JMIR Publications

ABSTRACT

Background:

Artificial intelligence (AI) and machine learning (ML) models are frequently developed in medical research to optimize patient care, yet they remain rarely utilized in clinical practice.

Objective:

The present study aims to understand the disconnect between model development and implementation by surveying physicians of all specialties across the United States.

Methods:

A HIPAA-compliant survey was emailed to residency coordinators at ACGME-accredited residency programs to distribute among attending physicians and resident physicians affiliated with their institution. Respondents were asked to identify and quantify the extent of their training, specialization, and the type and location of their practice. Physicians were then asked follow-up questions regarding AI in their practice: whether its use is permitted, whether they would use it if made available, primary reasons for using or not using AI, elements that would encourage its use, and ethical concerns.

Results:

Of the 941 physicians who responded to the survey, 384 (40.8%) were attending physicians, and 557 (59.2%) were resident physicians. The majority (81.9%) of physicians indicated they would adopt AI in clinical practice if given the opportunity. The most cited intended uses for AI were risk stratification, image segmentation/image analysis, and disease prognosis. The most common reservation, cited by the 18.1% of physicians who indicated that they would not use AI even if it were clinically accessible, was the potential to replicate human bias.

Conclusions:

The present study emphasizes that most academic physicians within the United States are open to adopting AI in their clinical practice. For AI to become clinically relevant, however, developers and physicians must work synergistically to design models that are accurate, accessible, and intuitive while thoroughly addressing ethical concerns associated with the implementation of AI into medicine.