Variability in the use of protective mechanical ventilation during general anesthesia

BACKGROUND: The purpose of this study was to determine whether significant variation exists in the use of protective ventilation across individual anesthesia providers and whether this difference can be explained by patient, procedure, and provider-related characteristics. METHODS: The cohort consisted of 262 anesthesia providers treating 57,372 patients at a tertiary care hospital between 2007 and 2014. Protective ventilation was defined as a median positive end-expiratory pressure of 5 cm H₂O or more, tidal volume of <10 mL/kg of predicted body weight and plateau pressure of <30 cm H₂O. Analysis was performed using mixed-effects logistic regression models with propensity scores to adjust for covariates. The definition of protective ventilation was modified in sensitivity analyses. RESULTS: In unadjusted analysis, the mean probability of administering protective ventilation was 53.8% (2.5th percentile of provider 19.9%, 97.5th percentile 80.8%). After adjustment for a large number of covariates, there was little change in the results with a mean probability of 51.1% (2.5th percentile 24.7%, 97.5th percentile 77.2%). The variations persisted when the thresholds for protective ventilation were changed. CONCLUSIONS: There was significant variability across individual anesthesia providers in the use of intraoperative protective mechanical ventilation. Our data suggest that this variability is highly driven by individual preference, rather than patient, procedure, or provider-related characteristics.