Source avec lien : Safety Science, 122, 2020/02/01. 10.1016/j.ssci.2019.104519
La radiothérapie est une technique importante pour traiter le cancer. En raison des risques professionnels élevés, le processus est soumis à des réglementations et des normes de sécurité strictes. Toutefois, ces normes n’imposent pas l’utilisation d’une méthode particulière d’analyse des dangers. Systems Theoretic Process Analysis (STPA) est une nouvelle méthode d’analyse des risques, essentiellement différente, basée sur la théorie des systèmes. Le présent document contribue à combler cette lacune grâce à une évaluation préliminaire de la STPA appliquée à un processus de radiothérapie à intensité modulée mature. La principale leçon que nous avons apprise est qu’une modélisation graphique et systémique du processus analysé, bien que difficile pour les débutants, est un instrument puissant pour saisir les dangers et même pour en identifier d’autres. Une analyse causale d’un sous-ensemble de ces dangers nouvellement découverts a conduit à des mesures d’atténuation des risques significatives et précieuses. Ces résultats suggèrent que la STPA peut être une option viable pour l’analyse de sécurité en radiothérapie.
Radiation therapy is an important technique to treat cancer. Due to the high occupational risks involved, the process is subject to severe safety regulations and standards. However, these standards do not mandate the usage of a particular hazard analysis method. The de facto methods currently used are the reliability theory-based Fault Tree Analysis (FTA) and Healthcare Failure Mode and Effects Analysis (HFMEA). Systems Theoretic Process Analysis (STPA) is a new, essentially different hazard analysis method, based on systems theory. Although successfully applied in many industries, there are only a few reports on STPA implementation in radiation therapy. This paper contributes to filling this gap with a preliminary assessment of STPA applied to a mature Intensity Modulated Radiation Therapy (IMRT) process. The analysis was conducted by a team consisting of two experts in radiation therapy and one software systems engineer, with little domain knowledge. 142 potentially unsafe control actions were identified and compared with the results of an earlier HFMEA. The main lesson we have learned is that a graphical, system-wise modeling of the analyzed process, although challenging for beginners, is a powerful instrument to catch the same and even other, new hazards. A causal analysis of a subset of these newly found hazards has led to meaningful and valuable risk mitigation measures. These results suggest considering STPA as a viable option for safety analysis in radiation therapy. We expect that this top-down, well-structured way of analysis can especially be advantageous for safety assessment in early design phases, when an HFMEA is not possible yet, because most of system’s implementation and behavior is still unknown.