Fuzzy Based Multimodality Clinical Alarm in ICU/CCU

In high workload areas such as the Intensive or Critical Care Units (ICU/CCU), clinicians are burdened with too many alarms and false alarms, leading to poor user response or no response to alarm signals, which in turn leads to serious patient safety concerns, adverse events such as injury and deaths. Even with the implementation of the IEC international alarm standard, alarm hazards are still seen as the top health technology hazard in healthcare institutions. There are numerous new developments in alarm technology, including in the areas of alarm detection and smart alarm design, aimed at improving the sensitivity and performance of alarm systems. However, there is still a lack of studies on the application of Human Factors Engineering (HFE) principles and AI in designing alarms for medical devices that could improve user response and ensure patient safety. This research aims to develop a fuzzy logic base multimodality clinical alarm monitor software to improve alarm response among the clinicians in ICU/CCU and the performance of the clinical alarm. The research involves testing, verifying, and validating the fuzzy-based multimodality clinical alarm to improve the performance of the alarm system. The proposed fuzzy alarm is compared to the medical professional interpretation of a patient physiological condition extracted from the MIMIC II database. The results show that the proposed fuzzy alarm can match the interpretation of medical professionals with high accuracy. In terms of sensitivity and specificity, the proposed alarms achieve good performance with blood pressure and heart rate specificity and sensitivity at 100%. Meanwhile, sensitivity and specificity for respiratory rate are at 97.59% and 99.68%, while sensitivity and specificity for oxygen saturation are at 100% and 98.04%, respectively. The research concluded that incorporating alarm information systems based on risk, human factor engineering principles, and fuzzy logic into the alarm system significantly improves user response while reducing alarm hazards and optimising the performance of the alarm system.