Patterns of multimorbidity in the general population of São Paulo, Brazil: a retrospective observational study

Abstract

BackgroundSeveral patterns of multimorbidity have been identified in the general population worldwide: chronic diseases tend to cluster into distinct forms of co-occurrence associated with higher mortality, disability, and functional decline. We aimed to depict these patterns in a large urban sample of the general population to establish the effect of individual and contextual factors for their occurrence. MethodsWe used data from the São Paulo Megacity Mental Health Survey. The investigation used the WHO World Mental Health Composite International Diagnostic Interview in a stratified multistage area probability sampling of 5037 individuals aged 18 years or older. Trained interviewers asked participants face-to-face about common chronic physical diseases (cardiovascular disease, diabetes, insomnia, headache or migraine, respiratory disease, chronic pain, arthritis, neurological disease, and cancer) and mental disorders (depression, anxiety), substance misuse, heavy drinking, and tobacco use. We investigated bivariate comorbidity through a tetrachoric correlation matrix. We assessed patterns of multimorbidity for all conditions by exploratory factor analysis. We then used multilevel models adjusted for sex, age, education, income, local residence, area-level income inequality, and violence, for each resulting factor to establish their effect in these patterns of morbidity. FindingsWe identified four patterns of multimorbidity through factor analysis: (1) anxiety, depression, or insomnia, head or migraine, and respiratory diseases (12·9% of data variance explained); (2) cardiovascular or metabolic conditions, chronic pain, and arthritis (10·6%); (3) substance misuse, heavy drinking, and tobacco use (9·6%); and (4) neurological disorders and cancer (7·6%). After controlling for individual factors, the first multimorbidity was most common among women, whereas the third combination was more prevalent in men. The second multimorbidity was associated with increased age and living in areas with high income inequality. The multilevel analysis also identified the importance of the place of residence for all four patterns of multimorbidity. InterpretationSignificant co-occurrence of chronic diseases in São Paolo can be subsumed into four main patterns of multimorbidity. The effects of sex, age, and place of residence in the clustering of common diseases suggest the interaction of complex individual and contextual factors. Recognition of these patterns could help to predict needs and organise health systems. FundingState of São Paulo Research Foundation, National Council for Scientific and Technological Development, Vitória Foundation of Science and Technology.