Diabetic retinopathy screening is recommended for children with Type I and Type II diabetes, yet screening rates remain low. New point-of-care diabetic retinopathy screening using autonomous artificial intelligence (AI) may now help to improve these lower screening rates. Researchers recently found that this technique is effective in terms of accuracy and cost.

A recent economic evaluation looked at AI screening data within the literature from 1994 to 2019, including: out-of-pocket cost for autonomous AI screening, ophthalmology visits and DR treatment, probability of undergoing standard retinal examination, relative odds of undergoing screening and sensitivity, specificity and diagnostic capability of conventional screening and autonomous AI screening.

The investigators reported that the expected true-positive proportions for standard ophthalmologic screening were 0.006 for Type I diabetes and 0.01 for Type II; for autonomous AI, the proportions were 0.03 for Type I and 0.04 for Type II.

If screening adherence sat at 20%, the researchers estimate that using autonomous AI would result in a higher mean patient payment ($8.52 for Type I and $10.85 for Type II) than conventional screening ($7.91 for Type I and $8.20 for Type II). However, they added that autonomous AI screening was the preferred strategy once screening adherence exceeded 23%.

Wolf RM, Channa R, Abramoff MD, et al. Cost-effectiveness of autonomous point-of-care diabetic retinopathy screening for pediatric patients with diabetes. JAMA Ophthalmol. September 3, 2020. [Epub ahead of print].