An instrumental variable analysis of the impact of practice guidelines on improving quality of care and diabetes-related outcomes in the elderly Medicare population.

The effect of the diabetes practice guideline recommending >or=2 HbA1c tests annually on diabetes-related outcomes was evaluated using Medicare claims data. The study population included 1998 and 1999 incident diabetes patients aged >or= 67 years, who were Medicare eligible and without known diabetes-related complications at baseline. Number of HbA1c tests was measured 1 year after diabetes incidence. All-cause death and diabetes complications were identified during follow-up, through December 2003. The analysis was conducted with an instrumental variable method and a bivariate probit model, controlling for individual, social, and health care system characteristics. Among 13 033 patients, 27.1% followed the practice guideline. Receiving >or=2 HbA1c tests annually was significantly associated with a decrease in probability of 28.8 percentage points for macrovascular complications, 28.7 for atherosclerotic heart disease, and 23.1 for chronic kidney disease or end-stage renal disease in the 4-year follow-up period.