SW-1D - What are possible approaches when the active comparator new users design (ACNU) is not easily applicable?

Background: The active comparator, new user design (ACNU) has been the gold-standard to estimate treatment effects for non-interventional studies due to the prevention by design of several types of bias, including depletion of susceptible and immortal time bias. In some instances, this design is not easily applicable. This can be due to challenges in identifying a suitable comparator, complex standard of care, or difficulties defining the index date of a treatment or a comparator. In these cases, alternative approaches could ensure the generation of reliable evidence.

Objectives: In this symposium, we aim to highlight circumstances where the ACNU design may not be directly applicable to answer the most relevant CER research question and will present alternative approaches to generate valid and reliable evidence. We will highlight advantages and potential shortcomings of these approaches and would conclude with recommendations and discussion on the optimal scenarios to adopt individual methods. This symposium will be particularly valuable for participants with experience in designing studies on treatment effects (e.g., effectiveness or safety for specific medications) who may have faced similar challenges. It will help the audience navigate various alternative options to the ACNU design and understand the benefits and risks of the individual methods presented.

Description: We will start by introducing the ACNU design for non-interventional studies, its advantages, and situations where this design cannot easily be applied (Elisabetta Patorno).
We will discuss the prevalent new user design as an extension of the ACNU allowing for the inclusion of initiators of the study drug who were previously exposed to the comparator treatment (Kristian Filion); then present clone censor weighting as a strategy to avoid the introduction of immortal time derived from misclassification if individuals into treatment groups (Xabier García de Albéniz). We will highlight high dimensional propensity scores methods as an alternative to prevalent new users’ design when confounder identification is challenging in a new therapeutic area (Tianze Jiao). Finally, we will discuss cases and potential approaches where a comparator cannot be easily identified by any possible approaches through examples in rare diseases (Anouk Déruaz Luyet). The speakers will highlight key aspects of each approach, including strengths in controlling bias and confounding, limitations, challenges in implementation, and suitability of different data sources for the specific purpose. We will conclude this symposium with a discussion and addressing questions from the audience.