From study to target populations: addressing external validity through the Lens of Transportability and Generalisability

Description

Randomised and observational studies each have unique strengths and limitations when estimating causal effects for a target population. While estimates derived from randomised controlled trials may possess internal validity, they often fail to adequately represent the target population due to restrictive inclusion criteria. Observational data are more likely to represent the target population, increasing the likelihood of external validity. However, evidence from observational studies is prone to potential bias due to unmeasured confounding. A framework has recently been proposed to address the external validity bias through evaluating generalisability and transportability. This session will explore the current developments.

[The topic of my talk was invited and its title was "Measurement invariance in outcomes: A problem for transportability?]
[Invited presentations for the Medical Statistics Section.]

Period	3 Sept 2025
Event title	Royal Statistical Society International Conference 2025
Event type	Conference
Location	Edinburgh, United KingdomShow on map