Challenges in predicting future high-cost patients for care management interventions.

Area of Science:

• Health Services Research
• Data Analytics
• Predictive Modeling

Background:

• Claims data segmentation is used to predict future hospitalizations.
• Understanding actuarial limitations in risk segmentation is crucial for healthcare management.

Purpose of the Study:

• To evaluate the accuracy of claims data segmentation in predicting Medicare hospitalizations.
• To illustrate the limitations of person-centered risk segmentation for identifying high-utilization beneficiaries.

Main Methods:

• Analysis of a 100% sample of Medicare beneficiaries from 2017-2018.
• Examining hospitalization rates for progressively narrower segments of heart failure patients.
• Graphical representation of 2018 hospitalization rates based on 2017 data.

Main Results:

• Narrower patient segments showed increased hospitalization rates.
• However, these segments accounted for a smaller percentage of total Medicare FFS hospitalizations.
• Over half of all patients across segments and the total population had no hospitalizations in 2018.

Conclusions:

• Claims data alone has limitations in predicting high-cost patients.
• Integrating clinician input and 'light touch' monitoring can improve prediction accuracy.
• Systemic strategies for utilization management and provider steering may be more effective than individual patient targeting.