Creating Historical Controls Using Data from a Previous Line of Treatment - Two Non-standard Approaches

Where medical interventions are licensed based on only uncontrolled study data (for example a single-arm trial), a common approach for estimating the incremental benefit is to compare the treatment to a 'historical control'; data collected from patients who did not receive the intervention. We illustrate with motivating examples two methods for the creation of historical controls where disease progression and overall survival are typically the key clinically meaningful endpoints. The first method utilises information routinely collected in a clinical trial programme: patients' time to disease progression on their previous line of treatment against which outcomes can be compared. The second uses published clinical outcomes for the prior line of treatment which can be extrapolated to estimate outcomes at the next line. As examples we use two pharmaceuticals licensed on the basis of uncontrolled clinical studies - idelalisib for double-refractory follicular lymphoma and ofatumumab for double-refractory chronic lymphocytic leukemia. Although subject to limitations that should be considered on a case-by-case basis, the methods may be appropriate when trying to quantify the clinical benefit of treatment based on limited and uncontrolled trial data. As a result, the methods can be used to inform health technology adoption decisions.