Kaya E, et al. Imaging Sci Dent. 2022 Jul;52:e34. https://doi.org/10.5624/isd.20220050

Background/Objectives: The delayed repair of failed medical devices, particularly when multiple devices are simultaneously unavailable, may adversely affect patient safety and healthcare continuity, making repair prioritization a critical hospital decision problem. This study develops an integrated multi-criteria decision-making (MCDM) framework for prioritizing failed medical devices for repair and evaluates the consistency and robustness of the resulting priorities. Its principal contribution is to address repair prioritization as a distinct operational decision problem by integrating deterministic and fuzzy MCDM approaches and robustness analyses within a unified framework. Methods: Seven evaluation criteria—patient safety risk (PSR), device function (DF), device criticality (DC), utilization frequency (UF), availability of alternative devices (AAD), resource impact (RSI), and device age (DA)—were considered. The framework was implemented in a hospital case study involving 32 medical devices using the Analytic Hierarchy Process (AHP), AHP-weighted scoring, the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS), Fuzzy Analytic Hierarchy Process (Fuzzy AHP), and Fuzzy Technique for Order Preference by Similarity to Ideal Solution (Fuzzy TOPSIS). Robustness was examined through ±10%, ±30%, and ±50% criterion-weight sensitivity analyses, criterion-removal analysis, comparative ranking analysis, and hospital expert-based AHP and class-based AHP scenarios. Results: In the baseline AHP-weighted scoring ranking, Anesthesia Machine (4.9233) and External Cardiac Pacemaker (4.8862) received the highest repair priorities. Comparisons with TOPSIS, Fuzzy TOPSIS, hospital expert-based AHP, and class-based AHP yielded Spearman’s rank correlation coefficient ($\rho$) values ranging from 0.8328 to 0.9952. Sensitivity analysis showed high overall ranking stability under substantial criterion-weight variations, while criterion-removal analysis also maintained high correlations ($\rho$ = 0.9901–0.9963), supporting the robustness of the overall ranking structure. Conclusions: The findings demonstrate that the framework can support hospital biomedical units in prioritizing failed devices, enhance the transparency and consistency of repair decisions, and contribute to more effective maintenance resource allocation.