Evaluation of the Clinical Impact of the Biofire Filmarray® Rapid Multiplex PCR Assay in Blood Culture Identification Combined with Antimicrobial Stewardship Intervention

Abstract

Poster Abstracts • OFID 2017:4 (Suppl 1) • S627 Conclusion. Verigene BC-GN, in combination with antibiotic stewardship, suc-cessfully improved time to effective antibiotic therapy among MDR GN organisms causing bacteremia. Background. Molecular testing has been shown to improve turnaround time (TAT) for identifying bloodborne pathogens. Early results can inform directed escal-ation or de-escalation of antimicrobial therapy. Paired with antibiotic stewardship, rapid pathogen identification has been shown to reduce antibiotic utilization and improve patient outcomes. However, many of these studies were in single site insti-tutions. We evaluated implementation of the BioFireÒ FilmArray Blood Culture Identification System (BCID) across 3 acute care facilities utilizing a central testing laboratory at Carolinas Healthcare. Methods. BCID testing was implemented over a 2-month period. A multidis-ciplinary team developed standard protocols for processing, transport and testing, with communication of results across teams of stewardship pharmacists. Standard algorithms were used across all facilities to guide antibiotic prescribing. Data were collected between January and May 2017 from three sources (BacTec, Theradoc and Cerner EMR). Positive bottles were tracked from the time of the positive bottle alert through pharmacist intervention. We evaluate rates of interventions and consistency using variance comparison tests. Results. 708 positive blood cultures were identified at 3 acute care facilities and tested using BCID. TAT from positive bottle to BCID result was 4.6 (95% CI 4.4–4.8) hours. 86.7% (614/708) were on appropriate empiric antimicrobials at the time of the BCID result. 28.0% (198/708) required a recommendation by a pharma-cist. 39.6% (78/197) had an escalation recommendation while 26.4% (52/197) had a de-escalation recommendation. There was no significant variation across shifts or sites except with de-escalation where variation was greater than 10% across sites (P = 0.02). Conclusion. BCID testing was successfully implemented across a large integrated healthcare system using central testing laboratory paired with a team of stewardship and virtual care pharmacists. Our strategy provided timely and reproducible results across facilities and shifts. Implementation of BCID allowed for more pathogen directed therapy at all facilities with variability in need for escalation and de-escalation of therapy between facilities. Background. Bloodstream infections are a major cause of morbidity and mor-tality worldwide, with favorable clinical outcomes associated with early optimal anti-biotic selection. Rapid diagnostics have become a key part in achieving this. Biofire Filmarray® was introduced at our institution for rapid blood culture (BC) identifica-tion, coupled with antimicrobial stewardship (AS) interventions. We aimed to assess the impact of this test on time to adequate antimicrobial therapy in a setting with pre-existing effective AS interventions. Methods. An observational retrospective chart review, pre and post study was performed. We reviewed adult positive BC before and after implementation of Biofire. Outcomes were: (1) time from BC result reported to health care provider to start of adequate antimicrobial therapy,(2) time to stopping antimicrobial therapy in BC thought to be contaminants, (3) time to any change in antimicrobial therapy and (4) a composite outcome of outcomes 1 and 2. A univariate Cox proportional hazards model was performed. Results. 326 positive BC were analyzed, 173 before and 153 after Biofire imple-mentation. At the time of healthcare provider notification, 77 were not on adequate antimicrobials, with median time to adequate therapy of 6.98 hours. (IQR 3.93–23.96) before and 6.1 hours. (IQR 1.84–20.95) after implementation, P = 0.48. There were 75 BC classified as contaminants and median time to stopping antimicrobials was 48.28 (IQR 18.56–89.36) vs. 45.25 hours. (IQR 15.12–100.60), P = 0.61. Time to any change in any antimicrobial therapy was similar with a median of 13.05 (IQR 4.00–36.77) vs. 10.90 hours. (IQR 2.97–31.10), P = 0.87. Analysis of the composite outcome revealed a median of 23.95 (6.29–58.50) vs. 14.82 (IQR 4.07–44.79) hours. (Hazard ratio 1.33, 95% confidence interval 0.96–1.84, P = 0.09). Conclusion. Implementation of the Biofire Filmarray® did not have a statistically significant effect on our composite outcome of time to adequate therapy and time to discontinuation in the case of contaminants. Our findings suggests that when added to other effective AS surveillance and interventions, the magnitude of the clinical impact of rapid PCR diagnostics for BC identification is minimal.