The explainable AI dilemma under knowledge imbalance in specialist AI for glaucoma referrals in primary care
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
View abstract on PubMed
Summary
This summary is machine-generated.Artificial Intelligence (AI) shows promise in identifying glaucoma patients needing urgent referrals. However, AI explanations did not improve human-AI team accuracy and introduced bias, highlighting challenges in integrating AI with clinical expertise.
Area Of Science
- Ophthalmology
- Artificial Intelligence in Healthcare
- Clinical Decision Support
Background
- Glaucoma patient referrals by primary eye care providers rely on clinical experience.
- Specialized Artificial Intelligence (AI) models for referrals are trained on clinical data but may not generalize to real-world practice.
- A knowledge gap exists between AI capabilities and practical clinical application.
Purpose Of The Study
- To investigate the effectiveness of AI explanations in assisting primary eye care providers with glaucoma referrals.
- To assess whether AI-human collaboration, aided by explanations, can improve referral accuracy compared to human judgment alone.
- To understand the impact of AI explanations on provider trust and decision-making biases.
Main Methods
- Development of AI models to identify urgent glaucoma referrals from routine eye care data.
- Creation of both intrinsic and post-hoc AI explanation methods.
- A user study involving 87 optometrists evaluating AI predictions with and without explanations.
Main Results
- Human-AI teams achieved 60% accuracy, outperforming humans alone (51%) but underperforming AI alone (80%).
- AI explanations did not enhance performance and introduced uncertainty, increasing over-reliance on incorrect AI recommendations.
- Both intrinsic and post-hoc explanations contributed to anchoring bias, causing participants to align more closely with AI referrals.
Conclusions
- Current AI explanation methods do not effectively support primary eye care providers in glaucoma referrals.
- Challenges remain in designing AI support systems that enhance human-AI collaboration without compromising clinical judgment or introducing bias.
- Further research is needed to develop effective mechanisms for AI integration that preserve human agency and surpass AI performance.
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
Related Concept Videos
Glaucoma is an eye condition characterized by increased intraocular pressure that damages the retina and optic nerve, leading to irreversible blindness if left untreated. The human eye has various components, including the cornea, iris, pupil, lens, and optic nerve. Aqueous humor is secreted by the epithelium of the ciliary body in the posterior chamber and flows through the trabecular meshwork and canal of Schlemm, maintaining normal intraocular pressure. The trabecular meshwork and the canal...
Angle-closure glaucoma, or closed-angle glaucoma, is an eye condition where the iris bulges out and blocks the iridocorneal angle, resulting in a buildup of aqueous humor and increased intraocular pressure. Immediate medical attention is necessary due to the sudden onset of symptoms. The treatment for angle-closure glaucoma includes short-term and long-term approaches. Short-term treatment involves using eye drops like pilocarpine to lower intraocular pressure by increasing aqueous humor...
In open-angle glaucoma, the iridocorneal angle remains open, but the trabecular meshwork becomes stiff, slowing down the outflow of aqueous humor. This causes a buildup of aqueous humor in the anterior chamber, leading to a sudden increase in intraocular pressure. The treatment for open-angle glaucoma focuses on reducing the elevated intraocular pressure by either decreasing the secretion of aqueous humor or increasing its outflow.
Drugs such as carbonic anhydrase inhibitors, α2- and...
Blinding is a commonly used method of not telling participants which treatment a subject is receiving. Blinding is a critical part of a randomized control trial or RCT. It reduces the bias that affects the results. In an RCT, blinding is used in the form of a placebo. A placebo effect occurs when untreated subjects falsely believe they have received the treatment and report improved symptoms. A placebo or a dummy treatment is administered to subjects to negate the bias caused by such an effect.