Medical Care
How Mayo Clinic Leverages Real-World Data for Precision Medicine
2024-12-09
Mayo Clinic clinicians are now delving into the potential use of healthcare-specific large language models. These models, accessed through a generative artificial intelligence chat application, aim to enhance patient care and improve clinical decisions. In a field where ChatGPT and Google Gemini may only provide relevant answers a fraction of the time, California-based Atropos Health claims its federated healthcare data network can offer detailed and accurate consultations even for the most obscure medical questions. By drawing on data from millions of patients, clinicians can gain valuable insights to inform treatment. For instance, when treating a patient with an unusual genetic condition predisposing them to a specific cardiac disease, such data can be crucial. Last year, Atropos launched a generative AI-enhanced platform with a chat-based interface called ChatRWD. Dr. Peter Noseworthy, chair of cardiac electrophysiology at Mayo Clinic, has begun testing it. The platform provides a Real World Data Score and a Real World Fitness Score for each dataset, helping users select the most suitable data. A study led by Saurabh Gombar showed that healthcare-specific LLMs like OpenEvidence and ChatRWD were able to produce actionable, reliable evidence 42% or 60% of the time, far exceeding general-purpose LLMs. Since 2022, Atropos has been collaborating with Mayo Clinic to develop data-driven methods through automated reports called Prognostograms. This allows physicians and researchers to access Mayo Clinic's deidentified data repository. For patients in critical care, the ability to find answers through the platform saves time. Observational clinical researchers often face long timelines to generate insights from real-world data. But with a tool like ChatRWD, they can get close to research-grade information. Atropos projects more than 200% growth in additional dataset availability over the next year. The power of patient data emerges with AI as it can capture the totality of patient outcomes that traditional medical research may miss. Rare or unusual conditions are better represented in large data samples. Mayo Clinic has been working on a decentralized clinical trial program to extend the reach of clinical trials. Access to clinical trials has exacerbated health disparities, and the goal is to bring more cures to more people. While ChatRWD is yet to be deployed at scale at Mayo Clinic, it has already provided some interesting insights. Andrea Fox is the senior editor of Healthcare IT News. Email: afox@himss.org. Healthcare IT News is a HIMSS Media publication.

The Power and Potential of ChatRWD

ChatRWD has emerged as a significant tool in the healthcare field. It allows clinicians to interact with real-world data in real-time and surface those insights directly at the point of care. This is a game-changer as it saves time and provides valuable information that can inform treatment decisions. For example, in critical care situations, where traditional means of determining treatment can take weeks, an AI-driven Prognostogram can complete the process in days. This not only benefits patients but also helps clinicians make more informed decisions quickly. The ability to access and analyze large amounts of real-world data through ChatRWD is a powerful asset that has the potential to transform healthcare.

Moreover, the scores provided by Atropos' platform, such as the Real World Data Score and the Real World Fitness Score, help users select the most appropriate dataset for their questions. This ensures the accuracy and reliability of the information obtained. The study led by Saurabh Gombar demonstrated the superiority of healthcare-specific LLMs like OpenEvidence and ChatRWD over general-purpose LLMs in terms of producing actionable, reliable evidence. This highlights the importance of specialized language models in healthcare.

The Impact on Clinical Trials

Mayo Clinic's decentralized clinical trial program is another area where ChatRWD can have a significant impact. By extending the reach of clinical trials beyond major academic medical centers, more patients can participate and benefit from potential cures. Access to clinical trials has been a challenge, especially for underrepresented populations and those in underserved rural communities. ChatRWD can help bridge this gap by providing a platform for data-driven research and decision-making. It allows clinicians and researchers to access deidentified data from Mayo Clinic and analyze it in real-time, leading to more inclusive and effective clinical trials.

However, while ChatRWD shows great promise, it is still in the process of being deployed at scale at Mayo Clinic. But even in its current form, it has already provided some interesting insights and has the potential to revolutionize healthcare. As Atropos projects significant growth in additional dataset availability, the possibilities for using ChatRWD and other healthcare-specific language models become even more exciting.

The Role of Real-World Data

Real-world data plays a crucial role in healthcare with the advent of AI. It allows clinicians to capture the totality of patient outcomes that may be missed in traditional medical research. For example, rare or unusual presentations of diseases are better represented in large data samples, providing a more comprehensive understanding of these conditions. This can lead to improved treatments for patients who are historically outside the reach of clinical trials. By using real-world clinical data, clinicians can gain a deeper understanding of patient responses to treatments and make more informed decisions.

Compared to just using ChatGPT or other general-purpose LLMs, real-world clinical data offers a more targeted and relevant approach. It is based on actual patient experiences and outcomes, making it a valuable resource for healthcare. The collaboration between Atropos and Mayo Clinic in developing data-driven methods through Prognostograms is a testament to the importance of real-world data in improving healthcare.

more stories
See more