To (1) explore clinical and demographic characteristics of users and nonusers of PPIs in 3 cohorts of institutionalized older people in Finland, and (2) compare the risk of death associated with use of PPIs in each setting. Cross-sectional assessment of 3 institutionalized cohorts with 1-year follow-up of all-cause mortality. A total of 1389 residents of 69 assisted living facilities (first cohort), 1004 residents of long term care hospitals (second cohort), and 425 residents in acute geriatric wards or in nursing homes (third cohort). Demographic, drug use, and diagnostic data were collected during structured assessments conducted by trained nurses or geriatricians. Cox proportional hazards models were used to compute unadjusted and adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between use of PPIs and mortality. In the assisted living facility, the mortality was 20.2% (n = 74) and 20.4% (n = 208) among users and nonusers of PPIs, respectively (P = 0.94). PPIs were not associated with mortality in unadjusted or adjusted analyses. In the long term care hospitals, use of PPIs was associated with increased mortality (HR, 1.36; 95% CI 1.04-1.77) when adjusted for age, sex, comorbidity, use of SSRIs, and malnutrition. In the acute geriatric wards and nursing homes, use of PPIs was associated with increased mortality (HR, 1.90; 95% CI 1.23-2.94) when adjusted for age, sex, comorbidity, delirium, and use of aspirin and SSRIs. PPIs were not associated with mortality among residents in assisted living facilities, but were associated with increased mortality in settings where residents experienced higher levels of disability and possible susceptibility to adverse drug events.