Toward improved analysis of concentration data: Embracing nondetects

Abstract

Various statistical tests on concentration data serve to support decision-making regarding characterization and monitoring of contaminated media, assessing exposure to a chemical, and quantifying the associated risks. However, the routine statistical protocols cannot be directly applied because of challenges arising from nondetects or left-censored observations, which are concentration measurements below the detection limit of measuring instruments. Despite the existence of techniques based on survival analysis that can adjust for nondetects, these are seldom taken into account properly. A comprehensive review of the literature showed that managing policies regarding analysis of censored data do not always agree and that guidance from regulatory agencies may be outdated. Therefore, researchers and practitioners commonly resort to the most convenient way of tackling the censored data problem by substituting nondetects with arbitrary constants prior to data analysis, although this is generally regarded as a bias-prone approach. Hoping to improve the interpretation of concentration data, the present article aims to familiarize researchers in different disciplines with the significance of left-censored observations and provides theoretical and computational recommendations (under both frequentist and Bayesian frameworks) for adequate analysis of censored data. In particular, the present article synthesizes key findings from previous research with respect to 3 noteworthy aspects of inferential statistics: estimation of descriptive statistics, hypothesis testing, and regression analysis.

Peripheral tryptophan, serotonin, kynurenine, and their metabolites in major depression: A case–control study

“…Nonparametric tests were used according to imputation strategy and metabolite distributions. 28 Step 1: Nonparametric Wilcoxon rank-sum tests were used to compare metabolite levels between patients and HC.…”

“…Median and IQR are not sensible to imputation methods if values below the LLOQ do not exceed 50% of median (all studied metabolites) and 25% of IQR (all studied metabolites except xanthurenic acid and 5‐hydroxytryptophan). Nonparametric tests were used according to imputation strategy and metabolite distributions …”

Prospects for finding Junge variability-lifetime relationships for micropollutants in the Danube river

“…The relative standard deviation s/m of concentration measurements in the Danube from JDS3-CW, biodegradation half-lives in the Danube estimated from our literature review and physicochemical properties of these nine substances are summarized in Tables S1 and S6 of the ESI. † A single arithmetic mean, standard deviation and corresponding s/m, were calculated for the nine micropollutants using concentration measurements above the LOQ combined with imputed data for measurements reported below the LOQ 28,29 to avoid bias in the estimation of means and standard deviations. A detailed example of the method for the imputation can be found in the ESI.…”

Mild To Severe Traumatic Brain Injury In Children: Altered Cytokines Reflect Severity

“…Where the sample was below the lower limit of detection the value of the lower limit of detection for that assay was substituted. [23] No samples reached an upper limit of detection. Statistics Descriptive statistics were used to describe the demographics of the cohort.…”