Assessment of Rapid-Blood-Culture-Identification Result Interpretation and Antibiotic Prescribing Practices

• Published in: Journal of clinical microbiology 2017-05, Vol.55 (5), p.1496-1507
• Main Authors: Donner, Linsey M, Campbell, W Scott, Lyden, Elizabeth, Van Schooneveld, Trevor C
• Format: Article
• Language: English
• Subjects: Anti-Bacterial Agents - therapeutic use; Bacteremia - diagnosis; Bacteremia - microbiology; Bacteriology; Blood Culture; Clinical Decision-Making - methods; Decision Support Systems, Clinical; Enterobacteriaceae; Enterobacteriaceae - classification; Enterobacteriaceae - drug effects; Humans; Inappropriate Prescribing - prevention & control; Inappropriate Prescribing - statistics & numerical data; Molecular Diagnostic Techniques - methods; Nebraska; Practice Patterns, Physicians; Staphylococcus; Staphylococcus - classification; Staphylococcus - drug effects; Surveys and Questionnaires
• ISSN: 0095-1137; 1098-660X
• DOI: 10.1128/JCM.02395-16

Rapid pathogen identification can alter antibiotic prescribing practices if interpreted correctly. Microbiology reporting can be difficult to understand, and new technology has made it more challenging. Nebraska Medicine recently implemented the BioFire FilmArray blood culture identification panel (BCID) coupled with stewardship-based education on interpretation. Physician BCID result interpretation and prescribing were assessed via an electronic survey, with a response rate of 40.8% (156/382 surveys). Seven questions required respondents to interpret BCID results, identify the most likely pathogen, and then choose therapy based on the results. The tallied correct responses resulted in a knowledge score. General linear models evaluated the effect of role, specialty, and utilization of the BCID interpretation guide on the mean knowledge score. The specialties of the respondents included 55.7% internal medicine, 19.7% family medicine, and 24.6% other. Roles included 41.1% residents, 5.0% fellows, and 53.9% faculty. Most reported that they reviewed antimicrobial susceptibility results (89.4%) and adjusted therapy accordingly (81.6%), while only 60% stated that they adjusted therapy based on BCID results. The correct response rates ranged from 52 to 86% for the interpretation questions. The most common errors included misinterpretation of and genus results. Neither role nor specialty was associated with total knowledge score in multivariate analysis ( = 0.13 and 0.47, respectively). In conclusion, physician interpretation of BCID results is suboptimal and can result in ineffective treatment or missed opportunity to narrow therapy. With the implementation of new technology, improved reporting practices of BCID results with clinical decision support tools providing interpretation guidance available at the point of care is recommended.