Our aim was to introduce the Hungarian pilot telemedical screening program for diabetic retinopathy (DRP) and also to evaluate the efficacy of non-mydriatic fundus photographs. A total of 502 eyes of 251 diabetic patients were photographed with a non-mydriatic digital fundus camera in a tertiary diabetology care center. These three 45°-field images were transmitted to the reading center via Internet, where they were graded by two independent ophthalmologists. After non-mydriatic photography (NM method), 28 patients were also examined in mydriasis by an ophthalmologist (O method) and were also photographed in mydriasis (M method). For the comparison of the three methods the kappa statistic was used. With non-mydriatic imaging of 502 eyes no retinopathy was found in 74.5%, DRP was detected in 15.5%, while 10.1% of the photos were ungradable. The rates of DRP severity levels were: 13.55% mild/moderate non-proliferative, 0.59% severe and 1.39% proliferative DRP. Comparing the results of the gradable non-mydriatic photos by the two independent graders, perfect intergrader agreement was found (k = 1.00). The measure of intermethod agreement between NM and M method was also perfect, with a kappa value of 1.00 (grader A and grader B). Based on the results of the O method, there were no misdiagnosed cases nor with the NM-, neither with the M method. Non-mydriatic cameras could be ideal tools of an extended countrywide retinopathy screening program which may serve to reduce the high prevalence of diabetes-related blindness in Hungary.
ASJC Scopus subject areas
- Health Informatics