Figure 3.

Eigenvector analysis of clinical features. Collected clinical features from Redcap data were assessed for all patients with less than 30% of data missing (n = 375). Missing data was imputed in R using multiple imputation by chained equations. Contribution to data variation is highlighted by the color legend and distance of a vector from the origin. Relationship of features are represented by directionality with direct relationships represented by same directionality, inverse relationship by opposite directionality, and no relationship as orthogonal directionality. Omitted features did not have high enough variance contribution to visualize.