Using Health and Family Data from the National Center for Health Statistics to Study Health Disparities

1
Using Health and Family Data from the National
Center for Health Statistics to Study Health
Disparities

• Anjani Chandra, Ph.D.

2
Agenda

• NCHS mission
• Overview of major NCHS data systems
• NHANES
• NHIS
• NHCS
• NVSS
• Closer look at data sets related to reproductive
  and perinatal health

3
NCHS Mission

• To monitor the nations health by providing data
  that are
• accurate
• relevant
• timely
• To support the broader CDC mission
• SAFER HEALTHIER PEOPLE

4
NCHS Data Are Used For

• Comparisons across populations, providers, and
  geographic areas
• Planning/assessing public health programs
• Identifying health problems, risk factors, and
  disease patterns

5
Quality Control at Every Stage

• Data collection
• Instrument design
• Interviewer (or data collector) training
• Data analysis
• Evaluation of data quality
• Information dissemination

6
Four Major Data Systems at NCHS

• National Health and Nutrition Examination Survey
• National Health Interview Survey
• National Health Care Survey
• National Vital Statistics System and surveys

7
NHANES

• Data sources
• Personal interviews (including ACASI)
• Physical examinations
• Laboratory tests
• Nutritional assessments

8
NHANES sample design

• 30,000 persons aged 2 months or older
• Oversample
• African Americans
• Mexican Americans
• Adolescents aged 12-19
• Older persons aged 60
• Pregnant women

OP96031
9
Mobile Examination Center (MEC) Trailer used for
NHANES
10
Some Uses of NHANES

• Estimate the prevalence and distribution of
  health conditions and related risk factors in the
  population
• Describe awareness, treatment and control of
  selected diseases
• Monitor trends in health, risk behaviors, and
  environmental exposures over time

OP96023
11
Some Uses of NHANES (contd)

• Study the relationship of diet, nutrition and
  health
• Develop growth charts for evaluating child
  development
• Explore emerging public health issues
• Establish a national probability sample of
  genetic material for future research

OP96024
12
(No Transcript)
13
(No Transcript)
14
National Health Interview Survey (NHIS)

• Data source
• Personal interviews with household members
• Sample
• 41,000 households annually
• (approximately 110,000 people)
• Oversample African Americans and Hispanics

15
NHIS Components

• Core questionnaire
• Topical modules
• Periodic modules

16
NHIS Core Questionnaire

• Collects minimal information on
• Health behaviors
• Chronic and acute conditions
• Health care utilization
• Health status and limitations

17
NHIS Modules

• Topical modules collect data on
• Emerging problems or specific health problems
• Cancer
• Healthy People 2000 and 2010 indicators
• Diabetes
• Periodic modules collect more detailed
  information on one of the four areas of health
  above.

18
NHIS Module Topics Have Included

• Health status and disability
• Insurance coverage
• Access to care
• Use of health services
• Immunization
• Health behaviors
• Family practices

19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
The National Health Care Survey(Original Surveys)

• National Hospital Discharge Survey
• National Ambulatory Medical Care Survey
• National Nursing Home Survey
• National Health Provider Inventory

23
The National Health Care Survey (Newer Surveys)

• National Hospital Ambulatory Medical Care Survey
• National Survey of Ambulatory Surgery
• National Home and Hospice Care Survey

24
Some Uses of NHCS Data

• Prevalence of diagnosis and treatment
• Trends in use of services
• Characteristics of patients and providers
• Patterns of disease
• Use of technology drugs, and other treatments
• Emergence of alternative care sites

25
(No Transcript)
26
(No Transcript)
27
Other NCHS Surveys

• National Immunization Survey
• State and Local Area Integrated Telephone Survey
  (SLAITS)
• National Longitudinal Survey of Aging (LSOA)

28
National Vital Statistics System

• Data source
• State vital registration
• Sample
• All births, deaths, and fetal deaths
• Counts of marriages and divorces

29
Hispanic and API Subgroups (since 1990s)

• Hispanic origin
• Mexican, Puerto Rican, Cuban, Central or South
  American, other/unknown Hispanic group
• Asian American or Pacific Islander
• Chinese, Japanese, Hawaiian, Filipino, Asian
  Indian, Korean, Vietnamese, Samoan, Guamanian,
  Other Asian or Pacific Islander1

1 Categories in green are available only for
selected States
30
Some Uses of Mortality Data

• Life expectancy
• Fetal and Infant mortality
• Occupational mortality
• Causes of death
• National Death Index (for mortality
  ascertainment)

31
Life Expectancy at Birth, by Race and Year
Source National Vital Statistics System -
Mortality, CDC, NCHS. Source Arias E (2002).
United States Life Tables, 2000. Natl Vital
Statistics Reports 51(3). Hyattsville, MD NCHS.
32
(No Transcript)
33
Revised Death Certificate

• Multiple race
• Education
• Marital status
• Place of death
• Pregnancy status

• If transportation injury, decedents role with
  respect to vehicles
• Tobacco use

34
Some Uses of Natality Data

• Birth rates
• Nonmarital births
• Births to teens
• Prenatal care and birthweight
• Cesarean delivery
• Obstetric complications
• Timing of births (incl. delayed childbearing)
• Pregnancy rates (along w/other data)

35
(No Transcript)
36
Pregnancy, Live Birth, Induced Abortion Fetal
Loss Rates, 1990 1995
Rate per 1,000 Women
1995
1990
Source Ventura et al, 2000
37
Revised Birth Certificate

• More detail on cigarette smoking
• Method of delivery
• More height weight info
• Congenital anomalies

• Infertility treatment
• WIC
• Infections during preg
• Maternal morbidity
• Breast feeding
• Payment for delivery

Separate worksheets for mother for facility, to
encourage collection from best data sources.
38
Certificate Revision Website(birth death)

• For everything you always wanted to know about
  the revision but were afraid to ask
• www.cdc.gov/nchs/vital_certs_rev.htm

39
Other Datasets for Perinatal Health Research

• In addition to the annual National Natality
  Datasets
• Linked Birth/Infant Death Files
• Matched Multiple Birth File (through RDC)
• Fetal Death Files

40
National Survey of Family Growth
41
Uses of NSFG data

• Collect periodic data on the intermediate
  variables
• Snapshot of U.S. fertility, family formation,
  and reproductive health
• Complement vital statistics
• Track national health objectives
• Evaluate health social policies
• Research demographic and public health

42
Intermediate variables
Intercourse variables

• Timing of first intercourse
• Percent of women who
• ever had intercourse
• Time spent in marriage
• (separation, divorce)
• Frequency of intercourse

Social factors

• Race/ethnicity
• Religion
• Labor force
• participation
• Education
• Income
• Access to
• health care
• Family
• background
• Community
• environment
• (economic,
• social, etc.)

Conception variables
Fertility (live births)

• Contraceptive use
• Sterilization
• Infertility

Pregnancy outcome (gestation) variables

• Miscarriage and
• stillbirth
• Induced abortion

43
NSFG sample (Cycles 1-5)

• Nationally representative sample of women
• Civilian non-institutionalized population of
  women 15-44
• Multi-stage probability sample
• Oversample black women and starting in 1995,
  Hispanic women
• 1988 1995 based on NHIS sample

44
NSFG Cycles 1-5
Cycle
Year
N
Women 15-44
1 1973 9,797 Ever-married 2 1976 8,611
Ever-married 3 1982 7,969 All 4 1988 8,450 All 4
(Tel) 1990 5,686 Reintv of 88 new
teens 5 1995 10,847 All
45
Innovations in 1995 NSFG, Cycle 5

• Major changes in questionnaire content and format
• Changed mode of data collection to CAPI and Audio
  CASI
• 20 token of appreciation paid to respondents --
  maintain response rates with longer, more
  sensitive interview

46
Selected Findings from the 1995 NSFG
47
Children Ever Born, by Womans Education, 1995
Source Abma et al, 1997, Table 4
48
Percent of Women 18-29 Who Had a Birth Before 18
Percent
Median Family Income at Block Level
Source 1995 National Survey of Family Growth,
Contextual Data Files
49
Father Figure During Adolescence U.S. women
15-44 in 1995
Percent
Type of Father Figure
Source Peterson, poster presented at PAA 1998
50
Percent Likelihood that 1st Marriage Dissolves
Within Ten Years
Percent
Age at marriage
Source Bramlett Mosher, 2002
51
Type of Contraceptive Used Among Women 15-44 Who
Currently Use Contraception U.S., 1995
Method
Source Piccinino Mosher, 1998
52
One-Year Failure Rates by Method US,
1995(Chance of Having an Unintended Pregnancy in
1 Year of Using the Method)
53
Current contraceptors 20-44 who are using female
sterilization
Source Mosher et al., 2003, Table 9
54
Percent of Babies Born 1990-94 who were
Breast-fed at all
Source Mosher et al., 2003, Table 15
55
Payment for Delivery for Womens Most Recent
Birth, 1991-5
Mothers Age at Delivery
Source Abma et al, 1997, Table 69
56
Percent of Recent Births that were Intended at
Conception
Age at birth
Source Abma et al., 1997, Table 14
57
Use of Family Planning or Medical Services in
Past Year Among Women 15-44 U.S., 1995
Percent
Source Abma et al, 1997, Table 74
58
Percent of Women 15-44 Ever Treated for PID
Percent
Age at first sex
Source Abma et al, 1997, Table 89
59
Percent Uninsured Among Women 15-44
Percent
Source 1995 National Survey of Family Growth
60
Percent of Unmarried Women Reporting 4 or More
Male Sexual Partners in Past Year, Interviewer
versus Self-administered mode
Percent
Source Abma et al, 1997, Table 27
61
Using the 1995 NSFG

• Several data files available
• Common CASEID allows linkage of information
  across files
• Access procedures vary due to confidentiality and
  sensitivity of data

62
1995 NSFG data files available

• Public Use Files
• Respondent file
• Pregnancy (interval) file
• Omitted Items File
• Fieldwork File
• Contextual Data Files

63
NSFG - variance estimation

• Complex sample design must be accounted for to
  make accurate inferences
• Requirements for accurate estimation in 1995 NSFG
• Fully adjusted weight (POST_WT)
• Collapsed strata variable (COL_STR)
• Panel identifier (PANEL)
• For reference and further details Series 2,
  124

64
NSFG Public Use Data Files

• Available on cartridge tape and CD-ROM (1982 NSFG
  CDROM coming soon)
• 1995 extensive documentation (also on web)
• Users Guide
• Codebook
• CAPI Reference Questionnaire

65
1995 NSFG - Contextual data files

• Restricted use (contact NCHS RDC)
• Community characteristics (over 1000)
• 4 levels of aggregation (state, county, census
  tract, block group)
• 3 points in time (1990, 1993, 1995)
• Rich potential for descriptive as well as
  detailed multilevel analyses.

66
NSFG related files available through NCHS
Research Data Center

• Contextual data files
• State level indicators (user-supplied)
• NHIS linked analyses (1988 1995)
• RDC Contact Kenneth Harris
• Email rdca_at_cdc.gov

67
Cycle 6 and beyond

• Shorter intervals between Cycles
• Cycle 6 in 2002
• Cycle 7 planned for 2005 or 2006
• Sample independent from NHIS
• Contextual data
• Inclusion of males 15-44
• Moving toward continuous interviewing in Cycle 7

68
Inclusion of males 15-44

• Other half of fertility and family formation
  equation
• 1996 Welfare Reform Act
• Fatherhood Initiative (federal interagency
  effort)
• Information collected directly

69
Male NSFG will shed light on

• Father involvement activities
• Child support
• Male reporting of
• pregnancies/births
• sexual partnerships
• contraceptive behavior
• infertility
• Male role in unintended pregnancy
• STD/HIV risk transmission

70
Female NSFG in Cycle 6

• All core topics maintained from Cycle 5
• Expanded Audio CASI
• Streamlined event histories
• Other selected enhancements based on
  questionnaire design research

71
2002 NSFG Preliminary Counts

• Female Male Total
• Total 7,646 4,929 12,575
• 15-19 1,150 1,122 2,272
• 20-24 1,364 938 2,302
• 25-44 5,132 2,869 8,001

Data collection ended in March 2003. Final
counts will be slightly lower, post-data editing
cleaning.
72
2002 NSFG (Cycle 6)

• Public use data files documentation target
  Summer 2004
• Web-based documentation
• CDROM for data files
• Contextual data about a year later
• Watch our webpage for updates
• www.cdc.gov/nchs/nsfg.htm

73
Quality Control Strategies in 2002 NSFG

• Full dress rehearsal Pretest in 2001 (n600)
• Extensive input from research policy world
• Cognitive testing of questionnaire components
  with key subpopulations
• Rigorous protocol for Spanish translation
• Interviewer training modules on cultural
  sensitivity and refusal aversion
• Verification interviews

74
Lets take a step back and consider other
concerns about surveys
75
All surveys are facing challenges

• People are harder to find and talk to
• People are more reluctant to participate
• ? Survey response rates suffer, or they cost more
  to maintain.

76
All surveys are facing challenges

• Computerization of survey instruments can
  simplify interviewers job to some extent, but
• Survey questionnaires tend to become more complex
• Innovative data collection strategies (including
  ACASI) require considerably higher cost in
  programming and testing
• ? Overall survey costs increase

77
All surveys are facing challenges

• Controlling survey costs requires
• Collection and use of process data in directing
  effort of interviewers in the field
• Real time modeling to balance cost control with
  response rate and error reduction goals
• YOU, as data users, need to be aware of these
  survey management issues, as they may impact on
  your research.

78
Questions?Comments?