Making Blood Culture Workups More Efficient

A major problem with blood culture workups is that it is often difficult to determine a true positive (pathogen) result from a false positive (contaminant) result. One solution is to adjust the testing algorithm in place. In 1970, it was shown that 11% of contaminated cultures had additional positive cultures, while 69% of clinically significant cultures had multiple positive cultures. In the proposed algorithm (summarized right), any
isolates that are typically contaminants that are not present in follow-up cultures are reported as contaminants. To confirm the accuracy of the new algorithm, a nurse epidemiologist and an infectious diseases physician reviewed the patient charts.

Overall, organisms considered to be contaminants (CNS, diphthroids, Micrococcus spp, Bacillus spp, and viridans streptococci) were identified in 495/1040 positive blood cultures. Of the 495, 286 were classified as contaminants due to negative follow-up cultures, 171 were investigated by a pathology resident, and 15 were classified as pathogens (viridans strep with positive follow-up cultures). The majority (62%) of the contaminants were identified as CNS.

Two types of errors, VM (very major) and M (major) were used to determine the success of the new algorithm. A "very major error" occurred if an isolate that was classified as a contaminant was determined to be clinically significant after chart review. A "major" error occurred if an isolate was classified as indeterminate or pathogenic and was later determined to be not clinically significant. VM errors occurred in 6.3% of cases and M errors occurred in 6.6% of cases. Further adjustment of this algorithm may decrease the error rate in the future, but the algorithm was determined "acceptable" by the authors of this study.

Source: http://jcm.asm.org/content/40/7/2437.full