Radiation incident learning in multi-site centers.

Area of Science:

• Healthcare Management
• Patient Safety
• Clinical Informatics

Background:

• Multi-site hospital systems face challenges in incident learning due to resource limits and varied site cultures.
• System-wide policy adherence is often inconsistent across different hospital locations.
• Effective cross-site learning and proactive risk management are crucial for enhancing patient safety.

Purpose of the Study:

• To introduce a novel, resource-efficient hybrid approach for incident learning in large, multi-site hospital systems.
• To facilitate cross-site learning and proactive risk management to improve patient safety.
• To address challenges posed by resource constraints and diverse site cultures in incident reporting.

Main Methods:

• A hybrid approach combining statistical and root cause analyses was applied to safety events from 2023-2024 across five hospital sites.
• Failure modes (FMs) were identified using keywords, with frequency and cross-site correlations visualized through word clouds.
• A proactive model was employed at a satellite clinic to predict and mitigate potential patient harm.

Main Results:

• Analysis of 228 events identified "carepath" and "contour" as the most frequent failure modes in treatment planning.
• "Contour" and "documentation" keywords appeared across four sites, indicating common issues.
• At a satellite clinic, the proactive model reduced the average risk priority number from 106 to 77 and the estimated affected patients from 27 to 20.

Conclusions:

• A comprehensive, system-wide hybrid approach was developed, enhancing incident learning efficiency across multi-site clinics.
• The approach facilitated benchmarking of site-specific safety cultures and enabled measurable cross-site learning.
• Novel word cloud analysis and a proactive model improved visualization of root causes and system-wide preparedness for patient safety.