Past research has demonstrated the high prevalence of depression in elderly. However, the most of studies followed the symptom trajectory of individuals diagnosed with depression in a clinical setting and few longitudinal studies have characterized the patterns of depression in older adults population-based.

To describe changing of depressive disorder in an elderly population-based over a 12-month period and to examine the influence of medical and psychosocial factors on the outcome.

Data come from a longitudinal ESA Study (Enquête sur la Santé des Aînés) of elderly community persons (n = 2752). Depression, including major and minor depression, measured using the DSM-IV criteria. Generalized estimating equations (GEE) were used to assess relations between participant characteristics at baseline and depression, 12 months later.

Among the 164 (5.9%) participants, who are depressed at baseline, 19.5% were continuously ill cases and 80.4% had recovered, 12 months later. Multivariate analyses showed that the risk of depression over the 12-month period was higher among for participants who were separated; living in rural region; with a great number of daily hassles, with high level of stress intensity, great number of chronic disease and with fair/poor perception of mental health.

Results support the hypothesis about medical and psychosocial factors as predictors over time of depression, in old persons. Using readily available prognostic factors (for example, high level of stress intensity, living in rural region, great number of chronic disease) could help direct treatment to elderly at highest risk of a poor prognosis.