Screening of diabetic retinopathy in primary care in Ukraine

Abstract

Purpose: To optimize the detection of DRP through the use of artificial intelligence in primary healthcare. Method: Screening of diabetic retinopathy (DR) was conducted at the Municipal non-commercial enterprise "Irpin city center of primary medical and sanitary aid" of Irpin city council of Kyiv region from 02.01.2024 to 17.03.2024. Patients with diabetes mellitus (DM) type 1 or 2 and people who did not have a diagnosis of DM and were directed to screening due to the presence of risk factors (RF) - age, overweight - were involved in the screening. Screening was carried out in stationary points arranged on the basis of health care institutions. All individuals underwent color photography of the fundus with a non-mydriatic fundus camera without pupil dilation. Fundus photography of both eyes was performed. The obtained images were immediately uploaded to the cloud environment for analysis which was carried out using artificial intelligence Retina-AI CheckEye©. The presence of DR and the stage of the disease were determined. Results: 609 patients (1218 eyes) participated in the screening. There were 148 patients with diabetes mellitus, 49 of them had poor-quality photographs of both eyes. The statistics were calculated for 99 patients with diabetes mellitus. The remaining 461 patients did not have a diagnosis of diabetes mellitus and were referred for screening due to the presence of risk factors. 65 patients of them had poor-quality photographs of both eyes. The statistics were calculated for 396 patients with risk factors. Patients were referred for screening by family medicine doctors - a total of 560 people. There were 49 self-referrals for screening. Conclusion: DR is much more common (18 times) in patients with established diabetes than in patients at risk. People with risk factors for diabetes should be screened for DR. Screening has identified DR for the first time in 83% of patients with established diabetes who were unaware of their presence.