Spatial variation in Aboriginal and Torres Strait Islander people’s access to primary health care

The report shows that overall, Australian Government funded Indigenous-specific primary health-care services appear to be well positioned relative to the geographic distribution of Aboriginal and Torres Strait Islander people and to the distribution of other GP services. However, there are a number of areas where Aboriginal and Torres Strait Islander people have very limited access to both Indigenous-specific services and GP services in general.

Records access and management on closure of a medical practice

In July 2014, a Melbourne general practice made headlines when the Australian Privacy Commissioner found that it had breached the Privacy Act 1988 (Cwlth) by failing to properly secure patient medical records.1 The practice had relocated, leaving its records behind in a garden shed; when a thief broke in, the records became accessible to the public. The case raises the question of how patient medical records should be stored and transferred at the time of a practice closure or other change in operations, such as physical relocation or retirement of a practitioner. Significant changes to practice operations occur frequently in Australia due to corporatisation, relocation or retirement. This article outlines the legal and regulatory requirements that govern how patient records are to be managed in such circumstances.

Overlapping requirements

When a practice closes, its medical records are subject to enforceable legal, regulatory and professional standards and rules. In the Australian Capital Territory, New South Wales and Victoria, specific laws govern the management of medical and other practice records.2–5 In other jurisdictions, there are privacy and information management laws of general application, including the federal Australian Privacy Principles (which replaced the former National Privacy Principles on 12 March 2014). These overlapping laws necessitate compliance with the Australian Privacy Principles, as well as the relevant requirements of state or territory law. Relevant non-legislative regulation includes the code of conduct of the Medical Board of Australia (the Code),6 and the Royal Australian College of General Practitioners’ accreditation standards (notably, standard 4.2)7 and Handbook for the management of health information in private medical practice.8 Other sources that are outside the scope of this article include practices’ private contracts and contractual terms of service, such as for indemnity insurance and with medical centre service providers.

Regulations and laws

The definitions of medical records and patient health information are somewhat fluid. They can include consultation notes, correspondence, diagnostic referrals and results, patient-directed correspondence or information, financial and appointment records, as well as policies and procedures. Ambiguous or undocumented ownership of medical records and patient health information can create significant risk for practice owners and practitioners.

Professional standards regulation

Requirements for the disposal, transfer and access of records on closure or relocation of a medical practice place personal responsibilities on individual medical practitioners and duties on corporate medical practice entities. A recent high-profile case was clear in its finding that the practice’s administrative services provider was responsible for the maintenance and control of up-to-date records.9

The Code advises careful health information management at all times, particularly at practice closure. Its provisions guide the assessment of medical practice in cases of disciplinary action against registered medical practitioners.6 According to Skene, “provisions of the Code may have legal effect, despite not being legislation or case law”, because they provide normative standards by which good medical practice is defined and against which accusations of questionable medical practice might be judged by the relevant disciplinary tribunal or admitted as evidence in court proceedings.10

Three statements in the Code are relevant here. At 8.4, it states that “[m]aintaining clear and accurate medical records is essential” to prevent unauthorised access, facilitate patients’ right to access information contained in their records, and prompt facilitation of patient-initiated requests of health information transfer. Second, the Code at 3.13 stipulates that medical practitioners adequately inform patients of a decision to end a professional relationship; arrangements to ensure continuing care of the patient ought to “include passing on relevant clinical information”. This statement applies equally to the ending of a professional relationship in the context of the closure or transfer of a practice as to the end of a professional relationship that has become ineffective or compromised. The Code states at 3.15 that advance notice should be given with facilitation of continuity of care via the transfer of records at practice closure. In addition to these three requirements, medical practitioners must follow applicable health information law in their own jurisdictions, with the law taking precedence when the two conflict. In short, medical practitioners are obliged to comply with the Code’s stipulations that medical practitioners are personally responsible for their patients’ records.6

Privacy and practice management law

The ACT, NSW and Victoria each have specific laws that govern the closure of practices.2,4,5 The ACT’s requirements are the most stringent, stipulating in Principle 11 that health practices which close, merge or relocate must inform ACT Health, publish a notice in a daily newspaper and give each consumer at least 30 days’ notice.2 This requirement applies to each health service practice in the ACT, including the premises and operators where or through which a person provides health services.

In NSW, individual practitioners are personally responsible for breaches of the Health Practitioner Regulation (New South Wales) Regulation, while for medical corporations, a medical practitioner must be appointed personally responsible for the corporation’s record keeping. In Victoria, the Health Records Act requires health service providers to publish a notice of closure and an explanation of their records management in a locally circulating newspaper. No advance warning of closure is required, but practices must comply with guidelines issues by the Health Services Commissioner.11

Where a practice closes without an identified practice to which records will be transferred, the requirements for record retention that apply in each jurisdiction remain, as do requirements of the Code while a medical practitioner remains registered and thus subject to the Code.

In short, despite the overlapping and complex nature of requirements for medical record storage, management and disposal at times of practice closure or major business disruption, the implications for patient continuity of care and practitioner responsibility are significant. Primary Health Networks might be well placed to facilitate record transfer or custody in similar cases of practice closure without transfer arrangements in place.

Febuxostat-associated rhabdomyolysis in chronic renal failure

Clinical record

A 68-year-old man of European descent presented to our emergency department with rhabdomyolysis and acute-on-chronic kidney disease. He had a history of stage 3 chronic kidney disease (CKD3) — based on 17 tests in the year before this admission, his estimated glomerular filtration rate (eGFR) was 35 ± 7 mL/min/1.73m2 (CKD3: eGFR = 30–59 mL/min/1.73m2), his serum creatinine concentration was 179 ± 42 µmol/L (reference interval [RI], 60–120 µmol/L) — and of polyarticular tophaceous gout, type 1 diabetes mellitus, hypertension, ischaemic heart disease (coronary artery bypass grafting in 1998) and peripheral vascular disease.

His gout, diagnosed 13 years before this presentation, had been treated with allopurinol and then colchicine. Both, however, caused anaphylactic reactions. He had had multiple short courses of prednisolone to treat acute attacks on a background of naproxen. A month before this presentation, naproxen was withdrawn and treatment with a new hypouricaemic drug, febuxostat(40 mg daily) initiated. This drug was obtained through the Special Access Scheme of the Therapeutic Goods Administration. Two doses of febuxostat were withheld 12 days before this admission when he was admitted to hospital for 8 days with Haemophilus influenzae pneumonia and acute-on-chronic kidney disease, with his serum creatinine concentration peaking at 245 µmol/L. He was treated with intravenous ceftriaxone and azithromycin for 5 days, after which his renal function values had returned to baseline levels. He was discharged home and prescribed oral amoxicillin, which was to be continued for 5 days.

When he presented to our emergency department, the man was lethargic, oliguric, dehydrated and with acute-on-chronic kidney disease (serum creatinine concentration, 669 µmol/L; eGFR, 7 mL/min/1.73m2), but he was haemodynamically stable. There were no symptoms suggesting infection, and he was afebrile. Elevated serum creatine kinase activity (48 200 U/L; RI, 20–200 U/L) and the presence of myoglobin in his urine were consistent with rhabdomyolysis. The man was hyperkalaemic (potassium, 6 mmol/L; RI, 3.5–5.0 mmol/L) with tall T waves in his electrocardiogram; this responded to oral resonium and nebulised salbutamol. His serum creatinine concentration continued to rise despite intravenous hydration, peaking 48 hours after admission at 833 µmol/L (eGFR, 6 mL/min/1.73m2). His bicarbonate levels declined from 22 mmol/L to 15 mmol/L (RI, 24–31 mmol/L), reflecting worsening metabolic acidosis.

At the time of his admission, he was taking (in addition to febuxostat) aspirin (100 mg daily), simvastatin (40 mg daily), gemfibrozil (600 mg twice daily), frusemide (40 mg daily), metoprolol (100 mg twice daily), moxonidine (200 mg twice daily), insulin and omeprazole (20 mg daily).

Febuxostat was considered to be the likely dominant precipitating factor and was withdrawn, as were simvastatin and gemfibrozil; he had used these two medications for 12 years, but febuxostat had only recently been prescribed. Further, application of the Naranjo adverse drug reaction (ADR) probability scale1 indicated that febuxostat possibly caused rhabdomyolysis in our patient.

Haemodialysis of the patient was commenced, and he had five cycles over the next 11 days. His renal function gradually returned to baseline (Figure). He was discharged on Day 23. Some months later, treatment with benzbromarone was started, a uricosuric drug that was also obtained through the Special Access Scheme. After taking benzbromarone for one-and-a-half months, his plasma urate concentration was 0.26 mmol/L (RI, 0.25–0.50 mmol/L).

Uncontrolled gout is a significant disorder because of the debilitating attacks of acute gout, the ensuing joint damage that causes pain, deformity and loss of function, and the organ damage involved, particularly renal dysfunction. The xanthine oxidase inhibitor allopurinol is an effective medication for reducing plasma urate concentrations to below 0.36 mmol/L, and this effect, if maintained, will almost always eliminate recurrent acute attacks of gout and the risk of joint and organ damage.2–4 As the active form of the drug is cleared exclusively by the kidney, the starting dose needed by patients with impaired renal function is lower.4 Risk of hypersensitivity, manifested as toxic epidermal necrolysis, is an uncommon but significant problem; a macular-papular rash is seen in as many as 2% of patients. These problems may be more common in those with impaired renal function or also using diuretic medications.5

There are alternative approaches for lowering urate levels in patients with hypersensitivity reactions. Uricosuric drugs, such as probenecid, are effective in patients with normal to moderately impaired renal function, but are not effective in those with severe renal impairment.6 Benzbromarone is a more effective uricosuric agent in individuals with renal impairment, but must be imported into Australia under the Special Access Scheme, as it is not registered in this country.7

The availability of a second xanthine oxidase inhibitor, febuxostat, is therefore welcome, especially for patients who do not tolerate allopurinol well. Febuxostat is registered in the United States and Europe, and was recently also registered in Australia (December 2014). The medication is generally well tolerated and dose adjustment is not necessary in patients with mild to moderate renal dysfunction. Liver function abnormalities and cardiovascular thrombotic reactions have been identified by postmarketing studies, but their incidence is very low. Only a few case reports have described hypersensitivity reactions. Cost is an issue, as febuxostat is more expensive than allopurinol, and should therefore not be the agent of first choice.

Rhabdomyolysis is noted as a rare side effect in the product information for febuxostat (following post-marketing experience), but there has only been a single published case report.8 In our patient, a serious additional decline in renal function was marked by substantially elevated creatine kinase activity, suggesting that rhabdomyolysis had caused this decline. It is likely that the concomitant statin and fibrate hypolipidaemic medications (ie, simvastatin and gemfibrozil) that the patient had taken uneventfully for several years contributed to his myositis. Dehydration associated with the patient’s recent pneumonia probably also contributed to renal deterioration.

It is notable that in our case and that of Kang and colleagues,8 patients with chronic kidney disease had been taking statins both before and together with febuxostat. The combination of chronic kidney disease and statin therapy may represent a risk for febuxostat-induced rhabdomyolysis and renal injury.

The options for reducing this patient’s plasma urate levels — a critical goal, given his deteriorating renal function, progressive joint damage, and recurrent and severely painful acute gout — are quite limited. Febuxostat is now clearly contraindicated. Probenecid has limited efficacy when the glomerular filtration rate falls below 30 mL/minute. Benzbromarone, which is more effective than probenecid in patients with impaired renal function, can be obtained in Australia under the Special Access Scheme, and is a reasonable option. However, the patient’s hepatic function would need to be carefully monitored, as the medication was withdrawn from the market in Europe and North America after rare reports of serious hepatic toxicity.7 Recombinant forms of uricase (such as rasburicase, pegloticase) would undoubtedly be effective in reducing urate levels, but repeated use would be prohibitively expensive; further, there is a risk of developing antibodies to pegloticase, which results in reduced efficacy and the possibility of adverse effects.9 Desensitisation to allopurinol, although complex, time-consuming and not without risk, also remains an option.10

Lessons from practice

Options for reducing plasma urate levels to prevent recurrent acute and tophaceous gout are limited, especially in patients with impaired renal function.
Febuxostat is an effective alternative xanthine oxidase inhibitor to allopurinol.
Although the product information for febuxostat indicates that there is no need for dose adjustment in patients with moderate renal impairment, prescribers need to be cautious.
Chronic kidney disease and concomitant statin therapy may represent a risk for febuxostat-induced rhabdomyolysis and subsequent renal injury.

Figure

Time course of serum creatine kinase activity and estimated glomerular filtration rate (eGFR) in our patient. Arrows: Dialysis was undertaken on Days 3, 7, 10, 12 and 15.

Readmissions after stroke: linked data from the Australian Stroke Clinical Registry and hospital databases

Understanding the factors that contribute to hospital readmission for patients who have had acute stroke could improve outcomes for these people. Estimates of the frequency of readmission to hospital within the first year of onset vary widely from 13% to 62%, in part depending on whether readmission is for any cause or for a stroke-specific diagnosis.1,2 Hospital readmission is also frequent (36%–48%) after transient ischaemic attack (TIA).3,4 Predictors of hospital readmission after a stroke include older age, multiple comorbidities, diabetes mellitus, longer length of stay, physician specialty for the index admission, and size and type of hospital.5 To our knowledge, no study has specifically examined predictors of hospital readmission after TIA.

The Australian Stroke Clinical Registry (AuSCR) was established in 2009 to collect prospective, continuous, patient-level data on the quality of acute stroke care and patient outcomes (http://www.auscr.com.au).6 Linkage of AuSCR data to routinely collected hospital data can provide a powerful quality improvement tool, giving the ability to understand variations in care and predictors of important outcomes. These data can then be used to design interventions to reduce variations in care delivery and improve recovery after stroke.

We aimed to assess the feasibility of linking data from AuSCR to routine hospital datasets in Victoria, and to determine the frequency of and factors associated with hospital readmission during the year after acute stroke or TIA.

Methods

Datasets

Data from one Victorian hospital using AuSCR were linked to the Victorian Admitted Episodes Dataset (VAED) and Victorian Emergency Minimum Dataset (VEMD). AuSCR includes information on all admitted patients with stroke or TIA at participating hospitals, based on a minimum dataset of personal information (eg, name, address, age, sex, type of stroke), quality of care indicators (eg, treated in a stroke unit) and outcomes between 90 and 180 days after a stroke (eg, quality of life).6 The VAED contains morbidity data on all admitted patients in Victorian public and private acute hospitals and includes a wide range of demographic, administrative and some clinical variables (eg, International Classification of Diseases, 10th revision, Australian modification [ICD-10-AM] diagnosis codes). The VEMD includes similar details for people who are treated at any 24-hour emergency department (ED) of a public hospital in Victoria.

Ethics approval for this project was obtained from the ethics committee at the participating hospital and the data custodians (AuSCR and the Victorian Department of Health).

Data linkage

Data from AuSCR were linked with the VAED and VEMD for the period from 15 June 2009 to 31 December 2010. Follow-up of patients in AuSCR continued to 30 June 2011.

Data were linked using a two-stage separation principle, whereby identifying AuSCR variables for patients at the participating hospital were submitted to the Victorian Department of Health. These variables included last name and first three letters of the first name, date of birth, sex, postcode, partial Medicare number, and hospital admission unit record number (as Victorian hospitals have separate patient identification numbering systems, individuals can have multiple hospital unit record numbers). The Victorian Data Linkages unit performed stepwise deterministic linkage of the AuSCR data to the VEMD and VAED, with a 3-year look-back period. The de-identified, linked Department of Health data were then returned to the hospital principal investigator (H D), and one of us (M K) merged the content data from AuSCR with the de-identified dataset using a unique project identifier for each patient.

Outcomes

The primary outcome was all-cause hospital readmission within 30 days, 6 months and 1 year from time of hospital discharge after the index admission for stroke or TIA (ie, the first registered event in AuSCR). Hospital readmission was defined as an admission to an acute care hospital in Victoria for any reason. All primary diagnoses recorded for presentation to an ED or hospital discharge were categorised using ICD-10-AM definitions.

Patient characteristics, social circumstances, health system factors, clinical processes of care and health outcomes derived from the datasets were compared by hospital readmission status, using the categories outlined elsewhere.7 Patient characteristics (eg, age, sex, place of birth), clinical processes of care (eg, admission to an acute stroke unit, use of thrombolysis) and health outcomes (eg, discharge destination) were obtained from AuSCR. Comorbidities were obtained from the VAED using ICD-10-AM codes (Appendix). The Charlson comorbidity index (CCI) score,8 which uses 19 conditions or diseases as a prognostic marker for poor outcome,9 was calculated from VAED data. Health system data were obtained from AuSCR (eg, length of stay), the VAED (eg, number of admissions before index event, prior stroke or TIA admission) and the VEMD (eg, number of ED presentations before index event, prior stroke or TIA presentation).

Statistical analysis

We used descriptive statistics to compare patients according to hospital readmission status, using the χ2 test for categorical variables and the Wilcoxon–Mann–Whitney rank-sum test for continuous variables. We used Nelson–Aalen cumulative hazard estimate curves to illustrate the timing of hospital readmission by stroke subtype.

Multivariable logistic regression models were used to explore factors associated with hospital readmission, which was defined as the dependent variable. The measure of stroke severity included whether a patient was able to walk on admission. Other independent variables were selected if they were statistically significant in univariable analyses, using P < 0.1 as the threshold. Assessments for collinearity were made and a condition index of 10–15 was considered acceptable.10 Multivariable results are reported as adjusted odds ratios (aORs) with 95% confidence intervals. Significance was set at P < 0.05. All analyses were undertaken with Stata (version 12.1, StataCorp).

Results

Of 788 patients registered in the AuSCR, 658 (83%) had a stroke (81% [534/658] ischaemic stroke; 18% [117/658] intracerebral haemorrhage; 1% [7/658] undetermined type) and 130 (17%) had a TIA. Their median age was 76 years (interquartile range, 66–84 years), 46% (359/781) were female, and 58% (427/738) were born in Australia. Of the AuSCR registrants, 655 (83%) were recorded as having their first-ever stroke or TIA event, while 133 (17%) had previously had a stroke or TIA.

The availability of the linkage variables between datasets was excellent (Box 1). AuSCR data were linked to the VAED or VEMD over three iterations. As records were matched, they were removed from the source datasets (VAED or VEMD). The final overall matched linkage achieved was 93% of AuSCR registrants in the VAED (736/788) and the VEMD (731/788). Of the 788 AuSCR registrants, 782 (99%) were linked to at least one of the VAED and VEMD (Box 1).

Fifteen per cent of patients (108/715) had an all-cause hospital readmission within 30 days. Readmissions increased to 36% (247/694) within 6 months and 42% (291/685) within 1 year. Diseases of the circulatory system were the most common reason for hospital readmission within 1 year, occurring in 20% of patients (56/286), including stroke or TIA in 12% (35/286) (Box 2). Patients with an index TIA were more likely than patients with stroke diagnoses to be readmitted within 1 year (Box 3).

Many patient characteristics were similar for those with and without hospital readmission within 1 year (Box 4). However, readmitted patients were more likely to have stroke risk factors such as hypercholesterolaemia or diabetes mellitus and greater overall comorbidity as defined by their CCI scores.

Differences in health system, clinical care and health outcome factors between patients who were and were not readmitted within 1 year are shown in Box 5. Most patients were managed in a stroke unit for their index event, and clinical processes of care between those with and without hospital readmission were consistent. However, readmitted patients were more likely to have had more ED presentations before their index event, compared with patients who were not readmitted.

Results of multivariable analyses for each time period are shown in Box 6. The factors that remained significantly associated with hospital readmission within 1 year were ≥ 2 ED presentations before the index event, a higher CCI score, and TIA being the reason for the index hospitalisation. The same factors were associated with hospital readmission within 6 months. Higher CCI score and multiple ED presentations were associated with readmission within 30 days.

Discussion

Australian data on factors related to hospital readmission for patients with stroke or TIA are limited. We found that data linkage between the AuSCR and routine hospital datasets was feasible and can identify determinants of hospital readmission for patients who have had stroke or TIA.6,11

We found that patients with multiple ED presentations before their initial hospitalisation were more likely to be readmitted to hospital over the next year than those with fewer than two ED presentations. To our knowledge, there are no other studies with similar analyses that have explored factors associated with hospital readmission after stroke or TIA.

Our findings are consistent with data from the United States,4 where a study involving 2802 patients found that those discharged from hospital after a TIA had a greater risk of readmission within 1 year compared with patients with ischaemic stroke (hazard ratio, 1.20; 95% CI, 1.02–1.42). However, as the reason for hospital readmission was based on self-report by either patient or proxy,4 there are some concerns about the reliability of these data.12

A strength of our study is the use of ICD-10-AM coding to categorise diagnoses and outcomes, which showed that diseases of the circulatory system were the most common reason for readmission. There were no differences in the causes of readmission for patients discharged with ischaemic stroke and those with other stroke or TIA diagnoses, but the proportion of readmissions was lower than in the US study.4 Overall, we found that 42% of patients who survived an initial hospitalisation for stroke or TIA were readmitted within the first year, which falls within the range previously reported (13%–62%).1,2,13–21 However, only 12% of patients were readmitted due to another stroke or TIA. This result is similar to that from an Australian study that used data from 1075 patients in the Hunter Area Heart and Stroke Register (13%),1 but lower than in a US study of 1818 veterans with stroke, which used multiple health care plan data sources (31%).2

In our adjusted analyses, we found that increased frequency of comorbid conditions, as measured by the CCI, was independently associated with readmission within 1 year. This differs from the Hunter Area Heart and Stroke Register study, which found that people presenting with stroke had an increased number of comorbidities such as hypertension (38%), but found no association of CCI with readmission within 1 year.1 However, our results are similar to other studies which found that patients with more comorbidities (CCI score ≥ 3)13 or higher comorbidity summary scores2 were more likely to be readmitted within 1 year.

A major limitation of our study is that the data were derived from only one hospital, resulting in a small sample size (788) compared with other data linkage studies (≥ 16 000).22 This also meant that the clinical processes of care received during hospitalisation were similar for the included patients, as 98% were admitted to the same stroke unit. Frequency of readmission within 1 year and processes of care are likely to differ between health services, and our findings may not be generalisable to other health services. Further, coding quality in other hospitals was not assessed. Readmissions to hospitals outside Victoria were not captured in our study, but we believe this number would be small.

Nevertheless, we have shown that linkage of AuSCR data with routinely collected hospital data is feasible. A larger ongoing study (Stroke12311) will assess cross-jurisdictional data linkage involving over 17 000 Australian stroke patients registered in AuSCR between 2009 and 2013. These linked data will provide a richer data source across a broader range of hospitals and locations for validating our preliminary findings of the frequency and determinants of hospital readmission after stroke or TIA.

1 Data linkage iterations*

Medicare8 = partial Medicare number. MedSuf = first three letters of first name. DOB = date of birth. VEMD = Victorian Emergency Minimum Dataset. URNo = unit record number. VAED = Victorian Admitted Episodes Dataset. * Availability of linkage variables from datasets: URNo, 100%; DOB, 99%; Sex, 99%–100%; Medicare8, 90%–92%; MedSuf, 90%–92%.

2 Primary diagnosis recorded for first readmission within 1 year,* by stroke subtype

		Stroke subtype
Cause of readmission (ICD-10-AM codes)	Total (n = 286)†	ICH (n = 37)	IS (n = 184)	TIA (n = 65)

Certain infectious and parasitic diseases (A00–B99)	7 (2.5%)	0	4 (2.2%)	3 (4.6%)
Neoplasms (C00–D49)	11 (3.9%)	1 (2.7%)	9 (4.9%)	1 (1.5%)
Diseases of the blood and blood-forming organs (D50–D89)	7 (2.4%)	0	7 (3.8%)	0
Endocrine, nutritional and metabolic diseases (E00–E89)	9 (3.2%)	1 (2.7%)	8 (4.3%)	0
Mental, behavioural and neurodevelopmental (F01–F99)	5 (1.8%)	1 (2.7%)	3 (1.6%)	1 (1.5%)
Diseases of the nervous system (G00–G99)	21 (7.3%)	2 (5.4%)	8 (4.4%)	11 (16.9%)‡
Diseases of the eye and adnexa (H00–H59)	11 (3.9%)	1 (2.7%)	7 (3.8%)	3 (4.6%)
Diseases of the ear and mastoid process (H60–H95)	1 (0.4%)	0	1 (0.5%)	0
Diseases of the circulatory system (I00–I99)	56 (19.6%)	10 (27.0%)	30 (16.3%)	16 (24.6%)
Cerebrovascular disorders	40 (13.9%)	5 (13.5%)	25 (13.6%)	10 (15.4%)
Stroke	25 (8.7%)	4 (10.8%)	17 (9.2%)	4 (6.2%)
TIA	10 (3.5%)	0	5 (2.7%)	5 (7.7%)
Heart failure	6 (2.1%)	0	3 (1.6%)	3 (4.6%)
Myocardial infarction	1 (0.4%)	0	1 (0.5%)	0
Atrial fibrillation	5 (1.8%)	0	2 (1.1%)	3 (4.6%)
Diseases of the respiratory system (J00–J99)	10 (3.5%)	1 (2.7%)	9 (4.9%)	0
Chronic pulmonary disease	3 (1.1%)	1 (2.7%)	2 (1.1%)	0
Diseases of the digestive system (K00–K95)	21 (7.3%)	1 (2.7%)	14 (7.6%)	6 (9.2%)
Diseases of the skin and subcutaneous tissue (L00–L99)	1 (0.4%)	0	1 (0.4%)	0
Diseases of the musculoskeletal system (M00–M99)	10 (3.5%)	1 (2.7%)	7 (3.8%)	2 (3.1%)
Diseases of the genitourinary system (N00–N99)	11 (3.9%)	2 (5.4%)	7 (3.8%)	2 (3.1%)
Congenital malformations and chromosomal (Q00–Q99)	5 (1.8%)	1 (2.7%)	3 (1.6%)	1 (1.5%)
Symptoms, signs and abnormal clinical (R00–R99)	44 (15.4%)	8 (21.6%)	28 (15.2%)	8 (12.3%)
Injury, poisoning and other consequences (S00–T88)	23 (8.0%)	0	17 (9.2%)	6 (9.2%)
Factors influencing health status and services (Z00–Z99)	21 (7.3%)	4 (10.8%)	15 (8.2%)	2 (3.1%)

ICD-10-AM = International Classification of Diseases, 10th revision, Australian modification. ICH = intracerebral haemorrhage. IS = ischaemic stroke. TIA = transient ischaemic attack. * Patients may have had more than one readmission, but only the primary diagnosis for the first readmission is shown. † Excludes five patients with stroke of undetermined type. ‡ P < 0.05.

3 Nelson–Aalen cumulative hazard estimate curves showing readmission in the first year after discharge from hospital for an index event

4 Comparison of patient characteristics by hospital readmission within 1 year

Characteristic (data source)

Readmitted (n = 291)

Not readmitted (n = 394)

Patient characteristics (AuSCR)

Median age, years (IQR)

76 (65–83)

75 (65–82)

0.68

Female

135/288 (46.9%)

166/390 (42.6%)

0.26

Born in Australia

158/267 (59.2%)

205/371 (55.3%)

0.55

Aboriginal and/or Torres Strait Islander

3/289 (1.0%)

3/392 (0.8%)

0.10

English spoken

242/289 (83.7%)

329/388 (84.8%)

0.68

Documented evidence of previous stroke

52/291 (17.9%)

55/394 (13.9%)

0.37

Pre-existing conditions (VAED)*

Atrial fibrillation

82/291 (28.2%)

83/344 (24.1%)

0.25

Hypercholesterolaemia

46/291 (15.8%)

30/344 (8.7%)

0.006

Hypertension

187/291 (64.3%)

206/344 (59.9%)

0.26

Diabetes

42/291 (14.4%)

14/344 (4.1%)

< 0.001

Angina

20/291 (6.9%)

13/344 (3.8%)

0.08

Smoking (current)

38/291 (13.1%)

49/344 (14.2%)

0.67

Obesity

12/291 (4.1%)

7/344 (2.0%)

0.12

Peripheral vascular disease

11/291 (3.8%)

3/344 (0.9%)

0.01

Congestive heart failure

32/291 (11.0%)

24/344 (6.9%)

0.08

Renal disease

42/291 (14.4%)

26/344 (7.6%)

0.005

Dementia

18/291 (6.2%)

24/344 (6.9%)

0.69

Mean Charlson comorbidity index score (SD)

2.8 (1.9)

2.4 (1.5)

0.02

Type of stroke (AuSCR)

Intracerebral haemorrhage

37/291 (12.7%)

42/394 (10.6%)

0.40

Ischaemic stroke

184/291 (63.2%)

286/394 (72.6%)

0.009

Transient ischaemic attack

65/291 (22.3%)

64/394 (16.2%)

0.04

Undetermined

5/291 (1.7%)

2/394 (0.5%)

0.12

Stroke severity variables (AuSCR)

Able to walk on admission

83/236 (35.2%)

125/358 (34.9%)

0.95

Cause of stroke (known)

115/291 (39.5%)

166/394 (42.1%)

0.49

Social circumstances (VAED)

Married or with partner before admission

186/291 (63.9%)

226/370 (61.1%)

0.46

Private patient in public hospital

92/291 (31.6%)

109/370 (29.5%)

0.55

AuSCR = Australian Stroke Clinical Registry. IQR = interquartile range. VAED = Victorian Admitted Episodes Dataset. VEMD = Victorian Emergency Minimum Dataset. * According to International Classification of Diseases, 10th revision, Australian modification codes, before and at index event.

5 Comparison of health system, clinical care and health outcome factors by hospital readmission within 1 year

Factor (data source)

Readmitted (n = 291)

Not readmitted (n = 394)

Health system (AuSCR)

Median length of hospital admission, days (IQR)

5 (2–10)

5 (3–10)

0.55

Transfer from another hospital

20/291 (6.9%)

14/394 (3.6%)

0.11

Stroke occurred while in hospital for another condition

6/291 (2.1%)

2/394 (0.5%)

0.09

Health system (VAED and VEMD)

Mean emergency presentations (SD)*

0.9 (1.7)

0.66 (1.1)

0.05

Two or more emergency presentations*

67/291 (23.0%)

66/394 (16.8%)

0.04

Two or more admissions*

94/291 (32.3%)

107/394 (27.1%)

0.14

Stroke emergency presentation*

14/291 (4.9%)

10/394 (2.5%)

0.13

TIA emergency presentation*

7/291 (2.4%)

8/394 (2.0%)

0.28

Stroke or TIA admission*

15/164 (9.2%)

13/158 (8.2%)

0.77

Clinical processes of care (AuSCR)

Admitted to a stroke unit

284/290 (97.9%)

388/393 (98.7%)

0.41

Received thrombolysis

22/184 (11.9%)

40/283 (14.1%)

0.49

Taking an antihypertensive agent at discharge

250/286 (87.4%)

331/380 (87.1%)

0.60

Received a care plan at discharge

56/286 (19.6%)

69/379 (18.2%)

0.07

Health outcomes (AuSCR)

Discharged to home

135/286 (47.2%)

160/378 (42.3%)

0.21

Discharged to aged care facility

14/286 (4.9%)

25/378 (6.6%)

0.35

Discharged to inpatient rehabilitation setting

116/286 (40.6%)

170/378 (44.9%)

0.26

AuSCR = Australian Stroke Clinical Registry. IQR = interquartile range. VAED = Victorian Admitted Episodes Dataset. VEMD = Victorian Emergency Minimum Dataset. TIA = transient ischaemic attack. * Before index event.

6 Factors associated with all-cause hospital readmission within 30 days, 6 months and 1 year after stroke or transient ischaemic attack (TIA)

Adjusted odds ratio (95% CI)*

Factor

30 days

6 months

1 year

Female

1.31 (0.83–2.07)

1.22 (0.85–1.74)

1.14 (0.80–1.62)

Higher Charlson comorbidity index score

1.25 (1.11–1.42)†

1.19 (1.07–1.32)†

Able to walk on admission

1.46 (0.85–2.52)

1.09 (0.72–1.67)

1.04 (0.69–1.56)

Documented evidence of previous stroke

0.89 (0.61–1.33)

1.01 (0.80–1.27)

1.01 (0.81–1.27)

TIA on index admission

1.77 (0.93–3.40)

1.98 (1.19–3.30)†

2.15 (1.30–3.56)†

Two or more emergency presentations before index event

2.09 (1.25–3.50)†

1.69 (1.09–2.61)†

1.57 (1.02–2.43)†

* Adjusted for all factors shown. † P < 0.05.

Geographical mobility of general practitioners in rural Australia

Key to improving the poorer health status that characterises people in rural areas is ensuring equitable access to appropriate health care.1–3 However, this requires recruiting and retaining an adequate supply of appropriate health workers, which is known to be difficult in rural and remote areas.4,5 While considerable research has been conducted on the factors and barriers that facilitate and impede medical workforce supply in rural areas, there is a dearth of quantitative empirical evidence relating to the dynamics of general practitioner mobility patterns — specifically, which doctors move where, at what frequency, and why.

Understanding GP mobility is important because of its impact on workforce availability — both in the origin area (place from which the doctor moved) and the destination area. Considerable investment is made by governments into health programs specifically oriented towards improving the recruitment and retention of doctors in rural areas, with the goal of maximising movement into and minimising movement away from rural areas.

Despite a large body of social sciences literature on both inter- and intraregional migration, its applicability to the health workforce is not clear. Unfortunately, research literature focusing specifically on medical and health workforce mobility is scant. Internationally, sentinel works related to doctors include one 20-year national study in the United States on the volume and location of rural moves, although covariate analysis was not reported.6 Subsequent publications from the same dataset have focused on mobility in and out of areas of high need and between four major regions of the US.7,8 Similarly, a few Canadian studies have focused specifically on interregional (large distance) migration patterns of doctors without a focus on rural areas per se.9,10 Much of the extant health mobility literature has concentrated on the international migration of doctors from developing to developed countries,11 focusing particularly on ethical issues, the impact of the loss of doctors on reduced access to health care in origin countries, and the roles of these international medical graduates (IMGs) in destination countries.12–14

Associations between mobility and covariates have rarely been quantified,9,15,16 with younger age being the dominant common factor linked with increased mobility. Much less has been written about the mobility of Australian doctors,17,18 and no specific mobility data have been published for the non-medical health workforce. The reasons for this lack of literature include limited access to data at a suitable geographical scale; lack of longitudinal studies from which to monitor doctors’ movements; inherent difficulties of tracking individual doctors without linked datasets; and insufficient numbers of moves to generate valid and reliable results.

In an attempt to redress this paucity of evidence, we aimed to describe the geographical mobility of GPs in Australia both within rural areas and between rural and metropolitan areas. We describe where doctors are moving to and from, how many doctors are moving, and the characteristics of doctors who move. Such research helps to provide the basis for better understanding the role of push and pull factors behind why doctors move and what has influenced their decision to move. This in turn assists policymakers to design policies targeting medical workforce maldistribution in rural and remote areas.

Methods

We used data from the large Medicine in Australia: Balancing Employment and Life (MABEL) survey, conducted within the Centre for Research Excellence in Medical Workforce Dynamics. MABEL is Australia’s national longitudinal survey of doctors, which collects similar data in annual waves from mostly the same panel of doctors (https://mabel.org.au). MABEL was approved by the University of Melbourne Faculty of Business and Economics Human Ethics Advisory Group (Reference 0709559) and the Monash University Standing Committee on Ethics in Research Involving Humans (Reference CF07/1102 – 2007000291).

Study participants

The first wave of the MABEL survey, in 2008, invited the participation of the entire medical workforce, and 3906 GPs (19% of Australia’s GP workforce) completed the initial survey. Subsequent annual waves of previous respondents saw a 70%–80% retention rate, including the annual addition of new GPs to the dataset and returning participants who missed at least 1 year. This study used data from waves 1–5 (2008 to 2012), comprising 3502 (wave 2), 3514 (wave 3), 3287 (wave 4) and 3361 (wave 5) responses. Detailed non-response bias was conducted for waves 1 and 2.19,20 The most notable observable bias was a significant increase in the number of responses from doctors in remote areas, attributable to a financial incentive ($100 honorarium) to maximise participation of these GPs. GP registrars were excluded, because many do not have autonomy over their work location during their fellowship training.

Locational measures

Locational data were geocoded to a specific town or suburb. Each GP’s self-reported work location (< 1% missing data) was used to calculate mobility, by comparing their location between each annual wave. Mobility was classified using the seven-category Modified Monash Model scale,21 which combines population size of settlements (< 5000; 5000–15 000; >15 000–50 000; and > 50 000) with the Australian Statistical Geography Standard — Remoteness Areas (ASGS-RA) classification, to define a geographical classification of most relevance to Australian GPs.22 Locational changes within the same rural town or within metropolitan areas were all classified as “no change”.

Statistical analysis

Analysis was conducted in two distinct parts. First, all GP respondents were analysed using only their origin location and destination location, aggregated using the Modified Monash Model. GPs who participated in all five waves thus contributed four origin–destination pairs. Second, mobility outcomes were assessed for their association with additional key covariates of age (< 40, 40–54, ≥ 55 years), sex, having a life partner, IMG status and location restrictions as part of their registration, business relationship within the practice, and length of stay in that location. For this analysis, mobility was categorised as no change (metropolitan), no change (rural), change from rural to metropolitan, change from metropolitan to rural and change within rural areas (where “rural” encompasses all six non-metropolitan categories, from regional to very remote). Annual “risk” of moving between rural and metropolitan locations was measured using total number of observed years. Panel (clustered) logit models were additionally used to measure the associations between these risk factors and either leaving rural areas (models 1 and 2) or leaving metropolitan areas (models 3 and 4). Length of stay was removed from models 2 and 4 because of its strong multicollinearity. All calculations were performed using StataSE 12 (StataCorp) with a 5% significance level.

Results

Between wave 1 and wave 5, a total of 5844 GPs completed at least one MABEL survey. Of these, 1810 GPs completed all five waves, providing 7240 mobility observations. A further 805 GPs missed one survey (2415 mobility observations), 786 completed three out of five waves (1572 observations) and 887 completed only two waves. Additionally, 1470 GPs completed only one MABEL survey, contributing no mobility data. In total, there were 12 114 mobility observations, which decreased to 10 900 after GP registrars were removed from the dataset.

Overall, fully trained GPs were observed to have a mobility rate of about 4.6% (507/10 900). In comparison, GP registrars had a mobility rate of 21% (253/1214).

Box 1 summarises the number of locational changes for the five waves (2008–2012). Cells along the main diagonal represent GPs who did not change their location between waves. This approximation of retention within each category shows a decreasing rate as the degree of geographical rurality or remoteness increases. In cells to the right of the diagonal, the destination location is increasingly remote compared with the origin location (eg, large rural to remote), and cells to the left of the diagonal capture GPs who have moved to decreasingly remote locations (eg, small rural to medium rural). The first row captures all GPs (103) who moved from a metropolitan origin to a non-metropolitan destination during the five-wave period; 98.6% (7015) stayed within a metropolitan location. The first column of Box 1 captures all GPs who moved from a regional, rural or remote origin to a metropolitan destination (133). Just under half of all observed location changes were between non-metropolitan and metropolitan locations (236). There were 478 observations originating from a remote or very remote location, with 417 (87%) GPs remaining within the same location in the next year, and most (45 [74%]) of the 61 movers remaining within a non-metropolitan area.

Aggregate counts of GPs and the characteristics of the movers and stayers are summarised in Box 2. The observed risk (per observed year) of moving to a non-metropolitan area was 1 in 75 for metropolitan GPs. In contrast, the risk of losing non-metropolitan GPs to metropolitan areas was 1 in 31. Of the 271 GPs who moved within non-metropolitan Australia, 77 moved to regional centres (population over 50 000), but only 24 left regional centres for a smaller rural or remote location. A further 18 GPs moved from a rural to a remote location and 35 moved from remote practice to small or large rural locations.

Box 2 also shows the characteristics of GPs who moved compared with GPs who stayed in their original location. There was a small increased risk of moving for the youngest group of GPs, while sex and having a life partner had minimal association with increased mobility. IMGs had an increased risk of moving; even more so for the subgroup who were restricted in their location choice. GPs who were also principals of their practice were much less likely to move, while contract employees were highly mobile away from regional, rural and remote areas.

Box 3 shows the association between observed significant location changes and GP characteristics, with two binary outcomes tested (leaving rural and leaving metropolitan practice). Younger rural GPs were significantly more likely to leave rural practice than older rural GPs. There were no other significant associations between GP mobility and age or between GP mobility and sex and family status. The risk of moving to a metropolitan area was 2.5 to three times higher for rural GPs in their first 3 years in a location than for those who had been in a location for 4 years or more. Both contract employees and salaried employees were highly likely to leave rural practice, while salaried employees were most likely to leave metropolitan areas. Compared with GPs in regional centres, those in small and medium rural towns were significantly more likely to leave rural practice, while GPs in very remote areas had a lower risk of moving to a metropolitan area. The omission of length of stay strengthened the mobility odds ratio of all employment types, compared with practice principal or partners, and the association between small population size and an increased risk of turnover in rural areas. Additionally, IMGs restricted in their practice location had a higher risk of moving than Australian-trained, unrestricted GPs.

Discussion

This study provides the first national evidence of rural GP mobility over an extended period. Moreover, we investigated whether individual- and practice-level covariates were associated with the propensity to move. We used the seven-category Modified Monash Model to show which groups of GPs exhibit the highest mobility and are most at risk of leaving rural practice, or most likely to leave metropolitan areas for rural practice.

GPs in small rural towns and remote areas had higher mobility rates. While remote and very remote GPs had the highest mobility rate, this group was not significantly at increased risk of leaving non-metropolitan practice completely. Rural GPs practising in small towns (less than 5000 residents) and in medium-sized towns (up to 15 000 residents) were most at risk of moving to metropolitan areas. These results further support the need for policies to better target GPs in small rural communities and differentiate them from GPs in large regional centres.21–23

GPs most at risk of moving, both from and to rural areas, are those who have only been in their current location for up to 3 years, similar to recent findings in rural New South Wales.24 That is, once a GP has been settled in either a rural or metropolitan location for at least 3 years they are less likely to move.

Additionally, younger GPs (under 40 years) and those working as either salaried or contract employees are more likely to be mobile. Sex and family status were not associated with mobility.

When more data become available, we plan to investigate whether there is any association between mobility and GPs’ satisfaction with the schools that are available for their dependents.

Unrestricted rural IMGs had a slightly but non-significantly increased risk of leaving rural practice compared with locally trained unrestricted GPs. Further investigation of the strength of association between mobility and changed restriction (overseas trained) or bonding (Australian trained) status is also planned.

Our study was strengthened by the removal of GP registrars. GP registrars frequently have minimal control over their training locations, and so their moves are not equivalent to observed moves of GPs who have independently chosen to practise in a specific location. GP registrars are highly mobile, both between metropolitan and rural areas as well as between different rural areas. In this study, GP registrars were observed to have a mobility rate about five times higher than the annual mobility rate for fully trained GPs.

The main limitation of this study was our use of a self-selected cohort. Annually, the MABEL survey includes about 16%–19% of all Australian GPs, with 75%–80% of participants returning each year (potential selection or attrition bias). Despite having five annual waves of data, the number of GPs observed changing work location was still relatively small. In total, only 236 GPs were observed moving in either direction between rural and metropolitan areas. Larger numbers of moves are observable using more complete datasets like the Australian Health Practitioner Regulation Agency dataset,18 but this approach only provides very limited covariate information and includes large bias from the highly mobile GP registrar subgroup. The true mobility rate of rural–urban relocations for Australia’s GP population may be different to the 3.2% (rural origin) and 1.3% (urban origin) observed annual rates in our study. More observed moves and sophisticated panel data analysis will be possible as additional MABEL data become available.

Increasing workforce supply and maintaining the existing rural medical workforce remains a key health issue in improving rural health in Australia. For several decades now, the Australian Government has made considerable investment in training more medical graduates, exposing these new doctors to more rural experience during their training, and increasing GP fellowships, rural bonded scholarships and rural retention payments.25 Nonetheless, little longitudinal evidence exists on how long to expect GPs to remain in different locations, what locations GPs move from and to, and personal and organisational factors associated with mobility. Most existing evidence comprises only cross-sectional data on retrospectively identified factors and prospective intention.

Using the best available GP data, this study helps to understand who is most likely to move each year, how often moves occur and where they might move to and from. In particular, these results both highlight and quantify the strong association between mobility propensity and increasing rurality and remoteness of practice locations. Such evidence is useful in guiding more effective targeting of rural health policies and workforce planning and incentives.

1 Summary of origin–destination work location changes* for all non-registrar general practitioners (annual survey, 2008–2012)

	Destination location, no. (%)
Origin location	Metropolitan	Regional centre	Large rural	Medium rural	Small rural	Remote	Very remote	Total observations (n = 10 900)

Metropolitan	7015 (98.6%)	38 (0.5%)	9 (0.1%)	11 (0.2%)	30 (0.4%)	10 (0.1%)	5 (0.1%)	7118 (100%)
Regional centre	15 (2.0%)	695 (94.7%)	4 (0.5%)	6 (0.8%)	7 (1.0%)	4 (0.5%)	3 (0.4%)	734 (100%)
Large rural	28 (3.5%)	20 (2.5%)	741 (91.6%)	6 (0.7%)	10 (1.2%)	2 (0.3%)	2 (0.3%)	809 (100%)
Medium rural	26 (3.7%)	21 (3.0%)	5 (0.7%)	633 (89.3%)	17 (2.4%)	3 (0.4%)	4 (0.6%)	709 (100%)
Small rural	48 (4.6%)	27 (2.6%)	14 (1.3%)	13 (1.2%)	943 (89.6%)	4 (0.4%)	3 (0.3%)	1052 (100%)
Remote	14 (4.0%)	3 (0.9%)	5 (1.4%)	7 (2.0%)	5 (1.4%)	311 (89.1%)	4 (1.2%)	349 (100%)
Very remote	2 (1.6%)	6 (4.7%)	2 (1.6%)	4 (3.1%)	3 (2.3%)	6 (4.7%)	106 (82.2%)	129 (100%)

* There were an additional 51 work location changes observed within non-metropolitan areas where the location category was unchanged.

2 Characteristics of general practitioners who remain in or change their work location (annual survey, 2008–2012)

						Per-year “risk”
Origin location characteristic	No change: metropolitan	Metropolitan to rural*	Rural to metropolitan	No change: rural	Moved within rural	Rural to metropolitan	Metropolitan to rural

Total observations	7015	103	133	3378	271	3.20% (1 in 31)	1.34% (1 in 75)
Age group
< 40	751 (11%)	16 (16%)	28 (21%)	335 (10%)	44 (17%)	6.3% (1 in 16)	1.9% (1 in 53)
40–54 years	3171 (46%)	41 (40%)	68 (52%)	1722 (52%)	136 (52%)	3.2% (1 in 31)	1.2% (1 in 85)
55+ years	3006 (43%)	45 (44%)	35 (27%)	1281 (38%)	84 (32%)	2.3% (1 in 44)	1.4% (1 in 73)
Sex and family circumstances
Male and partner	3007 (45%)	40 (42%)	64 (50%)	1835 (58%)	136 (54%)	2.9% (1 in 35)	1.2% (1 in 83)
Female and partner	2783 (42%)	38 (40%)	51 (40%)	1024 (32%)	87 (35%)	4.0% (1 in 25)	1.3% (1 in 79)
Male and no partner	306 (5%)	8 (8%)	5 (4%)	141 (4%)	17 (7%)	2.7% (1 in 37)	2.4% (1 in 42)
Female and no partner	517 (8%)	10 (10%)	8 (6%)	180 (6%)	10 (4%)	3.8% (1 in 27)	1.8% (1 in 57)
Training location and place restriction†
Local, unrestricted	5282 (80%)	66 (71%)	70 (54%)	2226 (70%)	145 (56%)	2.6% (1 in 38)	1.1% (1 in 87)
IMG, restricted	295 (4%)	11 (12%)	35 (27%)	416 (13%)	60 (23%)	6.2% (1 in 16)	3.3% (1 in 30)
IMG, unrestricted	1022 (15%)	16 (17%)	24 (19%)	519 (16%)	54 (21%)	3.6% (1 in 28)	1.4% (1 in 72)
Business relationship
Principal, partner	2119 (33%)	15 (16%)	16 (13%)	1211 (39%)	38 (18%)	1.2% (1 in 87)	0.6% (1 in 155)
Associate	664 (10%)	11 (12%)	13 (11%)	469 (15%)	20 (9%)	2.4% (1 in 41)	1.5% (1 in 67)
Salaried employee	371 (6%)	13 (14%)	15 (13%)	307 (10%)	36 (17%)	3.9% (1 in 26)	3.1% (1 in 32)
Contract employee	3261 (51%)	52 (57%)	76 (63%)	1086 (35%)	119 (56%)	5.4% (1 in 19)	1.5% (1 in 68)
Length of stay in origin location
≤ 1 year	533 (8%)	24 (26%)	40 (33%)	416 (13%)	68 (29%)	6.8% (1 in 15)	3.8% (1 in 26)
2–3 years	722 (11%)	18 (19%)	32 (26%)	414 (13%)	50 (21%)	5.8% (1 in 17)	2.2% (1 in 45)
4–6 years	894 (14%)	14 (15%)	11 (9%)	406 (13%)	47 (20%)	2.2% (1 in 46)	1.4% (1 in 70)
7–10 years	917 (14%)	8 (9%)	12 (10%)	396 (13%)	25 (11%)	2.6% (1 in 39)	0.8% (1 in 125)
11+ years	3491 (53%)	30 (32%)	26 (21%)	1521 (48%)	48 (20%)	1.5% (1 in 67)	0.8% (1 in 125)

IMG = international medical graduate. * Rural includes regional, rural or remote. † Locally trained and restricted doctors (that is, Australian-trained graduates who are bonded to initially work in a rural area) have been removed from this analysis because there were very few observations in this group.

3 Panel logit models of general practitioners who move between metropolitan and rural work locations (annual survey, 2008–2012)

Origin location characteristic

Leaving rural,* model 1† (OR [95% CI])

Leaving rural, model 2† (OR [95% CI])

Leaving metropolitan, model 3‡ (OR [95% CI])

Leaving metropolitan, model 4‡ (OR [95% CI])

Age group (reference: 55+ years)

< 40 years

2.06 (0.79–5.34)

2.85 (1.09–7.43)**

0.62 (0.30–1.27)

1.11 (0.57–2.14)

40–54 years

1.62 (0.75–3.48)

1.68 (0.76–3.72)

0.61 (0.35–1.04)

0.76 (0.45–1.26)

Sex and family status (reference: male and partner)

Female and partner

1.07 (0.54–2.10)

1.13 (0.56–2.27)

0.97 (0.57–1.63)

0.90 (0.54–1.50)

Male and no partner

1.35 (0.35–5.27)

1.51 (0.37–6.14)

1.32 (0.50–3.46)

1.56 (0.64–3.79)

Female and no partner

0.62 (0.15–2.47)

1.07 (0.29–3.91)

1.26 (0.56–2.80)

1.20 (0.54–2.65)

Training location and place restriction§ (reference: Australia, unrestricted)

International, restricted

0.91 (0.41–2.05)

1.57 (0.73–3.36)

1.59 (0.74–3.41)

2.65 (1.29–5.43)††

International, unrestricted

1.39 (0.58–3.32)

1.92 (0.79–4.66)

1.02 (0.54–1.92)

1.33 (0.73–2.41)

Business relationship (reference: principal, partner)

Associate

1.18 (0.39–3.61)

1.76 (0.58–5.40)

1.83 (0.76–4.37)

2.43 (1.05–5.60)**

Salaried employee

3.89 (1.15–13.16)**

7.23 (1.94–26.89)††

3.22 (1.33–7.82)††

4.94 (2.14–11.40)††

Contract employee

5.18 (1.98–13.56)††

8.38 (2.72–25.78)††

1.47 (0.74–2.92)

2.20 (1.16–4.19)**

Length of stay in origin location (reference: 11+ years)¶

≤ 1 year

3.40 (1.34–8.62)††

—

4.46 (2.11–9.42)††

—

2–3 years

3.73 (1.41–9.85)††

—

2.63 (1.25–5.53)**

—

4–6 years

1.03 (0.36–2.94)

—

2.16 (1.04–4.48)**

—

7–10 years

0.86 (0.30–2.48)

—

1.24 (0.54–2.85)

—

Modified Monash rural scale (reference: regional centre)

Large rural

2.25 (0.78–6.50)

2.68 (0.88–8.20)

Medium rural

3.04 (1.02–9.09)**

3.46 (1.09–10.99)**

Small rural

3.96 (1.41–11.14)††

5.03 (1.61–15.78)††

Remote

1.22 (0.34–4.34)

1.27 (0.34–4.81)

Very remote

0.20 (0.01–3.03)

0.23 (0.01–4.17)

* Rural includes regional, rural and remote. † Dependent variable outcome = moves from rural/remote to metropolitan (n = 133). ‡ Dependent variable outcome = moves from metropolitan to rural/remote (n = 103). § Locally trained and restricted doctors (that is, local graduates who are bonded to initially work in a rural area) have been removed from this analysis because there were very few observations in this group. ¶ Length of stay was removed from Models 2 and 4 because of its strong multicollinearity. ** P < 0.05. †† P < 0.01.

General practice after-hours incentive funding: a rationale for change

After-hours incentive funding was made available to accredited general practices in 1998 as a foundation component of the Practice Incentives Program (PIP). The PIP after-hours practice incentive payment was intended to “help resource a quality after hours service and compensate practices that make themselves available for longer hours, in recognition of the additional pressures this entails”.1 Funding for each participating practice was based on the formula shown in Box 1. The model thus predominantly focused on access to after-hours care and comprised four main components:

a practice’s standardised whole patient equivalent (SWPE) — the sum of the fractions of care provided to practice patients weighted for the age and sex of each patient
how that practice ensured 24/7 care, from arranged external provision (stream 1) to self-provision only (stream 3)
the location of the practice, based on its Rural, Remote and Metropolitan Areas (RRMA) classification
the value per SWPE ($2).

In 2010, the PIP was audited by the Australian National Audit Office (ANAO).2 In an analysis of data provided by Medicare, the ANAO estimated that 14.9% of practices were non-compliant with respect to after-hours incentive payments between the financial years 2005–06 and 2008–09, the highest level of non-compliance across 12 PIP components. The Practice Nurse Incentive Program payments, at 9.5% non-compliance, were second highest. The ANAO investigated the potential of identifying practices at high risk of non-compliance for after-hours incentive payments in 34 practices deemed “high risk”. These practices were deemed high risk because of little evidence of actual after-hours service provision based on Medicare billings, while as stream 3 practices under PIP (Box 1), they were supposed to provide 24/7 care. On further investigation of these practices by random after-hours calls, only half provided at least a phone number at which a practice doctor could be contacted. The ANAO considered that secondary sources of information were imperative in ensuring practice compliance.

After-hours incentive funding and Medicare Locals

Under the aegis of the Commonwealth Government’s 2011 national health reforms to promote local decision making, Medicare Locals were delegated responsibility for the funding and delivery of after-hours services in their constituencies from 1 July 2013. Each Medicare Local, including Tasmania Medicare Local (TML), had the opportunity to develop and/or implement the most applicable and relevant mechanism for their locale.

For TML, the local situation is complex as its constituency encompasses large urban practices through to small isolated practices across an entire state that is geographically challenging. Three factors led to TML’s decision to continue the existing PIP after-hours funding arrangements in the 2013–14 financial year while simultaneously developing a preferred mechanism to be implemented in subsequent years: the complexities of the service environment; the desire to implement a fair, transparent and auditable mechanism; and the immediate need to provide after-hours incentive funding.

As part of the development process of the new after-hours incentive funding model, the PIP mechanism was interrogated to gain an understanding of the groundswell of disaffection for it among Tasmanian general practitioners. In this article, we describe the determination of the drivers of the PIP after-hours incentive funding model and the implications of this mechanism when viewed in context of the available (objective) Tasmanian after-hours data.

Drivers of the Practice Incentives Program mechanism: practice size, funding stream and location

For a given-sized practice, the primary determinant of after-hours PIP was its stream as reflected in PIP payments calculated for a practice of 2000 SWPE by RRMA classification and stream (Box 2). The practice size of 2000 SWPE was chosen for simplicity and consistency with subsequent calculations. Stream could make up to a threefold difference in after-hours incentive payments for practices of the same SWPE and RRMA levels, as compared with a 1.5-fold difference for practices of a given SWPE and a given stream level located in the most disparate locales (RRMA 1/2 and RRMA 7).

Together, stream and location could give rise to a potential 4.5-fold difference in payments — in other words, the impacts are additive. For example, a practice with 2000 SWPE classified at RRMA 7 and stream 3 will have a PIP of $18 000 as compared with $4000 for the same-sized practice classified at RRMA 1/2 and stream 1.

The greatest impact on after-hours PIP, however, was practice size, with payments directly proportional to individual practice SWPE as reflected in PIP payments calculated for practices in RRMA category 1 by SWPE and stream (Box 3). For example, the after-hours PIP for a stream 1 practice classified at RRMA 1/2 with 2000 SWPE is $4000 as compared with $20 000 for a similarly classified practice with 10 000 SWPE; a fivefold difference in practice size giving rise to a fivefold difference in after-hours PIP.

Thus, the PIP after-hours incentive funding mechanism was a multiplicative model primarily driven by practice size (SWPE), and in which the impacts of the claimed method of provision of after-hours care (stream) and location were additive, but of decreasing importance.

Implications

For the PIP after-hours incentive funding mechanism to fulfil its stated aim, practice size should therefore be the primary determinant of the burden and pressures faced by practices after hours, followed by stream and location. Given that the role of SWPE in PIP after-hours funding has been a major source of contention among Tasmanian general practices, the relative importance of practice size to after-hours burden is open to debate. Core questions include:

how much does SWPE matter to after-hours on-call requirements and after-hours service provision?
is it really the individual practice SWPE that matters?
does one extra patient make that much difference during the on-call period or is the number of doctors available to share the burden of after-hours care a more relevant consideration?

Does size really matter? Insights from the 2013–14 Tasmanian After-Hours Practice Funding Scheme

The average full-time GP has been attributed a value of 1000 SWPEs annually.3 Given this accepted workload and to expedite subsequent analysis, the 92 Tasmanian practices in receipt of an after-hours incentive payment within the 2013–14 Tasmanian After-Hours Practice Funding Scheme (as at 31 December 2013), have been categorised into one of five SWPE bands, as specified in Box 4. The choice of SWPE bands in part reflects feedback received during the development of the Tasmanian General Practice After Hours Incentive Funding Model for 2014–15 that sustainable on-call care requires the availability of 4 or 5 full-time doctors. The breakdown was also supported by the distribution of Tasmanian general practices by practice size (SWPE) (individual data not shown due to commercial-in-confidence reasons). The average SWPE for Tasmanian practices was 3693.

Practices have also been classified under a new Tasmanian General Practice Location Classification (TGPLC), which was developed as a foundational element of the funding model given concerns about the applicability of the RRMA and other classifications in the Tasmanian context. The TGPLC is an eight-level classification: 1, major metropolitan centre; 2, major urban centre; 3, other urban centres; 4, urban fringe; 5, rural locations; 6, remote locations; 7, very remote locations; and 8, isolated.

Practice characteristics

In Tasmania, practice size becomes more restricted as level of remoteness increases. Large practices (SWPE bands 4 and 5) are only located in major urban locations (locations 1–3) under the TGPLC, whereas isolated areas only support small practices (SWPE band 1). Detailed data are not provided (due to commercial-in-confidence reasons). It is also evident that as practices become increasingly remote and isolated, the provision of 24/7 care becomes a necessity, with all practices in locations 7 and 8 of the TGPLC classified as stream 3. This observation arguably reflects the intent of stream 3, which was to support those practices that had no option but to provide 24/7 care.1 However, stream 3 practices occur in each TGPLC location category (1–8) and thus this categorisation does not of itself identify practices with unavoidable after-hours burden. In contrast, unavoidable after-hours burden is at least partially captured through a comprehensive and accurate location mechanism.

There was no obvious relationship between stream and practice size in Tasmanian practices providing after-hours care, with at least one practice from each SWPE band (1–5) represented in each stream (1–3). While any practice in SWPE band 3 or higher theoretically has the in-house capacity to provide sustainable 24/7 care (at least 4 or 5 full-time equivalent GPs), 23 of 35 stream 3 practices fall within SWPE band 1 or 2. These practices unquestionably face excessive on-call burden. Whether this burden is avoidable largely depends on the availability of viable alternative providers — that is, the existence of practices in the near vicinity to share the load.

Together, these data indicate the importance of determining the specific objectives of any after-hours incentive funding scheme and ensuring that the mechanism embodies those objectives — for example, support for (unavoidable) burden.

Performance by key determinants of Practice Incentives Program funding: practice size, funding stream and location

An analysis of objective data of after-hours service provision by general practices receiving after-hours incentive funding through TML provides further insights into the functioning of the 1998–2013 PIP after-hours funding scheme in relation to practice size, stream and location, and type of after-hours care.

Urgent after-hours attendances

Over the period 1 July 2013 to 31 December 2013, there is clear evidence that practice location, based on the TGPLC, has the strongest association with the number of urgent after-hours attendances per practice (Medicare item numbers 597 and 599), followed by stream. As assessed through simple linear regression, location explained 27% of the difference in urgent after-hours attendances between practices (P < 0.001) as compared with 9% for stream (P = 0.003). Practice size (SWPE band) had no explanatory value (R2 = 0.01; P = 0.39) (individual practice data not shown due to commercial-in-confidence reasons). The importance of practice location in the provision of urgent after-hours care is also reflected in the fact that the 18 practices located in non-urban localities (locations 5–8) accounted for almost two-thirds (926 of 1404) of urgent after-hours attendances.

On the basis of these results, and the fact that the most remote practices are smaller in size and stream 3 (on-call 24/7), rural and remote practices are, in general, facing greater burdens than urban practices after hours.

After-hours services to registered aged care facilities: a distinct form of after-hours activity

An interesting and marked contrast exists in relation to practice performance in the provision of care to immobile patients — for example, patients in residential aged care facilities (RACFs). First, the vast majority (2139 of 2423 [88.3%]) of after-hours RACF visits (Medicare item numbers 5010, 5028, 5049 and 5067) occurred in urban localities (locations 1–4) a finding consistent with the distribution of RACF bed numbers. Second, no relationship was found between increasing remoteness and number of after-hours RACF visits per practice (R2 = 0.00, P = 0.92), nor was a relationship found between number of after-hours RACF visits and a practice’s stream (R2 = 0.01, P = 0.36). Some stream 1 practices gave rise to some of the highest levels of RACF visits, while otherwise ensuring urgent on-call access. For some practices, this activity may be arising due to formal arrangements with a nearby RACF. A relationship was, however, observed between number of RACF visits per practice and SWPE band, although it was limited, with practice size explaining 6% of the difference in number of RACF visits per practice between practices (P = 0.02). Some of the smallest practices gave rise to some of the highest levels of RACF visits (data not shown due to commercial-in-confidence reasons). Again, this could be due to formal arrangements.

Summary

Tasmanian data indicate that practice location is the primary predictor of on-call burden — that is, the burden increases as remoteness increases. Conversely, RACF visits, which are greater in number than after-hours attendances, predominantly occur in urban locations, but location is not a predictor of individual practice activity. Neither practice size (SWPE band) nor practice stream appear to be strong independent predictors of the provision of any form of after-hours care, although there are small positive trends for numbers of RACF visits and urgent on-call attendances. This trend reflects the fact that across SWPE bands and streams, there is a spectrum of providers (those that do and do not provide urgent care) and practices (ranging from those that provide minimal after-hours RACF visits to those that provide extensive after-hours RACF visits). These results were underpinned by relationships between location and stream (all very remote and isolated practices being stream 3), and between location and practice size (large practices > 6000 SWPE only being located in major urban locations).

Conclusion

The PIP after-hours incentive funding mechanism operating as at 30 June 2013 did not preferentially support practices that provide after-hours care. An after-hours incentive funding mechanism that recognises those practices that have no alternative but to provide 24/7 care is needed. Use of streams or tiers in the mechanism is considered inappropriate, potentially amounting to a perverse incentive. RACF visits should be considered an important but distinct form of after-hours care in an incentive funding mechanism. Finally, demands for transparency and use of auditable data are well justified.

1 Calculation of the Practice Incentives Program AHPIP

AHPIP = $2.00 × SWPE × stream × (1.00 + RL)

Stream = 1 means that the practice ensures access to 24/7 care

Stream = 2 means that the practice ensures access to 24/7 care and provides minimum specified levels of care (based on SWPE and hours of after-hours care provision)

Stream = 3 means that the practice provides 24/7 care

RL1 = 0; RL2 = 0; RL3 = 0.15; RL4 = 0.20; RL5 = 0.40; RL6 = 0.25; RL7 = 0.50

AHPIP = after-hours practice incentive payment. SWPE = standardised whole patient equivalent (practice based). RL = rural loading (based on the Rural, Remote and Metropolitan Areas classification).

2 After-hours incentive payments for 2000 standardised whole patient equivalents by stream and Rural, Remote and Metropolitan Areas (RRMA) classification

3 After-hours incentive payments for Rural, Remote and Metropolitan Area 1 by stream and standardised whole patient equivalent (SWPE)

4 Proportion of Tasmanian general practices in the After Hours Practice Funding Scheme by standardised whole patient equivalent (SWPE) band

SWPE band	SWPEs	No. (%) of general practices (n = 92)

1	≤ 2000	31 (33.7%)
2	2001–4000	30 (32.6%)
3	4001–6000	15 (16.3%)
4	6001–8000	10 (10.9%)
5	8001 +	6 (6.5%)

Suppression clauses in university health research: case study of an Australian government contract negotiation

Government research contracts routinely contain suppression clauses. Have universities forgotten their role in promoting open enquiry?

In a 2006 survey of a random sample of public health academics in Australia (46% response fraction), 21% of the 302 respondents reported having personally experienced a funding-related suppression event in the preceding 5½ years; ie, a funder had invoked a clause in the funding contract “sanitising, delaying or prohibiting” the publication of research findings.1The study also showed that the incidence of sanitisation events had increased over time. According to the respondents, their work was targeted because it “… drew attention to failings in health services (48%), the health status of a vulnerable group (26%), or pointed to a harm in the environment (11%)”.1

These findings appear to reflect a worrying tendency of Australian governments to seek to control the conduct and reporting of public good research.2 In this article, I present an account of a recent contract negotiation that arose from a researcher-initiated grant application in a competitive round called by an Australian state government agency. This will be followed by my analysis of the negotiation process, and a proposed research agenda for investigating the extent, causes and implications of such practices.

The proposed contract and our responses

In 2012, I was the primary applicant on a modest (< $50 000) researcher-initiated grant application proposing a pilot study in an Australian hospital of an alcohol screening and brief intervention program. I was notified that my application had been successful, and this was followed by an email message and draft contract. Included in the email message was the statement: “Some clauses in the Agreement are non-negotiable, particularly those that relate to insurance requirements and Intellectual Property”.

I will discuss three clauses that my colleagues and I identified as problematic, together with the responses we sent by email to the funder. Identifying information has been removed.

(1) Intellectual property (IP)

The proposed contract (clause [x]):

… ownership of Intellectual Property in or in relation to Contract Material vests upon its creation in [the funder]. The University must, upon request by [the funder] do all things necessary to vest ownership and title of Intellectual Property in [the funder].

We responded:

We’re not comfortable with the notion that [the funder] would own the IP given the substantial in-kind contribution of the University in this case, eg, all of the investigator time (mine and Dr [X]’s). We will discuss with the University legal people but suggest either joint ownership of the IP or that we own it and grant [the funder] an irrevocable licence to use it (my preference).

(2) Publication

The proposed contract: (clause [y])

The University must not publish any articles, statements or any other information arising from this Agreement without [the funder]’s approval in writing beforehand.

Our response:

We can’t accept a restriction on the right to publish findings from this work. As the clause presently reads, [the funder] could deny permission to publish and this is at odds with the University of [X] Act 1989 which requires us to “promote free enquiry”.

We are more than happy to notify [the funder] of our intent to publish and to provide a copy of any paper we submit to a scientific journal. Accordingly, we propose an amendment to the clause along the following lines:

“The University will provide a copy of any articles, statements or any other information arising from this Agreement to [the funder] at least 28 days before any such item is published.”

(3) Termination for convenience

The proposed contract (clause [z]):

[The funder] may by notice in writing at any time terminate this Agreement for convenience, such termination to be effective immediately unless stated otherwise in the notice.

Our response:

This is not acceptable to the research team. Why should the purchaser of a research service be permitted to pull out of the agreement without notice or penalty? We cannot invest our time in the research without knowing that the work will be paid for.

The funder’s response to our comments

Re Clause [Intellectual Property]: … [this is] a [government legal entity]-funded small grants program so IP is vested in [the funder].

Re: Clause [Publications]: [the funder]’s approval is required in writing before publication.

If the IP and Publication are not acceptable then you may chose (sic) to relinquish the grant.

In a telephone call that followed this email exchange, I was told by the funder that this was the “standard contract”, and that no one had previously complained about these clauses.

The University’s position

I also forwarded the draft contract to the University legal office and asked them to examine it. I received the following response:

The contract is a nice clear contract. The only points which may be of conjecture are clauses [x] and [y] re IP and publishing respectively. Can you read those two clauses and let me know if you would seek changes to them? If you do, we can try and get those changes, but may not necessarily succeed. An example may be that you feel that the publishing rights are too fettered.

I was left with the clear impression that, if my colleagues and I did not object to these clauses, the University would sign the contract.

The outcome

Months later, after several email exchanges and telephone calls with the funder, I was notified that they were willing to accept all the amendments we had specified:

Clause [x] Intellectual Property: joint ownership of the IP is acceptable to [the funder].

Clause [y] Publications: changes acceptable.

Clause [z] Termination for Convenience: can be deleted.

In relation to the last clause, it was explained to me in a telephone conversation that it had been inadvertently “left in” from another contract.

This notification by the funder, sent by email, was followed by the statement:

Again I have been asked to stress that these changes are agreed to based on the nature of your application and do not set a precedent for any future contract negotiations.

What was special about our application was not explained.

Analysis

In summary:

A government agency presented us with a contract that would have enabled it to own all of the intellectual property generated from our research, prohibit publication of our findings, and to shut down the project without notice or explanation;
The University would probably have signed the contract containing those clauses;
After lengthy negotiation between the researchers and the funder, the funder relented on all three clauses.

Our experience is consistent with the findings that Yazahmeidi and Holman published in 2007.1 The statements by the funder that this was a standard contract and that clauses had been “left in” from a previous template, as well as the claim that it was unusual for such an arrangement to be challenged further indicate that suppression clauses may be common.

Australia’s universities were each established by Acts of Parliament that specify the object and functions of these institutions. For example, the University of Newcastle Act 1989 stipulates in Part 2, Section 6 (“Object and functions of University”):

(1) The object of the University is the promotion … of scholarship, research, free inquiry …

(2) The University has the following principal functions for the promotion of its object: … (d) the participation in public discourse.

In our case study, and in the cases documented by Yazahmeidi and Holman, universities were prepared to consent to contracts that explicitly limited the capacity for open enquiry and participation in public discourse. The critical context for our experience was that the funding we were seeking was not for the salary of an existing employee, so we could negotiate robustly, even at the risk of not being funded.

It should be noted that our project was uncontroversial; its aims could not threaten a vested interest, in contrast to other research into alcohol and health that we have undertaken (eg, the impact of controlling alcohol availability on antisocial behaviour3,4).

I recently presented our case study at an international conference and asked the audience of about 50 alcohol policy researchers about their experiences. Scientists from Australia, Canada, Switzerland, Sweden and the United Kingdom reported that they had firsthand experience of similarly restrictive contracts in their countries. Several United States scientists affirmed that this type of contract would be unacceptable to their institutions, which, they said, were vigilant about protecting publication and intellectual property rights.

Current policy on public good research contracting

The Australian Government’s Coordination Committee on Innovation has developed the “National principles of intellectual property management for publicly funded research”, the purpose of which is to “provide guidance for the ownership, promotion, dissemination, exploitation and, where appropriate, protection of Intellectual Property (IP) generated through Australian Government funded research by public sector institutions”, including “grants awarded by the ARC, NHMRC, and other government research funding schemes.”5 The latter category appears to include the researcher-initiated competitive grant scheme discussed in this case study, but it explicitly excludes research purchased by government through requests for proposals and tenders (eg, through AusTender, https://www.tenders.gov.au). How much health research is purchased through tenders is unknown.

The draft contract examined in this article was in direct conflict with these principles, which state that “Ownership and the associated rights of all IP generated as a result of Australian Government competitively funded research will initially be vested in the research institutions receiving and administering the grants as a way of recognising the inventive contribution made by the research institutions.”5

Neither the “National principles of intellectual property management for publicly funded research” nor the Australian Government Intellectual Property Manual6 discuss publication rights except in the context of IP. For example, there is advice in the latter document that government agencies should “[c]onsider a review period prior to any disclosure of materials by publication or transfer; provid[ing] opportunity to identify IP and explore protection options” (p. 77). In other words, there is no authoritative guidance about publication policies from the perspective of the public’s right to know the outcomes of public good research purchased with public funds. There may be value in the universities and state, territory and federal governments working together to produce model clauses for future research contracts.

Research questions

Our case study generates several research questions about the extent, causes and implications of the practices we have discussed here.

Extent

What proportion of government-funded health research in Australia is subject to suppression clauses?
How do contracting practices vary by jurisdiction, department and subject area?
What do university policies say on the matter of suppression clauses?
Do universities differ in their willingness to enter such contracts?
How common are such contracts in other countries?

Awareness

How aware are university research and legal officers of this problem?
How aware are researchers of the problem?
What do the public think about government agencies purchasing research in this way and of universities being party to such contracts?
Are government funding agencies aware of the potential harms of such research purchasing practices?

Implications for the scientific evidence base

What are the implications of suppression clauses for the evidence base?
Are publication clauses documented in reports of research findings?

Ethical and legal considerations

In what circumstances, if any, are governments justified in asserting a right to secrecy or selective release of health research findings?
What are universities’ and individual researchers’ legal and ethical responsibilities?
Is there scope for using provisions of federal and state freedom of information acts to identify instances of suppression?

Conclusion

It appears that governments in Australia commonly seek the right to limit the freedom of health scientists to report the findings of publicly funded research. It should be noted that there are no such restrictions on research funded by the major competitive grant schemes, those of the National Health and Medical Research Council and the Australian Research Council. Our findings relate to a large number of federal and state departments and agencies for which research is not their primary responsibility, and the problem we have discussed is unlikely to be limited to health research. Lessons from our experience include the fact that non-negotiable clauses sometimes become negotiable — if the researcher persists. Negotiating when the stakes are low (as in our case) and systematic efforts to raise the awareness of universities, funders and the public may help us move toward a more open research funding process. Policy is needed to guide the purchasing of public good research. This may be accomplished via negotiation between the universities and key federal and state agencies, and through extension of the existing policies (eg, those in the Government’s IP manual6) to cover the publication of research findings for the public good.

Fostering creativity and innovation in the health system: the role of doctors-in-training in biomedical innovation and entrepreneurship

Doctors-in-training are well positioned to continue Australia’s strong history of biomedical innovation and entrepreneurship

In Australia and overseas, there is growing interest in the development of biomedical innovation and entrepreneurship relating to improvement of diagnostics, treatments and health services.1,2 Innovation follows from the efforts of biotechnology and pharmaceutical organisations, academia and the health services sector. Entrepreneurship is then required to market innovations.

A lack of talented biomedical innovators and entrepreneurs limits positive change in health care. The Department of Health and Ageing’s McKeon review: strategic review of health and medical research — better health through research proposed ways to enhance research capacity and commercial and non-commercial pathways to innovation.1 The review strongly supported junior health professionals engaging in innovation research. These proposals were endorsed by the Australian Medical Association’s position statement on clinical academic pathways.3

Doctors-in-training (DITs) — medical students, interns, residents and specialty trainees — are well placed to develop their innovation and entrepreneurship skills, given their traditional medical skills and knowledge. They have skills in basic, clinical and public health science. DITs train at the coal face of the health care system, and are therefore exposed to the practical outcomes of policies, procedures and systems. They have well developed skills that are empirically associated with entrepreneurial success,4 such as team work, communication, productivity, accountability, responsibility, problem solving and data analysis.

Biomedicine in Australia has many innovators and entrepreneurs (eg, CSL, Cochlear and the inventors of the human papillomavirus vaccine), and Australia has a strong biotechnology sector. According to the Scientific American Worldview scorecard for 2013, Australia ranked seventh in the world for biotechnology (across health and non-health industries), up from tenth in the previous year, first for “best growth in public markets”, second for “greatest public company revenues” and second for “most public companies”.5 The 88 biotechnology companies listed on the Australian Securities Exchange are valued at more than $51 billion. However, there is room to further support innovation and entrepreneurship training systems for our DITs to further strengthen the Australian outlook, particularly when considering the plethora of international training programs that are offered.2

Engagement with biotechnology and pharmaceutical industries can generate benefits for academic research, including funding, in-kind resources (eg, high-throughput facilities), commercial exploitation of technology or intellectual property, and cross-fertilisation of knowledge and skills.

In this article, we explore the role of DITs as biomedical innovators and entrepreneurs.

International experience with training in biomedical innovation and entrepreneurship

The US-based Society of Physician Entrepreneurs (SoPE) recently launched their Innovation Scholar Program, which offers doctors hands-on bioentrepreneurship experience through practicums with biomedical companies.6 The program comprises a 1-year apprenticeship, in which the junior physician is linked to a biomedical company, a mentor and a project. Companies provide the scholar with a broad array of knowledge, skills and experiences in product development and commercialisation. The scholar may also engage in an innovation-related university curriculum, depending on the geographical and workplace situation.

A Society for International Bioentrepreneurship Education and Research was recently proposed, with a mission to advance international bioentrepreneurship education and research.

A new course titled Medical device design and innovation from Yale University offers students opportunities to develop innovative solutions to problems posed by physicians.7 This course pairs physician mentors with students from medicine, engineering, physics, chemistry and management. Also, massive open online courses (MOOCs) on innovation are freely available online and are gaining popularity among junior doctors; for example, Coursera — an education platform that partners with top universities and organisations worldwide to offer free online courses — has a course titled Healthcare innovation and entrepreneurship.8

Other pathways exist for doctors to engage in specific innovation and entrepreneurship training. For example, in the United Kingdom, Industrial CASE (Collaborative Awards in Science and Engineering) Studentships9 are offered to bioscience PhD students in a collaborative environment; partnerships are established between academic institutions and organisations in the private, public and civil society sectors, with a PhD supervisor supplied by each branch. This program is administered by the Biotechnology and Biological Sciences Research Council and, in 2015, 125 of these 4-year studentships were made available. This program is funded by academic and non-academic partners, but financial obligations of non-academic partners depend on the size of the company.

Drivers of increased training in biomedical innovation and entrepreneurship

The development of innovation and entrepreneurship-related skills is emphasised in physician competency framework of the Royal College of Physicians and Surgeons of Canada (the CanMEDS Framework),10 and is supported by the Australian Medical Council’s graduate outcome statements,11 the Australian curriculum framework for junior doctors12 and the Committee of Presidents of Medical Colleges’ position statement on the role of the medical specialist.13 Through exposure to the health system, practitioners develop an understanding of the complex systems in which innovation will be applied.

It has been shown that device manufacturers gain more premarket approvals from the patents of physician-founded firms than from those of non-physician-founded firms.14

National experience with training in biomedical innovation and entrepreneurship

The role of DITs in Australian biomedical innovation and entrepreneurship has received little attention from researchers, and not much is known about current levels of activity.

Quality and safety improvement projects for junior doctors are now offered, albeit in selected hospitals in Victoria and Western Australia. These seek to develop skills and leadership in medical service safety and quality improvement.15,16 The Medical Service Improvement Program in WA includes 3 months’ work on a quality improvement project.15 Throughout the program, junior doctors are offered supervision and mentorship, attend lectures, and visit non-health industry locations.

A recent review found that junior doctors entering the health care system are ideally placed to cultivate their interest and expertise in improving the health system.17 Collaborating with junior doctors rather than with their senior colleagues builds a strong quality culture. In the United States, quality improvement skills are now a core competency for accreditation of residents (ie, the vocational equivalent of DITs in the US) and this has stimulated quality improvement training in medical curricula.

The crisis in clinical academia and need for integrated pathways and structures for development of clinical academic careers has been widely discussed in Australia and overseas.18 There is a clear need for more medical practitioners to incorporate research, education and leadership dimensions into their clinical duties.1 Interest in clinical academia is high among medical graduates — up to 56% of Australian medical students express interest in including research in their careers.19 Several universities in Australia now offer dual MB BS/PhD degrees, with the aim of strengthening clinical academic training.

Benefits of increased training in biomedical innovation and entrepreneurship

Trainees, health services that employ trainees, the Australian community, research institutions and biotechnology companies all stand to benefit from greater development of innovation skills and knowledge. A number of academic articles explore the effectiveness of residents’ and trainees’ participation in curriculum redesign, clinical guideline implementation and quality improvement teams.20 Intuitively, health services are more likely to attract and retain a higher-quality workforce if they provide innovation-friendly positions.

Risks of increased training in biomedical innovation and entrepreneurship

Time away from clinical positions (part-time or full-time) means longer training, but evidence for compromised clinical learning as a result is slight.21 Greater non-clinical activities may draw medical practitioners out of clinical practice, but the effects of this remain to be empirically explored. The impact of a potential reduction in full-time equivalent DITs in clinical care owing to non-clinical activities (eg, research and innovation) should be tracked carefully. Population-based studies such as the Medicine in Australia: Balancing Employment and Life (MABEL) study enable such tracking.

Recommendations

To enhance innovation and entrepreneurship training for DITs, we recommend the following strategies.

Explore interest and engagement: Studies are required to better understand motivations for, perceived benefits of and detriments of engagement in innovation activities.

Encourage interest: A cultural change to enhance the focus on innovation in health care is needed. This could be be brought about by advocacy, mentorship, education programs and grant funding schemes. Interest in innovation should be encouraged throughout the medical career continuum. Coordinated and strong agency-led leadership is important. The Australian Academy of Health and Medical Sciences may be a suitable leadership organisation.

Develop structured education, mentorship and coaching programs: These would enable freer interchange between researchers and the biotechnology, pharmaceutical and investment industries, and would embed a greater commercialisation culture in research.

Enhance flexibility of training pathways: This would include opportunities for breaks in training and allocation of protected time for innovation-focused occupations, which could be counted towards training requirements.

Develop funding incentives to drive research and innovation: Universities, health and hospital systems need to be effectively incentivised to promote innovation.

Australia has a strong biotechnology sector, within which a stronger innovation-skilled medical workforce could be developed. In developing such a workforce, it is important to focus on training DITs as they are our future workforce.

State-based legal requirement for Schedule 8 prescriptions: why so complicated?

Inconsistent prescription requirements between Australian states and territories create unnecessary complexity for health professionals

In Australia, medicines defined as Schedule 8 (S8) under the Standard for the Uniform Scheduling of Medicines and Poisons are strictly regulated because of the high risk of misuse and/or physical and psychological dependence associated with them.1 They have to be prescribed, dispensed, documented and destroyed in specific ways that are in compliance with each state and territory’s different drug regulations. S8 medicines are under stricter control than Schedule 4 (S4) medicines (other prescription-only drugs), for which requirements have been standardised between states and territories.2,3

Australia has no central body to regulate the handling of S8 drugs. Although the Therapeutic Goods Administration (TGA) is the national body for the regulation of medicines, each state and territory self-regulates under the general principles established by the TGA and has its own interpretation and legislation regarding S8 drugs, resulting in varied prescribing requirements. The legal requirements for obtaining authority and writing prescriptions for S8 medicines are listed in Box 1 and Box 2: they are often difficult to find and are long and daunting to read.

Impact on practice

The establishment of a national registration agency, the Australian Health Practitioner Regulation Agency (AHPRA), in 2010 meant that Australian health professionals were allowed to freely practise in any state or territory. Greater mobility of health practitioners between jurisdictions has been accompanied by new problems.

First, to the best of our knowledge, prescribers newly relocated to a different state or who practise across more than one jurisdiction have no single, clear resource that documents the slight nuances in each state or territory’s regulations. Legal requirements for prescribing S8 drugs are not accessible in a prescriber-friendly manner. Pharmacists can guide prescribers on the regulations and legality of prescriptions; yet the same confusion applies to pharmacists who move interstate.

Second, travelling patients bringing an S8 prescription interstate might discover that a legal prescription in one state is not legal in another. The dispensing pharmacist would need to contact the medical practitioner in the patient’s home state to find a solution. If this could not be done, treatment would be delayed until a local prescription was obtained from a medical practitioner in the state the patient was visiting.

What can we do?

It may be impractical to unify health care legislation in Australia to eliminate the complexity. However, all states and territories could maintain individual regulations but unify the S8 legal requirements. Given that S4 requirements are standardised between the different states and territories, why are S8 requirements treated differently?

For the moment, resources highlighting state-based S8 requirements for prescribers should be made readily available. A comprehensive quick-reference guide, such as the table we provide here, minimises the ambiguity in legal requirements for health practitioners, and its use may also reduce the amount of time spent by pharmacists and doctors in correcting non-compliant prescriptions.

1 Current requirements of Australian states and territories for obtaining authority to prescribe Schedule 8 medicines

State or territory

Required authority

Australian Capital Territory4

Write “Standing short term approval” for treatment of less than 2 months. For treatment of longer than 2 months, write “CHO approval number” followed by approval number from the ACT Chief Health Officer

New South Wales5

From Director-General NSW Health for psychostimulants, alprazolam, methadone, buprenorphine, flunitrazepam and hydromorphone

Northern Territory6

None

Queensland7

None

South Australia8

From SA Minister for Health for more than 2 months of treatment13

Tasmania9

From Tas Secretary for Health for more than 2 months of treatment14,15 (1 month for alprazolam, prior approval for psychostimulants, fentanyl and hydromorphone)

Victoria10,11

May need a Drugs and Poisons Regulation Group permit to prescribe to drug-dependent patients

Western Australia12

From WA Department of Health Chief Executive Officer for more than 2 months of treatment

2 Current legal requirements for prescribing Schedule 8 medicines in each state of Australia

Australian Capital Territory4

New South Wales5

Northern Territory6

Queensland7

South Australia8

Tasmania9

Victoria10,11

Western Australia12

Prescriber

Name

✓

Address

✓

Phone no.

✓

Qualification

✓

Signature

Patient

Name

✓

H (with initials)

✓

Address

✓

Date of birth

✓

Medicine

Name

✓

H (description of the medicine)

✓

H (description of the medicine)

Form

✓

Not specified

✓

Not specified

Strength

✓

Not specified

Quantity

✓

H (in words and figures)

✓ (in words and figures)

H (in words and figures)

✓ (in words and figures)

H (in words and figures)

Direction

✓

No. of repeats

✓

H (in words and figures)

Interval for repeats

✓

Date

✓

Only one S8 drug per prescription*

Multiple items allowed

✓

Not specified

✓

Multiple items allowed

✓

✓ = required. x = not required. H = information that must be written in the doctor’s own handwriting. * Exceptions apply: different forms of the same drug are acceptable.

The spectacular recent trials of urgent neurointervention for acute stroke: fuel for a revolution

How should we redesign our stroke services in light of neurointerventional advances?

In 2013, neutral results from three trials of neurointervention for treating ischaemic stroke were simultaneously published — a triad of gloom.1–3 In just over 2 years since, five positive trials have been reported.4–8 What explains this extraordinary turnaround, and what are the implications for stroke services in Australia and around the world? The answers to these questions are surprising and reflect a mixture of science, technology and policy.

The roles of science, technology and policy

The science involved is the culmination of a decade of work on proving that brain imaging can identify the ischaemic penumbra — the area of the brain that has shut down and is on the path to infarction but, with successful reperfusion, is potentially salvageable. By recruiting patients with a favourable profile for reperfusion therapy (so-called target mismatch, where the ratio of perfusion lesion to established infarct is > 1.8, the perfusion lesion volume is > 15 mL, and the established infarct volume is < 70 mL),9 we are now able to identify those who are likely to respond well. In addition, computed tomography (CT) angiography is now widely available and can demonstrate major cerebral vessel occlusion — a clear target for therapy.9

In contrast to the neutral trials, the recent trials all used either advanced imaging to identify patients with the “reperfusion responder” profile or angiography to prove major vessel occlusion, or both, then randomly assigned this population of likely responders to receive endovascular reperfusion (usually in addition to alteplase thrombolysis) or standard acute stroke care. The combination therapy resulted in potent reperfusion and a dramatic treatment effect (Appendix), such that three of the five neurointervention trials were stopped early.

The technology is all about the device. In today’s fast-moving world, it is almost impossible to design, fund and complete a trial of a device without it becoming obsolete by the time the trial has finished — the fate of the previous studies.1–3 Unlike in coronary intervention, the thrombus or embolus in ischaemic stroke must be physically extracted, and the new generation of retrievable stents are a major advance in this regard. One of the recent trials, MR CLEAN, demonstrated that carotid stenting (for extracranial occlusion) was also required for 13% of the patients receiving intra-arterial treatment.4

Finally, the unanticipated influence of policy can have profound effects. One of the neutral trials, IMS III, was conducted in the United States at a time when neurointervention was generously compensated.1 This, together with the attractiveness of the technology (despite its lack of evidence), meant that most people were treated outside the trial. The difficulties in recruitment (only one or two patients per centre per year) and a possible selection bias of recruiting only “difficult” patients might have had an effect on the results of IMS III.

In contrast, the Dutch MR CLEAN trial provides an important lesson.4 All the neurointervention centres in the Netherlands participated in this trial, and from 2013 there was no reimbursement for people treated outside the trial. This allowed some 500 patients to be recruited from 16 sites in just over 3 years, compared with IMS III, which needed 7 years to recruit 656 patients from 58 sites. If trial-only reimbursement for unproven devices were enforced, it is likely that reliable data on efficacy would have been available much earlier, potentially saving hundreds, if not thousands, of lives.

Implications for practice

Implementing intravenous thrombolysis has been a difficult and protracted affair, and we are still researching the best ways to achieve it.10 However, there has been progress in Australia, thanks to the hard work of clinicians in the state stroke networks. For example, in New South Wales, the Agency for Clinical Innovation led the establishment of a state-wide stroke reperfusion strategy that involved training paramedics to screen for potential thrombolysis candidates with the FAST (Face, Arms, Speech, Time) test and fast-tracking potentially eligible patients to 24-hour thrombolysis centres.

Our challenge is how to redesign our stroke services and how best to build capacity in the neurointerventional workforce. What is the required infrastructure, support and training in advanced imaging selection that would work in our hospitals? The London model of hyperacute stroke centres might work in our capital cities, but the “drip and ship” model of starting thrombolysis followed by urgent transfer to a comprehensive stroke centre may be a solution for outer metropolitan and regional thrombolysis centres. It is abundantly clear that we will never be able to provide on-site neurointervention at all stroke thrombolysis-capable hospitals, nor achieve complete equity of access to endovascular therapy for stroke patients from rural and remote communities.

A potential solution to the shortage of neurointerventionalists is the emerging model to train neurologists in interventional neuroradiological skills. Given there is broad acceptance that a stroke physician (with appropriate training) does not necessarily have to be a neurologist, there is growing support in the US and Australia for the position that a neurointerventionalist does not have to be a radiologist, provided he or she has had appropriate training.11

The exact shape of future neurointerventional stroke centres is still uncertain, but the endovascular revolution has arrived, and the stroke community needs to work quickly to redesign stroke care services and build the workforce of specialists trained in endovascular therapies. Drivers for change will include the new national stroke care standard (launched on 10 June 2015)12 and the next revision of the national Clinical guidelines for stroke management.

Stroke medicine has come a long way from the nihilism of two decades ago, with numerous interventions now supported by high-level evidence. Immediate brain imaging will identify strokes that are due to haemorrhage, and rapid blood pressure lowering and stroke unit (or intensive) care are the mainstay of treatment for these patients, with surgery needed only for a select few. For patients with ischaemic stroke, revascularisation with appropriate intravenous thrombolysis should be sought, followed by advanced brain imaging to identify patients suitable for additional endovascular therapy.

What is the future? Colleagues in the US and Germany are exploring the utility of an ambulance with an onboard CT scanner (Box), with anecdotal reports of excellent responses to alteplase when given within minutes of major stroke onset. The search is also on for more effective intravenous thrombolytic drugs, with Australia leading an international trial of tenecteplase versus the current standard, alteplase.13 However, none of this will be effective without further public health interventions to improve awareness of stroke and the importance of immediately calling 000 for any suspected stroke patient. When it comes to stroke, time is brain.

Ambulance with a computed tomography scanner