Effect of a “handshake” stewardship program versus a formulary restriction policy on High-End antibiotic use, expenditure, antibiotic resistance, and patient outcome

Abstract

This study reports the effect of implementing an antibiotic stewardship program (ASP) based on the “handshake” strategy for 2 years on multiple endpoints compared with that in a preceding period when an antimicrobial restriction policy was only applied in the absence of a complete program in a tertiary-care Lebanese hospital. The studied endpoints were broad-spectrum antibiotic consumption, antibiotic expenditure, nosocomial bacteremia incidence rate, and patient outcome.An interrupted time series analysis was undertaken to assess the changes in the trend (ΔT) and level (ΔL) of the aforementioned endpoints among adult inpatients before (October 2013 to September 2015) and after the introduction of the ASP (October 2016 to September 2018).After the implementation of the “handshake” ASP, marked changes were observed in the consumption of broad-spectrum antibiotics. The mean use density levels for imipenem and meropenem decreased by 13.72% (P = 0.017), coupled with a decreasing rate of prescription (ΔT = −24.83 defined daily dose [DDD]/1,000 patient days [PD]/month; P = 0.02). Tigecycline use significantly decreased in level by 69.19% (P < 0.0001) and in trend (ΔT = −25.63 DDD/1,000 PD/month; P < 0.0001). A reduction in the use of colistin was also documented but did not reach statistical significance (ΔL = −8.71%, P = 0.56; ΔT = −5.51 DDD/1,000 PD/month = −5.5; P = 0.67). Antibiotic costs decreased by 24.6% after ASP implementation (P < 0.0001), and there was a distinct change from an increasing rate to a decreasing rate of expenditure (ΔT = −12.19 US dollars/PD/month; P = 0.002). The incidence rate of nosocomial bacteremia caused by carbapenem-resistant gram-negative bacteria (CRGNB) decreased by 34.84% (P = 0.13) coupled with a decreasing trend (ΔT = −0.23 cases/1,000 PD/month, P = 0.08). Specifically, a noticeable reduction in the incidence rate of bacteremia due to carbapenem-resistant Acinetobacter baumannii was documented (ΔL = −54.34%, P = 0.01; ΔT = −0.24 cases/1000 PD/month, P = 0.01). Regarding patient outcome, all-cause mortality rates did not increase in level or in rate (ΔL = −3.55%, P = 0.59; ΔT = −0.29 deaths/1000 PD/month, P = 0.6). The length of stay and 7-day readmission rate remained stable between the two periods.In conclusion, the “handshake” ASP succeeded in controlling the prescription rates of antibiotics and in decreasing the nosocomial bacteremia rates caused by CRGNB without compromising patient outcome in our facility. It also had an economic effect in reducing antibiotic costs compared with the previous restriction policy on antimicrobial dispensing.