Letters

To the Editor: I would like to report the results of a recent survey showing that junior doctors who undertook undergraduate training in a rural clinical school (RCS) felt better prepared for their new role as a doctor than those who did not. In Australia, the potential threat to the quality of clinical training because of increasing numbers of medical graduates justifies an interest in junior doctors’ perceptions of their learning environment and their level of preparedness for hospital work. Factors affecting this potential threat were investigated. In 2008, 428 junior doctors employed in Queensland training hospitals were asked to complete an online survey. Ethics approval was obtained from the Behavioural and Social Science Ethical Review Committee of the University of Queensland. During their undergraduate training, the majority of respondents had not studied at an RCS. The remaining respondents had attended a large university RCS (of about 1600 enrolled students) or a small university RCS (of about 650 enrolled students) in Queensland, or had attended an RCS of unknown size in another state (n = 9). A mixed methods cross-sectional design, based on two validated scales, was used: The Preparation for Hospital Practice Questionnaire (PHPQ) assessed respondents’ perceptions of the adequacy of undergraduate medical training to prepare them for hospital work;1 and The Postgraduate Hospital Educational Environment Measure (PHEEM) measured respondents’ hospital clinical teaching and learning environment.2 Of 167 responses received (a 39% response rate), 33 were incomplete, leaving 134 usable responses. The majority of respondents were women (60%), aged 26–29 years (53%) and graduates of the large university (79%). Forty-eight per cent had chosen to spend a portion of their undergraduate training at an RCS. A univariate analysis (using analysis of variance and analysis of covariance) made comparisons by age, sex, undergraduate university medical program (large or small Queensland university, university in another state) and attendance at an RCS. Overall responses to the PHEEM indicated that there was “room for improvement” in the autonomy, teaching and social support of the hospital learning environment. No significant differences were detected in either the PHEEM or PHPQ subscale scores by age, sex, postgraduate year or undergraduate university. Junior doctors who attended any RCS as part of their undergraduate education scored significantly higher on three out of eight scales of preparedness for hospital work (based on the PHPQ) (Box). A caveat is that confidence and feelings of preparedness do not equate to “competence”, but the study was not designed to assess competence. Confidence is also strongly related to temperament, which could influence self-selection choices such as factors associated with the size of medical school. These findings are congruent with other studies showing that students perceive undergraduate training outside a traditional metropolitan-based program to be of higher quality3 and that the performance of students trained in RCSs is comparable to that of their urban-based peers.4,5 Dispersing smaller cohorts of students to individual clinical schools (similar to the RCS model) may prove a successful strategy to cope with escalating numbers of undergraduate students. In theory, better prepared junior doctors should require less intensive supervision, contribute more to the hospital workload and alleviate the strain on already stretched educational resources in hospitals. In summary, given the imminent increase in medical graduates entering training hospitals, considerable work needs to be done to improve the quality of education provided for junior doctors in order to maintain a high-quality, supportive educational atmosphere. Comparison of Preparation for Hospital Practice Questionnaire (PHPQ) mean subscale scores between Queensland junior doctors who attended a rural clinical school (RCS) as part of their medical training and those who did not* Medical school attended Interpersonal skills Confidence Collaboration Management Science Prevention Holistic care Self-directed learning Any RCS (n = 65)† 3.97 (0.920) 4.00‡ (0.883) 4.28§ (0.735) 4.04§ (0.889) 4.08 (0.838) 4.56 (0.717) 4.56 (0.809) 4.44 (0.643) Large university RCS (n = 42) 3.84 (0.925) 3.82 (0.877) 4.15 (0.772) 3.74 (0.875) 3.95 (0.882) 4.46 (0.737) 4.43 (0.817) 4.36 (0.666) Small university RCS (n = 14) 4.28 (0.897) 4.68‡ (0.770) 4.68§ (0.583) 4.73‡ (0.691) 4.46 (0.692) 5.00§ (0.418) 5.01§ (0.558) 4.71 (0.641) No RCS (n = 69)¶ 3.87 (0.905) 3.58 (0.784) 3.98 (0.853) 3.69 (0.858) 3.88 (0.918) 4.55 (0.726) 4.47 (0.767) 4.43 (0.694) * Figures are mean score (SD). Rating scale of respondents’ perceptions of their preparedness for hospital work: 1 (very inadequately prepared) to 6 (very adequately prepared). † All junior doctors who attended an RCS during their undergraduate degree, including nine non-Queensland RCSs. ‡ Significant at P < 0.001 level. § Significant at P < 0.05 level. ¶ All junior doctors who did not attend an RCS during their undergraduate degree.

Diann S Eley

To the Editor: The general practitioners’ dilemma is acute. Women asking for mammography and/or testing for ovarian cancer; men asking for prostate-specific antigen (PSA) testing for prostate cancer. The debate, fuelled by uncertain knowledge, rages in the general medical journals,1,2 including contributions in the 17 August 2009 issue of the MJA.3,4 The PSA test is, of course, intimately associated with the name of its inventor, Thomas A Stamey. I am, therefore, somewhat surprised that neither of these recent MJA articles mentioned Stamey’s 2004 “recanting” of his proposed value of the test back in 1987: “current evidence from the last 10 years is convincing that the relationship between prostate cancer and serum PSA is tenuous at best, especially with serum PSA less than 10 ng/mL and perhaps even less than 22 ng/mL. This time is not the first we have had second thoughts regarding the usefulness of serum PSA in preoperatively reflecting prostate cancer”. He concludes: “any excuse to biopsy the prostate has an excellent, age dependent chance of being positive”.5 Perhaps the authors could tell your readers why they did not think that Stamey’s 20-year experience of PSA testing and his 2004 conclusions warranted a mention?

Peter C Arnold

In reply: The letter by Arnold highlights the practical dilemma facing busy doctors when communicating with patients about whether to screen for various cancers, prostate cancer in particular. There are no easy answers at present, although presentation of both sides of the argument regarding prostate cancer testing helps doctors and patients make a shared decision.1 We believe the following decision-aid show card, written in plain English, is an excellent resource to assist with this: <http://www.cancer.org.au/File/HealthProfessionals/GPprostateshowcard.pdf >.

David P Smith · Emily Banks · Mark S Clements · Robert A Gardiner · Bruce K Armstrong

To the Editor: In a recent letter by Kowal,1 reference was made to my previous letter2 commenting on an article by Andreasyan and Hoy.3 Kowal questions my suggestion that Andreasyan and Hoy’s work adds evidence to the body of work indicating the health benefits for Indigenous Australians of living in small remote communities. She then suggests that “rather than jumping the gun, we need more research into the effects of outstation life on health and wellbeing”. I think Kowal misses the point. I did not refer to “outstations”. I wrote of the many remote Indigenous communities, most of which have been established in the past 40 or so years, not all of which are called “outstations”. These communities now have their survival threatened by current government policies that support a few selected larger communities at the expense of many smaller communities. The evidence provided by Andreasyan and Hoy, building on other work, does not support policies that are likely to lead to a drift of people away from their homelands to large centralised settlements and the fringes of towns and cities. While the research evidence is not overwhelming, on balance there appears to be more evidence suggesting health benefits from helping people remain on their traditional lands than evidence to support policies that encourage them to migrate to larger settlements and towns. Kowal accuses those who cite this evidence, such as Tom Calma and myself, as jumping the gun. I suggest it is governments that are jumping the gun in imposing policies that are not evidence-based. While there is a need for more research and for careful scrutiny of the research, as undertaken appropriately by Zhao and colleagues,4 the evidence definitely points to concerns that current government policies may have a significant detrimental effect on Indigenous people.

David J Scrimgeour

To the Editor: A timely article by Yelland and colleagues in the 7 September 2009 issue of the Journal1 correctly identifies postnatal depression as a significant public health issue. However, as members of the consortium that developed the 2008 beyondblue National Action Plan for Perinatal Mental Health (NAP),2 we are concerned that the National Perinatal Depression Initiative is being considered by Yelland and colleagues without reference to the NAP, the document on which the federal government based its funding allocation. The NAP recommended not only identifying current depressive symptoms, but also — equally importantly — using a structured method to assess the broader psychosocial risk factors known to affect maternal and infant mental health.2 The NAP also clearly recommended that any psychosocial assessment be accompanied by adequate workforce training and supervision and integrated pathways to care, complemented by community awareness programs. Although Yelland and colleagues1 raise concerns about the potential harm of routine screening for depression in women during pregnancy, none of the participants in a recent Australian study of antenatal screening for depression reported feeling stigmatised, labelled or distressed after using the Edinburgh Postnatal Depression Scale (EPDS).3 Indeed, many reported feeling relieved and supported that additional care was offered. This reinforces the assertion made in the NAP that a well trained workforce is essential for conveying the purpose of routine assessment — that such assessment is not an endpoint nor a substitute for full diagnosis, but the beginning of an ongoing process that helps professionals to be aware of women who may need support and treatment. Multiple studies evaluating the EPDS against structured diagnostic interviews for detecting major depression have shown that the EPDS has very good psychometric properties for scores of 13 or more (indicative of possible depression) in English-speaking populations.4 Yelland and colleagues also express concern that a limited number of interventions will be recommended to help depressed women. In fact, the NAP recommended that the full range of treatment options be offered within an integrated framework of community, primary care and specialist services.2 A first step in a national approach to perinatal mental health is to develop evidence-based clinical practice guidelines for perinatal depression and related disorders. This task is now underway, a clear indication that we are well placed to begin translating knowledge into practice. Research across Australia continues to examine the efficacy of a range of psychosocial risk assessment models,5 barriers to the uptake of referral and treatment options, and the impact of psychosocial assessment on maternal outcomes. This work will help ensure that a national approach to perinatal mental health is embedded in an evidence-based and evaluative framework.

Marie-Paule Austin · Nicole Reilly · Jeannette Milgrom · Bryanne Barnett

To the Editor: The research article by Keogh and colleagues1 on uptake of genetic testing and insurance implications highlights the need to ensure that Australian insurance industry policy in relation to genetic testing does not increase the risk of adverse health outcomes. The authors called upon the insurance industry and the Human Genetics Advisory Committee (HGAC) of the National Health and Medical Research Council (NHMRC) to reconsider the use of genetic information in relation to insurance.1 Another recently published study has also raised concerns about the way genetic information is used in the insurance industry.2 The HGAC has had initial discussions with the insurance industry about: avoiding genetic discrimination in the insurance setting; developing policies and guidelines to ensure appropriate use of genetic test results in insurance underwriting; promoting genetic education and training in the financial industry; and providing support for consumers and health professionals wanting to challenge adverse decisions. Through the HGAC, the NHMRC will encourage the development of an evidence-based process for assessment of genetic tests for use by the insurance industry, to ensure that it meets the required standard for underwriting. This will, in turn, support appropriate, equitable consumer access to optimal health care.

Warwick P Anderson

To the Editor: Congratulations on the 21 September 2009 issue of the Journal, highlighting preventive cardiology in Australia. I believe that the “working class man” of the well known Jimmy Barnes song has trouble affording preventive cardiology care in Australia. Assuming that he is currently employed, he will not be entitled to a Health Care Card and will be paying full price for care in general practice. There will be a gap of $20 to $60 between what he pays and the Medicare rebate per visit. He will have to pay about $30 per prescription. Assuming that he is taking antiplatelet agents, statins, and a blood pressure agent such as an angiotensin-converting enzyme inhibitor, he would be paying about $90 per month for medication. On an average income of $1200 per week1 ($800 after tax), and assuming that he visits his general practitioner monthly, preventive cardiology care might cost him about 4% of his disposable income. If our working class man is earning $900 weekly ($600 after tax), then preventive cardiology care would cost him 5% of his disposable income. Of course, the costs are similar for a wealthy man. However, the costs represent a far smaller percentage of his income than for the working class man. In this regard, the costs of preventive medicine are similar to the poll tax imposed in England in the 14th century. This regressive capitation tax was raised to pay for the imperial wars in France. However, the disproportionate burden it imposed on the serfs resulted in such anger that it led to the Peasants’ Revolt in 1381. Until medicine is funded more equitably, there will be a major disincentive to pursue preventive cardiology care for poorer people, who experience cardiovascular disease at an earlier age than their wealthier counterparts.2 The problem is not which model to use for absolute risk assessment,3 but the money to apply it.

V Michael Jelinek

To the Editor: We describe two liver transplant patients who presented with unexplained dyspnoea and were subsequently found to have unilateral right diaphragmatic palsy, an uncommon complication of orthotopic liver transplantation.1-3 Both transplant recipients were male. One, aged 64 years, had a liver transplant in 2005 for hepatitis C-related chronic liver disease. The other, aged 66 years, had a liver transplant in 2004 for end-stage alcoholic liver cirrhosis. Both patients had presented with exertional dyspnoea several weeks after transplantation. Both were reformed smokers with no prior respiratory symptoms or established respiratory or cardiac condition. Preoperative pulmonary function tests had been essentially normal in both patients (Box 1). Chest x-rays of both patients during the postoperative convalescence period showed unilateral elevation of the right hemidiaphragm compared with the immediate pre-transplant images. (Images for Patient 1 are shown in Box 2.) Fluoroscopic study (the “sniff test”) and a computed tomography scan of the chest confirmed the presence of right hemidiaphragmatic palsy in both patients. Follow-up chest x-rays and pulmonary function tests over 2 years showed no significant improvement. In both patients, the postoperative clinical course over these 2 years was characterised by recurrent hospital admissions with hypoxaemia and intercurrent respiratory tract infections, some requiring supplemental oxygen therapy, non-invasive positive pressure ventilation and invasive ventilation in the intensive care unit. Currently, one of these patients is well, apart from dyspnoea on moderate exertion. The other patient died from a cause unrelated to his diaphragmatic palsy. Unilateral diaphragmatic palsy following liver transplantation is thought to be related to traumatic crush injury to the right phrenic nerve from a clamp placed on the inferior vena cava (IVC) during surgery.1 The proximity of the phrenic nerve to the IVC renders it very vulnerable to this type of injury from side-to-side cross-clamping of the suprahepatic IVC.1,2 In 2008, we modified our technique to avoid cross-clamping of the IVC by performing cavocavostomy, a type of “piggyback” technique that involves “side-biting” (partial clamping) of the retrohepatic IVC away from the diaphragm.4 Unilateral diaphragmatic palsy can reduce exercise tolerance5 and may place additional mechanical stress on ventilation, which could exacerbate hypoxaemia if these patients develop intercurrent pulmonary infections. Reporting on a series of patients with phrenic nerve injury after liver transplantation, McAlister and colleagues1 found right hemidiaphragmatic palsy in 38% of patients after transplantation, but most of the patients recovered their diaphragmatic function within 9 months. In contrast, our patients did not show any signs of recovery for over 2 years, indicating that loss of diaphragmatic function after liver transplantation may be longstanding or permanent. 1 Comparison of pulmonary function tests before and after orthotopic liver transplantation* Patient 1 Patient 2 Pulmonary function test Before transplant After transplant Difference Before transplant After transplant Difference FEV1 (% of predicted) 2.72 L (80%) 1.73 L (51%) – 29% 2.78 L (85%) 1.40 L (47%) – 38% FVC (% of predicted) 3.79 L (87%) 2.41 L (56%) – 31% 3.83 L (92%) 2.17 L (57%) – 35% FEV1/FVC 0.72 0.72 0.73 0.65 TLC (% of predicted) 5.81 L (87%) 4.28 L (62%) – 25% 7.09 L (110%) 5.18 L (86%) – 24% FEV1 = forced expiratory volume in 1 second. FVC = forced vital capacity. TLC = total lung capacity. * Tests were performed while the patients were in a clinically stable condition. 2 Erect chest x-rays before and after orthotopic liver transplantation, Patient 1 A: In 2005, before liver transplant. B: In 2006, 6 months after liver transplant. Note marked elevation of the right hemidiaphragm and presence of bilateral calcified pleural plaques.

H S Subhash · John W C Chen · Libby John · Jeffery J Bowden · Dimitar Sajkov · Peter Frith

To the Editor: There has been considerable debate about whether selective serotonin reuptake inhibitors (SSRIs) can induce suicidal thoughts and behaviour. Using data from the 2007 National Survey of Mental Health and Wellbeing (NSMHWB),1,2 we examined the relationship between SSRI use and suicidality. The NSMHWB was a nationally representative household survey of 8841 individuals aged 16–85 years. Respondents were interviewed face-to-face and they provided information to assess whether they met International Classification of Diseases (10th revision) criteria for a lifetime affective disorder, and had symptoms in the previous year; had experienced suicidality in the previous year; and had used SSRIs (and/or other psychotropic medications) in the previous fortnight (and, if so, whether they had been taking them for < 1, 1–2, 3–5 or > 5 months). We restricted our analyses to the 555 individuals with symptoms of an affective disorder in the previous year, and examined their suicidality over that year. In our first analysis, we compared those who had used SSRIs in the previous 2 weeks and had been taking them for any duration (n = 109) with those who had not used SSRIs in the previous 2 weeks (n = 446 [respondents in this group were not asked whether they had taken SSRIs at any other time]). Secondly, we compared the subgroup who had been taking SSRIs for more than 5 months (n = 80) with the same non-user group we used in the first analysis. In both analyses, SSRI users were no more likely than non-users to have seriously thought about suicide, made a suicide plan or made a suicide attempt (Box). These findings are consistent with two recent systematic reviews of studies of SSRI use and attempted or completed suicide.3,4 Our study results add to their findings because we considered a fuller range of suicidal thoughts and behaviour. Although the reviews concurred with our findings with respect to adults, they found some evidence for SSRI use increasing the risk of suicidality among children and adolescents. The NSMHWB had limitations, including a potential for recall and misclassification bias, and its inability to account for all possible confounders. Importantly, its cross-sectional nature precluded determining whether an individual’s SSRI use preceded or followed his or her suicidality. By restricting our second analysis to SSRI users who had used SSRIs for more than 5 months, we increased the likelihood that SSRI use occurred first, but we could not determine this conclusively. This would have been a problem had we found an association, because we could not infer the direction of causality. However, with no association demonstrated causality becomes a moot point. Our findings support the contention that SSRI use in adults with affective disorders is not associated with suicide risk. Nonetheless, clinical judgement is required in prescribing SSRIs. Suicidality in the previous year among SSRI users and non-users All users* (n = 109) Non-users† (n = 446) χ2 P Long-term users‡ (n = 80) Non-users† (n = 446) χ2 P Seriously thought about suicide 23% 15% 3.20 0.27 20% 15% 0.83 0.58 Planned suicide 9% 5% 1.76 0.31 4% 5% 0.45 0.41 Attempted suicide 8% 3% 5.23 0.10 2% 3% 0.52 0.52 SSRI = selective serotonin reuptake inhibitor. * SSRI use in previous 2 weeks and for any duration. † No SSRI use in previous 2 weeks (use at other times unknown). ‡ SSRI use in previous 2 weeks and for > 5 months.

Jane E Pirkis · Philip M Burgess · Amy K Johnston · Harvey A Whiteford

To the Editor: On behalf of the Australian Medical Council (AMC), I would like to comment on a recent article1 and two letters2,3 in the Journal about the content of the AMC’s Good medical practice: a code of conduct for doctors in Australia (“the Code”)4 and the consultation process used to support its development. The AMC developed the Code on behalf of state and territory medical boards. It was endorsed by the AMC directors and will be recommended to the new Medical Board of Australia, due to be established later this year. The Code was strengthened by a robust consultation process (supported financially by the Australian Government Department of Health and Ageing) and has been supported by members of the medical profession and the community.5,6 Myers2 seems unaware of the make-up of the AMC, which includes nominees of its many stakeholders (including medical boards). He also seems to be unaware that medical boards in Australia (including the Medical Practitioners Board of Victoria) have already issued codes of professional conduct and that the new national Code brings together and builds on the key elements of these existing codes. The core roles of the AMC are the assessment of international medical graduates on behalf of state and territory medical boards and the accreditation of Australian medical schools and medical colleges. It is not the role of the AMC to “evaluate . . . laws and regulations” or to “address the issue of the accountability of regulatory bodies”, as suggested by Myers. Komesaroff and Kerridge1 and Myers2 appear to believe that a code of professional conduct can be legally binding. Although disciplinary tribunals may use the Code as a guide in their task of assessing allegations of unprofessional conduct, the most important role of the Code is to guide doctors about professionally acceptable conduct. Such conduct is based on many elements, but must especially encompass conduct that is consistent with both professional and legal standards developed by the profession and by parliaments, respectively. It would be neither feasible nor useful to develop a code to guide doctors in meeting those standards in their daily work, without reference to the law and to ethical and professional standards. The Code is not designed to enforce particular kinds of outcomes, but does have an educational focus, and we believe it will contribute to informing and enriching practice. Perhaps Komesaroff, Kerridge and Myers would do well to familiarise themselves with the codes of conduct already in place around Australia and to recognise that they have been used for many years by medical boards to assess complaints about doctors’ conduct. The final code is available online.4 It will be interesting to observe how it is used in the years ahead.

Richard A Smallwood

To the Editor: Complementary and alternative medicine (CAM)1 has become increasingly popular, with Australians spending an estimated $1.8 billion on CAM per year.2 We conducted a prospective study to determine the prevalence and patterns of CAM use among Australian patients with osteoporosis and to identify demographic, socioeconomic and disease-specific features that predict its use. A recent Canadian study of 360 patients with osteoporosis showed that 57% used CAM; CAM users tended to be younger than non-users and better educated, but with a lower quality-of-life score for mental health.3 To our knowledge, there have been no similar studies carried out in Australia. We surveyed 202 randomly selected patients of a Sydney tertiary hospital osteoporosis clinic over a 10-month period (March to December 2007). The clinic sees a wide spectrum of people, from healthy postmenopausal women to patients with multiple comorbidities. At their routine clinic visits, participants completed a prospective questionnaire based on previous CAM therapy surveys.3,4 They were asked about use of CAM, reasons for use, cost, perceived benefits, household income, medical insurance status and educational background. Ethics approval for the survey was obtained from the hospital’s ethics review board. The mean age of participants was 68.5 years (SD, 10.9 years); 80% were women; and 56% were born overseas. We analysed responses using firstly a fairly broad “standard” definition of CAM (including all therapies listed in the Box), and then a more stringent definition that excluded hypnosis, multivitamin therapy, tai chi and yoga. CAM use was reported by 104 patients (51%) and 62 patients (31%) based on the standard and stringent definitions, respectively. Comparative popularity of the CAM therapies used by the surveyed patients is shown in the Box. Common reasons for CAM usage were its holistic approach (53%) and the perception that conventional therapy was providing inadequate pain control (29%). Seventy-three per cent of respondents did not consult a physician before starting CAM, and 23% stated that their treating specialists were unaware of their CAM use. Eighty-eight per cent paid for their therapies out of pocket and the mean cost per patient per month was $21. (Given that 2.2 million Australians have an osteoporosis-related condition5 and 51% of these patients use CAM, the estimated annual expenditure on CAM therapies by patients with osteoporosis is $696 million.) CAM users were more likely than non-CAM users to be university-educated (26% v 14%; P = 0.05); to be born in Asia (20% v 13%; P = 0.03); to have a lower lumbar spine bone mineral density T score (mean, − 2.35 SD v − 2.20 SD; P = 0.05); and to make more frequent clinic visits (mean T score, 1.8 v 1.5 visits /12 months; P = 0.03). More than half the respondents (57%) were unable to differentiate between the terms “osteoporosis” and “osteoarthritis”, especially those from a non-English-speaking background (64% compared with 46% of patients from English-speaking backgrounds; P = 0.02) and CAM non-users (57% compared with 41% of CAM users; P = 0.01). The proportion of patients using CAM in our study was comparable to the proportion in a similar Canadian cohort (51% v 57%).3 Patients who resort to CAM for pain relief may be suffering residual fracture pain or coexisting conditions, such as osteoarthritis; patients unable to differentiate between osteoporosis and osteoarthritis may be seeking relief from pain caused by the latter. Given the low rate of patient disclosure of CAM use to physicians in our study and the reasonable chance of interaction between CAM and conventional therapies (17%),3 physicians would be prudent to regularly and clearly discuss concurrent CAM use with patients (using interpreters when necessary) to avoid potentially harmful drug interactions and side effects. At the same time, physicians could recommend appropriate balance-training therapies,6 such as tai chi, which has been shown to improve bone health and prevent falls.7 Proportion of osteoporosis clinic patients who reported using various complementary and alternative medicine (CAM) therapies (n = 104) Therapy No. (%) of patients* Multivitamins 25 (24%) Fish oil 24 (23%) Acupuncture 20 (19%) Tai chi 15 (14%) Glucosamine 14 (13%) Yoga 13 (13%) Chiropractic/osteopathy 12 (12%) Naturopathy 6 (6%) Herbal therapy 5 (5%) Chinese medicine 3 (3%) Aromatherapy 2 (2%) Massage therapy 2 (2%) Homeopathy 1 (1%) Hypnosis 1 (1%) Minerals 1 (1%) Laser therapy 1 (1%) Any CAM (standard definition) † 104 (51%)‡ Any CAM (stringent definition) † 62 (31%)‡ * Percentages total > 100% because patients used multiple therapies. † The standard definition of CAM includes all therapies listed here; the stringent definition excludes hypnosis, multivitamin therapy, tai chi and yoga. ‡ These are percentages of the total sample (N = 202).

Jenson C S Mak · Steven Faux

To the Editor: Robertson and Kerridge1 criticised our article “Patient privacy versus protecting the patient and the health system from harm”2 based on an interpretation that we were advocating notification for all patients with somatisation disorders, but our recommendation for a confidential notification system pertained specifically to factitious disorder. We agree that “ignoring or failing to integrate mental health care in future health planning is to invite a higher burden of morbidity, mortality and cost”.1 Indeed, we do not wish to “[constrain] the patient’s interaction with the health system”, but rather believe that a more complete, accurate and easily available patient history would allow doctors to optimise such patients’ care within the system. Clinically appropriate resource allocation would substitute appropriate psychiatric/psychological and primary care for more costly and inappropriate (potentially harmful) emergency and procedural care, such as numerous cardiac catheterisations. The case described in our article illustrates well the higher burden of morbidity and cost that the patient, and the system, had to endure because of the failure of multiple health providers across a range of acute settings to diagnose and treat the patient’s primary illness.2 Avoidance of truthful disclosure on the part of the patient contributes to this diagnostic failure. An electronic medical record (EMR) notification in this kind of case would allow any given doctor to overcome the otherwise almost insurmountable barriers to collating such a patient’s history, and thus to be aware of, to balance and to manage the factitious disorder diagnosis — a notoriously difficult task.3 Robertson and Kerridge argue there is a lack of evidence for “costly” EMR systems. However, it has been found that “Hospitals with automated notes and records [have] fewer complications, lower mortality rates, and lower costs”.4 A truly private portable EMR should help all patients obtain more appropriate and cost-effective care, by reducing duplication of costly investigations and doctors’ time spent chasing records. Conversely, patients might reasonably abhor a privacy system that inadvertently results in duplicate computed tomography scans — the prior records being “private” and unavailable — when radiation exposure increases the risk of cancer.5 We would be the first to acknowledge the risk of “stigmatisation” and agree with safeguards to mitigate potential consequences, as mentioned in our article.2 However, in the case of this patient and others in a similar situation, we still believe that he, the doctors struggling to provide appropriate care, and the system deserve better, which certainly won’t happen with the status quo.

Dawn E DeWitt · Ravi Bhat · Stephanie Ward

To the Editor: ANZAME (the Australian and New Zealand Association for Medical Education) shares two significant concerns about medical education in Australia that were raised in articles published recently in the Journal.1,2 The first concern is the status of health professional education (HPE) research in Australia. In 2007, the ANZAME Committee of Management investigated the profile of HPE research with the National Health and Medical Research Council (NHMRC) and the Australian Research Council (ARC); both noted they would accept grant applications for such research. At that time, the ARC indicated that they did not have any HPE research expertise among their reviewers. A number of nominations, which included ANZAME members, were made, and the comments of Roberts and Conn1 suggest this has been successful. HPE research was further assisted by an Australian Bureau of Statistics “Field of Research” classification. Professor Brian Jolly was instrumental in achieving this classification code. Fortunately, the code occurs in the education section, as ARC guidelines state they will not fund medical or dental research. At each annual ANZAME conference, educational research is discussed and presented. While Roberts and Conn are correct in stating that other local conferences on education have arisen in recent years,1 this year’s ANZAME conference was the largest ever, and showcased a wide range of research, both completed and in progress. ANZAME also promotes research through seeding grants, by providing particular supports to students who undertake educational research, and by publishing educational research articles in the Association’s journal, Focus on Health Professional Education. The second concern is the need for significant increases in the number of health professional educators to meet the needs of increasing numbers of students and young professionals. Such educators are invariably swamped by the aspects of their work related to program delivery and service requirements, at the expense of research activity. Brooks’ proposal to establish academic centres2 may go some way to overcoming this. However, we need to develop the HPE expertise that will enable new and innovative educational programs to be developed, implemented and evaluated. How are we going to recruit and train HPE academics for these roles? Earlier this year, ANZAME put out a draft position statement to highlight some of the issues and to indicate our support for promoting academic pathways.3 Workshops at the recent ANZAME conference further refined a set of recommendations. It will take vision, time, commitment, and a real increase in funding to ensure we have the trained personnel in place to achieve the required outcomes. HPE in Australia and New Zealand has matured — it now needs support for ongoing research and careful workforce planning to ensure the high quality of health professional educators is maintained.

on behalf of the ANZAME Committee of Management

To the Editor: My compliments on the issue of the Journal on the survival of general practice (20 July 2009). But something was missing — a patient’s perspective. I hear from fellow Sydneysiders that they cannot find a general practitioner offering continuity of care. They tell me that: All recommended GPs have “closed books”. They can’t get an appointment on the day and have to wait until a few days later, unless it’s a “real emergency”. At big, “commercial” practices, they seldom see the same GP again — and must tell their story each time. Their GP (preferably female) is there some weekdays only. They can’t find a GP who does home visits. Neither “their” GP, nor any of the others in the practice, is available after hours — they must ring an emergency doctor with no access to their records. GPs want them out quickly with a prescription or referral. GPs say that, although it’s a simple procedure, it’s better done by a specialist. GPs want to start a “care plan”, even if there’s nothing much wrong. They’re not happy with attention from the nurse — they want to see the doctor. And so it goes. Back in 2006, the Australian Consumers Association, together with advice on choosing a GP, commented: The relationship you have with your doctor can be one of the most crucial in your life ... A positive ongoing relationship with your GP is extremely valuable, making it all the more important to choose the right one.1 Shouldn’t the MJA open its pages — on an issue which matters so much to them — to patients, to air their concerns? As it would be anecdotal, we would need evidence. A recent article in the BMJ surveyed British attempts at harnessing the views of patients.2 The authors’ opening comments were: “There is now a widespread realisation that patients’ views are not optional but essential to achieving high quality care.” Similarly, Australian patients’ expectations about weight management in general practice were discussed in the Journal in 2006.3 Is general practice’s survival not about adapting to meet patients’ requirements for high quality of service? In special issues devoted to the adaptation of general practice to future change, should not bodies like the Australian Consumers Association and some of the patient support groups, such as Arthritis Australia, be asked to contribute? Perhaps the MJA’s discussion is missing the wood for the trees?

Peter C Arnold

To the Editor: Hospitalisation rates for seasonal influenza are highest among young children and people aged over 65 years.1-3 Calculation of laboratory-confirmed infection rates is difficult because influenza testing is not consistently performed. Using diagnostic codes to identify hospitalisations for acute respiratory illness provides one measure of the relative burden of pandemic (H1N1) 2009 (pH1N1) virus infection compared with influenza virus infections in previous influenza seasons. On 26 May 2009, the first pH1N1 virus human infection in New South Wales was identified, and between June and August 2009, pH1N1 was the predominant circulating influenza virus strain (found in 15%–82% of influenza A-positive specimens).4 In July 2009, the Hunter New England Area Health Service (HNEAHS) (population, 866 000) in NSW established population-based surveillance for hospitalisations for acute respiratory illness at all 35 HNE public hospitals admitting patients with respiratory illness. Data on hospitalisations coded with eight International Classification of Diseases (10th revision) codes (Box 1) were analysed by age group and year. Data for 2007 and 2008 (relatively severe and relatively mild influenza seasons, respectively)5 were used for comparison. Monthly hospitalisation rates for acute respiratory illness were calculated for 1 January 2007 to 31 August 2009. Over the period June to August in 2007 and 2008, 1736 and 1267 hospitalisations for acute respiratory illness occurred, respectively, and monthly hospitalisation rates peaked during August in both years (859 [9.7/10 000 persons] in 2007 and 517 [5.8/10 000 persons] in 2008). The highest hospitalisation rates occurred in children aged 0–4 years (113 [20.6/10 000 persons] in 2007 and 37 [8.9/10 000 persons] in 2008) and in people aged ≥ 62 years (276 [16.3/10 000 persons] in 2007 and 212 [12.1/10 000 persons] in 2008). In 2009, over the same 3-month period (June to August), there were 2378 hospitalisations for acute respiratory illness, with monthly hospitalisation rates peaking in July (1029 [11.4/10 000 persons]). Between May 2009 (before pH1N1 circulation) and July 2009, the monthly hospitalisation rate for acute respiratory illness in the HNEAHS increased by 260% (from 385 [4.3/10 000 persons] to 1029 [11.4/10 000 persons]). During the same period, age-specific hospitalisation rates increased more than threefold among children aged 0–4 years (from 31 [5.5/10 000 persons] to 101 [17.9/10 000 persons]) and 5–16 years (from 36 [2.5/10 000 persons] to 140 [9.8/10 000 persons]), and more than sixfold among people aged 17–31 years (from 43 [2.6/10 000 persons] to 281 [16.9/10 000 persons]) (Box 2). In July 2009, age-specific hospitalisation rates among people aged 5–61 years exceeded respective peak age-specific hospitalisation rates for 2007, while hospitalisation rates among children aged 0–4 years and people aged ≥ 62 years remained below 2007 peak rates for their respective age groups. Although our analysis calculated hospitalisation rates for various types of respiratory illness (including infections with non-influenza viruses) and may be affected by changes in health care-seeking behaviour and coding practices, comparison over multiple seasons suggests several unique features of the 2009 influenza season in the HNE region. Compared with 2007 and 2008, the HNEAHS experienced a rapid, early increase in hospitalisations for acute respiratory illness, coinciding with the emergence of pH1N1. Furthermore, a disproportionate burden of hospitalisations occurred among people aged 5–61 years (particularly those aged 17–31 years), with a relatively lower burden among those aged ≥ 62 years. Ongoing surveillance will determine whether these age-specific trends continue in the 2010 influenza season. 1 Emergency department ICD-10 separation codes used for acute respiratory illness hospitalisation surveillance Code Disease J11 Influenza, virus not identified J11.1 Influenza with other respiratory manifestations, virus not identified J11.8 Influenza with other manifestations, virus not identified J12.9 Viral pneumonia, unspecified J15.9 Bacterial pneumonia, unspecified J18.9 Pneumonia, unspecified J22 Unspecified acute lower respiratory infection J96.9 Respiratory failure, unspecified ICD-10 = International Classification of Diseases (10th revision). 2 Monthly hospitalisation rates for acute respiratory illness in the Hunter New England Area Health Service region, May–August 2009, compared with peak monthly rates in 2007 and 2008* * Hospitalisation rates are per 10 000 residents of the Hunter New England Area Health Service region.

Fatimah S Dawood · Craig B Dalton · David N Durrheim · Kirsty G Hope

To the Editor: Bronchiolitis, most often associated with respiratory syncytial virus (RSV), is a major cause of hospitalisation in young children. Those with chronic lung and congenital heart disease (the latter affecting about 192 births annually in Western Australia) are at particularly high risk.2 Immunoprophylaxis with the RSV monoclonal antibody palivizumab is effective in reducing severe RSV-related hospitalisations, and monthly immunoprophylaxis is recommended in high-risk children.2,3 Monthly immunoprophylaxis is costly; therefore, the most cost-effective schedule follows the times of peak RSV activity4 — usually the winter months, May to October. Using the Western Australian Data Linkage System,5 we investigated the seasonality of bronchiolitis hospitalisations (International Classification of Diseases-10 code J21) from 1996 to 2005 as a proxy for RSV-related illness. Data specifically for RSV-related illness were considered unreliable because some children may not have been tested for RSV, test results may not have been documented on hospital discharge notes, or RSV immunofluorescence tests may have given false negative results. Furthermore, RSV codes (B97.4, J12.1, J20.5, J21.0) were not used by hospitals in WA until July 1999. We identified 11 988 hospitalisations for bronchiolitis throughout WA among 245 249 births. Most bronchiolitis admissions (81%) were in children younger than 12 months. In the Perth metropolitan region, there was a clear winter seasonal pattern, with hospitalisations peaking in July. However, in the Kimberley region in northern WA, there was a sustained bimodal seasonality, with a peak in April and second peak in August (Box). Moreover, only 51.5% (469) of bronchiolitis admissions in the Kimberley and 61.5% (444) in the Pilbara–Gascoyne (located in mid-north WA) occurred between May and October, as opposed to 84.3% (6354) in the metropolitan region. These data support an earlier implementation and longer dosing schedule with palivizumab for high-risk children in the Kimberley and Pilbara–Gascoyne than for those in Perth. Our study has some limitations. Not all bronchiolitis hospitalisations may be caused by RSV. However, when we investigated only those hospitalisations with a specific RSV code, the monthly distribution showed a similar pattern. Additionally, timing of RSV activity, and therefore bronchiolitis, may vary from year to year. Although the numbers were too small to allow separate analysis by calendar year, bronchiolitis hospitalisations in the Kimberley showed extended seasons in 8 of the 10 years. Our findings support the need for each jurisdiction to know its seasonal pattern of bronchiolitis and RSV hospitalisations, and to implement recommended palivizumab schedules accordingly. Such use of extended prophylactic regimens may well require its cost-effectiveness to be reconsidered. Our analysis highlights the relevance of population-based data linkage studies to clinical care policy. Monthly distribution of bronchiolitis hospitalisations by region of child’s birth, 1996–2005

Hannah C Moore · Anthony D Keil · Peter C Richmond · Deborah Lehmann

To the Editor: In 2008, considerable publicity was given to the diagnosis and treatment of prostate cancer in two Australian celebrities: retired Australian Football League player Sam Newman and radio announcer Alan Jones. Newman’s cancer was reported on 5–9 March 2008, culminating in a 12-minute item on the high-rating Australian television program 60 Minutes, when Newman said, “Be tested and enjoy life”. Jones’s announcement of his forthcoming surgery was reported on 3–4 July 2008, and his recovery on 16–27 July. We investigated whether this publicity was associated with changes in the rate of prostate specific antigen (PSA) testing. We used Medicare Australia data on claims for Medicare Benefits Schedule item 66655 (which allows “one annual PSA test for an individual without prior prostate disease”) to determine the mean number of PSA tests and age-standardised testing rate between January 2002 and March 2009. We fitted a linear regression model to the log of the age-standardised rates, modelling secular trends using natural splines with two degrees of freedom (splines were piecewise cubic within the data range and otherwise linear), with indicators set for the average effect of the second, third and fourth quarters and an indicator for the second quarter of 2008. We calculated predicted values and prediction intervals (which take into account both uncertainty in the predicted mean and individual variation), excluding observed data for the second quarter of 2008 (Box). The mean monthly number of PSA tests was 72 064 in the 12 months before April 2008, increasing to 107 633 in April 2008, and 90 813 in July 2008. The seasonally adjusted age-standardised rate of PSA tests rose 17% above the secular trend in the second quarter of 2008 (95% CI, 8%–27%), with the observed data point falling outside the prediction interval (Box). This represented an additional 39 000 PSA tests during the second quarter of 2008. These data show that prostate cancer screening increased significantly in the quarter after media coverage of Newman’s prostate cancer diagnosis. Screening seemed to have returned to the expected seasonal rate by the time of Jones’s diagnosis, in the third quarter of 2008. The impact of celebrity cancer diagnoses on population screening has been demonstrated previously, most recently by an increase in breast cancer screening following news of pop singer Kylie Minogue’s breast cancer diagnosis in 2005.2 In a survey of US men, almost a third stated they would be more likely to have a PSA test after hearing celebrities endorse prostate cancer screening.3 In addition, coverage of melanoma by 60 Minutes in 1987 was followed by a 167% increase in melanoma detection in the subsequent 3 months. Whether the additional 39 000 PSA tests performed in the second quarter of 2008 result in a surge in diagnoses of prostate cancer remains to be seen. Despite continuing uncertainty about the benefits of PSA testing, but unequivocal evidence of harms,,6 Australian men’s willingness to be screened for prostate cancer appears to have been boosted by this celebrity endorsement. Age-standardised* prostate-specific antigen testing rates in Australia (assessed quarterly) Prostate-specific antigen (PSA) testing rates showed a consistent seasonal pattern, with the highest rates in the third quarter (July–September) for all years except 2008, when they were highest in the second quarter. * Standardised to Australian 2001 population.

David P Smith · Mark S Clements · Melanie A Wakefield · Simon Chapman

To the Editor: The recent article by Lowthian and colleagues raises some important concerns regarding current management of warfarin therapy in Australia, especially the provision of quality warfarin education.1 While the study focused on the uncertainty surrounding who is responsible for this task, additional barriers to providing warfarin education include limited access to suitable resources and a tendency to rely on a single verbal counselling session and/or the supply of written material, often just before discharge from hospital.2 As Lowthian et al note, it is plausible that warfarin education was provided to the patients interviewed; it may simply have been inadequate to meet their needs or delivered at the wrong time or place. It is not always feasible for health care providers to reliably provide the necessary education in busy health care settings (eg, general practitioners when prescribing warfarin, pharmacists when supplying it). The authors note a potential role for practice nurses in providing this education; we would also like to highlight the role of accredited pharmacists, who can visit patients in their homes and provide targeted medication-related education. An annual Home Medicines Review can be provided for patients taking high-risk medications such as warfarin.3 Under the Fourth Community Pharmacy Agreement Research and Development Program, the Pharmacy Guild of Australia has funded two projects that are attempting to meet the education needs of patients taking warfarin. One project is trialling a comprehensive postdischarge service involving a series of follow-up home visits by trained accredited pharmacists, to provide not only warfarin education but also point-of-care international normalised ratio monitoring, in liaison with the patients’ GPs. The other project is piloting a process whereby community pharmacists can identify potential candidates for patient self-monitoring of warfarin and, with GP collaboration, organise suitable training, with ongoing support from the pharmacist and GP. Patient self-monitoring, although not widely practised in Australia, has been shown to provide the best clinical outcomes for suitable candidates.4 An important component of these two projects is a website, launched in late 2008, containing a range of free, downloadable educational resources for both consumers taking warfarin and health care professionals responsible for its management (http://www.anticoagulation.com.au). Education is crucial in achieving optimal health outcomes for patients taking warfarin, and health care professionals should cooperate to ensure their patients are well educated. This should involve regular reinforcement of consistent messages, and can also involve better use of existing services and educational tools.

Luke R E Bereznicki · Leanne Stafford · Ella C Jeffrey · Gregory M Peterson · Shane L Jackson

In reply: Bereznicki and colleagues have reinforced the importance of patient education in optimising warfarin safety, while highlighting the need for role clarification in its provision. We applaud the Pharmacy Guild of Australia’s sponsorship of research projects trialling different models of service delivery, such as the coordination of postdischarge warfarin home care by pharmacists and support of patient self-monitoring programs, where appropriate. The success of such care models will be enhanced by a collaborative partnership between the patient and all members of the warfarin care team.1 In addition, appraisal of patient suitability for programs such as patient self-monitoring requires systematic and regular evaluation of cognitive function, compliance and health literacy, to reduce the likelihood of adverse events.2-4 Likewise, ongoing education with resources tailored to individual patients’ capabilities will optimise patient safety.5

Judy A Lowthian · Basia O Diug · Sue M Evans · Ellen L Maxwell · Alison M Street · Leon Piterman · John J McNeil

To the Editor: Lowthian and colleagues ask who is responsible for the care of patients treated with warfarin therapy.1 In Melbourne, it seems that this devolves mainly upon the pathology laboratory, whereas in other states, general practitioners manage care for their own patients, in my own experience. Where management of warfarin dosing, based on the international normalised ratio (INR) result, is performed by the laboratory, this is done as a courtesy to GPs. This practice possibly evolved from competition between private laboratories and, as such, may have been intended to induce other pathology requests to the service provider. Larger metropolitan private pathology services may have sufficient pathologists with haematology training to provide warfarin dosing to patients, but this may not always apply in smaller laboratories. In some laboratories, warfarin dosing may be provided to patients by scientific staff. Although warfarin therapy can sometimes be challenging, its difficulties are not usually insurmountable. Problems for pathologists arise because they are expected to give a warfarin dose by telephone to a patient they do not know and have never seen. The patient’s phone number may be not supplied or incomplete, and the call may not be answered. If answered, a brief history of previous INR results, dosages and test intervals may or may not be elicited from a person who may have poor English, who may be the patient, a relative, friend or neighbour; or a message may have to be left. If the message is received, it may not be acted on, depending on memory, understanding or level of compliance. If the message is not received, the patient and the patient’s GP may be unreasonably irate, even though multiple attempts to contact the patient may have been made. It is not hard to understand why some GPs prefer to pass this often frustrating and time-consuming aspect of treatment to pathologists — who may also be pressed for time. Problems are further compounded if warfarin dosing devolves to pathology laboratory scientists, which may be required by their employer. Scientists would be, in effect, performing drug dosing without medical training or medical registration and may be exposing themselves to litigation risk, without carrying medical indemnity. In other Australian states, it seems that GPs do provide treatment and dosing of warfarin for their own patients. In my view, this is the safest and most satisfactory outcome for patients, pathologists and scientists. With the projected advent of point-of-care testing with reliable INR results2 and new anticoagulant alternatives1 that would enable easier management of anticoagulant therapy by GPs, this whole issue could potentially be resolved.

John C Roberts

To the Editor: I appreciated Brukner’s perspective on treatment of venous thromboembolism (VTE),1 but I do feel some additional considerations and clarifications are in order. Quotes from Kearon2 and Kearon and colleagues3 about the duration of therapy are missing an important caveat. The suggestion for longer duration and possibly “indefinite” anticoagulant therapy after a first deep vein thrombosis or pulmonary embolism is in reference to unprovoked VTE. Brukner1 mentions a plane flight from China to Australia 3 days before he experienced a symptomatic pulmonary embolism. According to guidelines issued by the American College of Chest Physicians, such a flight (although I cannot comment with certainty on its duration) is likely to represent a reversible minor risk factor for VTE — “prolonged travel (eg, > 8 h)”.3 Indeed, Kearon2 and the authors of a related research article to which he refers4 make the important distinction between provoked and unprovoked VTE. Kearon comments: For patients with a minor reversible risk factor, the risk of recurrence is about 5% in the first year after stopping anticoagulant therapy. This is considered low enough to justify stopping anticoagulant therapy at the end of 3 months.2 Additional factors and preferences are also important when considering prolonged anticoagulant therapy for individual patients. For instance, I can empathise with the 25-year-old who wants to discontinue warfarin therapy after 6 months following a first unprovoked VTE — another 50 years of warfarin therapy might seem a disheartening burden. And certainly we must take into account other variables that may increase a patient’s risk of bleeding complications while receiving warfarin therapy, such as older age, the presence of other acute or chronic illnesses and the use of other medications, particularly antiplatelet therapy.3 For someone such as Brukner,1 who claims that his anxiety about a recurrence of VTE is reduced by staying on warfarin therapy, I would probably say, “That’s OK by me, as long as you are aware of the risks and implications of provoked and unprovoked VTE”. For other patients with unprovoked VTE who may not want to continue indefinite anticoagulant therapy, there are other management options worth discussing. First, there is the possibility of further risk stratification for VTE recurrence by testing D-dimer levels. This strategy is receiving considerable study and validation as a useful tool for predicting VTE recurrence.5,6 Second, for patients who refuse to continue warfarin therapy but would consider other therapies for reducing their risk of having a recurrence of VTE, I would suggest taking a daily low dose of vitamin E, which has shown some benefit in reducing VTE.7 Finally, the ASPIRE (Aspirin to Prevent Recurrent Venous Thromboembolism) Trial, which is currently underway, should provide a definitive answer as to whether aspirin is a safe and effective option for secondary VTE prevention. If the trial supports the use of aspirin for this purpose, it would offer another VTE risk-reduction option for those who prefer not to take warfarin indefinitely.8

James A Fink

To the Editor: Brukner’s personal perspective in a recent issue of the Journal highlights the inherent difficulty of managing patients taking warfarin.1 Although Brukner advocates life-long warfarin therapy for patients who have experienced a pulmonary embolism, we believe few clinicians would be willing to expose their patients to the increased risk of haemorrhage associated with long-term warfarin therapy, especially if a transient risk factor for venous thromboembolism (VTE) existed (such as frequent air travel, in Brukner’s case). There is an abundance of conjecture regarding the appropriate duration of warfarin therapy for VTE, with guidelines and clinical intuition often differing. But it is possible that, in the future, decisions about the appropriate duration of warfarin therapy may be guided by the use of pharmacogenetics. Pharmacogenetics gives clinicians an insight into how a patient’s genetic profile affects his or her ability to metabolise drugs, thereby allowing doctors to better tailor the dose and duration of the patient’s medications. In the case of warfarin, metabolism of S-warfarin (the more potent enantiomer in the racemic warfarin mixture) occurs via the cytochrome P450 system, specifically enzyme CYP2C9.2 Studies have shown that patients who metabolise warfarin normally are homozygous for the wild-type allele CYP2C9*1, whereas patients with polymorphisms in the CYP2C9 allele have reduced warfarin metabolism and increased risk of bleeding.3 The anticoagulation effect of warfarin actually occurs via inhibition of the C1 subunit of the vitamin K 2,3-epoxide reductase complex (VKORC1).4 Numerous polymorphisms have been identified in VKORC1, and it is speculated that VKORC1 polymorphisms alone may account for a significant proportion of response variability to warfarin.4 Validated algorithms have been developed to combine information on a patient’s CYP2C9 and VKORC1 genotypes with factors such as age and body surface area to identify an appropriate warfarin regimen.5 Although the pharmacogenetic information required to employ this algorithm has not been assessed for cost-effectiveness, it is quite possible that clinicians will be able to use such pharmacogenetic information to their advantage in the future. This would allow doctors to identify with greater precision which patients are likely to benefit from warfarin (and how much is required, for how long), rather than relying on equivocal evidence and clinical intuition alone.

Amit K Verma · Ajai K Verma

To the Editor: In their national analysis, Robson and colleagues found less favourable outcomes among term singleton babies born in public hospitals than in private hospitals.1 Most health services research is non-randomised, and it is unrealistic to expect studies such as this to be as internally valid as a randomised controlled trial.2 As noted by others,3,4 in the absence of randomisation, it is difficult to untangle the myriad differences between mothers, babies, and the care provided in private and public hospitals (ie, the results might be subject to confounding). Robson et al statistically accounted for the potential confounding effect of maternal smoking (although these data were only available for about half the mothers), age, Indigenous status, parity, diabetes, hypertension, rurality and method of birth. We replicated their analysis using the Queensland Perinatal Data Collection (July 2005 – December 2007), which included virtually complete data on maternal smoking. Our analysis of 124 300 term singleton babies gave an adjusted odds ratio (AOR) for perinatal mortality of 2.0 (95% CI, 1.4–2.7), which is similar to that reported by Robson et al. Using more detailed data from the Queensland dataset, we statistically adjusted for other potential confounders — including pre-existing and gestational diabetes, pre-existing and pregnancy-induced hypertension, pre-eclampsia, eclampsia, antepartum haemorrhage, anaemia, depression, urinary tract infection, low birthweight (< 2500 g), socioeconomic status (based on area of usual residence5), alcohol and drug misuse, and artificial reproductive technology — and obtained an AOR of 2.1 (95% CI, 1.5–2.9). We are not implying that adding more and more variables to a statistical model is an appropriate way to account for confounding. Our aim is simply to show that Robson et al’s result is robust to statistical adjustment using the available data; this is not the same as saying the analysis is robust to confounding. We found that the higher perinatal mortality in public hospitals was greater for neonatal deaths (AOR, 3.1; 95% CI, 1.8–5.6) than for stillbirths (AOR, 1.6; 95% CI, 1.0–2.4). Excluding lethal congenital anomalies4 did not materially change the result (AOR, 1.9; 95% CI, 1.3–2.8). We also stratified our analysis by level of hospital — tertiary referral (neonatal intensive care unit), base (special care nursery), and community — and obtained a similar twofold mortality excess in each stratum. Although perinatal mortality is uncommon among term singleton babies (1 in 1000 in private hospitals versus 2 in 1000 in public hospitals), any excess risk should be investigated and the reasons for it understood. The results from Robson et al’s article might be due to confounding, but they should not be dismissed and should be investigated with more detailed clinical data. Even if all the excess risk is due to confounding, explicit confirmation of this would be extremely useful. To this end, the Statewide Maternity and Neonatal Clinical Network in Queensland Health will undertake clinical review and classification of term singleton deaths according to national guidelines6 and collaborate with the Australian Maternity Outcomes Surveillance System (AMOSS)7 to enhance prospective monitoring of late gestation perinatal deaths nationally.

Michael D Coory · Guan T Koh · Vicki Flenady · Maarten Kamp

To the Editor: Robson and colleagues found higher crude odds of perinatal death in public hospitals — an unsurprising finding, given the maternal demographics in the public system.1 However, after taking into account known risk factors for poor perinatal outcome by “adjusting for the potentially confounding variables available in the NPDC [National Perinatal Data Collection]”, the authors found that the odds ratio actually went up — implying that the pregnancies of mothers in the private system are higher risk. This is implausible. There are a number of problems with the analysis. The possibility that potentially important maternal information was not included has been acknowledged.1,2 Of greater concern is the omission of a history of low birthweight from the regression model. Since intrauterine growth retardation is a very strong risk factor for perinatal death,3 and since the rate of low birthweight in the public hospitals was double that in the private hospitals, this omission is puzzling. Furthermore, while the authors’ stated aim was to assess the effect of the private model of obstetrician-led interventional care on perinatal outcomes, they included the method of birth and hospital type as independent explanatory variables. Since the interventional nature of obstetric care is, as acknowledged by the authors, a key component of the model of care provided in a private hospital, adjusting for this variable will lessen the usefulness of the study in assessing the impact of private hospital care on perinatal outcomes. Put another way, in deciding what is best for her baby, a mother who is considering giving birth in the private system cannot cherry-pick only the non-interventional side of the obstetrician-led model of care — she must adopt the entire package. Including method of birth as a separate explanatory variable artificially inflates the apparent “riskiness” of the pregnancies of mothers presenting to private hospitals to give birth. These mothers had a 75% increased incidence of caesarean section; however, most of these procedures would be elective rather than emergency. Since this distinction was not made, the women in private hospitals appear high risk despite having had a caesarean section for a low-risk pregnancy. This will have the effect of making perinatal outcomes in private hospitals appear more favourable than they actually are. As currently analysed, the data are of little value to prospective mums and dads in making the first, and one of the most emotive, of the many decisions they need to make as parents.

Daniel C Chambers

To the Editor: Of the many problems with the study by Robson and colleagues,1 the one that concerns us most is the outcome of perinatal death. The headline that this is twice as high in public hospitals has the potential to scare those who have no choice about where they give birth. Both the public and the medical community have a right to expect that this outcome is rigorously and accurately represented. We do not believe this to be the case. The perinatal death rate is made up of stillbirths and neonatal deaths in the first 28 days of life. Stillbirths will comprise 70% of all perinatal deaths, and more than 90% of stillbirths after 37 weeks’ gestation will occur before the onset of labour.2,3 It cannot be excluded that this relates to model of care but, because less than 10% of these stillbirths will have occurred during delivery, it is most certainly not related to the type of birth hospital or the interventions performed during labour. Most likely, it reflects differences in intrinsic risk that have not been controlled for by Robson et al.1 The neonatal death component of this statistic is probably confounded by ascertainment bias. Most neonatal deaths after 37 weeks’ gestation occur for two reasons: intrapartum hypoxia–ischaemia and severe congenital abnormalities. When the latter are diagnosed antenatally, the mothers will often be transferred to tertiary public hospitals for a second opinion and further management. Babies with undiagnosed lethal abnormalities or severe intrapartum hypoxia will invariably be transferred postnatally to a tertiary public hospital neonatal intensive care unit, where death often occurs after withdrawal of intensive care. The midwives data collection form is filled in at the birth hospital, and the baby is categorised as “discharged”, “transferred” or “died”. The 2006 national perinatal data collection report specifically cautions: Ascertainment of neonatal deaths within 28 days of birth is likely to be incomplete. In particular, deaths occurring among babies transferred to another hospital, readmitted to hospital or dying at home may not be known to midwives who collect these data or staff who compile state and territory data.2 Thus, it is unlikely that the difference in perinatal mortality described by Robson et al1 has anything to do with the birth hospital. To represent it as such in the public arena is irresponsible. Further, for the MJA to misrepresent “perinatal mortality” as “neonatal mortality” in the media release about this article defies belief.

Nick Evans · Girvan Malcolm · Adrienne Gordon