3187 - Combining Risk Matrix and Network Theory to Improve Risk Analysis in Radiotherapy

Purpose/Objective(s): The aim of this study was to enhance the traditional risk analysis approach proposed by the International Atomic Energy Agency (IAEA) in its report "Aplicación del Método de la Matriz de Riesgo a la Radioterapia" (Viena, 2012), and more recently expanded in the SEFM, SEPR, and CSN Project MARRTA (2022). We integrated the risk matrix method with network theory to better understand and mitigate risks in radiotherapy services. We hypothesized that representing the system as a complex network of initiating events and barriers would provide a more comprehensive understanding of risk dynamics and improve management strategies.

Materials/Methods: A bipartite network was constructed, connecting initiating events to their corresponding safety barriers. Each connection was weighted based on the robustness of the barrier. An algorithm was developed to apply the risk matrix method to this network, calculating risk levels for each initiating event. Additionally, percolation theory was used to simulate the system's response to random failures and targeted attacks on critical barriers. The model analyzed changes in the distribution of initiating events across different risk levels and identified phases where specific risks dominated. Three real-world radiotherapy services were studied to compare the effectiveness of implementing less robust but accessible barriers versus highly robust and expensive ones.

Results: The network-based approach allowed for the identification of phase transitions based on the number of barrier failures, enabling the prediction of when the system enters phases dominated by high or very high risks. Simulations using percolation theory demonstrated that these transitions are influenced by the robustness of the barriers, occurring more rapidly when the most connected barriers are targeted. In real-world cases, incorporating less robust barriers significantly reduced overall risk more effectively than relying solely on highly robust barriers. These findings highlight the utility of the network approach in identifying risk mitigation strategies that are both practical and effective.

Conclusion: Integrating the risk matrix method with network theory provides a powerful tool for analyzing and managing risks in radiotherapy. This approach offers insights into the interactions between initiating events and barriers, enabling the identification of emergent risk patterns and the prediction of critical risk phases. The findings demonstrate the potential of this methodology to improve safety protocols in radiotherapy by informing the design of cost-effective and impactful risk mitigation strategies. This approach can be adapted to other complex systems to enhance their safety and reliability.