HCUP KID - Healthcare Cost and Utilization Project, Kids' Inpatient Database

General description

1. Database primary purpose - Provide national estimates of hospital inpatient stays for patients younger than 21 years of age.
2. Overall data type - Health outcomes
3. Dataset type - Cross-sectional
4. Data source - Claims
5. Data level - Patient level, but includes a hospital-level file
6. Geographic location of the data collection sites - United States
7. Sponsor, manager, or home institution - Agency for Healthcare Research and Quality's (AHRQ)
8. Date range - 2006, 2009, and 2012
9. Geolocation data - 2006 and 2009: hospital state postal code, modified FIPS state/county code, urban/rural code, region, zip code and address from AHA Annual Survey; 2012: location (urban/rural) of both hospital and patient
10. Dates - Year; admission day is on a weekend and month; discharge quarter
11. Hospital identifiers - 2006 and 2009: AHA hospital identifier, HCUP hospital identification number, and Hospital name from AHA Annual Survey; 2012: KID hospital number
12. Physicians identifiers - Provides de-identified physician identifiers, which can be used to distinguish between physicians. If the original physician identifier is based on a state license number or Universal Physician Identification Number (UPIN), then Physician number can be used to track a physician across hospitals. If the original physician identifier is based on hospital-specific identifiers, then it can only be used to track physicians within a hospital.
13. Longitudinal tracking - Track providers (see identifiers above)
14. Financial variables - Contains charge information and provides supplemental files containing cost-to-charge ratios.
15. Clinical areas of interest - all; Pediatric care
16. Number of records - Two to three million hospital stays each year.
17. Variables that are uniquely present in this dataset - The largest publicly available all-payer pediatric inpatient care database in the United States. KID's data enables analyses of rare conditions, such as congenital anomalies, as well as uncommon treatments, such as organ transplantation.
18. Database caveats and limitations - Not all data elements are available from every state, and not in all the years. Also, not all data elements are uniformly coded across states.

Applicable methods

1. Association methods, such as linear regression models (1, 2), logistic regression models (3, 4), hierarchical models (5), ANOVA (6)
2. Univariate statistical analysis (7, 8)
3. Propensity score (9, 10)
4. Sensitivity Analysis (11)

High-impact designs

• Assess temporal trends in condition-related hospitalizations (12)

• Association between regional factors and resources utilization for congenital disease (13)

Data dictionary

To access the data dictionary, click here

Variable categories

1. Patient demographics (e.g., age, neonatal age, sex, race, ethnicity, median household income, and median income for patient's ZIP Code)
2. Hospital discharge records (e.g., diagnosis-related group (DRG) in use on discharge date, elective and non-elective admission, number of diagnoses on discharge, LOS)
3. Charges (e.g., expected payer, total charges)
4. Diagnosis codes (ICD-9; including injury type)
5. Procedure codes
6. Hospital characteristics (e.g., ownership, bed size, STRATA, teaching status, number of hospital discharges, and hospital markets file defined by geopolitical boundaries, fixed radius, variable radius, and patient flow)
7. Disposition at discharge (e.g., discharge to home, transferred to [type of facility], left against medical advice)
8. Died during hospitalization
9. Severity and comorbidity measures (e.g., chronic condition indicators, AHRQ comorbidity measure)

Linkage to other datasets

• Linkages can be established with AHA for the years 2006 and 2009