In response to Ms. Nettles, we have no empirical evidence to suggest that decreasing physician involvement would contribute favorably to the effectiveness or cost-effectiveness of intensive diabetes management. Our study and a study by Peters and Davidson (1), as well as the Diabetes Control and Complications Trial (2), relied on the ongoing active collaboration between physicians and nurses to achieve the desirable outcomes obtained. To presume or conclude an economic benefit by decreasing physician involvement is not supported by our study or the current literature. Indeed, Peters and Davidson (1) demonstrated that patient noncompliance with physician-supervised nurse case management was common and was associated with worse outcomes. Accordingly, decreased physician involvement might be associated with lower participation in or adherence to nurse case management and worse population-based outcomes.

Although we cannot comment on behalf of the author of the editorial, we agree with the issues raised by Dr. Davidson. Targeting high-risk diabetic patients for short-term case management may enhance the efficiency of the intervention. Whether a nurse management intervention such as the one described is affordable or even cost-effective is still uncertain and can best be addressed by a rigorous economic evaluation. We have almost completed a cost-effectiveness analysis of nurse case management as implemented in our study and hope to contribute data to this issue soon.

Finally, in response to DeBusk and colleagues, we caution that specifically targeting "difficult" patients might not enhance the efficacy of nurse case management. Although older, sicker, and less socially advantaged patients are at greater risk for adverse outcomes, it is not clear that intensive intervention is equally feasible or effective. We agree that there is great potential for broader multiple risk factor interventions, particularly for the cardiovascular risk factors mentioned in their letter. Passive interventions such as letters to the primary care physicians to address hypertension, dyslipidemia, and smoking cessation were a part of our nurse case management intervention. However, none of the associated outcomes differed by study group. Future programs should implement expanded protocols and management algorithms to address other risk factors relevant to the disease state (1).