April 2020 will mark the 40th anniversary of Jay Herson’s paper “Patient Registration in a Cooperative Oncology Group”. That paper, published in the first volume of the journal Controlled Clinical Trials, dealt with sources of bias and randomization practices that existed at the time. Looking back, it is surprising to see how much has changed since 1980. Whilst randomization remains a key design feature for confirmatory trials aimed at new drug approval, many innoative non-randomized study designs have been proposed and used in recent years.
Randomized clinical trials have often been criticized for being conducted in conditions patient that do not reflect the real world. There is an acute need for more and larger pragmatic trials that reflect clinical practice, but because these trials are conducted with tight budget and limited control, they are often considered as yielding unreliable or even misleading evidence of treatment benefits and harms. In addition, the recent advent of precision medicine has revived controversies about the role of randomization as a core feature in clinical trial design, for testing treatments that clearly have better efficacy than the current standards of care.
Finally, the availability of large databases of real world evidence has also led to questioning the necessity to carry out expensive and lengthy randomized clinical trials, if sophisticated statistical techniques (e.g. based on causal inference) could inform the choice between therapeutic options.
This symposium will address the sources of bias that continue to threaten sound clinical research now and in the foreseeable future. Past achievements and the current landscape will be described along with realistic proposals to move the field forward.
The final programme will be published soon.
Magritte Museum, Brussels