Peer Reviewed: Computations of Confidence Intervals for Estimates in the United States National Hospital Discharge Survey, 1979–2000

Yao-Hua Luo Luo; Matthew Zack

Table 1

Differences in Layout of Relative Standard Error (RSE) Parameter Tables Before 1988 and During or After 1988, CD-ROM on National Hospital Discharge Survey Data, 1979–2000^a ^, ^b

1979–1987

Type of Statistic for Hospitalizations (Listed as Subtables)	Weighted Estimates	Demographic Characteristics (Listed as Variables)

		Region	Race	All Others
First- or all-listed diagnosis	5,000–40,000,000	*	*	*
Days of care	10,000–250,000,000	*	*	*
Procedure	5,000–30,000,000	*	*	*


1988–2000
Demographic Characteristics (Listed as Values of Variables)	Type of Statistic for Hospitalizations (Listed as Variables)

	First-listed Diagnosis		All-listed Diagnoses		Days of Care		Procedure

	A	B	A	B	A	B	A	B
Region	•	•	•	•	•	•	•	•
Race	•	•	•	•	•	•	•	•
Male	•	•	•	•	•	•	•	•
Female	•	•	•	•	•	•	•	•
Aged <15 years	•	•	•	•	•	•	•	•
Other	•	•	•	•	•	•	•	•
TOTAL	•	•	•	•	•	•	•	•

^aCD-ROM issued by the National Center for Health Statistics, 2002. Relative standard error (RSE) parameters are expressed as percentages before 1988; asterisks (*) represent possible RSE values during 1979–1987. In contrast, each statistic for 1988–2000 has two coefficients, A and B, which are derived from RSE curves.

^bBullets (•) represent coefficient values during 1988–2000.