Can we trust administrative data in joint arthroplasty? A validation study against the Italian Arthroplasty Registry data as a gold standard.

INTRODUCTION: Hospital Discharge Data (HDD) can be a valuable source of information for epidemiological research, but in Italy its accuracy in arthroplasty has not yet been determined on a large scale. The aim is to assess Italian HDD accuracy in reporting clinical information on hip/knee arthroplasties, using RIAP data collected by the Italian Arthroplasty Registry (RIAP) as a reference. METHODS: Coding systems for procedures and diagnoses in RIAP data and HDD for years 2007-2021 are mapped to a common list of items describing surgical procedures and related diagnoses. The ability of HDD in predicting procedures and diagnoses is evaluated by sensitivity, specificity and accuracy, while using RIAP data as a reference. RESULTS: Surgical procedures and causes for elective and urgent arthroplasties are predicted by HDD with at least 96% sensitivity. Performances drop when evaluating procedures and diagnoses at fine-grain level and for rare events. DISCUSSION: HDD reports reliable clinical information in arthroplasty and is an effective tool for epidemiological purposes. Nonetheless, a cautious approach must be considered when dealing with high-detail and rare events.