Overcoming Medical Overuse with AI Assistance: An Experimental Investigation
How to cite
Wang, Ziyi; Xue, Lian; Wei, Lijia. 2024. "Overcoming Medical Overuse with AI Assistance: An Experimental Investigation." AEA RCT Registry. AEARCTR-0013376. https://doi.org/10.1257/rct.13376-1.0
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
This study explores the potential of Artificial Intelligence (AI) to mitigate medical overuse by examining the decision-making processes of physicians in a controlled experimental setting. Medical overuse, characterized by unnecessary interventions that may harm patients and inflate healthcare costs, is a significant issue in both developed and developing healthcare systems. By integrating AI tools that offer evidence-based recommendations, this research aims to understand the extent to which AI can influence physician behavior and reduce overtreatment.
Our experimental design involves a lab-in-the-field setting within a medical school, where we manipulate both monetary incentives and the availability of AI assistance across different groups of medical students. We employ a two-by-two factorial design to assess how these variables impact the rates of overtreatment and the accuracy of clinical decisions. The study’s primary outcomes include the frequency of overtreatment, diagnosis accuracy rates, and the adoption rate of AI advice.
The findings are expected to contribute to the broader discourse on how AI can be lever-aged to enhance clinical decision-making and reduce inefficiencies in medical practices. By providing empirical evidence on the effects of AI interventions in reducing overtreatment, this study could inform policymakers and healthcare administrators about the merits of incorporating AI technologies in healthcare settings, potentially leading to improved healthcare efficiency and patient safety.
Experimental design
We conducted a lab-in-the-field experiment at a medical school affiliated with Central South Hospital, utilizing a two-by-three factorial design. The participants, referred to as prospective physicians, were recruited from this institution, where they were trained to become medical doctors.
Incentive Schemes (Between-Subject): The experiment varied the incentive structures for participants when making choices in a multiple-choice medicine prescription task. Three types of incentives were tested: flat, progressive, and regressive, affecting how participants were compensated for their choices.
AI Assistance (Within-Subject): We also varied the decision-making process by introducing an AI assistance feature. Participants made initial choices independently and could then opt to view AI-generated recommendations before finalizing their decisions.
Control Variables: To account for individual fixed effects and other potential confounders affecting the propensity for overtreatment, we implemented pre- and post-experiment assessments. These assessments collected data on doctors’ professional abilities, cognitive reflection (CRT), IQ, algorithm literacy, trust in algorithms, awareness of algorithms, and perceptions of algorithm fairness.
Design details
Cited by
No citation data available for this registration yet.