Baseline characteristics and management have changed over time in patients requiring mechanical ventilation; however, the impact of these changes on patient outcomes is unclear.To estimate whether mortality in mechanically ventilated patients has changed over time.Prospective cohort studies conducted in 1998, 2004, and 2010, including patients receiving mechanical ventilation for more than 12 hours in a 1-month period, from 927 units in 40 countries. To examine effects over time on mortality in intensive care units, we performed generalized estimating equation models.We included 18,302 patients. The reasons for initiating mechanical ventilation varied significantly among cohorts. Ventilatory management changed over time (P < 0.001), with increased use of noninvasive positive-pressure ventilation (5% in 1998 to 14% in 2010), a decrease in tidal volume (mean 8.8 ml/kg actual body weight [SD = 2.1] in 1998 to 6.9 ml/kg [SD = 1.9] in 2010), and an increase in applied positive end-expiratory pressure (mean 4.2 cm H2O [SD = 3.8] in 1998 to 7.0 cm of H2O [SD = 3.0] in 2010). Crude mortality in the intensive care unit decreased in 2010 compared with 1998 (28 versus 31%; odds ratio, 0.87; 95% confidence interval, 0.80-0.94), despite a similar complication rate. Hospital mortality decreased similarly. After adjusting for baseline and management variables, this difference remained significant (odds ratio, 0.78; 95% confidence interval, 0.67-0.92).Patient characteristics and ventilation practices have changed over time, and outcomes of mechanically ventilated patients have improved. Clinical trials registered with www.clinicaltrials.gov (NCT01093482).