Induction and impact of non-differential measurement error through use of regression calibration: Illustration in a real-data example of risk factors for white matter hyperintensity progression