Information Technology Reference
In-Depth Information
Table 17. Ranking of hospital quality using Charlson index
Hospital
Actual Mortality
Predictive Mortality
Difference
Rank
1
1.23
7.38
6.15
6
2
0
0
0
10
3
4.46
14.29
9.83
1
4
0.79
5.53
4.74
7
5
1.14
10.23
9.09
2
6
0.57
9.14
8.57
3
7
3.53
12.01
8.48
4
8
2.22
4.44
2.22
9
9
1.28
8.97
7.69
5
10
3.13
6.25
3.12
8
Figure 28. Predictive modeling for length of stay with APRDRG indices
we have discovered some “gaming” by some providers shifting patients into higher levels of severity in
order to increase the level the predicted mortality. This gaming indicates that one of the main problems
remains the lack of consistency in assigning patient conditions. A second problem with consistency is
assigning a level of severity to a patient when compared across the different indices. In the future, these
different indices should be reconciled.
dIscussIon
There appears to be just a few similarities between the severity level as defined by the Charlson Index
compared to the APRDRG indices of mortality and severity. Primarily, the Charlson Index restricts
itself to a few of the ICD9 codes while omitting all others. In contrast, the APRDRG uses a separate
 
Search WWH ::




Custom Search