Predictors of mortality in severe mental illness inpatients: A 10-year prospective cohort study

Individuals with severe mental illness have higher mortality rates than the general population. Metabolic parameters and characteristics of the psychiatric illness may play a role in determining this outcome. We aimed to identify predictors of mortality in individuals with severe mental illness.

This is a 10-year cohort study that included inpatients from a psychiatry unit of a tertiary hospital in Brazil. The primary outcome was death due to natural cause. First, we built a Directed Acyclic Graph to illustrate our research question and evaluate how variables were associated with each other according to literature. Then, we performed LASSO analysis followed by multivariable Cox proportional hazards regression models to determine mortality predictors.

We included 416 patients, of whom 21% died during the 10 year-follow-up. The most frequent cause of death was cardiovascular diseases (19.3%). In the final multivariable model, older age (HR = 1.066, 95% CI 1.043-1.090) and tobacco use (HR = 2.135, 95% CI 1.186-3.844) were predictors of mortality, whereas higher educational level (HR = 0.924, 95% CI 0.873-0.978) was a protective factor against mortality. No significant associations were found between metabolic syndrome and severity of psychiatric illness and mortality.

In this study, older age and tobacco use were predictors of mortality, whereas higher education was a protective factor. We found no association between metabolic syndrome and severity of psychiatric illness and mortality in our sample.