A Comparison of National Healthcare Safety Network (NHSN) Ventilator Associated Pneumonia (VAP) and Ventilator Associated Event (VAE) Definitions
Methods: We evaluated patients mechanically ventilated for > 2 days in six intensive care units at an academic medical center between December 2011 and November 2012 using the two surveillance definitions. Surveillance using the VAP definition was completed manually through the routine surveillance processes of trained infection prevention specialists (IPS). Surveillance using the new VAE definition was done retrospectively through a combination of automated and manual surveillance. IPS were blinded to the results from the existing VAP surveillance. Rates of infection were compared using chi-square analysis (Epi Info) and agreement between the two definitions was measured by kappa statistic (IBM SPSS).
Results: 1832 patients were ventilated for 14,904 ventilator (vent) days (median days/ patient = 5; range = 2-80). Twelve patients met VAP by the old definition, resulting in a rate of 0.81 cases/1000 vent days. 100 patients met VAE criteria (6.71/1000 vent days). These VAE included 57 cases of vent associated condition (3.82/1000 vent days), 20 infection-related vent associated complication (1.34/1000 vent days), and 23 possible or probable VAP (1.54/1000 vent days). The old VAP rate was significantly lower than the VAE rate (p<0.01), although not statistically different than the possible/probable VAP rate (p=0.06). Agreement between the old VAP definition and probable/possible VAP from the VAE definition was poor (kappa = 0.05).
Conclusion: The new VAE definitions resulted in higher event rates. There was poor agreement between the old and new surveillance definitions, likely related to the difference in surveillance starting points for the methods. Infection prevention departments instituting the new definition should be aware of the differences, and able to discuss them with hospital stakeholders.
P. Kremer, None
C. Sykora, None
M. Craighead, None
L. Ruhl, None
D. K. Warren, None
M. H. Kollef, None
H. M. Babcock, None
See more of: Oral Abstract Session