This paper discusses the challenges faced when implementing #ArtificialIntelligence tools in clinical practice and presents principles based on the authors' experiences to address these challenges. The authors provide an in-depth analysis of the difficulties encountered during the integration of #AI into #Healthcare settings, focusing on radiology as an example, and offer practical solutions to overcome these barriers. Key Messages: 1️⃣ Successful integration of AI tools in clinical practice requires addressing the technical, cultural, regulatory, and ethical aspects of implementation. 2️⃣ Collaboration between AI developers, clinicians, and other stakeholders is crucial for understanding and addressing the unique challenges faced in healthcare settings. 3️⃣ Establishing standards for AI performance, validation, and evaluation is necessary to ensure the safe and effective use of AI tools in clinical practice. 4️⃣ Appropriate education and training for healthcare professionals using AI tools can enhance their adoption and utilization while mitigating potential risks. 5️⃣ Ensuring data privacy, security, and compliance with regulatory requirements is essential for the successful deployment of AI solutions in healthcare. Why It Is Worth Reading: This paper is worth reading because it provides valuable insights into the practical challenges of implementing AI tools in healthcare settings, specifically #Radiology, from the perspective of experienced professionals. By discussing real-life examples and proposing feasible solutions, the authors contribute to the growing body of knowledge on AI integration in clinical practice. Reading this paper will help healthcare professionals, AI developers, and policymakers better understand the complexities of implementing AI tools in healthcare and work together to address the challenges and unlock the potential benefits of AI in improving patient care. ✍🏿 Bernardo Bizzo, Giridhar Dasegowda, MBBS, Christopher Bridge, Benjamin Miller, James M. Hillis, Mannudeep K. Kalra, Kimberly Durniak, Markus Stout, Thomas Schultz, Tarik Alkasab, Dr. Keith J. Dreyer. Addressing the Challenges of Implementing Artificial Intelligence Tools in Clinical Practice: Principles From Experience, Journal of the American College of Radiology, Volume 20, Issue 3, 2023, Pages 352-360, ISSN 1546-1440. DOI: 10.1016/j.jacr.2023.01.002. (Behind paywall)
Do we have a risk analysis framework for AI based solutions in healthcare - especially can it help us automate such an analysis ? I guess, it would help in development of a strategy for adoption ? There are many things we use everyday, which pose some risk, but we do not stop using them. So, do we have the leeway for a trade-off, at least for some opportunities.
AI should be viewed as a clinical decision tool that augments a practitioners care, such as wells criteria or Ottawa ankle rules. Great article and i feel we are aligned to the critical analysis competency within the medical standard of care. We should not blindly accept results / outputs and understand where limitations in the algorithm exist. Medico-legally you cannot defend because Skynet told me so. Much more to come in this space, but this publication and dialogue is necessary.
This post and article are interesting and intriguing. While there is a wide range of potential benefits of AI in healthcare, we must also acknowledge that there are numerous challenges that need to be addressed, and we currently do not have a clear solution for them. 🤔
This is an amazing...of course also how to solve problems related to an unscheduled or emergency situation that will be out of the machine learning ability, also the anaesthesiologists and critical care physiens should shoulder many data that will be processed by machine learning to have the prediction of certain complications or prognosis
Do we want A1 to be smarter than a human brain? To maybe take over our lives?
AI/machine learning in healthcare has a long way to go
Thank you for citing this
#4 is very much needed Jan
Chief Executive Officer & Founder at Cubismi, Inc.
1yTo clarify for anyone who needs it, this is from a #comprehensiveAI advocate. At a recent American College of Radiology meeting, Keith gave a presentation advocating for “Comprehensive AI” where we would continue (despite massive brittle failure rates) to pour $ into more and more “point solutions” for tasks. Knowing GE HealthCare AI orchestrator designs etc, will assume GE and closely aligned Radiology Partners is also pursuing this direction…. I predict that none of the goals outlined in this article will be achieved by the #comprehensiveAI groups, yet confident we will see it pushed and resourced for many more years to come. Who wins wilth this approach? Machine manufacturers, not patients nor physicians.