2011 OPEN FORUM Abstracts
MULTIDISCIPLINARY STRATEGY TO REDUCE VAP.
Brian Glynn, Raymond Malloy, Linda Earle; Thomas Jefferson University Hospital, Philadelphia, PA
INTRODUCTION: Ventilator- associated pneumonia (VAP) is the leading cause of preventable deaths due to infection and is associated with an increased LOS of at least 6 days, prolonged ventilator dependence and an estimated 40K dollar increase per hospital stay. BACKGROUND: In 2007, the Department of Respiratory Care joined a multidisciplinary working group at Thomas Jefferson University Hospital charged with reducing the incidence of VAP from 10.0 to 0 incidences per 1000 ventilator days in 7 adult intensive care units. METHODS: The Ventilator Associated Pneumonia Bundle promoted by the Institute of Healthcare Improvement was the foundation of our initiative. The VAP working group clarified definitions, standardized procedures, and employed educational strategies throughout all disciplines. The working group incorporated best practices for ventilator care and maintenance, with elements including but not limited to q4hour oral care, q 12 hour chlorhexidine oral rinse, early tracheotomy, head of the bed elevation and endotracheal tubes with subglottic suction. An on- line record of compliance with VAP prevention bundle was reviewed monthly. A discussion of the diagnosis of VAP was held with the attending intensivist to confirm each suspected diagnosis. RESULTS: The VAP rate decreased from 10.0 to 7.57 per 1000 ventilator days in the first 6 months. A further drop was seen when ownership of specific metrics was identified, compliance with the prevention bundle was improved and electronic ventilator order entry was begun. We now have a sustained VAP of 3.0 per 1000 ventilator days over 2 years ( with an increase of 50 ICU beds) translating to 29 lives saved and a minimum cost savings of 5.4 million healthcare dollars. CONCLUSION: This strategy will be expanded by examining the unique characteristics of each ICU to reach our goal of 0 VAPS. Sponsored Research - None