PURPOSE: We will provide an overview of different research questions (estimands) in trials with imperfect adherence or treatment switching, related causal comparative effectiveness methods, and their (sub-optimal) use in health technology assessment (HTA). Three speakers will discuss the validity and acceptance of available methods in HTA and will consider whether these methods are currently being used sub-optimally. DESCRIPTION: HTA should investigate both the effect of actual and intended treatment to guide clinical and policy decisions. Our focus is on situations with treatment switching or adherence influenced by time-varying post-randomization factors such as progression or side effects. In these cases, traditional methods may fail and causal inference methods must be used. Recently, HTA agencies have accepted and recommended using causal methods, and a paradigm shift is taking place. Selection of the appropriate adjustment method has become crucial to inform decision-making on patient access to innovative and cost-effective treatments. However, adjustment methods are often used poorly and their full potential is not being realized. This workshop has four parts: • Uwe Siebert will describe post-randomization confounding/selection bias in trials, introduce causal inference concepts and adjustment methods, and present case examples of biases occurring from inappropriate analyses. • Nicholas Latimer will demonstrate that the full potential of methods used to adjust for treatment switching is not being realized: He will explain how the use of adjustment methods can be improved by discussing key assumptions and application choices, and how these relate to health economic models. • Amanda Adler, Chair of NICE Technology Appraisal Committee B, will discuss the challenges of using statistical methods to address treatment switching in the HTA process using examples from NICE. • We will actively engage participants in the discussion and use the ISPOR polling app to collect the audience’s view on discussed issues.