Artificial intelligence and teleophthalmology are emerging as scalable ways to improve screening for diabetic retinal disease, but their economic value in U.S. health systems has been unclear. A recent study published in Nature sought to explore the value of different retinal screening options in different care settings.
Researchers built a five-year Markov model to compare AI-based retinal screening, teleophthalmology and exams by eyecare professionals. They looked at two settings: a small primary care network and a large integrated health system. They also tested different willingness-to-pay thresholds, ranging from $400 to $1,500 per patient over five years in the smaller network and $800 to $3,500 in the larger system.
The Pay Off
AI-based strategies completed about three times more screenings, found 3.6 to 3.8 times more true positives and got 7.5 to 8.0 times more patients started on treatment than eyecare professionals alone. Teleophthalmology was moderately effective.
Patient volume and willingness to pay impacted which option was the most cost-effective. In the primary care network, handheld camera teleretinal screening was favored at lower volumes, while handheld AI or stationary AI became better choices as volume rose. In the integrated health system, eyecare professional screening remained the best option at a lower willingness-to-pay level. At the higher level, handheld teleretinal screening or stationary AI was preferred depending on volume.
For eyecare professionals, the takeaway is that AI-based screening—especially handheld AI—can be cost-effective when patient volume is high and the system is willing to pay more per patient. The study suggests its value comes from real-time diagnostic feedback, which can lead to earlier detection and lower costs tied to more advanced diabetic retinal disease.
Read more of the latest AI in Eye Care news here
Author
