The Belgian policy of funding antimicrobial stewardship in hospitals and trends of selected quality indicators for antimicrobial use, 1999–2010: a longitudinal study

Abstract

Objectives: In order to improve antimicrobial (AM) use, a policy of providing technical and financial support to AM management teams (AMTs) was rolled out in all Belgian hospitals between 2002 and 2008. We aimed to analyse the association of this policy with AM use for the two indications accounting for the largest number of patients receiving AM: prophylaxis for major lower limb orthopaedic surgery and pneumonia. Design, setting, participants: We used patient-level data routinely collected in all Belgian acute care hospitals between 1999 and 2010. We modelled trends for selected quality indicators (QIs) using the year of AMT implementation in each hospital as the main 'change point', with fine-tuned case-mix adjustment. Of all admissions for lower limb orthopaedic surgery, and pneumonia between 1999 and 2010, 90% (325 094) and 95% (327 635), respectively, were found eligible for analyses. Outcomes: The surgery QI was defined as: cefazolin, dose in the expected range, and no use of other AM. For pneumonia, QIs were: ratio of oral/parenteral defined daily doses (DDD, O/P QI), and mean number of DDD minus penicillin, per 100 days of hospitalisation (DDD QI). Results: Between 1999 and 2010, the surgery QI improved from 59% to 71%, the O/P QI from 0.72 to 0.97, and the DDD QI from 96 to 64. Heterogeneity between hospitals was high. Overall, no association was found with the year of implementation of the AMT. Conclusions: Improvements have been observed but could not be related at the national level to the policy under study. However, these results cannot be extrapolated to other QIs for AM use in hospitals. Our findings do not question the need for AMT, nor the need for continuation of AMT funding. Several recommendations can be made in order to make the best of Belgium's unique political and financial commitments in that field. ; This study was funded by the Scientific Institute of Public Health (internal grant, 99 429 Euros, project number W3035.0116.1). Data have been generated as part of the routine work of the Ministry of Health and National Sick Fund. RB acknowledges support from a Methusalem research grant from the Flemish government not specific for this project.