Impact of An Electronic “Antibiotic Time Out” on Provider Prescribing Patterns

Abstract

BackgroundCurrent antibiotic stewardship guidelines suggest the use of an antibiotic time-out (ATO) 48–72 hours after antibiotic initiation to encourage review of empiric regimens once additional diagnostic information is available; however, the recommendation is based on low-quality evidence. Our objective was to retrospectively evaluate the impact of an electronic ATO alert within the EPIC™ electronic health record on provider prescribing patterns after implementation.MethodsThe ATO alerts were instituted in January 2017 and were triggered when an inpatient received Vancomycin (VAN), Piperacillin/Tazobactam (PT), Ceftriaxone (CEF), or a combination of them for >72 hours. We used an EPIC™ benchside report to identify ATO alerts between January and April 2017 and systematically reviewed charts from the last week of each month to identify de-escalation opportunities (DEO). Pediatric, bone marrow transplant, orthopedic, and cystic fibrosis patients were excluded. The primary outcome was de-escalation within 12 hours of the alert, defined as narrowing of spectrum or discontinuation of antibiotics.ResultsWe identified 805 alerts among 209 patients; 87 patients were excluded from analysis. Among 122 included patients, a median of 3 alerts were triggered per patient (470 in total). DEO was identified in 34.7% of alerts; de-escalation events (DEE) occurred in 34.3% of DEO. Table 1 lists alerts, DEO, and DEE by antibiotic. PT was the most frequently de-escalated antibiotic (46.4% [26/56] of DEE). De-escalation occurred more frequently among patients either actively followed (P < 0.01) or receiving new consultations by Infectious Disease (P = 0.04).Table 1.De-escalation opportunities and events by antibioticAntibioticAlerts(n = 470)De-escalation opportunities(n = 163)De-escalation events(n = 56)VAN130 (27.6%)23 (14.1%)8 (14.3%)PT137 (29.1%)66 (40.5%)26 (46.4%)CEF123 (26.2%)9 (5.5%)6 (10.7%)VAN/PT65 (13.8%)62 (38.0%)15 (26.8%)VAN/CEF15 (3.2%)3 (1.8%)1 (1.8%)ConclusionAn electronic ATO alert triggered only on the basis of drug and duration lacked specificity in identifying opportunities for antibiotic de-escalation. De-escalation occurred significantly more frequently with Infectious Disease team involvement. Additional study is required to identify how to best support de-escalation efforts.Disclosures J. S. Lewis II, Merck & Co.: Consultant, Consulting fee. J. C. McGregor, Merck & Co.: Grant Investigator, Research grant.