Overcoming Medical Overuse with AI Assistance: An Experimental Investigation

SSRN Electronic Journal, Journal Year: 2024, Volume and Issue: unknown

Published: Jan. 1, 2024

This study evaluates the effectiveness of Artificial Intelligence (AI) in mitigating medical overtreatment, a significant issue characterized by unnecessary interventions that inflate healthcare costs and pose risks to patients. We conducted lab-in-the-field experiment at school, utilizing novel prescription task, manipulating monetary incentives availability AI assistance among students using three-by-two factorial design. tested three incentive schemes: Flat (constant pay regardless treatment quantity), Progressive (pay increases with number treatments), Regressive (penalties for overtreatment) assess their influence on adoption assistance. Our findings demonstrate significantly reduced overtreatment rates—by up 62% conditions where (prospective) physician patient interests were most aligned. Diagnostic accuracy improved 17% 37%, depending scheme. Adoption advice was high, approximately half participants modifying decisions based input across all settings. For policy implications, we quantified (57%) non-monetary (43%) highlighted AI's potential mitigate enhance social welfare. results provide valuable insights administrators considering integration into systems.

Language: Английский

An Evaluation on the Potential of Large Language Models for Use in Trauma Triage

Emergency Care and Medicine, Journal Year: 2024, Volume and Issue: 1(4), P. 350 - 367

Published: Oct. 12, 2024

Large Language Models (LLMs) are becoming increasingly adopted in various industries worldwide. In particular, there is emerging research assessing the reliability of LLMs, such as ChatGPT, performing triaging decisions emergent settings. A unique aspect emergency process trauma triaging. This requires judicious consideration mechanism injury, severity patient stability, logistics location and type transport order to ensure patients have access appropriate timely care. Current issues overtriage undertriage highlight potential for use LLMs a complementary tool assist more accurate patient. Despite this, remains gap literature surrounding utility process. narrative review explores current evidence implementation Overall, highlights multifaceted applications especially settings, albeit with clear limitations ethical considerations, artificial hallucinations, biased outputs data privacy issues. There room rigorous into refining consistency capabilities ensuring their effective integration real-world improve outcomes resource utilisation.

Language: Английский

Overcoming Medical Overuse with Ai Assistance: An Experimental Investigation

Published: Jan. 1, 2024

Language: Английский