Inequalities and inequities in mental health and care

This dissertation aims at analyzing the relationship between socio-economic status (SES) and mental health and care. It attempts to understand how different socio-economic groups present unequal risk of mental disorders and to what extent different socio-economic groups use unequal quantity, type and quality of mental care. Since its earlier beginnings, psychiatric epidemiology has evidenced the association between socio-economic status and mental disorder. However, the numerous prevalence studies addressing depression have yielded inconsistent results. This calls for a thorough investigation of the sources of such heterogeneity. This dissertation attempts to achieve the following objectives: . To unfold methodological and contextual covariates influencing the SES/mental health relationship. . To assess the longitudinal influences of material deprivation on depression. . To assess the extent to which outpatient and inpatient mental care are fairly used. The methodological influences of socio-economic inequalities in mental health were tackled through a meta-analysis of previously published works. We built a database of previous published studies addressing the socio-economic factors of depression prevalence, incidence and persistence in adults population studies and being published in English, French, German and Spanish after 1979. The lower socio-economic group has 80% more prevalence of depression. Inequalities are more acute for persistent depression than for new episode. The results indicated that inequalities are much more pronounced when mental health is looked at from a subjective point of view or in terms of resulting disability. Social inequalities in mental health are also influenced by geographic context. Europe has a gradient 30% less pronounced than North-America. As the period of reference decreased, the gradient rose, suggesting that duration might be an explanatory factor. Geographical analysis of socio-economic inequalities in mortality is carried out with the death certificates of the Belgian National Institute of Statistics (NIS), covering all causes mortality and 11 specific mortality causes, from 1985 to 1993. Spatial concentration was computed through a Moran ? I. We compare a simultaneous autoregressive model with a weighted-least-square model. Findings show that spatial concentration is pervasive, that suicide and mortality by liver cirrhosis are among the most correlated causes of death. Getting rid of spatial autocorrelation leads to significant change in the relationship between deprivation and mortality, suggesting the influence of contextual effects on socio-economic inequalities. The difficulty to move from correlation to causation between SES and depression owes partly to the difficulty of disentangling the direct effect of socio-economic status from other ?and numerous- confounding factors such as family history, genetic endowment, cognitive abilities, early schooling experience, which, for most of them are rather stable overtime. The longitudinal analysis attempts to estimate the impact of time-varying socio-economic covariates on depression. The results show that material deprivation (and change of) does not affect the level or the risk of depression while social network does slightly. We found much stronger gradient with time invariant socio-economic factors such as educational level.[...] (ESP 3) -- UCL, 2002