Fig. 2

Mean within-diagnosis distance compared to mean between-diagnosis. The within-diagnosis means offer information on patient-to-patient variability within a diagnosis; between-diagnosis means offers information on the degree of separation between patients with one diagnosis from patients of another diagnosis. Mean inter-patient distances were highest for cosine and Jaccard metrics, lowest for augmented bipartite and augmented cosine metrics (post hoc Bonferroni test, p < .05). Within-diagnosis mean distances are lower than between-diagnosis mean distances for all metrics (post hoc Bonferroni test, p < .05)