PURPOSE: RCTs, especially those undertaken in oncology, often allow for patients to switch study treatments, particularly from under-performing ones. Statistical methodology (e.g. IPCW, RPSFTM) can be used to try to account for this treatment flexibility, with the main aim of reducing potential bias in estimates of comparative efficacy. However, treatment switching also affects many, if not most, observational studies. A related issue in these real-world studies is one where patients initiate treatment after a considerable delay following their initial diagnosis (e.g. due to lack of availability of the drug). In fact, whilst in HTA statistical methods are usually used to adjust for treatment switching in RCTs, the methods used mostly originated in the observational data literature. The aim of this workshop is to outline and discuss current approaches to handling treatment delays/switching in observational studies when estimates of comparative effectiveness are sought. DESCRIPTION: The challenges of interpreting comparative effectiveness in the presence of treatment delays and switching will be outlined. A broad overview of the analytical methodology that aims to deal with these challenges will be provided, using various case studies across a range of therapeutic areas. The workshop will also include discussion on the strengths and weaknesses of these approaches, from a multi-stakeholder perspective.