The Importance of Accounting for Time-Related Bias in Single-Arm Trials with Historical Controls (Part 2)
Q&A with Samy Suissa, PhD
In our previous blog post, we sat down with Dr. Samy Suissa, the co-founder and principal investigator of the Canadian Network for Observational Drug Effect Studies, to discuss the main themes from his recent journal article in Epidemiology. Namely, he argues that while single-arm trials with historical controls are gaining recognition in clinical trial design, pharma companies need to understand the potential biases that could exist before conducting these types of trials and seeking regulatory approval.
In part two of our Q&A, Dr. Suissa weighs in on how these issues can be avoided and what the evolution of real-world data might mean for the future of clinical trials.
Q: How do you design single-arm trials to avoid these issues?
A: You use the principles of epidemiology. Of course, patients must be matched on their clinical and other relevant characteristics – age, comorbidity, disease history, etc – possibly using propensity scores. As I explain in my paper, this is important to address confounding, but not enough to address selection bias. For this, what is important is to design the study by matching the patients at the same stage of disease and selecting the correct “matched” time point at which they would have received treatment had they been in the study arm. Thus, for example, if a patient received the new study treatment after two failures, they should be matched with a patient in the historical control group who received a standard treatment after two failures as well. There are tools in epidemiology that can help design a single-arm trial with historical controls that make sure the time element is properly matched.
Q: What kind of reaction have you received, from your paper?
A: Some pharmaceutical companies that are preparing to use single-arm trials for regulatory approval have reached out to me. After reading my paper, they will be incorporating the approach I proposed to minimize any time-related biases.
Q: Is there a way to bridge the gap between RCT’s and single-arm trials?
A: I think epidemiology is a good place to start. There are the trialists who believe only in randomized trials and the observational researchers who believe strongly in the value of observational research. My paper talks to the trialists and argues that there are cutting-edge methods in observational and epidemiological research to deal with some of the challenges that exist in these randomized trials. I’d like to think paper is a nice bridge between the two solitudes.
Q: How do you see the growth of real-world data impacting clinical trials going forward?
A: Clinical trials are fundamental to the advancement of knowledge on drug effectiveness. With the increased access and sophistication of real-world EMR and observational databases a hybrid is developing, namely the “pragmatic” trial, where patients are randomized within an EMR, with little control other than the usual physician follow-up. Researchers track progression and outcomes directly via from the EMR records. This hybrid approach has some methodological challenges but will make trials much more flexible. I believe that randomized pragmatic trials can link the strength of RCTs with the needs of single-arm trials and play a role in the future of clinical research.
Click here to learn how Panalgo’s IHD Analytics platform can support your RWE teams and deliver richer insights, faster.
Dr. Samy Suissa is a Co-Founder and Principal Investigator of the Canadian Network for Observational Drug Effect Studies (CNODES). He is Director of the Centre for Clinical Epidemiology, Lady Davis Institute for Medical Research at the Jewish General Hospital and Professor, Departments of Epidemiology and Biostatistics and of Medicine, McGill University, in Montreal, Canada. Dr. Suissa also heads the McGill Pharmacoepidemiology Research Unit. He was the founding Director of the Quebec Research Network on Medication Use. Dr. Suissa also sits on Panalgo’s Strategic Advisory Board.