Skip to main content
. 2017 Apr 24;19(3):385–408. doi: 10.1007/s10198-017-0891-9

Fig. 3.

Fig. 3

The tree adds external variables such as weekly median wage to the variables underlying model 1. Weekly median wage appears at the start of the tree as the predictively most significant variable, followed by the nurse-to-bed ratio and specialist. Other variables measuring the organisational feature of the process of stroke care occupy important positions in the tree, supporting the idea that organisational factors are among the most significant predictors of the clinical quality of care