Title: The case for specifying the "ideal" target trial

Abstract: The target trial is an increasingly popular conceptual device for guiding the design and analysis of observational studies that seek to perform causal inference. As tends to occur with concepts like this, there is variability in how certain aspects of the approach are understood, which may lead to potentially consequential differences in how the approach is taught, implemented, and interpreted in practice. In this paper, we consider two of these aspects: how the target trial should be specified, and relatedly, how the target trial fits within a formal causal inference framework. We first describe two challenges with what we call the "aligned" approach to target trial specification, which is common in evaluations of medical interventions using healthcare databases and seeks to specify a target trial that is closely aligned to the observational data so that all protocol components apart from randomization can be closely emulated. We then argue how these challenges can be circumvented by an approach that focusses on specifying the "ideal" target trial: a trial with certain idealized aspects that ensure it is relevant to answer the research question. In essence, this approach is applicable in a broader range of settings and enables systematic assessment of all potential sources of causal bias. Importantly, this view provides clarification for how the target trial approach fits within a formal causal inference framework.