Association between hobby engagement and frailty among adults aged 65 years and older in 15 countries: findings from four prospective cohort studies.

Background: While frailty is a major public health concern in ageing populations, evidence on its relationship with hobby engagement among older adults remains limited. Existing studies have largely been limited to single-country settings and specific hobby types, with little evidence on cross-national generalisability or whether engagement frequency is associated with additional benefit. Methods: We conducted secondary analyses of harmonised longitudinal data from four nationally representative ageing cohorts conducted across 15 countries between 2010 and 2023: the China Health and Retirement Longitudinal Study, the English Longitudinal Study of Ageing, the US Health and Retirement Study, and the Survey of Health, Ageing and Retirement in Europe. We included participants aged ≥65 with baseline data on hobby engagement, frailty, and covariates, and who had frailty assessed at follow-up using the frailty index (FI). We used linear mixed models and random-effects meta-analyses to estimate the association between hobby engagement and FI, with average marginal effects (AMEs) and their 95% confidence intervals (CIs) reported. We further examined whether this association varied across subpopulations and whether higher engagement frequency was associated with additional benefits, and performed sensitivity analyses to assess the robustness of our findings. Results: Overall, 43,346 participants from 15 countries were included in the binary hobby-engagement analysis. The hobby engagement frequency analysis included 31,538 participants from 14 countries, as harmonisable information on engagement frequency was unavailable in ELSA. Hobby engagement was associated with lower FI (pooled AME = - 0.044; 95% CI = -0.054, -0.034; P < 0.001). Similar associations were consistently observed across age, education, smoking history, alcohol drinking, and physical activity levels, with AMEs from -0.032 to -0.058 and P-values <0.05, with no statistically significant interactions detected. Analyses by frequency suggested that hobby engagement at least monthly was associated with a significantly lower FI compared with less than once per month (pooled AME = 0.003; 95% CI = -0.006, -0.001; P = 0.028), but higher frequencies, either weekly (P > 0.05) or daily (P > 0.05), did not provide any additional benefits. Sensitivity analyses confirmed the robustness of our findings. Conclusions: Hobby engagement was associated with lower subsequent FI among adults aged ≥65 years. The association appeared to plateau beyond a modest engagement level, suggesting that hobby engagement at an appropriate frequency may represent a generalisable and low-cost behavior associated with frailty risk.