Regulation Threatens the Progress of Personalized Medicine
A new class of treatments calls for a new regulatory framework.
- By
- November 03, 2025
- CBR - Health Care
When people have fallen ill throughout history, medical practitioners have tended to treat them with one-size-fits-all medications—a single medicine for everyone with Illness A, and another for all those with Illness B. That began to change at the beginning of the 21st century, as human genome sequencing opened new avenues of treatment.
Known as personalized medicine, the approach involves developing drugs for small groups of patients based on their genomic profiles. Among the many custom medicines that have emerged over the past 25 years are treatments for lung cancer, genetic diseases, and cardiovascular ailments. Many of these drugs are far more effective than their generic predecessors.
However, research by University of Chicago’s Marina Chiara Garassino and Kunle Odunsi, Northwestern’s Marciano Siniscalchi, and Chicago Booth’s Pietro Veronesi indicates that traditional drug testing and approval processes threaten to prevent many promising personalized medicines from ever reaching the market.
A central problem is that as clinical researchers target more and more genetic alterations affecting ever smaller patient populations, they’re struggling to assemble suitable groups to conduct the clinical trials required to obtain regulatory approvals. Garassino, Odunsi, Siniscalchi, and Veronesi propose an alternative: Rather than focus on the approval of individual drugs, create a personalized drug discovery process that is itself subject to testing and regulatory approval. Such a shift, they write, would lower costs and give drugmakers the financial incentives to encourage, rather than hinder, the development of personalized medicines.
Under the existing approval processes, the US Food and Drug Administration, the European Medicines Agency, and other regulators require drugmakers to obtain approvals for new therapies by subjecting them to randomized controlled trials. This research method involves administering a new therapy to a treatment group and the best available alternative to a control group, and then monitoring survival statistics over several years. To be approved, the drug must produce higher success rates than its predecessor.
The researchers used a simple model of investment decision-making to analyze the financial incentives drug companies have to conduct personalized drug discovery under the current regulatory regime. They focused on how personalized therapies are transforming the treatment of lung cancer, which affects about 220,000 people in the United States annually.
Why the drug approval process fails for rare cancers
Their findings indicate that as the number of patients a targeted therapy is suited to treat declines, the volume of clinical trials conducted to seek its approval also falls. They note, for example, that over a 10-year period, researchers conducted 103 clinical trials for treatments targeting a set of genomic mutations experienced by 14.5 percent of lung cancer patients. In contrast, a set of mutations affecting 0.5 percent of patients was the focus of only three treatments that reached clinical trials.
The researchers note a nearly linear pattern to this trend. Their model “suggests that the current regulatory infrastructure is not well-suited to promote research to identify new drugs for ever smaller fractions of the population,” they write.
They further find that if drug companies continue to hold the prices of personalized therapies steady as they develop more and more of them, industry profits are on track to evaporate entirely due to the high cost of randomized trials. If they raise prices to cover the cost of obtaining approvals, the overall financial burden will skyrocket to a level that is socially unsustainable. The result is that regulation, rather than science, is fast becoming the limiting factor in developing many life-saving treatments, they conclude.
The researchers propose that regulators shift their focus from approving individual drugs to approving a drug discovery process. This pipeline, they write, would be supported by new technologies, such as machine learning (ML) and artificial intelligence, that allow personalized drugs to be developed through simulation.
The process they propose for discovering cancer treatments has four steps:
1. Sequence tumor DNA and RNA to identify the genomic “driver” alterations that drive tumor growth.
2. Conduct the critical step in the personalized drug discovery process of rebuilding altered proteins by replacing the multiyear, lab-based experimental process with faster ML methods, such as one using Google DeepMind’s AlphaFold tool.
3. Use supercomputers to screen billions of molecules and identify those that can “block” altered proteins from triggering tumor growth.
4. Identify compounds suitable for humans by combining ML methods with rapidly increasing knowledge of the molecular structures of safe drugs.
Ongoing research at the University of Chicago and Argonne National Laboratory is further investigating the feasibility of this process, say Garassino, Odunsi, Siniscalchi, and Veronesi. And they note that other medical researchers have already been working on some of the necessary steps. As an example, they cite a 2023 study demonstrating for the first time that, by using Google DeepMind’s AlphaFold and AI, it is possible in a single month to identify a new inhibitor to treat the most common type of liver cancer. Additionally, the FDA is already working on regulatory guidance for the use of AI in drug development.
If the drug discovery process the researchers propose proves safe and effective via clinical trials, drugmakers could use it to create personalized drugs for small patient groups, which would likely lead to further cuts in the per-patient cost of developing personalized drugs and provide substantial benefits to society at large, the authors say.
Marina Chiara Garassino, Kunle Odunsi, Marciano Siniscalchi, and Pietro Veronesi, “On the Economic Infeasibility of Personalized Medicine, and a Solution,” Working paper, October 2025.
More from Chicago Booth Review
More from Chicago Booth
Your Privacy
We want to demonstrate our commitment to your privacy. Please review Chicago Booth's privacy notice, which provides information explaining how and why we collect particular information when you visit our website.