|
Abstract:
|
Although limit of detection (LOD) issues are ubiquitous in exposure assessment, replacing values below the LOD with a constant (e.g. LOD/2) continues to be the standard practice in environmental epidemiology. We consider the situation where, due to the practical logistics of data accrual, sampling, and resource constraints, exposure data are analyzed in multiple batches where the LODs and the proportions of censored observations differ across batches. Compounding this problem is the potential for non-random assignment of samples to each batch, often driven by enrollment patterns and biosample storage. We propose a likelihood-based multiple imputation strategy to impute observations that are below the LOD while simultaneously accounting for differential batch assignment. Our simulation study shows that the proposed method has superior estimation properties (i.e., bias, coverage, statistical efficiency) compared to standard alternatives, provided that underlying distributional assumptions are satisfied. We illustrate our method by analyzing data from a cohort study in Puerto Rico that is investigating the relationship between endocrine disruptor exposures and preterm birth.
|
