Historical control: Design, measures, and classic example

Historical control trials (HCTs) utilize retrospective data on a standard treatment, or observation alone, as a control. Indications for HCTs include disease states where no viable therapeutic alternative exists, and randomizing patients to observation or minimally effective therapy is unethical. While limitations arising from the inability to limiting confounders through randomization exist in HCTs, appropriate use of single-arm studies implementing historical controls has the advantage of treating all patients on trial with active therapy and spurring development and approval of novel therapies. In this chapter, we discuss the limitations of this trial approach, which can be effectively mitigated by carefully defining the disease space, limiting heterogeneity in patient and disease characteristics, and choosing clinically meaningful and reproducible endpoints. We apply these principles to a real-world example in urologic oncology, the BCG-unresponsive nonmuscle invasive bladder cancer space.