Statistical Approaches to Study Exposome-Health Associations in the Context of Repeated Exposure Data: A Simulation Study
Résumé
The exposome concept aims to consider all environmental stressors simultaneously. dimension of the data and correlation that may exist between exposures lead various statistical challenges. Some methodological studies have provided insight regarding efficiency specific modeling approaches in context assessed once for each subject. However, few considered situation which are repeatedly. Here, we conduct a simulation study compare performance assess exposome-health associations multiple exposure variables. Different scenarios were tested, assuming different types numbers exposure-outcome causal relationships. An application using real collected within INMA mother-child cohort (Spain) is also presented. In experiment, methods showed varying across scenarios, making it challenging recommend one-size-fits-all strategy. Generally, such as sparse partial least-squares deletion-substitution-addition algorithm tended outperform other tested (ExWAS, Elastic-Net, DLNM, or sNPLS). Notably, number true predictors increased, declined. absence clearly superior approach underscores additional challenges posed by repeated data, presence more complex structures interdependencies variables, highlights careful consideration essential when selecting appropriate method. this regard, provide recommendations based on expected scenario. Given heightened risk reporting false positive negative applying these techniques advise interpreting results with caution, particularly compromised contexts those limited sample size.
Domaines
Sciences du Vivant [q-bio]
Fichier principal
warembourg-et-al-2023-statistical-approaches-to-study-exposome-health-associations-in-the-context-of-repeated-exposure.pdf (2.86 Mo)
Télécharger le fichier
Origine | Publication financée par une institution |
---|---|
Commentaire | UPDATED with corrections https://pubs.acs.org/doi/10.1021/acs.est.3c10626 |