Benchmark dose contours for bivariate exposures : application to maternal drinking patterns and childhood cognition
Conference
65th ISI World Statistics Congress 2025
Format: CPS Abstract - WSC 2025
Keywords: benchmarking, child-health, non-parametric, risk_prediction
Session: CPS 16 - Experimental Design and Clinical Trials
Monday 6 October 4 p.m. - 5 p.m. (Europe/Amsterdam)
Abstract
Benchmark dose analysis is a statistical procedure for estimating an exposure level that is associated with a specified increased risk of an adverse health outcome. We consider the challenge of benchmark dose analysis in the context of a study aiming to relate patterns of drinking behaviour in expectant mothers (e.g. proportion of days spent drinking, and the amount of alcohol consumed/drinking day) to childhood cognition . We propose a flexible framework for benchmark analysis that allows a more nuanced assessment of risks associated with multi-dimensional exposure variables. The method entails fitting a generalized additive model (GAM) for the effect of the exposures on cognition while adjusting for potential confounders via suitable propensity scores to flexibly estimate the dose-response surface. From this model we obtain a benchmark dose contour that relates the two continuous exposure variables to an outcome. We illustrate our method using data assembled from six U.S. cohort studies that measured maternal reports of alcohol use during pregnancy, and measurements of cognitive function in their offspring. While our results provide important scientific insights regarding adverse effects associated with various prenatal drinking profiles, more generally the method may be useful in a broad range of settings involving exposures or mixtures of exposures that can be measured in several dimensions.