Fuzzy Analytical Hierarchy Process-based Risk Priority Number Approach in Failure Modes and Effects Analysis for...

Failure modes and effects analysis (FMEA) is commonly used to identify, prioritize, and mitigate potential failure modes (FMs) by assigning a risk priority number (RPN). However, traditional RPN-based FMEA has limitations, particularly when handling the degree of interdependency within processes. To address this, we propose a fuzzy analytical hierarchy process (AHP)-based RPN method, designed to prioritize FMs by accurately weighing risk factors in magnetic resonance imaging (MRI)-guided high-dose-rate brachytherapy (BT) for gynecologic (GYN) cancer.

A process map covering all steps was developed for MRI-based GYN BT, and potential FMs were identified. Evaluators were given 2 questionnaires, one for conventional FMEA and another for AHP evaluation. For the AHP method, substeps were grouped by job specialty, with identical weights applied to FMs within each specialty group. Fuzzy linguistic terms helped evaluators handle uncertainties, and final fuzzy AHP-based RPN values were calculated by applying weighted risk factor scores.

The process map included 6 primary steps, 33 substeps, and 82 FMs. In the fuzzy AHP RPN analysis, the top 5 FMs were identified as incorrect/suboptimal applicator insertion, incorrect applicator reconstruction, dose-volume histogram not meeting the physician's intent, incorrect/missing contours, and applicator/patient movement. By comparison, the conventional FMEA ranked the top 5 as incorrect/missing contours, incorrect/suboptimal applicator insertion, dose-volume histogram not meeting the physician's intent, applicator/patient movement, and incorrect applicator reconstruction. FMs with rank differences of 10 or more between methods were mostly related to applicator insertion and MRI.

This study demonstrates the feasibility and effectiveness of a fuzzy AHP-based RPN method for comprehensive FM prioritization, tailored to the clinical workflow of MRI-based GYN BT. Our findings provide a valuable reference for implementing fuzzy AHP-based risk assessment in MRI-guided BT.