Experimental Data on AI-Assisted Medical Decision-Making

Ziyi Wang from Wuhan University conducted a lab-in-the-field experiment evaluating AI's role in mitigating medical overtreatment. The study used a three-by-two factorial design, manipulating monetary incentives and AI availability among medical students performing a prescription task. Results show AI reduced overtreatment rates by up to 62% and improved diagnostic accuracy by 17% to 37%, depending on the incentive scheme.

Use Cases

• Analyzing the effect of incentive schemes (Flat, Progressive, Regressive) on AI adoption
• Evaluating AI's impact on overtreatment rates and diagnostic accuracy
• Studying behavioral modifications in medical decisions based on AI input
• Quantifying monetary and non-monetary incentives for overtreatment for policy design

Strengths

• Experimental design is detailed, employing a three-by-two factorial structure
• Results include specific quantified impacts: overtreatment reduction up to 62% and accuracy improvements of 17-37%

Limitations

• Dataset metadata is sparse; column names, row counts, and file formats are unknown
• License is closed, restricting access and reuse
• Last updated date is unknown, affecting assessment of freshness

Provenance

Source

Ziyi Wang, Wuhan University

Collection Method

Lab-in-the-field experiment at a medical school using a novel medical prescription task