Use of biomarkers in epidemiologic studies: minimizing the influence of measurement error in the study design and analysis

Abstract

The inclusion of biomarkers measured on the continuous scale, such as endogenous sex hormones or antioxidant levels, has become common in epidemiologic studies, and introduces additional sources of error that are specific to biomarkers. This includes error associated with specimen collection, processing, and storage; laboratory error (both within and between batch); and variability in the biomarker levels over time within an individual. In this review, we discuss and recommend study design and analytic strategies to deal with these sources of measurement error. In particular we describe methods to prevent or minimize some sources of error through appropriate sample collection and storage, communication with the laboratory, proper batching of samples, and participant matching. We also discuss how to quantify error related to biomarkers, focusing on issues of quality control, pilot studies, and how to measure within-person stability over time. Further, we discuss analytic issues for dealing with laboratory and within-person variability. Finally we recommend that journals standardize the reporting of biomarker assays in scientific manuscripts.