Value and challenges of using large clinical datasets from physiologic monitors to improve alarm systems safety: Suggestions for improvement

2^nd International Conference on Health Informatics and Technology

July 27-29, 2015 Valencia, Spain

Azizeh Sowan

Posters-Accepted Abstracts: J Health Med Informat

Abstract :

Clinical Alarms are the 2015 top technology hazards. Thousands of clinical alarm-related deaths are reported. Cardiac monitors are associated with the highest number of deaths and alarms. To date, the most commonly used techniques to quantify alarm data are observation, surveillance cameras and middleware. Large datasets of the audit log of modern physiologic monitoring devices have not yet been used for predictive modeling, capturing unsafe practices or guiding initiatives on alarm systems safety. We utilized the cardiac monitorsâ?? audit log as an objective data source to quantify alarms rate in a 20-week interventional project that took place in an intensive care unit. Interventions include changing alarm logarithms and education on alarm management. The audit log of the cardiac monitors is a very complex chronological record of all alarms data (e.g., priority, limits, time generated vs. ended) and 21 clinician actions (e.g., limit change). Results showed a total of 139,452 alarms (190 alarm/patient day) in the pre-intervention period compared to 136,104 alarms (163.4 alarm/patient day) in the post-intervention period (P>0.05). This presentation provides datadriven discussions of the advantages of the audit log, challenges of using it in alarm safety studies, critical considerations in presenting alarm data and suggestions for improving the logged data to be a useful source for clinicians, researchers, vendors and policy makers. Despite current challenges, large digitalized clinical datasets provide an objective, detailed data source of recorded alarmsâ?? events and types and user actions and hold a great promise in performance improvement.