Childhood adversity and disadvantage in relation to incident disability and multimorbidity: prospective analyses of ELSA and two SHARE inception samples.

BACKGROUND

Childhood health adversity and socioeconomic disadvantage may shape late-life health, but evidence for incident disability and multimorbidity remains limited. Cross-cohort inference is complicated by non-equivalent measures of childhood experiences and differences in timing. We estimated cohort-specific prospective associations in the English Longitudinal Study of Ageing (ELSA) and the Survey of Health, Ageing and Retirement in Europe (SHARE).

METHODS

ELSA childhood-health/material-adversity indicators and health were assessed at wave 3. SHARE material/cultural disadvantage was assessed at wave 3 or wave 7, with health landmarks at waves 4 and 8; the two inception samples were analyzed separately. Outcomes were the first observed onset of difficulty in at least one of six activities of daily living (ADL-6) and the first observed onset of at least two of seven disease groups (MM-7). Complementary log-log person-period models incorporated interval duration, repeated observations, available survey-design information and baseline weights, and stabilized inverse-probability-of-censoring weights; death was a competing event. Highest exposure categories were compared with zero, without assuming measurement equivalence.

RESULTS

Combined survey/attrition-weighted models included 4,290/3,424 ELSA participants, 13,863/9,064 SHARE wave-3 participants, and 23,026/12,560 SHARE wave-7 participants for ADL-6/MM-7, respectively. In the mortality-ascertained ELSA wave 4-6 window, highest-versus-zero hazard ratios (HRs) were 1.43 (95% confidence interval [CI] 1.09-1.87) for ADL-6 and 1.35 (1.02-1.80) for MM-7. After excluding ADL-6 onset observed at wave 4, the lagged ELSA HR was 1.03 (0.70-1.52) through wave 6. Over waves 4-9, HRs were 1.47 (1.19-1.82) and 1.24 (0.98-1.57). SHARE wave-3 HRs were 1.36 (1.08-1.71) and 1.34 (1.09-1.64); wave-7 HRs were 1.60 (1.06-2.43) and 0.93 (0.68-1.27). Standardized competing-risk estimates showed a higher risk of ADL-6 in the primary ELSA analysis and both SHARE inception-sample analyses. MM-7 risk differences were positive in ELSA and SHARE wave 3, whereas the SHARE wave-7 contrast was imprecise and compatible with no association.

CONCLUSION

The highest cohort-specific exposure category versus zero was associated with incident ADL disability in the primary ELSA and SHARE landmark analyses, but the ELSA association appeared strongest in the earliest interval and was not clearly sustained later. Associations with multimorbidity varied by observation window and SHARE inception sample. The findings neither establish measurement equivalence nor explain the observed cross-cohort heterogeneity.