Analysis · Technology
AI Freed Specialist Capacity by Changing Who Needed a Visit
By AWEI · AI-compiled · Published · Analysis prepared · 1 source · doi.org
At an eye hospital in Bangladesh, 331 patients completed their diabetic eye screening encounters without seeing a specialist. That change helped raise completed encounters per specialist hour by about 40%. The key was allowing an AI result to change the patient's route through the clinic.
Michael D. Abràmoff and colleagues reported the B-PRODUCTIVE trial in 2023. Researchers randomly assigned 105 clinic days to an AI-guided workflow or a control workflow at Deep Eye Care Foundation in Rangpur. Three retina specialists participated, alongside 993 patients eligible for AI screening. Randomizing days supported a causal test of the workflow's effect in this setting.
Every enrolled patient underwent AI imaging, including those in the control group. On control days, everyone then saw a specialist. On intervention days, patients whose AI result indicated no referable diabetic eye disease finished without that visit. Positive results and insufficient-quality images still led to specialist examination. This design tested what happened when staff acted on AI output.
Productivity among all clinic patients with diabetes rose from 1.14 to 1.59 completed encounters per specialist hour. The measure included encounters completed through AI alone; it did not mean doctors personally examined 40% more people. The clinic had more demand than capacity and no fixed appointment slots, allowing other patients to use the freed capacity.
The limits matter: this was one hospital, three specialists and one narrowly defined diagnostic task. Eligible patients had no relevant symptoms or prior diabetic eye disease diagnosis. Scheduled treatments did not increase significantly, and the trial did not establish better long-term vision outcomes or financial savings. Two authors disclosed ties to the manufacturer, which supplied the system and support free of charge.
For AI evaluation, the useful lesson is to measure completed work across the service. Here, changing patient routing increased capacity. Whether that translates elsewhere depends on eligibility, referral needs and whether freed capacity can actually be used.
Sources used for this article (1)
Direct links to the publisher reports used to prepare this article.
- Source 1
- Autonomous artificial intelligence increases real-world specialist clinic productivity in a cluster-randomized trial ['Michael D. Abràmoff', 'Noelle Whitestone', 'Jennifer L. Patnaik', 'Emily Rich', 'Munir Ahmed', 'Lutful Husain', 'Mohammad Yeadul Hassan', 'Md. Sajidul Huq Tanjil', 'DENA WEITZMAN', 'Tinglong Dai', '
