Background In 2001 the World Health Organisation (WHO) proposed a conceptual framework for describing functioning and disability: the International Classification of Functioning, Disability and Health (ICF). ICF conceives and organises disability as a combination of three components: Impairment, Activity and Participation. Few studies have adopted this comprehensive approach to measure disability. The object of this work is twofold: to assess the validity of the ICF framework to measure functioning and disability when it is applied to older population; and to investigate the role of disability in explaining all-cause mortality among individuals aged 50+.

Methods Data were drawn from the English Longitudinal Study of Ageing (ELSA). 9,715 subjects aged 50+ interviewed at wave 1 (2002–2003) and followed-up for about 10 years were studied. The items used to measure impairment, activity and participation were selected and agreed in double-blind by three researchers. Confirmatory factor analysis (CFA) was used to determine to what extent the ELSA data fit the three-factor WHO’s ICF framework. The association of disability with mortality was estimated using the measure of disability derived from the measurement model and applying discrete-time survival analysis in a general latent variable framework. All analyses were performed using Mplus.

Results From CFA we identified 13 items strongly associated (factor loadings >0.4) to Impairment (I), 19 to Activity (A) and 8 to Participation (P); we also identified a specific latent eyesight variable within the impairment domain, composed by three eyesight items. The second order CFA model showed that the three-factor (I, A and P) disability structure has an acceptable fit using data on a sample of older individuals (model fit: CFI = 0.945, TLI = 0.942, RMSEA = 0.042). From the discrete time survival analysis, we found strong and significant time-varying effect of disability on mortality, which persisted throughout the follow-up (OR in the first time-event equal to 3.26, p < 0.001; OR in the last time-event equal to 2.23.). Adjusting for demographic characteristics, socioeconomic factors and health-related behaviours the association was attenuated, but overall the effect of disability on mortality remained high and significant (OR in the first time-event equal to 1.94, p = 0.016; OR in the last time-event equal to 1.75, p < 0.001).

Conclusion Our work confirmed the validity of the ICF framework to capture disability among the older population and identified the more relevant factors for each component. Disability, framed in such a way, was strongly associated with mortality and the effect was quite stable throughout the course of the 10 year follow-up.