By Sam Kay, VP of Pharma, Basil Systems
Drug repurposing has long promised a faster, cheaper, and safer path to therapeutic discovery. In theory, existing compounds with known safety profiles could be redirected toward new diseases, bypassing early-stage development and de-risking regulatory approval. But in practice, most repurposing efforts still depend on luck, limited datasets, or outdated literature searches.
At Basil Systems, we believe it’s time to treat drug repurposing as a systematic science, not a serendipitous one. We’ve developed an integrated data platform that brings together three of the most powerful, yet underutilized, assets in drug development: regulatory labels, clinical trial registries, and real-world adverse event data. And in a recently completed retrospective validation study, this approach achieved 100% accuracy in predicting new approved indications across three off-patent compounds using only the data that was available before their label expansions.
That’s not a coincidence. It’s the result of building a better framework, one that respects the complexity of pharmaceutical evidence but brings modern signal detection and cross-validation to bear on the problem.
Why Repurposing Is Broken
The current state of drug repurposing is disjointed. Literature-based methods are limited to what’s already been published. Clinical trial analysis often suffers from timing bias—by the time a trial shows up in the public record, it’s already in progress. And adverse event signals, while valuable, are rarely examined in context with regulatory labels or dosing history. Each dataset operates in its own silo.
Traditional approaches also lack the rigor to filter out weak or coincidental associations. As a result, many repurposing claims fail to replicate, stall in development, or never gain regulatory traction. What’s missing is a method to triangulate across multiple sources of evidence, one that can prioritize truly actionable leads.
A New Method for Discovery
We built the Basil Intel platform to address this gap. Leveraging a dataset of over 600 million indexed life science records, the methodology integrates and targets adverse event reports, comprehensive clinical trial metadata, and regulatory label documents, harmonized across major global agencies. Using proprietary algorithms, the system detects signals within each dataset, then scores and ranks repurposing opportunities based on cross-source validation, regulatory precedent, and clinical relevance.
In our validation study, we applied this approach to three off-patent compounds with documented label expansions approved between 2015 and 2024. Using strict temporal controls, we simulated a real-world discovery environment—analyzing only data that would have been available prior to the new indication approvals.
The results were striking. In all three cases, the Basil platform ranked the approved new indication within the top three predicted candidates. The probability of achieving this outcome by chance was less than 0.1%. By comparison, adverse event analysis alone identified the correct indication in only 24% of cases. Clinical trial data alone performed slightly better, but was still limited by timing and scope. Regulatory label review, meanwhile, was not feasible as a standalone method, as it would have required thousands of hours of manual effort.
Evidence That Matters
What made the difference was integration. When we looked at the top-ranked repurposing opportunities flagged by Basil, nearly 90% had supportive signals across all three data sources. These were not speculative leads—they were robust, evidence-backed candidates grounded in real-world data, clinical activity, and regulatory context.
And the platform’s strength wasn’t limited to retrospective validation. When we ran a prospective analysis on five additional compounds currently undergoing label expansion efforts, the Basil platform identified aligned indications in four out of five cases—again, using only pre-approval data.
This suggests that our framework can function not just as a diagnostic of what’s already happened, but as a predictive tool for what’s likely to work in the future.
Practical Implications for Industry
For drug developers, this kind of validated, systematic repurposing pipeline could significantly shorten development timelines, reduce costs, and derisk portfolios. It’s particularly valuable for under-leveraged assets, molecules that are off-patent, out of favor, or stuck in therapeutic niches with limited commercial upside.
For regulatory and medical affairs teams, it offers a way to prioritize post-market opportunities backed by reproducible evidence. For payers and policymakers, it provides a rational framework for expanding access to therapies where unmet need is high and existing data may support new use.
In short, it creates an operating system for repurposing that’s scalable, evidence-driven, and built for real-world application.
What Comes Next
We’re continuing to validate the Basil platform across additional compounds and therapeutic areas, and are in discussions with partners ranging from global pharma to digital health innovators looking to augment their own pipelines with our analytics. We’re also exploring how this approach could be extended into rare disease landscapes, where drug development is historically slow and expensive.
Our goal isn’t just to find new uses for old drugs, it’s to bring structure and predictability to a space that’s been defined for too long by luck and anecdote.
The pharmaceutical industry has mountains of data. What’s been missing is a unified framework to make sense of it. At Basil, we’re building that framework. And for the first time, drug repurposing may finally be ready to live up to its potential.
