Background Intrinsic capacity (IC) is a relatively new concept for measuring health and capacity in older age. Social isolation has been associated with a greater risk of negative health outcomes for older people. However, the relationship between social isolation and IC has not yet been explicitly investigated, particularly in longitudinal analyses. Thus, this analysis aimed to explore whether social isolation was associated with the level and rate of change in IC over time, using the English Longitudinal Study of Ageing (ELSA).

Methods Social isolation and IC were measured at 3 time points (waves 2, 4, and 6) covering 8-years of follow-up. Item response theory was used to generate a total IC score at each time point from 14 indicators of capacity covering the 5 domains of IC: word recall and orientation in time (cognition); balance test, chair rise test, walking speed, lower mobility, and upper mobility (locomotion); self-rated eyesight and hearing (sensory); grip strength, BMI, and waist circumference (vitality); CES-D scale and Satisfaction with Life Scale (psychological) (Campbell et al., 2022). A social isolation score (0-7) was generated from seven indicators, all given 1 point: living alone, less than monthly contact with children, family, or friends, not being a member of any organisations, not working, and not volunteering. A latent growth curve model was used to model the trajectory of IC over the 3 waves and test whether social isolation significantly predicted the baseline level (intercept) and rate of change (slope) of IC score. Covariates included age, sex, education, wealth, smoking, alcohol consumption, physical activity, health conditions, and self-rated health. Full information maximum likelihood was used, resulting in a total sample of 7,690 individuals.

Results When fully adjusted, social isolation was significantly negatively associated with the intercept of IC score (b=-0.65 (SE 0.09), p<0.001) and was positively associated with the slope (b=0.05 (SE 0.02) p=0.008). Those with higher social isolation scores had lower IC scores but had a slower rate of decline than those with lower social isolation.

Conclusion Less isolated individuals tended to experience a better level of IC but saw a steeper decline over 8 years than socially isolated individuals. It is unlikely that being less isolated ‘causes’ a steeper decline in IC but that isolated individuals have ‘less to lose’ and remain at a consistently lower level of IC than non-isolated individuals who start with a higher level of IC and show a steeper decline over time.