Effect of a national antibiotic stewardship intervention in China targeting carbapenem overuse: An interrupted time-series analysis

Abstract

Objectives: To assess trends and patterns of carbapenem use and to evaluate the effects of a nationwide antibiotic stewardship policy to reduce carbapenem overuse.

Methods: In this quasi-experimental study, using longitudinal data from the national drug procurement database and interrupted time-series analyses with carbapenems as the intervention group and possible carbapenem substitutes as the comparison group, we evaluated the effects of a national stewardship policy on carbapenem consumption and expenditures, by region and types of healthcare institutions.

Results: The carbapenem procurement volume declined by -28.8% (95% CI -35.0 to -22.6) (-334.4 thousand defined daily doses [DDDs] per month), and carbapenem expenditures showed a relative reduction of -38.1% (-43.9 to -32.2). The gap between the use of carbapenems and each drug in the comparison group narrowed after the policy intervention, with an increase in tigecycline use (14.9 thousand DDDs [10.8-18.9]) and a slower decrease in use of certain third-generation cephalosporin combinations (-85.7 [-143.0 to -28.4]), penicillin combinations (-200.9 [-421.4-19.6]), and fourth-generation cephalosporins (-116.9 [-219.8 to -14.0]). Consumption was highest during the pre-policy period, and declines were largest following the intervention in the eastern region (-32.1%, -39.8 to -24.4) and among tertiary hospitals (-266.2 [-339.5 to -192.9] thousand DDDs per month).

Conclusion: This population-level drug utilization research represents the first comprehensive evaluation of the effectiveness of China's nationwide carbapenem stewardship. The national policy targeting carbapenem prescribing has led to a sustained reduction in carbapenem use with limited substitution. Effects varied geographically and were concentrated in tertiary and secondary hospitals.

Keywords: Antimicrobial stewardship; Carbapenem use; Healthcare institutions; Interrupted time-series analysis; Policy intervention.