Stewardship in Community Hospitals—Optimizing Outcomes and Resources (SCORE): A Cluster-Randomized Controlled Trial Investigating the Impact of Antibiotic Stewardship in 15 Small, Community Hospitals

Abstract

Background. Small, community hospitals (SCH, <200 beds) account for almost three fourths of US hospitals; most do not have antibiotic stewardship programs (ASPs). Forthcoming federal regulations will require all SCHs to have ASPs. Our objective was to compare the impact of 3 types of ASPs in a network of SCHs. Methods. Intermountain Healthcare has 15 SCHs within its network, all of which lacked ASPs prior to the study. The 15 SCHs were randomized to 1 of 3 ASPs (Table) with increasing infectious diseases (ID) support from a tertiary referral center. Antibiotic use data was obtained via the NHSN AU option. Adjusted antibiotic use rate ratios (RR) between the intervention period (15 months) and the baseline (31 months) trend were compared between Programs 2 and 3 versus Program 1 (reference group) using negative binomial segmented regression. We used fixed effects models to draw conclusions for the 15 studied SCHs and mixed models with random hospital effects for gen-eralizable inferences concerning the 3 ASPs. Results. Adjusted RRs comparing intervention antibiotic use to baseline are shown in Figure 1 (all antibiotics) and Figure 2 (broad-spectrum antibiotics). Under fixed-effects models, Program 2 SCHs did not significantly reduce total antibiotic use (RR 0.96, [0.83, 1.10]) while Program 3 SCHs reduced total use by 17% (RR 0.83 [0.72, 0.95]) compared to Program 1 SCHs. Program 2 and 3 SCHs reduced broad-spectrum use by 31% (RR 0.69 [0.53, 0.91]) and 27% (RR 0.73 [0.56, 0.95]), respectively, compared to Program 1 SCHs. Similar trends but with wider confidence intervals including the null hypothesis were obtained under mixed models.