Objectives International evidence suggests that rates of inability to work because of illness can change over time. We hypothesised that one reason for this is that the link between inability to work and common illnesses, such as musculoskeletal pain and mental illness, may also change over time. We have investigated this in a study based in one UK district.

Methods Five population surveys (spanning 2002–2010) of working-age people aged >50 years and ≤65 years were used. Work disability was defined as a single self-reported item ‘not working due to ill-health’. Presence of moderate–severe depressive symptoms was identified from the Mental Component Score of the Short Form-12, and pain from a full-body manikin. Data were analysed with multivariable logistic regression.

Results The proportion of people reporting work disability across the surveys declined, from 17.0% in 2002 to 12.1% in 2010. Those reporting work disability, one-third reported regional pain, one-half widespread pain (53%) and two-thirds moderate–severe depressive symptoms (68%). Both factors were independently associated with work disability; their co-occurrence was associated with an almost 20-fold increase in the odds of reporting work disability compared with those with neither condition.

Conclusions The association of work disability with musculoskeletal pain was stable over time; depressive symptoms became more prominent in persons reporting work disability, but overall prevalence of work disability declined. The frequency and impact of both musculoskeletal pain and depression highlight the need to move beyond symptom-directed approaches towards a more comprehensive model of health and vocational advice for people unable to work because of illness.