Pain and self-rated health in the very old: Individual participant data meta-analysis from five TULIP Consortium cohorts

Background In adults and older adults, pain is inversely associated with self-rated health (SRH) independent of health status. Less is known about this association in the very old. This paper aims to determine how pain is associated with SRH in very old people with and without adjustment for individual health status. Methods Using data from five cohorts of very old people aged 80+ from different cultural contexts (n=2,664), we separately fitted single- and multi-level logistic models of SRH to examine cross-sectional and longitudinal associations with pain, respectively. We performed random effects meta-analysis to combine cohort-specific odds ratios (ORs) of positive SRH (selecting response options ‘good’ and above) for those reporting any pain in the past month, at baseline and then over time for up to 10 years follow-up, with adjustment for sociodemographic (age, sex, income status) and individual health (multimorbidity, depressive symptoms, functional ability) variables. Results Over half of participants across cohorts (55.4%) reported having pain and 71.1% had positive SRH at baseline. At baseline, the odds of positive SRH were significantly lower in those with pain in all cohorts when we adjusted for sociodemographic variables (combined OR 0.37; 95% CI 0.28, 0.48), which persisted when we added health-related variables to the models (combined OR 0.42; 95% CI 0.34, 0.54). In longitudinal analysis, over time pain remained significantly associated with lower odds of positive SRH throughout follow-up (combined OR 0.45; 95% CI 0.35, 0.58) after adjusting for sociodemographic and individual health variables. Conclusion Pain is common in very old people, and it is associated with worse SRH independent of health status at baseline and over time. This inverse association is consistent across cultures and time, highlighting the importance of addressing pain in this age group.