42
since 1965. The NHDS is designed to provide information about characteristics of inpa-
tients (regardless of age) hospitalized and discharged from the hospitals in the survey.
Medical and administrative data—including date of birth, sex, race, ethnicity, marital
status, ZIP codes, dates of admission and discharge, discharge status, expected source of
payment, procedures, diagnoses, size of hospital, and hospital ownership—for approxi-
mately 300,000 hospital discharges are obtained from two sources. The first source uses
data that is manually-abstracted and transcribed by hospital and U.S. Bureau of the
Census staff from a manually-selected sample of hospital discharge records. Completed
forms are coded, computerized, and edited by NCHS. The second source uses a system-
atic sample of electronic hospital discharge files containing medical record data selected
from electronic files purchased from public and private organizations authorized by the
states. Approximately 10 percent of the abstracts are independently recoded with an
overall error of 0.6 percent for medical coding and 0.3 percent for administrative coding.
Approximately 40 percent of respondent hospitals provided data through the automated
system. Sample data are weighted to produce annual, nationally-representative estimates.
Routine analysis of NHDS data costs $100,000 to $150,000 per year.
Strengths: Nationally representative. Targets all ages. Excellent quality. Timely.
Economical. Uses International Classification of Diseases, Ninth Revision, Clinical
Modification (ICD-9-CM) rubrics to code medical diagnoses.
Limitations: Does not include ED data. Cannot distinguish first admissions from read-
missions in some states; thus, allowing for discharge rates, not injury rates. Measures
discharges and not individual patients (potential duplicates).
Recommendations: Analyze routinely to study incidence, trends, demographics, external
causes of injury, service use, hospital characteristics, expected source of payment, chief
complaint and diagnosis, medications, type of provider, and disposition of MTBI-related
hospitalizations in non-federal, short-stay hospitals in the United States.
Nationwide Inpatient Sample (NIS)
Characteristics: Nationally-representative, multi-state health data system based on all
hospital discharges from a stratified probability sample of non-federal, short-stay
hospitals (994 in 28 participant states in 2000). Sponsored by the Agency for Health-
care Research and Quality since 1988. Designed to approximate a 20 percent sample
of all non-federal, short-term, general and other specialty hospitals in the United States.
The NIS 2000 is a sample of hospitals that comprise about 80 percent of all hospital
discharges in the United States. States voluntarily report allowable electronic, coded
discharge data from all persons hospitalized regardless of age. Analysis of NIS data
costs $100,000 to $150,000 annually.