Designing and Analyzing Clinical Trials for Personalized Medicine

This group has been sunsetted As of

Parent DTF: Methods and Processes Domain Task Force


In medical research, the treatment paradigm is shifting toward personalized medicine, where the goal is to match patients to the treatments most likely to deliver benefit. Treatment effects in various subpopulations may provide some information about treatment effects in other subpopulations. This working group will develop methodology and best practice guidelines for clinical trial design where treatments are being investigated or tested in several groups of patients.

  • This working group will explore and develop statistical models that are from both frequentist and Bayesian in nature. Our development and testing will be done by simulating trials where there are a number of related treatments being tested in parallel. A specific example might be the exploration of the same drug but tested in different ways – for instance in different (sub) populations or different indications. Trials in personalized medicine will be explored where there might be a biomarker, or trials of a treatment where different populations (such as different risk factors) might have a different response to the treatment, or the same compound is being tested in different but related indications, for instance different pain indications or inflammations. Using simulations and numerous options for the statistical analysis we will explore and develop several possibilities, some of them straight forward, some of them novel. The Working Group will explore cases where the endpoint is continuous, dichotomous or time to event. Interim analyses will be specified at fixed intervals by time, the number of human subjects recruited or the number of events observed. The working group will explore cases where interim analyses depend on individual groups or the whole study can be stopped for success or futility. The working group will explore cases where the estimation of the response on the control and/or arms can be augmented using a hierarchical model to borrow from data from previous studies.

Key Deliverables:

  • Working Group will produce and disseminate a white paper that will serve as guidelines for designing and analyzing clinical trials for personalized medicine and testing heterogeneity of treatment effects.
  • PhD students will support the Working Group by providing programming assistance as part of their training within the KUMC department of biostatistics.

This group is sunsetted.