OBJECTIVES: Innovative drugs approved based on single-arm trials should be compared with external controls in health technology assessments. This study investigates a methodology that constructs external controls with better comparability by using real-world data (RWD).

METHODS: To compare with the single-arm trial of tisagenlecleucel, we created four scenarios for external controls by conducting a literature review and analyzing the claims data: (1) systematic literature review on salvage chemotherapy, (2) unadjusted retrospective cohort of relapsed or refractory diffuse large B-cell lymphoma, (3) retrospective cohort with adjusting for the time-related bias through including all potential exposure sets, (4) matching-adjusted indirect comparison (MAIC) for the scenario 3 cohort. We mimicked the eligibility criteria and index period of the single-arm trial to construct a comparable cohort, and measured patient characteristics, median overall survival (OS), and progression-free survival (PFS).

RESULTS: The OS of tisagenlecleucel was reported as 11.9 months in the single-arm trial, whereas that of external control from the literature review reported 6.3 months (scenario 1). The OS from the unadjusted retrospective cohort was 4.89 months (scenario 2). In scenario 3, patients' previous treatment history became similar to that of the single-arm trial, and the OS decreased to 4.50 months including more progressed records. After matching baseline characteristics (scenario 4), treatment history became equal to that of single-arm trial, and the OS was reduced to 4.34 months. The PFS showed a similar tendency to decrease survivals as comparability improved.

CONCLUSIONS: The comparability with single-arm trial showed a difference in the comparative effectiveness. It is necessary to ensure the comparability of patients based on MAIC, and the pharmacoepidemiological design to adjust for time-related bias.