Greetings
9/18/2018
This is my introductory blog for the NHLBI Innovative Clinical Trials Resource Educational component. In this space I will be writing about innovative trials, upcoming events, controversies, and anything else of interest in advancing clinical trial design. I would love if this space was used extensively for questions and debate, so questions and comments are most welcome as we get going.
For a very brief introduction to me, I’m a bit late coming into clinical trials. I have a PhD from Carnegie Mellon where I was trained as a computational Bayesian statistician, after which I worked on statistical methodology and biological applications. In 2010, I joined Berry Consultants and began learning clinical trials in earnest. Because of my computational background, I began by programming clinical trial simulations and watching the design of a large number of clinical trials. Eventually I began designing trials myself and I currently co-lead the consulting unit at Berry Consultants with my NHLBI ICTR counterpart Kristine Broglio. While I design trials in all areas with a variety of methodologies, I have particular expertise in methodologies relying on sharing information, such as basket trials (where information is shared across indication) and the utilization of historical data in a trial.
This week, we will hold our introductory webinar “An introduction to innovative trial design”. We’ve tried to organize this webinar by connecting an array of innovative techniques to the standard gold standard randomized clinical trial, illustrating what features of the gold standard trial are improved/adjusted/broken by each innovation, and how to weigh the costs and benefits of adopting a specific innovation. After this, you may not know much more about the details of implementing these designs, but you should have a good flavor of what is available, and the decision process involved in pursuing an innovative trial. In future webinars we will go into more details on a subset of the trial designs considered.
Our current standard clinical trial dates to the Streptomycin trial in Tuberculosis from 1946.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2091872/
This 1946 trial adds randomization to previously existing components to produce a recognizable modern trial.
The components of a gold standard trial admirably perform two goals – eliminating bias and providing conclusive, interpretable results. At the same time, the gold standard is resource intensive and may require us to ask narrower questions than we prefer. This creates a conflict for developing and adopting innovative trials. Much as medical research must tread very lightly to avoid harm to research patients, we desire to limit any “experimenting with experimentation” to avoid compromising our research agenda. No one wants an effective therapy derailed by a poor experiment. And yet experimentation has advanced since 1946 in terms of flexible sample sizes, incorporation of external information into a trial, enrichment designs, dose ranging, platform trials, adaptive arm dropping and adaptive randomization, and many other innovations. Experimental methods that were well understood prior to 1946, such as factorial designs, are not sufficiently adopted in current medical research.
Note that innovative trials are not just aimed at being smaller and cheaper, often more importantly innovations are aimed at asking broader, better questions. Such trials are often larger but provide a significantly increase in the information obtained while trying to minimize the marginal cost of that information, for example by allowing us to consider multiple populations instead of being forced to guess our ideal patient population under uncertainty. Enrichment designs, for example, allows you to consider multiple patient populations. The size of the trial is typically increased, and you need safeguards to account for incorporating subgroups into the analysis, but you will get a more robust answer at the end of the trial.
After the webinar, please feel free to ask further questions or make comments and we would be happy to expand on any of the topics discussed. We hope this public website and this blog will allow us to continue discussion on these topics.
– Kert Viele
