UC Berkeley Researchers Develop Machine Learning Methods to Accelerate Drug Development Using Real-World Evidence
核心洞察
UC Berkeley (搜索)'s Center for Targeted Machine Learning and Causal Inference is developing new methodologies that combine machine learning with biostatistics to make clinical trials faster and more affordable while maintaining scientific rigor.
The center has secured $4.7 million in funding from Novo Nordisk and Gilead Sciences to advance real-world evidence applications in drug development, following the FDA's framework established after the 21st Century Cures Act.
Researchers are creating statistical techniques to simulate placebo trials using real-world health data and replace traditional drug trial randomization, potentially reducing the decade-long timeline and billion-dollar costs of bringing new drugs to market.
More than 300 million people worldwide suffer from diseases without cures, including many cancers, Alzheimer's disease, Parkinson's disease, and over 30 million Americans with rare diseases lacking effective treatments. The challenge isn't lack of research effort—developing a new drug costs between $173 million and $2.6 billion, takes an average of 12 years, and only 12% of potential drugs receive FDA approval, according to National Institutes of Health estimates.
Researchers at UC Berkeley (搜索) School of Public Health are pioneering solutions to accelerate drug development through the Center for Targeted Machine Learning and Causal Inference (CTML (搜索)), which leverages real-world evidence and advanced statistical methods to streamline the approval process without compromising patient safety.
Transforming Drug Development Through Real-World Evidence
The research initiative stems from the 21st Century Cures Act of 2016, which prompted the FDA to create a framework for evaluating "real-world evidence"—clinical data from electronic health records, medical claims, and digital health technologies. While proponents saw potential for supporting new drug approvals and post-market studies, critics worried about replacing the gold standard of clinical trials with potentially unsafe alternatives.
"Clinical trials, particularly large Phase 2 and 3 trials, are one of the biggest reasons it can take more than a decade and over a billion dollars to bring a new drug to patients," said Dr. Michael C. Lu, Dean of UC Berkeley (搜索) School of Public Health. "By combining the power of machine learning with the rigor of modern biostatistics, our researchers are developing new methodologies—such as using real-world health data to simulate placebo trials—that could make clinical trials faster, more efficient, and more affordable while maintaining the highest standards of scientific rigor."
Advanced Statistical Methods Bridge Research Gap
CTML (搜索) specializes in causal inference—scientific inquiry using formal mathematical frameworks to establish cause-and-effect relationships rather than mere statistical associations. Targeted learning, a statistics subfield combining causal inference, machine learning, and statistical theory, provides the foundation for answering scientifically impactful questions with statistical confidence.
The center is co-directed by Dr. Maya L. Petersen, professor of biostatistics and epidemiology and co-director of the UCSF UCB Joint Program in Computational Precision Health; Dr. Mark van der Laan, professor of Biostatistics and Statistics; and Dr. Alan Hubbard, professor of Biostatistics.
Van der Laan, who developed the targeted learning field, has collaborated with the FDA since 2010, when the agency commissioned him and Dr. Susan Gruber to demonstrate targeted learning applications in drug safety studies. Both Van der Laan and Petersen have conducted workshops for FDA personnel on methods to expedite drug study and approval processes.
Major Industry Partnerships Drive Innovation
In 2020, Novo Nordisk provided a $3.2 million research gift to launch the Joint Initiative for Causal Inference (JICI (搜索)), expanding collaboration to include Copenhagen University, Oxford University, Harvard University, and University College London. JICI develops methods to draw causal conclusions from observational and clinical trial datasets using machine learning advances through targeted learning frameworks.
JICI (搜索) researchers create statistical techniques that replace drug trial randomization while enabling causal conclusions to support drug development programs. Their work reduces randomized clinical trial costs, integrates clinical trial data with observational data, produces valid randomized trial analyses, and develops techniques for accelerated candidate drug screening.
The initiative's success led to a new partnership with Gilead Sciences, which contributed $1.5 million over three years beginning in 2024 to investigate optimal real-world evidence applications.
Addressing Global Health Challenges
The center's mission addresses a critical healthcare gap where traditional drug development timelines leave millions without treatment options. By harnessing cutting-edge causal inference and targeted learning for robust discoveries and informed decision-making, CTML (搜索) aims to improve public health outcomes while maintaining regulatory standards.
"This is exactly why public research universities matter," Lu added. "CTML (搜索) exemplifies how the intellectual firepower of Berkeley can help solve some of the world's most pressing problems."