Cost-Effectiveness of Detecting and
Treating Diabetic Retinopathy

J.C. Javitt and L. Aiello
Ann Intern Med 124:164-69, 1996

Summary and Commentary by
Richard Eastman, MD

Objective. To determine the cost and cost-effectiveness of ophthalmologic screening and treatment of diabetic retinopathy.

Design. Computer modeling using probabilistic methods with hazard rates and treatment effects from population-based epidemiologic studies and clinical trials.

Measurements. Cost of screening and treatments were measured. The effect of screening and treatment on person-years of sight, and on life expectancy adjusted for quality of life (quality-adjusted life years [QALY]) was also assessed.

Results. Implementation of appropriate screening and treatment is predicted to save 710,800 person-years of sight more than the 413,200 person-years of sight saved by the current level of screening. The cost (in present-value dollars) would be $1,757 per person-year of sight saved, and $3,190 per QALY saved (insulin-dependent diabetes mellitus [IDDM] $1,996/QALY saved; non-insulin-dependent diabetes mellitus [NIDDM] $3,339/QALY saved).

Conclusions. Screening and treatment of diabetic eye disease saves sight and is highly cost-effective. The results support widespread screening and treatment for diabetic eye disease.

Commentary

Sight-threatening retinopathy is a common complication of diabetes. Some 40,000 patients per year are estimated to develop proliferative retinopathy with high risk characteristics, and 95,000 per year are estimated to develop macular edema.¹ Clinical trials published more than a decade ago proved that photocoagulation significantly reduces the risk of blindness.^2,3 Sadly, less than half of all patients with diabetes report having seen an ophthalmologist or having had a dilated eye examination within the preceding year.¹

Based on current estimates of diabetes prevalence⁴ and rates of blindness observed in Wisconsin between 1984 and 1992,¹ about 23,000 people with diabetes are estimated to become legally blind each year. The risk is highest in NIDDM patients not taking insulin (~18,000 cases/year), second-highest in NIDDM patients taking insulin

(~5,000 cases/year), and lowest in patients with IDDM (~ 400 cases/year). Hopefully, educational efforts by lay and professional organizations,⁵ and by the National Eye Health Education Program of the National Eye Institute will reverse the current state of affairs.

The authors look at the treatment of retinopathy from a health economics point of view, and conclude that screening and treatment of diabetic retinopathy is more cost-effective than many other medical interventions, most of which are universally available. Their analysis is conservative and sound. The message should be clear to all health-care providers: screening and appropriate treatment for diabetic retinopathy must be implemented throughout the U.S. health-care system.

Implementation of appropriate eye care is necessary, but one should not conclude that it is sufficient to prevent all cases of blindness. Careful inspection of the appendix to this article reveals some interesting caveats about eye disease in people with diabetes.

The risk of proliferative diabetic retinopathy (PDR) is highest in IDDM; NIDDM patients requiring insulin have risks that are only slightly lower. The risk is low in those with NIDDM who do not require insulin. Panretinal photocoagulation is highly effective, reducing the annual risk of blindness from 8.8% per year in untreated patients to 1.5% per year in those who have received photocoagulation, an 83% reduction in risk.

The situation is quite different with macular edema. Macular edema occurs earlier and at much higher rates in NIDDM than in IDDM. The risk is nil during the first 4 years of IDDM, while the rates are 4.6% and 7.8% per year in NIDDM patients taking and not taking insulin, respectively. The annual incidence 5–7 years after diagnosis of diabetes is 10 times greater in NIDDM patients not taking insulin, and 25 times greater in those taking insulin, compared with IDDM patients. Fur-thermore, macular edema is much less responsive to focal photocoagulation. Treatment of clinically significant macular edema reduces the annual risk of blindness from 5% per year in untreated patients to 3.3% per year, only a 34% reduction in risk.

Failure to note the above caveats might lead one to conclude that screening and treatment of diabetic eye disease will prevent all cases of blindness due to diabetic retinopathy. The reality is that screening and treatment can prevent most cases of blindness due to PDR, but less than half of the cases of blindness resulting from macular edema. A high residual rate of blindness due to macular edema will remain even if all patients are detected early through screening and are treated with appropriate focal photocoagulation.

The Diabetes Control and Comp-lications Trial (DCCT)6 showed that treatment of IDDM that achieved and maintained HbAlc levels of ~7% reduced the rate of development of significant nonproliferative retinopathy by 76% in those without retinopathy at the time of entry into the study. In the secondary treatment cohort (with baseline non-proliferative retinopathy), the risk of proliferative retinopathy was reduced by 47%. A similar study conducted in Japanese people with NIDDM⁷ showed nearly identical results. However, the effect of intensive treatment on macular edema in the DCCT was less, with only a 26% (not statistically significant) reduction in risk.⁶

These data are worrisome with respect to preventing macular edema-related blindness, particularly in NIDDM. The patients in the DCCT who developed macular edema had diabetes for an average of 9 years at the time of entry into the study, and all had non-proliferative retinopathy.6 Epidemiologic data suggest that patients develop NIDDM 4–7 years before clinical diagnosis,⁸ and numerous studies have shown that about 20% have non-proliferative retinopathy at the time of diagnosis of diabetes. This may explain, at least in part, the high rates of macular edema in those with NIDDM in the first 10 years after clinical diagnosis, compared with IDDM. The clinical trial data6 suggest that treatment with the goal of normoglycemia, instituted at the time of diagnosis, may have less impact on macular edema in NIDDM than on other microvascular complications, particularly in those who already have non-proliferative retinopathy.

A comprehensive program to reduce diabetes-related blindness should include screening and treatment of retinopathy, and diabetes treatment with the goal of normoglycemia. Eliminating macular edema-related blindness will require intervening earlier in the natural history of NIDDM, during the undiagnosed phase or before development of diabetes, or development of more effective treatments.

References

¹Klein R, Klein BEK: Vision disorders in diabetes. In Diabetes in America 2nd Edition. National Diabetes Data Group, Eds. Bethesda, MD, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, 1995, p. 293-338.

²Diabetic Retinopathy Study Group: Photocoagulation treatment of proliferative diabetic retinopathy: clinical application of the Diabetic Retinopathy Study (DRS) findings. Ophthalmol 88:583-600, 1981.

³ETDRS Research Group: Photocoagulation for macular edema. Arch Ophthalmol 103:1796-1806, 1985.

⁴National Diabetes Information Clearinghouse: Diabetes Statistics. NIH Publication No. 96-3926, 1995.

⁵American Diabetes Association: Position statement: Standards of medical care for patients with diabetes mellitus. Diabetes Care 19(Suppl 1):S8-15, 1996.

⁶The DCCT Research Group: The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 329:977-86, 1993.

⁷Ohkubo Y, Kishikawa H, Araki E, Miyata T, Isami S, Motoyoshi S, Kojima Y, Furuyoshi N, Shichiri M: Intensive insulin therapy prevents the progression of diabetic microvascular complications in Japanese patients with non-insulin-dependent diabetes mellitus: a randomized prospective 6-year study. Diabetes Res Clin Pract 28:103-17, 1995.

⁸Harris MI, Klein R, Welborn TA, Knuiman MW: Onset of NIDDM occurs at least 4-7 years before clinical diagnosis. Diabetes Care 15:815-19, 1992.

Richard Eastman, MD, is the director of the Division of Diabetes, Endo-crinology, and Metabolic Diseases at the National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, in Bethesda, Md.