BACKGROUND
As the availability and performance of AI-based clinical decision support (CDS) systems improve, physicians and other care providers poised to be on the front lines will be increasingly tasked with using these tools in patient care and incorporating their outputs into clinical decision-making processes. Vignette studies provide a means to explore emerging hypotheses regarding how factors may impact physician acceptance and usage of AI-based CDS tools, or to generate new hypotheses regarding how individuals may react to specific scenarios involving new applications where much is unknown. To best anticipate the decision making of physicians in clinical scenarios involving AI decision support tools, hypothesis-driven research is needed that enables scenario testing before the implementation and deployment of these tools.
OBJECTIVE
This paper describes a research protocol for an online vignette-based survey study that aims to understand predictors and causal factors of physician decision-making in the context of AI decision support tools. This paper focuses on the implementation of an original online survey that utilizes a novel sequential randomization technique.
METHODS
U.S.-based physicians who are listed in the most recent version of the American Medical Association (AMA) Physician Masterfile (PMF) will be recruited via email and mail (target n=420). Via an online survey, participants will respond to questionnaires regarding their demographics, professional experience, and attitudes toward/experience with AI/ML in medicine. They will then be randomly assigned, at three time points, to a three-part hypothetical vignette detailing a clinical scenario involving an AI decision support tool. Participants will be asked to respond to questions regarding their hypothetical decision-making as it relates to clinical risk, the amount of information provided about the AI, and the AI result, as described in the vignette.
RESULTS
The study is currently in progress and data collection is anticipated to be completed in 2024.
CONCLUSIONS
The online vignette study will provide information on how physicians may react to hypothetical scenarios that are based on emerging applications of AI/ML in health care settings. Our rationale for focusing on physicians is that these individuals will be faced with incorporating AI/ML algorithms to inform the evaluation and management of patients. This study will generate a better understanding of physician decision-making.