The number of people given AI-assisted screening tests to detect diabetic retinopathy is considerably low despite the potential benefits of the technology, research has demonstrated.
Academics from Stanford University and New York University School of Medicine have found that ophthalmologists rarely use AI-assisted screening tests to identify the eye condition.
As part of the study, the research team assessed the health data of millions of people living with diabetes and analysed what treatments they were given.
- AI-generated alerts reduce hospital patients’ risk of dying, study shows
- AI-powered eye exams improve screening rates among young people with diabetes
- Insights into inflammation and heart disease offer hope for better treatments
Diabetic retinopathy is a medical condition in which damage occurs to the retina due to diabetes.
It is a leading cause of blindness in developed countries and one of the lead causes of sight loss in the world, even though there are many new therapies and improved treatments for helping people live with diabetes, including laser therapy, surgery and anti-VEGF drugs.
Previous studies have revealed that less than two-thirds of people living with diabetes attend a yearly eye exam.
Traditionally, ophthalmologists conduct a dilated fundus exam to detect for diabetic retinopathy by taking pictures of the retina.
However, scientists have spotted that AI technology can spot the signs of diabetic retinopathy just as good, or in some cases better, than humans.
In 2021, the FDA approved AI technology to detect diabetic retinopathy because of the promise it showed.
- Breakthrough in restoring vision using stem cells
- Type 2 diabetes study makes flagship dataset available to researchers worldwide
- Study to detect type 2 diabetes from voice launches
In order to assess if ophthalmologists have started using AI technology to spot the signs of the condition, the research team looked at the TriNetX federated database – a document which includes data from 62 health care organisations in the US.
By looking at this database, the scientists were able to examine the health outcomes of 4,959,809 people living with diabetes.
They found that that 0.09% of eye exams in the US used AI-assisted screening tests, while usual methods increased by 185.4% during the same period.
This work is published in the journal JAMA Ophthalmology.
