TrialsgrambetaRCTs are beautiful

ACCESS

2024

AI for Children's diabetiC Eye ExamS — Autonomous Artificial Intelligence Increases Screening and Follow-up for Diabetic Retinopathy in Youth

In 164 youths with diabetes at a paediatric diabetes centre, an autonomous AI diabetic eye examination at the point of care increased diabetic eye exam completion within 6 months compared with scripted referral to an eye care provider with education.

Result

Diabetic eye exam completion rate within 6 months
Absolute difference
78 points
NNT
2

Absolute difference and NNT are derived from the event rates.

Also reported: Among 25/81 intervention participants with an abnormal result, 64% (16/25) completed follow-through with an eye care provider vs 22% in the control arm (P<0.001).

Original publication
Subtopic
Diabetes care
Design
Parallel-group RCT, single academic paediatric diabetes centre
Relevance
Important: one of several pillars.
Source
Wolf RM, et al. Nat Commun. 2024;15(1):421.doi 10.1038/s41467-023-44676-z

Discussion & critique

Shows autonomous AI closing a screening gap in a racially and ethnically diverse youth population where referral alone fails: the access argument for AI rather than the accuracy one. Single centre, 164 participants, and the endpoint is exam completion rather than retinopathy detected or vision preserved; because the comparator was a referral, the trial tests point-of-care delivery as much as the AI itself.