What does this project tell us?

This project describes some of the common treatment pathways people take following an acute coronary syndrome (ACS) hospitalisation. Linked administrative data enables effective secondary use of data to explore patient’s real-world journeys through the health system. It provides evidence on how many people appear to align with existing clinical guidelines and identifies high level characteristics of those who do and don’t take specific treatment pathways.

Monitoring patient’s real world ‘treatment pathways’ is important to understanding guideline implementation, particularly as guidelines change and evolve over time. The 2022 Australian Institute of Health and Welfare (AIHW) publication Medication use for secondary prevention after coronary heart disease hospitalisations: patient pathways using linked data (AIHW, 2022) provides further detail about variation in the use of medication recommended by guidelines (Chew et al. 2016) and highlights subpopulations that would benefit from additional ongoing support after their transition from hospital to the community.

While this project describes the health outcomes experienced by those in different treatment subgroups, interventional procedures and medication use are only some of the factors that impact health outcomes. For example, older people with multiple comorbidities may not undergo interventional procedures due to a high risk and may also be more likely to experience poor health outcomes. Further, as this data are collected for administrative purposes, with research being its secondary use, many of the factors that impact both treatment decisions and health outcomes are not captured. For example, there is limited information about clinical factors (such as blood pressure and blood lipid levels), adverse reactions or unacceptable side effects due to in-scope medications, and psychosocial factors. This project cannot identify causal relationships between treatment pathways and health outcomes, but instead provides a starting point for further research and improved understanding of the patient journey.

Limitations of the data

The NIHSI (version 2.0) includes data from 6 of the 8 jurisdictions of Australia and therefore results may not be generalisable to Western Australia and the Northern Territory. Further, private hospital data were available only for Victoria (to June 2017), Queensland and the Australian Capital Territory. As a result, readmissions and the capture of interventional procedures undertaken within 40 days of index hospitalisation will be underestimated. The inclusion of only emergency readmissions may somewhat mitigate the impact on the readmission outcome measure.

The reason for discontinuing a medication is not captured in the administrative data. For example, it may be due to a patient’s choosing, or may be done in consultation with a medical professional due to adverse side effects or a new therapeutic approach. In addition, an assumption of the analysis is that PBS dispensing records accurately represent medication use. However, it is possible that a person may be dispensed a medicine but does not take the full course or fails to take it as prescribed by their health professional.

Medications supplied in public hospitals are not available in this data. Time spent admitted to public hospitals was excluded from the measurement of persistence as the supply of medications in this setting would not be captured in the PBS data. The use of 40 days post index hospitalisation to measure initiation was selected to account for up to one month’s supply from before the hospitalisations and supplies from the hospital which are not captured in the PBS for all jurisdictions.

PBS data do not include any information about the intended dosage; as a result, this information was estimated for the cohort. Incorrect dosage assumptions may impact persistence estimates.

Aspirin, an antiplatelet agent, was excluded from the analysis as it is available over the counter and is therefore not comprehensively captured in the PBS data set. While the guidelines recommend dual antiplatelet therapy (aspirin and another antiplatelet), those who only took aspirin will be classified as not initiating and persisting to all 4 guideline indicated medication classes.

Comorbidity information is obtained from the admitted patient care data, from hospitalisations prior to the index and the index hospitalisation, and relies on valid and accurate coded data. The identification of comorbidities is likely to be an underestimate of disease burden in the population. Further, multimorbidity, which is an indicator of more complex health conditions and treatment needs is not captured.

In this analysis, Indigenous status was derived from the admitted patient care data recorded at the index hospitalisation only. This may be an underestimate of the First Nations population in the cohort.