Individual-level attribution modeling reveals that real-world semaglutide effectiveness is not a fixed property of the molecule, but an interaction between pharmacological exposure and modifiable care and support conditions, carrying implications for real-world evidence interpretation and clinical development.