Precision Medicine in Clinical Trials

Most clinical trials conducted are designed for the average patient. The FDA’s one-size-fits-all approach may be successful for most patients but not for all. A drug may work for one subject but not for another. Whether the patient will be reactive to a specific treatment also depends on his or her genes, lifestyle and environment. With advancements in technology and research, it is possible to determine specific characteristics that make a patient reactive to a given drug.

When a clinical trial results in a failure, it can cost incredible amounts of money to a pharmaceutical company. It is quite possible that a drug can be effective in a subset of a population for a particular indication, and trials conducted on the broader population can be too underpowered to prove the efficacy of the drug. Hence, it becomes significant to include particular groups of patients in trials. For example, BIO analyzed the probability of the success rate of clinical trials with and without the selection of Biomarkers. The results were evident that the success rate of clinical trials increase if the study population is selected based on biomarker results (as shown in the bar chart below).


The following are the key challenges in clinical trials with the introduction of the concept of personalized medication:

Finding discrete groups and subgroups of patients: It becomes difficult to find subgroups of patients in populations where a new drug is supposed to be most effective. It needs prior data on the drug as well as the disease. With the movement towards data sharing within the industry, it will become easier to get the data required for finding such groups.

Complex study Designs: Sometimes, to investigate the effective group of patients or effective dose of a drug, clinical trials need to follow a complex study design. Complex study designs are not easy to conduct from a logistic, standards, and analysis perspective. The following are a few complex study designs followed in recent clinical trials:

  • Longitudinal cohort studies
  • Basket trials
  • Umbrella trials
  • N-of-1 study design
  • Adaptive Trials

Statistics and Programming: There is not much available to refer to, for analysis and management of data of these complex designs. There is constant need to develop standards and availability of bio-informaticians to plan and analyze data from these studies.