‘Inevitable’ That AI Will Be Integrated Into Medical Care

Medscape’s 2024 survey of over 5000 practicing clinicians from six European countries has showed there is more enthusiasm than apprehension about the use of artificial intelligence (AI) in medicine. However, many doctors have reasonable concerns, and the survey reveals a worrying knowledge gap among physicians.

Experts predict that it is inevitable that AI will become integrated into mainstream healthcare within 5-10 years, and both physicians and experts agree that regulation and oversight are needed now.

Physicians’ balance of views is “notable” and “both practical and positive”, said Eric Sutherland, senior health economist for the Organisation for Economic Co-operation and Development (OECD)’s AI in Health expert group, speaking to Medscape Medical News. He said that the findings of Medscape’s European Doctors and AI report echo a survey of medical associations that the OECD will publish this month.

“There is a degree of caution, while there is also enthusiasm for the potential of AI to improve health outcomes,” he said. However, “the reality is that the digital age has brought great promise for doctors and health providers, but those promises have failed to live up to the hype.” For physicians, “there are real reasons to be apprehensive,” he said, including concerns about potential job loss, impacts on the nature of the job, effects on the provider-patient relationship, and concerns about the quality of AI solutions.

Physicians’ Concerns About AI

Some survey respondents were worried about the effects of AI on physician autonomy or replacing clinical judgment and expertise and about patients believing AI-generated misinformation or using AI for self-diagnosis.

While the consensus was that AI use by doctors in healthcare will reduce the risks for misdiagnosis, medical error, and malpractice, Sutherland highlighted the “not insignificant portion” predicting increased risks.

“It is reasonable to be concerned,” he said, otherwise we are “blindly trusting a technology” when we should be taking active steps to understand and address these risks.

A particular concern is AI being trained on poor-quality datasets that may have limited applicability. However, he believes that good quality data applied by medical professionals trained in using AI solutions “can reduce medical error and improve patient outcomes.”

“I am more concerned about the risks of inaction with respect to AI,” he said

Knowledge Deficit

Among survey respondents, only around 1 in 10 doctors feel “very knowledgeable” about AI, with many saying they are “not at all knowledgeable.” Fully 95% of all doctors thought it “very” or “somewhat” important to become educated about healthcare AI applications.

Low levels of clinician knowledge were “worrying and challenging,” while recognition of the importance of training is encouraging, said Maurício Alves, hospital administrator at the National School of Public Health at NOVA University Lisbon, Lisbon, Portugal. He described the survey as “relevant and necessary to put into perspective the level of knowledge, skills, and involvement of healthcare professionals.”

Managers and administrators understand the inevitability and potential advantages of AI and similarly see their own training as essential, he told Medscape Medical News.

Doctors Must Engage With AI

“Education is critical,” agreed Luigi De Angelis, MD, president of the Italian Society for Artificial Intelligence in Medicine. “Without deep engagement and understanding, doctors risk taking a passive role in AI integration, leading to suboptimal solutions.”

Although around 40% of respondents thought that healthcare should be an “early adopter” of AI, De Angelis warned against the current “wait-and-see” stance expressed by a majority of clinicians in all countries surveyed.

Future acceptance of AI among physicians will depend on their involvement in actively shaping these innovations, he said, otherwise “these technologies will evolve without essential clinical insight.”

AI Integration Inevitable

Over the next 5-10 years, many AI solutions “will become seamlessly integrated into everyday clinical activities across all medical disciplines,” De Angelis predicted. “Institutions, regulatory bodies, and medical associations should collaborate now to establish guidelines and best practices.”

Sutherland has made the point that AI is likely to reduce significantly the traditional 17-year gap from publication of new medical information to integration into medical practice. It will be “fundamentally transformational to how health is delivered, received, maintained, and improved,” he said.

“Conscious decisions” are needed about where AI solutions are most beneficial. Ultimately “it is up to us all — health professionals, policy makers, technologists, and the public — to decide what future we want and take the steps we can to get there. And we must do that together.”

Alves, who describes himself as a promoter of the possibilities of digital transformation and innovation in healthcare, agreed that in the near future, the use of healthcare AI will become commonplace. “Let us not be under any illusions: AI is here to stay.” He added that the “reality check” provided by the survey is crucial, and clinicians and providers must be “the main agents of transformation.” 

Sheena Meredith, MBBS, MPhil, is an established medical writer, editor, and consultant in healthcare communications, with extensive experience writing for medical professionals and the general public. She is qualified in medicine and in law and medical ethics.