Background There has been great interest in predictors of mortality in older cohorts in the international literature in recent years. The Lifeways Cross-Generation Cohort Study is a longitudinal birth cohort study including adult family members with detailed baseline objective and subjective health and socio-demographic measures, including a means-tested measure of General Medical Services (GMS) eligibility and comprehensive mortality follow-up to 2014. One of the chief study objectives was to establish the influence of healthcare availability on health outcomes.1
Methods 1223 Lifeways grandparents in 631 families participated in the baseline recruitment questionnaire and/or health examination. Mortality follow-up was achieved using national death records from the General Registry Office. A generalised linear mixed model estimated between couple and between family variance in odds of death to be 0, thus standard logistic regression was used. Predictors that were significant at univariable level (p < 0.10) and for which data were available on >75% of participants were entered into the multivariable models, with age, sex, smoking status retained a priori. Analysis was carried out using IBM SPSS Statistics v.20.
Results Mean age at baseline was 59.7 (37.2–89.7); 124 deaths were confirmed within the 10-year follow-up period. In univariable analysis, the odds ratios and 95% confidence intervals (in parentheses) for the significant (p < 0.10) variables were as follows: Socio-demographic factors: male sex, 2.043 (1.405–2.970); age, 1.149 (1.120–1.178); GMS eligibility, 4.579, (2.616–8.014) and being married, 0.486 (0.290–0.817). Clinical factors: worse mean self-rated health, 1.564 (1.149–2.129); waist circumference, 1.029 (1.013–1.045); total cholesterol, 0.749 (0.597–0.941); average systolic blood pressure (SBP), 1.010 (0.999–1.021); reported hypertension, 1.784 (1.221–2.605); reported diabetes, 4.373 (2.460–7.773); reported activity-limiting health condition, 2.470 (1.369–4.459) and frequency of moderate exercise, 0.857 (0.739–0.993). The following are the odds ratios and 95% confidence intervals of all variables in the final multivariable model (Nagelkerke R Square value of 0.340, n = 817, 77 deaths). Age, 1.166 (1.125–1.209); male sex, 1.242, (0.705–2.186); smoker, 2.940 (1.525–5.668), waist circumference, 1.034 (1.012–1.056); average SBP, 0.989 (0.975–1.004); total cholesterol, 0.882 (0.677–1.147), reported hypertension, 2.180 (1.219–3.897); and reported diabetes, 1.947 (0.854–4.438).
Conclusion The analysis confirms clearly that conventional clinical health status variables are significant predictors of mortality at 10-year follow-up in this longitudinal cohort of older Irish adults (explaining approximately 34% of the variance). GMS eligibility is also strongly associated with mortality, which may be because those with established health needs register themselves in order to avail of comprehensive means-tested healthcare. This has policy implications for healthcare provision and accessibility.
Acknowledgement Additional authors from the Lifeways Cohort Study Steering Group.
O’Mahony D, Fallon UB, Hannon F, Kloeckner K, Avalos G, Murphy AW, Kelleher CC; Lifeways Cross Generation Cohort Study Steering Group. The lifeways cross-generation study: design, recruitment and data management considerations. Ir Med J. 2007;100(8 Suppl):3–6