eVENT: An expert system for detecting ventilatory events during anesthesia

Despite the best intentions of clinicians many patients suffer adverse events during their medical care. As technology becomes more sophisticated and the amount of information generated increases, the risk that something goes wrong or is missed increases. The technology, work environment and clinical work flow needs to be designed to mitigate these risks. Technology has the potential to help clinician’s with tasks that humans do not perform well, such as vigilance, and reduce risk to patients.

Dr. J Mark Ansermino is a pediatric anesthesiologist with a background in health informatics and a particular interest in patient safety. He and his team are developing an expert system that automatically detects important changes in a patient’s status in the operating room. The clinician must divide his or her time between keeping an eye on the monitors, attending to the patient, and other tasks such as teaching students and giving drugs. Important events or evolving conditions can be missed. The technology will assist the busy clinician with attending to the huge amount of information generated by the monitors and bringing the important changes to the attention of the clinician. The system can also offer advice on what should be done about these changes and information on current treatment guidelines, providing the clinician with a recommended course of action in real time.

The current emphasis is on developing rules for a few dangerous events that affect breathing. This will be expanded to address other anesthesia-related events in the operating room. In this research, the system will be evaluated in both simulated and real clinical settings to determine if it is possible for clinicians to identify and respond to critical events more rapidly and reliably. This expert system will lend itself to promoting better decision making by both less skilled or experienced anesthesiologists during training or in situations where advanced training is not available.