Background. The National Healthcare Safety Network (NHSN) has recently supported efforts to shift surveillance away from ventilator associated pneumonia (VAP) to ventilator associated events (VAE) to decrease subjectivity in surveillance and increase clinical correlation. This study compares the results of an automated surveillance strategy using the new VAE definition to a prospectively performed clinical application of the definition.
Methods. All patients ventilated for 2 or more days in the medical and surgical intensive care units (ICUs) were evaluated by two methods: 1) retrospective surveillance using an automated algorithm combined with manual chart review following the NHSN VAE methodology and 2) prospective surveillance with some modifications of the NHSN methodology by pulmonary physicians in collaboration with the clinical team administering care to the patient at the bedside.
Results. Of the 1209 patients evaluated, 69 were found to have VAE by the retrospective surveillance while the prospective surveillance identified 67 events. 56 patients were determined to have VAE by both methods (kappa = 0.81, p = 0.04). There were 24 patients considered to be VAE by only one of the methods. Most discrepancies were the result of clinical disagreement with the NHSN VAE methodology.
Conclusion. There was good agreement between the study teams. Overall, a similar number of events were called by each method (69 vs 67). Awareness of the limitations of the surveillance definition for VAE can help infection prevention personnel in discussions with critical care partners about optimal use of these data.
Disclosures. S. Micek, Tetraphase: Grant Investigator, Research support M. Kollef, Cubist: Speaker's Bureau, Speaker honorarium; Basilea: Speaker's Bureau, Speaker honorarium; Merck: Speaker's Bureau, Speaker honorarium