Health services administration

Recent trends in general practice structure may have an impact on future patient care in Australia The organisational structure of general practice in Australia has changed considerably since the late 1990s.1 Prompted by health care system changes, practice cost increases and incentives, originally small‐scaled general practices have merged into larger entities.1 At present, 2% of Australian general practitioners work in solo private practices, while 20% of full‐time GPs and 33% of part‐time GPs are employed in large practices with six or more GPs.1 Practice ownership levels have changed as well. In 2020, 25% of Australian GPs were practice owners, a decline from 35% in 2008.2 Factors such as management responsibilities, financial burdens and a lack of work–life balance seem to discourage practice ownership.3 Concurrently, commercial ownership models emerged, ranging from ownership by other health professionals to corporate groups of publicly listed companies, with shareholders having little or no involvement in patient care or practice management.3 In 2000, there were six corporate groups; by 2020, three remained, which operate over 400 medical centres. Survey data from 2020 indicate that approximately 16% of GPs work in corporate‐owned practices.1 Co‐located services increased as well. In 2016, 81% of practices had pathology services on location versus 49% in 2010; for imaging, these figures are 30% versus 15% and for physiotherapy 50% versus 29% respectively.3,4 Trends towards increasing general practice size and corporatisation have also been reported in other advanced economies. In the United States, about 16% of primary health physicians are employed in corporate‐owned practices.5 Sweden has encouraged for‐profit facilities to increase provider choice and competition, resulting in 40% of general practices being privately owned.6 Corporatisation of general practice in New Zealand is rising as new generation GPs prefer a lower administrative burden.7 While national policies in the United Kingdom have encouraged partnerships and the integration of general practices to improve service quality and facilitate access, most practice groups have remained independent businesses (< 5% corporate‐owned).4 Corporate business models may have clear advantages for GPs, including financial security and a better work–life balance, and they might benefit patients as they offer a range of on‐site medical services and extended opening hours.3 However, concerns have been raised about the consequences of growing practice sizes and corporate‐owned general practices, in particular regarding corporate obligation to deliver revenue to their shareholders who do not participate in any aspect of patient care, putting profits before patient care.8 Moreover, some corporations currently active in Australian general practice operate across international borders and their cultures may be inimical to patient‐focused health care. The aim of this narrative is to synthesise the available information about the impact of changes to size and ownership on the delivery of patient care and the health care system. The review is based on articles and reports found through an extensive search of the (grey) literature, the reference lists of key articles and related articles, and via the “cited by” feature in Google Scholar. Only publications in English, published after 2000 and applicable to the developments in Australia were considered for inclusion. After screening, 14 publications were included (Supporting Information, table 1). Impact of general practice size and ownership model Independent of size and ownership, general practices should deliver high quality services, with commitments to access, continuity and quality of care and optimal use of health expenditure (Box 1). Impact on access to care Initially, most corporate practice models offered bulk‐billing; however, the introduction of the Medicare rebate freeze in 2013 resulted in a change to mixed‐ or private‐billing models.9 The expansion of GP practice consolidation and corporatisation is associated with less closely located competition, which may lead to further reductions in bulk‐billing and an increase in service fees and out‐of‐pocket expenses.10 While there is no direct evidence that larger practices and corporatisation lead to financial barriers to accessing care, access may be affected when multiple smaller practices conglomerate into larger, more centrally located practices, especially for people residing in regional and remote areas where there are already fewer GPs per person.1 The reported long waiting times in corporate medical centres and the shortage of viable transport options to more distant general practices may further impede access to medical care when required.11 Impact on continuity of care Continuity of care enhances patient satisfaction and trust, which improves the understanding and management of illnesses, adherence to treatment and may ultimately lower mortality rates.12 Results from an Australian online survey13 showed that while most participants (n = 2477 ) do have a usual GP (80%) or general practice (90%), 28% had sought care from multiple practices in the preceding year. Reasons for multiple practice visits were not provided.13 In larger practices, the “usual GP” was often replaced by “usual practice”, resulting in patients consulting GPs they had never met before.13 Similarly, higher levels of continuity of care were experienced in small practices (50%) compared with practices with seven or more GPs (31%), according to a study from the UK.14 A large UK study4 (n = 903 357) found that corporatised practices provided significantly worse overall patient experience compared with traditional GP practices and particularly in regards to consulting a preferred GP. In the qualitative aspect of an Australian study,11 participants visiting corporate practices reported a lack of continuity of care and that doctors were often not aware of their medical history. Furthermore, participants indicated that appointments were short, unsatisfactory and lacked follow‐up. Lower follow‐up rates were also found in for‐profit Swedish primary care centres for patients with chronic diseases compared with centres owned by not‐for‐profit organisations (71% v 75%).15 Impact on quality of care While the efficiency of management processes, co‐location of health services and extension of opening hours of corporate‐owned general practices undoubtedly accommodate the need of patients, research suggests that the quality of corporatised GP services may be worse compared with care delivered by traditional providers.11,16,17 In a UK cross‐sectional study16 (n = 971 232), out‐of‐hours services delivered by corporate organisations scored significantly lower for the timeliness of provided care and for physician and overall patient experience. In 2013, a survey11 of 617 Australian Capital Territory residents did not find a relationship between practice size and patient satisfaction. However, there was a difference between patient satisfaction and care delivered in non‐corporate practices (91%) and corporate practices (79%).11 A UK study17 (n = 8300) found that non‐traditional providers performed consistently worse (P < 0.01) on 15 out of 17 performance measures related to patient experience, cost and efficiency, access and clinical effectiveness. There are few Australian studies that investigate whether size or ownership affects patient outcomes. One of these,18 which focused on diabetes care in nearly 150 practices across three states between 2011 and 2014, found that neither practice size nor ownership type related to good glycaemic control and completion of annual cycles of diabetes care. Impact on health expenditure In 2018–19, about $8.0 billion (6%) of the total government health expenditure related to benefit paid for non‐referred medical services, which include GP services, an increase of 75% in the past decade Box 2).19 The concurrent 48% government expenditure increase per capita on received GP services has been linked to the rise in chronic diseases, resulting in a multitude of services delivered to single patients.1,19 It has been argued that the observed increase might be associated with the practice of overservicing to meet income targets, or with GPs over‐referring to commercially related and co‐located services.22 An analysis of 8 years of general practice activity data, found that practice size was positively associated with pathology ordering: practices with five to ten full‐time equivalent (FTE) GPs had a nearly eight times higher rate of ordering than those with fewer than two FTE GPs.23 An investigation of the impact of pathology co‐location with similar data found no direct association between co‐location and increased test ordering, although the clinical appropriateness of tests was not investigated.24 In addition, some corporate GPs reported feeling pressured by their employers to increase services billed and the occasional use of provider numbers for clinically unnecessary services or added consultation items without the physician’s consent.8 There is also evidence of potential underuse of GP services. Up to 40% of emergency department presentations could have been managed in a general practice setting, and a clear relationship was found with lower levels of continuity of care provided.4,25 Large and corporate‐owned practices seem to be associated with overservicing and lower levels of continuity of care and may therefore negatively affect health costs. A considerable amount of secondary care expenditures might thus be saved each year through effective primary health care delivery. Conclusion There is a growing trend of general practice consolidation with co‐located pathology, imaging and allied health services. This trend is expected to continue as 50% of current practice owners plan to retire within 10 years and nearly 60% of non‐owning GPs are “not at all interested” in becoming practice owners in the future.1 While there is little Australian evidence that worse clinical care is delivered in privately or corporate‐owned general practices, there is also no evidence that care is better. Hence, if general practice in Australia is to navigate the future changes in practice size, ownership and increasingly co‐located service organisation, more Australian research and potentially regulation are needed to track and control what this means for patient care in terms not only of patient experience but also for health outcomes in general practice. Box 1 – Key elements of general practice service provision Box 2 – Government health expenditures and Medicare benefits paid per capita in the financial years 2008–09 and 2018–1919 2008–09 2018–19 Increase (%) Estimated resident population Australia 21 249 199 24 992 747 17.6% All government health expenditure ($)1,3 78.6* 133.6* 70.0%† Benefit paid all services ($) 14.3* 24.1* 68.9%† Benefit paid out‐of‐hospital non‐referred services (GP services) 4.6* 8.0* 74.6%† Benefit paid per capita for total Medicare 671 963 43.5% Benefit paid per capita for GP services 218 322 47.7% Average of GP services per capita 5.3 6.3 18.9% GP = general practice. * Billion. † The increase in percentage is calculated before rounding the expenditures to one decimal point in billions. Source: Australian Institute of Health and Welfare.20,21

Caroline Moel-Mandel · Vijaya Sundararajan

International use of acute kidney injury care bundles, including e-alerts, represents a potential pathway for significant improvement in acute kidney injury management in Australia

Anna C Bendall · Sven‐Jean Tan · Emily J See · Nigel D Toussaint

Australia has a once-in-a-lifetime opportunity to create a healthy, sustainable, equitable and prosperous future by taking bold action to build back better, fairer and greener after the coronavirus pandemic

Sandro Demaio

To the Editor: In an article on the Human Rights Act 2019 passed by the Parliament of Queensland, Brolan1 noted the Act was “historic but not without challenge”. This challenge is manifest in the case of prisoners. In 2007, the Queensland Coroner recommended prisoners have access to clean injecting equipment.2 We described in 2009 the threat to prisoners’ health of ongoing breaches in infection control,3 which was later evident in the cluster of coronavirus disease 2019 (COVID‐19) cases in the Wacol Youth Detention Centre in Brisbane. In 2018, The Medical Journal of Australia documented the precarious state of harm minimisation in Australia’s prisons. With reference to Queensland, there was only mention to the elimination of hepatitis C infection from one prison and the fact that opiate replacement therapy was not available to all prisoners.4 Furthermore, despite some initial success to improve hepatitis C infection rates among Queensland prisoners,5 they have gone backwards, with reportedly high rates of post‐treatment reinfection in Queensland prisons. How is it possible that Queensland continues to stand out as a model of health service deprivation? Evidence that opiate replacement therapy can be life‐saving for prisoners is conclusive.6 Human rights are universal. The right to health provision and health protection cannot be, and in fact has not yet been, effectively negotiated for or by the community’s most disempowered individuals. Despite human rights protections since 2004, the Australian Capital Territory’s dismal experience7 challenges not just Queensland but all Australians.

Michael Levy · Daniel Mogg

Junior doctors face uncertain career progression and futures as a consequence of the pandemic

Kate Johnston · Chloe Tyson · Indra Danny · Lois Meyer

Patients receive similar treatment and have similar outcomes whether their initial hospital has cardiac catheterisation facilities or not

Peter L Thompson

Improving integrated care and systematically targeting major cancer and non- cancer causes of morbidity and mortality could yield major benefits

Emily Banks · Grace Joshy

The COVID‐19 pandemic has necessitated rapid uptake and use of telehealth, unmasking a number of concerns potentially not previously contemplated by clinicians, patients and legislators In the physical distancing climate of coronavirus disease 2019 (COVID‐19), the ubiquity of virtual communications in medical practice generates a number of challenges. Consultation via telehealth allows for creation of audiovisual documentation of the clinical interaction as well as observation by unseen parties from each participant’s perspective, either in real time or subsequently via review of any recordings. It is necessary for clinicians to i) obtain informed consent for clinician‐led recordings, ii) be aware of potential patient‐generated recordings (both declared and undeclared), and iii) meet legal, privacy and storage requirements pertaining to health information arising from a virtual consultation. Consent to participation Observing next of kin or third parties to a virtual telehealth consultation must be introduced to the treating clinician in a manner consistent with an in‐person consultation, whereby such an individual would, with the patient’s consent, attend the consultation with the patient. In considering the clinician’s screen, consent for clinician participation is implied, but should be specifically broadened where appropriate to allow for the presence of clinical observers. Indeed, the clinician’s duty of confidentiality still applies to telehealth consultations, necessitating awareness of others within earshot or visual proximity to the consultation. Implications of virtual participation The benefits of a virtual consultation include participation and collaboration with members of the patient’s family previously unable to participate, as well as increased access to health care for patients with particular physical challenges or vulnerabilities, including vulnerability to infection with COVID‐19. Interviewing a patient in their home adds rare insights for a clinician not typically engaged in home visits, including opportunities for environmental observation, which may be of clinical value. The home setting allows for involvement of parties (seen and unseen) potentially contrary to the patient’s best interests. Pertinent examples include family violence or elder abuse contexts, where presence of offenders may jeopardise the clinical encounter and may pose direct risks to the patient in the periconsultation period and subsequently via covert audio or video footage. A 2020 article provided insights on screening questions for detecting and navigating potential abuse during telehealth consultations in the setting of a COVID‐19‐related domestic violence epidemic secondary to government‐imposed social restrictions.1 Beyond clinical value, novel forms of documentation (including audiovisual recording) generated within the consultation may benefit research, education, billing and coding, subject to appropriate ethical and consent obligations. The content of a traditional clinical consultation episode is limited to the parties in the room and, to a defined extent, other parties (via review of written documentation). In the telehealth context, a wide audience can potentially review video footage of the consultation, as if they were there, for an indefinite period. This may have implications for the practicalities and duration of storage required of such material, its latent role as discoverable documentary evidence in future litigation (particularly given the persuasive nature of audiovisual documentation), and in substantiation of episodic care funding. Clinical interactions may incorporate questions or discussions that, while appropriate sequentially, may appear inappropriate, deficient, discourteous or misleading if taken out of context or distilled to a single statement or query. Recordings, and their potential edits, could be used by patients in a maladaptive manner, engender abnormal illness behaviour, or make a participant consciously or unconsciously feel the need to perform or otherwise change clinical interactions. Recordings by the patient The likelihood of a patient recording a clinical encounter is much higher in the age of telehealth, when secret recording is increasingly possible. The legality of recording a private conversation without consent depends on the state or territory where the person undertaking the recording resides, as surveillance legislation is largely a matter for these jurisdictions (Box 1). In New South Wales, South Australia, Tasmania, Western Australia and the Australian Capital Territory, it is an offence to record a private conversation. This was upheld in NSW in Toth v Director of Public Prosecutions, where it was held that a patient secretly recording a consultation with a general practitioner was an offence.2 However, in Victoria, Queensland and the Northern Territory, it is lawful to record a private conversation without consent if you are a party to the conversation.3,4,5 In all jurisdictions, it is generally not permissible to publish or communicate information secretly recorded. However, exceptions exist; for example, in Victoria, the prohibition on publication or communication of information secretly recorded does not apply to subsequent use in the course of legal or disciplinary proceedings.6 Courts may be more receptive to the notion of undisclosed recordings for defensive purposes where there is a reasonable belief that a recording might be necessary to address a substantive harm. Thus, in certain jurisdictions, patients can secretly record a consultation without the consent of the clinician and this recording may be used in legal or disciplinary proceedings. These risks are best described as emerging given the widespread use of telehealth and the paucity of reported examples of recording. It should also be stressed that when practitioners are behaving professionally and meeting the appropriate standard of care, the medico‐legal risk of patient recordings is minimal. Practical measures to prevent patients from secretly recording screens include disabling the in‐built recording functions in telehealth platforms, using platforms lacking this recording option, and employing programs preventing screen recording or superimposing watermarks including publication preclusion. However, such measures will not prevent another party from recording a consultation with an additional device. Provision of documented restrictions to the patient at the time of any patient‐generated recording and co‐recording by the clinician (to ensure record integrity) may be of value. However, an automated message before consultation commencement expressly stating the clinician does not consent to screen recording (intending to effect a licence agreement or permit a gag order) is unlikely to achieve this in jurisdictions allowing patients to record without the clinician’s permission. Recordings by the clinician Key to understanding and managing both consent and any recordings is the status of these recordings at law. The definition of health information as defined by the Privacy Act 1988 (Cth) s 6FA is broad, including not only information pertaining to someone’s health but also personal information collected to provide, or in providing, a health service to an individual.7 Interpreted literally, any information pertaining to a patient that is recorded, irrespective of consent, may be considered health information with requirements for storage in compliance with the relevant state or territory health records and/or freedom of information legislation (Box 2). In Australia, under the Privacy Act 1988 (Cth) as well as relevant state and territory legislation, a patient’s medical records will generally be held and owned by the clinician or health care organisation, but patients are entitled to access and take a copy of their records. However, concepts of data sovereignty are changing.8 Patient‐driven and centralised health records (such as collaborative digital hospital files and My Health Record) are contemporary examples of this, with reduced clarity about the roles and responsibilities of potential contributors (including the patient) to a medical record as well as the ownership of that information. Various jurisdictions within Australia legislate minimum periods for medical record‐keeping, generally 7 years from the date of the last record entry for adults and until the age of 25 years for children. Many variations exist, based on state or territory, whether the records reside in a public or private institution, or relate to public health, quality improvement, disability, implants or artificial devices, sexual assault counselling, or child protection. Efficient and safe storage of electronic health information by clinicians, including telehealth recordings, is increasingly challenging. Considerations include provisions regulating onshore versus offshore and cloud‐based storage technicalities, including encryption inherent in the platform of choice, preventing evolving real‐time threats to health information security (including via insurance and strategic risk mitigation), and compliance with legislated security requirements. The omnipresence of personal digital devices, including smartphones, has irrevocably altered the role and prevalence of clinical photography, videography and digital team communication tools, constantly generating much data, not all of which are routinely stored by health services or clinicians relying on them to guide clinical decisions. Clear documentation of consent to recording of digital information by clinicians is important, and that consent should extend to the purpose of the recording. When the patient provides their consent, the use of the recording should be limited to that purpose.9 Recording of telehealth discussions between health care workers, including multidisciplinary meetings and case conferences, engenders further challenges. Recordings may be helpful for updating absent clinicians, minute taking, education or documentation. However, such recordings constitute health information, necessitating compliance with management and storage requirements applicable to a virtual consultation. In the public sector, patients may have access to recordings under freedom of information legislation, potentially resulting in significant alterations to the dynamic and tone of the discussion. This is a complex area of law which varies among jurisdictions but is worth keeping in mind. Where there is uncertainty, proactive discussion with medical indemnity providers may be invaluable, especially given the heterogeneity of legal obligations upon clinicians across jurisdictions. Conclusion In Australia, the COVID‐19 pandemic has necessitated rapid uptake and use of telehealth. This has unmasked a number of concerns potentially not previously contemplated by legislators, patients and clinicians, particularly concerning the recording of clinical consultations and thereby the creation of health information, with extensive associated data management and security compliance challenges. Recording of clinical conversations or processes may enhance patient and clinician participation, self‐reference, research, education and funding. In certain jurisdictions, however, clinical consultations or meetings may be lawfully recorded with or without participants’ knowledge, and may later be accessible to the patient, including for use in future legal or disciplinary proceedings, potentially stifling candid discussion. This and the challenging obligations relating to data management technicalities represent real risks for clinicians and health services. It is incumbent upon health care providers and lawmakers alike to consider these issues in a practical context, ensuring that telehealth is not only a useful tool but a safe and effective one. Box 1 – Legislation governing covert recordings State or territory Legislation pertaining to recording Australian Capital Territory Listening Devices Act 1992 (ACT) New South Wales Surveillance Devices Act 2007 (NSW) Northern Territory Surveillance Devices Act 2007 (NT) Queensland Invasion of Privacy Act 1971 (QLD) South Australia Listening and Surveillance Devices Act 1972 (SA) Tasmania Listening Devices Act 1991 (TAS) Victoria Surveillance Devices Act 1999 (VIC) Western Australia Surveillance Devices Act 1998 (WA) Box 2 – Legislation governing health information management Jurisdiction Legislation governing health information management (not including legislated regulations) Federal Privacy Act 1988 (Cth); Personally Controlled Electronic Health Records (Consequential Amendments) Act 2012 (Cth); My Health Records Act 2012 (Cth); Freedom of Information Act 1982 (Cth) State or territory Australian Capital Territory Health Records (Privacy and Access) Act 1997 (ACT) New South Wales Health Records and Information Privacy Act 2002 (NSW) Northern Territory Health Services Act 2014 (NT); Information Act 2002 (NT) Queensland Information Privacy Act 2009 (QLD); Right to Information Act 2009 (Qld); Public Records Act 2002 (QLD) South Australia Freedom of Information Act 1991 (SA); State Records Act 1997 (SA) Tasmania Personal Information Protection Act 2004 (TAS); Right to Information Act 2009 (TAS) Victoria Health Records Act 2001 (VIC); Privacy and Data Protection Act 2014 (VIC); Freedom of Information Act 1982 (VIC); Public Records Act 1973 (VIC) Western Australia Freedom of Information Act 1992 (WA); State Records Act 2000 (WA)

Caitlin C Farmer · Sam C Pang · Dev Kevat · Jessica Dean · Danielle Panaccio · Patrick D Mahar

In June 2015, 41% of Australian medical specialists were women,1 but only 28% of those in senior or leadership positions.2 Academic research is important for obtaining tenure and promotion in medicine. First authorship on publications is typically granted to junior authors and last authorship to directing senior authors. The proportion of women among first authors in six prominent American medical journals increased from 5.9% in 1970 to 29.3% in 2004, and for last authorship from 3.7% to 19.3%.3 However, a 2016 study found that the proportion of authors who were women in high impact medical journals had plateaued or declined since 2009.4 Examining Australian patterns of authorship could help identify barriers to the academic advancement of women in medicine. We identified in PubMed all journal articles published during 2005–2018 by the eight journals associated with peak bodies of Australian medical practitioners, and used the validated genderize. R tool to determine the probable gender of authors’ first names.5 We used Poisson regression to analyse first and last authorship (male = 0, female = 1) by year; we report the statistical significance of the deviation of the regression slope (B‐value) from zero. The relationship between number of authors and gender were assessed by linear regression, including an interaction term between gender and time. Formal ethics approval was not required for this analysis of publicly available data. Gender could be determined with at least 50% probability for the first authors of 26 621 of 27 804 articles (96%) and the last authors of 26 972 (97%). Between 2005 and 2018, the proportion of women among first authors in the eight journals increased from 522 of 1600 (32.6%) to 899 of 2391 (37.6%; P < 0.001); the proportion among last authors did not change (28.0%). The proportions of women among both first and last authors increased significantly in the Journal of Paediatrics and Child Health, the Australian and New Zealand Journal of Obstetrics and Gynaecology, and the Medical Journal of Australia, as did those of first authors (but not last authors) in the Australian and New Zealand Journal of Public Health, Emergency Medicine Australasia, the Australian and New Zealand Journal of Psychiatry, and the Australian and New Zealand Journal of Surgery; the proportions of neither changed significantly in Australian Family Physician (Box; Supporting Information, table 1). The mean number of authors on publications with women as first authors (3.8; standard deviation [SD], 2.4) was higher than for those with men as first authors (3.3; SD, 2.4; P < 0.001). The difference between author numbers was smaller, but statistically significant, with respect to last author gender (women: mean number of authors, 3.6; SD, 2.4; men: 3.5; SD, 2.4; P = 0.045) (Supporting Information, tables 2, 3). Our study did not distinguish between research, review, and other journal article types. While our findings may reflect overall involvement of women in research, they do not specifically define gender proportions among leaders of high impact academic research programs. The increase in the proportion of first authors of Australian medical journal articles who are women may reflect the rise in the proportion of female doctors from 33% to 43% between January 2006 and December 2018.1 It is also possible that women, under‐represented in their specialties, feel greater pressure than men to publish as first authors for purposes of career progression.2 Our data indicate that the proportion of women as first authors has increased, but that of last authorship has grown only in some specialities. Box – Proportions of women as first and last authors of articles in selected Australian medical journals, 2005–2018* * The raw data are included in the online Supporting Information, tables 4 and 5. † From 2018: the Australian Journal of General Practice.

Matthew J Lennon · Rose Kennedy · Hannah Ryan · Dennis R Neuen · Melissa Godwin

Putting the patient at the core of personalised cancer care delivery remains the elusive final frontier Over 20 years ago, the Australian House of Representatives Inquiry into the management of breast cancer recommended that cancer care should be delivered using a multidisciplinary approach.1 Ten years later, an article published in this Journal articulated how to put multidisciplinary care into practice,2 paving the way for the concept to be embedded into clinical cancer practice and policy of today.3 One of the key recommendations made in the article, and since adopted as national policy, was for the patient to be included “as a member of the multidisciplinary team”. But as of today, multidisciplinary care does not routinely include input from patients themselves. Patients do not attend multidisciplinary meetings. Rather, their circumstances are discussed and treatment recommendations are made. They may subsequently make a shared decision with the clinician, but their input tends to occur after the multidisciplinary discussion and it is uncommon for the patients’ perspectives to systematically inform these discussions. Putting the patient at the core of personalised cancer care delivery thus remains the elusive final frontier. The potential benefits of such an approach are well established and include a greater alignment of care with individual patient goals, better understanding of needs, and better patient outcomes and satisfaction.4 Health care providers are poor surrogates for assessment of patients’ symptoms, needs and experiences and patient‐reported outcomes (PROs) collection is a way to systematically integrate patients’ perspectives into assessment, treatment planning, and ongoing monitoring.5 PROs report on patients’ subjective perception of health, functional status, unmet needs, and quality of life and are collected directly from patients either online, via a smartphone, or through paper‐based means. While there are many sets of questions that are relevant to any patient with cancer, specific questions can be tailored to particular cancer types, populations, or different phases of cancer trajectory.6 PROs as a concept are not new and not unique to cancer. However, while they have been used extensively in research, including clinical trials, their adoption in routine clinical care has received much less attention, with only one article on the topic published in this Journal over 10 years ago.7 This article summarises the current evidence supporting the use of PROs in cancer as an example of their potential of relevance to broader health care delivery, and argues for their routine adoption into practice. The evidence for the utility of PROs in cancer care is compelling. A single‐institution randomised controlled study of 766 patients included participants with multiple metastatic cancers and assigned them to a usual care group or a group that reported on their symptoms using an electronic portal.8 The study found that patients reporting PROs had longer survival, with a median prolongation of 5.2 months; comparable in effect size to many effective, novel cancer therapies.8 In Ontario, where PROs have been routinely collected since 2007, administrative data analysis has shown survival improvement irrespective of the phase of cancer treatment, as well as reduced hospitalisation and emergency department presentations.9 Two systematic reviews concluded there was strong evidence that implementation of PROs improves patient–provider communication and patient satisfaction.10,11 PROs are recommended by the Australian Commission on Safety and Quality in Health Care, have been advocated for by the Clinical Oncology Society of Australia, and have been included as a critical element of care in the Australian Digital Health in Cancer Care Roadmap.12 But to date, their adoption into routine clinical practice has been limited. Like multidisciplinary care 20 years ago, a significant barrier to their integration has been the ability of the health system to operationalise routine collection and response to PROs data. In 2020, we finally have technology for efficient, real‐time collection, reporting of, and response to PROs through customisable portals and dashboards and integration with the electronic medical records. But technology alone is not sufficient and its roll‐out, especially with regards to electronic medical records, has been slow and fragmented. Similar to the approach to multidisciplinary care,2 it is time to articulate the principles and outcomes necessary to integrate PROs into the routine clinical workflow (Box 1). Where multidisciplinary care called for a core team of experts, the PROs collection requires a core dataset. While a dataset using a generic PROs measure, such as the Edmonton Symptom Assessment System Revised (ESAS‐r) used in Canada, may be most appropriate for screening for unmet needs in any clinical setting, more specific measures may be required for assessment of different cancer types, different phases of disease (ie, at diagnosis v end of life) or for different populations, such as Indigenous patients.13 There is a need for a clear communication framework involving relevant heath care providers in a timely fashion, with feedback communicated to the patient. The process must be accessible to patients irrespective of technology, rurality, remoteness or language barriers. Lastly, the collection of PROs needs to be underpinned by agreed standards that clearly articulate and support the role of the patient in this process. While the barriers to adoption of PROs in clinical practice are significant,14 they are not insurmountable. System redesign may be required to integrate PROs collection and feedback into the routine workflow, with clear pathways to inform a standardised approach. PROs collection should not become an additional task but rather be considered part of a realignment of workload and services to meet patients’ needs, reduce care variation, and optimise resource utilisation. A systematic approach to identification of needs is critical to supporting self‐management, an essential component of patient care, as it assists the patient in knowing what symptoms are unexpected, what to report, and how to seek support when needed.15 Data from existing PROs systems show that rapid, real‐time feedback to health care providers facilitates timely response, reducing the likelihood of issues escalating or remaining unaddressed. In many cases, this response may only require reassurance and/or advice on self‐management, with only the more severe issues necessitating referral and/or hospital admission. With advances in technology, patients and health care providers can obtain visual summaries of trends over time that may assist further in decision making, while aggregated data derived from individual cases can be used to drive health system improvement and plan services to meet demand. Are we ready for this final frontier? PROs can become an important part of value‐based care delivery with support through relevant drivers, such as reimbursement and accreditation. But we need to build capacity through training and a community of practice to share learnings, resources and tools. Many tools already exist (Box 2), with technology fully capable to support rapid processing of data and linkage to electronic health records; although it is important to note that PROs collection can be achieved using paper‐based surveys or simple reporting such as text messaging. As we reflect on the 20 years of multidisciplinary cancer care in Australia, it is worth remembering that multidisciplinary care is not just about multidisciplinary meetings. Similarly, PROs are not just about PROs collection tools. Together, these two complementary approaches put into practice the principle of personalised care. It is the focus on the patient that enables us to realise the full potential of the multidisciplinary care through framing multidisciplinary recommendations in the context of what the patient identifies as their main issues, needs or concerns. It is time to reach this final frontier and make personalisation of cancer care through PROs an achievable standard in Australia. Box 1 – Principles of clinical use of patient‐reported outcomes (PROs) in cancer care Principle Outcome Core data Agreed core dataset appropriate for patient, population or setting Communication framework Relevant information is communicated to relevant team members in a timely fashion through agreed channels Access Systems established to ensure access for all users irrespective of distance, technology, language, literacy level, or completion method Standards of care Patients and clinicians are aware of PROs and support their use at key points in clinical care pathway through: best practice protocols; stratified alert systems; reporting framework; professional development opportunities; adequate resourcing to allow collection, scoring, review, response and feedback and re‐screening; and feedback and co‐design opportunities Patient involvement Information and education Self‐management support Feedback and co‐design opportunities Box 2 – Examples of clinical practice resources for patient‐reported outcomes (PROs) use in routine care Australian Commission on Safety and Quality in Health Care (https://www.safetyandquality.gov.au/our-work/indicators-measurement-and-reporting/patient-reported-outcome-measures) General information on PROs, including list of validated measures and guides for implementation International Consortium for Health Outcomes Management (https://www.ichom.org/) Multiple resources and standard datasets for multiple conditions including cancer Cancer Care Ontario, Patient Reported Outcomes and Symptom Management Program (https://www.cancercareontario.ca/en/cancer-care-ontario/programs/clinical-services/patient-reported-outcomes-symptom-management) A system of PROs screening for patients with cancer, including clinical pathways and guides for patients and health care providers Clinical Oncology Society of Australia, PROs Think Tank report (https://www.cosa.org.au/media/332504/cosa_pros_think_tank_report_final.pdf) An overview of evidence, current practice and recommendations for PROs uptake into cancer care in Australia

Clinical Oncology Society of Australia (COSA) Patient Reported Outcomes Working Group

There is little legal recourse in Australia for someone who is denied care on resource grounds, particularly in the context of a pandemic

Michelle A Gunn · Fiona J McDonald

To the Editor: Fetene and colleagues1 describe refusal to prescribe or dispense prescriptions to some of their study cohort of people who inject drugs (PWIDs) and concern that their mental health treatment needs may not be met due to SafeScript, Victoria’s real time prescription monitoring system. SafeScript is a live electronic database providing information about the prescribing and dispensing of monitored medicines to each patient, instantly available in real time on a prescriber’s or pharmacist’s desktop (https://www2.health.vic.gov.au/safescript). SafeScript alerts prescribers and pharmacists to the risk of uncoordinated treatment by multiple providers or to the overdose risk of drug–drug interactions. Prescribers and pharmacists may have responded appropriately by offering more effective treatments instead of providing a continued supply of the medication. For instance, benzodiazepines are not recommended for first line or prolonged treatment of anxiety. Many people who misuse drugs have comorbid mental health disorders and need medical support. SafeScript is helping to identify this group of at‐risk patients so they can receive the appropriate medical treatment they require. SafeScript provides a clinical decision support system for prescribers and pharmacists, enabling more informed decisions for safer prescribing or dispensing of high risk monitored medicines. By providing proactive alerts, strong real time prescription monitoring systems, such as SafeScript, reduce overdose deaths from prescription opioids2 and decrease the number of opioid prescriptions, diversion, and opioid‐related morbidity and substance use disorder outcomes.3 Since the rollout of SafeScript, the number of multiple provider episodes and the average morphine equivalent dose have both been trending gradually but consistently downwards. SafeScript was designed with the lessons learnt from the United States. The implementation of this system included several measures to encourage and support a professional response for high risk patients, such as podcasts and face to face and online training, in which more than 4500 prescribers and pharmacists have participated. Furthermore, the SafeScript implementation included upgrading the Drug and Alcohol Clinical Advisory Service, providing trained general practitioner clinical advisors to offer peer support, a consumer pharmaceutical helpline, and increased funding to support professionals and consumers manage benzodiazepine problems. In addition, $273.1 million were invested in drug treatment, support and harm reduction services in 2019–2020,4 representing a 65% increase in investment through the last five Victorian state budgets. This new initiative helps prescribers and pharmacists provide the appropriate clinical care and professional response needed by high risk patients.

Malcolm Dobbin

Objectives: To describe patterns of care in New South Wales for men with prostate cancer, and to ascertain factors associated with receiving different types of treatment. Design: Individual patient data record linkage study. Setting, participants: 4003 New South Wales men aged 45 years or more enrolled in the population‐based 45 and Up Study in whom prostate cancer was first diagnosed during 2006–2013. Main outcome measures: Prostate cancer treatment type received; factors statistically associated with treatment received; proportions of patients who consulted radiation oncologists prior to treatment. Results: In total, 1619 of 4003 patients underwent radical prostatectomy (40%), 893 external beam radiotherapy (EBRT) (22%), 183 brachytherapy (5%), 87 chemotherapy (2%), 373 androgen deprivation therapy alone (9%), and 848 no active treatment (21%). 205 of 1628 patients who had radical prostatectomies (13%) had radiation oncology consultations prior to surgery. Radical prostatectomy was more likely for patients aged 45–59 years, with regional stage disease, living 100 km or more from the nearest radiotherapy centre, having partners, or having private health insurance, while lower physical functioning, obesity, and living in areas of greater socio‐economic disadvantage reduced the likelihood. EBRT was more likely for patients aged 70–79 years, with non‐localised or unknown stage disease, living less than 100 km from the nearest radiotherapy centre, or not having private health insurance, while the likelihood was lower for patients aged 45–59 years or more than 80 years and for those who had several comorbid conditions. Conclusions: Men with prostate cancer were twice as likely to have radical prostatectomy as to receive EBRT, and fewer than one in seven had consulted radiation oncologists prior to prostatectomy. The treatment received was influenced by several socio‐demographic factors. Given the treatment‐specific side effects and costs, policies that affect access to different treatments for prostate cancer should be reviewed.

Mei Ling Yap · Dianne L O'Connell · David E Goldsbury · Marianne F Weber · David P Smith · Michael B Barton

To the Editor: As the coronavirus disease 2019 (COVID‐19) pandemic spread across Europe, we worked in the intensive care unit (ICU) of a Swedish university hospital. We share our experiences and offer some thoughts regarding Sweden’s pandemic response. The decentralised Swedish health system works on three levels (Box). These traditional divisions may partially account for the lack of coordination between care services in the initial phases of the pandemic, where large numbers of deaths occurred in care homes. As the pandemic intensified, safety checks were implemented to protect residents of aged care facilities. This resulted in a quick containment of infections, although tragically too late for many. Our health care region received the fourth highest number of hospitalisations in Sweden.1 We were privileged to work in a system that was well organised, without political conflict and with pre‐existing disaster plans that were quickly converted to pandemic plans. A pandemic‐specific leadership established a centralised inventory and oversaw the acquisition and distribution of beds, staffing, medical equipment, essential drugs, personal protective equipment and disinfection agents. An eight‐step plan ensured a rapid escalation of regional ICU capacity. Intermediate care units were opened, reducing demand for ICU beds. Projected numbers of patients were calculated daily, based on models provided from the Public Health Agency of Sweden and local data. Anaesthesia and intensive care are a combined speciality in Sweden. This enabled the rapid deployment of a large workforce of anaesthetists and nurse anaesthetists to ICUs. Despite these resources, our tripled ICU capacity meant significant staffing challenges, with additional difficulties because of staff illnesses and quarantines. Our impression is that the Swedish response has been controlled and planned for the long term. Daily public announcements from the Public Health Agency became a regular part of our lives and Swedes were generally compliant with recommendations regarding physical distancing and hygiene routines. We are perplexed by reports in the media that life went on as usual in Sweden. In fact, life was very different. Most people worked from home, large numbers were furloughed, many institutions were closed and public events were cancelled. Travel was discouraged and fell dramatically.2 What sets the Swedish approach apart from others is that these measures were largely voluntary, with generally good public support. We avoided an overwhelming wave of patients with an undercapacity of ICU beds, as seen in many other countries. We maintained normal criteria for ICU admissions. This is notable given that Sweden has the second lowest number of ICU beds per capita in Europe.3,4 Results from intensive care are encouraging, with mortality rates generally lower than previously reported.5 Challenges included staff burnout, a shortage of usual sedatives and lack of clinical experience with this new disease, resulting in the use of futile and potentially harmful treatments. However, guidance from a national group of senior clinicians provided regular recommendations6 and there was excellent compliance with advice from regulatory authorities. Up to 70% of elective surgeries were cancelled during the first half of 2020. Cancer‐related surgeries continued to be prioritised during the pandemic, but the longer term effects of cancelled surgeries, outpatient clinics and altered illness behaviour are not known. We are heartbroken at our inability to provide enough comfort to relatives of our patients who succumbed to COVID‐19 when hospital visits were prohibited. As two Australian emigrants working in a Swedish ICU, we are humbled by our ability to contribute to the care of patients during the pandemic. Our Australian medical training instilled in us a sense of duty, tempering any feelings of helplessness. We applaud the tenacity of our Swedish colleagues. We wish our Australian colleagues well and hope that Australia will be protected from the horrors of COVID‐19. Box – Decentralised organisation of the Swedish health care system

Michelle S Chew · Thomas Halliday

To the Editor: Palliative radiotherapy is effective for symptomatic management of bone metastases in cancer patients. However, it may take 2–4 weeks after completion of radiotherapy to achieve maximal clinical response.1 Radiotherapy can be delivered as a single fraction treatment (SFRT), or over a more protracted course of multifraction treatment (MFRT).2 Randomised trials have consistently shown that SFRT and MFRT provide equally effective symptom control,3 and SFRT is associated with lower medical and societal cost,4 allowing for better health services utilisation. Hence, in patients with poor prognosis, the use of SFRT over MFRT should be encouraged to minimise the time patients spend on treatment at the end of life without compromising efficacy. Using the population‐based Victorian Cancer Registry data linked to the Victorian Radiotherapy Minimum Data Set, we evaluated the use of SFRT for bone metastases at the end of life. The study sample included all cancer patients who received radiotherapy for bone metastases between 2013 and 2016, and died within 30 days of commencing radiotherapy. The primary outcome was SFRT use and the associated factors. The Cochrane–Armitage test for trend was used to evaluate temporal changes in SFRT use over time. Logistic regression was used to evaluate factors associated with SFRT use; variables with a P value below 0.1 in univariate analyses were included in multivariate model, which employed the robust standard error, with analyses clustered on patient identifiers to allow for clustering of patients who had multiple courses of radiotherapy. The study was approved by the Austin Health Human Research Ethics Committee (LNR/18/Austin/34). A total of 1069 patients received 1359 courses of radiotherapy for bone metastases at the end of life, of which 396 courses (29%) were SFRT, and 963 (71%) were MFRT (Box). There was no significant change in SFRT use over time: from 30% in 2013 to 32% in 2016. SFRT was more commonly used closer to death: 49%, 29% and 25% of radiotherapy courses delivered within 7 days, 8–14 days, and 15–30 days of death, respectively. There were large institutional provider variations in SFRT use: 33% and 19% of radiotherapy delivered in public and private institutions, respectively. In multivariate analyses, the site of bone metastases, time between radiotherapy and death, and treatment institution type were independently associated with SFRT use. Overall, in this large Victorian population‐based study, less than one in three courses of radiotherapy for bone metastases at the end of life were SFRT, and about one in two courses of radiotherapy delivered in the last week of life were MFRT, meaning that these cancer patients spent multiple days in their final week of life receiving radiotherapy. Acknowledging that estimation of prognosis towards the end of life can be difficult, there are models (eg, the TEACHH model)5 that can be useful in guiding clinicians in this process. Nonetheless, given the large body of evidence supporting the use of SFRT for bone metastases, there is a need to raise awareness of the recommendation to use of SFRT instead of MFRT, especially at the end of life, among radiation oncologists, other health professionals and patients. This can be achieved through health education initiatives such as the Choosing Wisely campaign (https://www.choosingwisely.org.au/). Box – Factors associated with single fraction palliative radiotherapy (SFRT) for bone metastases at the end of life in Victoria, 2013–2016 (1359 courses of radiotherapy) Variable SFRT (n = 396, 29%) MFRT (n = 963, 71%) Multivariate analysis (odds ratio [95%CI]) P Age (years) Mean (SD) 71.7 (11.9) 70.1 (12.3) < 60 62 (24%) 192 (76%) 1 60–69 95 (27%) 252 (73%) 1.04 (0.67–1.61) 0.9 70–79 149 (32%) 312 (68%) 1.29 (0.83–1.99) 0.3 ≥ 80 90 (30%) 207 (70%) 1.17 (0.73–1.88) 0.5 Sex Men 247 (28%) 622 (72%) Women 149 (30%) 341 (70%) Primary cancer type Lung 163 (30%) 373 (70%) Prostate 49 (29%) 118 (71%) Breast 32 (29%) 77 (71%) Gastrointestinal 53 (28%) 138 (72%) Melanoma 21 (25%) 62 (75%) Other 78 (29%) 195 (71%) Target site of radiotherapy Spine 202 (27%) 545 (73%) 1 Skull 9 (6%) 147 (94%) 0.15 (0.06–0.38) <0.001 Rib 35 (51%) 33 (49%) 3.82 (2.10–6.95) <0.001 Shoulder 35 (49%) 36 (51%) 2.80 (1.44–5.42) 0.002 Hip 19 (36%) 34 (64%) 1.67 (0.78–3.54) 0.2 Pelvic bone 20 (27%) 54 (73%) 1.10 (0.57–2.14) 0.8 Extremities 42 (52%) 39 (48%) 3.04 (1.74–5.29) <0.001 Multiple site 34 (31%) 75 (69%) 1.25 (0.73–2.14) 0.4 Time between radiotherapy start date and death 1–7 days 92 (49%) 97 (51%) 1 8–14 days 99 (29%) 246 (71%) 0.40 (0.25–0.65) < 0.001 15–30 days 205 (25%) 620 (75%) 0.33 (0.21–0.51) < 0.001 Socio‐economic status 1st quintile (most disadvantaged) 103 (35%) 194 (65%) 1 2nd quintile 63 (32%) 137 (69%) 0.88 (0.55–1.40) 0.6 3rd quintile 70 (26%) 203 (74%) 0.73 (0.46–1.17) 0.2 4th quintile 55 (22%) 194 (78%) 0.63 (0.39–1.02) 0.06 5th quintile (least disadvantaged) 105 (31%) 235 (69%) 1.00 (0.62–1.61) 0.9 Remoteness of area of residency Major city 262 (28%) 668 (72%) Inner regional 109 (31%) 241 (69%) Outer regional/ remote 25 (32%) 54 (68%) Treatment institution type Public 317 (33%) 633 (67%) 1 Private 79 (19%) 330 (81%) 0.44 (0.29–0.65) < 0.001 Treatment institution location Metropolitan 280 (27%) 746 (73%) 1 Regional 116 (35%) 217 (65%) 1.02 (0.71‐1.47) 0.9 Year of radiotherapy 2013 115 (30%) 262 (70%) 2014 93 (28%) 243 (72%) 2015 84 (26%) 241 (74%) 2016 104 (32%) 217 (68%) MFRT = multifraction radiotherapy.

Wee Loon Ong · Farshad Foroudi · Roger L Milne · Jeremy L Millar

Challenges with data infrastructure are affecting medical workforce research and access to medical care Access to high quality medical care can save lives and help reduce the consequences of the growing burden of chronic disease. However, the delivery of this care relies on a well trained health and medical workforce organised to optimally respond to community need, working in supportive work environments within models of care that are fit for purpose, with minimal geographic or financial barriers to access for all communities. There has been a long term need in Australia for coordinated, evidence-informed workforce policies. However, for many years the development of the medical workforce has been shaped by self‐regulation and market forces. Short term and uncoordinated workforce planning has generated cycles of contraction and expansion of training places, sporadic regulation, and recent policy dilemmas.1,2 Most recently, a dramatic increase in numbers of graduates from Australian medical schools has occurred in the absence of clear plans as to how to use these additional doctors to optimally meet community need. Early data suggest that flooding the market with more graduates has not addressed persistent rural shortages, with insufficient numbers willing or able to navigate a career pathway to work in areas of need.3,4 Oversupply continues to be an issue in some specialties (eg, emergency medicine or cardiothoracic surgery) while shortages persist in others such as general practice and psychiatry.5 Over‐reliance on international medical graduates continues in many rural communities,1 while the fierce competition for accredited training places in some specialties leaves many junior doctors caught in the middle.6 Furthermore, Australian doctors are increasingly reporting burnout and mental health problems,7 with significant negative effects on productivity and patient safety.8 With these problems seeming to defy solutions,9 it is not surprising that there have been calls to add the work–life balance of clinicians to the Institute for Healthcare Improvement’s set of principles to guide optimising health system performance (optimal patient experience, improved population health and reducing costs).10 In light of these issues, the development of Australia’s new National Medical Workforce Strategy (NMWS) scoping framework and consultation process for the final strategy is welcome. The NMWS is being designed to frame the development and coordination of national medical workforce policies to address our pervasive workforce challenges: geographic maldistribution; specialty over‐ and undersupply; the balance of generalists and specialists; Indigenous and culturally safe workplaces; doctor work readiness; and changing models of care.11 One of the six principles of the NMWS is to “[a]pply an evidence‐based approach wherever possible, drawing on data and information from all stakeholders”.11 Data on the medical workforce Achieving an evidence‐based approach to workforce policy requires more high quality longitudinal and linkable data that is both broad across different doctor groups and rich in doctor characteristics, compared with what is currently available (Box). Institutional bias, fragmentation, inconsistent definitions and restricted access provide substantial barriers to our ability to use those data for the social good. Few available sources offer a long term, holistic and objective view of the medical workforce: professional training bodies can only use data sourced from relatively brief periods of postgraduate training; the Department of Health relies on Medicare billing data and raw counts of medical practitioners through the Australian Health Practitioner Regulation Agency; and the states are limited to poor data on salaried, generally hospital‐based practitioners. Data that are made available to researchers are overly aggregated, especially geographically, often preventing useful evidence from emerging about medical workforce behaviours, training outcomes, career choices and treatment patterns. Many sources remain closely guarded by training and service providers and governments, such as surveys regularly completed by doctors on registration with the Australian Health Practitioner Regulation Agency (including the new national medical training survey12) or with individual colleges. Where data are controlled by individual agencies, there is minimal potential for multipurpose use and no process for linkage to other sources. Hence, it is impossible to understand and track career pathways of doctors even though these are a key element of policy. The analysis of workforce data to generate evidence from these multiple sources has been relatively unsophisticated and preoccupied with the simple modelling of supply and demand — ignoring how practitioner behaviours, and the drivers of those behaviours, influence workforce numbers and practitioner quality. Although these data can be used to count and describe trends, they mostly cannot be used to understand why decisions are being made and how services are driven, which are essential for designing policy. The Australian community deserves a broader understanding as to how different policies and programs are addressing their needs. Lack of this understanding has been a major contributor to the decisions that have led to the current situation of workforce oversupply.16 Neither are health workforce data linked to patient‐level data — a factor overlooked in the NMWS scoping framework — that is, data on inputs are not linked to data on activities, outputs and health outcomes, making it impossible to determine how workforce and policy changes affect community needs and population health. Any policies aimed at the medical workforce should at least examine their effects on patients. Finally, the availability of administrative medical workforce data to researchers is at an all‐time low. There was a reduction in funding of the Medical Schools Outcomes Database in 2015 and the withdrawal of funding (from 2016) for the Australian Institute of Health and Welfare to produce health workforce statistics. The Bettering the Evaluation and Care of Health (BEACH) study14 was also discontinued as the only data on the clinical activities of general practitioners. Adding to the challenge, the internationally unique Medicine in Australia: Balancing Employment and Life (MABEL) panel survey of 9–10 000 doctors per year ceased in 2019 after 11 annual waves of data collection.13 Moreover, researchers skilled in using health workforce data will be difficult to sustain without addressing the availability of data, and this expertise will soon dissipate, adding to severe reductions of health workforce analytical staff at the Commonwealth level when Health Workforce Australia ceased in 2014. It is notable that the new National Health Information Strategy makes no mention of health workforce data.17 Despite its ongoing reliance on competitive grant funding, MABEL data have played a key role in national medical workforce policy over the past 11 years. It was a World Bank exemplar of health workforce data collection internationally,18 and continues to guide the distribution of over $1 billion funding to regional health care through its use in the design of the Modified Monash Model (used to classify which geographical areas are eligible to receive increased funding), as well as supporting the design of rural health workforce programs. Unlike other datasets (Box), MABEL data transcended traditional divides of salaried and private practice, different doctor types, career stages and career trajectories as the basis for supporting policy and program decision making at a national scale. The future for medical workforce research The medical workforce represents the backbone of the health care system and a major public investment, yet despite the large gap between supply and community need, the scope of available data does not support evidence-informed decision making. While existing administrative and registration minimum data support national medical workforce planning, they are unable to give insights into doctors’ career and clinical decisions. With the pressures on the health care system and medical workforce at an all‐time high, we believe that the Australian community deserves better insights into how different medical workforce policies and programs are promoting access to equitable, high quality care. We need to know more about the doctors being produced from long and expensive taxpayer‐funded training programs, as well as why they choose disciplines, practice locations and practice patterns. Moreover, there is a growing awareness of the importance of maintaining the health and wellbeing of this workforce, but available national data to underpin key policies to prevent poor mental health are missing. We propose that any reforms to the Australian health care workforce must be informed by robust evidence. The collection and availability of this evidence needs to be at the forefront of policy and planning, embedded within objectives of key national strategies such as the NMWS and National Health Information Strategy. Future medical workforce data strategies need to be institutionally neutral, guided by a research strategy including agreed priority research questions with resources to conduct the research, and underpinned by openness and data sharing. Healthy national medical workforce data are fundamental to achieving healthy doctors and communities. Box – Available national datasets on the medical workforce* table#t1 tbody td:nth-child(n+2) P. Pleft { text-align: center; } Data source Unit record data available to external researchers Unique identifier to enable linkage over time Data linked to patients Rich data on doctor characteristics Doctors grouped by organisation (practice, hospital) Doctor group Medicare provider file/MBS With consent Yes Yes No No Private practice Medical college surveys No Yes (but some surveys anonymous) No No No Vocational trainees and Fellows National medical training survey12 No No (anonymous) No No No Pre‐vocational and vocational MABEL13 With consent Yes No Yes Yes All medical practitioners BEACH14 No NA (random sample of GPs each year) Yes Yes Yes GPs National Health Workforce Dataset15 AHPRA registration data No Yes No No No All medical practitioners AHPRA registration survey No (table builder available) No No No No All medical practitioners Medical Education and Training dataset No No No No No Pre‐vocational and vocational BEACH = Bettering the Evaluation and Care of Health; GP = general practitioner; MABEL = Medicine in Australia: Balancing Employment and Life; MBS = Medicare Benefits Schedule; NA = not applicable. * States and territories also have their own data collections for the public hospital workforce, but these vary in what is collected and are not available to external researchers. Many hospitals in recent years also conduct surveys of health and wellbeing. Many clinical registries, epidemiological datasets, hospital separation data, and electronic medical record data focus on patients and do not include doctor identifiers or characteristics.

Grant M Russell · Matthew R McGrail · Belinda O’Sullivan · Anthony Scott

Evidence to inform conversations on Aboriginal health issues — in response to COVID‐19 and beyond Nearly everyone has been affected in some way by the coronavirus disease 2019 (COVID‐19) pandemic, and it is a public health risk for Aboriginal peoples and communities.1 The impacts of the pandemic are pervasive, wide‐ranging and continue to affect people and communities differently. Concerns about the indirect impacts of COVID‐19, caused by missed, delayed and avoided health care — not as a direct consequence of COVID‐19 infections — are shared internationally.2,3,4 While the prevalence of COVID‐19 in New South Wales remains low,5 local data show significant changes in health utilisation across the state. During the 4‐month period from March to June 2020, compared with the same period in 2019, face‐to‐face primary care consultations decreased by 22.1%, breast screen activity by 51.5%, ambulance incidents by 7.2%, emergency department visits by 13.9%, public hospital inpatient episodes by 14.3%, and public hospital planned surgical activity by 32.6%.6 Such decreases are not unique to NSW.7 Before COVID‐19, Aboriginal people faced health disadvantages and inequitable access to health care. Any decrease in health care access for Aboriginal people through missed, delayed or avoided health care may lead to further adverse health outcomes and inequities.1,4,8 In recent months, we came together as a group of 12 Aboriginal community members from across NSW to share our experiences and perspectives regarding the indirect impacts of COVID‐19. We live and work on Eora, Wilyakali, Bundjalung, Yuin and Gumbaynggirr lands. The discussions occurred over three separate sessions, each held a week apart between 24 August and 1 September 2020. Six members of the group (DF, CP, PO, BO, DL and KB) captured the key messages identified from the talks and synthesised the findings into three main themes: community supporting the community; the social determinants of health; and access to health care. These conversations were hosted and supported by the Critical Intelligence Unit established as part of the NSW Health COVID‐19 response and the Agency for Clinical Innovation (TDB). Illustrative quotes shared by the co‐authors have been selected to demonstrate salient points. The term “mob” has been used throughout to identify who we are and where we are from — our connection to our shared identity as Aboriginal people. Community supporting the community is a real strength — in the pandemic, and always In responding to COVID‐19, we see that Aboriginal organisations are coming together, more than ever, to create a movement that will continue to inform positive change to address Aboriginal health issues. Mob are proud of how they are keeping each other safe. It is a point of pride that has strengthened community. Our mob are concerned about the safety of others and our elders. (CP) Aboriginal leaders and Aboriginal community controlled health services are active in responding to COVID‐19, drawing on experiences from the 2009 HINI influenza pandemic and implementing culturally appropriate resources.9 The pandemic has been disruptive, and community events and gatherings have been cancelled because of important and legitimate public health concerns. However, this does impact our community approach to health care, cultural practices and connection to country.1,10 Our mob aren’t able to connect for sorry business and funerals, marriages and births. The provision of our health care, along with the provision of our social and emotional wellbeing, has changed. And connectivity is the main ingredient for our mob to stay healthy. This is the biggest barrier. (CP) Social determinants of health for Aboriginal people Social determinants are the conditions in which people are born, grow, live and age, and how these factors influence our health and determine health inequalities.11 Cultural determinants of health such as connection to country (land and water), traditional practices and kinship systems promote resilience and support social and emotional wellbeing for Aboriginal peoples and communities.10,12 The COVID‐19 pandemic is likely to amplify the social determinants of health,13,14 and our concern is these determinants will continue to affect access to health care and increase health inequalities. Based on our own lived experiences and anecdotal community feedback, we are hearing that food security has increased for some Aboriginal people in response to COVID‐19. People are fearful of going into large shopping centres — fearful of catching COVID‐19. In some rural and remote areas, local shops are pushing up their prices, and people are left with no choice but to buy cheaper (and often less healthy) options to feed their families. Increase in government payments has resulted in the one and only shop in community providing food jamming their prices up. The price of food and water is beyond compare when you are paying $10 for a loaf of bread. Because of COVID‐19, people don’t want to come into town to do their shopping. (DL) We are concerned that restricted access to health care in response to border closures will impact the health and wellbeing of Aboriginal peoples. Some communities are being hit hard. To give a raw example, people are being refused medical treatment and are driving 600–800 km just to get any sort of medication or treatment around their health. (DL) We are also concerned that a lack of cultural safety displayed during COVID‐19 will lead to Aboriginal people being confronted with racism when trying to access health care.15 COVID‐19 has made accessing health care even more difficult Deciding to seek health care is difficult, and for some Aboriginal people, access to care has become more challenging during COVID‐19 with reduced availability of services. Many doctors and services have temporarily shut their doors to new patients, and this is likely to have a profound impact on people’s health. More generally, there have been efforts to overcome access challenges posed by COVID‐19 through the use of telehealth and virtual care. In our opinion, telehealth for diagnosis and e‐prescribing can be useful; however, there are challenges to using telehealth such as limited access to equipment and internet connection, and reluctance from some people to disclose personal information over a device. When we look at the provision of health care for our mob, one of the biggest barriers is having to sit in front of a computer. And talk to a computer, rather than a human connection. Our mob like to connect and have a yarn. (CP) Our view is that paying attention to the intersections of culture and diversity is essential to understanding the indirect impacts of COVID‐19. Within Aboriginal communities, there are minority groups who are significantly affected by COVID‐19. Minority groups include people with existing chronic conditions, people with disabilities, people experiencing homelessness, people living in rural and remote areas, and people who identify as lesbian, gay, bisexual, transgender, queer, asexual and questioning. Sistergirl and brotherboy are terms used for gender diverse people within some Aboriginal or Torres Strait Islander communities.16 If the mob aren’t receiving health related treatment, how this is feeding into direct or indirect impacts on disabilities. And how we can pick this up through the health system as disability is not in closing the gap. If we aren’t addressing it at a higher level, we are never going to address it at the ground level. (DL) We are also concerned about an increase in risk for our older people living with disability. These risks have been outlined by Aboriginal people with disability and their representative organisations, advocates and allies in international and national calls to action for governments to ensure Aboriginal disability‐inclusive public health, social and economic responses to the pandemic that put our mob at the forefront of any future planning in the health system.17 The recent drought, bushfires and now COVID‐19 are compounding risk factors for mental health issues and suicide. There is concern that some government measures to control the spread of COVID‐19 are triggering for mob — especially for those with trauma histories.18 We know mental health issues and suicide rates are high for our peoples,8,19 and we are concerned this level of disadvantage will worsen in response to COVID‐19. We support the recommendations made by the Centre of Best Practice in Aboriginal and Torres Strait Islander Suicide Prevention at the University of Western Australia to manage COVID‐19 recovery and address adverse impacts.19 The recommendations focus on the right to self‐determination, the health and mental health workforce, social and cultural determinants of health, digital and telehealth inclusion, and evaluation that includes Indigenous data sovereignty. These recommendations directly align with our lived experiences and were running themes throughout our discussions and overall assessment of the indirect impacts of COVID‐19 in our communities across NSW. Where to next? We prepared this article to inform future conversations on Aboriginal health issues in response to the COVID‐19 pandemic and beyond. Our view is that drawing on the lived experience and realities of Aboriginal peoples, taking firm action on the social determinants of health and working collaboratively with Aboriginal peoples and communities is the most effective way to address the indirect impacts of COVID‐19.

David Follent · Cory Paulson · Phillip Orcher · Barbara O'Neill · Debbie Lee · Karl Briscoe · Tara L Dimopoulos‐Bick

A 73-year-old man underwent a right total right knee replacement for osteoarthritis

Suran L Fernando · Marc J Capon · Sarah L Green · Michael J Boyle

Careful planning is required to ensure new resources for mental health lead to better consumer care Both the Productivity Commission into Mental Health1 and the Royal Commission into Victoria’s Mental Health System (RCVMHS)2 acknowledge that mental health services have been in the grip of a protracted resourcing drought. The RCVMHS identified that, until recently, Victoria’s public mental health services have fared particularly badly,3 with the 2015–2016 per capita funding to public mental health services being the lowest of any state. The impacts of chronic under‐resourcing, including a predominant focus on managing risk, underutilisation of evidence‐based therapies, and a lack of individualised care, were all reported by service users in testimonies highlighted in the RCVMHS interim report.2 After years of stagnation, in 2020 Victorian mental health services experienced some funding growth.4 These resources have allowed our team to strengthen existing services; bring on new staff, including expanding our lived experience workforce; and initiate new programs, including the Hospital Outreach Post‐Suicidal Engagement (HOPE) initiative, intensive community packages of care, the pre‐hospital response of mental health and paramedic team (PROMPT), and a mental health, alcohol and other drugs hub in our emergency department. In addition, our service is currently planning new mental health beds and rolling out an innovative hospital‐in‐the‐home program stemming from the recommendations of the RCVMHS interim report. We anticipate the final recommendations, due in February 2021, will bring even more new resources. Like sudden heavy rain on degraded soils after drought, such an inundation is welcome, but not without its own risks and challenges. Our recent implementation efforts have highlighted several challenges in managing rapid funding growth, including issues with human resources, leadership capacity, change management competency and stakeholder engagement, which will need consideration across the system to ensure services can translate funding into better consumer care. Human resources Delivering care requires staff who are difficult to find. There are insufficient mental health nurses5 and psychiatrists6 to fill current roles, particularly in regional and rural areas, and with an ageing workforce,7 problems with staffing are predicted to worsen.6 While international recruitment may assist, these processes come with lengthy delays and high administrative burdens. As a consequence, program implementation may be delayed due to recruitment challenges, or may adversely affect the operations of other service areas when clinicians shift between roles. Significant investment in training and recruitment pathways are required to prevent workforce shortages becoming a barrier to the pace of reform. Diversification of disciplines, grades and programs that encourage qualified staff to enter mental health care will all be required.1 Leadership capacity and competency in change management The presence of effective leadership and competency in change management principles are critical for successful health reform.8,9 Rapid growth can stretch existing leadership capacity. Health leaders already face significant challenges at system (eg, demographic changes, increasing demands, advancing technology), organisational (eg, human resources, changes to organisational structures and processes, intensification of frontline and middle management roles) and individual levels (eg, lack of role clarity, lack of training in managerial and leadership capabilities).10 For leaders already managing the challenge of daily operations, additional responsibilities to enact significant reform quickly carries a risk of overload. A key potential outcome of overload is loss of focus on the consumer.11 Therefore, as the pace of desired reform increases, a focus on developing current and future mental health leaders should be prioritised.2,12 Engaging stakeholders meaningfully Meaningful stakeholder engagement is crucial to successful change.9 As outlined by the RCVMHS, the “necessary changes to the mental health system cannot … be achieved by government alone”.2 The implementation of new programs requires engagement with diverse stakeholders including consumers, carers, staff, hospital executives, government and industrial bodies. Such stakeholders often have different interests, and forging and maintaining alignment is critical to progress. Meaningful engagement is vital to achieving this and subsequent success,9 but it is also resource intensive. Enacting multiple reforms quickly carries the risk that meaningful stakeholder engagement may be sacrificed. For this reason, timelines for delivery need to balance urgency with getting things right. The appropriate urgency for implementation articulated in the initial recommendations from the RCVMHS (eg, operation of an additional 170 acute mental health beds by mid‐2022) will only be achieved if those leading the change are supported and those affected by the change are included. Harnessing a generational opportunity for reform The RCVMHS calls for transformational change2 to our mental health system. The pace of this change and the other challenges involved must be managed carefully by system leaders to ensure that the intended reform occurs and results in provision of better care to the community.

Steven Moylan

Rapid advances in AI in ophthalmology are a harbinger of things to come for other fields of medicine The autonomous detection and triage of eye disease, or even accurate estimations of gender, age, and blood pressure from a simple retinal photo, may sound like the realms of science fiction, but advances in artificial intelligence (AI) have already made this a reality.1 Ophthalmology is at the vanguard of the development and clinical application of AI. Advances in the field may provide useful insights into the application of this technology in health care more broadly. Artificial intelligence Once described as the capacity of intelligent machines to imitate human intelligence and behaviour, AI now describes many theories and practices used to achieve computer intelligence (Box 1).2 Machine learning is an application of AI that uses algorithms or statistical models to make decisions or predictions. Complex patterns and relationships are learned from data to generate an outcome.2 Machine learning traditionally relies on the extraction of features from the data by human operators which then serve as input variables to optimise algorithm performance. The performance of these systems is constrained by the features that are recognised as important by humans. In contrast, artificial neural networks are an advanced method of machine learning able to extract features without explicit programming.2 Deep learning is the construction of multiple layers of artificial neural networks which can identify features in data that are not recognisable by humans. Although deep learning systems may be powerful, they lack human‐crafted inputs, meaning that large quantities of data are typically required to train algorithms. Artificial intelligence in ophthalmology As a discipline, ophthalmology is at the forefront of AI system development and translation in clinical practice. Leading uses of the technology include detecting, classifying and triaging a range of diseases, such as diabetic retinopathy, age‐related macular degeneration (AMD), glaucoma, retinopathy of prematurity, and retinal vein occlusion, from clinical images.3 The increasing global burden of eye diseases, coupled with the development of new therapies for previously untreatable conditions, has served as a major driver for AI innovation in ophthalmology. As a case in point, there are presently over 430 million people living with diabetes, most of whom require annual or biennial screening for retinopathy using retinal photography. This vast demand for diabetic eye screening services has stimulated the development of AI algorithms to identify sight‐threatening disease. Several algorithms have achieved performance that meets or exceeds that of human experts.4,5 Accordingly, in 2018, the United States Food and Drug Administration approved an AI system to detect referable diabetic retinopathy from retinal photographs, the first autonomous diagnostic system to be approved in any field of medicine.6 Advances in deep learning have extended to other imaging modalities that are commonly used in ophthalmology. Ocular coherence tomography is an imaging technology that produces highly detailed, depth‐resolved images of the retina. A recent collaboration between researchers and clinicians at Google DeepMind, Moorfields Eye Hospital and University College London culminated in the development of a deep learning system capable of detecting and triaging more than 50 different retinal conditions at levels equivalent to a panel of experienced ophthalmologists.7 AI systems with the capacity to detect a wide range of diseases, such as this, are likely to be most useful in clinical practice. A highly anticipated innovation is the development of AI systems capable of accurate disease prediction. Such tools could assist in managing patient expectations, improve the quality of care and reduce treatment costs.3 In ophthalmology, prediction models have been trained to personalise re‐treatment intervals for patients with neovascular AMD,8 predict progression from early to late AMD,9 estimate the extent of future visual field defects in patients with glaucoma,10 and predict diabetic retinopathy progression.11 Although these models presently achieve only moderate levels of accuracy, their performance has been shown to be superior to humans in several studies.3,8 Future advances in the accuracy of prediction models will likely come from the use of large longitudinal datasets drawing on multiple data sources, together with the development of more advanced AI systems.3 Despite these significant advances, AI systems are not in widespread clinical use and in some cases real‐world performance has been inferior compared with in silico validation.2,3 Training and validation of deep learning algorithms with large, representative data (eg, data from people of different ethnicities) acquired using multiple devices (eg, different retinal camera models) and data collection protocols (eg, retinal photographs acquired with and without pupil dilation) are key to achieving clinical applicability.4,5 This approach was used in the development of deep learning systems for retinal photographic screening for diabetic retinopathy, AMD and glaucoma which are now being used in large scale screening programs in Singapore and China.4,5 In these programs, AI is used to identify images without evidence of disease, so that human graders can focus their efforts on the images of those with disease, enabling improved efficiency and cost savings.12 Challenges to the clinical adoption of artificial intelligence Several obstacles to the adoption of AI in health care remain. The training of deep learning systems requires access to large amounts of medical data which has significant implications relating to privacy and data protection. In the context of ophthalmology, this is particularly pertinent, as the retinal vasculature may be considered biometric data, making it impossible to completely anonymise retinal photographs.3 Furthermore, characteristics that are not visible to human examiners, such as age and sex, can now be accurately predicted from a single retinal photograph using deep learning.1 Several recent major breaches of data protection laws relating to AI system development have already come to light.13 While individual patient data used to train an algorithm do not remain within the system, incorrect handling and sharing of data may lead to patients withdrawing consent to the use of their data under General Data Protection Regulation laws. It is not certain how data withdrawal requests will be dealt with when an individual’s data have been used in the process of training a deep learning system. Accordingly, developments in AI need to be accompanied by advanced data protection and security measures. Another challenge to the acceptance of deep learning algorithms in medicine is the difficulty in determining the basis for clinical decisions made by these systems, informally described as the “black box” problem. Visualisation tools have been developed to assist clinicians by highlighting the salient image features that contribute to the AI system classification (Box 2).12 This has the potential to create trust in system‐generated decisions, particularly if the features correspond with those used by experienced clinicians for clinical decision making.14 Interpretability is particularly important when considering legal liability in the event of patient harm arising from the use of AI in medicine. In traditional malpractice cases, a physician may be asked to justify the basis for a particular clinical decision and this is then considered in light of conventional medical practice.15 In comparison, challenges in identifying the basis for a given decision made by AI might pose problems for clinicians whose actions were based on that decision. The extent to which the clinician, as opposed to the technology manufacturer, should be held accountable for harm arising from AI use is a subject of intense debate.15 Factors such as the manner in which these AI systems are used and their classification as either products or software are likely to have important bearings on how cases are litigated.15 Further challenges for existing regulatory frameworks come from algorithms that continue to learn and evolve over time.15 Understanding how a given system is trained, its accuracy, and its operational limits is of great importance. Oversampling of a particular population or disease severity during training has the potential to introduce bias.4 Therefore, consideration of performance thresholds will help to inform appropriate use of AI systems. The Australian Government, through the CSIRO and Data61;16 the Australian Council of Learned Academies;17 the Australian Academy of Health and Medical Sciences;18 and specialty groups, such as the Royal Australian and New Zealand College of Radiologists,19 have made significant efforts to develop frameworks and policies for the effective and ethical development of AI. These consultative works have highlighted key priorities, including building a specialist AI workforce, ensuring effective data governance and enabling trust in AI through transparency and appropriate safety standards. Through targeted investment in research and development, Australia is aiming to advance its AI competitiveness. These framework documents provide guidance for developers, clinicians and health care consumers to navigate this rapidly evolving field. Broad dissemination of these documents should form part of a wider public engagement and education campaign to ensure that AI is developed and used in a considered and careful manner in health care. Rapid advances in AI in ophthalmology are a harbinger of things to come for other fields of medicine. While these technologies may eventually lead to more efficient, cost‐effective and safer health care, they are not a panacea in isolation. The successful integration of AI into health systems will need to first consider patient needs, ethical challenges and the performance limits of individual systems. Box 1 – Relationship between artificial intelligence and its subtypes Box 2 – Original retinal photograph of right eye with macular degeneration (A). Heat map of image A showing visualisation of traditional features associated with macular degeneration, such as central scarring (B). Original retinal photograph of left eye with referable diabetic retinopathy (C). Heat map of image C showing visualisation of traditional features, such as micro‐aneurysms and haemorrhages (D)

Jane Scheetz · Mingguang He · Peter Wijngaarden

The disability burden of non-fatal disease is not reflected in allocation of grants

Stephen E Gilbert · Rachelle Buchbinder · Ian A Harris · Christopher G Maher

Investigator‐initiated trials run by clinical trial networks provide net economic benefits to health systems Delivery of optimal health care relies on evidence from randomised clinical trials, among other factors, to inform best practice. While the generation of such evidence requires resources, both national and international assessments of health and economic benefits resulting from medical research indicate large returns on investment.1,2,3 In Australia, during the decade 2006–2015, more than 10 000 clinical trials were conducted through Australian clinical trials networks (CTNs), including more than 5 million participants, ranking Australia in the top tier of clinical trial activity.4 Industry‐funded clinical trials accounted for an estimated $930 million of the total $1.1 billion spent annually on clinical trials, with National Health and Medical Research Council (NHMRC) funding accounting for about $164 million annually.4 While the proportion of funding for non‐industry‐sponsored investigator‐initiated clinical trials (IITs) is relatively small, these studies account for more than half of Australia’s clinical trial activity.4 This study funding balance is similar to what is reported elsewhere.5 In Australia, IITs conducted by Australasian CTNs have had a major impact on the improvement of health care quality and outcomes around the world.6,7 IITs are designed and conducted by independent clinicians and academic researchers to generate clinical evidence to improve health care. Benefits are multilayered and not restricted to the discovery of new therapies. Much of the benefit comes from identifying and addressing uncertainty in existing practices; evaluating a range of treatment options that address key unanswered questions free of commercial imperatives, identifying alternative and potentially more efficient diagnostic strategies; and identifying more effective models of care or expensive interventions that are no more active than the lower cost alternative. Australasia has large, geographically dispersed CTNs across multiple clinical areas,8 with many more having been launched since the original report (personal communication Australian Clinical Trials Alliance [ACTA]). Between one‐quarter and one‐third of all Australian Government‐funded NHMRC support for clinical trials between 2004 and 2014 was awarded to IITs conducted by an established CTN.8 CTNs ensure clinically important, high priority and relevant research questions are appropriately conducted and provide efficiency through established infrastructure. Within Australasia, CTNs are widely regarded as key drivers of innovation and represent good value for public investment.8 Although the Australian Government invests in IITs and the CTNs that coordinate them, their value has not been well characterised. Governments are increasingly looking to systematically integrate activities that generate high quality evidence (such as IITs) with other aspects of the health care system (such as measurement of health outcomes or development of safety and quality policies) to build self‐improving, sustainable systems (Box). Understanding the potential return on investment is therefore paramount. In 2015, ACTA and the NHMRC profiled 37 established CTNs in Australia.8 Subsequently, a cost–benefit analysis for the profiled networks was calculated for those that i) were operational for more than 10 years; ii) had conducted more than five high impact peer‐reviewed IITs where an influence (or potential influence) on clinical practice and/or policy were identified (maturity); iii) received a significant proportion of funding from Australian funders (local investment); and iv) were available to participate (feasibility) in this analysis.9 Three CTNs that had conducted a total of 25 IITs were included in the analysis: the Australasian Stroke Trials Network (ASTN), the Interdisciplinary Maternal Perinatal Australasian Collaborative Trials (IMPACT) Network, and the Australian and New Zealand Intensive Care Society Clinical Trials Group (ANZICS CTG). Gross economic benefits across these CTNs were almost $2 billion, with the majority due to improvements in patient health outcomes ($1.4 billion), and 30% due to avoided health service costs — $453 million from the difference in outcomes and $127 million from differences in service costs. Gross costs, which included the cost of running the CTN, coordinating centre costs and the cost of running the entire IIT program in each CTN, were about $335 million, with most of those costs being for the IIT program itself (accounting for 73% of total costs). The benefit to cost ratio was 5.8:1 if findings from the 25 IITs were implemented in 65% of the eligible Australian population for one year.9 Similar findings have been reported internationally, with studies in the United States reporting a benefit to cost ratio of 4.2:1 over 10 years.3 In the United Kingdom, randomised clinical trials funded under the National Institute for Health Research health technology assessment program were expected to have a net benefit of £3 billion, with just 12% of this benefit required to cover the costs for all research undertaken.10 In the Australian analysis, funding provided to run a portfolio of IITs did not cover the total costs within either a CTN or at an individual IIT level, and in‐kind support was relied upon to make up the shortfall. The NHMRC funding received by all Australasian CTNs between 2004 and 2014 was represented by just 9% of the $2 billion gross benefit.9 The magnitude of avoided health care costs appears large, reflecting the size of health care expenditure. The Australian analysis highlighted the importance of in‐kind support within CTNs not only to sustain the viability of the CTNs but for their ability to conduct individual IITs.9 The total quantum of site level, in‐kind support could not be quantified accurately during the study. However, this support was described as being finite, at capacity in many instances, and at risk of exhaustion. From a sustainability perspective, the reliance on in‐kind support is concerning, and undermines the timeliness, volume and international competitiveness of clinical research in Australasia. Anecdotal evidence from interviews suggested that site level in‐kind support represents up to a 50% increase in trial funding. Late‐phase IITs conducted by CTNs deliver better health outcomes and health service value through a variety of mechanisms. Importantly, IITs play a critical role in addressing clinically significant questions, influencing guidelines, and identifying ways to improve safety and quality and opportunities for more efficient resource use. As stated in a scoping review, IITs “can also yield a substantial knowledge return on investment for hospitals and institutions that actively engage in trials, including the following: more skilled clinicians and increased research capacity, improved patient outcomes, and better health system performance. Also, the difference in cost of care for trial and non‐trial patients can be negligible”.11 Large increases in the benefit to cost ratio could be realised through relatively small increases in implementation rates. Research to identify the barriers and enablers of trial implementation should allow IITs to be translated more effectively into frontline health care delivery. But, intuitively, the conduct of potentially practice‐changing IITs through CTNs is likely to enhance implementation rates, as these virtual, nationwide consortia of clinicians are likely to involve a majority of the relevant clinical community. Hence, the reasonable assumption that clinicians who participate in IITs are more likely to implement trial results in their own practice and to translate new knowledge to their clinical colleagues. What we do not yet know is the extent to which IITs translate into routine practice. This is rarely measured or monitored in Australia. Measures of implementation should be incorporated routinely into IIT design, particularly for randomised clinical trials that are arguably more likely to result in clinically significant and potentially practice‐changing findings. Notwithstanding the clear economic benefit demonstrated for the 25 trials conducted by the selected group of three CTNs, it might be possible to reduce trial costs further. The overall cost of trials is a complex, multilayered issue, particularly as small pilot studies are often required to demonstrate the feasibility of recruitment. But combined with the push to answer key questions more quickly especially for the seriously or critically ill patients, such considerations have been drivers in implementing newer adaptive trial designs, which have flexible sample sizes that might reasonably be expected to reduce clinical trial costs.12 The analysis conducted of the three selected CTNs represents the first such analysis conducted of the role of CTNs in the Australian health sector. Despite the limitations of the analysis, it is clear further investment in existing CTNs, as well as therapeutic areas for which there are no CTNs at present, is warranted. This needs to be done in a manner that seeks operational efficiencies, including consolidation of infrastructure and the means to ensure engagement with geographically dispersed health services to improve patient access to trials across communities.11 In conclusion, there is potentially enormous, and arguably untapped, value in investing in IITs conducted by CTNs, as they provide net benefits to health care systems. However, the exact return on investment is contingent on the level of implementation. Further work in this regard is warranted. So, where to from here? High quality health systems are reliant on a strong clinical trials sector. In particular, the role of IITs run by CTNs is paramount in order to address clinically important questions, especially those that relate to health care variation. Clinical trial infrastructure needs to be strengthened, and we must endeavour to reduce reliance on in‐kind funding to ensure that the sector remains viable. Finally, we must strive to maximise implementation of trial findings to optimise current investment in the sector. Box – A self‐improving, sustainable health care system

The joint ACTA/ACSQHC Working Group

The Royal Commission into Aged Care Quality and Safety has highlighted the high rates of polypharmacy and potential medication‐related harm in residential aged care facilities (RACFs) in Australia.1 Residential medication management review (RMMR) is a government‐funded service for facilitating quality use of medicines in RACFs.2 Previous studies have found that RMMRs by accredited pharmacists and general practitioners identify a mean of 2.7–3.9 medication‐related problems per resident, and 45–84% of pharmacists’ recommendations were accepted by GPs.3 Guidelines recommend that residents should generally receive an RMMR on entering an RACF and when their clinical circumstances change,4 but annual claims data5,6 and recent research indicate that not all residents receive RMMRs.7 We examined time to first RMMR after RACF entry by analysing data for the national historical cohort of the Registry of Senior Australians (ROSA).7 In ROSA, de‐identified data collected during aged care eligibility assessments are linked to information about government‐subsidised aged care services, general practice and allied health services subsidised under the Medicare Benefits Schedule (MBS), medicines subsidised under the Pharmaceutical Benefits Scheme (PBS), and the Australian Institute of Health and Welfare National Death Index.8 Non‐Indigenous people aged 65 years or more who first entered permanent residential care during 1 January 2012 – 31 December 2015, had received an entry‐into‐care assessment within 100 days, and had received at least one PBS‐subsidised medication during the preceding year were included. Recipients of Department of Veterans’ Affairs‐funded services and people who had previously undergone RMMRs (eg, during transition care) were excluded. The cumulative incidence function was used to determine time to first MBS claim lodged by GPs for RMMRs (item code 903) or Home Medicines Reviews (HMRs) (item code 900) after entry to permanent residential care, adjusted for competing events (death, or permanent departure from the first RACF for another reason) using the Fine–Gray method,9 with follow‐up to 31 December 2016. Statistical analyses were undertaken in SAS 9.4. The University of South Australia (reference, 200489) and Australian Institute of Health and Welfare (reference, E02018/1/418) Human Research Ethics Committees provided ethics approval for the study. A total of 176 390 residents in 2799 RACFs were followed for a median 479 days (interquartile range [IQR], 149–858 days). Median age at entry was 84 years (IQR, 79–88 years), 108 908 were women (61.7%), and 84 864 were living with dementia (48.1%). In the year preceding entry, residents received a median of 11 unique prescription medications (IQR, 8–16 medications); 109 765 (62.2%) had received at least one high risk medication (as defined by the United States Institute for Safe Medication Practices10), and 7912 (4.5%) had received HMRs in the 12 months prior to RACF entry. By three months after RACF entry, 19.1% of residents (Wald 95% confidence interval [CI], 18.9–19.3%) had received RMMRs, 11.8% (95% CI, 11.6–11.9%) had died without RMMRs, and 5.7% (95% CI, 5.6–5.8%) had left their RACF for other reasons without RMMRs. At 12 months, 43.1% (95% CI, 42.8–43.3%) had received RMMRs, 20.6% (95% CI, 20.5–20.8%) had died without RMMRs, and 9.0% (95% CI, 8.8–9.1%) had left without receiving RMMRs. By 24 months, 49.7% (95% CI, 49.5–50.0%) had received RMMRs, 25.8% (95% CI, 25.6–26.0%) had died without RMMRs, and 10.2% (95% CI, 10.1–10.4%) had left their first RACF for other reasons without receiving RMMRs (Box). The high burden of medication use at the time of RACF entry suggests that most residents could have benefited from RMMRs, but MBS claims for RMMRs were lodged for fewer than one in five residents within three months of RACF entry, and fewer than one in two within two years. Our findings are generalisable to all older Australians entering RACFs, as ROSA captures data for all people aged 65 years or more who access government‐subsidised permanent residential aged care in Australia. We could not determine why residents were not referred for RMMRs, nor the impact of recent program changes2 on RMMR uptake and resident outcomes. In 2014–15, fewer GP medication review claims were reimbursed under the MBS (54 803 RMMRs, 63 872 HMRs) than pharmacist claims (93 517 RMMRs, 72 607 HMRs).5,6 Analysing GP claims may underestimate the number of RMMR reports prepared by pharmacists because GP claims are submitted after the medication management plan is discussed with the resident or family, while pharmacist claims are submitted after the report is sent to the GP.7 MBS claims may not be lodged if the full RMMR process cannot be completed (eg, because the resident died, their clinical circumstances had changed, or the RMMR report was not received or followed up), or claiming may be overlooked. Linkage with pharmacist claims data at the individual resident level could facilitate investigation of these limitations. Despite RMMRs being a key means for minimising medication‐related harm, MBS claims for RMMRs are lodged for only a fraction of residents who enter RACFs. The potential underuse of the program may be a missed opportunity for identifying and resolving medication‐related problems in Australian RACFs. Box – Stacked cumulative incidence function for time to first residential medication management review (RMMR), for first two years of permanent residential care* RACF = residential aged care facility. * For 176 390 residents (in 2799 residential aged facilities) included in the Registry of Senior Australians.8

Janet K Sluggett · J Simon Bell · Catherine Lang · Megan Corlis · Craig Whitehead · Steven L Wesselingh · Maria C Inacio