Objective: To determine whether monitoring and feedback of blood culture contamination rates to phlebotomists would reduce the overall contamination rate.
Design: Before and after interventional study.
Setting: Blood cultures collected by venipuncture by phlebotomists at Foothills Hospital, Calgary, a tertiary care teaching hospital.
Intervention: Feedback of contamination rates calculated from a laboratory definition of blood culture contamination. The definition was based on isolation of typical skin organisms from a single blood sample when two samples were obtained.
Main outcome measure: Reduction in the laboratory-defined contamination rate in the second year.
Results: Of 8462 cultures collected by phlebotomists in the prefeedback year, 224 (2.6%) were contaminated, compared to 131 (1.4%) of 9282 cultures in the postfeedback year. There was a rise in the total number of positive cultures regarded as significant but a fall in the number of coagulase-negative staphylococci that were regarded as significant by our definition. The rate of contamination in blood cultures collected by nonphlebotomists did not change.
Conclusions: The contamination rate decreased after feedback. Our definition of contamination was imperfect and could be improved, but it was valuable in achieving a real reduction in blood culture contamination.