Socioeconomic position and functional dependency trajectories across six longitudinal ageing cohorts

Socioeconomic position (SEP) has been associated with functional health in later life; however, whether these associations extend to specific patterns of functional deterioration, recovery, and mortality across different ageing cohorts remains unclear. We analyzed harmonized longitudinal data from six ageing cohorts representing 22 countries: the China Health and Retirement Longitudinal Study (CHARLS), English Longitudinal Study of Ageing (ELSA), US Health and Retirement Study (HRS), Korean Longitudinal Study of Ageing (KLoSA), Mexican Health and Aging Study (MHAS), and Survey of Health, Ageing, and Retirement in Europe (SHARE). The analysis included 104,834 adults aged ≥ 50 years and 454,747 person-wave observations. SEP was defined using a four-level index that combined baseline educational attainment and household wealth. Cohort-specific Cox proportional hazards models were used to examine incident limitations in activities of daily living (ADL) and instrumental activities of daily living (IADL). Continuous-time multistate Markov models estimated transitions among no, mild, moderate, and severe dependency, and where sufficiently reliable mortality data were available, death. Compared with high SEP, low SEP was associated with a higher hazard of incident ADL limitation in all six cohorts. For incident IADL limitation, positive associations were observed in most cohorts, whereas the low-versus-high estimate in MHAS was not statistically significant (hazard ratio 1.31, 95% confidence interval 0.84–2.06). Distribution-standardised relative-index-of-inequality estimates were greater than 1 in all cohort–outcome combinations, ranging from 1.44 to 2.49 for ADL limitation and from 1.34 to 2.78 for IADL limitation. In the multistate analyses, low SEP was associated with higher estimated transition intensities from no dependency to mild dependency and from mild to moderate dependency across the six cohorts, although the magnitude and precision of the estimates varied. At 5 years, the probability of remaining without dependency was 0.641 for low SEP versus 0.873 for high SEP in CHARLS, 0.658 versus 0.841 in HRS, and 0.750 versus 0.890 in ELSA. Alternative SEP specifications produced similar graded patterns in most of the cohorts. In the strict-progressive sensitivity analysis, 46 of 54 deterioration-transition comparisons remained in the same direction as in the primary models, although some mortality estimates were imprecise, and KLoSA confidence intervals were unavailable. Lower SEP was associated with a higher incidence of functional limitations and less favourable subsequent functional state distributions in most cohorts. However, the magnitude and pattern of deterioration, recovery, and mortality associations varied across cohorts and dependency states, indicating that SEP-related differences in functional trajectories were not uniform across settings.