Risk Adjustment Based Upon Resource Utilization - Text Mining Techniques for Healthcare Provider Quality Determination - page 278

Table 16. Length of stay by hospital for cardiovascular surgery

Table of DSHOSPID by DS_LOS_Level
DSHOSPID DS_LOS_Level(Disease Staging: Length of Stay Level) Total

Frequency
Row Pct
Col Pct

3 4 5

1 25

7.69
15.53

284
87.38
24.03

16
4.92
13.01

325

2 2

18.18
1.24

8
72.73
0.68

1
9.09
0.81

11

3

11
9.82
6.83

89
79.46
7.53

12
10.71
9.76

112

4 24

9.49
14.91

211
83.40
17.85

18
7.11
14.63

253

5 5

5.68
3.11

72
81.82
6.09

11
12.50
8.94

88

6 20

11.43
12.42

140
80.00
11.84

15
8.57
12.20

175

7

42
14.84
26.09

217
76.68
18.36

24
8.48
19.51

283

8 4

8.89
2.48

37
82.22
3.13

4
8.89
3.25

45

9 19

24.36
11.80

54
69.23
4.57

5
6.41
4.07

78

10 9

9.38
5.59

70
72.92
5.92

17
17.71
13.82

96

Total

161 1182 123 1466

in very serious condition, and will require considerable resources to treat. However, with so many of
the patients in the same index level, it will be extremely difficult to use this measure to predict patient
outcomes, and to distinguish between providers in terms of quality.
We compare these same three indices to the ten hospitals that perform cardiovascular surgery (Table
16). Hospital #7 has the highest proportion in the most severe category. This reconfirms the results from
the APRDRG indices showing that #7 has the most severe patients while performing the most difficult
surgeries compared to the other hospitals. However, hospital #1 also has a high proportion in category 4
in spite of having less severe surgeries. Table 17 shows the mortality index. Note that 36% of the patients
in hospital #7 are in the highest mortality category; this is far higher than for any other hospital.
Table 18 shows the resource demand level. Note again that for the cardiovascular patients, there are