|

New study: AI use grows in healthcare, but patient trust lags behind

Hanken's Associate Professor Sippo Rossi standing beside a window at Hanken.
While the use of AI in healthcare continues to grow, patients’ trust has not kept pace. A new study shows that many people still prefer doctors and traditional risk assessment tools to AI-assisted recommendations.

Published in the Journal of Medical Internet Research, the study examined how patients perceive medical recommendations in varying clinical scenarios. 

“We wanted to see how the source of the recommendation affects a patient's trust and their intention to follow the recommendations, depending on whether they know the recommendation comes from a doctor alone, a doctor using a risk calculator, or a doctor using AI”, says Sippo Rossi, assistant professor at Hanken School of Economics, and one of the researchers behind the study.

The researchers examined whether differences in trust and willingness to follow recommendations were linked not only to the source of the recommendation and clinical scenario, but also to demographic characteristics. 

“As it was a US-based study, it brought its own demographic characteristics. We measured factors such as perceived discrimination in medical settings, numeracy and health literacy. People with higher levels of numeracy and health literacy were more likely to trust the recommendations and act on them, while those reporting discrimination also had lower trust, which is consistent with previous research”, says Rossi.

The study also revealed generational differences. In the US sample, participants belonging to Generation Z were less likely to follow medical recommendations than baby boomers. However, in a separate, not yet published European study, Rossi found the opposite pattern, with younger people showing greater intent to follow recommendations.

“In the US context, my assumption is that younger people may be more disillusioned with the healthcare system. In the Nordic context, younger people may simply have greater confidence in the healthcare system.”

Lower trust in depression advice

The researchers examined six clinical scenarios in the study: sepsis, lung cancer, colon cancer, RSV vaccination, heart failure and depression.

“One finding was that people trusted recommendations related to depression and heart failure less and were less likely to follow them. Depression may be seen as a more human and behavioural issue, while the heart failure scenario was framed as low-risk. Other conditions such as cancers and sepsis may be perceived as more measurable and urgent.”

Rossi says that it's important to note that the researchers did not mention generative AI specifically in the scenarios. Medical AI is often more conservative and focused on well-studied applications, such as recognising cancers from medical images. He notes that for patients to understand how AI arrives at its recommendations, more education about the technology will be needed.

“Many people may not realise that these systems are often not generative AI. There are generative AI applications in healthcare, but in many cases the technologies are completely different. Providing short explanations, bullet points and clear information about how conclusions were reached could help reduce uncertainty and anxiety.”

Unlike patients, many medical professionals seem to be quite optimistic about AI. They see it as an additional tool rather than a replacement, according to Rossi based on his work with medical practitioners and researchers.

A divided view of AI

The researchers collected the data in 2024, and Rossi says that it would be interesting to see whether people's feelings about AI have changed.

“I see increasing scepticism towards AI, especially among younger generations. Society seems divided. Some believe AI will solve major problems such as cancer, while others fear its impact on jobs and society.”

He points out that some have predicted for decades that radiologists would be replaced by AI. In reality, we are nowhere near that.

“The ideal situation is that routine cases are solved faster while doctors use their expertise and time for the more demanding cases.”

Text and photo: 
Jessica Gustafsson