A clinical case manager, who works closely with patients on the lifestyle and medication changes required to achieve glycemic control, is a prominent feature of most successful diabetes clinics or programs. The case manager is usually a nurse or nurse practitioner with additional training or experience in diabetes care and in techniques to help patients become more capable self-managers of their illness. In this issue, Aubert and colleagues [8] report the results of the first randomized trial testing the effects of a primary care-based case management program for diabetic patients. The intervention achieved significant reductions in hemoglobin A_1c values and improvements in self-reported health.

The credibility of the results is supported by the random allocation of patients within the same practices to different interventions. This precludes the possibility that baseline or secular differences in the quality of care of providers or practices could have influenced the results. Although there was considerable loss to follow-up, the investigators used conservative approaches to examine the impact of the losses on study results and still found significant differences in key outcomes. The external validity of the findings is supported by the representative nature of the patients, who were recruited from a managed care diabetes registry; the location of the study (busy private practices); and the reliance on existing personnel at health maintenance organizations to create the guidelines and carry out the intervention.

Case management by nurses was shown to be effective in a nonrandomized trial of diabetes care [9] and in randomized trials involving other major chronic conditions. Comprehensive nurse case management programs have reduced cardiovascular risk factors in patients with established coronary heart disease [10], rehospitalizations in patients hospitalized for congestive heart failure [11], and hospitalizations [12] and nursing home admissions [13] among community-dwelling older adults.

These interventions work, in my view, by creating new, more responsive systems of care for chronically ill patients rather than just adding a new provider. The interventions all have common features. First, the program addresses the critical role that patients play in managing their illness by assuring that case managers are trained educators with ready access to a defined set of high-quality educational offerings. Second, the delivery of care is redesigned to give patients more time with the case manager, access to a broader array of resources and expertise, and closer follow-up. Most often, critical aspects of care-care planning, home monitoring, medication regulation, and self-management support-are delegated to the nurse-case manager, who assures close follow-up, often by telephone [14, 15]. Third, explicit guidelines and regular communication with clinical experts who are integral to the intervention guide clinical management. Finally, the collection and organization of relevant data for individual patients and populations support population-based care, reminders, and feedback. These system elements are consistent with the changes found in other studies of chronic illness care to be associated with better outcomes [16].

Aubert and colleagues [8] have shown once again that diabetes care can be substantially improved by enhancing primary care, not bypassing it [17-19]. They used personnel and resources already available within a health maintenance organization to create comprehensive system change capable of substantially improving key clinical outcomes. But is such an intensive intervention affordable? Aubert and colleagues and others have estimated that a nurse-case manager can care for a caseload of 250 patients. A health system serving 100 000 patients is likely to have 3000 or more diabetic patients. If case management were part of ongoing care for most patients with chronic illness, such a system would require 12 to 15 case managers for diabetes alone, which is probably not affordable.

Most health systems are experimenting with case management interventions for common chronic illnesses but are limiting them to the highest-risk patients [20]. An important feature of the study by Aubert and colleagues is the lack of preintervention screening or targeting. They recruited all participants from a diabetes registry that included patients whose initial hemoglobin A_1c values would be considered to be under reasonably good control. The intervention achieved significant reductions in hemoglobin A_1c values across the full spectrum of baseline values. Given what we know about the epidemiology of diabetes and its complications, we should not deprive patients with lower hemoglobin A_1c values of the benefits of better glycemic control.

A sizable proportion of patients, especially younger ones, in the nurse case management arm were lost to follow-up. Although this might be viewed as a deficiency of the program given the intensity of follow-up, it might also reflect the fact that most of the reduction in glycemic control was seen within the first 6 months. This suggests that intensive nurse case management may be instituted for several months to stabilize the treatment regimen, build the patient's skills and confidence as self-manager, and improve glycemic control. Nurse case management would then be continued or reinstituted only for patients experiencing difficulty. This approach ensures that all patients receive solid grounding in the management of their disease while controlling cost by limiting the number of nurses required.

The study by Aubert and colleagues adds to the growing literature showing that organized systems of chronic illness care that include nurse case management can substantially improve important outcomes in major chronic illnesses. It is critical that these interventions be examined and understood fully because they represent more than the addition of a new provider to the practice team. Approaches that provide more meaningful visits for diabetic patients and more intensive follow-up, ensure closer adherence to evidence-based guidelines, meet the self-management needs of patients, and ensure that meaningful clinical information on individual patients and the larger population of patients are readily available may improve outcomes regardless of the exact combination of providers. Because they provide guideline-driven clinical management and sophisticated self-management support, nurse-case managers may be the most efficient nidus for building more effective systems of care for chronic illness.