Overview: This repository contains the code and data used to assess the transportability of a cardiovascular disease (CVD) prediction model from the Framingham Heart Study to the NHANES dataset. We followed the methodology outlined in the reference paper "Transporting a Prediction Model for Use in a New Target Population" (doi: 10.1093/aje/kwac128).
Key Findings: Our analysis revealed promising results, indicating that the CVD prediction model can be successfully transported to the NHANES target population. It demonstrated high predictive accuracy in both male and female subsets of the NHANES data, highlighting its potential for real-world applicability. Moreover, our gender-specific analysis underscored the importance of considering sex-based differences in CVD risk factors, providing valuable insights for model customization.
Limitations: While our findings are promising, it's important to acknowledge certain limitations. The assumption of source population representativeness, simplifications in the simulation study, and the reliance on a single evaluation metric (Brier Score) should be considered.