One of the most serious complications of diabetes is damage to the back of the eye resulting in loss of vision. This is called diabetic retinopathy. Approximately 35% of people with diabetes are at risk of diabetic retinopathy. More than 10% are at risk of more severe vision-threatening disease.
Not enough people with diabetes have their eyes checked
Twice as many patients with diabetes report fear of vision loss over any other complication. However, recommended annual vision screening is rarely completed. In fact, 21% of patients with diabetes around the world have never completed a screening. In the U.S., only 60% of the patients receive yearly dilated examinations. These rates are even lower among low-income patients of minority race and ethnicity. Early detection and intervention are important for people with diabetes because they can prevent blindness.
Artificial intelligence reduces need to dilate pupils
In a recent study, the use of artificial intelligence (AI) to detect eye disease due to diabetes was assessed in 942 adults across 15 primary care and eye care facilities. The participants already had diabetes-related eye disease. The researchers compared the AI system with a gold-standard of retinal photographs assessed by experts. They found that the level of agreement between the AI system and the experts was high. They also found that it was not necessary to dilate the pupils to get a good view of the back of the eye for 87% of eyes.
These results suggest that an AI system can accurately detect eye disease due to diabetes. This could improve vision without the need for doctor oversight or need for dilation in most people. This system has the potential to increase screening for eye disease due to diabetes especially in underserved communities.