Ventilator-associated pneumonia: depends on your definition

Abstract

Reduction of hospital-acquired infections is a patient safety goal and regularly monitored by Performance Improvement committees. There is discordance between the ventilator-associated pneumonia (VAP) rate reported by the Infection Control Committee (ICC) and that observed by our Trauma Service. To investigate this difference, a retrospective evaluation of cases of VAP diagnosed on a single service was undertaken. A prospectively collected database was queried for VAP in intensive care unit patients between January 2010 and June 2011. This was compared with the list of mechanically ventilated patients provided by the ICC. Comparison for criteria used to diagnose pneumonia, ventilator day of the diagnosis, was recorded. The ICC identified two VAPs from 136 potential patients compared with the Trauma Service identifying 36 VAPs. A difference in diagnostic criteria between the ICC and the Trauma Service focused on use of the National Nosocomial Infection Survey (NNIS) algorithm versus quantitative microbiology from bronchoalveolar lavage specimens. Thirty-five of 36 Trauma Service VAPs were not identified as VAPs by the NNIS algorithm as a result of the chest radiographs. Application of differing definitions of VAP results in markedly different VAP rates. The difference has significant implications as infection rates are increasingly reported as a quality metric.