Research Article
Disentangling the Association of Hydroxychloroquine Treatment with Mortality in Covid-19 Hospitalized Patients through Hierarchical Clustering
Figure 2
Characteristics of sample according to the two clusters identified.Comparison of the continuous variables used for hierarchical clustering—including (a) age (years), (b) BMI (Kg/m2), (c) eGFR (mL/min/1.73 m2), and (d) C-reactive protein plasma levels (mg/L, log-scale) between the two clusters of Covid-19 patients identified, namely, the low (green) and the high risk (red) cluster. Here, these variables are represented through boxplots, with boxes showing the interquartile ranges (IQR = Q1-Q3), continuous lines showing the whole distribution range from Q1 – 1.5∗IQR through Q3 + 1.5∗IQR, and dots showing more extreme values in the dataset.
(a) |
(b) |
(c) |
(d) |