Using record linkage to test representativeness of an ageing cohort

Abstract

Background with rationale The numbers of older people are rising faster than any other group in society and Governments need to develop policies that will help older people stay as healthy and independent as possible for as long as possible. To do this they need to know what shapes the lives and health of older people and how they respond to their changing social, financial and health circumstances. Many countries are now investing in large and expensive ageing cohorts as omnibus studies to better understand the ageing processes and perhaps indicate how society can best prepare for an ageing population. These ageing cohorts (of people aged 50 and over) are invariably drawn from population-wide sampling frames and claim to be representative of the older populations, despite limitations in the sampling frame and modest response rates. Aim The aim of this study was to examine the representativeness of the Northern Ireland Cohort for the Longitudinal Study of Ageing (NICOLA) cohort through linkage to administrative health and demographic data. Specifically, to determine if there were any age, sex, or area of residence differences in NICOLA participants and whether they were healthier than the rest of the population in terms of recent hospital attendance or proximity to care home admission and/or death. Logistic regression will be used to compare characteristics of participants and the rest of the Northern Ireland population aged 50 and over. Data/Methods An identifier for each of the 8504 NICOLA Wave 1 participants was attached to the Health-card registration system which contains the HCN (the unique identifier) and thence to databases holding hospital admissions, mortality records and admissions to a care home. Results The data are currently in the safe setting, with analysis starting this week, and results expected early autumn.