The ability to monitor the appropriateness of hospital antibiotic use is a key element of an effective antibiotic stewardship program. A set of 11 generic quality indicators (QIs) was previously developed to assess the quality of antibiotic use in hospitalized adults treated for a bacterial infection. The primary aim of the current study was to assess the clinimetric properties of these QIs (nine process and two structure indicators) in daily clinical practice.
In a cross-sectional point-prevalence survey, performed in 2011 and 2012, 1890 inpatients from 22 hospitals in the Netherlands treated with antibiotics for a suspected bacterial infection were included, and data were extracted from medical records. In this cohort we tested the measurability, applicability, reliability, room for improvement and case mix stability of the previously developed QIs.
Low applicability (≤10% of reviewed patients) was found for the QIs ‘therapeutic drug monitoring’, ‘adapting antibiotics to renal function’ and ‘discontinue empirical therapy in case of lack of clinical and/or microbiological evidence of infection’. For the latter, we also found a low inter-observer agreement (kappa <0.4). One QI showed low improvement potential. The remaining seven QIs had sound clinimetric properties. Case-mix correction was necessary for most process QIs. For all QIs, we found ample room for improvement and large variation between hospitals.
Establishing the clinimetric properties was essential, as four of the 11 previously selected QIs showed unsatisfactory properties in this practice test. Since the quality of antibiotic use and the process of documenting data is changing over time and may vary per country, QIs should always be tested in practice first.
