96. Comparison of Primary Care Provider Types in the Management of Diabetes Mellitus
A Shih, SUNY at Stony Brook, Dept of Preventive Medicine; M Graber, Northport VAMC and SUNY at Stony Brook, Dept of Medicine; H Mehdizadeh, Northport VAMC; RE Czaja, Northport VAMC
Objectives: A central theme of medical reform over the past decade has been the development of organizational models based on primary care. The rush to establish this model in health care systems has created a substantial shortage of primary care physicians. In the short term, the urgent need for primary care providers has been met in many systems by utilizing nonphysician clinicians (e.g., physician assistants and nurse practitioners) and medical subspecialists (e.g., cardiologists and gastroenterologists) for the delivery of primary care. The objective of this study is to compare the management of diabetes mellitus by different primary care provider types- nurse practitioners (NPs), generalist physicians (GPs), and medical subspecialist physicians (SPs) practicing outside their subspecialty- using HbA1c outcomes, referral patterns, compliance with treatment guidelines, and prescription patterns.
Methods: This study utilized a cross-sectional analysis of data from October 1 1998 to November 30 1999 from an outpatient primary care program within a Veteran's Affairs Medical Center (VAMC). The study subjects were a total of 1740 diabetic patients enrolled in the Northport VAMC Primary Care Program and 44 primary care providers- 10 NPs, 14 GPs, and 20 SPs. Multivariate logistic regression and chi-square tests of associations were the primary tests used to analyze the differences in HbA1c outcomes, referral patterns, compliance with treatment guidelines and prescription patterns.
Results: The different primary care provider types had similar patients, referral patterns, and HbA1c outcomes. In multivariate logistic regression analysis, when compared to GPs, the odds ratio for severely uncontrolled HbA1c (HbA1c > 9%) for NPs was 1.15 (95% CI, .79-1.69), and for SPs, 1.00 (95% CI, .70-1.42). However, there were significant differences in compliance with treatment guidelines. In the performance of recommended labs for diabetic patients, NPs had higher performance rates in five of six tests measured; the difference was statistically significant for three of those tests. The GPs' rates exceeded the SPs' rates for all six tests measured; again, the difference was statistically significant for three of those tests.
Conclusions: Our findings suggest that these three types of primary care providers are equally effective in achieving glycemic control but differ in their compliance with diabetes clinical care guidelines.
Impact: The findings generally support the use of nonphysician clinicians and subspecialist physicians in the provision of primary care. These findings should be considered in future design and redesign of primary care programs.