Background: National hospital antimicrobial consumption data, based on reimbursement records, provide valuable insights into antimicrobial use but lack clinical context. Linking these data with patient demographics, clinical indications, and hospitalisation details allows for assessing the appropriateness of antimicrobial use. This study aims to identify diagnoses associated with the highest prescription of antimicrobials in Belgian hospitals, facilitate inter-hospital benchmarking, and evaluate prescribing practices in relation to clinical guidelines.

Methods: The AntiMicrobial consumption data linked with DIAgnoses (AMDIA) pilot project used retrospective data from 2016 and 2017 in Belgian hospitals. Hospital antimicrobial reimbursement data from the National Institute for Health and Disability insurance were linked to Minimal Hospital Data (MHD) from the Federal Public Service for Public Health at the patient-level using a unique hospital stay identifier. The MHD dataset included patient demographics (age, gender, pre-admission location), hospitalization details (diagnoses, procedures, ward type, length of stay), and antimicrobial data (unique identifier for drug classification, ATC-code, volume, administration route). Antimicrobial consumption was classified and quantified following WHO ATC/DDD methodology. Diagnoses were classified using All Patients Refined Diagnosis Related Group codes. Diagnoses with the largest number of hospitalizations involving antibiotics were analysed, focusing on adult patients in acute care hospitals. The proportion of hospital stays involving antibiotics and broad-spectrum antibiotics was calculated for each diagnosis. The ESAC-Net definition for hospital broad-spectrum antibacterials was used.

Results: In 2017, Belgian acute care hospitals recorded 1,585,010 hospitalizations, with antibiotics prescribed in half of these cases (50.2%; 795,100). Hip replacements accounted for the highest number of hospitalizations involving antibiotics (3.7%; 29,502), with antibiotics prescribed in 98% of these cases and a median DDD of 4 per stay. Vaginal deliveries ranked second (3.2%; 25,441), with antibiotics prescribed in 30% of the cases and a median DDD of 0 per stay. Broad-spectrum antibiotics were used in 8.4% of hip replacements and 0.3% of vaginal deliveries.

Conclusion: This study identifies areas for antimicrobial stewardship by highligthting priority diagnoses in Belgian hospitals. Due to delayed access to data, real-time evaluation is limited, efforts to obtain up-to-date data are ongoing to enhance monitoring and support stewardships.