Comparison of 2 systems for clinical practice profiling in diabetic care: Medical records versus claims and administrative data

Objectives: To (1) describe the rate of glycosylated hemoglobin (HbA(1c)) testing and control in a primary care clinic at an academic medical center; (2) compare academic medical center and health maintenance organization (HMO) data for a subgroup of the same patients; and (3) discuss the challenges to accurate clinical practice profiling. Study Design: Retrospective analysis of medical records and claims and administrative data. Patients and Methods: We reviewed the medical records of 300 patients with diabetes mellitus from the Primary Care practice at an academic medical center (AMC). The HMO and AMC were in the southeastern United States. The study adhered to the National Committee for Quality Assurance's Health Plan Employer Data and Information Set 3.0 (1998) guidelines for collection of the numerator and denominator data. Results: Overall, 275 (91.7%) of 300 patients underwent a documented HbA(1c) test (range, 77.8%-98.0% across the 5 Primary Care Associates clinic physicians). Of these 275 patients, 206 (74.9%; range, 52.4%-84.7%) were under "good control" (defined by the National Committee for Quality Assurance as an HbA(1c) level of less than or equal to9.5%). The mean (SD) HbA(1c) level was 8.45% (1.98%). Data from the HMO documented a rate of HbA(1c) testing of 36.3% (45/124). compared with 92.7% (115/124) based on case note review at the academic medical center. Conclusions: Current administrative and claims-based information systems have inherent weaknesses if used for performance measurement. Reliance solely on medical record review is time and cost prohibitive. To ensure complete reporting of mandated "quality measures" will necessitate the tracking of data across different healthcare systems.