Skip to page content


Saving Lives with AI: Using the Deterioration Index Predictive Model to Help Patients Sooner

A Novant Health facility used a predictive model to identify hospitalized patients who were at risk of deterioration more quickly, reducing mortality by 22% and saving an estimated 153 lives over 11 months.

Proactively identifying patients who are showing signs of deterioration is critical to preventing morbidity and mortality for acute care patients. Predictive models, which use statistical analysis to find patterns, can take in a vast amount of data and quickly turn it into actionable information about individual patient risk for clinicians. Studies show that by considering more data points than traditional methods, on top of learning from patient and population data, models can offer more personalized and accurate predictions.

Novant Health New Hanover Regional Medical Center deployed and operationalized Epic’s Deterioration Index, a predictive algorithm that uses data like demographic information and lab results in its score calculation to proactively identify patients at risk of deterioration.

Read the full article on EpicShare.