Text Mining and Patient Severity Clusters - Text Mining Techniques for Healthcare Provider Quality Determination - page 305

Figure 11. Charlson index by procedure cluster for index values 6-10
Figure 12. Decision tree to predict Charlson level
Figure 12 shows the relationship of the diagnosis clusters to each of the codes that define the Charlson
Index; Figures 13, 14, and 15 show the procedure clusters.
It is not surprising that a much higher proportion of patients with liver disease are in cluster 4, which
is focused on alcohol abuse as such abuse does damage the liver. An even higher proportion of patients
with severe liver disease are in cluster 4. In addition, both acute myocardial infarction and congestive
heart failure are concentrated in cluster 9, which tends to identify with heart disease. One of the reasons
that the Charlson Index conditions are scattered across the clusters is because many are in fact related.
Many patients with renal disease also have heart disease, and conversely. Also, a higher proportion of