Italy
December 30, 2023Benevolent AI
January 31, 2024How AI will – and won’t – change health care in 2024
The vast advancement in the field of AI during 2023 makes for a promising 2024, where an intersection of AI and the world's top medical systems will pave the way to a future where data-driven insights catalyze a paradigm shift in healthcare delivery...
The vast advancement in the field of AI during 2023 makes for a promising 2024, where an intersection of AI and the world’s top medical systems will pave the way to a future where data-driven insights catalyze a paradigm shift in healthcare delivery. AI has the potential to distill data from the best medical systems and transform it into enhanced and personalized care accessible to all. AI is shaping a healthcare landscape that prioritizes innovation, efficiency, and inclusivity for the benefit of communities worldwide.
How AI Could Turn Data from the Best Medical Systems into Better Care for All
Muhammad Mamdani, Vice President of Data Science and Advanced Analytics at Unity Health Toronto, explains a nuanced perspective on the role of artificial intelligence (AI) in transforming healthcare in 2024. Mamdani is also a professor at the University of Toronto and director of the Temerty Centre for Artificial Intelligence Research and Education in Medicine. He emphasizes the gradual integration of AI into clinical decision-making processes. He also shares his optimism for AI’s potential to improve patient care through innovations like chatbot-style solutions for health-related inquiries and post-discharge communication. However, Mamdani also addresses challenges, including the need for improved information sharing among healthcare providers, ethical considerations regarding privacy and algorithmic biases, and the financial burden of deploying AI solutions. Despite acknowledging apprehensions, Mamdani advocates for a thoughtful and responsible deployment of AI in healthcare, emphasizing its inevitable integration for the greater benefit of patients.(1)
Muhammad Mamdani, Vice President of Data Science and Advanced Analytics at Unity Health Toronto, explains a nuanced perspective on the role of artificial intelligence (AI) in transforming healthcare in 2024. Mamdani is also a professor at the University of Toronto and director of the Temerty Centre for Artificial Intelligence Research and Education in Medicine. He emphasizes the gradual integration of AI into clinical decision-making processes. He also shares his optimism for AI’s potential to improve patient care through innovations like chatbot-style solutions for health-related inquiries and post-discharge communication.
However, Mamdani also addresses challenges, including the need for improved information sharing among healthcare providers, ethical considerations regarding privacy and algorithmic biases, and the financial burden of deploying AI solutions. Despite acknowledging apprehensions, Mamdani advocates for a thoughtful and responsible deployment of AI in healthcare, emphasizing its inevitable integration for the greater benefit of patients.(1)
Impact of a deep learning sepsis prediction model on quality of care and survival
In a significant breakthrough, researchers at the University of California San Diego School of Medicine have employed an artificial intelligence (AI) model called COMPOSER to predict sepsis infection in high-risk patients at the emergency departments of UC San Diego Health. The AI system, working silently behind the scenes, continuously analyzes over 150 patient characteristics upon arrival, including vital signs, test results, medical history, and demographics. If high-risk factors for sepsis are identified, the system alerts nursing staff through the hospital’s electronic health record (EHR), allowing the medical team to initiate timely interventions. The study, published in npj Digital Medicine, demonstrates a 17% reduction in mortality with the implementation of COMPOSER.(3)
UC San Diego Health, a trailblazer in AI healthcare, has integrated the AI model into various hospital in-patient units and plans to extend its use to the newly established UC San Diego Health East Campus. The research, supported by the National Institutes of Health and the Joan and Irwin Jacobs Center for Health Innovation, underscores the potential of AI to enhance patient outcomes in emergency settings.(3)
In a significant breakthrough, researchers at the University of California San Diego School of Medicine have employed an artificial intelligence (AI) model called COMPOSER to predict sepsis infection in high-risk patients at the emergency departments of UC San Diego Health. The AI system, working silently behind the scenes, continuously analyzes over 150 patient characteristics upon arrival, including vital signs, test results, medical history, and demographics. (3)
If high-risk factors for sepsis are identified, the system alerts nursing staff through the hospital’s electronic health record (EHR), allowing the medical team to initiate timely interventions. The study, published in npj Digital Medicine, demonstrates a 17% reduction in mortality with the implementation of COMPOSER.UC San Diego Health, a trailblazer in AI healthcare, has integrated the AI model into various hospital in-patient units and plans to extend its use to the newly established UC San Diego Health East Campus. The research, supported by the National Institutes of Health and the Joan and Irwin Jacobs Center for Health Innovation, underscores the potential of AI to enhance patient outcomes in emergency settings.(3)
AI in medicine, new therapeutics for Alzheimer's, and GLP-1 for weight loss
In a recent AMA Update video and podcast, Dr. Kirsten Bibbins-Domingo, Editor-in-Chief of JAMA and the JAMA Network, discusses the accomplishments and future goals for JAMA in 2024. Dr. Bibbins-Domingo highlights the success of launching new initiatives, such as the first in-person JAMA Summit, bringing experts together to discuss evidence generation and clinical trials. Looking ahead, she emphasizes the importance of journals being a trusted source of science and improving communication through quicker and more accessible platforms. The discussion also touches on the acceleration of new therapeutics, including weight loss drugs and AI-related advancements. (4)
Dr. Bibbins-Domingo expresses excitement about upcoming JAMA Summits and emphasizes the role of journals in communicating science effectively, addressing the demand for multimedia content, and interacting with enthusiastic audiences interested in staying informed about the latest developments in science and medicine. The AMA Update concludes with Dr. Bibbins-Domingo’s personal ambition to execute ambitious projects for the journal in the coming year.(4)
In a recent AMA Update video and podcast, Dr. Kirsten Bibbins-Domingo, Editor-in-Chief of JAMA and the JAMA Network, discusses the accomplishments and future goals for JAMA in 2024. Dr. Bibbins-Domingo highlights the success of launching new initiatives, such as the first in-person JAMA Summit, bringing experts together to discuss evidence generation and clinical trials. (4)
Looking ahead, she emphasizes the importance of journals being a trusted source of science and improving communication through quicker and more accessible platforms. The discussion also touches on the acceleration of new therapeutics, including weight loss drugs and AI-related advancements.Dr. Bibbins-Domingo expresses excitement about upcoming JAMA Summits and emphasizes the role of journals in communicating science effectively, addressing the demand for multimedia content, and interacting with enthusiastic audiences interested in staying informed about the latest developments in science and medicine. The AMA Update concludes with Dr. Bibbins-Domingo’s personal ambition to execute ambitious projects for the journal in the coming year.(4)
Google AI has better bedside manner than human doctors — and makes better diagnoses
A Google-developed chatbot named Articulate Medical Intelligence Explorer (AMIE), based on a large language model (LLM), demonstrated the ability to match or surpass human doctors’ performance in conversing with simulated patients and suggesting possible diagnoses based on medical history. The experimental chatbot outperformed board-certified primary-care physicians in diagnosing respiratory and cardiovascular conditions, and it ranked higher in empathy during medical interviews. Although the chatbot is still experimental and has not been tested on real health problems, researchers suggest it could eventually contribute to democratizing healthcare. However, caution is advised, and the tool should not replace interactions with physicians, as medicine involves more than just collecting information. The study is yet to undergo peer review.(5)
A Google-developed chatbot named Articulate Medical Intelligence Explorer (AMIE), based on a large language model (LLM), demonstrated the ability to match or surpass human doctors’ performance in conversing with simulated patients and suggesting possible diagnoses based on medical history. The experimental chatbot outperformed board-certified primary-care physicians in diagnosing respiratory and cardiovascular conditions, and it ranked higher in empathy during medical interviews. (5)
Although the chatbot is still experimental and has not been tested on real health problems, researchers suggest it could eventually contribute to democratizing healthcare. However, caution is advised, and the tool should not replace interactions with physicians, as medicine involves more than just collecting information. The study is yet to undergo peer review.(5)
An AI revolution is brewing in medicine. What will it look like?
AI tools in medicine, particularly in radiology, have been discussed and developed for several years. However, many clinicians are still cautious about fully embracing these tools due to various limitations and concerns. AI tools are often developed for specific tasks and may not provide a comprehensive interpretation of medical examinations. Additionally, the performance and safety of these tools need to be closely monitored, and external validation needs to be obtained to ensure accurate results. The current approach involves adding more specialized AI tools, leading to challenges in integration and deployment within healthcare systems.
The article mentions a growing interest in a new approach called generalist medical AI. Inspired by large language models like ChatGPT, these models are trained on massive datasets and can be adapted for various tasks. Unlike specific AI tools, generalist models aim to assess anomalies comprehensively and provide a more holistic diagnosis. The goal is not to replace physicians but to enhance their capabilities, especially in areas where AI can excel.
Despite the potential benefits, significant challenges are ahead, including the need for rigorous testing, addressing biases, and ensuring the safety and effectiveness of these models in real-world clinical settings. The journey towards implementing generalist medical AI for clinical care is still in its early stages, and more research and development are needed before widespread adoption. (6)
AI tools in medicine, particularly in radiology, have been discussed and developed for several years. However, many clinicians are still cautious about fully embracing these tools due to various limitations and concerns. AI tools are often developed for specific tasks and may not provide a comprehensive interpretation of medical examinations. Additionally, the performance and safety of these tools need to be closely monitored, and external validation needs to be obtained to ensure accurate results. The current approach involves adding more specialized AI tools, leading to challenges in integration and deployment within healthcare systems. (6)
The article mentions a growing interest in a new approach called generalist medical AI. Inspired by large language models like ChatGPT, these models are trained on massive datasets and can be adapted for various tasks. Unlike specific AI tools, generalist models aim to assess anomalies comprehensively and provide a more holistic diagnosis. The goal is not to replace physicians but to enhance their capabilities, especially in areas where AI can excel.
Despite the potential benefits, significant challenges are ahead, including the need for rigorous testing, addressing biases, and ensuring the safety and effectiveness of these models in real-world clinical settings. The journey towards implementing generalist medical AI for clinical care is still in its early stages, and more research and development are needed before widespread adoption.(6)