Abstract

OBJECTIVES: To quantify case detection and blindness prevention attainable through screening for diabetic retinopathy in a district population. DESIGN: Literature review including a pooled estimate of screening test sensitivity, and quantitative modelling, including sensitivity analyses. SETTING AND PATIENTS: The diabetic population of a typical district health authority or health board. MAIN RESULTS: Evidence suggests that in a British general practice based diabetic population, prevalence of retinopathy requiring treatment would be between 1% and 6%; annual incidence of blindness among diabetics with retinopathy requiring treatment would be between 6% and 9%; sensitivity of screening tests in detecting retinopathy requiring treatment would be between 50% and 88%; and treatment could prevent 77% of expected cases of blindness. Of those screened, about 4% would be correctly detected as requiring treatment during an initial screening round, but this yield could decrease to about 1% in subsequent annual screening rounds. Of those treated, about 6% would be prevented from going blind within a year of treatment and 34% within 10 years of treatment. CONCLUSIONS: Screening and early treatment of diabetic retinopathy can prevent substantial disability. The effectiveness and efficiency of screening could be enhanced by improving the performance of current tests or increasing use of mydriatic retinal photography, and by increasing uptake, particularly among diabetics at greatest risk.