National Center for Health Statistics DCC CENTERS FOR DISEASE CONTROL AND PREVENTION Changes in Race Differentials: The Impact of the New OMB Standards.

National Center for Health Statistics

DC CCENTERS FOR DISEASE CONTROL

AND PREVENTION

Changes in Race Differentials: The Impact of the New OMB

Standards on Health Data in the NHIS

Jacqueline Wilson LucasDivision of Health Interview Statistics

What is OMB Directive 15?

• Standard for data collection on race and ethnicity in the federal statistical system

• Implemented by the Office of Management and Budget (OMB) in 1977

• Minimum reporting categories:

• White, Black, Asian and Pacific Islander (API) and American Indian/Alaskan Native (AIAN)

Why change OMB Directive 15?

• Demography of the country is changing

• Resistance to reporting a single race for people with more than one

• Need for federal statistical systems to keep pace with the changing population - most importantly the Decennial Census

What are the new standards for federal race and ethnicity data collection

• Revised categories:

• White, Black/African American, American Indian or Alaska Native (AIAN), Native Hawaiian or Other Pacific Islander (NHOPI) [new], Asian [new]

• Hispanic origin (Hispanic/Latino)

• Ask prior to and separately from race

Revised guidelines, cont’d

• Data collection and tabulation

• Data systems must allow respondents to report more than one race

• Data systems must tabulate and report information on multiple race persons, provided data meet agency standards

Why changes to race data are important

• Among the most commonly used demographic variables in analysis of health data

• Many health outcomes of interest - hypertension, diabetes, cancer morbidity and mortality - differ by race

• Important to know whether changes in health outcomes by race over time are due to changes in behavior or changes in the way we measure race

What are the most important effects of the new standards?

• Changes in tabulation and presentation of data shifts in people reported in particular categories

• Changes in trend data monitoring new groups creates breaks in data

• Changes in the interpretation of data for racial/ethnic groups need to understand the effect of reporting and interpreting data for groups whose composition may be changing over time

How are race data collected in the NHIS?

• Self-reported race and ethnicity collected since 1976; since 1982 for all household members

• Ethnicity asked in one question, race in two:

• Group or groups that best represent HH member’s race

• Which mentioned BEST describes race (for those with more than one race)

Distribution of racial groups in the NHIS, 1997-1999

0% 50% 100%

1997

1998

1999

Surv

ey

year

Percent

WhiteBlackAIANAsian/API*NHOPI*Other raceMultiple race

SOURCE: CDC/NCHS: National Health Interview Survey; weighted estimates.

Age distribution of single race groups, 1998 NHIS

0

20

40

60

80

White Black AIAN API

Per

cen

t Under 17

18-24

25-44

45-64

65+

SOURCE: CDC/NCHS, 1998 NHIS (weighted data)

Age distribution of multiple race groups, 1998 NHIS

0

20

40

60

80

AIAN/White API/White Black/White Other comb

Per

cen

t

Under 17

18-24

25-44

45-64

65+

SOURCE: CDC/NCHS, 1998 NHIS (weighted data)

95% confidence intervals for private health insurance coverage, 1998 NHIS

0

20

40

60

80

100

White AIAN AIAN/White

Upper bound

Lower bound

Point estimate

SOURCE: CDC/NCHS, 1998 NHIS (weighted data)

95% confidence intervals for private health insurance coverage, 1998 NHIS

0

20

40

60

80

100

White Black Black/White

Upper bound

Lower bound

Point estimate

SOURCE: CDC/NCHS, 1998 NHIS (weighted data)

95% confidence intervals for private health insurance coverage, 1998 NHIS

0

20

40

60

80

100

White API API/White

Upper bound

Lower bound

Point estimate

SOURCE: CDC/NCHS, 1998 NHIS (weighted data)

95% confidence intervals for respondent assessed health status as “excellent or very good”, 1998 NHIS

0

20

40

60

80

100

White AIAN AIAN/White

Upper bound

Lower bound

Point estimate

SOURCE: CDC/NCHS, 1998 NHIS (weighted data)

95% confidence intervals for respondent assessed health status as “excellent or very good, 1998 NHIS

0

20

40

60

80

100

White Black Black/White

Upper bound

Lower bound

Point estimate

SOURCE: CDC/NCHS, 1998 NHIS (weighted data)

Summary

• measuring new population groups - NHOPI and the multiple race groups - whose characteristics and patterns of illness and disease appear to be distinct and must be studied further

• racial/ethnic identity - fluid, not fixed - changes our concept of race; substantive meaning of primary race

• relationship between race and health in epidemiologic analyses increasingly more complex

Limitations

• Sample size and precision of estimates

• Limited ability to examine large range of health outcomes - to look at data for adults and children will likely require 3 years or more of data combined to get stable estimates

Limitations (Cont.)

• Differences between multiple race groups and their single race counterparts may be explained by SES and other factors not controlled for in these analyses:

• e.g., age-adjustment possible, but age stratification not possible (in part because of age structure of some multiple race groups)

Availability of multiple race data for analysis

• NHIS Public Use Files (on CD-ROM, data tape, web) do not contain detailed information on multiple race groups

• confidentiality and data reliability issues

• geographic and familial clustering

• Special Request Files

• in-house data files containing suppressed information are available through Research Data Center

Where to find out more about the OMB race standards:

http://www.whitehouse.gov/OMB/inforeg:

Statistical Policy Section - Data on Race and Ethnicity

• question wording (Section II, Appendix B)

• surveys and administrative records (Section IIIb)

• bridging methods (Section V, Appendix D)