A.L. Peters, A.P. Legorreta,
R.C. Ossorio, and M.B. Davidson
Diabetes Care 19:601-606, 1996.

Summary and Commentary
by Irl B. Hirsch, MD

Objective. To document the quality of diabetes care provided to patients in a large health maintenance organization (HMO) during a 1-year period.

Design. Review of 353 randomly selected charts of 14,539 diabetic members of a large California HMO.

Subjects. The mean age of the study population was 55.2 years. Of the 353 patients, 30 had insulin-dependent diabetes mellitus, and 56% were male. For the patients with non-insulin-dependent diabetes mellitus, 12% were on diet alone, 60% received sulfonylureas, 15% received insulin, and 12% received a combination of insulin and a sulfonylurea.

Measurements. Measurements included visit frequency and documented foot exam; frequency and levels of glycated hemoglobin, blood pressure, and finger-stick glucose; and initial and final values of fasting lipid levels and random blood glucose concentration. All of these were compared to the Ameri-can Diabetes Association standards of care.

Results. Overall, patients averaged 4.5 visits to their primary care provider (PCP). A blood pressure was measured at least once during the study period in 86% of the patients, while foot exams occurred at 6% of all visits.

Referrals were uncommon: 22% were referred to an ophthalmologist, 5% to a podiatrist, 10% to a dietitian, and 5% to an endocrinologist. Eight percent attended some sort of diabetes education class.

Glycated hemoglobin was only measured in 44% of the patients during the study period, and 39% of those measured had one or more level >10%. Fasting glucose was measured in 35% of patients. There were no differences in the fasting glucose levels before and after the study period (191 vs. 183 mg/dl).

Total cholesterol was measured in 56% of the population, but low-density lipoprotein and high-density lipoprotein cholesterol were measured in less than one-third of the patients. Fifty-two percent had no assessment of proteinuria, and 92% did not have a documented foot exam.

Conclusions. Diabetes management in this HMO is inadequate, and this creates an even greater future burden on the health-care system and negative consequences for patients.

Commentary
This report confirms what findings from numerous other studies^1-5 have concluded: diabetes care provided in the population being studied does not meet national standards. This particular study was not designed to compare the care practices for these HMO patients to care practices for patients in fee-for-service (FFS) settings in the same communities, whether by the same or different physicians. Despite the concerns of many, there actually are no data to date concluding that HMO diabetes care is inferior to FFS care.

Of course, health-care economists are less interested in endpoints such as number of glycated hemoglobins measured and foot examinations. Of more importance are outcomes such as emergency room visits, hospitalizations, blindness, lower extremity amputation, dialysis, and death.

When practice patterns do not meet acceptable national standards, or when routine services or supplies are intentionally withheld (e.g., specialty referrals when indicated, strips for home blood glucose monitoring, or routine assessment for microalbuminuria), is it accurate to extrapolate that major diabetes-related outcomes suffer? Unfortunately, we must rely on the now-familiar term of “evidence-based medicine” to guide us, and most of the practice surveys such as this one by Peters and colleagues do not have the statistical power (or duration of observation) to measure the “non-surrogate” endpoints, such as blindness or mortality.

Is it appropriate to extrapolate the results of this study to all HMOs? The results of this survey probably do not surprise many. On the other hand, the authors correctly state their belief that the care provided to these patients is likely representative of the care provided around the state of California, no matter which type of insurance is used. Certainly, the results are consistent with the findings of numerous other surveys.^1-5

The fundamental issue, therefore, is not the method of payment to the physician, but the quality of care provided to all patients with diabetes. This, of course, directs the blame to our medical education system, and not necessarily to the form of insurance used.

What is unfortunate, however, is that an HMO has a better opportunity to influence practice patterns of physicians than the traditional FFS system. To date, data do not exist demonstrating better diabetes care from providers in HMOs. On the contrary, recent reports in the news media suggest that economic incentives are the overriding influence of care, particularly for “for profit” HMOs.

The data from this study also deserve further scrutiny. Fifty-six percent of patients did not receive even one glycated hemoglobin measurement, 52% had no assessment for proteinuria, and 92% did not have a documented foot exam. This is similar to other surveys, and few would disagree that it is abysmal.^1,2

Of more concern is the fact that 86% of patients had a blood pressure measurement during the study period. That means 14% did not. Hypertension is estimated to be responsible for 30–75% of all diabetes-related complications, and more than 2.5 million Americans have both diabetes and hypertension.⁶ The authors suggested that these physicians provided “. . . adequate care in terms of blood pressure measurement and treatment, compared to the inadequate diabetic care provided.” I consider not measuring the blood pressure in any patient clinically significant due to the importance of treating this serious condition.

What should we conclude from these data? First, chart reviews, a more accurate assessment of practice patterns than physician recall from surveys, show much less diabetes care provided compared to the latter. This is an important point when evaluating these types of studies. Next, although HMOs have perhaps a better opportunity to change practice patterns for improvement of care, data showing that they have do not exist. On the other hand, at least for diabetes, data also do not exist that care in HMOs is worse. Finally, overall care for diabetes in the United States is poor, and steps need to be taken to improve it. Most likely, initiatives will have to come from medical schools and residency programs, since at least up until now, our traditional mechanisms of continuing medical education apparently have been ineffective.⁷

References

¹Miller KL, Hirsch IB: Physicians’ practices in screening for the development of diabetic retinopathy and the use of glycosylated hemoglobin levels. Diabetes Care 17:1495-97, 1994.

²Marrero DG: Current effectiveness of diabetes health care in the U.S. Diabetes Reviews 2:292-309, 1994.

³Jacques CHM, Jones RL, Houts P, Bauer LC, Dwyer KM, Lynch JC, Casale TSM: Reported practice behaviors for medical care of patients with diabetes mellitus by primary-care physicians in Pennsylvania. Diabetes Care 14:712-17, 1991.

⁴Kenny SJ, Smith PJ, Goldschmid MG, Newman JM, Herman WH: Survey of physician practice behaviors related to diabetes mellitus in the U.S. Diabetes Care 16:1507-10, 1993.

⁵Bechner RJ, Cowie CC, Howie J, Herman WH, Will JC, Harris MI: Ophthalmic examination among adults with diagnosed diabetes mellitus. J Am Med Assoc 270:1714-18, 1993.

⁶American Diabetes Association : Consensus statement: Treatment of hypertension in diabetes. Diabetes Care 19(Suppl 1):S107-13, 1996.

⁷Hiss RG, Davis WK: Intensified glycemic control and changes in training and continuing education of physicians. Diabetes Reviews 2:310-21, 1994.

Irl B. Hirsch, MD, is an associate professor of medicine at the University of Washington School of Medicine, in Seattle.