Nurse on computer


Predicting Patients at Risk of Becoming Hospital Super-Utilizers


Mat Reidhead

Mat Reidhead

Vice President of Research and Analytics



HIDI HealthStats


  • Care Coordination
  • Research


care coordination HIDI HealthStats research

Programs designed to intervene with hospital super-utilizers to reduce costs and improve outcomes commonly rely on predictive models to identify patients at highest risk of excess future hospital utilization, target scarce interventional resources and provide supplemental clinical and social decision support. For sustainability, super-utilizer intervention programs must exhibit more in direct cost savings than direct project investment. Successful care coordination models designed to reduce unnecessary utilization among super-utilizers have found that programmatic costs were more than offset by reduced uncompensated inpatient days and emergency department visits. Robust risk prediction modeling increases the likelihood of positive programmatic ROI, particularly in accountable care settings when the intervening provider also bears part or all of the financial risk for avoidable health care utilization.

While various criteria are used to identify hospital super-utilizers, it is estimated that they collectively account for four to eight percent of all patients, and 21 to 28 percent of all ED utilization. Super-utilizers typically have a variety of clinical, social and behavioral complexities that add extreme difficulty in the design of interventions designed to reduce their ED utilization. Successful interventions have focused on pain management, mental illness and substance abuse among patients by investing in integrated care delivery models that provide medical, social and mental health supports.

During fiscal year 2016 in Missouri, 20,655 patients with ten or more ED and inpatient visits accounted for 1.6 percent of patients, 12 percent of all inpatient and ED visits, and nearly $3 billion in hospital charges. With the exception of only 36 of these patients, all were identified as having at least one clinical, social or behavioral risk factor and 69 percent had all three risk factors. Nearly nine out of ten were diagnosed with a chronic disease during the year and 71 percent were diagnosed with multiple chronic conditions. Ninety percent were identified as having behavioral risk factors — more than half had more than one. And, 86 percent had a socioeconomic risk factor — nearly half had multiple social risk determinants. Super-utilizers with all three of the risk categories experienced 36 percent more hospital visits during fiscal year 2016 compared to super-utilizers with just one.

Read the Report

HIDI HealthStats are applied research briefs developed from HIDI data assets targeting relevant topics related to health policy and population health.

Back to Top