The Charlson Comorbidity Index - Text Mining Techniques for Healthcare Provider Quality Determination - page 165

1: 74.0%
0: 26.0%
It is clear from the remaining part of the decision tree that the prediction is based largely upon the
age and the level of the Charlson Index.
IF 86.5 <= Age in years at admission
AND charlson < 0.5
THEN
NODE: 27
N: 3397
1: 69.0%
0: 31.0%
IF 2.5 <= charlson
AND 47.5 <= Age in years at admission < 70.5
THEN
NODE: 29
N: 10808
1: 74.6%
0: 25.4%
This next node is one of two where sex plays a role. If the patient is male, between 73.5 and 86.5
years of age with a 0 Charlson Index, then the mortality in the subgroup is almost 60%.
IF Indicator of sex EQUALS 0
AND 73.5 <= Age in years at admission < 86.5
AND charlson < 0.5
THEN
NODE: 42
N: 2915
1: 58.8%
0: 41.2%
IF 47.5 <= Age in years at admission < 57.5
AND 0.5 <= charlson < 2.5
THEN
NODE: 46
N: 5349
1: 43.1%
0: 56.9%
The next node is the second one where sex plays a role. If the patient is between 72.5 and 80.5 years
of age, is female, and has a Charlson Index greater than zero, then the subgroup has a 46% mortality
rate.