Background. An association between disablement and late-life depression is often reported in cross-sectional studies. However, many lack effect sizes, and do not control for confounding. Therefore, it is difficult both to quantify the overall impact of poor health on depression and to understand which aspects are most salient.Methods. A catchment area survey of all over 65-year-old residents of an electoral district in London, UK, using a population register derived from a door-to-door census was undertaken. Depression was measured using SHORT-CARE, and the consequences of disease classified according to the WHO International Classification of Impairments, Disabilities and Handicaps.Results. Six hundred and fifty-four subjects were interviewed out of an older population of 889. The prevalence of SHORT-CARE pervasive depression was 17%. Impairment, disability and, particularly, handicap were strongly associated with depression. The adjusted odds ratio for depression in the most handicapped quartile compared with the least was 24·2 (8·8–66·6). The population attributable fraction (PAF) for depression attributable to handicap was 0·78. The PAFs for recent life events and female gender were much lower. Handicap explained most of the depression associated with individual impairments and disabilities. Adjusting for handicap abolished or weakened the associations between depression and social support, income, older age, female gender and living alone.Conclusions. Even given some uncertainty in distinguishing handicap and depression as constructs, and the impossibility of deciding direction of causality, it seems likely that handicap is of central significance to late-life depression. Handicap may be more amenable to intervention than either impairment or disability.