How we helped Idorsia prove treatment superiority using AI and real-world evidence

Challenge: Demonstrating superiority to the standard of care using RWE

Idorsia, a global biopharmaceutical company, set out to demonstrate that competing insomnia treatments underperformed compared to their own drug in terms of patient-reported outcomes (PROMs). Traditionally, the gold standard to prove such claims would be double-blind RCTs. However, running RCTs for each competing treatment would have been financially unviable and lacked flexibility.

At the same time, relying on Real-World Evidence (RWE) was not feasible with established industry methodologies, as patient-reported outcomes are rarely captured in structured data, making it impossible to compare treatments using conventional approaches. This created a major barrier: how to generate credible, data-driven insights without relying on impractical RCTs or incomplete RWE datasets.

How we helped

Our primary objective was to develop a patient-reported outcome measure (PROM) and conduct subsequent causal studies to demonstrate the effectiveness of various treatments. The single approach to make the project possible was to quantify the signal from doctor notes in RWE data through a modelled abstract ontology, which hadn’t been done prior to our study.

Together with our client, we designed an entirely new methodology to define PROMs by extracting hidden signals from unstructured physician notes.

We built DiSMOL, an AI framework that leverages natural language processing to analyze clinical notes and detect signs of daytime impairment in insomnia patients, something that diagnosis codes alone often miss. At the core of the solution was an abstract ontology, designed to quantify PROM-like signals from doctors’ free text.

This innovation enabled us to measure treatment effects across different therapies, while also unlocking new opportunities for RWE studies that were previously considered impossible. The methodology and results were later peer-reviewed and published in Nature Communications Medicine, marking a significant scientific milestone.

The impact: Bridging clinical research and data science

By combining data science and domain expertise, we helped Idorsia transform previously unused, unstructured data into actionable real-world evidence, proving that AI-driven RWE can complement, and in some cases replace, costly clinical trials.

• Set a new industry standard →  introduced a novel way of defining PROMs in RWE studies.
• Enabled future research → created a methodology that can be applied to other therapeutic areas.
• Commercial impact → strengthened our client’s market position by supporting their commercial strategy with robust RWE evidence.