Over 50% of Australians have at least one chronic disease. Stroke is a complex and costly disease. People living with stroke, require management of risk factors to prevent future cardiovascular events, and often have other chronic diseases and impairment. Of concern, people living with stroke report long-term unmet needs, particularly beyond 6 months post-stroke [1, 2].

These long-term unmet needs relate to a wide range of impairments in body function (e.g., fatigue, memory/concentration, speech) and activity and participation limitation (e.g., mobility, leisure, employment), which are associated with lower quality of life and increased risk of cardiovascular disease due to lifestyle impacts [3, 4]. As such, it is recommended that people living with stroke receive annual reviews to identify unmet needs and referral to a multi-disciplinary healthcare team to address identified problems [5, 6]. However, this care can be difficult to co-ordinate due to organisational and budgetary boundaries between hospital and community care which negatively impact transfer of care and provision of community based services [5].

Since 1999, the Australian government has invested substantial funds for policies that promote enhanced models of primary care. Australian Primary Care Practitioners, also known as General Practitioners, are incentivised through Medicare funded policies to provide chronic disease management and coordinate multidisciplinary care planning. These items are targeted at facilitating care, in accordance with clinical guidelines, using a coordinated, multi-disciplinary approach and promoting self-management of chronic disease [7]. Additional funding beyond that available for a standard consultation is provided via specific Medicare claim items. Funding items are also provided for reviewing these plans every three months, ensuring that care is responsive to patients’ changing needs [8]. Those requiring multi-disciplinary team-based support can also access Medicare subsidised allied health services, which may help address their long-term unmet rehabilitation and secondary prevention needs. Despite serving an important function, Primary Care Practitioner use of these guidelines and utilisation of these care types in a large, nationally representative sample, with reference to the impairment profile of survivors, is currently unknown.

We aimed to describe the use of Medicare funded chronic disease management, and multidisciplinary care coordination policies for people with stroke living in the community and how this varied by impairment status.

This is a retrospective, observational, cohort-based data linkage study. The cohort was derived from the Australian Stroke Clinical Registry (AuSCR), and included adults aged $\ge$18 years admitted to one of 26 participating hospitals from five Australian states between April 1, 2010 and December 31, 2014. Patient-level registry data were linked to Medicare and pharmaceutical claims data to June 31, 2016 [9, 10]. For this analysis we restricted the cohort to those with stroke (i.e., excluding those with Transient Ischaemic Attack [TIA]) who completed AuSCR six-month follow-up questionnaires and were living in the community at the time of completion. We excluded those living in residential aged care as they are not eligible for these items.

Ethics approvals for this study were obtained from Monash University (#7864) and the Australian Institute of Health and Welfare (#EO2017/1/346). Approvals were also received from the AuSCR Research Task Group, the AuSCR Steering Committee, and the Queensland Public Health Act. The AuSCR holds approval for collecting data via an opt-out model of consent, with a waiver of consent for those who died in hospital.

Of the initial AuSCR cohort (N = 17,980), 93% were linked with Medicare and 5432 were eligible for inclusion in analysis (Fig. 1). Of these, 72% were aged $\ge$65 years, 44% were female, 86% had experienced ischaemic stroke, and 57% had severe stroke (Table 1). AuSCR registrants who complete the follow-up survey have been shown to be demographically and clinically similar to non-responders [17].

Fig. 1.

Flow diagram of participants included in our latent class analysis. AuSCR, Australian Stroke Clinical Registry; MBS, Medicare Benefit Schedule; DVA, Department of Veterans Affairs; TIA, Transient Ischaemic Attack.

More than 99% of the study cohort saw a Primary Care Practitioners within 18 months of stroke and 13% had $\ge$1 claim for a medical specialist. During the study period 39% had a claim for a chronic disease management plan or review, 39% had a claim for multidisciplinary care coordination or review and 46% had a claim for Medicare-subsidised allied health care. Of those who claimed Medicare-funded allied health services, 47% claimed podiatry, 20% claimed physiotherapy services and 4% had a claim for services relevant to secondary prevention (exercise physiology or dietetics). Other types of allied health services such as speech pathology or occupational therapy were claimed by less than 2%.

For the LCA, two models converged, a model with: (1) two-level EQ-5D-3L indicators (none vs moderate/severe) and two classes, and (2) three-level indicators (none, moderate, severe) and three classes (Supplementary Table 2). The two-class model demonstrated a slightly lower AIC (29,462) than the three-class model (38,506). However, the three-class model was chosen as it produced groupings of patients with impairment that were clinically meaningful. Guided by the pattern in which patients within each class reported their impairment within each EQ-5D-3L domain (Table 1), we defined class 1 as patients with minimal impairment (n = 2576), class 2 as moderate impairment (n = 2293) and class 3 as severe impairment (n = 563). For example, those who indicated on the EQ-5D that they had no problems with mobility had a 99.6% odds of being allocated to the minimal impairment group whereas those confined to bed had a 100% odds of being allocated to the severe impairment group (Supplementary Table 3). An increase in age, stroke severity and number of comorbidities was also observed with increasing severity (Table 1).

A smaller proportion of those with severe impairment claimed a chronic disease management plan or review in the 18 months following stroke (27%) than those with minimal (37%) or moderate (43%) impairment (Table 2, Fig. 2). Those with minimal impairment were less likely to have a claim for multidisciplinary care coordination or review than those with moderate or severe impairment (32% vs 45%, 44%). Podiatry claims were more common in those with moderate impairment (28%) than other classes (16% minimal, 22% severe). Few differences were observed for other allied health services except for non-pharmacological secondary prevention (dietetics or exercise physiology), where a greater proportion of those with minimal impairment had a claim for these services compared to those in the other two impairment groups (Table 2, Fig. 2).

Fig. 2.

Proportion (%) of cohort with Medicare claim items for Medicare funded enhanced primary care by impairment group.

In multivariable logistic regression, those with moderate impairment were more likely to claim a chronic disease management plan or review (adjusted Odds Ratio (aOR): 1.24, 95% Confidence Interval (CI): 1.10, 1.40) than those with minimal impairment (Table 3). However, those with severe impairment were almost 40% less likely to claim this care type (aOR: 0.61, 95% CI: 0.49, 0.75). Those with moderate (aOR 1.53, 95% CI 1.35, 1.73) or severe impairment (aOR 1.49, 95% CI: 1.23, 1.81) were more likely to claim multidisciplinary care coordination or review than those with minimal impairment. Results for allied health care were similar to those for multidisciplinary care coordination. The overall results remained similar when analyses were restricted to those surviving 18 months post-stroke (Supplementary Table 4).

Using population linked data, we have identified important gaps in how Medicare funded enhanced primary care policies are used to support people living with stroke. Only 39% of people living with stroke had a Medicare claim for chronic disease management or multidisciplinary care coordination in the 18 months following stroke. Those with greatest impairment, who are likely to benefit most, were less likely to have claimed a chronic disease management plan than those with lesser impairments. There was little variation in claims for Medicare subsidised allied health services based on impairment, with claims predominantly used for podiatry, indicating that these items are not being optimally prescribed. The exception being allied health claims for secondary prevention services which were most likely to be accessed by those with minimal impairment.

Enhanced primary care policies have the potential to comprehensively support people with stroke to be managed according to clinical guidelines [6]. This potential benefit is supported by evidence of improved adherence to recommended care for diabetes [18] and reduced hospital readmissions for heart failure in Veteran populations [19]. Despite these potential benefits, our results highlight that Medicare incentivised policies directed at enhanced primary care are underutilised in people living with stroke. It is also concerning that claims for chronic disease management plans varied by impairment group with those with severe impairment being most disadvantaged. This is particularly concerning given the greater prevalence of cardiovascular disease in those with disability or frailty [20, 21]. Sub-optimal management of those with impairment may contribute to the increase in hospitalisation rates, as observed in a prior study, compared to those without impairment [17]. We provide preliminary evidence of gaps in primary care following stroke which may contribute to variability in health outcomes and inequity in access to recommended care.

According to clinical guidelines people with chronic impairments following stroke should be reviewed annually, and if identified as requiring further rehabilitation be referred to therapy services to set new goals and improve function [6]. In Australia, where annual follow-up with a stroke specialist is not standard practice, multidisciplinary care coordination provides an opportunity for Primary Care Practitioners to perform this role. These arrangements have been associated with greater self-assessed quality of care [22] and can support these processes to improve outcomes for people living with long-term impairment [23, 24]. The greater rates of use in those with moderate and severe impairment indicates that these Medicare items are generally being claimed appropriately, despite the overall uptake being sub-optimal.

Multidisciplinary care coordination plans only allow claims for five visits per year across all eligible allied health services [8]. Nevertheless, there is potential for these to enhance or subsidise allied health services provided through other avenues such as private insurance or community health. Consideration should be given by Primary Care Practitioners and patients as to which allied health services are most appropriate to address the patient’s health and impairment needs. This is important because referral to the most appropriate profession can lead to better patient outcomes, including reduced preventable hospital readmissions [25]. In our study, the most commonly claimed profession was podiatry, being more than double the second most commonly claimed profession, physiotherapy. While this finding is consistent with previous evaluations in older Australian populations, the lack of diversity in allied health prescribing is concerning [25]. In prior Australian surveys 84% of people living with stroke reported a range of unmet health needs [1, 2] indicating the need for improved access to allied health care that addresses stroke related impairments. Although it is possible that registrants were accessing allied health services through other avenues not captured in Medicare claims, our results suggest that Medicare-subsidised allied health items are not being optimally used to meet the needs of people living with stroke.

In addition to facilitating ongoing rehabilitation following stroke, the multidisciplinary care coordination plan can be used to refer for non-pharmacological management of risk factors for recurrent stroke such as high cholesterol, high blood pressure and physical inactivity [26]. Guidelines recommend that all people living stroke should receive non-pharmacological management of risk factors, including advice and support on maintaining a healthy diet and recommended physical activity levels [6]. Similar to variations in access to chronic disease management plans, those with severe impairment were less likely to claim for dietetic and exercise physiology services than those with mild or moderate impairment, further disadvantaging this group with regards to secondary prevention management. Again, this suggests potential biases or disproportionate gaps in secondary prevention management in those with severe impairments.

The strengths of this study include a large nationally representative sample of people with stroke from urban and rural locations with collection of quality of life outcome data to classify impairment at 90–180 days post stroke. The availability of personal information (name, surname, date of birth) from AuSCR supported person-level linkages with Medicare claims data for reliable exposure classification. The main limitation is that our findings are restricted to services claimed through Medicare. Some of the variability in care observed in those with severe impairment may be due to a greater probability of people within our Medicare group receiving end of life care. However, this is likely to account for only a small proportion of our cohort as we excluded those residing in residential care facilities.

The use of Medicare funded policies has potential to support Primary Care Practitioners to manage people with stroke according to clinical guidelines and fill gaps in long-term care for those with ongoing impairments. Despite this, uptake is poor for those living with stroke and variability exists with regards to how these items are prescribed. As such, it is unlikely that these models of care are being optimally used to address the ongoing rehabilitation needs of people with stroke, or assist in the prevention of secondary stroke. Further work is underway, through the PRECISE study [27] to determine the impacts on long-term patient outcomes, with potential to modify uptake and facilitate expansion of these programs by government to better address the needs of people living with stroke.

All authors contributed to the interpretation of the data and critical revisions of the manuscript for intellectual content. DAS was also responsible for drafting the manuscript and DU and TAC performed the statistical analysis. NEA was responsible for the initial conceptualisation and design of this study and NEA, NAL, MFK, DAC, AGT and VS acquired project funding to facilitate acquisition and linkage of data and provided critical input into writing the manuscript. All authors approved the manuscript for submission.

Ethics approvals for this study were obtained from Monash University (#7864) and the Australian Institute of Health and Welfare (#EO2017/1/346). Approvals were also received from the AuSCR Research Task Group, the AuSCR Steering Committee, and the Queensland Public Health Act.

We acknowledge the members of the AuSCR Steering Committee, staff from The George Institute for Global Health and the Florey Institute of Neuroscience and Mental Health, who manage the AuSCR; the hospital clinicians and patients who contributed data to the AuSCR; and the data linkage team from the Australian Institute of Health and Welfare who linked and provided the Medicare and pharmaceutical data for this project.

This study was funded by the Stroke Foundation (Australia) through a competitive Seed Grant. The following authors received Research Fellowship support from the National Health and Medical Research Council or the National Heart Foundation of Australia during this research: NEA (1072053), MFK (1109426), AGT (1042600), DAC (1063761 co-funded Heart Foundation; 1154273). NAL received Research Fellowship support from the National Heart Foundation of Australia during this research (102055). The AuSCR is supported by funds or resources from the Australian Government, the Florey Institute of Neuroscience and Mental Health, the Stroke Foundation (Australia), Monash University, consumer donations, and industry.

The authors declare no conflict of interest.