In modern hospitals, vital signs are continuously monitored with a variety of medical devices. Many of these devices are configured with threshold alarms, which are considerably limited:
- Monitors only raise alarms when the threshold is crossed
- Monitors are oblivious to each other
- Monitors typically don't use patient information to customize alarms
- Monitors do not provide detailed rational for alarms
As a result, most patient monitors produce many false alarms, which have been shown to have an adverse affect on patient care. For example, large numbers of false alarms has been shown to increase the likelihood of a nurse ignoring a true alarm.
Our insight is that many of these false alarms could be suppressed if devices had information from each other. For example, a respiration rate monitor alarm could be suppressed if a patient's oxygen levels and blood pressure were both normal. Combining information in this way would create smart alarms which could improve alarm accuracy, improving patient care.