Date created: | Last Updated:

Category: Project

Description: Non-detectable (ND) and outlying concentration values (OV) are a common challenge in biomarker research. However, general best practices on how to aptly deal with the affected cases are still missing. The high methodological heterogeneity in biomarker-oriented research, and the statistical bias in some of the applied methods may compromise the robustness, comparability, and generalizability of research findings. Here, We consider a model that views ND and OV as censored data, for instance due to measurement error cutoffs. In simulations with lognormal distributed data, we compare the performance of six methods, ranging from simple commonly used to more sophisticated imputation procedures, in four scenarios with varying patterns of censored values as well as for a broad range of cutoffs. We thereby introduce a novel algorithm using imputation from the censored intervals of a fitted lognormal distribution.