Identifying undercompensated groups defined by multiple attributes in risk adjustment.

Area of Science:

• Health economics
• Health services research
• Data science in healthcare

Background:

• Existing methods for evaluating health plan payment risk adjustment are often arbitrary and inconsistent.
• Identifying undercompensated patient groups, especially those with multiple attributes, is crucial for equitable healthcare.
• Current risk adjustment formulas may inadvertently create incentives for insurers to discriminate against certain patient populations.

Purpose of the Study:

• To develop a systematic approach for identifying undercompensated groups in health plan payment risk adjustment.
• To improve upon existing arbitrary and inconsistent evaluation methods.
• To pinpoint groups defined by multiple attributes that are systematically undercompensated.

Main Methods:

• Developed a 'group importance' measure within the random forests algorithm, extending the concept of variable importance.
• Utilized 2016-2018 IBM MarketScan and 2015-2018 Medicare claims and enrollment data for analysis.
• Evaluated two risk adjustment scenarios: the individual health insurance Marketplaces formula and the Medicare formula.

Main Results:

• Identified previously unrecognized undercompensated groups with multiple chronic conditions in the Marketplaces risk adjustment.
• Observed overcompensation for groups without chronic conditions in the Marketplaces.
• Found that the magnitude of undercompensation for multi-attribute groups significantly exceeds that for single-attribute groups.
• No consistent under- or overcompensation patterns were detected for complex groups in the Medicare risk adjustment formula.

Conclusions:

• The novel method effectively identifies complex undercompensated groups in health plan payment risk adjustment.
• Undercompensation in risk adjustment can incentivize insurer discrimination against vulnerable patient groups.
• This research offers policymakers insights into potential discrimination targets and a pathway toward more equitable health coverage.