Mortality in hospital patients with and without mental disorders: A data linkage cohort study

Objective: In the general population, people with mental disorders have increased mortality. This association, however, has not been explored in a population who at some time were inpatients of a public hospital. Methods: The sampling frame was patients admitted to an Australian regional public hospital 1996-2010. Those with known mental disorder were compared with an equal number of randomly selected patients without known mental disorder, matched for age, sex, and year of admission. Mortality outcomes were determined by linkage of hospital data and the National Death Index. Results: We identified 15,356 patients with mental disorder and 15,356 without known mental disorder, 25.2% of the former and 17.3% of the latter died during the study period. The odds ratio (OR) for death in those with mental disorder was 2.20 (95% confidence interval: 2.01-2.41) after adjusting for confounders, and their mean age at death was 4.6 years younger (p < 0.001). Other independent risk factors for mortality were being Indigenous (OR 1.72, 1.32-2.24), not partnered (OR 1.55, 1.36-1.76) or having multiple comorbidities (OR 1.65, 1.43-1.90). In the model that included multiple interactions, the ORs for death in Indigenous patients with mental disorder were markedly higher (6.6-9.5), regardless of other risk factors. Discussion: Among patients with a history of public hospital admission, those with mental disorders have higher mortality than those without mental disorders. This gap in life expectancy mandates increased attention, by clinicians and health services alike, to preventable causes of death in people with mental illness.