S A Hulme1, A Tin-U, K J Hardy, P W Joyce. 1. Diabetes Centre, Whiston Hospital, Prescot and Rennie Eye Clinic, St Helens Hospital, St Helens, UK. shirley.hulme@sthkhealth.nhs.uk
Abstract
BACKGROUND: Debate exists about the optimum way to screen for diabetic retinopathy. Cameras produce a permanent record, but offer patients less choice about when and where to be screened. Optometrists offer flexibility but sensitivity and specificity of schemes have varied considerably, perhaps because of variability in screening methodology and that there is frequently no quality assurance programme. AIMS: To audit our district-wide (population 340000) screening programme for diabetic retinopathy against national targets: sensitivity > 80%, specificity > 95% and referral to review < 3 months. METHODS: Trained optometrists performed slit-lamp examination with Volk lenses (78 dioptre) with standardized reporting. Audit was by ophthalmologist with slit-lamp and Volk lenses through dilated pupils. RESULTS: We examined 872 eyes of 439 patients; 64% were normal, 29% background diabetic retinopathy, 7% sight-threatening eye disease (STED). Sixty-three percent of patients were seen within 6 months of the original screen. Of these, sensitivity for any retinopathy was 72%, specificity 77%, positive predictive value (PPV) 53%, negative predictive value (NPV) 88%. For STED, in this group, sensitivity was 87% and specificity 91%, PPV 30%, NPV 99%. Median interval referral to ophthalmological review was 11.5 weeks with 73% reviewed in under the 13-week target. Of those referred 25% received laser therapy. Eleven patients found to have referable eye disease at their initial screen were not referred to an ophthalmologist by their GP. CONCLUSIONS: We conclude that effective district-wide screening for diabetic retinopathy by optometrists using slit-lamp and Volk lenses is possible; however, only 36% of identified people with diabetes in the district were screened over a 4-year period.
BACKGROUND: Debate exists about the optimum way to screen for diabetic retinopathy. Cameras produce a permanent record, but offer patients less choice about when and where to be screened. Optometrists offer flexibility but sensitivity and specificity of schemes have varied considerably, perhaps because of variability in screening methodology and that there is frequently no quality assurance programme. AIMS: To audit our district-wide (population 340000) screening programme for diabetic retinopathy against national targets: sensitivity > 80%, specificity > 95% and referral to review < 3 months. METHODS: Trained optometrists performed slit-lamp examination with Volk lenses (78 dioptre) with standardized reporting. Audit was by ophthalmologist with slit-lamp and Volk lenses through dilated pupils. RESULTS: We examined 872 eyes of 439 patients; 64% were normal, 29% background diabetic retinopathy, 7% sight-threatening eye disease (STED). Sixty-three percent of patients were seen within 6 months of the original screen. Of these, sensitivity for any retinopathy was 72%, specificity 77%, positive predictive value (PPV) 53%, negative predictive value (NPV) 88%. For STED, in this group, sensitivity was 87% and specificity 91%, PPV 30%, NPV 99%. Median interval referral to ophthalmological review was 11.5 weeks with 73% reviewed in under the 13-week target. Of those referred 25% received laser therapy. Eleven patients found to have referable eye disease at their initial screen were not referred to an ophthalmologist by their GP. CONCLUSIONS: We conclude that effective district-wide screening for diabetic retinopathy by optometrists using slit-lamp and Volk lenses is possible; however, only 36% of identified people with diabetes in the district were screened over a 4-year period.