Text Mining and Patient Severity Clusters - Text Mining Techniques for Healthcare Provider Quality Determination - page 298

pared to some of the other clusters. Note that diseases of the esophagus are fairly common across the
clusters. The mortality is given in Table 6.
In this second example, 4<8<2<5<10<7<1<6<9<3. Since patients in cluster 3 have some serious
and high risk infections, it is reasonable to conclude that they also have the highest risk of mortality,
followed by the patients in Cluster 9 with severe heart problems. Patients with problems of drug abuse
have the least risk of mortality, as do those patients chiefly with problems in the female reproductive
system. Roughly in the center, with 4% mortality, are the patients with problems in the abdomen and
the digestive system.
Table 7 shows the relationship to length of stay and total charges. For length of stay, the ordering
is 8<5<2<1<9<10<7<3<4<6. For total charges, the ordering is 4<8<2<1<5<7<10<3<9<6. Again, note

Table 6. Mortality by revised clusters

Table of allclusters by DIED
allclusters DIED Total

Frequency
Row Pct
Col Pct

0 1

1 34495

97.23
11.63

981
2.77
12.08

35476

2 30432

99.51
10.26

151
0.49
1.86

30583

3

59074
94.84
19.91

3216
5.16
39.61

62290

4 9822

99.92
3.31

8
0.08
0.10

9830

5 28354

99.05
9.56

271
0.95
3.34

28625

6 7723

96.95
2.60

243
3.05
2.99

7966

7

32750
98.75
11.04

415
1.25
5.11

33165

8 9327

99.89
3.14

10
0.11
0.12

9337

9 52697

95.51
17.76

2478
4.49
30.52

55175

10 31971

98.93
10.78

347
1.07
4.27

32318

Total 296645 8120 304765

Frequency Missing = 151