Heart surgeons see more potential in AI than colleagues from other fields
Cardiac surgeons may be more open to the use of artificial intelligence (AI) than some of their colleagues, according to new findings published in The Annals of Thoracic Surgery.[1]
“Neural networks and machine learning have been present in medicine for some time, but now, with the introduction of natural language processing, complex large language models have been fashioned,” wrote first author Joseph J. Platz, MD, an assistant professor in the division of cardiothoracic surgery at Saint Louis University, and colleagues. “This has made the application of AI technology available to non-experts, including a wide spectrum of healthcare personnel such as patients, nurses, administrators and physicians. In cardiothoracic surgery, AI's role is poised to be transformative not just in administrative tasks, but also in the clinical realm, potentially enhancing diagnostic accuracy, patient screening, tailored treatment, surgical precision and post-operative care, including predictive outcome modeling.”
Platz et al. surveyed more than 400 cardiothoracic surgeons about the use of AI in cardiothoracic surgery. All participants were members of the Asian Society for Cardiovascular and Thoracic Surgery, European Association for Cardio-Thoracic Surgery, European Society of Thoracic Surgeons and Society of Thoracic Surgeons. Responses were received between May and June 2024. More than 81% of responses were from men.
While 50.5% of respondents were adult cardiac surgeons, 41.5% were general thoracic surgeons, 5.4% were congenital surgeons and 2.6% landed in the “other or unknown” category. Also, 60.6% of respondents are based in North America, 22.4% are based in Europe, 10.6% are based in Asia and 6.4% landed in the “other or unknown” category.
Overall, a majority of surgeons said they see AI playing a significant role in image interpretation (73%) and patient screening (68%) in the next five years. When it came to intraoperative decision making and more technical tasks, however, surgeons remain less convinced.
One key takeaway from these survey responses was the fact that cardiac surgeons appear to view AI “more favorably” than general thoracic surgeons when it comes to preoperative evaluations and patient selection. Cardiac surgeons also appear to be more confident in AI’s ability to provide value when it comes to preoperative decision making and postoperative patient management.
The survey also found that surgeons are unsure about the use of AI for certain clinical tasks. When asked if AI will be able to make “accurate independent clinical management decisions you rely on,” just 33.2% of respondents said they find that to be “very likely” or “extremely likely.”
When it comes to administrative tasks, however, surgeons seem to see a lot of potential in AI’s ability to make an impact on patient care. Coding/billing and clinical note writing are two areas where a majority of surgeons said AI will provide value. This enthusiasm was more prevalent among cardiac surgeons than other respondents.
The survey also explored whether or not surgeons feel that AI could be a threat to the healthcare workforce in the next five years. Even though respondents believe AI will be more accurate as time goes on, they do not appear to feel threatened in any way when it comes to job security. One common concern, however, is the fact that AI may increase a health systems legal liability if mistakes are made.
Cardiac surgeons stood out yet again by being less concerned about risks to patient privacy than general thoracic surgeons.
“As with the rest of society, AI integration is certain to occur in the field of CT surgery,” the authors wrote. “This survey was not for the purpose of highlighting this process but rather as a means of assisting the specialty in incorporating AI technology in a measured, prospective fashion. We felt that the delineation of the attitudes and opinions of cardiothoracic surgeons regarding the role of AI would provide a good foundation for specialty leadership to begin deliberations and decision-making regarding both the site and process of its introduction. Identifying the attitudes, hopes, beliefs, and fears of practitioners should facilitate future policy decisions regarding which clinical and administrative domains will prove most fruitful for early AI introduction, and which should be delayed.”
Click here for the full analysis.
Michael has more than 19 years of experience as a professional writer and editor. He has written at length about cardiology, radiology, artificial intelligence and other key healthcare topics.