Аннотация
BACKGROUND: Adverse events associated with flexible ureteroscopes are increasingly reported to postmarket surveillance systems; however, the lack of standardized harm classification limits reproducibility and actionable insights. OBJECTIVES: To develop and reliability-test a reproducible taxonomy for classifying patient harm and death in Food and Drug Administration Manufacturer and User Facility Device Experience (MAUDE) reports (2020-2025) and to characterize mechanisms and severity profiles. METHODS: = 39), 148 unique reports (including 1 death) were analyzed. Two independent clinicians coded reports using a prespecified taxonomy for (1) primary harm category, (2) harm mechanism, and (3) severity grade. Inter-rater reliability was quantified using unweighted and weighted Cohen's kappa (κ). RESULTS: Inter-rater reliability was substantial for the primary harm category (κ = 0.82; 95% confidence interval [CI] 0.76-0.88; 86% agreement) and severity grade (weighted κ = 0.79; 95% CI 0.71-0.87) but moderate for mechanism (κ = 0.61). Traumatic (31.1%) and infectious (30.4%) categories predominated. Operational harm (clinical consequences without tissue injury) accounted for 20.3%. Most events (69.4%) were moderate or severe. While breakage/detachment was the most common mechanism (27.0%), 48.0% of reports lacked sufficient detail for classification. Reports involving reprocessing/contamination were disproportionately severe (63.6%). The single death involved septic shock with findings suggestive of device bioburden. CONCLUSIONS: This reliability-tested taxonomy enables reproducible harm phenotyping in passive surveillance. However, nearly half of MAUDE narratives lack sufficient detail for mechanism coding. As a passive system lacking exposure denominators, MAUDE supports analysis of reporting composition but not incidence estimation or valid between-device comparisons.