Information Technology Reference
In-Depth Information
Table 15. Mortality by number of diagnosis codes for cardiovascular patients
Table of Number of Codes by DIED
Number of Codes
DIED(Died during h
ospitalization)
Total
Frequency
Row Pct
Col Pct
0
1
2
3
100.00
0.21
0
0.00
0.00
3
4
3
100.00
0.21
0
0.00
0.00
3
6
14
100.00
0.97
0
0.00
0.00
14
8
36
100.00
2.50
0
0.00
0.00
36
10
68
100.00
4.73
0
0.00
0.00
68
11
110
100.00
7.65
0
0.00
0.00
110
12
117
99.15
8.14
1
0.85
3.57
118
13
178
99.44
12.38
1
0.56
3.57
179
14
408
97.61
28.37
10
2.39
35.71
418
15
501
96.91
34.84
16
3.09
57.14
517
Total
1438
28
1466
that the error rate is extremely high. An early study examined the lack of accuracy in relationship to
lost revenue to treat the infection. Now that providers are reducing or eliminating reimbursements for
such treatment, the reliability must be questioned even more. (Massanari, Wilkerson, Streed, & Walter
J Hierholzer, 1987) There will be every incentive to under-report such infections.
A study performed almost fifteen years after that of Massanari, et. al. indicates that accuracy is still
an issue in the reporting of nosocomial infection. (Vegni et al., 2004) Both studies found that just slightly
more than 50% of nosocomial infections were accurately reported. A specific ICD9 code does not ex-
ist for a diagnosis of nosocomial infection. The question, then, is how can a nosocomial infection be
identified through billing data? Table 18 reproduces the diagnosis clusters from Chapter 8, modifying
the last column to identify the potential for a nosocomial infection. If it is not nosocomial, it is com-
munity acquired. However, this information is provided by a domain expert as a nosocomial infection
Search WWH ::
Custom Search