| Literature DB >> 15246165 |
Naomar Almeida-Filho1, Ines Lessa, Lucélia Magalhães, Maria Jenny Araújo, Estela Aquino, Sherman A James, Ichiro Kawachi.
Abstract
We conducted a study of the association between gender, race/ethnicity, and social class and prevalence of depressive disorders in an urban sample (N = 2302) in Bahia, Brazil. Individual mental health status was assessed by the PSAD/QMPA scale. Family SES and head of household's schooling and occupation were taken as components for a 4-level social class scale. Race/ethnicity (white, moreno, mulatto, black) was assessed with a combination of self-designation and a system of racial classification. The overall 12-month prevalence of depressive symptoms was 12%, with a female:male ratio of 2:1. Divorced/widowed persons showed the highest prevalence and single the lowest. There was a negative correlation with education: the ratio college educated:illiterate was 4:1. This gradient was stronger for women than men. There was no F:M difference in depression among Whites, upper-middle classes, college-educated, or illiterate. Prevalence ratios for single, widowed and Blacks were well above the overall pattern. Regarding race/ethnicity, higher prevalences of depression were concentrated in the Moreno and Mulatto subgroups. There was a consistent social class and gender interaction, along all race/ethnicity strata. Three-way interaction analyses found strong gender effect for poor and working-class groups, for all race/ethnicity strata but Whites. Black poor yielded the strongest gender effect of all (up to nine-fold). We conclude that even in a highly unequal context such as Bahia, Blacks, Mulattos and women were protected from depression by placement into the local dominant classes; and that the social meaning of ethnic-gender-generation diversity varies with being unemployed or underemployed, poor or miserable, urban or rural, migrant or non-migrant.Entities:
Mesh:
Year: 2004 PMID: 15246165 DOI: 10.1016/j.socscimed.2003.11.037
Source DB: PubMed Journal: Soc Sci Med ISSN: 0277-9536 Impact factor: 4.634