Trends in Blood Culture Contamination: A College of American Pathologists Q-Tracks Study of 356 Institutions

Abstract Context.—Blood culture contamination extends hospital stays and increases the cost of care. Objectives.—To measure blood culture contamination rates in a large number of institutions over time and to elucidate practice patterns and demographic factors associated with sustained reduction in contamination rates. Design.—Longitudinal cohort study of 356 clinical laboratories that provided quarterly data about blood culture results, using a uniform definition of contamination. Mixed linear model analysis of the 1999 through 2003 data set. Results.—Blood culture contamination was significantly higher in institutions that used nonlaboratory personnel to collect blood (P = .03) and significantly lower in facilities that used a dedicated phlebotomy team (P < .001). Higher volume of blood collection was significantly associated with lower contamination rates (P < .001). Continued participation in the Q-Tracks monitoring program was associated with significant and progressive reduction in contamination rates. By the fifth year of participation, the median institution had reduced its blood culture contamination rate by 0.67% (P < .001). Conclusions.—Institutions that use decentralized patient-centered personnel rather than dedicated phlebotomy teams to collect blood cultures experience significantly higher contamination rates. Long-term monitoring of contamination is associated with sustained improvement in performance.