Real-world data is crucial for healthcare tech developers, as it reflects actual patient experiences, treatments, and outcomes in diverse environments, unlike controlled clinical trials, according to Atropos Health's Brigham Hyde. However, many developers face challenges in accessing this data. An April study revealed that of over 500 studies on large language models in healthcare, only 5% used real-world patient data. To tackle this issue, Atropos Health announced that AI developers can now train their models using its real-world evidence network. Founded in 2020 as a Stanford spinout, Atropos provides real-world clinical data to physicians. In 2023, it launched an evidence network featuring more than 300 million patient records from EHRs, claims data, and patient registries. The network comprises “dozens” of members, including developers, practitioners, researchers, data holders, and tech companies, offering a comprehensive view of disease progress and treatment efficacy across diverse populations. AI developers can now integrate their models with the network's infrastructure, facilitated by Atropos’ GENEVA OS platform.

This platform transforms real-world data into clinical evidence, offering fast, data-driven answers to complex medical questions. “GENEVA OS allows developers to train, test, and validate predictive models on standardized patient-level data, speeding model development while ensuring transparency, bias detection, and accuracy, ” Hyde explained. This infrastructure aims to accelerate AI development and increase reliability, driving innovations that enhance patient care. Hyde mentioned use cases for AI tools on the network include clinical trial simulation, patient journey mapping, cost of care estimation, and outcome prediction. Validated models can be deployed to Atropos’ partners, such as health systems and pharmaceutical companies. QuantHealth, a network member, uses AI to help pharmaceutical companies develop treatments more efficiently. Their CEO, Orr Inbar, praised Atropos’ platform for enabling his company to refine their product effectively. “Optimizing clinical trials through patient-level simulations is challenging, prompting us to enhance our AI platform and data frameworks, ” Inbar stated. “This evolution has helped 7 of the top 20 pharma companies improve their trials and clinical programs for excellence. ” QuantHealth now performs real-time simulations and deploys AI models at the point of care, opening new opportunities for its pharma clients, according to Inbar. Photo: metamorworks, Getty Images