Variable impact of an antimicrobial stewardship programme in three intensive care units: time-series analysis of 2012–2017 surveillance data

Abstract

Preprescription authorization (PPA) and postprescription review with feedback (PPRF) were successively implemented in 2012 and 2016 in our 1500-bed hospital. The impact of PPA and PPRF on carbapenems use and resistance levels of Pseudomonas aeruginosa was assessed in three intensive care units (ICUs). Carbapenems use (in DDDs/1000 occupied bed-days) and resistance of P.aeruginosa (percentage of non-susceptible (I+R) isolates to imipenem and/or meropenem) were analysed using a controlled interrupted time-series method. Two periods were compared: 2012-2015 (PPA) and 2016-2017 (PPA+PPRF). Models were adjusted on the annual incidence of extended-spectrum β-lactamase-producing enterobacteriacae. Carbapenem use was stable over the PPA period in all ICUs, with a significant change of slope over the PPA+PPRF period only in ICU1 (β2= -12.8, 95% confidence interval (CI)= -19.5 to -6.1). There was a switch from imipenem to meropenem during the PPA period in all three units. Resistances of P.aeruginosa were stable over the study period in ICU1 and ICU2, and significantly decreased over the PPA+PPRF period in ICU3 (β2= -0.18, CI= -0.3 to -0.03). In real-life conditions and with the same antimicrobial stewardship programme (AMSP) led by a single team, the impact of PPRF was heterogeneous between ICUs. Factors driving the impact of AMSPs should be further assessed in comparable settings through real-life data, to target where they could prove cost-effective.