RESUMEN
BACKGROUND: This paper explores the applicability of a psychosocial model of depression in an African setting. METHOD: Of a random sample of women (n = 172) from a Zimbabwean township, 79 had a severe life event in the year before interview. Twenty-nine who had an onset of depression were compared with 50 who did not. RESULTS: Having crisis support following a severe life event reduced the risk of onset of depression. The effect of crisis support was confounded by the number of severe events in the study year, but just persisted following adjustment for number of severe events and for socio-demographic factors. Women received crisis support more often from relatives than from partners or friends. The following variables were independently associated with onset of depression following a severe life event: number of life events in the previous 6 months, age, being in formal employment, having crisis support and separation from mother in childhood. CONCLUSION: This study supports the ubiquity of the social support construct. For Zimbabwe, the data suggest that crisis support from family members may be of particular importance in protecting against onset of depression.