Ideal and Actual Impact of Rapid Diagnostic Testing and Antibiotic Stewardship on Antibiotic Prescribing and Clinical Outcomes in Children With Positive Blood Cultures.

Abstract

Implementing matrix-assisted laser desorption ionization-time of flight and multiplex polymerase chain reaction has been associated with decreased mortality and hospital length of stay in adults, but the impact in pediatrics is less understood. This pre-post quasi-experimental study compared antibiotic prescribing for positive blood cultures in patients ≤21 years of age collected in 2012 (preintervention) and in 2015 (after matrix-assisted laser desorption ionization-time of flight/multiplex polymerase chain reaction). Time to effective and optimal antimicrobial therapy was evaluated using Cox proportional hazards regression. Time to ideal optimal therapy was estimated as the earliest potential initiation of optimal therapy. Antibiotic use and clinical outcomes were measured. There were 242 and 192 positive monomicrobial blood cultures in 2012 and 2015, respectively. Postintervention, time to optimal therapy (73.8 vs. 48.8 hours; P < 0.001) and organism identification (55.6 vs. 29.5 hours; P < 0.001) were reduced, and patients were more likely to receive optimal therapy by 7 days (hazard ratio, 1.85; P < 0.001). In the ideal scenario in 2015, there was an 8.8-hour delay in initiating optimal therapy based on the time that sufficient microbiologic data were available. Postintervention, time to effective therapy (2.8 vs. 2.7 hours; P = 0.782) and clinical outcomes did not differ. Unnecessary antibiotic duration for probable contaminants (skin flora) (43.1 vs. 29.7 hours; P = 0.027), vancomycin for methicillin-sensitive Staphylococcus aureus (54.0 vs. 41.3 hours; P = 0.008) and nonpenicillin/ampicillin antibiotics for group A Streptococcus, group B Streptococcus and Enterococcus faecalis (87.2 vs. 33.4 hours; P < 0.001) were reduced postintervention. Rapid diagnostics reduced time to optimal antimicrobial therapy and unnecessary antibiotic use without worse clinical outcomes.