DIA 2020 Global Annual Meeting

A synthetic control arm (SCA) has been suggested as possibly advantageous as an external control to aid in the interpretation of clinical trials where a randomized concurrent control is clinically unethical, practically infeasible, or unacceptable to patients. Because the SCA is made up of patient level data from the real world or from historical clinical trials, there is opportunity to match the patient composition of the SCA to that of the experimentally treated group in baseline demographics and disease specific characteristics. This may be a considerable advantage over traditional external controls such as benchmarking with static results from medical literature or clinical intuition with populations that may not be sufficiently similar to the experimentally treated group. Speakers will provide a brief introduction to SCA and will discuss a variety of statistical approaches for the creation and use of SCAs. Discussion may include the use of clinical trials data and real-world data for construction of SCAs, various statistical matching and weighting methods for creation of the SCA, the possible impact of unobserved or unavailable historical information on the balancing process and the treatment effect, and others.