Linking Medicaid Claims Data and Safety Net Clinic Electronic Health Record (EHR) Data to Obtain a More Complete Picture of Diabetes Prevention Services among Vulnerable Populations, 2005-2007

Background: EHR data can bridge gaps in insurance claims databases, especially for uninsured and underinsured populations. OCHIN, a community health center-controlled network, built and maintains a fully-integrated EHR system linking 30 safety net health center organizations across five states, with a single medical record for each patient across all sites.

Objectives: To study receipt of preventive services among OCHIN's adult diabetic population in Oregon, comparing services documented only in Medicaid claims or only in OCHIN data, among patients with Medicaid identification numbers.

Methods: Our study population was 4,240 persons with ≥2 diabetes-associated clinic visits in 2004-2005, ≥1 visit in 2006 and in 2007, to confirm continuity. We made individual-level linkages between OCHIN EHR data and Medicaid data for a subsample of 2,103 adults with Medicaid identification numbers, then identified services documented in only one dataset versus in both.

Results: In 2005-2007, 79% of the study population received ≥1 LDL screening. Among the subsample with Medicaid identification, 278 had an LDL screen in the Medicaid data but not OCHIN's data, 615 in OCHIN's data but not in Medicaid. Sixty-nine percent received >1 flu shot; 49 had one documented in Medicaid data but not OCHIN's data, and 667 in OCHIN but not Medicaid. Forty-seven percent received >1 nephropathy screening; 245 documented in Medicaid data but not OCHIN, and 521 in OCHIN data but not Medicaid. Ninety-three percent of the population received >1 hemoglobin A1c screening; 132 documented in Medicaid but not OCHIN data, and 606 in OCHIN but not Medicaid.

Conclusions: Networked safety net clinics' EHR data provide an unprecedented opportunity to obtain utilization data from uninsured and underserved populations. By better understanding the limitations inherent in claims and EHR data, we can use these data sources more effectively to conduct research to inform policies aimed to achieve social justice imperatives.