|
| |
|
|
|
| |
|
Data Collection Methods
CMS/JCAHO Process Quality
Measure Notes:
The 2005 Hospital Compare database available on the Centers for Medicare & Medicaid Services website was used for the CMS/JCAHO Process Quality Measures. This database was downloaded on September 22, 2006, and contains Iowa and national hospital performance measures for the time period January 2005 through December 2005.
AHRQ Quality Indicator Notes:
Iowa Rate computations: Iowa’s 2002-2005 State Inpatient Database (SID) datasets were used to compute AHRQ performance measures for each year individually, and for all four years combined. Each dataset was limited to include only acute inpatient cases.
2004 U.S. Rate computation: The 2004 Healthcare Cost and Utilization Project (HCUP) Nationwide Inpatient Database (NIS) was used to compute AHRQ performance measures for the year 2004. The NIS comprises all inpatient discharges from a 20% stratified sample of U.S. community hospitals in 37 participating states. This dataset includes Iowa.
AHRQ’s PSI software Version 3.0 Revision “a,” and IQI software Version 3.0, were used to generate estimates for both Iowa and U.S. rates.
For IQI risk adjustment the AHRQ-supplied APR-DRG Limited Version 20 Revision “a” was used for both Iowa and U.S. rates. For PSI risk adjustment, the HCUP Comorbidity software Version 3.1 was used for both Iowa and U.S. rates.
Computational considerations: In order to make similar comparisons, both the Iowa and U.S. data sets were limited to 11 total diagnosis codes - 9 diagnoses and 2 external cause/place of injury codes. A total of 6 procedure codes were used in the computations.
Within the Iowa SID, the variable DISP was used to create the disposition variable “DISP” used by the AHRQ software. Within the NIS, a combination of patient disposition variables DISPUB92 (UB-92 form coding) and DISPUNIFORM (uniform coding used beginning in 1988) were used to create the disposition variable “DISP” used by the AHRQ QI software. Because the NIS had a substantial amount of missing data within the DISPUB92 variable, those values that were missing in DISPUB92, but present in variable DISPUNIFORM, were coded to the variable DISP as follows. Thus, the amount of missing values within the NIS DISP variable was reduced to 1.27%. The Iowa SID has 0% missing values for the variable DISP.
| DISPUB92 |
DISPUNIFORM |
Coded as AHRQ Variable DISP |
| Missing |
Missing |
Missing |
| Missing |
A |
Missing |
| Missing |
1 (Routine) |
1 (Routine) |
| Missing |
2 (Transfer to short term hospital) |
2 (Short-term hospital) |
| Missing |
5 (Other transfers) |
Missing |
| Missing |
6 (Home health care) |
6 (Home health care) |
| Missing |
7 (Against medical advice) |
7 (Against medical advice) |
| Missing |
20 (Died in hospital) |
20 (Died in the hospital) |
| Missing |
99 (Discharged alive, destination unknown) |
Missing |
|
|
|
| |
