|
Abstract:
|
Pooling biomarker exposure data across multiple studies in a participant level meta-analysis enables more precise estimation of biomarker-disease associations. However, biomarker measurements often require calibration to a reference assay prior to pooling due to assay and laboratory variability across studies. This talk discusses the development of two broad categories of methods for assessing pooled and calibrated biomarker data. The first, named regression calibration, implements study-specific calibration models to estimate missing reference lab measurements. The second, named the repeated measures approach, accounts for within-person and within-study measurement correlation to perform calibration. Each method is illustrated in an application evaluating the relationship between blood circulating vitamin D levels and stroke in a pooling project of cohort studies.
|
