1124 - Toward Safer Radiation Therapy: Leveraging the Insights from an In-House and a National Radiation Oncology Incident Learning System

Purpose/Objective(s): To evaluate longitudinal patient-safety trends within departmental and national incident learning systems (ILS), identify actionable risk-mitigation strategies, and benchmark therapeutic radiation incident rates per treatment fraction against published data to assess prevention effectiveness.

Materials/Methods:

Our practice used an internal ILS for a decade before transitioning to the national Radiation Oncology Incident Learning System (RO-ILS) in 2017. Historical events were reclassified according to the RO-ILS taxonomy to enable longitudinal analysis. To contextualize national trends, we subtracted our institutional event counts from aggregate RO-ILS data from 2017 onward, thereby generating a comparator dataset representing all other participating practices. National safety patterns observed over the decade were qualitatively compared with our previously implemented risk-mitigation strategies, including prospective peer review prior to treatment planning and formal cancellation (“no-fly”) policies for rushed treatment starts. Because treatment-fraction–normalized therapeutic radiation incident rates are not available within RO-ILS, we calculated institutional rates and compared them with corresponding data from seven contemporary North American publications.

Results:

Although our facilities comprised 1.8% of the 781 nationally enrolled practices, they contributed 11.3% of aggregate RO-ILS events through 2024. The proportion of therapeutic radiation incident reports was similar between our practice and others and declined over time. Near-misses and unsafe conditions were more prevalent among other facilities (52.3%) than in ours (10.5%, p<0.01), whereas operational/process improvement events were significantly more common in our practice (78.8% vs 36.2%, p<0.01) and increased over time. Across seven North American publications, the mean therapeutic incident rate per treatment fraction was 0.00091 ± 0.00051 compared with 0.00028 ± 0.00014 over 14 years in our practice. National RO-ILS themed analyses identified peer review and mitigation of rushing as key risk-reduction strategies, aligning with our established safety initiatives. Survey feedback from our multidisciplinary QA committee indicated that both RO-ILS and our internal ILS supported a non-punitive safety culture with comparable usability and impact; RO-ILS was viewed as more streamlined and nationally connected, whereas the internal system was more granular but more resource-intensive.

Conclusion: Sustained engagement with incident learning systems was associated with persistently low therapeutic radiation incident rates despite organizational growth. As incidents reaching patients declined, increasing reporting of operational issues provided opportunities for earlier intervention. National RO-ILS themed analyses on peer review and rushing reinforce the importance of these locally implemented strategies in maintaining system reliability.