Letters

In developing its first seven standards for implementation by all health service providers in Queensland, the Health Quality and Complaints Commission (HQCC)1 considered the questions raised by Brand and colleagues2 regarding health care standards. We believe that the HQCC standards fit the framework suggested by Brand et al in that they are regulated, the measures of processes and outcomes are quantitative, and the criteria used for their development are evidence-based. In seeking to minimise conceptual confusion, the HQCC has “regulated” existing clinical guidelines and health standards with the aim of improving the quality of health services by requiring providers to establish systems to monitor and report on key aspects of care. The standards address the following areas: Review of hospital-related deaths; Management of acute myocardial infarction on and following discharge; Surgical safety, including antibiotic prophylaxis, prevention of venous thromboembolism, and correct surgery; Hand hygiene; Credentialling and scope of clinical practice; Complaints management; and The duty of health providers to improve the quality of care. The HQCC is now establishing a responsive regulatory framework to monitor compliance with the requirement for all health service providers to implement and maintain quality improvement processes. Brand and colleagues are critical of initiatives that burden providers with data collection and potentially distract from efforts to improve quality.2 But we contend that, without collection, analysis and review of data, the capacity to improve quality is limited. Since July 2007, the HQCC has required all acute care facilities to regularly report against the standards. Our intention is not to make the HQCC into a data repository, but rather to ensure that providers have the ability and the motivation to measure and monitor their own performance. Nevertheless, as a consequence, a unique dataset has been created that reflects wide variation across the state. This approach aligns with the model of metaregulation (or enforced self-regulation) and triple-loop learning espoused by Healy and Braithwaite.3 Ultimately, the purpose of health care standards is to improve the quality of care and drive a culture of quality improvement. Although the HQCC designed its compliance framework to evaluate the impact of its standards on quality and culture of the health system over an extended period, there are already promising indications of a positive effect after only 12 months.

Teresa A Lynne

To the Editor: The article by Statham and colleagues raises important issues about the accuracy of diagnosis by primary eye care providers, with all professionals in the study recording a diagnostic accuracy rate of less than 50%.1 From the general practice perspective, the authors raise a number of important contributors — lack of equipment, ophthalmological expertise and time. Additional factors, such as undergraduate and postgraduate exposure, and targeted training in the diagnosis of sight-threatening acute eye conditions, are also crucial considerations. From a postgraduate point of view, the Royal Australian College of General Practitioners offers a comprehensive continuing professional development program to support broad-based GP training, including in eye disease.2 The Master of Medicine (GP) offered by the University of Queensland also includes a dedicated subject on primary eye care, with particular emphasis on sight-threatening presentations.3 The Division of General Practice in which one of the sites in the report by Statham and colleagues1 sits is currently working with that hospital’s Department of Ophthalmology to institute an education/continuing professional development reform program to better target hospital eye referrals from primary care (Shelley Kleinhans, Health Systems Improvement Program Manager, GPpartners [Brisbane North] Division of General Practice, personal communication). It is very important to describe accurately the dimension of missed diagnosis within primary care — the ensuing challenge is to address it by harnessing the significant momentum within the primary care community for quality improvement.

Claire L Jackson

To the Editor: Between 1997 and 2006, suicide rates fell in all mainland Australian states and territories.1 Despite a rise in the small-population jurisdictions of Tasmania (1997–2006)1 and an earlier rise in the Northern Territory (1981–2002),2 the overall unadjusted national suicide rate has fallen steadily, from 14.7 per 100 000 in 1997 to 9.06 per 100 000 in 2006, the year for which the most recent statistics are available. Trends in suicide occurrence are usually examined with reference to suicide rates. However, because there has been population growth in all parts of Australia, examination of the actual number of suicide deaths in each region is also useful to ascertain whether the demographic changes accompanying population growth can alter unadjusted suicide rates. Annual suicide mortality statistics by state and territory from 1975 to 20061,3,4 show that the number of suicides peaked in every mainland state in 1997 or 1998. Since then, the number of suicides has fallen in New South Wales by 46.5%, in Victoria by 33.5% and in Queensland by 36.4% (Box). There has been larger variation in the number of suicides over time in NSW than in other states. Suicides in NSW rose from under 600 per year (approximately 11 per 100 000 population) in the early 1980s to a peak of 935 suicides in 1997 (14.9 per 100 000) and then fell to 504 suicides (7.4 per 100 000) in 2006. Data available to 2002 indicate that the decline in suicides occurred in almost every NSW health service area and was mainly due to a reduction in the number of suicide deaths in males, including young males.5 By 2006, NSW had the lowest suicide rate in Australia of 7.69 per 100 000 compared with 9.73 per 100 000 for other states and territories. The reasons for the steeper rise and subsequent fall in suicide in NSW compared with other states are unclear, but warrant further investigation with a combined clinical and epidemiological approach. The decline in suicide in NSW coincided with a change to the Mental Health Act 1990 in 1997 that broadened criteria for involuntary care and allowed more people to be treated. However, this factor alone would not explain the extent of the decrease nor the continued decline over a decade. It is possible that programs to prevent suicide or measures to improve access to psychiatric care in NSW have been more successful than those in other parts of Australia. Annual number of suicides in Australian states and territories, 1975–2006* * Data are from the Australian Bureau of Statistics.1,3,4

Matthew M Large · Olav B Nielssen · Steven M Lackersteen

To the Editor: We evaluated follow-up by SMS (short message service) text messaging of contacts of a patient with meningococcal disease. An 18-year-old woman from south-western Sydney was diagnosed with invasive meningococcal disease in July 2008 after presenting to hospital with a rash that appeared after a 2-day prodromal illness. The Sydney South West Public Health Unit identified the patient’s household and similar contacts, and arranged for these individuals to be treated with clearance antibiotics. The patient had visited a bar with friends 3 days before symptom onset. The extent of contact with people in this social network did not warrant treating them with clearance antibiotics. However, it was appropriate to warn them about meningococcal disease as recommended by national guidelines.1 A list of mobile phone numbers of 14 people who visited the bar with the patient was compiled by one of her friends. A text message was sent 2 days after the patient’s diagnosis to everyone on the list via a broadcast messaging service: Message from public health. A friend of yours has meningococcal disease. Watch out for symptoms. Please read the fact sheet at http://www.health.nsw.gov.au/factsheets/infectious/meningococcal.html or call 9515 9420. The message sender appeared as “SMS4U”. Two weeks later, one of us (J E C) made up to three attempts to telephone each of the contacts, explaining that this was a follow-up about a text message they may have received from the Public Health Unit. Contacts were asked whether or not they remembered receiving the message, had viewed the website, and found the information helpful. Twelve were contacted (six men, six women; age range, 18–24 years); all remembered receiving the message, nine looked at the website, and 11 found the message helpful. All were happy to receive the information this way. Some knew of their friend’s illness through other social contacts. This is the first time we have used SMS to communicate information to social contacts of a patient with meningococcal disease. To our knowledge, this is the first reported use of SMS for this purpose, although email and the Internet have been used previously.2 SMS communication appeared highly acceptable to these young people and provided useful information, but it may be less useful in other age groups. SMS has been used successfully in other health contexts — appointment and vaccination reminders3,4 and diabetes education.5 It enables delivery of a concise, timely and consistent message that can easily be broadcast to large groups. There are potential pitfalls: limited information can be conveyed; there is uncertainty regarding whether the message is received (the broadcast service we used provided a “successful send” receipt but not a “message opened” receipt); those without mobile phones cannot be contacted; those without Internet access cannot access web-based resources; and some recipients may not understand the message. The authority of a message from SMS4U (the only available option) was also of concern. We did not exploit the capability of forwarding an SMS message and, by doing so, “snowballing” the information. This could be valuable for alerting large contact networks. Our study was small, and we recommend further evaluation of SMS communication in larger groups.

Johanne E Cochrane · Chris Lowbridge · Patrick Maywood · Stephen J Conaty

To the Editor: We concur with Aratchige and colleagues that mumps in young adults is a “forgotten” disease,1 and believe that mumps control in Australia has suffered from both the successes and failures of our measles elimination program. Among residents of Sydney’s eastern and southern suburbs, 100 cases of mumps were notified in the second half of 2007. Sixty-three per cent of those who contracted the disease were aged 20–29 years, and 65% were male. This compares with an average of 13.6 cases (range, 4–32 cases) notified per annum from 1999 to 2006. During the second half of 2007, one institution managed three cases of severe orchitis in men aged 25–29 years whose diagnosis was confirmed by a positive mumps IgM test. In all three, initial fever and transient parotitis were followed after 7–10 days by severe testicular pain and swelling. Fever and testicular pain continued for a further 1–2 weeks, precluding their return to work. None had been vaccinated. Although mumps vaccine was introduced in Australia in 1980, mumps control has not been an explicit priority compared with measles.2 It seems that public health authorities in industrialised countries have assumed that measles control efforts based on two doses of the measles–mumps–rubella (MMR) vaccine would lead to simultaneous mumps control. While doubt has been cast over the effectiveness of this approach and raised the possibility of a three-dose schedule,3 we agree with the view of Schmid and colleagues that public health authorities should focus on adequate vaccination coverage and adherence to the recommended two-dose MMR vaccination scheme.4 The Australian birth cohort reported by Aratchige et al to have a dip in mumps immunity was the cohort born in the years 1978–1982.1 This group may have avoided natural measles (as well as mumps), missed the Measles Control Campaign in 1998 (which targeted primary-school children with MMR vaccine), and was subject to an ineffective national effort in 2001 to target young adults with MMR vaccine.2 At the time of the 2007 mumps outbreak, this cohort was aged 25–29 years and was the hardest hit. Concerted action to raise the level of two-dose coverage among young adults is urgently needed. Novel strategies exist for targeting this highly communication-aware age group through convergent Internet and mobile phone technologies. Social network sites such as MySpace and Facebook are heavily used by young people, and the proportion of mobile phones with Internet access is increasing. Sporting clubs and major entertainment events are another avenue to be considered with respect to both their physical and virtual locations (eg, posters at the clubs or events, advertisements on their websites). A comprehensive guide has recently been produced for Internet-based prevention of sexually transmitted diseases.5 It is well and truly time to adapt such methods to the promotion of MMR vaccination.

Mark J Ferson · Pam Konecny

To the Editor: The University of Sydney recently established a Chair of Radiology and appointed Professor Ming Wang as the first full Professor of Radiology in New South Wales. This long overdue appointment resulted from a bequest of Arthur Parker-Hughes, after whom the Chair is named. Likewise, it was largely through the generosity of Edgar John Rouse that the first university department of radiology in Australia was established at the University of Melbourne in 1965; I was appointed Foundation Professor. Considering the role of radiology in modern medicine, it is remarkable that to date in Australia, establishing Chairs of Radiology depends largely on private sponsorship. Soon after Roentgen’s discovery of x-rays in 1895, Scandinavian countries promoted the triad of medicine, surgery, and roentgenology, as it was then designated, as the basis of clinical management. Radiology departments were nurtured in their universities, and were leaders in research. European medical schools followed suit, and since about 1960, university radiology departments in the United States have been at the forefront of research. In Australia, university clinical departments developed relatively late, and, when they did, the Australian Universities Commission recognised the need to provide space and basic staffing for these new departments to achieve the desired academic standard.1 In recent decades, development of new university radiology departments has languished. Established in 1975, the radiology department at Flinders University closed in 2002. The radiology department at the University of Queensland began in 1977. Currently, medical schools rely on busy radiologists employed by teaching hospitals for academic input, and provide them with various adjunct titles. The range of diagnostic imaging modalities and interventional radiological procedures continues to expand, providing significant research opportunities. Medical science students should understand what is available, the benefits and limitations, inherent risks, and should appreciate the economic burden on the community from inappropriate use. Also, radiology provides an excellent means of teaching basic medical subjects, such as anatomy and pathology. Considering the importance of radiology in the health system, and how its academic status is recognised by leading overseas universities, surely it is reasonable that each medical school in Australia should include a department of radiology, or, at least, a full Professor of Radiology, financed primarily from within the university.

William S C Hare

To the Editor: The author of “High levels of confusion for cholesterol awareness campaigns”1 identifies my comment as the source of her perplexity. My remark, which Hall quotes in relation to the “Test the Nation” campaign, was actually given in response to a question about the Pfizer-sponsored “National Cholesterol Awareness Campaign”, which ran simultaneously. It seemed too inconsequential to request correction of the relevant newspaper article,2 because public health guidelines differ in regard to the target population for lipid testing. My comment reflected conservative Australian guidelines.3 It is well known that other sources recommend more widespread testing of adults.4 Hall’s confusion was a rhetorical device to justify her discussion of “condition branding” and to “explore the motivations” of the campaigns. Hall’s article attacks two programs that promote diet and lifestyle management of cardiovascular risk. It also criticises pharmacological treatment, thus eliminating all available options to address this important problem. The article undermines the tenuous availability in Australia of plant sterol-containing products, such as yoghurt, but provides no alternative strategies. It fails to take responsibility for its potential negative impact on the implementation of nutritional and other life-saving interventions. It is disconcerting that such a negative article has emanated from a so-called Centre for Health Initiatives. The cholesterol awareness campaigns are likened to “an unnecessary focus on an unimportant health problem” and disparagingly compared with a program devoted to public awareness of fungal nail infections. Understandably, the latter led to professional irritation and frustration. By contrast, the formal involvement of the Royal Australian College of General Practitioners in the Test the Nation–National Cholesterol Education Program of Australia (NCEPA) represents an effort to ensure that the initiative was justified, that the content was relevant, and that the logistics were attuned to primary care practice. The participation of professional organisations reinforced the quality, relevance and independence of the information provided. The dietary advice that was distributed by the NCEPA was widely acclaimed. Dyslipidaemia accounts for 49% of the attributable risk of coronary heart disease.4 Full implementation of risk factor guidelines could massively reduce cardiovascular disease,5 but public and professional adherence to guidelines is suboptimal.6 Chen and colleagues report that, in addition to those with diabetes or coronary heart disease, over 700 000 Australians are at high risk,7 but most Australians are unaware of the consequences. It seems extraordinary that anyone with a professed interest in public health could be so opposed to public education about risk factor management. Consequently, Hall’s article itself generates further confusion. The health sector is in the process of responding to calls for greater independence from commercial interests. Both cholesterol programs illustrate the implementation of many of the suggested changes. Unfortunately, Hall’s article suggests an open-ended list of demands that will be impossible to satisfy. Calls for further change need to be more constructive and clearly enunciate realistic proposals, supported by evidence that the net impact on Australian health care has been (or will be) beneficial. The article implies restrictions that would be impractical in other sectors. Standards for the interaction between industry and the health sector should be an example to emulate, rather than a soft target that loses step with normal practice.

David R Sullivan

In reply: My article1 was submitted for debate when two separate, industry-sponsored cholesterol awareness campaigns were simultaneously targeting the Australian public. There is potential for public confusion following exposure to concurrent campaigns with differing sponsors, creative techniques, and messages such as “Test the Nation”. My article did not diminish the importance of cholesterol screening (nor of the prevention of or treatment for hyperlipidaemia), and indeed it reiterated the National Heart Foundation guidelines. Its intent was to raise debate about industry-sponsored disease awareness campaigns, as there is growing concern in Australia about “disease mongering”2 and the evidence that this is occurring in the cholesterol market in the United States.3 I agree with Sullivan that the Australian public needs education about asymptomatic risk factors, including hypertension and hyperlipidaemia. Ideally, this would include clear information and non-emotive marketing techniques to convey who is most at risk, as well as transparent disclosure of sponsor interests.4 Contrary to Sullivan’s charge, I believe there are many opportunities for quality health education and behaviour change programs, several of which the Centre for Health Initiatives is currently undertaking.5 The intent of my article was to generate critical analysis of industry-sponsored campaigns in order to improve their public health benefit.

Danika V Hall

To the Editor: The recent article by Penington, highlighting the apparent neglect by Australian hospitals of actively participating in research over the past two decades,1 is both timely and concerning. During this time, many of our hospitals have seen themselves increasingly as clinical service providers, with teaching and research perceived as additional costs rather than contributions to their status, to the quality of patient care and to clinical and scientific discovery. My point is not simply to add weight to Penington’s eloquent historical, contemporary and strategic analysis of the nexus between hospitals and universities, but to explore areas that he touched on that need more detailed examination. I refer to the potential link between hospitals and community-based primary care. As Penington points out, “health care is increasingly provided outside hospitals”, yet this vital link between hospitals and primary care is often defunct, particularly when it comes to general practice. When Penington refers to hospitals working with “general practice networks” to meet the demands of an ageing population and chronic disease, we need to ask: which hospitals and which networks? In Melbourne, we have health networks that include groups of hospitals with extended primary care and community facilities and responsibilities. Driven by casemix funding, the hospitals or health networks have a vested interest in primary care to ensure short patient stays. However, not all states are the same and not all hospitals have similar links with the community. The other issue of concern is that general practice networks are often politicised and factionalised. Divisions of General Practice do not speak with a single voice, and their state-based organisations and national body do not always represent the views of regional Divisions. Added to the mix of 120 Divisions, we have 22 regional general practice training providers, the colleges (Royal Australian College of General Practitioners, Australian College of Rural and Remote Medicine), the Rural Doctors Association, the Australian Association for Academic General Practice, the Australian Medical Association, etc. Integration of the research, teaching and training activities of general practice or primary health care with hospital networks can only be achieved at a regional level. We need to think globally, but act regionally. This calls for the formation of new regional consortia including universities, health networks (hospitals), Divisions and regional training providers (responsible for general practice registrar training) to work together on national clinical and health service research agendas in large and well defined geographic regions. If the National Health and Hospitals Reform Commission is to take its role seriously, it will need to move beyond the confines of traditional hospital settings and explore opportunities in the community.

Leon Piterman

To the Editor: Penington identifies the appointment of a National Health and Hospitals Reform Commission (NHHRC) and the suspension of 5-year Australian Health Care Agreements as a “once in a generation” opportunity to rediscover university teaching hospitals for Australia.1 He identifies changes to health funding in Australia since 1975, the growth of “cost shifting” between federal and state governments, and reduced funding for university functions in hospitals as important contributors to the decline of university teaching hospitals in Australia.1 We agree with this analysis, but propose that the privatisation of many outpatient clinics as a result of cost shifting has had a disastrous effect on the clinical training of medical students, residents and registrars. Moreover, the reduced funding of university functions in hospitals has been replaced in “teaching” and community hospitals by industry funding, and the perception that industry has “bought” patients for their “research” agenda by providing data management services through per capita payments and gifts or perks for clinicians. The infrastructure sustaining clinical research should not be so reliant on industry. Funding from the federal government (in partnership with state governments) is needed to nurture independent research in university teaching hospitals. Penington highlights the fact that the key element of the university teaching hospital model was leadership of all units by academic clinicians with questioning minds. Since 1975, we believe leadership of units in “teaching” hospitals has changed such that very few are now led by academic clinicians. We recently experienced a lack of interest by medical specialists in supporting clinical research that had been approved and funded by the National Health and Medical Research Council (NHMRC). The project will evaluate doctor–patient communication about treatment options in oncology, including clinical trials. Participating doctors were required to post letters of invitation to patients and audio-record one consultation per patient recruited (20 per doctor). Specialists in all major teaching hospitals in New South Wales and Victoria were contacted; 17 of 41 specialists contacted in NSW (41%) and 15 of 52 contacted in Victoria (29%) have agreed to participate. The most common reasons doctors gave for not participating was that they were too busy or that there was no reward for participation. It is notable that no specialist from two major teaching hospitals — one in Sydney (3 contacted) and one in Melbourne (6 contacted) — agreed to participate. A concerning theme is that the motivation to participate in trials is driven by financial incentives rather than the importance of the question being addressed or interest in supporting novel investigator-initiated clinical research. A starting point for improving the quantum and calibre of “independent” clinical research in university teaching hospitals would be to support clinical research infrastructure by providing per capita payments for recruited patients — a model used by industry. Rebuilding independent research capacity in university hospitals will improve the standard of research in this country, and foster clinical research training. We believe reviving the university teaching hospital model in Australia is an important task for the NHHRC.

Rachel F Dear · Martin H N Tattersall

To the Editor: A causal connection between alcoholic mothers and developmental delays and physical abnormalities in their babies was identified in the 1970s and termed fetal alcohol syndrome (FAS). Other less extreme but still disabling effects fall under the umbrella term of fetal alcohol spectrum disorder (FASD).1 Some studies have found higher prevalences of FAS among indigenous children in several countries, including Australia.2-4 However, none are as high as those cited in a webcast video program produced by the Rural Health Education Foundation and accredited by (and examinable for professional development points awarded by) the Royal Australian College of General Practitioners, Australian College of Rural and Remote Medicine, Pharmaceutical Society of Australia, Royal College of Nursing Australia, and the Australian Physiotherapy Association.5 In this program, an Indigenous Australian health worker states that 540 out of 614 children aged under 12 in an (unnamed) Indigenous community are “already showing signs of primary and secondary disabilities associated with FAS and FASD”.5 These findings are not sourced and therefore not verifiable. We believe that unsubstantiated claims such as this can fuel racism against Indigenous children. Research conducted within a Queensland Aboriginal community school found that teachers were using information such as that provided in the video program to explain students’ poor school performances, when no formal diagnoses of FASD had been made for the children.6 It is racially discriminatory to impute a lifelong and incurable disability to Indigenous children when no teratogenic condition has been diagnosed. The prevalence of FAS and FASD has not been comprehensively established in Indigenous or non-Indigenous communities in Australia. There are other reasons why Indigenous students might not be succeeding in school, such as hearing impairments, being taught in Standard English (which is not their first language), or being assessed with culturally and linguistically biased school and IQ tests.6,7 Stigmatising them as intellectually impaired can lead to low self-esteem, behavioural problems, and absences from school. These outcomes have been noted here and overseas,7 yet some educationalists persist in blaming prenatal factors (including “bad genes”) rather than addressing the more difficult issues of systemic racism in the educational setting. In light of the federal government’s campaign to protect Indigenous children and to encourage their educational potential, as well as its general attack on binge drinking, it is essential to fund programs that address FAS and FASD in both Indigenous and non-Indigenous communities. Further, all governments need to support the dissemination of clear and substantiated information on this preventable cause of intellectual impairment.

Loretta R de Plevitz · Judith S Gould · Terrina M Smith

In reply: The Rural Health Education Foundation is sorry that a statement on its live-to-air, interactive program was interpreted as being racist. We can see how this has occurred and have added an addendum to the program’s website description to avoid any future misinterpretation of what the presenter was intending to communicate.1 The research on fetal alcohol spectrum disorder (FASD) quoted in the Foundation’s program by an Australian Aboriginal health worker is currently unpublished, and was undertaken in 2000 during a fieldwork placement as a requirement for a Master of Applied Epidemiology (Indigenous Health).2 This research identified that 540 out of 614 children aged under 12 in an (unnamed) Indigenous community had prenatal exposure to alcohol that exceeded the National Health and Medical Research Council (NHMRC) recommendations on alcohol consumption during pregnancy,3 and were subsequently at risk of primary and secondary disabilities associated with fetal alcohol syndrome (FAS) and FASD. During the program’s live discussion, the panel member incorrectly stated that the research cohort was already showing signs of primary and secondary disability related to FAS and FASD. Two individual projects within the Masters research contributed to the findings. The first, examining the risks of maternal alcohol use for child physical and psychological development, involved retrospective, longitudinal analysis of a data subset from an existing study over a 5-year period of 8556 women who received antenatal care, with subsequent follow-up of the mothers and their children when the children were 5 years old. The second was a descriptive study involving all women in the (unnamed) Indigenous community who gave birth within a 5-year period immediately before the research. Medical record audit and focus groups (talking circles) were conducted. This research is currently being expanded in the context of PhD studies, and the candidate intends to submit the new findings for publication in the near future. The video program content was developed in consultation with a group of health professionals with expertise in the area of Australian Indigenous health and FASD. At all times, the Rural Health Education Foundation seeks to provide positive examples of “what works” in its location-based filmed case studies. The Foundation and its representatives in no way meant to infer racism or discriminate against this (or any other) group of Indigenous Australians.

Brian D Bowring · Amanda Little

To the Editor: The randomised controlled trial recently reported by Blashki and colleagues does not support their hypothesis.1 The drop-out rates in both arms of the trial were very high — only 62% of general practitioners in the intervention group and 54% in the control group completed the trial. One cannot have any confidence in their conclusion that a short training course can improve GP skills in the provision of cognitive behavioural strategies (CBS). For example, what if the 38% of GPs in the intervention group who dropped out actually deteriorated in their CBS skills and therefore declined to be videotaped? Furthermore, only 56 of 1021 GPs in Victoria were willing to enrol in the trial. The authors concluded that their findings could only be applied to GPs who have a special interest in mental health. Perhaps the low participation rate indicates another more relevant idea — that GPs have had enough of “training models” being imposed on their lives. A recent systematic review has confirmed the low impact that educational training programs have on GPs for the management of mental health problems.2

Marjan Kljakovic

In reply: Our words in conclusion to our article were carefully chosen as: “Competency in CBS [cognitive behavioural strategies] in highly motivated GPs [general practitioners] can be improved by a brief training intervention”1 (italics added) — not that all such interventions will lead to improvements for all GPs, but that well designed and conducted training for selected GPs can do so. Research so far leaves open the possibility of large enough effect sizes for GP mental health training to be relevant to policy.2 The review cited by Kljakovic was limited in scope and noted the poor quality of studies included.3 A drop-out bias in our study, as proposed, seems most unlikely to us. Rather than being imposed, this training model was developed with GPs, by GPs and for GPs, and so might achieve better results than previous interventions studied. Funding for GP participation such as that more commonly available in drug studies might have increased participation. It is true that a great deal of training has been offered to GPs, and we hold that our study shows that such training can lead to GPs significantly improving their skills in this area. GP training should be considered within multifaceted interventions to improve primary mental health care.4

Grant A Blashki · Leon Piterman · Graham N Meadows · David M Clarke · Vasuki Prabaharan · Jane M Gunn · Fiona K Judd

To the Editor: Sondergaard is essentially correct in his criticism of the Australian and New Zealand College of Anaesthetists (ANZCA).1 And ANZCA president Leona Wilson’s obfuscatory response to the criticism provided me with little reassurance.2 We all know that our health care system is “highly complex” and involves multiple jurisdictions, but these points have no relevance in determining whether Sondergaard is adequately trained and competent to work in Australia as a specialist. To decide that someone with his history and qualifications cannot give anaesthesia unsupervised is patently nonsensical, as is the insistence that experienced overseas specialists must take the College’s final exams. All of us who were Fellows of the ANZCA precursor, the Faculty of Anaesthetists of the Royal Australasian College of Surgeons, were granted automatic Fellowship of the ANZCA on its formation in 1992, as were some senior practitioners who had never sat the Faculty exams. The College can, it seems, arbitrarily waive the exam requirement for some, and it regularly awards Honorary Fellowships to distinguished overseas visitors. However, these doctors are not the competent working clinicians with overseas qualifications who would, if they could, take up vacant positions in rural areas, such as Katoomba just outside Sydney. Here, in October last year, a woman in labour was turned away from a hospital for want of an anaesthetist and gave birth in an ambulance by the roadside. I worked as a specialist anaesthetist in Sweden for nearly 2 years and can verify that Scandinavia produces competent anaesthetists. The attitude of the ANZCA to overseas-trained specialists seems elitist and denies the Australian people access to the services of competent people who happen to have learned this essential specialty elsewhere.

James F Wilkinson

To the Editor: The incidence of type 1 diabetes mellitus (T1DM) is increasing in Australia.1,2 There is also general consensus that the incidence of diabetic ketoacidosis (DKA) is increasing in children, as noted in an Australian study.3 We conducted a retrospective audit of the referral pattern of patients with newly diagnosed T1DM presenting to the Children’s Hospital at Westmead, a tertiary referral centre serving the population of western Sydney. Referral data were available for 191 of 204 patients with newly diagnosed T1DM admitted to the hospital between January 2003 and December 2004. Most patients (150; 79%) had presented to their general practitioner before admission to hospital, and the remainder had initially presented to an emergency department. In the former group, the diagnosis of diabetes was indicated in referral letters or admission notes for 128 patients (85%), while a diagnosis other than diabetes (eg, gastroenteritis, urinary tract infection, sepsis) was made for 22 patients (15%). DKA was less common among patients whose referral letter indicated a diagnosis of diabetes compared with those with an alternative or no diagnosis or without a referral letter (27% v 47%; P < 0.001). Most patients (105; 70%) were referred to an emergency department within 24 hours of presentation to the GP, and their rate of DKA was lower than in those referred after 24 hours (31% v 51%; P = 0.03). These data suggest that better understanding by primary carers of the symptoms of new-onset T1DM and earlier referral are significantly associated with reduced risk of DKA. Most patients who first saw a GP (125; 83%) had initial investigations arranged; bedside urinalysis and/or measurement of fingerprick blood glucose levels were performed in 66%, while 31% were sent for formal blood tests. Patients who had bedside investigations performed had a significantly lower rate of DKA than those who had only formal blood tests or no investigations performed (26% v 52%; P = 0.002). It is noteworthy that, among patients who first saw a GP, 23 (15%) were diagnosed with diabetes but were not referred to an emergency department within 24 hours. The reasons for this are unclear but may be due to the GP waiting for confirmatory blood test results. The Australasian Paediatric Endocrine Group and International Society for Pediatric and Adolescent Diabetes guidelines recommend immediate referral for suspected new-onset T1DM, as DKA is fatal if left untreated.4 A public awareness campaign conducted in Italy in the 1990s was successful in reducing the incidence of DKA in children with newly diagnosed T1DM.5 Australian communities might benefit from a similar campaign to encourage prompt identification of symptoms of diabetes in childhood, prompt bedside investigations, and immediate referral to hospital for definitive care.

Maria E Craig · Catherine H Wong · Joanna Alexander · Ann M Maguire · Martin Silink

To the Editor: Gilbert and colleagues1 report thyroid function test results in a large number of pregnant women in Western Australia during the first trimester. While assessment of thyroid status is increasingly important in pregnancy, they do not present a strong enough argument for their reference ranges to be adopted. Their controls consisted of only 100 blood donors, and it is not clear whether these were age-matched with patients. Differences between pregnant and non-pregnant thyroid hormone ranges were too small to justify use of separate ranges. We assume from the article that the controls were not screened for thyroid antibodies. Prevalence of thyroid autoimmunity is high in women of reproductive age, whether or not they are pregnant.2 Serum thyrotropin (TSH) concentration is reduced in up to 20% of women during their first trimester, often with modestly increased thyroid hormones. The thyroid-stimulatory effect of human chorionic gonadotropin may help ensure adequate thyroxine delivery to the fetus. It is surely more important for clinicians to understand this than to have reference ranges that conceal normal physiological changes. Gilbert et al do not state whether patients with multiple or assisted-conception pregnancies were included — both are more likely to have abnormal thyroid test results.2 Their detection limit for TSH and the lower limit of normal differed by only 0.01 mU/L — they could therefore not reliably distinguish low TSH from suppressed TSH. They screened only 61% of pregnant women in WA. It is inconceivable that there was not a selection bias, as current guidelines3 advocate only screening high-risk groups such as those with a history of thyroid disease or previous poor obstetric outcome. Ethnic differences in TSH levels have been reported. However, data from the United States National Health and Nutrition Examination Survey (NHANES) suggest that TSH levels in Hispanics are no different to those of white people,4 contrary to what is suggested by Gilbert et al.1 Increased miscarriage risk may relate to autoimmunity itself, rather than altered thyroid function. The study by Negro et al5 is, to date, the only one showing a decrease in miscarriages when thyroxine is given to thyroid antibody-positive women. However, the TSH level before thyroxine was given was comfortably within the normal range reported by Gilbert et al. Publications in this complex area are only informative if they tell us something about thyroid physiology or about diagnosis and management of thyroid disorders. While laboratories must validate their reference ranges, it is unlikely that those reported by Gilbert et al could be generalised to the ethnically diverse and geographically dispersed Australian population. Also, as described,1 patients would have to be screened routinely for thyroid antibodies to ensure that the quoted ranges were applicable.

Richard L Kennedy · Usman H Malabu · David Porter

To the Editor: We read with interest Giles and colleagues’ recent article, which examined the adoption of strategies to reduce perinatal transmission of HIV infection in Australia.1 They found that uptake of strategies to reduce perinatal HIV transmission had increased, with widespread use of antiretroviral therapy (ART) and breastfeeding avoidance. The authors also noted that caesarean birth was a strategy less commonly utilised by women with HIV infection. They made particular comment about the caesarean delivery rate for women known to have HIV infection in Western Australia. It was disappointing that the authors did not refer to our recent publication describing the low rate of perinatal HIV transmission in WA using an individualised delivery modality policy.2 In our consecutive series of 56 pregnancies between 1991 and 2005, 48 (86%) were managed by a multidisciplinary team, with 98% (47/48) of women receiving ART (one woman actively declined this intervention). Only one baby in the group who received care through the multidisciplinary team acquired perinatal HIV. This pregnancy occurred in 1991 in a woman with advanced disease who received zidovudine monotherapy, a situation not applicable today. Elective caesarean delivery was based on either obstetric indications or a high HIV RNA level; 75% of women in our series had a vaginal delivery. The findings of our study were of particular note because 39% of mothers were Aboriginal, and predominantly from rural and remote regions of WA. Although the patient numbers in our study were small, the current international evidence does not support mandatory caesarean delivery for women receiving ART with undetectable plasma HIV RNA.3 The risk of vertical transmission in this circumstance is low, and caesarean birth is associated with short- and long-term morbidity (most notably, placenta accreta). Recent series have shown a trend of increasing vaginal birth rates among women with well controlled HIV infection.4,5 When infection is well controlled, we believe that the mode of delivery should be individualised, and vaginal birth should be an option for women who desire this delivery method. It is disappointing that Giles and colleagues appear to imply that the low caesarean delivery rate in WA is a reflection of suboptimal HIV care processes, rather than evidence-based practice.

Marisa T Gilles · Martyn A French · Jan E Dickinson

In reply: We were interested in Gilles and colleagues’ response to our analysis of the uptake of interventions to prevent perinatal HIV transmission in Australia. As Gilles et al state, the reported rates of perinatal HIV transmission are low in Western Australia, where the choice of delivery modality is individualised. We agree that the additional benefit of elective caesarean section in women being treated with highly active antiretroviral therapy with an undetectable viral load is not known. We also agree that elective caesarean section is associated with potential risks, and women should have a choice regarding mode of delivery. This choice should be informed by other obstetric factors, maternal viral load, and the clinical setting in which delivery takes place. Geographic variation in such factors is likely, and will certainly contribute to differences across states in the uptake of preventive interventions. Ongoing national surveillance will help ensure that women with HIV infection and their children benefit as much as possible from evidence-based obstetric practices.

Michelle L Giles · Ann M McDonald · Elizabeth J Elliott · John B Ziegler · Margaret E Hellard · Sharon R Lewin · John M Kaldor

To the Editor: Wilhelm and colleagues falsely concluded in their recent study that general practitioners in metropolitan Sydney and rural New South Wales had a low rate of recognition of psychological problems overall.1 Furthermore, Wilhelm et al took GPs’ judgements of the presence of psychological problems as the benchmark for “caseness” because of the difference between GP practice and psychiatric practice in the process of assessing psychological problems in consultation. My disagreement lies with what the researchers meant by “overall” and by “caseness”. The rate of recognition of caseness of psychological problems by GPs will vary according to the nature of the cases under consideration. In their study, Wilhelm et al found that they had complete data on 76% of their patients. Our work in New Zealand found that a major variable that influenced diagnostic behaviour within consultations was the frequency with which patients had previously consulted their GP.2 The more frequently the patient had been seen in the previous year, the more likely the GP was to diagnose a mental disorder. A second variable found in our research was the presence or absence of disability in the patient.3 GPs were less sensitive to the presence of mental disorders if there was little concomitant disability, and in sub-threshold cases, the presence of disability increased the chance of GPs identifying clinically significant symptoms. In general practice, the “new patient” is a different kind of case than the frequent attendee. Similarly, a patient diagnosed with depression who is seriously disabled is a different kind of case to the more common kind found in general practice — namely, a patient diagnosed with depression but with little or no disability. It would not be surprising if Wilhelm et al were to find that among the 20% of patients overall in whom GPs identified psychological problems, many were “typical cases” seen by GPs — namely, frequent attendees and those with disability.

Marjan Kljakovic

In reply: I must apologise for the inclusion of a comma in the first sentence of the conclusion in our article’s abstract, which changes the sense of the sentence.1 That was my oversight. It should read “Low rates of recognition of psychological problems by GPs [general practitioners] and infrequent treatment for those presenting with somatic symptoms ...”, meaning that there are low rates of recognition and treatment in patients with somatic symptoms rather than in patients overall. We were reflecting the need for more recognition of how to deal with depression and anxiety in the presence of somatisation. We think the 12-item Somatic and Psychological HEalth REport (SPHERE-12) is a useful instrument, but that it has an intentionally low “caseness” threshold and needs to have some other tool to increase clinical relevance. Kljakovic also comments on our use of GP judgement as a benchmark for caseness. The thrust of our article was to see how GPs make judgements and which of three different types of screening tool may assist them. This is not to say that GP judgement is an overall “gold standard” for caseness in an epidemiological sense. It is certainly true that new patients are very different from those who are frequent attendees and/or well known to the GP. The screening tools are probably more useful in the first instance or when there is a change in the patient’s mood. However, we wished to test these measures across the range of people seen by each GP, and the individual GPs were given the results from their own practices. The feedback from GPs was that these tools did prove helpful in drawing their attention to people they already knew about and also in identifying some that they did not. Such screens can also save time by ensuring that certain questions are routinely asked and responses are tracked, so the GP can see the results, reflect on them, and go on to ask other questions that build on this information, helping to make better use of the “face to face” time rather than having to run through them in the interview.

Kay A Wilhelm

To the Editor: A 56-year-old woman was recently admitted with recurrent large-volume haemoptysis associated with left-sided tearing thoracic pain. Growing up on a sheep farm in rural New South Wales, she had been diagnosed at age 8 years with pulmonary hydatidosis, which remained dormant on periodic clinical assessments. However, 2 years before presentation she started to cough up gelatinous material containing scolices of Echinococcus granulosus. Surgery was declined at that time due to the anticipated complexity of the operation and associated high perioperative risk. Long-term anthelmintic therapy was commenced. On admission, a computed tomography (CT) scan showed a contained aortic pseudoaneurysm (Box), consistent with rupture of the aorta into the hydatid cyst. Other images showed the cyst containing gas, indicating communication with the airway. After stabilisation, the patient was transferred to a cardiothoracic centre. A left upper lobectomy with dissection and removal of the mediastinal cyst was undertaken through a median sternotomy, and the aortic fistula was successfully repaired using bovine pericardial strips. Intraoperatively, there was no evidence of pericardial involvement. Histopathological examination showed a disrupted and degenerate hydatid cyst without a germinal layer and no protoscolices. The patient received albendazole for 6 months after surgery and her recovery was uneventful. Hydatid disease is caused by the intestinal parasitic tapeworm E. granulosus which, in Australia, is most prevalent in the eastern half of NSW at higher altitudes.1 Symptomatic intramural aortic-wall hydatidosis causing aortic-wall rupture and pseudoaneurysm formation has been described in fewer than a dozen cases.2 Hypotheses for arterial-wall invasion include dissemination during cardiac surgery; entry through vasa vasorum or pre-existing small intimal tears or aneurysms; or partial incorporation of the aortic wall into the hydatid pericyst.2,3 We identified four case reports in adults describing fistula formation between a pulmonary hydatid cyst and the aorta, including three European cases3-5 and one South African case (published twice).6,7 All patients were middle-aged men: two presented with chest pain and large-volume haemoptysis, one with anaphylactic shock and bilateral ischaemic lower limbs from aortic-wall hydatid cyst emboli, and one with a cyst eroding the abdominal aorta (found incidentally during surgery for a coeliac trunk aneurysm). One patient died during removal and another patient after removal of the primary cyst. Haemoptysis in pulmonary hydatid disease is a common presenting symptom. Mechanisms include pressure erosion of a bronchus, obstructive infection, cyst rupture or — very rarely, and emphasised in our case — erosion of a major vascular structure. Aortobronchial fistula resulting from pulmonary hydatidosis A: Contrast-enhanced thoracic computed tomography scan showing consolidation and cavitation within the left upper lobe. A multiloculated 4 cm diameter peripherally calcified hydatid cyst was present in the medial aspect of the left upper lobe (white arrows), penetrating under the aortopulmonary window. Contrast medium extended posteriorly into the base of the lesion, suggestive of an aortic leak (black arrow). B: Coronal reconstruction of the aortic arch demonstrated a round collection of contrast medium with a 1.4 cm base (black arrow), consistent with a contained saccular aortic pseudoaneurysm.

Stefan Buchholz · David Sowden · Troy Stapleton · Peter Pohlner · Craig Wright

To the Editor: In their recent article, Chen and colleagues argued for an increase in the number of Australians being treated with lipid-lowering drugs.1 It would appear pertinent to question the economic and therapeutic value of such an increase. The daily number of doses of statins in Australia increased by 1218% over the decade to 2005.2 Australia has considerably greater use of serum lipid-lowering agents than other Organisation for Economic Co-operation and Development countries, with annual costs in 2004 of $1.61 billion.2 Four years later, following the introduction of rosuvastatin and recent media reports regarding reductions in heart attacks and strokes, this figure must have escalated. Of necessity, the recommendation for the use of lipid-lowering therapy is largely based on extrapolation from tightly controlled clinical trials to clinical practice. This postulate has been questioned by a number of authorities, essentially due to eligibility requirements of trials excluding 40% of men and 80% of women,3 or, more importantly, because of failure to reach target levels, ranging in the world literature from 21% to 73% (references available from the author). This failure may be due to inadequate dosing or poor compliance, with non-compliance noted to be in the order of 35% at 2 years.2 In a submission to the Australian Government’s inquiry into health funding, I posed the hypothesis that if patients were required to undertake appropriate lifestyle changes before initiation of pharmaceutical intervention, annual savings of $130 million could be anticipated.4 Such an activity would also show benefits in reducing hypertension and obesity and improving glucose control in diabetic patients. Low high-density lipoprotein (HDL) levels and moderate elevation of triglycerides (to an extent that the triglyceride–HDL ratio is greater than 2) are associated with a preponderance of type B low-density lipoprotein (LDL) particles, which are known to be highly atherogenic. High-intensity interval training over a 6-week period increases the HDL levels in patients with initial levels < 1 mmol/L and reduces triglyceride levels, to the extent that the triglyceride–HDL ratio is significantly reduced to less than 2.5 Frequent requests to pharmaceutical companies for data regarding the effect of their preparations on elevating HDL levels in patients with levels < 1 mmol/L have been fruitless. In Australia, the incidence of coronary heart disease events decreased from 1994 to 2005, by 32% for men and 34% for women.6 Similar trends were observed for deaths from coronary heart disease and stroke.6 Year-by-year analysis of these trends fails to demonstrate any particular response in any one year. During the period 1997–2005, the increase in defined daily doses of statins rose from 20 per 1000 population per day in 1997 to nearly 180 per 1000 per day in 2005.2 One might have thought that this increase in statin therapy would have resulted in a far greater reduction in the incidence of cardiovascular disease and deaths than previously. However, this was not so — the trend continued unchanged in the “post-statin era” and of similar magnitude to the “pre-statin era”. In the words of the late Professor Julius Sumner Miller, “Why is it so?”

Michael A Neaverson

A 5-month-old boy with known congenital varicella syndrome presented to our hospital emergency department with generalised herpes zoster (shingles). The child was born in Australia. His Sri Lankan-born mother had developed chickenpox in the second trimester of pregnancy. Examination and investigation of the child at birth for complications of congenital varicella syndrome had revealed only skin changes on the left thigh (Box, A). A new vesicular rash had evolved over 3 days, initially involving right T8 (Box, B) and L4–5 (Box, C) dermatomes, then progressing to cover the entire body. There was no clinical evidence of visceral involvement. Cicatricial scarring had replaced the congenital skin changes (Box, D). Varicella zoster virus was isolated from vesicular fluid. Oral valaciclovir was prescribed for 7 days because the generalised nature of the rash demonstrated an insufficient immune response to varicella reactivation. The symptoms rapidly resolved, with no new scarring. A maternal chickenpox infection during pregnancy can be severe and life-threatening, and can also cause in-utero infection, which may be fatal or result in congenital abnormalities.1 Important features of congenital varicella syndrome include: dermatomal cicatricial scarring (highlighted by this patient); limb defects; intrauterine growth restriction; ophthalmological defects (chorioretinitis, optic atrophy, cataract); gastrointestinal or genitourinary abnormalities; neurological defects (developmental delay, seizures, deafness, limb paralysis, microcephaly).1 Shingles is caused by reactivation of varicella zoster virus. Childhood shingles is uncommon (incidence, 0.05%/year), rarely indicates primary immunodeficiency,2 and occurs more frequently following congenital infection (4.4%/year)2 or chickenpox infection during infancy (0.4%/year).2 Antiviral treatment of shingles in immunocompetent young children is not usually recommended, as complications (including post-herpetic neuralgia) are rare.3 Chickenpox has been mainly a childhood disease in Australia, but in many tropical countries it predominantly affects adults. Immigrants to Australia from these regions (including the mother of our patient) may remain susceptible to varicella infection.4 At least 5% of Australian women of childbearing age were born in tropical countries.5 Varicella vaccine is highly effective and is listed on the National Immunisation Program Schedule6 for childhood immunisation. Lowering the community prevalence of varicella infection by routine childhood immunisation can help protect non-immune adults and immunocompromised patients.1 Opportunistic, proactive identification and immunisation of non-immune adults, particularly for both prospective parents before pregnancy, could help prevent serious consequences. Cicatricial scarring and shingles in a 5-month-old infant with congenital varicella syndrome

Elizabeth K Nairn · Joshua Wolf · Jim P Buttery

To the Editor: Cardiac rehabilitation (CR) is an underutilised evidence-based treatment.1 Between 1 March 1998 and 28 February 1999, we surveyed 1933 patients aged 20 to 85 years discharged from public hospitals in the Hunter region with principal discharge diagnoses of acute myocardial infarction, unstable angina pectoris, congestive heart failure, and ischaemic heart disease. Patients undergoing coronary artery bypass graft surgery and percutaneous coronary intervention were also included. Among the 1202 respondents (62%), 493 (41%) reported being referred to CR, 309 (26%) reported attending at least one session, and 233 (19%) reported completing all or all but one session.2 The factors associated with referral were younger age, previous participation in CR, admission to a hospital providing CR, a discharge diagnosis of acute myocardial infarction, and coronary artery bypass surgery.3 We provide the following data, pertaining to non-referred patients, within the context of recent government initiatives to improve access to evidence-based treatments. Fifty-seven per cent of respondents (688) had not been referred to CR (2% did not answer this question), 645 of whom had not attended previously. The median age of these 645 people was 70 years. Most were male (64%), married (62%), had not completed high school (54%), were not in full-time employment (81%), had not been admitted to a hospital that offers CR (55%), did not have a discharge diagnosis of acute myocardial infarction (78%), and had not undergone revascularisation (92%). These 645 patients were asked if they thought they would have benefited from attendance at an outpatient CR program. Of the 380 patients who did not think they would have benefited, 41% (157) reported having at least three coronary risk factors, 39% (150) were interested in further services, and 26% (100) reported participating in at least one risk-factor-specific secondary-prevention program (Box 1). In conclusion, many patients who are not referred for CR reported having multiple coronary risk factors, yet few felt they would have benefited from attending CR or had participated in any alternative risk-factor-specific programs. We agree that system factors resulting in failure to refer should be investigated and rectified,1 but our data suggest that many non-referred patients would not attend if invited. This highlights the importance of research testing the efficacy of alternative models of CR in the Australian setting,4,5 and the need for research assessing the effectiveness of these programs in routine health services delivery. 1 Coronary risk factors, and opinions on the need for and participation in risk-factor-specific secondary-prevention programs Felt cardiac rehabilitation would have been beneficial* Total Yes No Number of patients 645 143 (22%) 380 (59%) Number of self-reported coronary risk factors None 60 (9%) 8 (6%) 37 (10%) One 137 (21%) 32 (22%) 73 (19%) Two 180 (28%) 45 (31%) 108 (28%) Three or more 254 (39%) 55 (38%) 157 (41%) Felt the need for further services 321 (50%) 119 (83%) 150 (39%) Chose one or more of the following options: Information on how to prevent or manage further heart trouble 280 (43%) 112 (78%) 124 (33%) Help with how to cope with emotional issues arising from heart problems 169 (26%) 82 (57%) 61 (16%) Exercise classes 130 (20%) 76 (53%) 34 (9%) Nutrition classes 124 (19%) 71 (50%) 34 (9%) Quit-smoking programs 51 (8%) 21 (15%) 21 (6%) Help with stress management 142 (22%) 69 (48%) 51 (13%) Help with getting back to work 35 (5%) 25 (17%) 10 (3%) Undertook risk-factor-specific secondary prevention 187 (29%) 50 (35%) 100 (26%) Participated in the following programs: Home exercise plan provided by hospital 62 (10%) 15 (10%) 34 (9%) Other home-based exercise program 55 (9%) 10 (7%) 31 (8%) Fitness-centre program 7 (1%) 2 (1%) 2 (0.5%) Diet/nutrition program provided by hospital 85 (13%) 27 (19%) 46 (12%) Other diet/nutrition program 57 (9%) 14 (10%) 27 (7%) Quit-smoking program 25 (4%) 7 (5%) 15 (4%) * 19% of non-referred respondents (122/645) did not answer this question.

Natalie A Johnson · Kerry J Inder · Amanda L Nagle · John H Wiggers