Figure 4.

Decision tree analysis for predicting immune reconstitution inflammatory syndrome (IRIS). To create the conditional inference decision tree for predicting IRIS, we used 10 baseline biomarkers (body mass index, CD4 and CD8 T-cell counts, human immunodeficiency virus RNA, HGB, white blood cell count, platelets, glucose, D-dimer, and C-reactive protein) and site excluding patients who died without experiencing IRIS, using R the party package (version 3.5). Potential splits are included in the tree model if they met a Bonferroni-adjusted threshold for statistical significance. Ovals indicate a split in the prediction rule on a specific variable, along with the corresponding P value. Each square shows the percentage of observations within that branch that met the decisional criteria (in this case HGB value) followed by the proportion of those patients who experienced IRIS. The darkest shaded square indicates a higher proportion of patients with IRIS. The lightest shade shaded square indicates the lowest proportion of patients with IRIS. Abbreviation: HGB, hemoglobin.