Equivalence of regression calibration methods in main study/external validation study designs