Defining a study population using enhanced reporting of Aboriginality and the effects on study outcomes

Area of Science:

• Public Health
• Biostatistics
• Epidemiology

Background:

• Under-reporting of Aboriginal and Torres Strait Islander people in health datasets hinders accurate population health assessment.
• Probabilistic linkage of datasets and enhanced reporting methods can adjust for under-reporting.

Purpose of the Study:

• To develop and evaluate algorithms for enhancing the reporting of Aboriginal status in birth data.
• To define a cohort of Aboriginal women and compare characteristics with those solely based on recorded status.
• To assess the impact of different reporting algorithms on within-group comparisons for Aboriginal people.

Main Methods:

• Three inclusive algorithms were applied to New South Wales birth data (2010-2014) to identify singleton births to Aboriginal or Torres Strait Islander women.
• Demographic characteristics of four derived study populations were compared.
• Associations between smoking in pregnancy and perinatal outcomes were analyzed using rates and relative risks.

Main Results:

• Women identified through enhanced reporting were older, less disadvantaged, and more urbanized than those with recorded Aboriginal status.
• While smoking rates and some perinatal outcomes varied across populations, relative risks for these outcomes were consistent.
• Estimates of within-group relative risks remained reliable across different study populations.

Conclusions:

• Enhanced reporting of Indigenous status in health datasets can improve cohort definition.
• Estimates of within-group relative risks for perinatal outcomes are robust regardless of the algorithm used for enhanced reporting.
• This methodology supports more accurate health research for Aboriginal and Torres Strait Islander populations.