# Stroke incidence in Indigenous, minority populations: a review of methods for studying stroke in Aboriginal and Torres Strait Islander Australians

## TL;DR

This paper reviews methods for measuring stroke incidence in Aboriginal and Torres Strait Islander Australians, highlighting challenges and recommending data linkage approaches for better insights.

## Contribution

The paper proposes a shift toward population-linked data for stroke studies in Indigenous populations, emphasizing practicality over traditional methods.

## Key findings

- Gold-standard methods have small sample sizes for Aboriginal patients.
- Hospital-based studies provide larger samples but face selection bias.
- Population-linked data allows for subgroup analysis but lacks clinical detail.

## Abstract

Declining worldwide or national stroke incidence rates are not always mirrored in disadvantaged, minority populations. Logistical barriers exist for effective measurement of incidence in minority populations; such data are required to identify targets for culturally appropriate interventions. In this comparative review, we aimed to examine whether “gold-standard” methodologies of stroke incidence studies are most effective for minority populations. We compared three studies of stroke incidence in Aboriginal Australians, each using different methodologies of case ascertainment. In Study 1, “gold-standard” population-based methods were used, while in Study 2, a retrospective hospital-based cohort design was utilized, and in Study 3, whole-of-population linked hospital and mortality data was employed. Study 1 captured both in-hospital and out-of-hospital stroke events but had a small sample size for Aboriginal patients. Study 2 provided a larger sample size while still allowing for clinical and radiological subtyping of stroke but was subject to selection bias and was limited to hospitalized cases. Study 3 had a large sample size and allowed for subgroup analysis, though lacked clinical adjudication and had large proportions of ‘undetermined stroke'. Despite diagnostic imprecision, we recommend a paradigm shift in measuring stroke incidence in Indigenous, minority populations, favoring the use of whole-of-population data linkage including non-hospitalized stroke deaths, over resource-intensive prospective methods, where more suitable for the target population.