Knight et al1 recently published in this journal a systematic literature review and meta-analysis of published evidence regarding the effect of disease management programs for patients with diabetes mellitus on processes and outcomes of care. The authors found 24 studies published between 1987 and 2001 evaluating glycemic control and concluded that there was a modest, but significant, reduction of A1C levels of 0.5%. However, at least 9 published studies (contained in references 2-5, only 1 of which was included in the review by Knight et al) demonstrate much better results from utilizing nurses following detailed treatment algorithms under the supervision of a physician. The change in A1C levels under the nurses' care ranged from -3.7% to -0.8%, compared to -1.5% to +0.8% in appropriately matched controls receiving usual care. In all of these studies, the fall in A1C levels was statistically different from the change in the controls. In only 5 of the 24 studies in the review by Knight et al was there a statistical difference in A1C changes between patients followed in a disease management program and their controls.
A number of barriers to the adoption and implementation of new scientific findings to patient care have been identified.2 These include: (1) healthcare provider knowledge; (2) communication between patient and healthcare provider; (3) attitudes and beliefs of the patient, community/culture, healthcare provider and healthcare system; (4) racial and ethnic disparities; (5) variation in settings, including the healthcare system; (6) clinical traditions; (7) socioeconomic status; and (8) cost. It is really not surprising that this disease management approach (ie, nurses following detailed treatment algorithms under the supervision of a physician experienced in diabetes care) improves outcomes. First, the algorithms by definition represent enhanced provider knowledge because they are formulated by diabetes specialists. Second, nurses are generally able to communicate better with patients than physicians are because nurses are more sensitive to the attitudes and beliefs of patients and their communities and culture. Finally and very importantly, nurses have more time to devote to diabetes care than do physicians because not only are nurses not directly responsible for other patient problems as is the physician, but they also usually don't have to see as many patients per unit time.
There is really little evidence that other disease management approaches improve outcome measures (although process measures often do).2 Using nurses to make clinical treatment decisions (under appropriate supervision) has not been widely recognized as an effective disease management approach in diabetes. Isn't it time it was and we establish such systems accordingly? Cost is often cited as a consideration, but economic studies have documented a subsequent cost savings.
Mayer B. Davidson, MD
Los Angeles, California
We agree with Davidson that use of properly trained, supported, and supervised physician extenders, including nurses, may be a cost-effective method of delivering good diabetes care, although evidence supporting this view from randomized clinical trials is sparse. We also agree that glycemic management in many patients with diabetes is less than optimal.
In our study, we defined disease management in a manner we believe to be consistent with interventions referred to as disease management that have been widely implemented in recent years. We searched for, and included, studies based on explicit criteria, including criteria for study design and reporting of estimates of effect. Others may define disease management differently, or include studies based on other criteria, so it is important to interpret results of our review and others in light of the (hopefully) explicit definitions and criteria used in such reviews. Indeed, under the broadest definition, almost anything clinicians do could be defined as disease management.
Such issues should not detract from the important messages conveyed by Davidson calling for improved glycemic management of patients with diabetes. Nor should these issues detract from results of our review.
Kevin Knight, MD, MPH
Los Angeles, California
Enkhe Badamgarav, MD, MPH
Cerner Health Insights