Stakeholder Perspectives on the Use of Artificial Intelligence in Medicine

Author(s)

Crossnohere N1, Wang G1, Bose-Brill S1, Bridges J2
1The Ohio State University, Columbus, OH, USA, 2Ohio State University College of Medicine, Columbus, OH, USA

OBJECTIVES: Artificial intelligence and machine learning (AI-ML) are fundamentally transforming the delivery of healthcare and the practice of medicine. The uptake of AI-ML however, is hindered by translational concerns. We sought to assess community and professional perspectives regarding barriers on the development, application, and oversight of AI-ML in medicine.

METHODS: We conducted 15 semi-structured interviews at a large, mid-western academic medical. Interviewees included patient advocates (n=2), clinicians (n=4), health system administrators (n=2), AI-ML tool developers and investors (n=2), and researchers (n=5). Interview transcriptions were analyzed using thematic analysis to identify overarching barriers on the use of AI-ML in medicine.

RESULTS: Three themes regarding barriers to the use of AI-ML emerged from interviews. The first theme described challenges in defining and explaining AI. Stakeholders not only had different understandings of what AI-ML was, may (and particularly patient advocates) also experienced difficulty in describing the concept either within or outside the field of medicine. The second barrier was biases in data, which encompassed concerns raised by stakeholders regarding racial and other biases present in the datasets used to generate AI-ML tools. The third barrier was the perceived lack of transdisciplinary collaboration on the development of AI-ML tools. Stakeholder expressed that tool developers must “wear multiple hats” and be able to have both technical expertise in the development of AI-ML tools, while also understanding the relevant clinical context.

CONCLUSIONS: AI-ML tools have nearly limitless potential in medicine, but stakeholders expressed that the uptake of AI-ML was hindered by the lack of a shared understanding of AI-ML is, concerns regarding potential data bias, and lack of transdisciplinary development of AI-ML tools. Engaging both community and professional perspectives on the topic of AI-ML is essential to understanding and in turn ameliorating the full-spectrum of barriers the technology faces.

Conference/Value in Health Info

2022-05, ISPOR 2022, Washington, DC, USA

Value in Health, Volume 25, Issue 6, S1 (June 2022)

Code

HSD69

Topic

Methodological & Statistical Research, Patient-Centered Research

Topic Subcategory

Artificial Intelligence, Machine Learning, Predictive Analytics, Patient Engagement

Disease

No Additional Disease & Conditions/Specialized Treatment Areas

Explore Related HEOR by Topic


Your browser is out-of-date

ISPOR recommends that you update your browser for more security, speed and the best experience on ispor.org. Update my browser now

×