular abnormalities in a patient with tenofovir-associated Fanconi syndrome A: Prominent changes of acute tubular necrosis (haematoxylin and eosin stain; original magnification, × 4). B: Acute tubular necrosis with proximal tubular eosinophillic inclusions (arrows) representing giant mitochondria visible by light microscopy (haematoxylin and eosin stain; original magnification, × 10).

David M Gracey · Mangalee Fernando · John Ziegler · Christopher P White · Jeffrey J Post

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

To the Editor: I am disappointed by Williams and colleagues’ blinkered view on “conflict of interest”.1 Their focus is solely on that of influence by profit-making medical supplies companies, often loosely referred to as “big pharma”. However, many careers are made and lost by government-influenced appointments and funding. The careers of some doctors in government-funded positions may hinge on whether they support the health department stance on a particular guideline. Rarely have I seen a future promotion, government research funding, committee appointment or professorial sinecure listed as a conflict of interest. This, I believe, is the real elephant in the room.

Peter J McLaren

To the Editor: We read with interest the article by Pinto and colleagues regarding colonisation of reusable tourniquets by multiresistant organisms (MROs).1 We highlight that surgical adhesive tape also has the potential to act as a significant fomite in health care settings. We collected partially used surgical tape rolls from several clinical areas of three hospitals in the Hunter New England Area Health Service. Using hands disinfected with alcohol gel, tape rolls from different locations in each area were placed into 21 clean collection bags (up to three tapes per bag). Tapes from each batch were placed in 21 sterile containers with 15 mL of brain–heart infusion broth and incubated overnight at 35°C in carbon dioxide. The broths were subcultured onto Columbia horse-blood agar (Oxoid Australia, Adelaide, SA), MacConkey agar (Oxoid) and differential selective media to detect vancomycin-resistant enterococci (VRE) (chromID VRE; bioMérieux, Marcy L’Étoile, France), methicillin-resistant Staphylococcus aureus (MRSA) (Brilliance MRSA; Oxoid) and multiresistant gram-negative bacteria (chromID ESBL; bioMérieux). A multiplex tandem polymerase chain reaction assay (MRSA4; AusDiagnostics, Sydney, NSW) to detect MRSA and methicillin-susceptible S. aureus (MSSA) was also performed on all broth cultures. Routine species level identification was performed (VITEK MS; bioMérieux). Susceptibility was determined in accordance with Clinical and Laboratory Standards Institute criteria.2 In 11 of the 21 tape batches, MRSA and/or VRE were identified. Of these, four were positive for MRSA and 10 for VRE, with three positive for both. MSSA was identified in two, both in association with VRE. All batches showed evidence of contamination with other bacteria such as Bacillus cereus, coagulase-negative staphylococci, non-multiresistant Enterobacteriaceae, Pseudomonas spp, Acinetobacter spp and other enterococci. Our results indicate that surgical adhesive tapes are frequently contaminated with MROs. Interpretation of these results is limited by the small number of tapes and clinical areas sampled, and the difficulty of proving a relationship to clinical infection. However, items such as intravenous cannulae, surgical drains and wound dressings are frequently fixed using surgical adhesive tape. This may lead to colonisation and subsequent infection. Furthermore, tape rolls are often left lying on contaminated surfaces, are handled by multiple individuals and cannot be disinfected. Surgical adhesive tape is a potential reservoir of pathogenic bacteria3 and fungi4 and was implicated in a prolonged S. aureus outbreak in a neonatal unit.5 The role of surgical tape as a potential fomite was reported in 19746 but has not been widely acknowledged since. Removing the outer layer of the tape roll is unlikely to reduce contamination, given visible contamination of the side of many rolls (Figure).3 Short rolls of surgical adhesive tape should be supplied in sealed packets and used for individual patients, only after hand disinfection, and discarded after use.

Patrick N A Harris · Chris Ashhurst-Smith · Sandy J Berenger · Alison Shoobert · John K Ferguson

To the Editor: Antimicrobial resistance has been identified as a major concern in Australia, particularly as few new antimicrobial agents are being developed.1 Studies suggest that up to half of antimicrobial agents prescribed in hospitals are inappropriate.2-4 Antimicrobial stewardship interventions, including dissemination of clinical guidelines and restrictions on antimicrobial formularies, may not be fully able to account for the complex indications for antimicrobial use in hospitalised patients. We instituted a rapid clinical audit and feedback system of patients on one of 14 restricted antimicrobial agents as a component of antimicrobial stewardship activities in 2011 at the Alfred Hospital in Melbourne. Multidisciplinary antimicrobial stewardship ward rounds involving a senior clinical pharmacist, an infectious diseases (ID) registrar and an ID physician were performed on weekdays. Patients in units (respiratory, haematology/bone marrow transplantation, burns and intensive care) who had existing liaison ID services were excluded from our analysis. Patients included in the analysis were prescribed a restricted antimicrobial, either for an indication outside hospital policies, or where approval had not been obtained through a web-based antimicrobial approval system. Between January and April 2011, 473 patients were identified as requiring review by the antimicrobial stewardship team. In total, 236 recommendations (Box) were made for 158 patients across all 18 units (73% surgical, 27% medical). For other patients, antimicrobial use was deemed clinically justified or the antimicrobial agent had been ceased at the time of review. Recommended changes to therapy involved ceftriaxone (n = 70), piperacillin/tazobactam (n = 23), ciprofloxacin (n = 22) and vancomycin (n = 18). A formal referral to the ID consultation service was made for 11 of the 236 patients (5%). Recommendations were followed in 78% (184/236) of cases; acceptance of recommendations was higher when review involved the ID physician (146/176, 83%) than when it did not (38/60, 63%; P = 0.002). Rapid clinical review by a multidisciplinary antimicrobial stewardship team was able to assess large numbers of patients requiring restricted antimicrobial agents. Postprescribing evaluation has been reported mainly in North American studies, but has not been reported in Australian hospitals.5 In most patients, agreement was reached with the treating clinicians to cease or reduce the use of antimicrobial agents, suggesting that their ongoing use was not clinically justified. In addition, regular ward rounds by the team supported compliance with the antimicrobial approval system, provided education to junior medical staff and identified antimicrobial use protocols that had no basis in evidence. Proportion of the antimicrobial stewardship team’s recommendations accepted Recommendation Accepted Stop antimicrobial agent 72% (56/78) Change drug dose 75% (33/44) De-escalate antimicrobial cover 83% (25/30) Change to oral antimicrobial agents 90% (26/29) Change drug to alternative 88% (21/24) Formal infectious diseases consult 64% (7/11) Initiate antimicrobial 73% (8/11) Additional management 100% (4/4) Additional diagnostic procedures 67% (2/3) Therapeutic drug monitoring 100% (2/2) Total 78% (184/236)

Kelly A Cairns · Adam W J Jenney · Sushena Krishnaswamy · Michael J Dooley · Orla Morrissey · Sharon R Lewin · Allen C Cheng

To the Editor: The median age of Australian patients at first diagnosis of cancer is 67.8 years.1 In advanced, incurable cancer, goals of treatment include symptom control for all patients and prolongation of survival by weeks to months in a subgroup. In older people, treatment decisions can be complicated by comorbidities, polypharmacy, frailty and cognitive impairment. Few studies have investigated older patients’ information needs and preferences for involvement in decisions about their care.2,3 We performed an exploratory study to investigate whether the health status of older cancer patients predicted their information needs, decision preferences and their oncologists’ treatment recommendations. Concordance between patients’ stated preferences and the perceptions of their oncologists was also measured. Fifty outpatients with advanced lung (n = 30) or bowel (n = 20) cancer, with a mean age of 66.9 years (range, 51–91 years) participated. Fourteen patients (28%) were older than 70 years and 80% were diagnosed with advanced cancer within the previous 4 months. Patients’ health status was measured using the Vulnerable Elders Survey (VES) 13, a validated questionnaire used to identify older persons at risk of health decline (indicated by scores of ≥ 3).4 Role preferences were elicited using the Control Preferences Scale.5 Thirteen patients (26%) had VES 13 scores of ≥ 3. Of these, five (38%) were over 70 years. Twenty-six of 49 patients (53%) wanted a passive role in decision making, and 29 of 50 patients (58%) wanted prognostic information. Age and VES 13 scores did not predict patients’ role preferences or desire for prognostic information. Oncologists were less likely to recommend chemotherapy for patients over 70 years (25% v 75%, P = 0.04) or for those who had VES 13 scores of 3 or above (20% v 80%, P = 0.02). Concordance between patients’ participation preferences and oncologist perceptions was 54%. Until there is more evidence from larger studies of patient preferences, oncologists should ask patients their preferences about decision making and prognostic information. Patients with advanced cancer, by characteristic and preference Characteristic Prefer non-passive decision control* Prefer prognostic information Total patients (n = 50) 23 29 Age 50–60 years 6 7 Age 61–70 years 11 14 Age 71–80 years 5 5 Age > 80 years 1 3 Male 10 18 Female 13 11 Born in Australia 15 14 Born outside Australia 8 15 Bowel cancer 13 19 Lung cancer 10 10 0 comorbidities† 3 5 1 comorbidity† 3 5 2 comorbidities† 4 7 ≥ 3 comorbidities† 12 12 0 concomitant medications‡ 1 4 1 concomitant medication‡ 2 3 2 concomitant medications‡ 7 5 ≥ 3 concomitant medications‡ 13 15 Lives alone 8 9 Lives with someone 15 20 Lives in aged care facility 0 0 ECOG PS = 0–1 20 24 ECOG PS = 2–3 2 4 Married 9 13 Not married 14 16 Primary education§ 3 5 Secondary education§ 14 14 Tertiary education§ 5 10 VES 13 score = 0–2¶ 17 19 VES 13 score = > 3¶ 6 10 ECOG PS = Eastern Cooperative Oncology Group performance status; higher score = poorer status. VES = Vulnerable Elders Survey. * Total in passive and non-passive categories was 49 (one patient gave no preference). † Missing data for one patient. ‡ Missing data for three patients. § Missing data for two patients. ¶ VES 13 scores range from 0–10; scores > 3 indicate vulnerability to health decline.

Lakshmi P Venkateswaran · Phyllis N Butow · Jesse Jansen · Nicholas R C Wilcken · Mark K Wong · Rina Hui · George Szonyi · Val J Gebski · Vasi Naganathan · Lisa G Horvath · Martin H N Tattersall

To the Editor: Multidisciplinary meetings (MDMs) are recognised as crucial to best practice in oncology, and governments have designated MDMs as key priorities in cancer care.1,2 Improvement in overall survival has been promoted as a potential benefit of multidisciplinary care, with little supporting evidence. However, consensus MDM decisions are worthless unless action is subsequently implemented. This retrospective audit aimed to assess whether MDM recommendations were implemented, as an indirect measure of the MDM process on patient outcomes. MDM records for patients discussed at the Austin Health (Melbourne) Uro-Oncology, Upper GI (gastrointestinal) and Colorectal Cancer MDMs between February and April 2010 were reviewed, and consensus recommendations were compared with treatment plans documented in the medical record. Reasons for change in management were classified as (1) new clinical information unavailable to MDM; (2) patient’s choice; (3) comorbidities or patient performance status; or (4) not otherwise specified. The Box summarises the results. Overall concordance was 76% (152/201 records). In 41 records (20%) in which an MDM discussion was documented, data were not available due to missing or incomplete MDM documentation in the medical record. Excluding records where data were unavailable, concordance was 95% (152/160 records), with discordance due to new clinical information (5), comorbidities or performance status (2) and patient choice (1). These results highlight a few issues. Documentation of MDM discussion and consensus decision is required for the MDM recommendations to be delivered to patients and communicated to other clinicians. This documentation may have been missed due to differences in recording systems (paper versus electronic; designated person recording) or MDM structure (pre-specified agenda versus ad-hoc discussion), which need to be addressed. The high concordance rate (where documentation was available) suggests that MDM management recommendations are being delivered and acted upon, and so effectively promote high-quality, evidence-based care. This assumes that the MDM always uses evidence-based medicine and includes all relevant members of the treating team. This aspect was not analysed as part of this audit. However, in support of this assumption, we found that neoadjuvant chemotherapy for stage T3 or greater muscle-invasive bladder cancer was not given before institution of MDMs in 2007 (0/19 potential patients), but is now regularly discussed at the Uro-Oncology MDM and is administered where appropriate (7/20 patients between 2007 and 2010). This suggests a positive effect of this MDM on access to a therapy known to improve survival. Our findings suggest that MDM recommendations are usually acted upon and improve patients’ access to appropriate treatment.3 Results of an audit of records of Uro-Oncology, Upper GI (gastrointestinal) and Colorectal Cancer multidisciplinary meetings, Austin Health, February – April 2010 Total no. of records Data not available, no. (%) Overall concordance, no. (%) Concordance when data available, no. (%) Discordance, no. (%) Uro-Oncology 118 32 (27%) 85 (72%) 85 (99%) 1 (1%) Upper GI 40 3 (8%) 34 (85%) 34 (92%) 3 (8%) Colorectal 43 6 (14%) 33 (77%) 33 (89%) 4 (9%) Total 201 41 (20%) 152 (76%) 152 (95%) 8 (4%)

George H Au-Yeung · Ahmad Aly · Andrew Bui · Carmel M Vermeltfoort · Ian D Davis

To the Editor: Return to driving following a stroke is a complex issue. Austroads provides general guidelines1 and the National Stroke Foundation recommends a process including off-road and on-road driving tests.2 On-road assessments conducted by occupational therapists are considered the gold standard for decision making on return to driving following a stroke. Limited access is available to off-road and on-road tests across Australia, with few occupational therapists qualified to assess driving ability. Additionally, a range of off-road assessments are used for patients who have had a stroke. General practitioners, who are instrumental in managing return to driving, often base their decisions on limited information regarding functional status, particularly in terms of vision, cognition and perception. Rehabilitation physicians generally have access to more detailed information from allied health staff on which to base their decisions on driving ability. I investigated whether rehabilitation physicians’ recommendations on driving following a stroke were associated with patients’ performance on two objective tools, which could be used in general practice to assist with decision making. Participants were recruited at two rehabilitation services in Adelaide, South Australia, using the following inclusion criteria: had been diagnosed with stroke; had driven before the stroke; were aged over 18 years; and had provided written informed consent. I performed two assessments. The first was the Useful Field of View (UFOV) assessment;3 this is a computer-administered assessment that analyses processing speed, divided attention and selective attention, and takes 20 minutes to complete. The second was the Stroke Drivers Screening Assessment (SDSA);4 this consists of three tests conducted at a table — dot cancellation, compass recognition and road sign recognition — and takes 45–60 minutes to complete. Both assessments have been validated in patients who have had a stroke by comparison to on-road assessment.5 Treating rehabilitation physicians, blinded to assessment results, were contacted to obtain their recommendations on driving ability based on their clinical assessment and feedback from allied health staff at a case conference. A total of 123 participants (98 men [80%]) were recruited, and diagnoses included 53 right hemisphere strokes (43%), 63 left hemisphere strokes (51%) and 7 other strokes (6%). The participants’ mean age was 67.3 years (SD, 13.5 years), median period since injury was 42 days (range, 7–2190 days) and mean amount of driving experience was 48 years (SD, 14.9 years). Results of the SDSA (n = 120) and UFOV assessment (n = 123) were significantly associated with rehabilitation physicians’ recommendations on driving (Box). This suggests that referring patients who have had a stroke for one of these assessments would provide GPs with objective information to guide decision making on driving. With the introduction of Medicare Locals, resources that allow GPs to refer patients for standardised off-road driving tests should be considered. Rehabilitation physicians’ recommendations on driving for patients with a stroke diagnosis and results of two off-road assessments Rehabilitation physicians’ recommendations (number of patients) Not medically fit to return to driving On-road assessment required Return to driving — no on-road assessment required P SDSA results Pass (n = 61) 5 36 20 0.001* Fail (n = 59) 24 29 6 UFOV results Processing speed Pass (n = 97) 18 55 24 0.02* Fail (n = 26) 13 11 2 Divided attention Pass (n = 59) 10 33 16 0.22 Fail (n = 62) 19 33 10 Selective attention Pass (n = 82) 11 49 22 0.001* Fail (n = 39) 18 17 4 Overall risk category Pass (n = 80) 12 46 22 0.007* Fail (n = 41) 17 20 4 SDSA = Stroke Drivers Screening Assessment. UFOV = Useful Field of View. * P values of < 0.05 were considered significant and indicate an association between rehabilitation physicians’ recommendations and results of off-road assessments.

Stacey R George

To the Editor: We noted with interest the recent publication by Bismark and colleagues1 about complaint-prone doctors in Victoria. This research makes a valuable contribution to the important task of identifying practitioners who are at high risk of delivering poor quality health care. Similar analysis of more than 32 000 complaints received in Queensland from 1992 to 2010 has been undertaken by the Health Quality and Complaints Commission (HQCC). A summary of this work, which was performed in collaboration with the Medical Board of Queensland, was presented recently.2 Between July 2006 and June 2010, only 5% of medical practitioners registered as practising in Queensland were the subject of a complaint in a given year. Medical practitioners with multiple complaints (four or more complaints over the period; 0.7% of all medical practitioners) have substantially more complaints (> 2 SD) than their peers (average number of complaints, 1.6). An analysis of a subgroup of 117 doctors who were each the subject of multiple complaints revealed that they were more likely to be men, to have specialist rather than general registration, and to have an Australian rather than a non-Australian first qualification. Surgeons made up over 40% of this subgroup and were more likely to be the subject of a complaint than were other specialists. Further work is underway to clarify the level of complaints relative to the number of registered practitioners within each specialty. Although there are differences in the samples and analytical methods used in the HQCC study compared with those used by Bismark et al, there are striking similarities in the central findings: a small proportion of doctors produce a disproportionate number of complaints, and past complaint history is predictive of future complaint risk. Health complaints commissions in Australian states are working closely with the Australian Health Practitioner Regulation Agency to explore remedial or preventive opportunities. This task needs to involve the health care industry and professional colleges, as these complaint-prone practitioners are often well known by their colleagues or employers but, equally often, not adequately managed. As there is evidence that unprofessional medical student behaviour predicts both unprofessional doctor behaviour and risk of disciplinary activity,3-4 medical schools should also be engaged in this task.

Matt Vance · Michael Ward · David McKenzie

To the Editor: Were the views expressed by Conway and Savulescu really opposing?1,2 I turned to those pages with much interest, only to find that it seemed as if they would have been in agreement, had you put the scenarios they each describe to both of them. Conway discussed the conflict between doctor and patient about acceptability of the recommended management (the “best interests” of a dying child). In contrast, Savulescu talked of irrational prejudices held by doctors on the grounds of race, sex and other factors. Indeed, unless I missed some subtle points, I could not see any opposition in their arguments. I was, therefore, stumped when your online poll asked me to vote on the question: “Do doctors have the right to refuse certain treatments on the grounds of personal conscience?” Is the question asking about Conway’s point about what the doctor sincerely, on medical grounds, considers to be in the patient’s best interests, or is the question asking about a doctor’s refusal to attend to a patient on the basis of some irrational prejudice? Horses of quite different colours — racing in different races.

Peter C Arnold

To the Editor: Alcohol misuse is one of the leading causes of preventable death, illness and injury in Australia because too many Australians drink too much alcohol too often. The evidence is clear on what are the most effective strategies to curb alcohol misuse at a population level.1 By far the most effective of these is increasing the price of alcohol by increasing alcohol taxes. Aside from some positive features, such as lower tax on low-alcohol and mid-strength beer and the higher tax on alcopops, the alcohol taxation regimen in Australia is flawed from both an economic and public health perspective. A review of the tax system led by Secretary to the Treasury, Ken Henry, concluded that “current taxes on beer, wine and spirits are incoherent”, and recommended taxation reform using a volumetric tax.2 The response from the federal government to this review has echoed the failures of previous Australian governments to avoid increasing taxes. An independent coalition called the National Alliance for Action on Action (NAAA) has been formed with the goal of reducing alcohol-related harm.3 The NAAA represents a broad-based alliance of major health and community organisations that advocate for more effective responses to Australia’s drinking problems. Recognising that there is no single solution, the NAAA has focused on three priority areas: alcohol pricing and taxation; alcohol marketing and promotion; and alcohol availability. The Australian Government’s national tax forum held on 4–5 October 2011 in Canberra was an ideal opportunity for an informed discussion about alcohol tax reform. However, despite acknowledgement by the government that the current system does not effectively target the health and social costs of alcohol abuse, alcohol taxation was not part of the tax forum agenda.4 This was disappointing. At an NAAA day of action on 6 July in Canberra, assurances were given by Treasurer Wayne Swan and Health Minister Nicola Roxon that alcohol tax would be discussed at the forum. These assurances were aired publicly and in a number of media forums. The forum discussion document indicates that this promise was not honoured.5 This complex issue has not been resolved. We feel it is vital for the health and wellbeing of the Australian community that the reforms recommended by the Henry review remain on the political agenda.

Christopher M Doran · Wayne D Hall · Brian R Vandenberg · Todd A Harper · Jane E Martin · Mike Daube

To the Editor: I read with interest the contrasting perspectives by Pirotta1 and Dwyer2 on the ethics of prescribing complementary and alternative medicine (CAM) interventions, and the subsequent letters on this subject published in the 17 October 2011 issue of the Journal. Individual CAM interventions should be assessed with the same evidence-based approach we would use for any other intervention. Furthermore, any evidence should be interpreted with the same caution as we do for conventional medicines. The efficacy of an intervention should be judged by the quality of the clinical evidence in the context of its scientific plausibility. Whether an intervention should be recommended depends on the context of the other known evidence-based management strategies and the patient’s individual clinical scenario. It is concerning that the supporters of the use of CAM in general practice appear to be willing to make recommendations for treatment in the absence of quality empirical evidence. Both Pirotta and Kotsirilos make the often-used argument that absence of evidence is not evidence of ineffectiveness.1,3 Pirotta suggests that we can turn to tradition1 and Kotsirilos implies that consumer demand indicates effectiveness.3 It should be acknowledged that both tradition and popularity are unreliable forms of evidence. The absence of quality evidence for an intervention should be a major barrier to recommending it, especially when known, effective alternatives exist. This is regardless of the philosophical tradition of the intervention. Even when there is evidence for efficacy, care must be taken in avoiding overreaching conclusions. For example, Kotsirilos interprets the relevant Cochrane review4 as supporting the use of cranberry for prevention of recurrent urinary tract infections in young women.3 That therapeutic recommendation is unjustified. Although the systematic review did find some evidence for cranberry, it noted problems with its quality and the lack of clarity of dosage and administration. The conclusion of the review in 2008 was that “further properly designed studies with relevant outcomes are needed”.4 A well designed randomised controlled trial was published earlier this year; it does not support the use of cranberry for this indication.5

Chun Wah M Tam

To the Editor: We recently reviewed the individual policies of seven Australian tertiary maternity hospitals, from across all states, regarding neonatal vitamin D supplementation. An Australian 2006 consensus statement concerning the treatment and prevention of vitamin D deficiency in children and neonates identified the need to implement vitamin D supplementation in neonates born to mothers with serum 25-hydroxyvitamin D (25-OHD) levels of ≤ 50 nmol/L.1 Despite this, we found that a number of institutions have recently adopted a policy of treating infants of mothers with an antenatal serum 25-OHD level of < 75 nmol/L. Treatment of the neonate is with oral stoss therapy (50 000 IU vitamin D as a single dose) and/or daily oral vitamin D therapy (1000 IU) until cessation of breastfeeding. It should be noted that previous research has suggested that a substantial proportion of Australian women of childbearing age have a serum 25-OHD level < 75 nmol/L.2 Vitamin D is a pleiotropic hormone that influences the expression of more than 200 human genes.3 It has a wide range of biological actions, and the full implications of vitamin D supplementation during early life are unknown. While it is clear that very low levels of vitamin D are associated with abnormal bone development and the risk of hypocalcaemic seizures, it is also possible that vitamin D supplementation may have adverse effects. For example, we and others have found that a history of vitamin D supplementation during early life may be associated with an increase in allergic outcomes such as asthma and hayfever in later life.4,5 These are observational findings, limited by potential confounding and reverse causation, but they highlight the potential concern of changes in medical practice without an accompanying updated evidence base. There is an urgent need for improved data and further debate before implementing a public health policy affecting as many as half of all Australian infants. In the interim, we would advocate treatment of neonates only if they fall within the 2006 guidelines (mother’s serum 25-OHD ≤ 50 nmol/L).

Kate M McCloskey · Natalie Wright · Anne-Louise Ponsonby · Peter J Vuillermin

To the Editor: The article by Baade and colleagues1 is valuable, but further research on clinical interventions and outcomes, as well as collaboration with other stakeholders, is required to improve care. The main result of Baade et al’s analysis of Queensland-wide population-based data was that overall survival of patients with rectal cancer decreases as distance from radiotherapy services increases. The article is important because it documents variation in overall survival of patients with rectal cancer according to where they live. Documenting outcome variation is a necessary first step along the path to improving service delivery. As Baade et al discuss, the next step is to identify the cause(s) of the variation. This next step could usefully investigate the quality of surgery. Non-randomised studies have shown that high-quality surgery with total mesorectal excision (TME) improves overall survival in patients with rectal cancer.2 In contrast, neoadjuvant or adjuvant radiotherapy has not yet been shown to improve overall survival, although, as Baade et al note, randomised controlled trials (RCTs) have shown reduced rates of local recurrence, with a corresponding improvement in disease-free survival.1 However, some experts have questioned the extent to which decreases in local recurrence rates, as measured in RCTs, lead to improvements in overall survival (given high-quality TME).3,4 Also, radiotherapy is not without risk (eg, sexual dysfunction, incontinence, bowel obstruction). Whether an individual patient (in consultation with his or her doctor) decides to have radiotherapy will depend on how he or she judges the trade-off between benefits and harms.4 In other words, identifying the cause(s) of variation in overall survival for patients with rectal cancer will not be easy because of the subtleties of everyday clinical practice. Moreover, the subsequent and final step in the pathway — implementing interventions to reduce unwarranted variation — is also not easy. We could make faster progress if there were closer collaboration among researchers in universities or institutes (who can generate new knowledge), budget holders in government (who can do something about modifiable causes of variation), and clinicians (who will be at the leading edge of any changes). A large amount of predictable health-services research is needed, and collaboration among researchers, budget holders and clinicians is as important as developing investigator-initiated ideas for research projects. Faster optimisation of service delivery could be achieved if budget holders were less suspicious of the motives of research groups,5 and if research groups put more emphasis on asking budget holders, “How can we help?”

Michael D Coory

To the Editor: We find ourselves in the curious position of criticising an article that supports improving access to radiotherapy. Baade and colleagues looked at cause-specific survival in all patients aged 20–79 years diagnosed with rectal cancer in Queensland.1 They found a 6% increase in cause-specific mortality risk for each 100 km increment in distance from the nearest radiotherapy facility. They concluded that the apparent poorer outcomes were due to a failure to receive radiotherapy. However, we suggest that their approach was significantly flawed in a number of ways. We have difficulty in understanding why, when information about the use of radiotherapy is available, this key explanatory factor was not actually measured. The size of the benefit they attribute to greater use of radiotherapy exceeds the expected improvements attributable to adjuvant radiotherapy compared with none at all.2 The authors state that pathological staging, another factor central to their analysis, is a reliable surrogate for clinical staging. However, they appear not to have made any allowance for the fact that neoadjuvant treatment leads to pathological down-staging. Finally, the main outcome, cause-specific survival, was taken from death certificates and is likely to be inaccurate. Use of radiotherapy is suboptimal in Australia.3 If the authors could identify the impact of distance on access to radiotherapy, this would be very valuable for cancer service planning. Clinical input is desirable in future investigations in this area.

Sean A Bydder · Nigel A Spry

In reply: We thank Bydder and Spry for their comments, although we suggest that they have misunderstood our study and its conclusions. We appreciate the opportunity to clarify our results. Our study provides clear evidence that, after rectal cancer diagnosis, differences in survival are directly correlated with how far patients live from radiotherapy facilities.1 We did not conclude, as stated by Bydder and Spry, that “the apparent poorer outcomes were due to a failure to receive radiotherapy”. This explanation would be simplistic and ignores the many, varied and complex factors (demographic, socioeconomic, geographical and clinical) that together contribute to poorer survival in regional and rural communities. Cause-of-death codes were based on case reviews by specialised coders using clinical information, in addition to the death certificate. Information on clinical stage is not available in population-based cancer registries and, for some cancers (including colorectal cancer), staging based on pathology reports has been shown to be a reliable substitute.2 Population-based data on radiation treatment are not available in Queensland. Finally, whether or not distance from radiotherapy facilities has a direct impact on access to radiation treatment, while outside the scope of our study, is an important question and, with our clinical collaborators, is indeed one focus of the next phase of our work.

Peter D Baade · Paramita Dasgupta · Joanne F Aitken · Gavin Turrell

To The Editor: Bellomo quite rightly points out that medicine is now in an era when it is recognised that the amount of evidence being generated far surpasses most individual doctors’ ability to adequately deal with it.1 He suggests that “knowledge management may now be one of the major challenges of modern medicine”. However, is this a problem only needing to be confronted by those in medicine? I would suggest that it needs to be confronted by the Australian health care system as a whole, because a learning health care system is attainable only through implementation of system-wide change.2 In the recent final report of the National Health and Hospitals Reform Commission, three out of the five levers for action required to create an agile and self-improving health system involve elements of knowledge management.3 Hence, there is now much that needs to be done to achieve appropriate knowledge management across the Australian health care sector. We first need to recognise that knowledge is not only evidence that is gleaned from scientific studies (explicit knowledge) but also involves knowing what others are doing (tacit knowledge), as well as building on experience.4 The term “knowledge management” refers to the capacity to manage all of this at an organisational level. Further, we need to determine whether knowledge management is a problem that can only be solved within Australia by more resources being given to national bodies,1 or whether more innovative approaches are required. As knowledge management is now an issue for all organisations in the health sector, there is much to be gained from adopting an organisational learning focus2-4 as well as by understanding that an alteration in values (and culture) is required.4,5 Widespread acceptance of the need to identify and disseminate best practices will be required, even if it involves using evidence from international organisations to prevent both redundancy and duplication of effort locally.2,4

Deborah J Verran

To the Editor: Health data can have an important role in alerting product safety regulators to consumer product-related injuries. Such injuries are a significant public health concern, with an estimated 173 000 incidents occurring each year in Australia, many of which require medical treatment.1 The new Australian Consumer Law (ACL; http://www.consumerlaw. gov.au), enacted in January 2011, increases safety requirements and recognises that industry members have an important injury prevention role. The ACL requires suppliers to report to the Australian Government when they become aware of serious injuries, illnesses or deaths associated with products they supply.2,3 Monitoring product safety issues under the ACL therefore relies on systematic reporting of injuries by consumers to suppliers, and by suppliers to safety regulators, although the extent of compliance is unknown. Given that medical treatment is an indicator of injuries serious enough to warrant mandatory reporting, monitoring of products involved in injuries that require treatment in hospital emergency departments is a logical place to focus initial attention. Australia currently has an array of injury data, such as emergency department data, morbidity and mortality data and specialised injury surveillance collections, that could be used for this purpose, without the need for new, expensive data collections. We conducted a pilot study of product-related injuries in children in Queensland, which identified significant potential for using existing injury data more effectively in product safety surveillance.4 A comprehensive national evaluation of routinely collected health data would enable product safety regulators to better understand and use these data. With a lack of exposure to the ACL in the health sector, clinical staff and injured parties may not be aware of reasons for and mechanisms of reporting product-related injuries. It is important that medical professionals are made aware of the law and informed about actions they can take to support this system. Engaging emergency department staff could be a first step for product safety regulators. This could be as simple as a well publicised free-call number or a dedicated email address that clinicians can use to access information or report injuries quickly, without the need to complete large amounts of paperwork. There are substantial opportunities for strengthening product safety surveillance in Australia using existing routine health information systems and timely reporting of clinically significant incidents. This is critical for ensuring a more strategic approach to emerging product safety issues, prioritising efforts and evaluating the efficacy of product safety initiatives, to reduce preventable injuries and deaths related to unsafe consumer products.

Kirsten McKenzie · Ruth A Barker · Deborah A Scott · Dave A Strachan