Screening for diabetic retinopathy is a crucial recommendation for children diagnosed with both type 1 diabetes (T1D) and type 2 diabetes (T2D). Despite its importance, current screening rates remain suboptimal. The advent of point-of-care diabetic retinopathy screening, leveraging autonomous artificial intelligence (AI), offers a promising solution by providing immediate results directly within the clinic. However, understanding the cost-effectiveness of this innovative approach in comparison to traditional eye care professional (ECP) examinations is essential for informed healthcare decisions.
A recent economic evaluation rigorously assessed the cost-effectiveness of utilizing AI-driven Point Of Care Diagnosis For Diabetes Retinopathy detection and subsequent treatment in pediatric diabetic patients. The study compared this method against standard screening protocols conducted by an eye care professional. This analysis drew upon parameter estimates from extensive literature spanning 1994 to 2019, with assessments conducted between March 2019 and January 2020. Key parameters included the out-of-pocket expenses associated with autonomous AI screening, ophthalmology appointments, and diabetic retinopathy treatment. It also considered the probability of patients undergoing standard retinal examinations, the relative likelihood of screening participation, and the diagnostic accuracy (sensitivity, specificity, and diagnosability) of both ECP screenings and autonomous AI screenings.
The core metrics evaluated were patient-borne costs or savings linked to diabetic retinopathy screening examinations. Cost-effectiveness was determined by analyzing the costs or savings relative to the number of true-positive cases accurately identified by each screening method. The study revealed significant differences in true-positive detection rates. Standard ECP screening yielded expected true-positive proportions of 0.006 for T1D and 0.01 for T2D. In contrast, autonomous AI screening demonstrated notably higher expected true-positive proportions, at 0.03 for T1D and 0.04 for T2D. Interestingly, under a base case scenario with a 20% adherence rate to screening, autonomous AI initially indicated a slightly higher mean patient payment ($8.52 for T1D and $10.85 for T2D) compared to conventional ECP screening ($7.91 for T1D and $8.20 for T2D). However, the findings shifted dramatically with increased adherence. Autonomous AI screening emerged as the preferred, more cost-effective strategy when patient adherence to diabetic retinopathy screening reached or exceeded 23%.
In conclusion, this economic evaluation strongly suggests that point-of-care diabetic retinopathy screening systems powered by autonomous AI represent an effective and potentially cost-saving approach for children with diabetes and their families, particularly when recommended screening adherence rates are met or surpassed. This highlights the potential of AI-driven diagnostics to improve access and efficiency in diabetic retinopathy screening within pediatric populations.
