Automated Diagnostic Test for Diabetic Retinopathy in Brazilian Mass Screening

Overview

In Brazil 10% of the adult population has diabetes. Of these, 39.0% are undiagnosed, at risk for developing complications such as diabetic retinopathy (DR). Due to the increasing prevalence of diabetes and high percentage of patients with uncontrolled disease, cost-effective tools are needed with focused attention on diabetes prevention and management in the current health system. The automatic retinopathy detection can enlarge the screening, reducing the workload and costs compared to manual image graders.

In the South and Central America Region, an estimated 9.4% of the adult population (20-79 years) has diabetes in 2015, and Brazil is the first country in number of people with diabetes. Of these, 39.0% are undiagnosed, at risk for developing complications such as diabetic retinopathy (DR)(1). The rising number of people with diabetes in the world has become a real challenge for the public health system to provide care for patients with DR and for people with diabetes at risk for this complication(2). A large proportion of patients with diabetes was inadequately controlled in Brazil, which may contribute to increased rates of diabetic complications (3). The detection of any degree of DR may result in improved medical monitoring and optimization of risk factors, delaying the progression of the disease(4). In Brazil, the great demand in the public service causes delay in early diagnosis, worsening health status of patients with diabetic retinopathy and increasing the cost of their treatment. Due to the increasing prevalence of diabetes and high percentage of patients with uncontrolled disease, cost-effective tools are needed with focused attention on diabetes prevention and management in the current health system. Several studies have shown that systematic screening for DR is an effective way of prevention (5). Furthermore, the automatic retinopathy detection can enlarge the screening, reducing the workload and costs compared to manual image graders(6). There are no reports on automated DR detection in Brazilian population, especially in screening campaigns with large-scale diagnosis.

Conditions

Interventions

Eligibility

Locations (1)

Outcomes