Letters

To the Editor: I applaud Leach and colleagues for their recent article examining lost opportunities with Australia’s health workforce.1 There are many lost opportunities in the area of rural health. National and international studies have documented that a health professional with rural origins is more likely to return to a rural area to work than a colleague who grew up in the city.2,3 My research over the past 4 years looks predominantly at primary school students and their parents, teachers and governesses who reside over 800 000 km2 of outback Australia and study by distance education. My research is mainly into primary school-aged students because it is generally agreed among career development theorists that students start to shape ideas about occupations long before they reach high school.4 The students I studied initially had little knowledge or interest in health careers (relevant educational activities were subsequently provided to the school community in 2008). I identified five main factors inhibiting their knowledge and interest in health careers: the severe chronic shortage of health professionals in rural and remote Australia (and therefore a shortage of positive health professional role models); the limited career pool of their parents (most of whom are involved in the agricultural industry); the isolation in which they live; the limited information about the health professions provided by the school (their window to the outside world, which did not make up for their environmental shortfall); and parental perceptions. Two factors influencing parental perceptions deserve mention: awareness of the unhealthy workloads and the high degree of burnout among rural medical practitioners (they didn’t want this for their children); and the negative images portrayed in the media (eg, reports about health workers’ pay). My research focused solely on developing an awareness of the health professions as possible career options for these students.5 However, other factors also prevent students who live in rural and remote Australia becoming health professionals: inequality of primary and secondary educational opportunities;6 falling access, participation, retention and success rates in tertiary education;7 and the financial, emotional, family and social costs of attending tertiary courses a long way from home.8 We are all aware of the hardships of urban-raised health professionals undertaking rural medical placements, but do we understand equally the hardships of rural and remote students studying in the city? If Australia is genuine about wishing to solve the rural health workforce crisis, students from rural and remote areas must be provided with real opportunities to become rural health professionals, because the evidence suggests that the long-term survival of rural health services may depend on the recruitment of rural students.

Susan M Gorton

To the Editor: In response to Hall and colleagues,1 the Royal Australasian College of Surgeons supports any measures that have been proven to successfully reduce death and injury in young drivers. Raising the minimum legal drinking age (MLDA) to 21 years has been shown to significantly decrease road crash deaths in the United States.1 The College agrees with these authors that there would be major political obstacles and very little public support in Australia to increasing the MLDA; however, should the politicians and the public see first hand the devastating effects of alcohol on young people that our surgeons see on an all-too-regular basis, the mindset might change significantly. Hall and colleagues state other ways that we can achieve further reductions in road crash deaths — extending the zero-tolerance laws for young drivers until age 22 years, as it is in Victoria, or until 25 years for even further reductions.1 The College certainly supports this, particularly as evidence is building that the physical maturation of the part of the human brain that assesses risk and controls impulsive behaviour is not complete until age 25 years in men.2-4 The Trauma Committee is most concerned about alcohol-related trauma and will explore this issue at the annual Trauma Committee workshop, during the College’s Trauma Week. The workshop, entitled “Alcohol and injury”, will be held at the College in Melbourne on 18 November 2010.

Daryl R Wall

To the Editor: Nelson and colleagues1 referred to the limited community data from the Melbourne Medical Deputising Service (MMDS) that was analysed by Carville et al.2 MMDS consultations also represent general practice consultations, although the majority of MMDS consultations occur after hours. We can now provide an update on MMDS consultations from January 1998 to June 2010, stratified by age. De-identified data were extracted from the MMDS database for diagnoses that included the terms “chicken pox” or “varicella” and “shingles” or “zoster”, and patient age. Using the total consultations as the denominator, we calculated the crude and age-specific rates of varicella (chickenpox) and herpes zoster (HZ [shingles]) per 1000 consultations by week, and present the results here by year. These updated data support the conclusions reached by both groups of researchers that there has been a decrease in varicella cases and a rise in HZ cases in Australian general practice consultations since the introduction of a varicella vaccine in 2000. We had previously shown a decrease in hospitalisations and MMDS consultations for varicella and an uncertain effect on HZ up to 2007 after the introduction of varicella vaccine.2 The overall rate of varicella-related MMDS consultations continued to decline from 2007 to 2010, with an annual average rate of 3.3/1000 consultations in 2000 decreasing to 1.5/1000 consultations in 2007 (P < 0.001, 2000–2007) and to 1.0/1000 consultations for the first half of 2010 (P = 0.043, 2007–2010). Decreasing rates of varicella-related consultation were seen in all age groups, although an apparent increase in the consultation rate for children aged less than 5 years in 2010 might be explained by a summer peak in varicella infection and incomplete annual data (Box 1).3 The trend in HZ-related MMDS consultations showed an increase in the annual average rate from 1.7/1000 consultations in 2000 to 2.7/1000 consultations in 2007 (P < 0.001, 2000–2007) and to 3.4/1000 consultations for the first half of 2010 (P = 0.020, 2007–2010). There was a substantial increase in the HZ-related consultation rates for people aged 70–79 years and 80+ years (Box 2). Our updated data support the recommendation for adding vaccination against HZ to the vaccine schedule for older Australians,1 although the optimal age at which this should occur remains to be determined. 1 Varicella (chickenpox)-related consultation rates, by age group, 1998 to June 2010* * Data from the Melbourne Medical Deputising Service database. 2 Herpes zoster (HZ [shingles])-related consultation rates, by age group, 1998 to June 2010* * Data from the Melbourne Medical Deputising Service database.

Kristina A Grant · Kylie S Carville · Heath A Kelly

To the Editor: The article in the Journal by Leach and colleagues highlights workforce shortages in the Australian health care sectors.1 The authors identify 12 clinical professions for which data from the 2006 Australian Bureau of Statistics census show poor retention rates, and warn of the implications for the Australian population of these future shortfalls in the health workforce. However, in our opinion, the imminent workforce shortage in the pathology sector (rather than that in the higher profile specialties) is likely to represent the vulnerable point of first failure of the health care system in Australia. Pathology and diagnostic laboratory medicine are major areas in the health system, with about 70% of medical decisions being based on laboratory test results.2 About 13% of Australia’s federal health budget is allocated to pathology testing, and a similar percentage is allocated from state health budgets. Within pathology laboratories, medical scientists form the single largest professional group, although they are usually unrecognised and overlooked in workforce surveys, as is the case in the article by Leach and colleagues. Medical scientists face these same problems of poor retention within the health workforce, leading to the risk of future shortfalls in laboratory capacity and an inability to provide adequate diagnostic support for the medical system. The potential laboratory capacity shortfall has been accelerated by government policies leading to reduced funding for pathology, and the consequent failure to invest in workforce recruitment, training and development. Such an approach may deliver short-term savings, but is leading to an unsustainable environment due to the lack of an appropriately qualified and trained workforce.3 This problem is not restricted to Australia, but highlights the need for a sustainable and long-term approach to ensuring access to relevant health resources for all Australians.

Tony Badrick · Leslie Burnett · Peter E Graham · Peter P Vervaart

To the Editor: We read with interest the article by Seed and colleagues1 about the collection and distribution of blood from two donors diagnosed with Plasmodium vivax malaria following travel to Papua New Guinea (PNG). Although the potentially infectious blood components were recalled before transfusion, this case underscores the complexity of managing the risk of transfusion-transmitted malaria (TTM). Three points in particular are worth noting. First, the donors were asymptomatically harbouring infection at donation following routine foreign travel. (Between 1963 and 1999 in the United States, only one donor was implicated in TTM following routine travel, compared with more than 30 donors with lengthy residence in Sub-Saharan Africa.2) Second, both donors acquired malaria despite observing Australian guidelines for chemoprophylaxis. Finally, both were non-reactive on the enzyme immunoassay (EIA) used to screen their donations (Malaria EIA, NewLabs, Newmarket, United Kingdom) after intervals of 4 and 13 months between return from PNG and their donations. Interestingly, chemoprophylaxis for these donors was both the cause and a potential solution for the “near-miss” event. The primary prophylactic regimens used were sufficient to modify the primary infection such that antibody response to the blood-stage antigens, on which the EIA is based, occurred only after relapse months later. Had the donors taken primaquine following their travel, the parasitaemia triggered by the hepatic hypnozoite forms that characterize P. vivax infections might well have been prevented. Although primaquine is not universally recommended for terminal prophylaxis, it is consensually recommended for terminal prophylaxis for travellers who have had “intense” or “significant” exposure to P. vivax or Plasmodium ovale, such as PNG would offer.3 Seed and colleagues suggest the need for an exception to blood service management practices for donors who have recently visited a geographic area with distinct epidemiological risk for malaria. Noting the disproportionate risk associated with travel to PNG, they suggest excluding donors with a history of recent visits to PNG from routine testing for malaria (and excluding their donations from fresh component production for an appropriate period). In the United States, regulators are similarly considering carving out an exception for travellers to Mexico,4 a low-risk country where the areas associated with most (approximately 75%) malaria deferrals report near-zero malaria risk (our unpublished data). US data suggest that more than 45 000 donors will be recovered annually if the Mexican state of Quintana Roo is exempted from the deferral requirements, and regulators are weighing this benefit against the exquisitely low added risk. The outcome of this deliberation is pending, but the fact that risk gradients within a given risk category are prompting consideration of differentiated management in two countries is notable.

Bryan R Spencer · Louis M Katz

To the Editor: Children have been seen as a key priority group for pandemic (H1N1) 2009 influenza (“swine flu”) vaccination. In Australia and New Zealand, children aged 0–4 years had the highest population rate of intensive care unit admissions for swine flu.1 From 3 December 2009, the Australian Government provided free H1N1-specific influenza vaccine for all Australians aged 6 months and older. However, it was not known how parents and other caregivers would respond to the offer of vaccination. We sought to determine caregivers’ intentions regarding whether their child would receive the pandemic vaccine. The study was approved by the human research ethics committee of the Children’s Hospital at Westmead and was conducted in November and December 2009, immediately before, and alongside, the commencement of the pandemic vaccination program for children. A paper-based questionnaire (with an alternative web-based option) was sent to caregivers of children aged 6 months to 5 years who were attending 16 long-day-care centres across metropolitan Sydney. It included questions about attitudes, behaviour, intentions and beliefs regarding swine flu, seasonal influenza and vaccines. We analysed responses using SPSS, version 17 (SPSS Inc, Chicago, Ill, USA) and conducted univariate analysis (as factors had high collinearity) to search for factors associated with caregivers’ intention for their child to receive the pandemic vaccine. The response rate was 44% (431/972). Most respondents had families of two children (47%) or one child (41%); 90% were mothers; and 48% had a postgraduate education. Caregivers were asked to report in relation to their eldest child attending the day-care centre (mean age, 38 months). Three children (out of 427 responses; 0.7%) had already received the pandemic vaccine; 23% of caregivers (92/400) said they would have their child vaccinated; 54% (217/400) were unsure; and 22% (87/400) would not. Intentions regarding seasonal influenza vaccination were similar. Factors associated with caregivers’ intention to have their child vaccinated with pandemic vaccine are shown in the Box. Those with the strongest associations included caregivers’ intention to have their child vaccinated against seasonal influenza in 2010 and belief that seasonal influenza vaccine is completely safe or only a slight risk. Factors not significantly associated with respondents’ intention to have their child vaccinated with the pandemic vaccine were the respondent’s age group, sex, education level, language spoken at home, and number of children in the household. The study indicated that at the commencement of the Australian pandemic influenza vaccination program for children, there was significant uncertainty among this sample of relatively highly educated respondents. The proportion of respondents intending to have their children vaccinated was far lower than recent Australian estimates (6% of children aged 4 years and under).2 Our findings suggest that, despite the acknowledged severity of pandemic influenza, respondents’ concerns about vaccine safety were influencing their intentions. Indeed, the program commenced in a context of public debate about the response to the influenza pandemic and the vaccination program, including concerns about the safety of using multi-dose vials for vaccine delivery.3,4 Among respondents intending for their child to have the pandemic vaccine, the association with the child having had a previous influenza vaccination suggests that, having once taken up vaccination, respondents were more likely to intend to do so again. More recent events in Australia leading to suspension of use of all three 2010 seasonal influenza vaccines for children under 5 years of age are likely to further increase safety concerns.5 While two vaccines have since been reinstated, these events present a significant challenge for future influenza vaccination programs. Providers are likely to have an important role in communicating recommendations and addressing caregivers’ concerns. Factors associated with caregivers’ intentions for their child to receive the pandemic (H1N1) 2009 influenza vaccine Survey responses Intention “yes” (n = 92) Number OR (95% CI) Would you have your child vaccinated against seasonal flu in 2010? (n = 396) No 8 1.00 Unsure 33 2.29* (1.11–5.15) Yes 51 17.49* (7.55–40.50) How safe do you think the seasonal flu vaccine is for children aged 1 to 5 years? (n = 396) Moderate/high risk 7 1.00 Don’t know 19 2.04 (0.82–5.07) Completely safe/slight risk 66 6.64* (3.00–15.22) How safe do you think the seasonal flu vaccine is for babies aged 6 to 11 months? (n = 395) Moderate/high risk 21 1.00 Don’t know 36 1.35 (0.75–2.44) Completely safe/slight risk 35 4.32* (2.27–8.22) Did your child receive seasonal flu vaccination in 2009? (n = 394) No 77 1.00 Yes 15 3.48* (1.65–7.36) How concerned do you feel about your child catching the flu? (n = 396) Not at all/ a little/ moderately 31 1.00 Very/extremely 61 2.46* (1.46–4.17) Does your child have any medical conditions requiring ongoing visits to a health care professional? (n = 396) No 76 1.00 Yes 16 2.00† (1.03–3.87) The flu vaccine can give you the flu. (n = 396) Agree 24 1.00 Don’t know 23 1.18 (0.62–2.23) Disagree 45 1.82† (1.04–3.20) OR = odds ratio. * P ≤ 0.01. † P ≤ 0.05.

Julie Leask · Maria Yui Kwan Chow · Catherine King · Robert Booy

To the Editor: The World Health Organization Surgical Safety Checklist (Checklist) has recently been editorialised by Merry and Barraclough.1 We are all keen to see perioperative mortality and morbidity reduced to a minimum, and one cannot disagree with the intent of the Checklist, which was tested in a pilot study2 involving almost 8000 patients in eight hospitals (in disparate developed and developing countries). The study showed a reduction in the death rate associated with surgery from 1.5% to 0.8% (or 15 per 1000 down to eight per 1000). Impressive indeed! My concerns are firstly with process. One hospital in Brisbane has already introduced the Checklist without the constructive engagement of those at the clinical coalface, contrary to what proponents of the Checklist advocate, and it is being used in a way which distracts personnel during the induction of anaesthesia. To those familiar with accident analysis, this is a good example of poor process implementation by a hospital administration and its managers. Secondly, I am concerned by the use of the study of Haynes and colleagues2 to underwrite the Checklist. Results of the study would be more meaningful if the reduction rates in mortality in the hospitals in the developed countries (Canada, United States, United Kingdom and New Zealand) were separated from those in the developing countries. Or would this have shown the developed countries’ reduction rate to be statistically non-significant? Are we to believe that use of the Checklist could save seven lives per 1000 procedures in Australia, or 15 000 to 21 000 lives across the three million procedures undertaken here annually? Total figures for perioperative mortality across all procedures in all operative categories remain a matter for conjecture. However, in looking at the high-risk group of cardiac surgical operations in Victorian public hospitals,3 we can see a published mortality rate of 10 per 1000 for coronary artery bypass or isolated valve surgery, only a little higher than that seen after the introduction of the Checklist across a broad surgical population. We should be careful to see that spin does not become embedded within medicine, even where it is associated with what is probably a good idea.

James P Bradley

In reply: Thank you for the opportunity to respond to Bradley’s letter. We agree that constructive engagement of clinicians is essential in adoption of the World Health Organization Surgical Safety Checklist (the Checklist) — that point is clearly made in our editorial.1 Bradley’s anecdotal comment about use of the Checklist causing distraction during the induction of anaesthesia is difficult to evaluate in the absence of more detail. Surely any competent anaesthetist committed to patient safety could liaise with other team members to find an acceptable time to fulfil the entirely reasonable expectation of ensuring that relevant safety checks have been done? Furthermore, the study of Haynes and colleagues2 provides considerable reassurance about the possibility of unintended harm from use of the Checklist. Justification for the Checklist does not depend on the above-mentioned study alone: the Checklist was developed by an extensive process of literature review and consultation. It reflects expert consensus, and its endorsement by the Australian and New Zealand College of Anaesthetists, the Royal Australasian College of Surgeons and many other organisations adds to this consensus. There is no suggestion that the Checklist will save any particular number of lives in any given country — rather, its use is likely to reduce the number of patients harmed during surgery, and this is surely worthwhile.

Alan F Merry · Bruce H Barraclough

To the Editor: The editorial by Runciman in the Journal was cogent and timely.1 The Garling report on acute care services in New South Wales public hospitals highlights areas where critical changes need to be made in order to improve patient safety.2 It is widely recognised that junior doctors play a vital role at the coalface in many key areas, including assessment and escalation of care for deteriorating patients and the handover of patient care, both in hospital between shifts and also from hospital to the community. In NSW, these processes are currently receiving much needed attention in order to ensure patient safety within hospital settings. As Runciman so rightly identifies, junior doctors also play a critical role in “transformational” cultural change. In public hospitals such as my own, it has been fascinating to observe the small ways in which a junior doctor can influence senior clinicians in fields such as compliance with hand hygiene protocols, and in querying transfusion requests which do not conform to National Health and Medical Research Council clinical guidelines. We need to continue to mentor junior doctors so their messages about change and improvement are well delivered and well received. With the move of junior doctor pre-vocational training into community settings (the Prevocational General Practice Placements Program) there are increasing opportunities for change on a larger scale. In the future, patients will not be the only agents who move from community to hospital settings and back again. Junior doctors will be employed across both settings and will be an important vector for change and improvement. We will also need to encourage supervision by senior doctors who engage with and understand the health system, not just the hospital system. Providing junior doctors with appropriate tools to drive change (eg, training in graded assertiveness and situational awareness), encouraging them to view their role in health system improvement at the whole-of-community level, and providing leadership across the system, give us every reason to believe that Runciman’s vision of transformational change is possible, and indeed likely.

Joanna R Sutherland

To the Editor: I read, with encouragement and hope, Runciman’s timely editorial concerning the pressing need for patient safety education to become an integral part of medical undergraduate curricula.1 I also support his plea for patient safety to now be treated as a full-blooded science. His description of medicine as “islands of excellence in a sea of mediocrity” is both apt and perhaps still somewhat generous. In many health care ranks (in common with the rest of the human race) there continues to be a reluctance at the personal level to admit the natural and ever-present human tendency to err.2 Consequently, opportunities continue to be lost to learn from analysis of these mistakes by reporting their mechanisms honestly in detail, either anonymously or openly in a trusting work environment. Sadly, the ongoing barriers of a blame culture (often accompanied by outcome bias), misplaced ego and lack of understanding of such fundamental issues as error cascades and latent errors work against these simple and powerful safety improvement steps. The preoccupation of sections of our medicolegal system with blaming and shaming usually innocent health care individuals for system failures is also a barrier to such reporting. While Runciman suggests a gradual awakening of appreciation of such system faults, lip service still abounds and basic human error concepts continue to be uncomprehended or overlooked, even by some of the more senior and influential members of our profession. This works against the education of junior doctors. As Runciman urges, the only way to deal with this now is to give the young, fertile medical undergraduate minds the opportunity to grasp these fundamental concepts at the beginning of their clinical lives. It is their patients who will benefit as well as their own peace of mind and, importantly, their own humility, throughout their years ahead. The repeated objection that undergraduate curricula are already overloaded is not acceptable. The science of patient safety is no longer an optional or minor subject for medicine. It should be woven intimately throughout the undergraduate curriculum.

John A H Williamson

To the Editor: The pace of improvement in the safety and quality of health care has been disappointing. However, the recent enthusiasm for medical education providing a new solution is troubling.1,2 The safety and quality movement has been regrettably dismissive of the need for creation of evidence.3 Indeed, it now seems that the focus on “dramatic” errors has been a distraction from the major volume of harm due to the poor care of patients with chronic disease.4 There is substantial literature on the development of professionalism, which attests to the difficulty of teaching medical students non-technical skills (or of selecting students with “appropriate” attitudes). There are three major reasons for this difficulty. The first is the problem of defining the desirable standard — whether of safety or professionalism.5,6 The second is the so called “hidden curriculum” — the ubiquity of error, unsafe practice and unprofessional behaviour that students are exposed to in clinical environments. Third, many of the safety and quality issues and skills just do not seem relevant to the students, despite the care that has been taken in the development of safety and quality curricula. To provide “experience that is arranged to be both motivating in the moment and consequential for capabilities in the long run”7 will require a new and more pragmatic approach to education. We need to select safety and quality issues that all doctors can influence by exercise of their personal knowledge and skills. This is the stuff of medical education. Errors in clinical reasoning8 form an excellent teaching topic, but many safety problems are simply too hard for medical students (neither do we expect medical students to learn the details of intensive care or casemix funding). Provision of appropriate experiential learning requires links to the people who understand and manage governance and safety in the clinical workplace. Rather than a stand-alone communication syllabus, the medical students in Queensland, for example, should receive the same teaching about clinical handover as the junior (and senior) medical staff of Queensland Health. Handover is then perceived as a highly relevant skill. If competencies in safety for medical stu-dents are to be required they must be linked to both fuller competencies for graduate training and the introduction of safer workplace practices. Without these parallel and more important activities, the proposed reform in medical education is likely to be disappointing in its impact.

Christine M Jorm

To the Editor: We report a case of measles in a 24-year-old man who returned from the Fédération Internationale de Football Association (FIFA) World Cup in South Africa in July 2010. Despite Australian health alerts about measles in South Africa,1 the patient had received no pretravel medical advice or vaccinations. Six days after returning home to the Northern Territory, the patient developed fever, headache and myalgia. The following day he developed vomiting, diarrhoea, and productive cough, with a widespread rash appearing the subsequent day. He visited two general practitioners, was prescribed doxycycline and then admitted to hospital on Day 6 of his illness. On examination, his temperature was 38.9ºC; pulse, 112 beats per minute; blood pressure, 133/72 mmHg; and oxygen saturation, 96% on room air. He had conjunctivitis, a widespread blanching maculopapular rash involving his face, trunk, limbs, hands and feet (Box) and cervical lymphadenopathy. He had bilateral basal lung crackles and tender hepatomegaly. Investigations showed thrombocytopenia (platelet count, 148 × 109/L; reference range, 150–450 × 109/L); hyponatraemia (sodium concentration, 131 mmol/L; reference range, 132–142 mmol/L); and abnormal liver function test results (alanine transaminase concentration, 417 U/L [reference range, < 40 U/L]; alkaline phosphatase concentration, 236 U/L [reference range, 39–117 U/L]). His chest x-ray was normal. The following day, measles virus RNA was detected from a throat swab, and the patient was put into respiratory isolation and therapy with doxycycline ceased. He made a full recovery. The patient reported receiving childhood vaccinations, and while he thought he may have received one measles, mumps and rubella vaccination, he had not had two. Follow-up was required for 84 identified contacts, with no measles cases subsequently notified in the NT. The FIFA World Cup is the world’s largest single-sport event, with an attendance this year of over 3 million people. Mass gatherings may be associated with outbreaks of communicable diseases such as meningococcal disease, measles and pandemic (H1N1) 2009 influenza, in addition to an increased risk of sexually transmitted diseases.2 These risks should be considered when seeing patients who have travelled to such events, in addition to country-specific health risks. Recent data from the GeoSentinal surveillance network showed that a systemic febrile illness was the most common presenting syndrome among travellers returning from South Africa. Most of these cases of illness (54.5%) were due to spotted fever group rickettsiosis.3 The risk of acquiring rickettsiosis increases among travellers visiting game parks, with an incidence of African tick bite fever (Rickettsia africae) among short-term safari tourists of 4.0%–5.3%.4 Measles has rarely been reported in travellers returning from South Africa,2 but the country is in the midst of a measles epidemic, with 17 354 confirmed cases between January 2009 and 12 August 2010.5 Measles presents with fever, cough, rhinorrhoea and conjunctivitis, followed by a widespread rash. The incubation period is usually 7–10 days, but may be up to 18 days. The virus is highly infectious, from 5 days before to 4 days after the onset of rash. Young adults from non-endemic countries such as Australia are at particular risk, as they may only have had one childhood vaccination for measles, with consequent inadequate protection. Although measles has been eliminated in Australia,6 sporadic outbreaks continue to occur,7 and travellers returning from overseas create an ongoing potential for the re-establishment of endemic measles. Given the public health implications of a delayed diagnosis, doctors must be alert to possible cases of measles in travellers returning from endemic countries. The patient’s widespread maculopapular blanching rash

Bridget E Barber · Krispin M Hajkowicz · Vicki L Krause · Kevin G Freeman · Bart J Currie

To the Editor: Allergen avoidance is the mainstay of food allergy management. Consumers with food allergies rely on accurate labelling of foods to avoid ingestion of allergens and subsequent allergic reactions. Current Australian legislation states that ingredients derived from common allergens (peanuts, tree nuts, eggs, wheat, cows milk, soy, fish, shellfish and sesame) must be clearly labelled.1 However, use of shared processing facilities can result in cross-contamination of other ingredients with these allergens. This has led to the use of advisory statements such as “may contain traces of” by manufacturers. A recent Food Standards Australia New Zealand survey found that consumers with food allergies are frustrated by such labelling.2 There is also confusion among the medical profession about whether to advise patients with food allergies to avoid all foods with allergen avoidance advisory statements. The perception by some in the general population and in the medical community is that these statements are so widely used that avoidance would be overly prohibitive. However, there are currently no published data on the extent of advisory labelling use in Australia. We aimed to assess the prevalence of advisory labelling for three common food allergens — peanuts, tree nuts and eggs — on the packages of products for which these allergens were not listed as ingredients. Products containing one type of tree nut (eg, macadamia nut) could still have advisory labelling for other tree nuts (eg, almond). All products were therefore examined for advisory labelling for any tree nut which was not listed as an ingredient. Packages of non-refrigerated processed foods were examined between August and September 2008 at a large supermarket in Melbourne. All product types were examined within each category (eg, for the savoury biscuits category, we examined rice crackers, flavoured wheat crackers and water crackers) and, for each product type, one flavour per brand was selected for examination. Single-ingredient foods, such as flour, sugar, fruit and vegetables, were excluded. Advisory statements included, but were not limited to, “may contain traces of”, “processed on the same line as” and “made on equipment that also processes”. Overall, 761 products were examined. Of these, 384 (50%) carried an advisory statement for one or more tree nuts. Of 737 products that did not list peanut as an ingredient, 348 (47%) carried an advisory statement regarding peanut. Of 641 products that did not list egg as an ingredient, 146 (23%) carried an advisory label for egg. The presence of advisory statements varied between categories of food, with sweet biscuits most likely to carry labelling for peanut and tree nuts and bakery items most likely to carry labelling for egg (Box). Advisory statements have been widely adopted by manufacturers across a range of products, and are likely to limit food choices for consumers with food allergies who avoid all foods labelled with advisory statements. Unfortunately, there is no evidence on the frequency of trace allergen contamination in Australian products, although studies in the United States found that only 10% of 179 products with advisory labelling for peanut contained detectable levels3 and that, as for Australia, this labelling was widely used for some product categories.4 There is also no evidence regarding the proportion of consumers with food allergies who will develop an allergic reaction (including anaphylaxis) to trace contamination of food. Scientific assessment of the risks posed to consumers with food allergies by trace contamination is urgently required. This evidence would allow the development of informative labelling guidelines, including changes to legislation where required, to allow consumers with food allergies to safely manage their allergies, hopefully without the need to avoid entire categories of common foods. Allergen avoidance advisory statements on packaged processed foods at a large supermarket in Melbourne, August–September 2008 Number (%) of products with an advisory statement* Category of food Peanut Tree nuts Egg Sweet biscuits (n = 130) 117 (93%) 120 (92%) 48 (70%) Chocolates (n = 60) 43 (80%) 49 (82%) 2 (4%) Bakery items (eg, cakes) (n = 35) 24 (71%) 30 (86%) 10 (71%) Muesli bars and snack bars (n = 27) 13 (67%) 20 (74%) 4 (15%) Dinner bases and stocks (n = 32) 19 (59%) 9 (28%) 5 (17%) Savoury biscuits (n = 41) 23 (56%) 23 (56%) 20 (51%) Lollies (n = 55) 29 (56%) 25 (45%) 1 (2%) Breakfast cereals (n = 63) 25 (41%) 37 (59%) 3 (5%) Instant noodles (n = 18) 7 (39%) 6 (33%) 8 (50%) Pasta sauces (n = 15) 5 (33%) 4 (27%) 0 Bread (n = 16) 5 (31%) 5 (31%) 5 (31%) Soups (n = 20) 3 (15%) 3 (15%) 4 (21%) Cake mixes (n = 30) 4 (13%) 20 (67%) 11 (58%) Tinned meals (n = 17) 2 (12%) 2 (12%) 1 (7%) Baby foods (n = 30) 3 (10%) 3 (10%) 1 (4%) Pasta (n = 13) 0 0 4 (39%) Chips (n = 20) 0 0 0 Other (eg, tinned fish, breadcrumbs, sauces, custard powder) (n = 139) 26 (19%) 28 (20%) 19 (14%) * Number of products that had an advisory statement but did not have the allergen of interest listed as an ingredient. The denominators used to calculate percentages were the numbers of products within each category of food that did not have the allergen of interest listed as an ingredient.

Jennifer J Koplin · Nicholas J Osborne · Katrina J Allen

To the Editor: Prescribing and dispensing generic medicines is an option to reduce costs in the community and is also common practice in public hospitals. In addition to the issues discussed by McLachlan,1 we have noted a concerning trend in the “branding” of many new generic medicines that has the potential to add to the confusion for patients, prescribers and pharmacists. Medicines with special release properties are branded with suffixes as a reminder of their longer duration of action, such as Sustained Release (eg, Tramal SR; CSL Limited) or eXtended Release (eg, Efexor-XR; Wyeth Australia). Different formulations may also use a suffix to indicate distinguishing properties, such as “dispersible” in Rulide D (Sanofi-Aventis) or “osmotic release oral system” in Adalat OROS (Bayer Schering). However, these suffixes, while meaningful, can be a source of misunderstanding about dosing intervals and length of action, leading to errors.2 A standard nomenclature does not exist, even for formulation descriptors. Adding to this confusion, generic medicine manufacturers are now marketing products with prefixes or suffixes, not to denote a modified formulation but to place their “brand” on the medicine. Ascent Pharmaceuticals has three different suffixes/prefixes for its generic products, reflecting the names of previous manufacturers (eg, Quinapril-GA, GN-Carvedilol), and two suffixes for simvastatin (GA and DP). Spirit Pharmaceuticals adds its name to some products (eg, Simvastatin-Spirit). Taking an oral medication history becomes challenging when patients state they use “Spirit” medication for their cholesterol, “GN” tablets for their heart and “GA” tablets for their high blood pressure. With the many generic “brands” now available (eg, 22 simvastatin products), packs with similar labelling lined up on the pharmacy or patient’s shelves increase the risk of wrong selection. Although no reports have been published to date, similarities in brand prefixes may cause medication errors in electronic prescribing and dispensing systems, when an incorrect medication is selected from a dropdown menu. Entering “Apo” in some systems selects more than 40 products manufactured by Apotex, which uses a naming pattern of the generic drug prefixed by Apo (eg, Apo-Alendronate), as well as APO-go (Hospira), which is apomorphine (and does not contain “go”). This situation could be avoided in future if the National E-Health Transition Authority’s Australian Medicines Terminology and its editorial principles are adopted in systems. These rules require that all medicines are represented by descriptions that list the generic name first, with the sponsor’s name following in parentheses (Paul Frosdick, Chief Terminologist, National E-Health Transition Authority, personal communication, 25 August 2010). The Food and Drug Administration in the United States and the Therapeutic Goods Administration in Australia have developed documents outlining approved medicines terminology,2,3 but there are no official lists of approved suffixes or prefixes. The Institute for Safe Medication Practices, a non-profit US-certified patient safety organisation internationally regarded as an expert in medication safety, has recognised this problem and maintains a list of products with drug name suffixes and their meanings.4 With more generic options available, good communication between medical and pharmacy clinicians and patients is essential to clarify the indication, along with the specific name, of medications prescribed. Pharmaceutical manufacturers need to consider the impact of meaningless prefixes and suffixes, which add to the confusion and potentially contribute to medication errors. National authorities should improve overall governance of labelling of generic medicines, to prevent errors reaching patients.

Linda V Graudins · Michael J Dooley

To the Editor: In recent years, medical school selection criteria have continued to evolve as new information and evidence have come to light.1,2 Increasingly, data are becoming available that shed light on local selection processes, and we commend Laurence and colleagues for their contribution to this.3 Diversity in selection processes is desirable because of the varying educational cultures and emphases between medical schools, and for maintaining variety among students who gain entry. The pursuit of “ideal” evidence-based selection criteria ought to continue, although the effect that a single, standard national selection process would have on the diversity of students entering medicine is unknown. However, a national approach to offering places for medical school could have considerable advantages. Mutual recognition of admission processes by medical schools is one approach suggested by Laurence et al.3 This may not have as significant an impact on student diversity as might a single selection process, and it could reduce the substantial financial burden that the application process places on both applicants and medical schools.3 Students from low socioeconomic backgrounds are often unable to apply to interstate medical schools due to the travel costs associated with attending interviews. This both reduces the number of opportunities they have of gaining a place at a medical school, and prevents them from gaining valuable interview experience, further reducing the likelihood of a successful application.3,4 Mutual recognition of and sharing of selection information would potentially level the playing field for applicants who cannot afford to attend multiple interviews. Students commonly receive medical school offers as late as midway through the first semester — a consequence of students receiving offers from multiple universities, and universities having to make a number of rounds of offers until their quotas are complete. Universities already rank their applicants, and, if students were able to submit their medical school preferences, applications could be coordinated nationally through a centralised admissions system. This would result in fewer rounds of offers and a shorter admission process, reducing the cost for universities and enabling students to enrol earlier, with greater certainty. This may avoid the need for students to relocate and miss out on vital parts of their first year of study.

Christopher X J Wong · Adam J Nelson · Ross L Roberts-Thomson

To the Editor: In their recent editorial, Bower and colleagues effectively summarised the problem of apparent governmental indifference to congenital anomalies.1 This is not unique to Australia and probably exists worldwide. National systems to collect congenital anomalies data were set up in many countries in the mid 1960s, including the National Congenital Anomaly System in the United Kingdom and the Canadian Congenital Anomalies Surveillance System. This followed the thalidomide tragedy and exemplified that it often takes an acute crisis to stimulate politicians into action. However, due to a lack of leadership, foresight and finances,2-4 these systems gradually declined to the extent that they became of very little value, lacking in accurate ascertainment and pregnancy termination data. As a result, regional registries were set up in England and Wales, and Canada was left with only two provinces (British Columbia and Alberta) collecting data. Prevention is one of the new driving forces for collecting good data, and the advent of using folic acid to effectively reduce neural tube defects brought a new urgency to the need for comprehensive useable data. Accordingly, the Canadian government set up a task force and formed a new entity in 2002, the Canadian Congenital Anomalies Surveillance Network, with a mandate to provide logistical and financial help to all 10 provinces and three territories. While progress has been slow, it has been very encouraging, with three additional provinces and one territory developing new surveillance systems this fiscal year (April 2010 – March 2011). The Network has set up guidelines and standards to enable all provinces and territories to collect data in a comparable format,5 which can then be forwarded to a central database in the national capital, Ottawa. The quality of the data should be improved because they are gathered at a local level. This model could be adapted for use in Australia because, according to Bower et al,1 a nucleus of good data from at least three states already exists.

R Brian Lowry

To the Editor: I agree with Emery and colleagues1 that family history can add much to downstream clinical interventions that provide tangible benefits to the presenting patient and his or her kin. The pedigree chart has some advantages over simple narrative recording of the same details, including the ability to instantly visualise relationships between individuals and the ease with which the chart can be updated and annotated.2 There is a familial aggregation (commonly an affected first-degree relative) in up to 25% of presenting cancer patients, while around 5% will harbour a highly penetrant genetic predisposition to cancer. In the cancer clinic, an acceptably detailed family cancer pedigree can typically be obtained in even less time than the 30 minutes suggested by Emery et al.1 Steps in drawing pedigrees have been outlined elsewhere2,3 and typically involve collecting information such as simple demographics, naming and symbolising different cancer types, and recording age of onset and age of death (if relevant) for the different individuals, starting with the presenting patient. The time required to construct a three-generation pedigree would typically be around 10 minutes, making it an attractive addition to routine history taking. If a cancer pattern emerges, the pedigree should be extended as far as possible. Of course, diagnosis verification (through death registries, etc) would be required before surveillance, prophylaxis and therapy decisions are addressed in the familial cancer setting. A convenient refresher for doctors who do not routinely draw pedigrees might be to first construct their own family tree, with reference to the symbols and relationship illustrators commonly used.2,3 It is likely that pedigrees in most clinics will be drawn by hand, at least initially, although there are software programs for pedigree creation available (eg, Family Tree Builder; <http://www.myheritage.com/family-tree-builder>). Over the years, I have informally asked my specialist oncology trainees to routinely construct family pedigrees. I do not recall, among these highly clinically skilled doctors, one that was able to correctly draw a family pedigree until we had worked on it together. Given the benefits of family pedigree analysis across many disease categories, a little practice in pedigree drawing may be a useful exercise for many of us. Even without the requisite confirmation of disease status of the individuals represented, a simple pedigree chart, combined with quick reference to information on the potential clinical significance of any patterns it shows (eg, by consulting National Health and Medical Research Council guidelines),4 can help prioritise referrals to busy familial cancer clinics.

Michael J McKay

To the Editor: Langlands and colleagues show that family history is poorly taken in patients presenting to an acute medical unit at a major tertiary hospital.1 However, their article and its companion editorials2,3 do not adequately stress the settings in which family history taking may be both easily achievable and most cost-effective. The nihilism that Thomas and Thompson2 convey about recording family history in the acute setting is of more concern, given that a tertiary hospital may offer the best opportunity to initiate the process of case detection for a number of heritable and lethal diseases, such as autosomal dominant familial hypercholesterolaemia (FH), the most common monogenic cause of premature coronary artery disease (CAD). We previously demonstrated that a family history of cardiovascular disease was almost never recorded by coronary care unit medical staff at Royal Perth Hospital.4 For 509 patients aged < 60 years presenting with symptomatic CAD to the coronary care unit, we found that 70% had insufficient clinical data documented in the medical records to enable a diagnosis of FH. In a follow-up study of 103 patients with premature CAD admitted to the coronary care unit, a nurse practitioner was able to record a positive family history of premature cardiovascular disease in a primary relative for 43% of patients, of whom 95% had phenotypic FH based on a recognised clinical diagnostic tool (the Dutch Lipid Clinic Network score).5 Patients detected in this way in Western Australia are now referred to a statewide FH program run by staff from a lipid clinic.5 In this program, where detailed pedigree drawing and family tracing is performed by trained nurses, we have found a causative mutation for FH in up to 85% of patients with a clinical phenotype strongly suggestive of FH. Additionally, we find that for every index case so detected, we can additionally diagnose at least three new cases of FH in the patient’s relatives, many of whom are young. This method of case detection and subsequent treatment with cholesterol-lowering therapies is highly cost-effective and, more importantly, enables therapy to be targeted at younger patients, thereby maximising the potential for preventing CAD. Our experience illustrates that for a lethal condition such as FH, an accurate family history recorded by a nurse can spark a cascade of action that leads to a definitive diagnosis of FH in the index patient and the subsequent detection of otherwise undiagnosed FH in the community, with significant associated cost savings.6 Hence, we propose that nursing staff can efficiently bridge this gap in medical care while we are getting our house in order by training medical staff to effectively take a family history.

Timothy R Bates · Elissa B Poulter · Frank M van Bockxmeer · Gerald F Watts

To the Editor: I read the article by Langlands and colleagues1 with some dismay, and a sinking heart. Their report is further evidence of the dangers of moving away from the basic skills of comprehensive history taking and performing a detailed physical examination. Interestingly, in the United States, the debate regarding performing a physical examination has come full circle, from virtually ignoring its importance to now telling us how vital it is and how to perform it.2 But it is the taking of a comprehensive history that, as one of my mentors told me, “is where the money is”, and a detailed family history is an integral part of this. Eliciting a detailed family history is arguably more important for paediatric patients, who have a longer potential life span and hence have more to gain from this information. In my paediatric practice, for example, I see numerous overweight children, some of whom have a strong family history of hypercholesterolaemia, vascular disease or type 2 diabetes mellitus, which places them at considerable risk of cardiovascular disease in their adult years. A detailed family history may also “unmask” the genetic contribution to a child’s history of deafness or learning disability. Time constraint is the main impediment to taking a comprehensive family history, but it is worth keeping in mind that it is time well spent and that, in the paediatric population, it may make a significant contribution to the long-term health of the child. With the impending advent of personalised genomic screening, there will be an even greater imperative to formalise the gathering of family history details.3,4

Simon E P Hauser

To the Editor: Pridmore1 describes a case in which a man’s recent actions suggested suicidal intent. The man told police he had no ongoing suicidal plans and they took his words at face value. He subsequently killed himself. The High Court exonerated the police of any responsibility, a decision that seemed based on two premises: that suicide does not presuppose mental disorder; and that “There is no general common law duty of care to rescue a person from harm, including self-harm”.2 There is sufficient grey in both those inter-related premises to make a black-and-white judgment suspect. I share Pridmore’s position that suicide does not always equal mental disorder, although one wonders whether situational crises in individuals with subtle vulnerabilities could be subsumed under such a label. More important is the question of individual autonomy. The reason we have no common law requiring us to “rescue” another adult is because we set such a high value on autonomy. Although we assume a person’s competence, our curiosity about it should be aroused when people behave in unexpected ways. When the behaviour is strikingly different, and has potential for serious harm, are we not obliged to intervene or procure assessment? This is, I would have thought, a moral rather than legal or medical concern. This is what we would do for a child wandering on the road or a demented person lost at night. We might be free of legal or medical censure for ignoring them, but we would be embarrassed to publicly admit our failure to act if we might have done something useful at the time. Making serious preparation to kill yourself is strikingly different behaviour, and should raise questions about both your competence and autonomy. Perhaps not 100% but surely more than 50% of such people have a disorder. A person preparing to suicide is “more likely than not” mentally ill, and that is how the “common person” would surely see it: “guilty”, as it were, till proven innocent. Expertise is called for to make that final determination. The police certainly do not have the expertise. Why, then, would they not seek it? Furthermore, if the “rational, cooperative and responsible” man in this case had indicated his plans to “rationally” commit suicide to escape an intolerable predicament, it is hard to believe the police would have walked away, even though he may have been competent to make such a decision!

Paul T Dignam

To the Editor: The editorial on the Human Variome Project by Cotton and Macrae1 neatly lays out the reasons for government funding for gene mutation databases for inherited diseases. Most of these diseases are rare, but collectively, they are common. The article highlights the key principles of detecting gene mutations and establishing pathogenicity in order to offer individuals (or couples) relevant health information for themselves and/or their (future) offspring. Cotton and Macrae mention a number of specific diseases, but do not mention the commonest life-shortening inherited disease affecting Australian children — cystic fibrosis (CF). Far from being theoretical, nearly all of the principles outlined by Cotton and Macrae are already in place for CF, including clinical databases (in Australia, the Australian Cystic Fibrosis Data Registry), an international gene mutation database (at http://www.sickkids.on.ca, which is contributed to by Australian genetics laboratories) and programs to offer carrier screening to the population. Unfortunately there is very little government funding for these initiatives, so they are not coordinated. In particular, screening for CF carriers in the population, which is of considerable clinical utility, has only small, fee-for-service programs that reach very few people.2,3 These programs are inequitable in that many people are unaware of the existence of such screening programs and, of those who are, many cannot afford the cost of testing. CF provides an excellent model for the development of a coordinated approach to inherited disease screening and, given that 800 000 Australians are carriers of CF mutations, funding CF screening should be a major government priority.

R John Massie · Martin B Delatycki

To the Editor: Costello and Murphy’s lament1 about the discoverer of prostate-specific antigen (PSA), Richard Ablin, describing PSA testing as a “hugely expensive public health disaster”2 contains several egregious claims that require correction. They write that “since the introduction of PSA testing in the 1980s, we have seen a 25% reduction in mortality” from the disease. Thirty years ago (in 1980), before PSA testing was possible, the age-adjusted mortality rate from prostate cancer in Australia was 33.4/100 000. In 2007, it was 31.0/100 000, a decline of 7.2%. Over the same period, prostate cancer incidence rose 110%, from 80.8/100 000 to 170/100 000,3 thanks to the aggressive promotion of PSA testing. Recent New South Wales data show that 3 years after radical prostatectomy, 77.4% of men are impotent and 12.3% have urinary incontinence, compared with 22.3% and 1.0%, respectively, of controls.4 Many of these men are, to use Costello and Murphy’s word, “overtreated”1 — they underwent unnecessary surgery and now live with the consequences. From a 2009 European trial,5 the take-home message for a man being tested today is: There is a one in 50 chance that, in 2019 or later, he will be spared death from a cancer that would otherwise have killed him. And there is a 49 in 50 chance that he will have been treated unnecessarily for a cancer that was never a threat to his life.6 This is what Ablin called a public health disaster. With reference to this European trial,5 Costello and Murphy claim that “PSA testing has led to greatly reduced mortality”. This “great reduction” was from 4.2 to 3.3 deaths per 10 000 person-years. Costello and Murphy propose that “those with a PSA level well below the median for men in their 40s . . . (the vast majority at this stage) could be reassured . . .”, but below the median lie half the men, not the vast majority. This policy would mean that the other half of men aged 40 (not the small minority, as implied by Costello and Murphy) would be above the threshold. They would not be reassured; presumably they would be offered more frequent testing and follow-up. Labelling half the population of men aged 40 as higher risk has enormous implications for the men (anxiety, inconvenience, cost) and the health system that would be called on to fund more appointments and tests. Such a proposal does indeed sound like a public health disaster.

Simon Chapman · Alexandra Barratt

In reply: Chapman and Barratt’s statements about prostate cancer highlight the dramatic difference between the views of sociologists and the views of clinicians who deal daily with the burden of prostate cancer diagnosis, treatment and mortality. In 1851, prostate cancer was considered a rare disease.1 Chapman and Barratt’s views seem to remain consistent with this thesis. However, in 2010 in Australia, prostate cancer is the most common cancer diagnosed in men, and the most common cancer causing death in men. Our own data do not support the contention that treatment for prostate cancer produces dreadful outcomes. We can reassure men that there is a high likelihood of cure through early treatment for prostate cancer, with a less than 5% chance of becoming incontinent and a 70% chance of retaining erectile function.2 Clearly, Chapman and Barratt have not taken into account a recently reported Swedish study which showed that in 20 000 men randomly allocated to prostate-specific antigen (PSA) screening or a control group between 1994 and 2008, there was a 50% reduction in mortality from prostate cancer for men in the screened arm.3 The numbers needed to screen (293) and treat (12) from this study are almost exactly the same as those from breast cancer screening studies. We repeat our recommendation for the early use of a single PSA test4 in men aged in their 40s. This is most beneficial because the background noise from benign prostatic hyperplasia development does not occur, and PSA level is very discriminatory for detecting those who go on to develop significant prostate cancer.

Anthony J Costello · Declan Murphy

To the Editor: The long-term functional outcomes in the three cases of swimming pool filter-induced transrectal evisceration described by Price and colleagues1 are excellent and significantly better than many other cases described in the literature. However, it may be possible to improve further on such results, or at least decrease short-term morbidity, by expediting the reduction of the eviscerated bowel. In all three cases described,1 the children presented initially to a local hospital and were subsequently transferred to a tertiary care facility before operative reduction was initiated. It is likely that earlier operative reduction of the eviscerated bowel would decrease secondary venous congestion of the prolapsed segment, thus minimising further ischaemic changes already initiated by the mesenteric arterial and/or venous tear. This in turn may improve the perfusion of the affected segment; lessen the development of hypothermia, hypovolaemia and sepsis; and possibly increase the final length of viable bowel. A “damage control” laparotomy2,3 performed at an appropriate hospital of initial presentation (one with a surgeon and anaesthetist available), with the principal aim of early reduction of the prolapsed bowel into the abdominal cavity, may improve outcomes in such cases.

Phillip J Carson

To the Editor: We read with interest the recent meta-analysis of rates and methods of suicide in Australia by Large and Nielssen.1 Previous research has shown a marked increase in hanging suicides in women in South Australia over the 15-year period 1986 to 2000,2 which concurs with the findings of Large and Nielssen at a national level between 1988 and 2007.1 One of the features of this increase in hanging suicides, which was not specifically addressed in their excellent overview, relates to trends in specific age groups such as the young. A previous analysis of suicides in individuals aged 16 years and under and for whom autopsies were performed at the Forensic Science Centre (now Forensic Science South Australia; FSSA), in Adelaide,3 included a total of 19 cases of suicide for the 5 years 1985 to 1989, of which seven deaths (37%) were from hanging. The age range of the hanging victims was 14–16 years (mean age, 15 years), with a male to female ratio of 6:1. This compares with a total of 10 suicides among young people, recorded in FSSA files for the 5 years 2005 to 2009, of which nine (90%) were by hanging. The age range of the hanging victims in this group was 10–16 years (mean age, 14.7 years), with a male to female ratio of 4:5. The difference between these two periods was statistically significant (P < 0.005).3 Thus, although suicide in those aged 16 years and under remains an uncommon event,3 developments in deliberate self-harm among the general population may also be reflected in the young. While there was a fall in the total number of young suicide victims, from 19 to 10, between the two periods in the SA study, this has not been matched by a decline in hanging suicides (ie, a significantly greater percentage of suicides in the young in SA now involve self-suspension).3 This finding may be of use to those studying specific issues and trends in youth suicide.

Roger W Byard · Amy Austin · Corinna van den Heuvel