Membership in a Diabetes Care Network Improves General Practitioners’ Practices for HbA_1c and Microalbuminuria Monitoring: A Cohort Study Among Patients With Type 2 Diabetes

To prevent complications of diabetes, clinical practice guidelines recommend specific frequencies of several checkups. Following these guidelines can decrease all-cause mortality and cardiovascular morbidity (1). France, like other countries, has sought to improve diabetes management, specifically by establishing diabetes care networks (DCNs) that coordinate care, provide continuing medical education for health care professionals, and educate network patients. Nonetheless, less than 5% of all patients with type 2 diabetes participate in these networks. Our aim was to study whether general practitioners (GPs) belonging to a DCN adhered more closely to guidelines for diabetes monitoring than GPs not belonging to a DCN for all their patients with type 2 diabetes and not only those in a DCN. Analyses focused on orders for two diabetes-specific examinations: HbA_1c and microalbuminuria testing.

Using health insurance reimbursement databases in southeastern France, we included 468 GPs in two networks and 468 nonnetwork GPs in the same geographical area, matched one to one by propensity scores (Table 1). We followed their patients treated with oral hypoglycemic agents who were aged 18–79 years from 2008 through 2011 (n = 11,832 and 10,976, respectively) and compared the frequencies of HbA_1c and microalbuminuria examinations in these two cohorts. Each patient’s mean annual number of examinations was calculated for his or her study monitoring period; a dichotomous variable (met or did not meet the threshold in the French guidelines) was then calculated for each test. Associations with GP network membership were analyzed by mixed logistic multivariate regression that took the matching design into account and adjusted for patients’ and GPs’ characteristics (Table 1).

After matching, GPs’ characteristics were similar in both groups, and the patient cohorts differed only slightly. Mean follow-up was 3.1 years in both cohorts. Monitoring was more frequent for both HbA_1c and microalbuminuria in patients of network GPs than nonnetwork GPs, even after adjustment for patients’ and GPs’ characteristics (Table 1).

Lack of information about GPs’ reasons for joining a DCN might cause endogenic bias in our study. Including a variable probably related to these motivations (i.e., GPs’ proportion of patients with diabetes) in the propensity score should, however, have limited such bias. Our results concur with those of previous studies of the impact of disease management programs on the frequency of checkups for type 2 diabetes (2–4). These reports, however, studied only patients belonging to these programs. The strength of the associations between GPs’ membership in a DCN and the frequency of diabetes checkups observed in our study for all patients with diabetes is about the same as that observed in a previous French study of network patients (5). This suggests that GPs belonging to a DCN adhered more closely to guidelines for HbA_1c and microalbuminuria monitoring than other GPs for all their patients with diabetes and not only network patients. Despite modest size effects, the impact on public health care may be significant. Further research is needed in other settings to confirm these results, for they may have important implications for the cost-efficiency of DCNs in France and elsewhere.