Medicaid coverage accuracy in electronic health records

Area of Science:

• Health Services Research
• Health Informatics
• Public Health

Background:

• Health insurance is crucial for accessing healthcare and improving health outcomes.
• Accurate documentation of health insurance, particularly Medicaid, in electronic health records (EHRs) is essential for monitoring coverage.
• Understanding factors influencing EHR data accuracy for insurance is vital for researchers and healthcare providers.

Purpose of the Study:

• To evaluate the validity of EHR data for tracking longitudinal Medicaid coverage.
• To assess how patient demographics, visit types, and clinic characteristics affect the accuracy of documented Medicaid coverage in EHRs.

Main Methods:

• Retrospective, observational study comparing EHR data with Medicaid enrollment data (gold standard).
• Included adult patients with Medicaid ID and at least one clinic visit between 2013-2014 from Oregon community health centers.
• Analyzed visit-level and patient-level agreement using statistical correspondence measures (kappa, sensitivity, specificity).

Main Results:

• Excellent agreement was observed between EHR and Medicaid data for health insurance information (kappa >0.80, sensitivity >0.80, specificity >0.85).
• Lower agreement was noted for visits with non-English language preference or missing income information.
• Patient-level agreement was lower for Black patients and higher for older patients in primary care settings.

Conclusions:

• Community health center EHR data serve as a valid source for Medicaid coverage information.
• Variations in agreement highlight the need for careful consideration of patient and visit characteristics when utilizing EHR insurance data.
• Findings inform researchers and clinic staff on the reliability and limitations of EHR data for insurance monitoring.