Root-cause and failure-mode analyses with an in-house incident reporting system: Comparison of its structure with ASTRO RO-ILS

Autor: Catherine Riehl, Jason Meier, Gina M. Goode, Henry Chou, Beatrice Bloom, A. Kapur, Louis Potters
Rok vydání: 2014
Předmět:
Zdroj: Journal of Clinical Oncology. 32:163-163
ISSN: 1527-7755
0732-183X
Popis: 163 Background: In industries with well-established processes for failure analyses, incident reporting systems facilitate analyzing root causes of adverse or near-miss events, and the dissemination of lessons learned. Such a system is now being promoted in radiation medicine as patient safety and care quality come into focus in clinical practice. Methods: An incident reporting system has been in place in our department since 2010 for staff to report actual or near-miss events. A quality management committee was concurrently established to review reported incidents on a weekly basis. A taxonomy of root causes was set up to analyze and categorize root causes of incidents. This taxonomy is compared to that of the 2012 consensus recommendations for incident learning database (Ford et al.) which was the basis of ASTRO RO-ILS currently in prototype phase. Results: To date, over 2,000 incidents have been analyzed in our in-house system. The majority of the reported incidents pertain to procedural mis-steps or miss-cues in the process leading up to patient treatment, resulting in delays. Identifying the root causes has helped re-engineer or fine-tune department processes and has resulted in smoother clinical operations (e.g., obtaining third-party imaging prior to making treatment appointments) as well as more safety checks (e.g., pace-maker alerts.) Our taxonomy of root causes is multi-dimensional and quite specific to clinical radiation oncology, e.g., drilling down to specific morbidities. In comparison, the taxonomy of causes in the RO-ILS is more generic and at a higher level (e.g., procedural or technical issues). The RO-ILS has at most second-degree branches compared with our taxonomy of 6-8 degrees in depth. Conclusions: Incident reporting is vital in analyzing and learning from reported events. Process changes in our clinic has been aided by the incident database and a hierarchical taxonomy between 6 to 8 degrees in root causes. The RO-ILS system in prototype is only 2 degrees deep. While the RO-ILS system appears capable of cataloging data from multiple centers, it is not clear that it will be able to understand or affect meaningful process changes given its lack of depth.
Databáze: OpenAIRE