Using the nominal group process for achieving consensus among several persons informed about a given topic [1], the forum of approximately 30 individuals identified 64 potential recommendations for change and then ordered them by rank. The forum leadership subsequently combined similar issues and, using the weights assigned to various items by the nominal group process, made a prioritized list of recommended changes (Table 1).

First Priority: Universal Access to Comprehensive Preventive Services

Most diabetes care, for both insulin-dependent diabetes mellitus (IDDM) and NIDDM, that occurs before the symptomatic onset of a specific diabetic complication is secondary and tertiary prevention. Prevention, as a component of health care, occurs at three levels: primary, secondary, and tertiary. Primary prevention works to prevent the disease from occurring at all, as with immunization for a contagious illness. Secondary prevention addresses the onset of disease complications once the disease is established, as with control of hypertension to delay or eliminate secondary renal disease. Tertiary prevention refers to modification of the effects of a complication, as with detecting and treating proliferative and preproliferative retinopathy to prevent blindness in diabetic patients. Primary prevention of diabetes is currently the subject of large-scale investigative studies supported by the National Institutes of Health. The concepts of secondary and tertiary prevention in diabetes, however, are supported by existing data from several previous studies, the most recent and impressive of which is the Diabetes Control and Complications Trial (DCCT) [2].

Diabetes care is not "treatment" in the classic sense—a maneuver to eliminate or reduce an unpleasant symptom or eliminate or reduce a pathophysiologic state—but rather a constellation of patient- and provider-initiated self-care steps maintained over the patient's lifetime to prevent something undesirable from occurring. Diabetes, as a chronic multisystemic condition in which patient self-management day to day and year to year constitutes 90% or more of the overall program, needs a long-term prevention program. Preventive services are generally not well handled by the current U.S. health care delivery system, which is largely designed for (and does very well with) acute and episodic illness. Preventive strategies get lost in the faster-paced, symptom-driven arena of acute and episodic care; they neither demand nor get physician attention and are weakly reimbursed. The generally asymptomatic character of diabetes in the precomplication phase does not provide a natural force for patients to seek necessary care. It is relatively easy for patients to approach their condition casually, and, conversely, it is relatively difficult for patients to approach their condition aggressively to avoid problems they perceive to be years away. In addition, preparing the patient to assume responsibility for self-management requires substantial interaction with other members of the health care team besides the physician, in particular, the nurse-educator and the dietitian. The frequent lack of availability of and reimbursement for these professionals for nonhospitalized diabetic patients is problematic.

Because of the chronicity of NIDDM and its recognized onset in the fifth and sixth decades of life, a disjunction exists between third-party payers who might pay for services during the preventive phase of diabetes and those who do pay for services after complications ensue. Payers who might provide for the preventive phase of diabetes care (much of which occurs before age 65) have small incentive to aggressively support a complications-delaying strategy, as another payer, the federal government, ends up paying for most of the cost of care after complications arise and for disability status resulting from those complications at whatever age they arise. This disjunction between the "prevention payer" and the "complications payer" was thought by the conference attendees to be a serious barrier to providing intensive management to patients with NIDDM.

The forum recommended that the U.S. health care delivery system be amended so that the important secondary preventive care that diabetic patients need through most of their illness be available to them regardless of employment status, insurance coverage, and other factors that, if not addressed, would result in the uneven availability of preventive services. The public health sector has long had a mandate to promote prevention, as well as a significant role in promoting management of chronic disease, and might be the appropriate agency to offer the preventive approach to diabetes that is now missing.

Second Priority: Co-management between Primary and Specialty Care Using an Integrated Data System for Patients with NIDDM

Forum members recognized that 90% to 95% of diabetes care in the United States is directed by the primary care physician, with the remaining fraction directed by physicians with special interest and training in diabetes and other metabolic disorders. With the current emphasis on primary care in the U.S. health care delivery system, it is unlikely that this ratio will change in the foreseeable future. However, most patients with diabetes would be well served during their illness if they had access to specialists as complications develop. In addition to specialty care for complications, a system of "shared care" between primary care physicians and diabetes specialists has been proposed [3]. The specialist and specialty team would review the overall status of the progress of diabetes care at periodic intervals, reset goals and directions as appropriate, and make recommendations to the primary care physician on how to carry out this plan. Making decisions about starting insulin therapy, identifying and recommending strategies for managing cardiovascular disease risk factors, and periodically screening for diabetes complications (for example, detection of microalbuminuria) are examples of how diabetes specialists could interact with primary care physicians and share responsibility for patients' management. A common data system for everyone offering care to a given patient would integrate this care into a seamless system and would also enhance reporting procedures and facilitate clinical research.

The validity of shared care is clear as it applies to IDDM, and physicians at both ends of a shared-care spectrum recognize this. The applicability of the concept to NIDDM has been, until recently, less clear. The forum recognized, however, that NIDDM in many ways is a more complex illness than IDDM; management approaches are more controversial, intensification of therapy may actually be harder to accomplish, and comorbidities are more numerous. A co-management system would permit responsibility for ongoing comprehensive care of patients with NIDDM to remain with the primary care physician but would provide the benefits of a specialty team to assist the primary care physician (and the patient) at points during the illness when their input would be positive.

Third Priority: Other Important Modifications to the Current Health Care Delivery System for Patients with NIDDM

Increasing Patient and Public Awareness of the Seriousness of NIDDM

Non–insulin-dependent diabetes mellitus is not considered or treated as a serious condition by many primary care physicians, particularly those whose original training antedates the relatively recent trend toward intensification of management of most diabetic patients. The disease lacks the dramatic symptoms and clinical events that characterize IDDM. The insidious onset, generally quiescent nature, and frequency of NIDDM in the older population have combined to tag this disease as mild. The common phrase used by clinicians toward patients with newly discovered NIDDM—a "touch of sugar"—captures this attitude very well.

Public opinion, however, can be changed about conditions that appear to be mild, are quiescent for long periods, but nevertheless are serious. The National High Blood Pressure Education Program [4] and National Cholesterol Education Program [5] are clear examples of successful public education programs that have motivated the general public to seek detection and treatment as indicated of these often asymptomatic and previously lightly regarded conditions. For both hypertension and hypercholesterolemia, the availability of a specific number for the upper limit of normal has attracted attention and motivated behavior. This tactic has not been used with diabetes. As a result, many physicians and their patients are appropriately concerned about a systolic blood pressure of 240 or a cholesterol value of 240 but are not particularly motivated to correct a blood glucose value in the same range.

Special Training in NIDDM Management for Primary Care Physicians and Residents in Primary Care Specialties

Most residency programs in the primary care specialties—internal medicine, family practice, and pediatrics—assign their trainees to one-half day per week in a continuity clinic throughout their 3-year program. Currently, in most residency programs, experience with diabetic patients occurs during an intensive but short-lived rotation on the endocrinology and metabolism service, and these services are often preoccupied with patients with IDDM. Given the frequency of NIDDM in the professional career of the internist and family physician (the second most common principal diagnosis for office visits to internists [6]) and the national trend to include greater experience with ambulatory patients (as opposed to the traditional, nearly exclusive, inpatient focus), it would seem reasonable to alter the curriculum of internal medicine and family practice residents to intensify their training in this condition. The forum recommended that training program directors in these specialties make a deliberate effort to include substantial numbers of patients with NIDDM in the continuity clinics of each of their house officers.

The forum also recommended that the continuing education modalities serving primary care physicians assume the responsibility for updating this large physician population on intensive management for NIDDM. These modalities include the medical literature, review articles in the medical literature, local and national continuing medical education programs, hospital-based medical staff continuing education activities (grand rounds, clinical conferences, and so forth), and protocol development within managed care systems.

A proposal for a restructuring of primary care residencies and continuing education programs for practicing primary care physicians to amplify training in diabetes has recently been made [7]. This proposal was included in a series of recommendations that emanated from the scientific diabetes community after publication of the DCCT results [2], recommendations on how the DCCT results could be incorporated into clinical practice. Although the DCCT specifically addressed issues concerning patients with IDDM, the recommended changes in graduate and continuing education for primary care physicians are equally applicable to NIDDM.

Developing and Disseminating Standards of Care for NIDDM

The American Diabetes Association periodically develops and publishes its recommended guidelines for the treatment of diabetes [8]. These valuable position statements have been developed by panels of authorities in the field of diabetes and reflect expert opinion on various aspects of diabetes care. However, these standards are published in the diabetes literature and seen by only a small fraction of primary care physicians. The forum recommended that the National Institute of Diabetes and Digestive and Kidney Disease (NIDDK) impanel a consensus conference on the standards of care for NIDDM and publish the results widely in the medical literature. In addition, the NIDDK should exert its influence to include this topic in major national physician professional meetings (such as those of the American Medical Association, American College of Physicians, and American Academy of Family Physicians). This suggestion was included in the report to the director of the NIDDK, submitted by representatives from the Diabetes Research and Training Centers advising the director on methods for implementing the findings of the DCCT on a broad scale [3].

Establishing a Central Coordinating Agency for All Aspects of Diabetes Care

The forum members recognized that many of the recommendations listed above would be inefficiently and incompletely implemented unless a national coordinating agency was given that responsibility. In other words, a deliberate effort at translation would be required. Translation, as originally conceived by the National Commission on Diabetes [9], is an active process of disseminating new scientific information to all patients who might benefit from it through identification and removal of barriers to such dissemination. A proclamation that a well-done scientific study (such as the DCCT) should be widely adopted will not ensure that it is. Adoption and dissemination take effort and direction (and usually a substantial research and development process), and these forces do not arise spontaneously. They must be created. Creation of an agency to coordinate the efforts of public and private agencies to modify health care for diabetic patients, implement the valuable findings of recent research, and recommend areas of needed research and development was deemed essential.

The forum identified the need for a coordinating agency to accomplish these tasks but did not pursue discussion of governance, authority, or funding of such a group. However, the authors of Metabolic Control Matters [3], the report submitted to the NIDDK on methods to implement the findings of the DCCT, have recently (subsequent to the NIDDM conferences described in this supplement) recommended that the NIDDK establish a National Diabetes Education Program similar to the National High Blood Pressure and National Cholesterol Education Programs previously described. The National Diabetes Education Program could supply the coordinating function proposed by the forum.

Developing Outcome-Based Incentives for Patients and Providers

Outcome standards can be developed to set target goals for the management of NIDDM. The development and testing of outcome standards are in their infancy, but several studies have pointed the way toward establishing standards. Target blood glucose levels are an obvious example, modified according to factors such as age, comorbidity, and obesity, all of which have been shown to be associated with higher blood glucose levels. A further step is developing standards for preventable microvascular complications in a group of patients belonging to a practice, an independent practice association, a health maintenance organization, a state entity, or some other medical care entity.

Another outcome measure is quality of life or functional status. From data collected in the Medical Outcome Study [10] and the PORT study on NIDDM [11], it is now possible to develop a range of function within which the average patient should fall, with a confidence interval or a standard deviation indicative of the range below which a patient with comprehensive, coordinated, high-quality technical and interpersonal care should not fall under ordinary circumstances (unpublished data). For example, in these studies, using Short Form 36 (SF36) from the Medical Outcome Study [10], the mean score for physical function was 68 compared with 84 in a generally healthy population. These standards are derived from samples of more than 5000 patients. This number can be further modified for age and for the presence of coexisting conditions and complications. The accumulated experience can be synthesized to produce outcome standards that can be used as benchmarks to assess quality and complement guidelines in the care of patients with NIDDM.

Dr. Greenfield: New England Medical Center, Box 345, 750 Washington Street, Boston, MA 02111.