Can Claims Data Algorithms Identify the Physician of Record?

Area of Science:

• Health Services Research
• Health Informatics
• Geriatric Medicine

Background:

• Administrative claims data are commonly used to identify primary care physicians (PCPs).
• The accuracy of these claims-based algorithms in the US Medicare population remains unassessed.

Purpose of the Study:

• To evaluate the agreement between PCPs identified by claims-based algorithms and the PCP of record in electronic health record (EHR) data.
• To assess the validity of claims-based algorithms for identifying PCPs in older adults with diabetes.

Main Methods:

• Utilized EHR and Medicare claims data from 3,658 Medicare fee-for-service beneficiaries aged 65+ with diabetes.
• Employed algorithms based on visit plurality and majority, with tie-breakers including last visit, cost, or time span.
• Analyzed 15,624 patient-years of data.

Main Results:

• Algorithms focusing on primary care visits showed higher agreement (78.0%-85.9%) compared to those using all visits (25.4%-63.3%).
• Lower agreement was observed in nonwhite individuals, Medicaid enrollees, patients with PCP changes, and those with over 10 physician visits.

Conclusions:

• Claims-based algorithms focusing on primary care visits may improve PCP identification accuracy.
• These algorithms exhibit reduced performance in vulnerable populations and individuals with fragmented care.
• Further research is needed to refine algorithms for diverse and complex patient groups.