5 key takeaways from a patient survey about AI in radiology
What do patients think about the use of artificial intelligence (AI) in radiology? A new study published in the Journal of the American College of Radiology addressed that very question.
“At present, it is still unknown how patients view the developments of AI in radiology in terms of awareness of this topic, uncertainties, and expectations,” wrote author Marieke Haan, PhD, of the department of sociology at the University of Groningen in the Netherlands, and colleagues. “This knowledge is crucial to define preconditions for the development of AI systems for different clinical purposes and how they should be used in routine radiology practice. It may also provide insight into whether there is a need for patient education on this topic.”
The authors surveyed 20 patients with a mean age of 63.8 years old who were scheduled for an outpatient CT scan. These are five important findings from that survey:
1. Patients appear to prefer humans over AI
Respondents indicated that their first preference is for humans to read radiology results, but they can possibly be swayed when told how effective AI can be.
“When scientific research shows that computers are indeed superior to humans, most patients indicate that they would rather let the computer do the work than the radiologist,” the authors wrote.
Patients also indicated that they know human radiologists “have flaws and may therefore not always perform optimally when doing their job.” On the other hand, they “value the experience of the radiologist” and “are unsure” about the quality of work performed by AI.
A key finding in this area was that patients don’t want to have their results exclusively read by algorithms.
“They indicate that AI in radiology is a nice instrument to double-check the conclusions of the radiologist, but only using AI could result in restricted views with wrong diagnoses,” the authors wrote.
2. Patients are unsure how AI is implemented in radiology
Previous research has found the general public doesn’t fully grasp what, exactly, radiologists are or what they do. Haan et al. noted that there is also uncertainty about how AI would be used in the interpretation of medical images.
“Patients express uncertainties about who is involved in the procedure at what stage when implementing AI,” the authors wrote. “To them, the roles of the radiologist, the radiology technicians operating the scan equipment, and the referring physician in relationship to the use of AI are not completely clear. Patients also say that it is unclear to them how AI would affect the procedure of scanning and receiving results.”
3. Patients think AI could lead to faster interpretation times, lower costs
Survey respondents said they see efficiency as a clear benefit of AI being implemented in radiology workflow, indicating that it should lead to faster interpretation times, reduced waiting times and lower overall costs.
4. Patients fear AI will reduce interaction with healthcare providers
“Patients express their concerns about depersonalized procedures in which patients become numbers,” the authors wrote, noting that “personal interaction when receiving information about the results of a scan is important to them.”
5. Patients don’t appear fully comfortable with AI being responsible for their healthcare
The lack of accountability associated with AI was a significant issue for survey respondents.
“Patients note that radiologists can be held accountable for their mistakes, and they wonder who can be held responsible for errors made by computers,” the authors wrote. “Some patients report that a computer is just a ‘giant calculator’ or a ‘dead thing’ and that humans will always be responsible.”
The authors noted that this makes it clear how important it is for the ethics of AI to be a major point of discussion in the years ahead.
Michael has more than 18 years of experience as a professional writer and editor. He has written at length about cardiology, radiology, artificial intelligence and other key healthcare topics.