Artificial intelligence (AI) has become a more acceptable and widespread tool in a variety of areas. Increasingly, ChatGPT is a popular tool in various educational, research, and clinical settings.
This article focuses on a fascinating preprint called, “Harnessing the Power of ChatGPT in Cardiovascular Medicine: Innovations, Challenges, and Future Directions”. This research was done by Marc Leon, Chawannuch Ruaengsri, Glenn J. Pelletier, Daniel Miguel Bethencourt, Masafumi Shibata, Manuel Quiroz Flores, and Yasuhiro Shudo from Stanford University.
ChatGPT and cardiovascular medicine
ChatGPT was launched in 2022 by Open AI and exceeded 100 million active users within two months. In the few years since its release, it has seen significant upgrades and widespread adoption. With each version, the capabilities of AI when it comes to medical knowledge have increased.
In this article, the authors set out to explore the use of artificial intelligence and, in particular, ChatGPT, in medical research, practice, and education related to cardiovascular diseases (CVDs). It can be used for all kinds of tasks to enhance the efficiency and accuracy of testing, diagnostics, and treatment.
With a review of retrospective data, the authors explain the different ways in which ChatGPT can enhance treatment outcomes in CVDs.
In their own words:
“This comprehensive review aims to systematically evaluate the innovative applications of ChatGPT in the CVD field, addressing current challenges, discussing future directions for development, and providing valuable insights for the vast patient population and the extensive community of professionals committed to cardiovascular health.”
Harnessing ChatGPT
So, how can ChatGPT be harnessed to improve patient outcomes in cardiovascular diseases?
This article outlines three broad considerations related to the use of ChatGPT in cardiovascular medicine: innovative applications, persistent challenges, and future directions.
Innovative applications using ChatGPT
The innovative applications discussed include enhancing clinical decision-making, innovating medical education, and advancing cardiovascular research. This article emphasizes the different ways that researchers, physicians, and other healthcare workers can benefit from the support of ChatGPT in their everyday work environment.
The authors include comparisons of the decisions recommended by ChatGPT versus the clinicians themselves. They report that the knowledge base of ChatGPT is similar to that of doctors in training. Even without advanced training, it consistently achieved higher than a passing mark on an American medical licensing board. Further, ChatGPT made accurate assessments and recommendations in the majority of cases. It even outscored the diagnostic work of doctors.
They also consider the role of AI in education for healthcare professionals and patients alike. Patients themselves often use online tools, like a Google search or, increasingly, ChatGPT, to do their own research. Some might even use it as a substitute for a healthcare clinician to assess symptoms before seeking out medical advice.
Harnessing AI to improve health literacy and alleviate educational and language barriers can have a significant impact. ChatGPT can improve communication between doctors and patients. One example of this includes helping to simplify and translate often dense medical literature to plainer language. Another is reviewing large amounts of data and providing summaries for public consumption.
Persistent Challenges
The challenges covered in this article include inaccuracies, ethical concerns, and a potential over-reliance on AI tools.
In particular, the authors emphasize the importance of safety for patients and research participants. There are life-threatening consequences of incorrect or late detection or delays in testing and treatments, because of ChatGPT or a human-made error. Therefore, strict ethical guidelines must be designed and followed to provide the most accurate results.
The article also points out potential limitations of this technology, particularly when it comes to being able to offer tailored treatment plans and understand more complicated medical cases.
Future Directions
Ultimately, the authors recommend a focus on real-world and context-specific clinical trials and studies of patients and clinicians dealing with diagnostics and treatment of CVDs. Based on their review of the literature and relevant case studies, the article concludes with five key recommendations:
- Expand and improve the training materials used for ChatGPT;
- Further develop customized ChatGPT models for different medical specializations;
- Ensure appropriate human oversight when using ChatGPT;
- Use real-world clinical trials;
- Implement rigorous ethical guidelines.
From Preprints.org to publication
Many preprints have gone on to be formally published in many peer-reviewed journals. Originally posted on preprints.org on August 28, 2024, this article went on to be published in MDPI’s Journal of Clinical Medicine on October 29, 2024. Preprints provide a valuable step in the academic process, allowing authors to get valuable feedback on their work. These improved preprints go on to find homes in journals, validating the findings through expert evaluation. Publishing a preprint allows the rapid dissemination of findings to a broad audience, encouraging early feedback and fostering scientific collaboration before formal peer review is complete.
Preprints.org publishes articles from all fields of research that report scientifically sound original research or present a comprehensive review of a field.
