Overview
Real-world evidence is increasingly used to inform healthcare decisions across countries, but differences in patient populations and healthcare systems can make it difficult to apply findings from one setting to another. Transportability analyses aim to address this by predicting outcomes in new populations using models developed elsewhere.
In this study, researchers evaluated how well a survival model developed using US data performed in an independent sample of patients with multiple myeloma. Using the Flatiron Health Research Database, they trained a model to predict overall survival and tested its performance in a separate US cohort. The model demonstrated strong accuracy, with close alignment between predicted and observed survival outcomes.
Why this matters
These findings highlight the importance of validating models before applying them to new populations. By ensuring strong model performance, researchers can more confidently use real-world evidence across countries, supporting more informed global healthcare decision-making.