Abstract
Health Technology Assessment (HTA) for reimbursement of all new cancer drugs in the European Union (EU) will be evaluated for all member states by a central European HTA starting in 2025. EU HTA guidelines for applicants under these new regulations put the focus on meta-analysis of aggregated randomized trial data and are in contrast with the growing number of cancer drug approvals by drug regulators that are based on single arm studies and the needs in the rapidly evolving field of oncological drug development. We advocate to broaden the methodological approaches for HTA by including observational data based causal inference methodology and target trial emulation into the assessments of comparative effectiveness. Causal inference estimates causal estimands, effect measures that reflect a population level effect in terms of contrasts of counterfactual outcomes in the same patients and are directly measured in the target population by modeling of hypothetical intervention. Target trial emulation allows with the use of causal inference to estimate causal effects by mimicking pragmatic trials that evolve apart from randomization like a trial. We illustrate the potential of causal inference for HTA and provide an introduction into causal inference methodology for health scientists involved in HTA.