Why Are Clinical Trial Terminations Increasing?
Insights from BI Board Member Rich Harrison
Introduction
Clinical trial terminations are rising, an unexpected trend given the rapid advances in AI, bioinformatics, and predictive modeling that are supposed to make drug development more informed and less risky.
A recent Nature Reviews Drug Discovery analysis of industry-sponsored phase II and phase III trials from 2013 to 2023 found that the number of terminated trials has more than doubled over the past decade, increasing from 209 (11% of all trials) in 2013 to 435 (23% of all trials) in 2023. The analysis focuses specifically on trials that were actively terminated by sponsors before completion, rather than including trials that completed but later failed to meet endpoints.
These late-stage terminations represent significant lost time, capital, and scientific effort across the drug development ecosystem. Behind those numbers are also real patients who enrolled in trials hoping for a better outcome, only to have those studies end before completion.
One of the authors of this study is Rich Harrison, a member of the Bridge Informatics board. We spoke with Rich about the role bioinformatics can and should play in more rigorously validating biology before clinical trials begin, helping teams enter the clinic with stronger evidence and lower risk.
For Rich, the troubling part is that many of these risks could potentially be addressed much earlier in development.
The role of bioinformatics in strategy and business terminations
The study reveals that many trials are terminated for strategic or business reasons, followed by lack of efficacy and enrollment challenges. This “strategic or business” category, which accounts for 36% of all terminations in the study, includes competitive pressures, portfolio reprioritization, and the well-documented problem of “herding”, where multiple companies pursue the same target simultaneously.
This is an area where bioinformatics is underutilized but shouldn’t be. Computational tools can map the competitive landscape in real time, flagging how crowded a market space is before a company commits to a costly late-stage program. Indication expansion modeling can identify alternative uses for a drug candidate early, reducing the risk that a competitor’s success renders an entire program commercially unviable.
Better upstream intelligence, powered by bioinformatics, could help make more informed portfolio decisions and avoid late stage clinical trial terminations.
The role of bioinformatics in biomarkers and earlier validation
As bioinformaticians, we spend a lot of time thinking about how data can de-risk decisions. Rich believes this is exactly where bioinformatics can make a major difference.
“Bioinformatics would go a long way here,” Rich told us. “The more validation that goes in early, the less risk you carry into clinical trials.”
In other words, stronger computational and molecular evidence before entering the clinic can dramatically improve confidence in a drug target, mechanism, or patient population.
One of the biggest opportunities lies in biomarker-driven validation. With deeper genomic and molecular analysis, teams can better confirm whether a target is truly implicated in disease and whether a therapy is likely to produce meaningful outcomes.
If more work were done earlier, leveraging bioinformatics alongside biomarkers, developers could enter clinical trials with stronger evidence and clearer hypotheses. Without that early validation, teams risk discovering fundamental issues only after significant investment has already been made.
As Rich put it, skipping that step can become “an expensive risk.”
A role for bioinformatics in reducing trial failure
For those of us working in bioinformatics, the takeaway is clear. The work done upstream, target validation, biomarker discovery, and patient stratification, has a direct impact on whether therapies succeed or fail in the clinic.
The more rigorously we validate biology before trials begin, the better chance we have of avoiding costly late-stage terminations.
And as Rich’s research highlights, strengthening that early evidence may be one of the most effective ways to improve the efficiency of drug development.
Read the full Nature article here: Analysis of phase II and phase III clinical trial terminations from 2013 to 2023
Conclusion
Rich’s perspective is especially valuable to us at Bridge Informatics. As a member of our board, he helps guide how we design our services to better meet the needs of the industry. His experience in drug development and clinical research helps ensure that the solutions we build focus on what matters most to our clients: reducing risk early, validating biology more rigorously, and improving the odds of success in the clinic.
Insights like those in this analysis reinforce why stronger bioinformatics and biomarker-driven validation are so important, not just for individual programs, but for the efficiency of the entire drug development ecosystem.
Click here to meet with a member of our team and explore ways we can work together.