The new platform, trained on 100 million patient records, aims to improve accuracy in trial simulation and support data-driven design decisions.

QuantHealth has launched what it describes as the first clinical trial-specific foundation model, trained on more than 100 million patient records and 30 billion drug-patient interactions. Known as LRDM v1.0 (Large Real-World Drug Model), the system enables simulation and optimization of trial designs using real-world therapeutic data and transformer-based AI architectures.

The model is designed to support sponsors in addressing persistent inefficiencies in clinical research, particularly during the trial phase of drug development – where over 90% of candidate therapies fail and where the majority of R&D investment is spent. According to QuantHealth, the tool can be used to predict patient outcomes, assess drug effectiveness, and optimize inclusion criteria before studies begin.

“This research is now bearing fruit, with the first version of LRDM now live, enabling us to leverage far more data, with deeper granularity, enhancing the accuracy of our models and unlocking new capabilities,” said Orr Inbar, CEO and co-founder of QuantHealth.

“With new therapies being more precise and more complex, this kind of deep modelling is crucial to advancing the state of care.”

The LRDM v1.0 model is expected to become available on AWS Marketplace and can also be accessed through QuantHealth’s self-service platform or via API. It supports trial designs across therapeutic areas including oncology, cardiovascular disease, and autoimmune conditions, and is compatible with a range of modalities such as small molecules, biologics, and cell and gene therapies.

QuantHealth says it currently works with eight of the top 20 pharmaceutical companies, helping teams run dozens of trials annually across all development phases. Its newest model features 138 million parameters and leverages 500 times more data than its predecessor.

“LRDM v1.0 can capture clinical patterns at a much greater temporal resolution and reason about drug mechanisms at a far greater depth,” said Eran Barash, head of AI and machine learning at QuantHealth.

“It also enables generalization from large populations to rare diseases and small subpopulations, which is particularly valuable in precision medicine.”

AWS is collaborating with QuantHealth to expand access to the tool. “The collaboration between AWS and QuantHealth will further enable pharma companies to leverage clinical foundation models for their use cases, increasing the fidelity of their clinical programs and improving the probability of success in clinical trials,” said Dan Sheeran, general manager for health care and life sciences at AWS.