53 - Effect of Rapid Identification of Bloodstream Isolates on Antibiotic Management Using a Pharmacist-based Treatment Algorithm

Background: The use of rapid molecular diagnostic testing to identify microorganisms and resistance markers has great potential to optimize medical care and assist with antimicrobial stewardship.  We implemented the Verigene bloodstream infection testing panel along with a pharmacy notification system to clinicians and assessed efficacy of the system.

Methods: In November 2019, we implemented the Verigene gram positive and negative panels for patients with positive blood cultures. Our antimicrobial stewardship committee developed a recommended treatment algorithm for pharmacists to use when notified of Verigene results.  The first positive bottle per patient and per admission was tested.  Subsequent positive bottles were not tested on the Verigene unless a different morphology was noted on the gram stain.  A gram stain was performed on all positive cultures and this result was called to the patient's nurse (if inpatient) and the covering physician was notified of the result.  After the Verigene result was available, an assigned pharmacist was notified of these results (organism identification and resistance markers if identified).  Pharmacists notified covering physicians of the test results and the recommended antibiotic management. Pharmacists documented the frequency that the test result changed the antibiotic management, including escalation, de-escalation or no change in therapy.  The data from the first six months was summarized.

Results: From 11/19/19-5/18/20, a total of 575 test results were called into the pharmacist (average 3.2/day).  Among these, 165 (28.7%) were considered likely contaminants, 106 had no change in therapy and 59 had antibiotic de-escalation.  Among the remaining 410 patients, 156 had de-escalation, 53 had escalation, 30 were not on any antibiotics and appropriate antibiotics were started.  Overall, antibiotic management changed in 298/575 (51.8%) of isolates run by Verigene in our institution including 215 (37.4%) de-escalations.  The most frequent antibiotics that were stopped included vancomycin (142) and cefepime (53).

Conclusion: Our pharmacist-based algorithm for notification and treatment recommendation based on Verigene results was highly successful in optimizing antibiotic management and improving antimicrobial stewardship in our institution.