Letters

To the Editor: The research letter by Hespe and colleagues1 on cardiovascular disease prevention is itself a missed opportunity to illuminate the complexities of person‐focused management of patients in general practice. While it provides a snapshot on cardiovascular disease prevention, it does not offer any exploration of the veracity or otherwise of these findings. Aggregate decontextualised and — as acknowledged — limited data ultimately fail to identify the true nature of the problem. Simply focusing on easily extractable data from computerised medical record systems, without linkage to the unique features and context of the person to whom these data belong, necessarily results in a distorted picture. Big data has the potential to inform only if it is appropriately interpreted and may be useful in process monitoring. Such data have a limited role in assessing general practitioner performance and outcomes of care.2 One must always remember that guidelines are nothing more than guides, which must be appropriately adapted to the unique circumstances of each patient. A more relevant research question would be: how appropriately or inappropriately are preventive treatments applied? This question addresses both overtreatment and undertreatment.3 As a binary question, however, it fails to ask more important contextual questions such as whether the patient can cope with the demands of the treatment, whether the treatment decision is a truly informed one, and whether it fits the needs and expectations of the patient given other health concerns. It ignores entirely the impact of a therapeutic alliance on actual health outcomes.4 Finally, the implied need for a hawkish attitude to prevention must take account of the fact that no intervention is risk‐free. Our obligation clearly states: primum non nocere. Research aiming to improve understanding of the interdependencies inherent in each and every consultation is urgently needed. The health and wellbeing outcomes of medical care are far less determined by biomedical interventions than by the contextual stressors in a person’s life.5 Providing general practice with the tools and resources to truly address the complexities of our patients’ needs is of utmost urgency.

Joachim P Sturmberg · Carmel M Martin

To the Editor: The incidence of syphilis is dramatically rising in Australia. As such, previously rare sequelae like syphilitic hepatitis are occurring more frequently, supported by a growing number of case reports in the literature.1 We present a typical case of syphilitic hepatitis and review the evolving at‐risk populations, to raise awareness of this potentially fatal yet highly treatable disease. A 34‐year‐old Caucasian man presented with a maculopapular rash over the trunk and limbs associated with abdominal discomfort, anorexia and fatigue. He had been treated for early syphilis 2 years previously, with serological evidence of response. He had no history of human immunodeficiency virus infection. He was married with children and denied other sexual relationships. Aside from the aforementioned rash, his examination was unremarkable. His alkaline phosphatase level was 684 U/L (reference interval [RI], 50–130 U/L), γ‐glutamyl transpeptidase was 942 U/L (RI, < 55 U/L), alanine aminotransferase was 429 U/L (RI, < 45 U/L), and aspartate aminotransferase was 181 U/L (RI, 5–35 U/L); bilirubin was 18 μmol/L (RI, < 20 μmol/L) and C‐reactive protein (CRP) was 55 mg/L (RI, < 5 mg/L). Abdominal ultrasound and extensive liver screen results were normal. Treponema pallidum particle agglutination assay and rapid plasma reagin test results were reactive at a 1:32 titre, prompting a diagnosis of syphilitic hepatitis. He was administered 2.4 million units of benzathine benzylpenicillin intramuscularly once‐weekly for 3 weeks. His rash, symptoms and liver function tests resolved within 6 weeks. The Australian notification rate of syphilis increased from 5.0 to 18.3 per 100 000 population between 2010 and 2017. Women aged 15–19 years experienced a tenfold increase in incidence over this period. In 2017, women from remote areas were 27.3 times more likely to contract syphilis than those from major cities, while Aboriginal and Torres Strait Islander people were 6.6 times more likely to contract syphilis than non‐Indigenous Australians.2 Despite this disproportionate rise in vulnerable populations, rates remain highest among men who have sex with men and those with human immunodeficiency virus infection.3,4 Re‐infection is common, as was found in our patient. Described as the “great imitator”, secondary syphilis can be difficult to diagnose; consequently, the true incidence of syphilitic hepatitis is not known. While our case illustrates a typical presentation, cases of fulminant liver failure have been described.5 Rash (78%), anorexia (57%) and fatigue (57%) are the most common presenting symptoms.1 Marked elevation of alkaline phosphatase and γ‐glutamyl transpeptidase, coupled with milder elevation of alanine aminotransferase and aspartate aminotransferase are the most common laboratory findings, with hyperbilirubinaemia present only in severe cases. A liver biopsy is not essential for diagnosis, but may reveal inflammatory infiltration of the bile duct, hepatic granulomas or, less commonly, intrahepatic spirochetes via immunohistochemical staining.1 It is imperative that clinicians consider ordering treponemal serology in high risk patients fitting this presentation, as prompt treatment with penicillin leads to rapid disease resolution and aversion of tertiary complications, including death.

Matthew Smale · William R Connell · Julien D Schulberg

To the Editor: The explanation of inference from confidence intervals by Hemming and Taljaard is interesting but unfortunately incorrect.1 The authors may have fallen for the confidence interval variation of the P value fallacy — the mistaken idea that the P value (or confidence interval) can capture both the long term outcomes of an experiment, as commonly reflected in the phrase “a trend to significance (P = 0.06)”, and the evidential meaning of a single result.2 In a frequentist approach, the P value follows from the null hypothesis, which is either accepted or rejected. The calculation of the P value proceeds only because we have accepted the null hypothesis to be true. Are Hemming and Taljaard confusing Bayesian and frequentist inferential methods?3 The difference between Bayesian and frequentist logic is analogous to the diagnosis of measles for a hypothetical patient presenting with a fever and a rash.4 With frequentist logic, we would consult a text book (the correct textbook being a key assumption), and base our diagnostic inference on a hypothetical cohort of 100 patients presenting to us with an identical rash and fever, to state that 95 of them would have measles. We would not be able to state which of this hypothetical group of individuals had measles. Moreover, a diagnosis of “a trend to measles (P = 0.06)” does not exist in the real world. By contrast, with Bayesian logic, our hunch (the prior) that the patient in front of us has measles is firmed up (the posterior) by knowing that there is a measles outbreak in the community (the evidence). Unfortunately, the thinking commonly found in association with P values and 95% confidence intervals, and suggestions that directive conclusions from randomised trials are achievable from borderline P values, leads to terms such as “a trend to significance” for findings from studies that are underpowered.5

James C Hurley

To the Editor: A 35‐year‐old man born in Dire Dawa, Ethiopia, with childhood exposure to swimming in rivers, migrated to Australia 18 years ago. He presented with recurrent gastrointestinal bleeding. His index gastroscopy revealed portal hypertensive gastropathy and large oesophageal varices with high risk stigmata of recent bleeding requiring banding. Abdominal ultrasonography and transient elastography excluded liver cirrhosis. Six weeks later, he re‐presented with recurrent severe haematemesis associated with dark maroon rectal bleeding. Repeat urgent gastroscopy and flexi‐sigmoidoscopy revealed oesophageal varices without active bleeding. On sigmoidoscopy, a large amount of dark blood was seen, presumed to be related to rapid transit from recent oesophageal variceal bleeding. In the next 48 hours, a repeat colonoscopy was performed. Colonoscopy identified yellowish nodules throughout the colon with diffuse telangiectasia (Box, A and B). Mucosa was oedematous and friable. Aphthous ulcers were seen in the transverse colon and biopsies were obtained. No polyps or focal source of colonic bleeding were evident. Histopathology confirmed the presence of cystic ova resembling Schistosoma in the lamina propria immediately adjacent to crypts. Although classical granuloma formation was absent, aggregation of eosinophils was seen around a ruptured ovum (Box, C and D). Indirect assay via serology testing was positive for Schistosoma mansoni antibody (titre of 1:640). His eosinophil count was normal and stool microscopy was negative. He was treated with praziquantel. Four months later, a follow‐up gastroscopy revealed stable appearance of grade 1 oesophageal varices without high risk features. Routine variceal surveillance had been organised but not further colonoscopies. Schistosomiasis affects over 200 million people worldwide but is not acquired in Australia. In a national survey in Ethiopia, 37.3 million individuals were living in endemic areas.1 Schistosoma mansoni is the commonest species to cause intestinal and hepatic schistosomiasis. In the 2016 national census, there were 11 795 people in Australia who were born in Ethiopia, 64.3% (7584) of whom were Australian citizens.2 Most screening data for schistosomiasis in Australia are based on African refugees, with 37% in Newcastle, 38% in Hobart and 12% in Melbourne.3 Despite anecdotal knowledge, there are no published cases of non‐cirrhotic portal hypertension related to schistosomiasis in Australia. As a multiracial country with high immigration and tourism, increased recognition in Australia is paramount. Box – Colonoscopy showing widespread yellowish nodules (arrows) with oedematous and friable mucosa (A) and diffuse telangiectasia (B); and histopathology showing aggregation of eosinophils surrounding a Schistosoma ovum(C) and a cross‐sectional image of a Schistosoma ovum (D)

Julia Lim · Shweta Sharma · Damian Dowling

To the Editor: We read with interest the article by Kault,1 who carried out an analysis on superspreaders, asymptomatic cases, and coronavirus disease 2019 (COVID‐19) elimination. Although all efforts made for preventing or containing the COVID‐19 pandemic are certainly welcome, we raise doubts on some basic aspects used for constructing the prediction model and which do not seem to be evidence‐based. In the risk model of COVID‐19 re‐emergence after release of restrictive measures (eg, lockdowns), Kault made some erroneous assumptions, including the fact that asymptomatic subjects may be as infectious as symptomatic patients with COVID‐19.1 This hypothesis seems to be contradicted by several lines of evidence. First, a meta‐analysis published in 2020 concluded that the rate of asymptomatic transmission of severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) infection is 35% lower compared with COVID‐19 patients with symptomatic illness.2 This has also been clearly explained in a seminal study showing that the viral load is the highest in concomitance with symptoms onset, so that the infectiousness of pre‐symptomatic or asymptomatic individuals is probably low.3 Notably, the impact of pre‐symptomatic SARS‐CoV‐2 transmission seems also rather limited, whereby the secondary attack rate was found to account for only 15% of all secondary COVID‐19 cases.4 A second aspect that needs to be highlighted is that presuming that 50% of SARS‐CoV‐2‐positive patients are asymptomatic may also be formally incorrect. Beside the fact that the asymptomatic SARS‐CoV‐2‐positive rate varies greatly depending on many genetic, demographic (ie, age, sex and ethnic origin) and even clinical (eg, time course of disease, comorbidities) variables, an analysis in the official database of the Italian National Institute of Health reveals, for example, that the rate of asymptomatic subjects with SARS‐CoV‐2 infection approximates 70%.5 Combined with lower infectiousness, the high prevalence of asymptomatic subjects bearing SARS‐CoV‐2 infection after release of restrictive practices (eg, lifting of lockdowns) would persuade us to conclude that the possible impact of asymptomatic superspreaders on SARS‐CoV‐2 transmission would be low and perhaps insufficient to influence or guide future policies aimed at restricting individual freedom.

Camilla Mattiuzzi · Giuseppe Lippi

To the Editor: Recruitment and retention of a sustainable rural health workforce was one of four issues highlighted by Osborne in a recent MJA supplement.1 Chapter 4 of the Supplement describes a need for longitudinal methods to evaluate recruitment and retention of nursing and allied health professionals, noting challenges around scale and links to policy.2 Comparative efforts examining the medical workforce in Australia are more advanced (eg, the Medical Schools Outcomes Database). Chapter 5 concludes there is a need for a longitudinal, linked database to address rural workforce planning that utilises public data sources, noting medicine was covered by all primary data sources identified, yet only three covered all health professions.3 The Nursing and Allied Health Graduate Outcome Tracking (NAHGOT) study is a research collaboration between the University of Newcastle, Monash University and Deakin University that addresses issues of scale and relevance to national health workforce policy. NAHGOT links Australian Health Practitioner Regulation Agency practice location data (the outcome) with university administrative records (explanatory variables including placement location and duration), and is complemented by the national Student Experience and Graduate Outcomes Surveys. Further complementing the NAHGOT database are the publicly available Socio‐Economic Indexes for Areas at varying spatial resolutions (eg, Statistical Areas Level 1 to Level 4), a general measure of spatial access (Australian Statistical Geography Standard — Remoteness Areas), and a workforce specific model used to define Distribution Priority Areas (Modified Monash Model). Across the nursing and allied health disciplines offered by the three universities, data from a new cohort of about 2000 first year students are captured annually and securely stored in a central, de‐identified repository. Database access is currently restricted to participating institutions, with the longer term intent to permit external data extractions based upon predefined study protocols through a formal process. To our knowledge, NAHGOT is the largest tracking study of this type and growing, with three other universities soon to join the collaboration, enabling expansion into Queensland and South Australia. All participating universities are funded by the Rural Health Multidisciplinary Training Program, with the study design reflecting the objectives of the federal Department of Health. Universities are the logical choice to undertake tracking of graduate outcomes at scale. Unlike other data sources, universities hold admission and professional placement data not available elsewhere. The protocol will soon be available4 and the first peer reviewed publications from NAHGOT have now been published,5,6 and as the project expands it is anticipated it will become a major contributor to workforce planning and augment established efforts in medicine.

Vincent L Versace · Tony Smith · Keith Sutton

To the Editor: Victoria introduced voluntary assisted dying in June 2019, historic legislation which aimed to enable terminally ill people in limited circumstances to end their life with autonomy, compassion and support. The Voluntary Assisted Dying Act 20171 is considered the most conservative and safe legislation in the world, with 68 safeguards. We compared recent Victorian reporting data2,3 with a 12‐month period of data on Oregon’s Death with Dignity Act (1997),4 as the jurisdictions have similar population sizes and there is no comparable jurisdiction currently operating in Australia. To access voluntary assisted dying in Victoria a person must: have an incurable and advanced disease, illness or medical condition that is expected to cause death within 6 months (or within 12 months for a neurodegenerative condition); be experiencing suffering, which the person considers intolerable; have decision‐making capacity in relation to voluntary assisted dying; be an adult, aged 18 years or older; and be an Australian citizen or permanent resident who has lived in Victoria for at least the past 12 months. Eligibility requirements for Oregonians are comparable; however, there is no requirement to demonstrate intolerable suffering. Based on a total number of 46 581 deaths in Victoria during the 2019–2020 period, 0.3% of deaths can be attributed to voluntary assisted dying.5 In both Victoria and Oregon, the most frequently reported reasons for requesting voluntary assisted dying and dying with dignity were loss of autonomy and losing control of body functions; additionally, loss of dignity and being unable to engage in activities that make life enjoyable were also cited. Data show that: Victorian applicants were aged between 32 and 100 years, with an average age of 71 years.2,3 In comparison, the average age in Oregon was 74 years, with a range of applicants from 18 to over 85 years.4 In Victoria, 44% of applicants were female, 55% were male and 1% selected “self‐described” as their gender,2,3 comparable to Oregon (41% female and 59% male).4 Between 19 June 2019 and 30 June 2020, 231 permits were issued. Of these, 104 (45%) self‐administered the medication, while another 20 (9%) had the medication administered by a medical practitioner (Box).2,3 Oregon data from 2019 (population, 4.2 million) revealed that 290 applicants were prescribed medication, with 150 (59%) self‐administering;4 in Victoria (population, 6.3 million), 54% self‐administered, while another 61 (21%) did not take the medication and died of other causes. In addition to the first 12 months of Victorian data, we also included the cumulative totals to February 2021 (Box). The Victorian Act requires three requests be made to a registered medical practitioner for an individual to access voluntary assisted dying. Medical practitioners can apply for either a self‐administration or practitioner administration permit at any one time. Time from first to last request occurred within 11 days for 25% of applicants and 19 days for 50% of applicants.2 Voluntary assisted dying and dying with dignity represent an option for individuals to choose the manner and timing of their death during the terminal phase of illness. Continued community awareness and conversations about end of life options are essential to ensuring that voluntary assisted dying is available to eligible Victorians who seek access. Box – Victorian voluntary assisted dying data snapshot4,5 Total Stage Status 19 June 2019–30 June 2020 To February 2021 Eligibility First assessment by coordinating medical practitioner Eligible 341 562 Ineligible 7 19 Second assessment by consulting medical practitioner Eligible 297 483 Ineligible 4 8 Permit applications Self‐administration permit Issued 201 350 Not issued 32 44 Practitioner administration permit Issued 30 55 Not issued 9 16 Withdrawn: case withdrawn from portal by medical practitioner or upon notification of death of the applicant 134 239 Medications dispensed For self‐administration 154 281 Confirmed deaths Medication administered Self‐administered 104 184 Administered by practitioner 20 40

Kate Furness · Donna Markham · Tamica Sturgess · Margaret O’Connor

To the Editor: The Medical Research Future Fund (MRFF) should be a policy triumph for the Australian Government, tackling unmet clinical needs through transformative research that saves lives, creates jobs and strengthens the industry.1 However, to deliver these objectives, medical research findings must be translated from the laboratory and library to achieve impact on the clinic, community and companies. The capacity for efficient health research translation has been limited in Australia.2 The MRFF provided $20 million per year, finishing in 2021, to pump‐prime Australia’s ten health research translation centres designated by the National Health and Medical Research Council, in which we work. The Translation Centres bring health services and consumers together with health researchers from universities and medical research institutes. Distributed across the nation, the Centres have come together to form the Australian Health Research Alliance (AHRA), providing a “go to” destination for those seeking expertise in health research translation.3 We want to see that research‐ and translation‐trained practitioners in medicine, nursing, allied health disciplines, clinical laboratories, pharmacy, health informatics and other frontline services have time stably funded to deploy their skills in a role we describe as “research translators”.4 Alongside conventional health care duties, such staff will have dedicated time to engage consumers, recruit participants to clinical research, partner with industry, prove the relevance of research to their service, promote best evidenced practice, and champion the adoption of innovation. While overseas governments invest heavily in such roles, funding time for frontline clinical staff to deliver research and translation alongside their clinical work, Australia does not. For example, the National Institute of Health Research in the United Kingdom commits approximately $20 per citizen per year, with an impressive impact on lives saved, jobs created and industry invigorated.5 Australia should address this translational workforce gap or risk failure to achieve full beneficial impact from the MRFF. About $65 million per year (about 10% of the MRFF investment income) would support a cadre of research translators broadly proportional to the provision from comparable research budgets overseas, with AHRA’s Centres best placed to provide efficient coordination. Thus, the MRFF would be powered for success by investing in research translators through AHRA, ensuring the triumph of an exciting new policy that promises benefit to all Australians.

John Savill · Christopher Levi · Gary Geelhoed

To the Editor: The broadly accepted definition of a cancer survivor recognises that the survivorship begins at diagnosis.1 However, survivorship care pathways conventionally begin at completion of active treatment, presenting a challenge for addressing survivorship needs at diagnosis and for people living with incurable cancer.2 A revision of the concept of cancer survivorship is needed, placing the survivor at the centre of a dynamic experience of life after a cancer diagnosis and opening up the survivorship experience to persons at any stage of cancer and at any phase of their disease trajectory. Until now, clinical care guidelines and models of survivorship have typically not included consumer input, but rather have been developed principally through health professional expert consensus.3,4 In a novel approach from 2019–2020, a panel of 47 experts and consumers across Australia and New Zealand came together to define six key domains of survivorship care in a Prostate Cancer Survivorship Essentials Framework:5 health promotion and advocacy, shared management, vigilance, personal agency, care coordination, and evidence‐based survivorship interventions. These six domains reached high consensus as being essential, with the 26 elements within domains all rated as high importance. Almost one‐third of the 47‐member panel were cancer survivors working collaboratively with medical, allied health and nursing expert representatives. The degree of consensus in such a broad coalition is remarkable, underscoring the validity of the approach that reflects the lived experience driven by survivors’ preferences. Importantly, the central domain related to personal agency of a survivor as a key element that linked all others (Box) and all domains were framed around outcomes that mattered for the patient (eg, empowerment, information, shared decision making, care coordination, symptom management). While the framework was developed for prostate cancer survivorship, none of the elements were unique to prostate cancer, highlighting the potential relevance of the framework to other cancers. More broadly, this approach aligns with existing models of chronic disease management and frameworks of consumer engagement in care that are fundamental to the delivery of health care in Australia and New Zealand. We believe the essentials framework is applicable to other adult cancer patient cohorts and presents an opportunity to move forward on cancer survivorship in Australia, taking forward a unique consumer–practitioner model where the survivor is not just the passive object of care but an actor in their own health and an empowered and supported agent of change. Box – Prostate Cancer Survivorship Essentials Framework

Jeff Dunn · Bogda Koczwara · Suzanne Chambers

To the Editor: The toll of group A Streptococcus is dramatically unappreciated, despite increasing evidence of its burden.1 In Australia and New Zealand, we recently demonstrated that group A streptococcal throat and skin infections cause a sizable burden at the population level — cellulitis is the main contributor to the total burden of all group A streptococcal diseases and acute rheumatic fever and rheumatic heart disease contribute disproportionately relative to their frequency of occurrence.2,3 At a global level, the burden of group A Streptococcus is not abating. Global Burden of Disease data suggest that incident cases and deaths due to rheumatic heart disease alone have surpassed those of meningitis (Box). In 2019, more than 85% of rheumatic heart disease cases occurred among people aged under 35 years.4 No other group A streptococcal‐specific endpoints are available from the Global Burden of Disease data, yet all‐cause cellulitis was ranked the 24th most frequently occurring condition in high income countries in 2019.4 Group A Streptococcus causes outbreaks of poststreptococcal glomerulonephritis, contributing to the burden of chronic renal disease, and it is estimated to be the fifth most lethal pathogen on the planet, behind the human immunodeficiency virus (HIV), Mycobacterium tuberculosis, Plasmodium falciparum and S. pneumoniae, yet expenditure on vaccine development is only 0.17% of that spent on vaccines for HIV infection, malaria and tuberculosis.5 The divergence in numbers of cases and deaths due to group A Streptococcus compared with meningitis partially demonstrates the value of vaccination. Another major benefit of vaccination is a substantial reduction in antibiotic consumption. Indeed, pharyngitis is a major driver of antibiotic consumption globally, and an estimated 17% of antibiotic prescriptions for pharyngitis among children in the United States could be prevented by a group A Streptococcus vaccine.6 Two major initiatives aim to progress vaccine development. The Australian Strep A Vaccine Initiative (ASAVI) and the Strep A Vaccine Global Consortium (SAVAC) are addressing technical and investment barriers and leading at least one of the current vaccine candidates to an efficacy trial for pharyngitis prevention by 2024.5 An effective vaccine may prevent health and economic burdens due to the full range of group A streptococcal diseases and associated antibiotic consumption. Box – Estimated number of new cases (left) and deaths (right) due to meningitis and rheumatic heart disease globally* * Data obtained from the Global Burden of Disease study 2019.4

Jeffrey W Cannon · Julie Bennett · Michael G Baker · Jonathan R Carapetis

To the Editor: In their article, Blakely and colleagues1 describe an infectious disease model for simulating the effect of a lockdown on the transmission of severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2). Although we cannot say this work determined pandemic policy, two of the authors have described their close collaboration with the Victorian Government, culminating in the release of a road map to reopening2 based directly on, and released alongside, their modelling.3 The model is stochastic and agent‐based, with 2500 individuals moving around a model space. When both an infected and a susceptible person land on the same patch, there is a probability of transmission. Some individuals are marked as being essential workers; population homogeneity is otherwise assumed.4 Models are necessarily abstractions from reality; it is neither possible nor relevant to include every population group. The question is whether the model effectively captures the dynamics of infection. The combination of model type and population structure has a surprising result. People in the model can only be infected by moving around, and a lockdown is simulated by a reduction in the pace and frequency of movement. At a technical level, the model’s mechanics guarantee the effectiveness of a population‐wide lockdown because it most extensively reduces movement. It is hardly surprising that Blakely and colleagues refer to a lockdown as an “opportunity”.1 The assumption of population homogeneity is robust to exceptions, but only to a point. Using official data, we estimate that, in Victoria, the odds of an aged care worker becoming infected were almost 12 times that of the general population (odds ratio [OR], 11.81; 95% CI, 11.76–11.87). For health care workers, the odds were more than three times higher (OR, 3.19; 95% CI, 3.14–3.23).5 At this level of contact and risk heterogeneity, the model cannot reflect the true virus dynamics. Throughout the period covered by the model predictions, interventions targeted at health care settings were implemented. These interventions, such as closing hospital tea rooms and changing aged care working conditions, cannot be factored into the model predictions because health and aged care workers are not included in the model. By failing to specifically consider the populations that drove the epidemic or the interventions targeted at those populations, any ultimate concurrence between the actual and predicted numbers can only be attributable to chance.

Bradley R Crammond · Vishaal Kishore

To the Editor: We read the recent article from Haines and colleagues1 with interest and noted that it is featured on the front cover of the print version of the Journal. We agree that the Bupa model for provision of general practitioner services to residents of aged care facilities has promise, as illustrated by this study. However, we wish to point out that, despite these promising findings, this Bupa model has been terminated. The arrangement at the time of the Haines study (2012–2014), whereby GPs were employed by the company as part of a broader care model to provide medical services to residents of Bupa aged care facilities, no longer exists — as known among GPs and in the industry. That care model was reviewed and Bupa GPs were taken off salary when it was observed that GP Medicare billings did not match or exceeded their income. They were advised that they could continue only as contractors, with financial reimbursement being made through their Medicare billings alone. A handful of GPs apparently have remained on salary and they are those who are able to ensure that their billings cover their income. Most GPs chose to sever their ties once the focus of the company shifted. Some remain as contractors, with the acknowledgement that their relationship within the care home has changed. The responsibilities of education, meeting attendance and audits, which are not eligible for a Medicare rebate, are no longer performed. The termination of Bupa GPs as salaried employees coincided with other changes within the care staff structure. We interpret this to mean that promising models of practice that may improve the health of people living in residential aged care facilities can be compromised by corporate decision making that has motivations in addition to, or even at variance with, provision of good health care. Our interpretation is based on the contamination‐adjusted intention‐to‐treat analyses presented by Haines and colleagues.

Ian D Cameron · Helen Steinke · Susan E Kurrle

To the Editor: In a recent MJA article, Dyer and colleagues1 stated, “Australia provides institutional long term care for almost 20% of the population aged ≥ 80 years, and 6% of those aged ≥ 65 years. This places Australia as the nation with the highest proportion of older people living in institutional care compared with 11 other nations”. However, the Australian numbers quoted are not comparable to the other 31 countries in the Organisation for Economic Co‐operation and Development (OECD) database.2 The “almost 20% of the population aged ≥ 80” refers to all persons in residential aged care at some time over the course of a year, whereas other countries mainly report point‐in‐time data. Only two countries (Australia and Greece) submitted data relating to residents over an entire one‐year period. In 2018, the Australian point‐in‐time rate for people aged ≥ 80 years was 13.9%.3 Point‐in‐time or census data are the appropriate numerator for calculating age‐specific usage rates for residential care, the standard method used in Australia for national purposes and the most common metric internationally. For people aged ≥ 65 years, the point‐in‐time rate was 4.5%4 rather than the 6.0% reported by Dyer et al. This is the difference between the 174 875 residents aged ≥ 65 years living in residential aged care at 30 June 2018 and the 234 617 who had lived in residential aged care during the 2017–18 financial year. For residents aged ≥ 80 years, the comparable figures were 137 537 (point‐in‐time) and 186 952 (the financial year).3 The Box presents statistics from the OECD countries used by Dyer and colleagues but incorporating point‐in‐time data for Australia.3,4,5 For persons aged ≥ 65 years, Switzerland has the highest rate at 5.6%. Seven countries then pack closely in between Australia (4.5%) and Denmark (3.9%). Given OECD data have inevitable limitations for comparative purposes, based on differences in national systems and reporting, these seven countries have similar levels of use. For people aged ≥ 80 years, rates are highest in Switzerland (16%), New Zealand (14%) and Australia (13.9%). Four additional countries sit close behind: the Netherlands, Sweden, Denmark and Canada, all in the 12.3–12.7% range. Nationally, the accurate 2018 usage rates for Australia are 4.5% for people aged ≥ 65 years and 13.9% for those aged ≥ 80 years. Internationally, Australia is a relatively high provider of residential aged care, but there are a number of countries with similar or higher levels of provision. Although there is insufficient evidence to claim that Australia is over‐reliant on residential care, the authors’ argument for increased investment in community‐based care is nonetheless an important one. Box – People aged ≥ 65 and ≥ 80 years using long term care as a percentage of total age group — international comparison of selected Organisation for Economic Co‐operation and Development (OECD) countries (2018) Percentage of people aged ≥ 65 years Percentage of people aged ≥ 80 years Switzerland 5.6% 16.0% New Zealand 4.3% 14.0% Australia 4.5% 13.9% The Netherlands 4.2% 12.7% Denmark 3.9% 12.7% Sweden 4.3% 12.6% Canada 4.0% 12.3% Germany 4.1% 10.7% Korea 2.7% 9.0% Japan 2.6% 7.2% United States 2.4% 6.1% Poland 0.8% 1.6% Data sources: For Australia, point‐in‐time usage data were extracted from the Australian Institute of Health and Welfare GEN Aged Care Data;3 for all other countries, data were extracted from OECD Statistics.5 Data are for 2018, or the most recent OECD data available in the case of Denmark (2014), the Netherlands (2017) and the US (2016). OECD data definitions indicate the point in time for Canada, Germany, the Netherlands, Poland, Switzerland and the US, a one‐month reference period for New Zealand and Japan, and are inadequately specified for Denmark, Sweden and Korea.2

Diane M Gibson

To the Editor: We are two international medical graduates who have been inducting, assessing and mentoring international medical graduates for many years. We found the article by Yeoman and colleagues1 very interesting and agree with their conclusions. However, one of the most important innovations by the Australian Medical Council (AMC) — the introduction of the workplace‐based assessment (WBA) program2 — was not mentioned. The current clinical examination by the AMC tests the clinical competency of the candidates. What is needed is an assessment of performance, as it is well known that many international medical graduates struggle in the workplace after passing the clinical examination. The AMC introduced the WBA program and is now conducting this assessment in nine accredited sites after it was pioneered in Newcastle2 in 2010. These sites are accredited by the AMC and the program replaces the AMC examination. The WBA programs use various tools, including mini clinical assessments, case‐based discussions, multisource feedbacks, and directly observed procedural skills. All these tools are well validated. The assessment duration varies from 6 to 12 months. The variability in the blueprint of the assessments and the duration of the programs provide the opportunity to test their reliability. The results are reviewed by the AMC before candidates are awarded the AMC certificate. The lessons learned from the WBA program can be and are being used in undergraduate and postgraduate settings. The AMC is to be congratulated for introducing this innovation in medical education and assessment.

Balakrishnan (Kichu) R Nair · Mulavana Parvathy

To the Editor: Sexual incidents involving patients in hospitals are prevalent and are distressing for patients and staff alike, but they are poorly managed.1 Such incidents are frequently perpetrated by people with acute mental illness (eg, mania, psychosis), substance misuse, personality vulnerabilities, and cognitive impairment (eg, delirium, dementia). Sexual incidents span the age range, although lack of staff's understanding of sexuality and sexual behaviour in older adults2 mandates practice improvement. Sexual safety is defined in health care settings as “recognition, maintenance and mutual respect of the physical [including sexual], psychological, emotional and spiritual boundaries between people”.3 Sexual incidents include sexual offences (criminal offences such as sexual assault) and inappropriate sexual behaviour, ranging from suggestive language to removing clothing, exposure, and public masturbation. State public mental health services have sexual safety policies and guidelines, but no such policies exist for general hospitals. The scholarly literature is similarly limited. Our experience of sexual safety in health care can be described as “a neglected area, there is no training, no one is talking about it and there seems little interest until something happens,” particularly involving older people.1 A hypothetical case based on clinical experience (Box) demonstrates the trivialisation and minimisation of sexual incidents in health care and the associated risks and ramifications. This phenomenon is aligned with the “dirty old man” stereotype, a reflection of ageist societal attitudes towards sex in older people as inappropriate, shameful or funny.4 Consequences of these attitudes for clinical management include poor documentation and communication, inconsistent responses to patient behaviour, and multiple victims, particularly staff. Sexual harassment is both under‐reported by staff5 and dealt with superficially, despite having significant effects on physical and psychological health and burnout.6 These issues are compounded in aged care, where staff ignore or minimise the impact of sexual behaviour, which is often excused by cognitive impairment4 or dismissed as harmless due to frailty or advanced age.5 Aged care staff may feel shame, guilt, confusion and even responsibility for causing the behaviour.5 Conversely, a tension exists between a cognitively impaired person’s right to sexual expression and the health care worker’s right to a safe workplace. Education of health care staff and development of guidelines with response pathways are needed to ensure sexual behaviour is understood and dealt with consistently and respectfully.4 In this situation, ageism trivialises risk and harm, which may have an adverse impact on a safe health care environment. Box – Hypothetical case based on clinical experience describing the trivialisation and minimisation of sexual incidents in health care and the associated risks and ramifications Warren is a 73‐year‐old single man treated in a geriatric ward in a general hospital for delirium. He has a history of cognitive impairment due to long term alcohol misuse and has a guardian for medical decision making. The delirium is slow to resolve and after a few weeks he starts propositioning nursing staff for sex and making sexual comments. Warren often has an erection when nurses assist with personal care and he masturbates on his bed, visible to patients, visitors and staff. He is moved to a single room and a minimum of two staff are present for any clinical interaction. Three female staff members are grabbed on their breasts while providing Warren with care. This is inconsistently documented in his medical record. His sexual behaviour is discussed with some derision in staff handovers. There is no disclosure of his behaviour to non‐ward staff involved in his care (eg, consulting teams), among whom further assaults occur. A security guard is stationed at his door to keep him in his room. He is commenced on a specific serotonin reuptake inhibitor with the aim of reducing libido. Warren continues to be sexually disinhibited in behaviour and comments, a barrier to securing a nursing home placement. General principles: Management must include behaviour assessment and addressing underlying issues (eg, delirium, unmet sexual needs) with practical environmental and governance measures such as formal handover of behaviour between shifts and for consulting teams and clinicians, clear and easily accessible management plans, and support for staff, visitors and other patients. Consider the use of sexual harassment measurement tools for reporting and awareness raising.

Anne PF Wand · Carmelle Peisah

To the Editor: Hoffman and colleagues’1 excellent review highlights the tightrope women walk when pursuing both a career and children. It is clear that, internationally, inflexible workplace policies as well as more insidious outdated attitudes towards working women cause significant damage to society as a whole. While mothers evidently bear the brunt, such discrimination also negatively affects fathers and we would argue that the real issue is “parenthood and medicine” rather than just “motherhood”. A culture that presents balancing a career and children as being a goal that only women should struggle with is itself part of the underlying problem. In recent generations, the traditional expectation for women to assume the role of primary childcarer has undergone a paradigm shift. It is increasingly commonplace for men to take paternity leave and even to return to part‐time work in order to achieve greater presence in their family life. A growing number of same‐sex couples are also choosing to have a family and face additional, unique challenges. Unfortunately, public policy is lagging far behind; in Australia, partners are entitled to only 2 weeks of paid leave after the birth of a child.2 Better access to spousal leave exists overseas, particularly in Scandinavia; for example, in Sweden, parents can share up to 480 days off work after the birth or adoption of a child, divided as however suits the individuals.3 In the United Kingdom, additional paternity leave of up to 50 weeks is available after the standard 2 weeks.4 A study from 2013 found uptake of the additional leave had been disappointingly low, highlighting poor awareness, practical deterrents, concerns around impact on finances and training, and fear of negative perceptions as key underlying reasons.5 With the majority of doctors in Australia having children at some point in their careers, it is time for further research, education and policy change to support all individuals.

Elizabeth Wootton · Gerard Forrest