Background
Appropriate utilisation of oral anticoagulants (OACs) reduces stroke risk in patients with atrial fibrillation (AF) (Aguilar & Hart, 2005). The vitamin K antagonist, warfarin, has been the mainstay of anticoagulation in AF for over two decades. It decreases the risk of stroke by almost two-thirds (Aguilar & Hart, 2005). However, it has a narrow therapeutic index and is associated with problematic drug and food interactions that require monitoring and dose adjustments. The direct-acting oral anticoagulants (DOACs) are at least non-inferior to warfarin in efficacy and safety (Connolly et al., 2009; Granger et al., 2011; Patel et al., 2011). In Australia, three DOACs (rivaroxaban, dabigatran, and apixaban) were listed for Commonwealth subsidy under the Pharmaceutical Benefits Scheme (PBS) for non-valvular AF in 2013; since then their use has markedly increased (Admassie, Chalmers & Bereznicki, 2017; Alamneh, Chalmers & Bereznicki, 2017; Drug Utilisation Sub-Committee (DUSC), June 2016; Pol, Curtis, Ramukumar & Bittinger, 2018). In contrast, the prescribing of warfarin has declined (Admassie, Chalmers & Bereznicki, 2017; Alamneh, Chalmers & Bereznicki, 2017; Drug Utilisation Sub-Committee (DUSC), June 2016; Pol, Curtis, Ramukumar & Bittinger, 2018).
Recent studies on the utilisation of OAC have highlighted both underuse and overuse in patients with AF in Australia (Admassie, Chalmers & Bereznicki, 2017; Alamneh, Chalmers & Bereznicki, 2017). The Tasmanian AF Study observed prescribing practice from 2011-2015, and reported that 55% and 63% of eligible AF patients with a high stroke risk were prescribed an OAC before and after DOACs were listed on the PBS, respectively (Admassie, Chalmers & Bereznicki, 2017). This study, however, involved only hospitalised patients, who might have been more co-morbid than those managed in primary care; the results therefore may not have reflected OAC prescribing rates in general practices nationally. The current AF prescribing patterns, in relation to stroke risk, in the Australian primary care setting remain unknown.
The primary objective of this study was to investigate the proportion of Australian primary care patients with AF prescribed an OAC according to their stroke risk, and temporal trends in prescribing patterns. The secondary objective was to examine variation in OAC prescribing between general practices.
Methods
Data for this study was obtained from NPS MedicineWise’s dataset, MedicineInsight. This is the largest and the most representative general practice dataset available to researchers in Australia. In 2018, 671 (8.3%) of the 8,065 general practices in Australia had been recruited by NPS MedicineWise (Busingye et al., 2019; MedicineInsight, 2018).
MedicineInsight uses a third-party tool that extracts, de-identifies and securely transmits patient data each week to its secure data repository. The extraction tool allows developing a longitudinal database of patients in general practices. The data that MedicineInsight collects from general practices include patient demographics, diagnoses, pathology test results, prescribed medications and reasons for encounter. However, specific patient identifiers, such as patient name, address, and date of birth, are not included in this dataset (MedicineInsight, 2018).
We performed ten sequential cross-sectional analyses of data on 1 September every year (census date) from 01 September 2009 to 01 September 2018. Patients with a recorded diagnosis of non-valvular AF were included in each analysis if (i) they were aged 18 years or older and not deceased on or before the census date, (ii) their recorded AF diagnosis date was at least 4 months before the census date, (iii) they had had three or more recorded general practice visits in the previous two years and at least one of these visits was in the last six months, and (iv) they had been registered in the general practice’s electronic records at least one year before the census date.
We defined patients with AF as being prescribed an OAC (warfarin, dabigatran, rivaroxaban or apixaban) or antiplatelet agent (clopidogrel, ticagrelor, aspirin, ticlopidine, prasugrel, dipyridamole, abciximab, eptifibatide or tirofiban) when they had at least one recorded prescription, dated within 365 days before the census date. Aspirin is available without a prescription, but we could only capture prescribed data.
For most of our study period, guidelines recommended using the CHA2DS2-VASc score (congestive heart failure (1 point), hypertension (1 point), age ≥ 75 years (2 points), diabetes mellitus (1 point), stroke/transient ischaemic attack (TIA) (2 points), vascular disease (1 point), age 65-74 (1 point) and sex female (1 point)) for assessing stroke risk and treatment eligibility in patients with AF (Steffel et al., 2018). Patients with AF were stratified as low risk when CHA2DS2-VASc was 0 and male or 1 and female, moderate risk with CHA2DS2-VASc =1 and male, and high risk with CHA2DS2-VASc ≥2 (Steffel et al., 2018). The proportion of patients who were prescribed an OAC, antiplatelets alone, or neither were calculated with 95% confidence interval (CI) each year on 1 September from 01 September 2009 through 01 September 2018. Temporal trends were shown using graphs and a Cochran-Armitage test for trend was used to determine if any observed trends were statistically significant. Similarly, the proportion of patients with moderate to high stroke risk (CHA2DS2-VASc ≥1 and male or CHA2DS2-VASc ≥2) or low stroke risk (CHA2DS2-VASc =0 for male or CHA2DS2-VASc =1 for female) who were prescribed an OAC was calculated each year for each practice site. All practice sites that contributed data at least for a year were included. Prescribing rates were ranked into quintiles and used as an indicator of general practice sites’ prescribing performance. The variation between the highest- and lowest-prescribing practice quintiles each year was calculated as a prescribing gap.
Socio-economic indexes for areas (SEIFA) quintile is an index developed by the Australian Bureau of Statistics (ABS) and ranks areas in Australia from 1 (most disadvantaged area) to 5 (most advantaged area) (Australian Bureau of Statistics, 2018). The ABS categorise rurality into five categories using the Accessibility/Remoteness Index of Australia (ARIA) score. These categories are major cities (ARIA 0-0.20), inner regional (0.21-2.40), outer regional (2.41-5.92), remote (5.93-10.53), and very remote (10.54-15) (Australian Statistical Geography Standard (ASGS), 2017); we collapsed remote and very remote areas into one group. SAS software (SAS version 9.4, SAS Institute Inc., Cary, NC, USA) was used for all data analyses, and a two-sided p-value <0.05 was considered statistically significant.
Ethics approval was obtained from the Tasmanian Health and Medical Human Research Ethics Committee (H0017648). We also obtained approval to conduct this study from the MedicineInsight independent Data Governance Committee (2018-033). Patients were not identifiable, and individual patient consent was waived for our ethics application.