Effects of clinical pathway implementation on antibiotic prescriptions for pediatric community-acquired pneumonia.

Italian pediatric antimicrobial prescription rates are among the highest in Europe. As a first step in an Antimicrobial Stewardship Program, we implemented a Clinical Pathway (CP) for Community Acquired Pneumonia with the aim of decreasing overall prescription of antibiotics, especially broad-spectrum.The CP was implemented on 10/01/2015. We collected antibiotic prescribing and outcomes data from children aged 3 months-15 years diagnosed with CAP from 10/15/2014 to 04/15/2015 (pre-intervention period) and from 10/15/2015 to 04/15/2016 (post-intervention period). We assessed antibiotic prescription differences pre- and post-CP, including rates, breadth of spectrum, and duration of therapy. We also compared length of hospital stay for inpatients and treatment failure for inpatients and outpatients. Chi-square and Fisher's exact test were used to compare categorical variables and Wilcoxon rank sum test was used to compare quantitative outcomes.120 pre- and 86 post-intervention clinic visits were identified with a diagnosis of CAP. In outpatients, we observed a decrease in broad-spectrum regimens (50% pre-CP vs. 26.8% post-CP, p = 0.02), in particular macrolides, and an increase in narrow-spectrum (amoxicillin) post-CP. Post-CP children received fewer antibiotic courses (median DOT from 10 pre-CP to 8 post-CP, p<0.0001) for fewer days (median LOT from 10 pre-CP to 8 post-CP, p<0.0001) than their pre-CP counterparts. Physicians prescribed narrow-spectrum monotherapy more frequently than broad-spectrum combination therapy (DOT/LOT ratio 1.157 pre-CP vs. 1.065 post-CP). No difference in treatment failure was reported before and after implementation (2.3% pre-CP vs. 11.8% post-CP, p = 0.29). Among inpatients we also noted a decrease in broad-spectrum regimens (100% pre-CP vs. 66.7% post-CP, p = 0.02) and the introduction of narrow-spectrum regimens (0% pre-CP vs. 33.3% post-CP, p = 0.02) post-CP. Hospitalized patients received fewer antibiotic courses post-CP (median DOT from 18.5 pre-CP to 10 post-CP, p = 0.004), while there was no statistical difference in length of therapy (median LOT from 11 pre-CP to 10 post-CP, p = 0.06). Days of broad spectrum therapy were notably lower post-CP (median bsDOT from 17 pre-CP to 4.5 post-CP, p