RESUMO
PURPOSE: Analysis of the validity of the non-mydriatic camera in detecting diabetic retinopathy and severity levels which can be referred to the ophthalmologist. Analysis of risk indicators of diabetic retinopathy which help increase the sensitivity of the screening: development time of prolonged diabetes (more than 20 years); poor metabolic control (HbA1C > 10%). METHOD: A cross-sectional study was designed to determine the validity of a diagnostic test (non-mydriatic fundus camera Topcon model TRC-NW6S). As criteria for evaluation we used validity indicators, predictive values, indicators of clinical use, and concordance. SETTING: Primary Care. Patients with severe cataracts were excluded. The sample size was 376 eyes of 188 type-2 diabetic patients. RESULTS: With reference to the distribution of diabetic retinopathy we found that 41.25% had a variable level of retinopathy. The validity indicators analysed improved as the number of fundus fields studied increased, obtaining sensitivity levels of greater than 75%; specificity greater than 95%; positive predictive values greater than 90% and almost perfect concordance when two or three fundus fields were used. CONCLUSIONS: The analysis of retinal photography in digital format can be considered a valid method with respect to the gold standard, and the severity indicators analysed can play a guiding role in obtaining greater prevalences of derivable diabetic retinopathy, but cannot be used as an isolated test for making clinical decisions.