Determine the inclusion criteria, endpoints, and study sizes with the highest chances of success.
Reduce or eliminate the need to run a concurrent control arm by incorporating synthetic control patients to model the standard of care.
Analyze trial results to identify patient subgroups that received more benefit from a treatment.
Our models are trained on curated longitudinal patient datasets originating from historical control arms, patient registries, and/or real world datasets.
We have regulatory grade data processing and model training frameworks that are fully reproducible and auditable.
We use patent-pending algorithms to simulate complete, longitudinal synthetic control patient records.
Unlearn uses proprietary machine learning algorithms to simulate synthetic control patients.
Each synthetic control patient serves as a virtual twin to an individual patient in the treatment population. Virtual twins de-risk clinical trials by eliminating confounding variables and increasing statistical power.
Unlearn’s synthetic control patients are indistinguishable from real patients receiving the standard of care or placebo.
Complete synthetic control patient records allow sponsors to perform the same in-depth analyses as in a traditional randomized controlled trial.