Model Finds AI Diagnosis of Pediatric Diabetic Retinopathy Effective, Cost-Saving

September 9, 2020

Point-of-care diabetic retinopathy screening using autonomous artificial intelligence (AI) systems is effective and cost-saving for children with diabetes and their caregivers.

At recommended adherence rates, point-of-care diabetic retinopathy (DR) screening using autonomous artificial intelligence (AI) systems is effective and cost-saving for children with diabetes and their caregivers, according to research published in JAMA Ophthalmology.

Among youths with type 1 diabetes (T1D) and type 2 diabetes (T2D), although the prevalence of DR ranges from 4% to 13%, the risk of developing the conditions remains high.

“DR is present in up to 50% of patients with T1D 28 years or more after diagnosis and already present at the time of diagnosis in 12% to 19% of patients with T2D,” the authors wrote.

Both the American Diabetes Association and the American Academy of Ophthalmology recommend yearly screening for DR; however, adherence to this guidance remains low. The FDA recently approved the first autonomous AI diagnostic system to detect DR in adults. The process consists of an operator guided by AI, who takes retinal images via a nonmydriatic fundus camera. Images are then assessed in real time at the point of care (POC) to determines DR status. 

To estimate cost savings of using autonomous AI DR screening in youths compared with a clinic-based ophthalmoscopy examination by an eye care professional (ECP), the researchers developed a patient-oriented decision model. “Using decision analysis, we modeled the cost-effectiveness of detecting and treating DR and its sequelae among children with diabetes,” the investigators wrote.

They defined effectiveness as the proportion of DR cases diagnosed while cost was defined as patient and family out-of-pocket cost. The model also assumed a child (under age 21) had previously diagnosed T1D or T2D and received regular care but had no history of known eye disease. As screening is recommended on an annual basis, the interval of focus was 1 year.

“The focus of the model was the diagnosis of clinically actionable, referable DR,” the authors wrote. “The outcome is expressed as probability of having DR (true positive) given a screening strategy.”

Based on literature, the researchers estimated the out-of-pocket cost for an ECP visit ranged from $35 to $500 depending on the copayment and insurance deductible. The base case cost for AI screening was $0, but a high cost of $100 was included for AI screening, as the researchers anticipated additional charges depending on insurance coverage.

Depending on the patient’s age and severity of the disease, the medical cost of treating DR can range between $0 and $10,000. “In the baseline scenario, we estimated that 8% of patients would possibly reach a maximum deductible of $600 (maximum of $10,000 in sensitivity analysis),” the researchers said, while “in the model, we used a base case value of $94, which was weighted by the probabilities of DR severity and the consequent likelihood of treatment.”

Analyses revealed:

  • The expected true-positive proportions for standard ophthalmologic screening by an ECP were 0.006 for T1D and 0.01 for T2D
  • The expected true-positive proportions for autonomous AI were 0.03 for T1D and 0.04 for T2D
  • Base case scenario of 20% adherence estimated that use of autonomous AI would result in a higher mean patient payment ($8.52 for T1D and $10.85 for T2D) than conventional ECP screening ($7.91 and $8.20, respectively)
  • When at least 23% of patients adhered to DR screening guidelines, autonomous AI screening was associated with cost savings and detected more cases (difference in true-positive rate of 0.027 in favor of autonomous AI)
  • As adherence increased, more patients underwent ECP screening, and the mean patient payment for ECP testing continued to increase. Thus, in the low-adherence baseline scenario, ECP screening is the preferred option.

Because the analysis is a patient-perspective model, it did not address system, societal, or additional third-party costs. However, “from the patient perspective, POC autonomous AI screening is advantageous because it can be performed in the diabetes clinic setting, saving the patient and their caregiver another physician visit in most cases,” the authors wrote.

In addition, lower out-of-pocket costs and reduced cost-sharing tend to improve adherence and outcomes in diabetes care.

“Future cost-effectiveness and cost-savings analysis models should analyze the use of these innovative AI systems from the health care system perspective,” the researchers concluded.

Reference

Wolf RM, Channa R, Abramoff MD, Lehmann HP. Cost-effectiveness of autonomous point-of-care diabetic retinopathy screening for pediatric patients with diabetes. JAMA Opthalmol. Published online September 3, 2020. doi:10.1001/jamaophthalmol.2020.3190