Artificial Intelligence (AI)

Improving Patient Outcomes Using Artificial Intelligence

 A Proactive vs Reactive Approach: How JHAH is Using AI Cognitive Computing Models to Identify Deteriorating Patients Sooner

With the implementation of JHAH’s state-of-the-art Electronic Health Record (EHR) back in 2018, the ability to leverage Artificial Intelligence (AI) through the introduction of Cognitive Computing Models that monitor patient populations and calculate a patient’s risk for a particular outcome has been applied. These AI capabilities provide JHAH with advanced technical abilities to proactively identify hospitalized patients at risk from sepsis, deterioration, emergency services, and other factors, and introduce interventions sooner to improve patient safety and outcomes.

Epic’s Deterioration Index is just one important model introduced at JHAH which has improved the identification of patients at risk of deteriorating while admitted as inpatients. Prior to the introduction of the Deterioration Index, hospitals including JHAH relied on manual assessments such as Modified Early Warning Scores (MEWS) which required clinicians to document information, often available elsewhere in the patient chart. This required duplicate documentation to determine a patient’s acuity and risk of deterioration. Overall, Epic’s Deterioration Index has been shown to flag 14 percent more patients at risk of deterioration with 12 percent fewer false positives than MEWS. Given that assessments such as MEWS were completed at prescribed time intervals, by the time a patient was deemed to be at risk, the interventions tended to be reactive since the patient was already deteriorating.

With the introduction of Epic’s Deterioration Index AI Cognitive Computing Model, our clinicians are able to take a proactive approach to identify deteriorating patients. This cognitive computing model uses multiple additional data points and an algorithm, that is much more accurate than the simple arithmetic used by other early warning systems, and turns it into actionable information. Essentially, the Deterioration Index updates dynamically based on the latest variables available in a patient’s chart providing clinicians with the most up-to-date risk assessment allowing for earlier interventions and better outcomes.

“The Deterioration Index is exciting because it moves the electronic record from a simple replacement of paper towards fulfilling its potential of enabling real improvements in the quality and safety of patient care,” shared Dr. Michael Walsh, Chief Executive at JHAH.

JHAH will continue to explore and introduce additional Cognitive Computing Models and AI capabilities in other areas to improve patient outcomes.