A computationally efficient formal method for discovering simultaneous masking in medical alarm

Numerous patient injuries and deaths have been caused by medical practitioners failing to respond to medical alarms. Simultaneous masking, where concurrently sounding medical alarms result in one or more being unhearable, is partially responsible for this problem. In previous work, we introduced a computational formal method capable of proving (formally verifying) if masking could occur in a modeled configuration of medical alarms. However, the scalability of the method limited the applicability and completeness of its analyses. In the work presented here, we show how we re-implemented the method to address these shortcomings. We evaluated the detection capabilities and scalability of the new version of the method with a series of realistic and synthetic case studies. Our results show that the new version of the method replicates and improves detection capabilities compared to the legacy method and does so with significant reductions in verification times. We discuss the patient safety implications of our results and explore directions for future research.