Objective: To estimate incidence of newly diagnosed mental disorders among ICU patients.
Design: Retrospective-matched cohort study using a population-based administrative database.
Setting: Manitoba, Canada.
Participants: A total of 49,439 ICU patients admitted between 2000 and 2012 were compared with two control groups (hospitalized: n = 146,968 and general population: n = 141,937), matched on age (± 2 yr), sex, region of residence, and hospitalization year.
Intervention: None.
Measurements and main results: Incident mental disorders (mood, anxiety, substance use, personality, posttraumatic stress disorder, schizophrenia, and psychotic disorders) not diagnosed during the 5-year period before the index ICU or hospital admission date (including matched general population group), but diagnosed during the subsequent 5-year period. Multivariable survival models adjusted for sociodemographic variables, Charlson comorbidity index, admission diagnostic category, and number of ICU and non-ICU exposures. ICU cohort had a 14.5% (95% CI, 14.0-15.0) and 42.7% (95% CI, 42.0-43.5) age- and sex-standardized incidence of any diagnosed mental disorder at 1 and 5 years post-ICU exposure, respectively. In multivariable analysis, ICU cohort had increased risk of any diagnosed mental disorder at all time points versus the hospitalized cohort (year 5: adjusted hazard ratio, 2.00; 95% CI, 1.80-2.23) and the general population cohort (year 5: adjusted hazard ratio, 3.52; 95% CI, 3.23-3.83). A newly diagnosed mental disorder was associated with younger age, female sex, more recent admitting years, presence of preexisting comorbidities, and repeat ICU admission.
Conclusions: ICU admission is associated with an increased incidence of mood, anxiety, substance use, and personality disorders over a 5-year period.