Evaluating the Effectiveness of an Antimicrobial Stewardship Program on Reducing the Incidence Rate of Healthcare-Associated Clostridium difficile Infection: A Non-Randomized, Stepped Wedge, Single-Site, Observational Study

BACKGROUND: The incidence rate of healthcare-associated Clostridium difficile infection (HA-CDI) is estimated at 1 in 100 patients. Antibiotic exposure is the most consistently reported risk factor for HA-CDI. Strategies to reduce the risk of HA-CDI have focused on reducing antibiotic utilization. Prospective audit and feedback is a commonly used antimicrobial stewardship intervention (ASi). The impact of this ASi on risk of HA-CDI is equivocal. This study examines the effectiveness of a prospective audit and feedback ASi on reducing the risk of HA-CDI. METHODS: Single-site, 339 bed community-hospital in Barrie, Ontario, Canada. Primary outcome is HA-CDI incidence rate. Daily prospective and audit ASi is the exposure variable. ASi implemented across 6 wards in a non-randomized, stepped wedge design. Criteria for ASi; any intravenous antibiotic use for ≥ 48 hrs, any oral fluoroquinolone or oral second generation cephalosporin use for ≥ 48 hrs, or any antimicrobial use for ≥ 5 days. HA-CDI cases and model covariates were aggregated by ward, year and month starting September 2008 and ending February 2016. Multi-level mixed effect negative binomial regression analysis was used to model the primary outcome, with intercept and slope coefficients for ward-level random effects estimated. Other covariates tested for inclusion in the final model were derived from previously published risk factors. Deviance residuals were used to assess the model’s goodness-of-fit. FINDINGS: The dataset included 486 observation periods, of which 350 were control periods and 136 were intervention periods. After accounting for all other model covariates, the estimated overall ASi incidence rate ratio (IRR) was 0.48 (95% 0.30, 0.79). The ASi effect was independent of antimicrobial utilization. The ASi did not seem to reduce the risk of Clostridium difficile infection on the surgery wards (IRR 0.87, 95% CI 0.45, 1.69) compared to the medicine wards (IRR 0.42, 95% CI 0.28, 0.63). The ward-level burden of Clostridium difficile as measured by the ward’s previous month’s total CDI cases (CDI Lag) and the ward’s current month’s community-associated CDI cases (CA-CDI) was significantly associated with an increased risk of HA-CDI, with the estimated CDI Lag IRR of 1.21 (95% 1.15, 1.28) and the estimated CA-CDI IRR of 1.10 (95% CI 1.01, 1.20). The ward-level random intercept and slope coefficients were not significant. The final model demonstrated good fit. CONCLUSIONS: In this study, a daily prospective audit and feedback ASi resulted in a significant reduction in the risk of HA-CDI on the medicine wards, however, this effect was independent of an overall reduction in antibiotic utilization. In addition, the ward-level burden of Clostridium difficile was shown to significantly increase the risk of HA-CDI, reinforcing the importance of the environment as a source of HA-CDI.