For how we live.
An autonomous AI for diabetic retinopathy hit 87.2% sensitivity in primary care. A smartphone acuity test agreed with a reference chart to within 0.07 logMAR. Both are triage, not diagnosis.
Systems that give a diagnosis with no clinician in the loop post high sensitivity in trials. In real-world use, one in six images was too poor to grade — and the strongest numbers come from company-run studies.
Matched against DPP-4 inhibitors, GLP-1 users had slightly more diabetic retinopathy and more retinopathy procedures. Vision impairment and blindness did not differ — and the coding for those is poor.
A supplement approved on 26 August adds SURPASS-CVOT to the prescribing information and rewrites the eye warning so it is addressed to patients with a history of retinopathy rather than to patients with diabetes.