Validation of administrative definitions of invasive mechanical ventilation across 30 intensive care units

Many insights into the etiology and outcomes of respiratory failure have been gleaned from studies using administrative data. However, the validity of identifying patients receiving invasive mechanical ventilation (IMV) in such datasets is not well established. One widely cited study in support of claims-based IMV definitions was limited to three centers within one Canadian provincial system and included only 46 ventilated patients. Another single-center study addressed only the positive predictive value of administrative definitions. We sought to assess a more complete array of operating characteristics for several administrative definitions of IMV and to characterize the patients best identified by these definitions among a larger and more diverse sample.

National Institute of General Medical Sciences
National Heart, Lung, and Blood Institute