Predicting Patients at Risk of Becoming Hospital Super-Utilizers
Key Findings
Throughout 2018, HIDI will deploy a super-utilizer predictive risk model using near-real time ADT feeds to assist hospitals and other providers in care coordination efforts.
More than 20,000 Missourians visited a hospital between 10 and 384 times during fiscal year 2016.
The majority of these hospital super-utilizers were uninsured or covered by Medicaid — their utilization accounted for nearly $3 billion in associated hospital charges during the year.
Based on these risk factors, a predictive model was developed using HIDI data to prospectively identify patients at high risk of becoming hospital super-utilizers.
The model exhibited 96% discriminant ability and strong external validity on a randomly selected, independent sample of 655,000 hospital patients in fiscal year 2016.
