Personalized Trials Show Potential to Transform Medicine

Every time you see a doctor, they collect information on your medical chart, giving them a longitudinal sketch of your individual health status. But what if the data collected during these clinical encounters was treated like its own scientific study to provide insights to guide treatment and advance knowledge?

Ken Cheung, professor of biostatistics at Columbia Mailman School, is a guest editor of a special fall issue of Harvard Data Science Review that explores the potential of these “personalized trials.” Previously known as “n-of-1 trials,” the broad use of single-subject studies could serve as the cornerstone of personalized medicine. The issue features contributions from several Columbia faculty, including Cheung. (A full list of guest editors and authors is available online.)

“The time has come to introduce the scientific method to the clinical encounter when the efficacy of a therapy is in question,” Cheung and co-authors write in the issue’s opening essay. Upsides of personalized trials include improved management of health conditions, such as through the more precise use of therapies and avoidance of unhelpful medications. The result, they write, “could revolutionize the way [clinicians] interact with patients and set the stage for a truly transformative approach to precision therapeutics.”

In the future, scientists and clinicians could draw on a database of similar personalized trials to generate knowledge and clinical insights. While realizing this future would require overcoming many logistical and cultural barriers, doing so presents a unique opportunity “for systematic, rather than subjective, data collection,” the authors write.

One barrier has been the term “n-of-1 trials,” which research by the authors shows is disliked by patients and clinicians alike. Patients say the term makes them feel like objectified and without agency. Focus groups settled on the more user-friendly term “personalized trials.” (Funding to support research in the opening essay was provided by the National Institutes of Health through grants LM012836 and AG063786.)

Another barrier is the heavy use of randomized controlled trials, seen as the “gold standard” in clinical trials—especially those involving drug development. These trials rely on unverifiable assumptions about whether all people in a study group respond to a given therapy in the same way. In fact, it is common knowledge that patients respond to therapies in different ways based on their medical history and other factors. “Rather than asking why individuals respond differently to the same treatment or need a different amount of the same treatment, a patient-centric approach asks what treatment works best for one particular individual,” the authors write.

Articles in the special issue cover personalized trial methods, design considerations, and their promise for improving health care; data analytic and statistical; and best practices for reporting, ethics, conduct, and payment of personalized trials. The issue concludes with several examples of personalized trials.

Among Columbia-affiliated authors, Hiroshi Mitsumoto, the Wesley J. Howe Professor of Neurology at Columbia University at The Neurological Institute of New York and NewYork-Presbyterian Hospital/Columbia University Medical Center, is co-author with Ken Cheung of an article on evaluating personalized trials in rare diseases. Naihua Duan, Professor of Biostatistics (in Psychiatry) in the Department of Psychiatry at Columbia University Irving Medical Center is co-author of articles on personalized trials as a promising paradigm for individualized healthcare and on the implementation of personalized trials in research and practice.

“Each of the contributors to this special issue makes a strong case for the broad use of personalized trials,” says Cheung. “We see a bright future ahead.”