A study of thousands of patients across dozens of countries suggests general support for the use of AI in medicine, although trust is a key factor and attitudes differ according to sex, wellness, and technological literacy.
The findings from six continents stress the importance of tailoring the implementation of AI to the unique characteristics of local patients and populations.
The research, in JAMA Network Open, revealed that women, those in poorer health, and people less technologically adept were generally less supportive of AI in healthcare.
Overall, patients had a preference towards AI systems that were explainable and for decision making to be led by physicians.
“This multinational cross-sectional study represents, to our knowledge, the most extensive and comprehensive survey to date of patient attitudes toward AI in health care worldwide,” asserted researcher Felix Busch, PhD, from the Technical University of Munich, and co-workers.
“With 13,806 participants from 43 countries, our findings provide a multifaceted understanding of patients’ preferences, trust, and concerns about AI in health care.
“The results illustrated a nuanced landscape of attitudes, with most patients expressing support for the use of AI in health care while also articulating concerns about its implementation.”
AI technologies offer the promise of tackling pressing health care challenges including staff shortages, high administrative costs, and economic constraints, the authors observe.
In clinical practice, its applications range from assisting with image-based diagnoses to personalizing treatment strategies and predicting risk factors and therapy responses.
Beyond direct patient care, AI aids the discovery and development of drugs, while streamlining administrative tasks such as data extraction, curation, and report structuring.
Noting the importance of patient acceptance for its sustainable adoption, the team conducted a large-scale survey of hospital patients to investigate trust, concerns, and preferences toward AI use in health care and diagnostics.
The survey was administered to a sample of patients at 74 COMFORT network hospitals in 43 countries, across diverse sociodemographic circumstances.
Overall, the study included 8951 (64.8%) patients in the Global North and 4855 (35.2%) in the Global South, according to the definition by the United Nations Finance Center for South-South Cooperation.
There was a predominantly favorable view of AI in health care, with 57.6% of respondents expressing a positive attitude. However, attitudes did differ according to demographic characteristics, health status, and technological literacy.
Women had fewer positive attitudes towards the use of AI in medicine than men, at 55.6% vs 59.1%, respectively. Patients with poorer health status also showed fewer positive attitudes toward AI use in medicine than those with very good health, with rather negative views recorded in a corresponding 29.2% vs 5.3%.
By contrast, greater AI knowledge and more frequent use of technological devices were associated with more positive attitudes relating to trust in AI.
Overall, fewer than half of the participants expressed positive attitudes regarding all items relating to trust in AI. The lowest level of trust was found for the accuracy of AI in providing information regarding treatment responses, with just 41.8% trusting AI.
More than seven in 10 patients preferred explainable AI and physician-led decision making, even if it meant slightly compromised accuracy.
The authors noted that multiple concerns were voiced despite the generally favorable views.
“More than half of patients indicated apprehensions about data security, reduced physician-patient interaction, and potential increases in health care costs,” they reported.
“These findings underscore the need for a balanced approach to AI implementation in health care, one that addresses patient concerns while leveraging the potential benefits of AI technology.”
