Application of A Causal Discovery Model to Study The Effect of Iron Supplementation in Children with Iron Deficiency Anemia

Abstract

Most clinical studies are descriptive, measuring different parameters that are associated with certain treatment effects, while causal relations cannot be confirmed. Computational models aid researchers to make predictions of causality and help to focus on the most relevant part of the data. This study used a computational model to find a causal link between iron supplementation effectiveness, RTI, and systemic inflammation parameters, and gut microbiome profiles. We used a causal discovery algorithm on a randomized controlled trial dataset (n=72) of 6 month-old infants from Kenya. We also used correlation and partial correlation to determine the causal effect of any causal link. We found that (1) expression of the Transferrin Receptor (TfR) has a positive causal link with the serum TfR-Ferritin ratio, whereasserumFerritin levels have a negative causal link to TfR expression. (2) C-Reactive Protein (CRP) together with IL-8 and IL-1B have a positive causal relation with IL-6. (3) No causal link between iron supplementation and gut microbiome profile. The first and second result is in accordance with the currentbiological research findings. While the third result shows no causality model, the skeleton might give information for future studies on understanding the gut microbiome profile. Computer modeling helped to uncover causality between clinical parameters in iron deficiency anemia children with iron-micronutrient supplementation. This could lead to more focused studies to better understand the iron supplementation practice as well as the biological mechanism of RTI, gut microbiome alteration, and iron supplementation.