Characteristics and outcomes of WIC participants and nonparticipants analysis of the 1988 national maternal and infant health survey

              w
COMPLETED A ?t> ft'Trf
United State*
Department of
Agriculture
Food and
Consumer
Service
Office of
Analysis and
Evaluation
Characteristics and
Outcomes of WIC
Participants and
Nonparticipants:
*
iS-dvss^
Analysis of the 1988
National Maternal
Health Survey
Contract No.:
Subcontract No:
MPR Reference No.:
53-3198-9-0-033
1-524-4790
7939-011
CHARACTERISTICS AND OUTCOMES OF
W1C PARTICIPANTS AND NONPARTICIPANTS:
ANALYSIS OF THE 1988 NATIONAL MATERNAL
AND INFANT HEALTH SURVEY
March 1995
Authors:
Anne Gordon
Lyle Nelson
Submitted to:
U.S. Department of Agriculture
Food and Nutrition Service
Office of Analysis and Evaluation
3101 Park Center Drive, 2nd Floor
Alexandria, VA 22302
Project Officer:
Janet Tognetti Schiller
Submitted by:
Mathematica Policy Research, Inc.
P.O. Box 2393
Princeton, NJ. 08543-2393
(609) 799-3535
Under subcontract to the Research
Triangle Institute
Contract Amount: $838,810
Subcontract Amount: $318,331
Fully Competitive
BLANK PAGE
M
ACKNOWLEDGMENTS
The report reflects the efforts of many individuals at Mathematica Policy Research and at the
Food and Nutrition Service. Barbara Devaney, MPR's project director for the WIC Modeling and
Analytic Projects, provided review and guidance during all phases of the preparation of this report,
wrote the Executive Summary (with Jim Ohls), and wrote portions of the introduction and literature
review. The literature review is also partly based upon materials prepared by Sheena McConncll.
Cara Hendricks, Daisy Ewell and Dexter Chu constructed the analysis files and programmed the
analyses. Nancy Whelan edited the final report, and it was produced by Jill Miller, Cindy Castro,
Debra Jones, Monica Capizzi and Marjorie Mitchell.
At the Food and Nutrition Service, Janet Tognetti Schiller, Jeffery Wilde, Donna Blum and Jay
Hirschman offered valuable insights and comments. Chester Scott and other staff at the National
Center for Health Statistics provided information on the NMIHS data. Linda Adair of the University
of North Carolina at Chapel Hill consulted on the design of the infant feeding analysis. Rick
Williams (formerly of the Research Triangle Institute) provided assistance in the use of SUDAAN
to estimate standard errors. Several MPR colleagues provided assistance with technical issues,
including Randy Brown, John Hall, Chuck Metc&Jf, Jim Ohls, and Peter Schochet. The assistance
and comments of all of these individuals are gratefully acknowledged. Nonetheless, the authors are
fully responsible for the findings and conclusions in this report.
Anne Gordon
Lyle Nelson
in
BLANK PAGE
W
CONTENTS
Chapter Page
EXECUTIVE SUMMARY xiii
I INTRODUCTION 1
BACKGROUND ON THE WIC PROGRAM 2
STUDY OBJECTIVES 4
II REVIEW OF PREVIOUS LITERATURE 7
STUDIES OF THE EFFECTS OF PRENATAL WIC PARTICI-PATION
7
Major Evaluations 7
Evaluation Reviews 13
STUDIES OF THE EFFECTS OF WIC ON INFANTS AND
CHILDREN 15
Effects of WIC on the Utilization of Health Care Services 15
III DATA AND METHODOLOGY 19
THE DATA 19
METHODOLOGICAL ISSUES 21
Selecting a Comparison Group 22
Determining the Analysis Sample 24
Weighting Sample Observations and Computing Standard Errors 24
IV CHARACTERISTICS OF WIC PARTICIPANTS AND
NONPARTICIPANTS 27
PARTICIPATION AND EXPERIENCES IN THE WIC
PROGRAM 27
WIC Participation 28
Program Experiences of Prenatal WIC Participants 30
Differences in the Experiences of Black and White Prenatal
WIC Participants 33
CONTENTS (continued)
Chapter Page
IV PRENATAL WIC PARTICIPANTS AND NONPARTICIPANTS 33
(continued)
Demographic and Socioeconomic Characteristics 34
Prenatal Care 38
Behavioral Risk Factors-Alcohol Consumption, Cigarette
Smoking, and Drug Use 42
Previous Pregnancies and Previous WIC Participation 45
Pregnancy Outcomes 49
POSTPARTUM AND INFANT WIC PARTICIPANTS AND
NONPARTICIPANTS 49
Demographic and Socioeconomic Characteristics of Postpartum
WIC Participants and Nonparticipants 51
Demographic and Socioeconomic Characteristics of Infant WIC
Participants and Nonparticipants 55
Infant Health Care Utilization and Health Status 67
Infant Feeding Practices 74
V MULTTVARIATE ANALYSIS OF THE RELATIONSHIP
BETWEEN PRENATAL WIC PARTICIPATION AND BIRTH
OUTCOMES 93
OVERVIEW AND SUMMARY 93
Methodological Issues 93
Summary 94
EMPIRICAL RESULTS FROM THE BASIC MODEL 95
Effects on Birthweight 95
Effects on Gestational Age 98
Effects on the Incidence of Adverse Birth Outcomes 98
Interpreting the Results from the Basic Model 104
RESULTS FOR MODELS USING ALTERNATIVE
MEASURES OF WIC PARTICIPATION 106
Addressing the Relationship between WIC Enrollment and
Pregnancy Duration 106
Addressing the Dose-Response Issue 121
SELECTION BIAS 124
AREAS FOR FURTHER RESEARCH 126
VI
CONTENTS (continued)
Chapter Page
REFERENCES 127
APPENDIX A: MULTIVARIATE ANALYSIS METHODOLOGY 131
APPENDIX B: ESTIMATED COEFFICIENTS IN LOGIT
MODELS OF THE EFFECT OF PRENATAL
WIC PARTICIPATION ON THE INCIDENCE
OF ADVERSE BIRTH OUTCOMES 139
VII
BLANK PAGE
v<\\
TABLES
Table Page
II.1 SUMMARY OF MAJOR WIC EVALUATIONS 8
IV. 1 WIC PARTICIPATION AND DURATION OF PARTICIPA-TION
BY PREGNANT WOMEN, INFANTS, AND
POSTPARTUM WOMEN 29
IV.2 PRENAYAL WIC PARTICIPANTS' EXPERIENCE WITH
THE WIC PROGRAM 31
IV.3 DEMOGRAPHIC AND SOCIOECONOMIC CHARACTERISTICS OF
PRENATAL WIC PARTICIPANTS, INCOME-ELIGIBLE
NONPARTICIPANTS, AND HIGHER-INCOME
NONPARTICIPANTS 35
IV.4 THE USE AND SOURCE OF PRENATAL CARE 39
IV.5 ALCOHOL CONSUMPTION, CIGARETTE SMOKING, AND
DRUG USE BEFORE AND DURING PREGNANCY 43
IV.6 PREVIOUS PREGNANCIES AND PREVIOUS WIC
PARTICIPATION 46
IV.7 PREGNANCY OUTCOMES FOR PRENATAL WIC PAR-TICIPANTS,
INCOME-ELIGIBLE NONPARTICIPANTS,
AND HIGHER-INCOME NONPARTICIPANTS 50
IV.8 DEMOGRAPHIC AND SOCIOECONOMIC CHARACTERISTICS OF
POSTPARTUM WIC PARTICIPANTS, INCOME-ELIGIBLE
NONPARTICIPANTS, AND HIGHER-INCOME
NONPARTICIPANTS 52
IV.9 DEMOGRAPHIC AND SOCIOECONOMIC CHARACTER-ISTICS
OF POSTPARTUM WIC PARTICIPANTS WHO
WERE AND WERE NOT PRENATAL WIC PARTICIPANTS 56
IV. 10 DEMOGRAPHIC AND SOCIOECONOMIC CHARACTERISTICS OF
INFANT WIC PARTICIPANTS, INCOME-ELIGIBLE
NONPARTICIPANTS, AND HIGHER-INCOME
NONPARTICIPANTS 59
IV.l 1 DEMOGRAPHIC AND SOCIOECONOMIC CHARACTER-ISTICS
OF INFANT WIC PARTICIPANTS WHO WERE
AND WERE NOT PRENATAL WIC PARTICIPANTS 63
IX
TABLES (continued)
Table Page
IV.12 INFANT HEALTH STATUS AND HEALTH CARE UTILIZATION
DURING THE SIX MONTHS AFTER BIRTH 68
IV.13 COMPLIANCE WITH INFANT FEEDING GUIDELINES OF
THE COMMITTEE ON NUTRITION OF THE AMERICAN
ACADEMY OF PEDIATRICS (CN-AAP) 78
IV. 14 INFANT FEEDING PRACTICES REPORTED BY
MOTHERS OF INFANT WIC PARTICIPANTS, INCOME-ELIGIBLE
NONPARTICIPANTS, AND HIGHER-INCOME
NONPARTICIPANTS 81
IV. 15 INFANT FEEDING PRACTICES USED BY MOTHERS OF
INFANTS WHO WERE NOT BREASTFED 86
IV.16 INFANT FEEDING PRACTICES USED BY MOTHERS OF
INFANTS WHO WERE EVER BREASTFED 89
V.l THE EFFECT OF PRENATAL WIC PARTICIPATION ON
BIRTHWEIGHT: BASIC MODEL %
V.2 THE EFFECT OF PRENATAL WIC PARTICIPATION ON
BIRTHWEIGHT BY RACE 99
V.3 THE EFFECT OF PRENATAL WIC PARTICIPATION ON
GESTATIONAL AGE: BASIC MODEL 102
V.4 EFFECTS OF PRENATAL WIC PARTICIPATION ON THE
PERCENTAGE OF UNFAVORABLE BIRTH OUTCOMES:
BASIC MODEL 105
V.5 EFFECTS OF PRENATAL WIC PARTICIPATION UNDER
ALTERNATIVE DEFINITIONS OF PARTICIPATION 108
V.6 EFFECT OF PRENATAL WIC PARTICIPATION ON
BIRTHWEIGHT, CONTROLLING FOR GESTATIONAL
AGE 109
V.7 EFFECT OF PRENATAL WIC PARTICIPATION ON
BIRTHWEIGHT: FOUR COHORTS DEFINED BY GESTA-TIONAL
AGE THRESHOLDS 113
V.8 EFFECT OF PRENATAL WIC PARTICIPATION ON
GESTATIONAL AGE: THREE COHORTS DEFINED BY
GESTATIONAL AGE THRESHOLDS 116
TABLES (continued)
Table Page
V.9 EFFECTS OF PRENATAL WIC PARTICIPATION ON THE
PERCENTAGE OF UNFAVORABLE BIRTH OUTCOMES:
FOUR COHORTS DEFINED BY GESTATIONAL AGE
THRESHOLDS 119
V.10 EFFECTS OF PRENATAL WIC PARTICIPATION AND OF
FIRST TRIMESTER WIC ENROLLMENT: FOUR
COHORTS DEFINED BY GESTATIONAL AGE
THRESHOLDS 123
A.1 INDEPENDENT VARIABLES FOR MULTTVARIATE
ANALYSIS OF THE EFFECTS OF PRENATAL WIC
PARTICIPATION ON BIRTH OUTCOMES 137
B.l ESTIMATED LOGIT COEFFICIENTS IN MODELS OF THE
EFFECT OF PRENATAL WIC PARTICIPATION ON THE
INCIDENCE OF LOW BIRTHWEIGHT, VERY LOW BIRTH-WEIGHT,
AND PRETERM BIRTH 141
B.2 ESTIMATED LOGIT COEFFICIENTS IN MODELS OF THE
EFFECT OF PRENATAL WIC PARTICIPATION ON NEO-NATAL
DEATH AND INFANT DEATH 144
XI
BLANK PAGE
\ "\
XM
EXECUTIVE SUMMARY
The Special Supplemental Food Program for Women, Infants, and
Children (WIC) provides supplemental foods, nutrition education, and
social service referrals to low-income pregnant, breastfeeding, and
postpartum women, to infants, and to children up to 5 years of age. This
study examines the background characteristics, pregnancy experiences, and
birth and infant outcomes of WIC participants, income-eligible
nonparticipants, and higher-income nonparticipants, using data from the
1988 National Maternal and Infant Health Survey (NMIHS). The
National Center for Health Statistics (NCHS) sponsored the NMIHS and
collected data from a national sample of 9,953 women who experienced
a live birth in 1988. The objectives of this study are to (1) examine
patterns of WIC participation among pregnant and postpartum women,
and infants; (2) conduct a descriptive analysis of the characteristics and
experiences of WIC participants, income-eligible nonparticipants. and all
other nonparticipants; and (3) estimate the effects of prenatal WIC
participation on birth outcomes.
CHARACTER-ISTICS
OF WIC
PARTICIPANTS
Approximately 62 percent of pregnant women who were income-eligible
for WIC are estimated to have participated in the program in
1988. Among prenatal WIC participants, 51 percent reported enrolling
in their first trimester and another 35 percent in their second trimester.
About 94 percent of prenatal WIC participants reported having
received nutritional advice at the WIC center, 66 percent remembered
being advised to get prenatal care, 64 percent reported being advised to
breastfeed their infants, and 62 percent reported having been advised
to avoid illegal drugs during pregnancy.
In general, the WIC program serves individuals who are disadvantaged on
average, even relative to the rest of the income-eligible population:
Prenatal WIC participants were more likely to be teenagers, less
likely to have a high school diploma, more likely to be black or
Hispanic, and less likely to be married or living with the baby's
father than were income-eligible nonparticipants or higher-income
nonparticipants.
WIC participants were less likely to have been employed in the
previous year, and were more likely to depend on Medicaid and
to lack private insurance than were nonparticipants.
Fathers of the WIC participants' babies were younger, less
educated, more likely to be black or Hispanic, and less likely to
xiii
have been employed than were fathers of the babies of income-eligible
nonparticipants and higher-income nonparticipants.
Prenatal WIC participants and income-eligible nonparticipants did not
differ significantly in terms of household size or mean household income,
but WIC participants were more likely to depend on public assistance
income. A larger proportion of prenatal WIC participants had incomes
below the poverty level (56 versus 45 percent).
Virtually all prenatal WIC participants had newborns who participated in
WIC as infants, and 77 percent were enrolled postpartum. About 30
percent of infant WIC participants and 16 percent of postpartum women
participants had not been enrolled prenatally.
PRENATAL
CARE AND
HEALTH
BEHAVIORS
Only slightly more than half of both WIC participants and income-eligible
nonparticipants received adequate levels of prenatal care,
compared with 82 percent of higher-income nonparticipants. Prenatal
WIC participants, however, were less likely than income-eligible
nonparticipants to receive inadequate levels of prenatal care in several
respects: WIC participants were less likely to receive no prenatal care
(1.3 percent for WIC participants versus 4.9 percent for
nonparticipants), and WIC participants who received prenatal care
were less likely to receive inadequate care, as measured by the Ressner
Index for the adequacy of care.
Prenatal WIC participants were more likely than both groups of
nonparticipants to have received their prenatal care from county or city
health departments, community health centers, or hospital clinics, and
were less likely to have used private doctors or HMOs. WIC participants
were much more dependent than income-eligible nonparticipants on
Medicaid or other government assistance for payment for prenatal care:
51 percent of prenatal WIC participants used Medicaid as compared with
26 percent of nonparticipants. Income-eligible nonparticipants were more
likely than prenatal WIC participants to pay for care with their own funds
or through private insurance. Prenatal WIC participants were more likely
to be hospitalized during pregnancy than income-eligible nonparticipants;
18 percent of participants and 14 percent of income-eligible
nonparticipants were hospitalized during pregnancy.
Participants differed little from income-eligible nonparticipants in their
reported use of alcohol, cigarettes and illegal drugs. Most WIC
participants (86 percent) reported not drinking alcohol during pregnancy,
and only 4 percent reported more than 1 drink per week. Although 37
xrv
percent of participants smoked before the pregnancy, one-fifth of this
group quit during the pregnancy, and most reduced smoking to some
extent. About 3 percent of prenatal WIC participants reported using
marijuana during the pregnancy, and 1 percent reported using cocaine.
INFANT
HEALTH CARE
USE AND
HEALTH
STATUS
In the first six months of life, infant WIC participants visited the doctor
5.9 times on average, with 3.6 visits for well-baby care and 2.5 visits for
care for an illness or injury. Income-eligible nonparticipants had similar
numbers of physician visits. WIC participants were less likely to receive
their well-baby care from a private physician than income-eligible
nonparticipants, and more likely to receive care from health
departments, community health centers or hospital clinics. Infant WIC
participants were also more likely to have their health care paid for by
Medicaid, and less likely to have care paid for by private insurance or
their family's resources.
Infant WIC participants did not differ from income-eligible
nonparticipants in the number of vaccinations received during the first 6
months of life. Both low-income groups, however, were less likely to
receive any vaccinations, or to receive the appropriate number of
vaccinations, than were higher-income nonparticipants.
Among infant WIC participants, 83 percent received a polio vaccination
in the first 6 months, and 61 percent received two or three vaccinations
(the recommended number). Similarly, 93 percent of infant WIC
participants received a diphtheria/tetanus/pertussis vaccination, and 78
percent received the recommended two or three vaccinations. Since 1988,
WIC agencies have stepped up efforts to increase immunization coverage
among infant WIC participants.
WIC infants did not differ significantly from income-eligible
nonparticipants in the prevalence of hospitalizations after birth, but they
were significantly more likely than higher-income nonparticipants to have
been hospitalized. Ten percent of infant WIC participants and 8.9
percent of income-eligible nonparticipants were re-hospitalized after birth,
but only 5.3 percent of higher-income infants were re-hospitalized after
birth.
xv
INFANT The feeding practices used by mothers of infant WIC participants were
FEEDING compared with guidelines for infant feeding from the Committee on
PRACTICES Nutrition of the American Academy of Pediatrics (CN-AAP).
Although breastfeeding is the preferred method of infant feeding, the
CN-AAP guidelines state that either breastfeeding or use of
appropriate infant formula is acceptable in the first 6 months, but not
cow's milk. Furthermore, the CN-AAP does not recommend
introduction of solid foods until the infant is 4 months old: solid foods
may be introduced any time between 4 and 6 months of age. depending
on the maturity of the infant.
About 85 percent of mothers of WIC participants fed their infants
according to the guidelines in the First month, but this prop rtion fell to
25 percent in the fourth month, since most mothers introduced .solid foods
before the fifth month. Nearly all mothers of WIC participants were in
compliance with the guidelines in months 5 and 6 (95 percent and 91
percent), but small proportions reported feeding their infants cow's milk
in these months.
In general, mothers of WIC participants and mothers of income-eligible
nonparticipants followed the guidelines to a similar extent. However,
mothers of WIC participants were more likely to feed their infants
appropriately in months 5 and 6, since mothers of income-eligible
nonparticipants were more likely to start feeding their infants cow's milk
during these months.
Mothers of infant WIC participants were less likely than mothers of
income-eligible nonparticipants to breastfeed their babies, and they
stopped breastfeeding sooner, and were more likely to supplement
breastfeeding with formula. Only 36 percent of mothers of WIC infants
were breastfeeding in their first month, compared with 51 percent of
mothers of income-eligible nonparticipants; only 10 percent were
breastfeeding in the sixth month after birth, compared with 21 percent of
mothers of income-eligible nonparticipants. Provision of infant formula
may attract nonbreastfeeding mothers to WIC. Furthermore, these 1988
data predate legislative changes that required WIC agencies to set aside
substantial resources for breastfeeding promotion.
The findings concerning infant feeding practices are descriptive and do
not control for the more disadvantaged socioeconomic characteristics of
WIC participants. Using the NMIHS data, Schwartz et al. (1992) found
no significant association between WIC participation and breastfeeding
initiation or duration, after controlling for differences in the
socioeconomic characteristics of WIC participants and income-eligible
nonparticipants.
xvi
PRENATAL W1C
PARTICIPATION
AND BIRTH
OUTCOMES
The mean birthweight of newborns of prenatal WIC participants is
3,284 grams (7.23 pounds), and 8.2 percent of newborns born to WIC
participants were low birthweight (less than 2.500 grams or 5.5 pounds).
Prenatal WIC participation is associated with higher average newborn
birthweight and gestational age. Although the simultaneous relationship
between prenatal WIC participation and the duration of piegnancy makes
it difficult to estimate the relationship between WIC participation and
birth outcomes reliably, it is possible to derive plausible upper and lower
bounds on these relationships. Prenatal WIC participation is associated
with an increase in average birthweight of between 25 and 68 grams
(approximately 1 to 2 percent of average birthweight). This estimate is of
a similar magnitude to estimates found in previous studies. Similarly,
prenatal WIC participation is associated with an increase in gestational
age of between one-fourth and one-half of a week.
Furthermore, prenatal WIC participation significantly reduces the
extremes of low birthweight and preterm birth. WIC participation reduces
the percentage of low birthweight births by between 1 and 3 percentage
points (from a mean of 10.8 percent for income-eligible nonparticipants).
and reduces the percentage of preterm births between 2.4 and 3.6
percentage points (from a mean of 14.2 percent for income-eligible
nonparticipants). However, prenatal WIC participation is not related to
neonatal or infant mortality.
XVII
I. INTRODUCTION
Inadequate nutrition and its consequences are a public health concern in
the United States. While on average, the diets of most Americans provide
sufficient quantities of most nutrients, some subgroups of the population
have special nutritional needs that are often not met. In particular,
pregnant women, infants, and children require adequate nutrition during
critical growth and development periods. To add,ess the need of these
subgroups, the Special Supplemental Food Program for Women. Infants,
and Children (WIC) provides supplemental foods, nutrition education, and
health care and social service referrals to low-income pregnant,
breastfeeding, and postpartum women, to infants, and to children up to
five years of age who are at nutritional risk. Supplemental foods and
nutrition education are intended to improve the nutritional status of low-income
pregnant women, which, in turn, is expected to improve pregnancy
outcomes. For infants and children, the supplemental foods arc expected
to reduce the prevalence of anemia and to improve physical and mental
growth and development. The WIC Program also aims to promote good
health care by referring participants to health care providers.
This report presents results from an analysis of data from the 1988
National Maternal and Infant Health Survey (NMIHS). The NMIHS
collected data from a national sample of women who experienced a live
birth, fetal death, or infant death in 1988. The objectives of this report
are to (1) examine patterns of WIC participation among pregnant women,
postpartum women, and infants; (2) conduct a descriptive analysis of the
characteristics and experience of WIC participants, income-eligible
nonparticipants, and higher-income nonparticipants; and (3) estimate the
effects of prenatal WIC participation on birth outcomes.
This report is organized in five chapters. The remainder of this chapter
provides background on the WIC Program and discusses in detail the
objectives of this study. Chapter II highlights findings from previous
research on the effects of WIC participation. Chapter III descrihes the
NMIHS data and our methodology. Chapter IV presents the findings
from a descriptive analysis of the patterns of WIC participation and the
characteristics of WIC participants, income-eligible nonparticipants, and
higher-income (ineligible) nonparticipants. The final chapter presents
results from a multivariate analysis of the effects of WIC participation on
birth outcomes.
BACKGROUND
ON THE WIC
PROGRAM
In 1969, the White House Conference on Food, Nutrition, and Health
recommended that special attention be given to the nutritional needs of
pregnant women and preschool children at nutritional risk. As a result,
in September 1972, Congress authorized the WIC Program as a two-year
pilot program. The program was designed to provide nutritional
screening, food assistance, nutrition education, and health and social
service referrals for low-income pregnant and postpartum women, their
infants, and children up to age five. Since its inception, the WIC
Program has grown dramatically. In fiscal year 198() (FY80), the
program served 1.9 million women and children at a cost of $725
million; in FY91, the program served an estimated 5.2 million women
and children at a cost of $2.3 billion.
The WIC Program is administered nationally by the Food and Nutrition
Service (FNS) of the U.S. Department of Agriculture (USDA) and at the
state level by a designated state agency, usually the state health
department. Congress sets funding annually, and the available funds arc
allocated to the states on the basis of a formula that accounts for the
number and percentage of eligible women being served, among other
factors. WIC is not an entitlement program, and states may not have
sufficient funds to serve all eligible persons who apply for benefits.
Federal regulations thus require that the states establish priority systems
to ensure that scarce program resources are fairly allocated and reach
those most in need. The priority system should be operated statewide and
satisfy the broad federal requirements described below:
Priority I. Pregnant women, breastfeeding women, and infants
at nutritional risk as demonstrated by hematological or
anthropometric assessments or by other documented
nutritionally related medical conditions
Priority II. Infants up to 6 months of age who were born to
women who were WIC recipients during pregnancy or who can
be documented as at nutritional risk during pregnancy
Priority III. Children at nutritional risk as demonstrated by
hematological or anthropometric measurements or other
documented medical conditions
Priority IV. Pregnant women, breastfeeding women, and
infants at nutritional risk because of an inadequate dietary
pattern
Priority V. Children at nutritional risk because of an
inadequate dietary pattern
Priority VI. Postpartum women, not breastfeeding, at
nutritional risk based on cither medical or dietary criteria
Priority VII. Previously certified participants whose nutritional
status might regress without the continued provision of
supplemental foods
The state agencies are required to use Priorities I through VI, and, at
their option, they can include Priority VII.
Program eligibility depends on categorical eligibility, income level, and
evidence of nutritional risk. To be categorically eligible for WIC, the
applicant must be (1) a pregnant woman, (2) a breastfeeding woman less
than one year postpartum, (3) a nonbrcastfeeding woman less than 6
months postpartum, (4) an infant up to 1 year of age, or (5) a child
between 1 and 5 years of age. States have the option to set income
eligibility between 1(K) and 185 percent of the federal poverty level,
provided that the income level is no lower than the income level for free
or reduced-price health services. All states have set income eligibility at
185 percent of the poverty level. Participants must be determined to be
at nutritional risk through a medical or nutritional assessment. Risk
factors include both medical risks, such as anemia, extremes of leanness
or obesity, high or low maternal age, or poor pregnancy history, and
dietary risks resulting from poor dietary patterns. Specific risks are
defined by the states within broad federal guidelines.
The WIC Program provides three types of benefits: (1) supplemental
food, (2) nutrition education, and (3) referrals to health care providers.
Supplemental food is usually provided in the form of a "food instrument"
(either a voucher or check), which can be exchanged for food in a store.
The food instrument lists the quantities of specific foods (including brand
names in some cases) that it can be used to purchase. The 1978
amendments to the Child Nutrition Act (PL 95-627) mandated that
supplemental foods should contain the nutrients that are lacking in the
diets of the populations targeted by WIC protein, vitamin A. vitamin C,
calcium, and iron.
Nutrition education is offered as a means of improving the nutritional
status of participants. The local agencies must spend at least one-sixth of
WIC administrative funds on nutrition education and counseling. At least
two nutrition education sessions must be provided in each 6-month
3
certification period (or at the rate of once per quarter for infants certified
for a period in excess of 6 months). However, participants cannot be
denied food supplements if they do not attend the nutrition education
sessions.
To qualify as a WIC provider, the local agency must show that accessible
health care facilities for low-income women, infants, and children arc
available. Clients must be advised about the types of health care
available, the location of health care facilities, how to receive health care,
and why it is useful. Routine health services and/or pcdiatric care arc
provided at nearly half of the WIC service sites (Williams ct al. 1990).
However, WIC funds cannot be used directly to provide health care to
participants.
STUDY The objectives of this study are to (1) examine patterns of WIC
OBJECTIVES participation among pregnant and postpartum women, and infants up to
6 months postpartum, (2) compare the characteristics and experience of
WIC participants, income-eligible nonparticipants, and higher-income
(ineligible) nonparticipants, and (3) estimate the effects of WIC
participation on birth outcomes such as birthweight, gcstational age,
and infant mortality. Specifically, the analysis addresses the following
four research questions:
1. How do prenatal WIC participants, income-eligible
nonparticipants, and higher-income nonparticipants compare
with respect to the following?
• Demographic and socioeconomic characteristics
• Behavior and experience before and during the target
pregnancy (health behaviors, pregnancy history, experiences
in the WIC Program)
• Outcomes of the target pregnancy
• Behavior and experience of the prenatal WIC participants
and their infants following delivery (infant health status,
infant health care utilization, infant feeding, postpartum
maternal and infant WIC participation)
2. How do postpartum WIC participants, income-eligible
nonparticipants, and higher-inc nc nonparticipants compare
4
with respect to demographic and socioeconomic
characteristics?
3. How do infant WIC participants, income-eligible
nonparticipants, and higher-income nonparticir ints compare
with respect to the following?
• Demographic and socioeconomic characteristics
• Infant health status and health care utilization
• Infant feeding practices
4. What are the effects of prenatal WIC participation on birth
outcomes including birthwcight; gestational age; and the
incidence of low birthweight, very low birthwcight, preterm
birth, neonatal mortality, and infant mortality through 6
months after birth?
In answering these questions, the analysis uses the 1988 NMIHS. The
WIC program has expanded since 1988 and it is possible that the
composition of the participant population may have changed since the
period covered by the NMIHS. However, the NMIHS remains the best
data source currently available to address these questions. The NMIHS
is a rich, nationally represcntaitvc database that includes data on maternal
and family characteristics, WIC participation, and maternal and infant
outcomes. Many variables in the NMIHS are not available in other data
sources that distinguish WIC participants. Furthermore, it is the most up-to-
date data available.
BLANK PAGE
II. REVIEW OF PREVIOUS LITERATURE
Since its inception in 1972, the WIC Program has grown immensely in
part because, at least on an intuitive level, the provision of food
supplements to at-risk pregnant women, infants, and children seems like
it should result in improved pregnancy and health outcomes. However,
as noted by Kennedy et al. (1982), although the benefit of nutritional
supplementation has been demonstrated in underdeveloped countries, its
efficacy in industrialized countries is continually questioned. In addition,
both the size and growth of the WIC Program have caused policymakers
and lawmakers, as well as the scientific community, to focus a great deal
of attention on quantifying its benefits in order to determine whether, as
a publicly funded intervention, the WIC Program indeed provides a
measurable net benefit to society.
Thus, the 1970s and 1980s have witnessed numerous evaluations of the
WIC Program that vary in scope and the outcomes examined. Many
evaluations have been performed at the state and local level, and some on
the national level. Outcomes examined have included birthweight, fetal
and neonatal mortality, medical conditions and nutritional status in the
mother and infant, and Medicaid cost savings at and around birth for
prenatal WIC participants and their newborns. Some evaluations have
also sought to determine whether WIC differentially affected particular
subgroups of the population (such as nonwhites or those with high-risk
conditions or risk behaviors like smoking) or whether the intensity of WIC
participation, such as the number of months in the WIC Program or the
number of WIC food instruments redeemed, influenced its effectiveness.
Each evaluation has been scrutinized and criticized by subsequent
investigators. This chapter reviews previous evaluations of the WIC
Program that are relevant to topics covered in this report, focusing on
studies of the effects of prenatal participation in the WIC Program, but
also briefly considering studies of the effects of WIC on health care
utilization among infants and children.
STUDIES OF
THE EFFECTS
OF PRENATAL
WIC
PARTICIPATipN
Studies of the effects of WIC participation during the prenatal period
include several major evaluations as well as reviews of these
evaluations. Both the evaluations and reviews are discussed below.
Major
Evaluations
The methodologies and findings of eight major evaluations of the WIC
Program published between 1979 and 1991 are summarized in Table
II. 1. The following information is restricted to those aspects of the
studies concerning the effects of prenatal WIC participation.
00 TABLE U.I
SUMMARY OF MAJOR WIC EVALUATIONS
Authors (Year Published) Comparison Group(s)
Selected Outcomes
I xi mined
Summary of Reported Impacts
for Outcomes Relevant
to Current Study Data Source
J. Edozien, B. Switzer, and R.
Bryan (1979)
WIC participants at enrollment
with characteristics of current
WIC participant group
Birthweight
Infant mortality
Growth of children
Increased birthweight More than 50,000 women, infants,
and children in 19 WIC projects in
14 slates; clinical examinations and
laboratory tests carried out between
1973 and 1976
E Kennedy, S. Gershofr, R.
Reed, and J. Austin (1982)
M. Kotelchuck. J. Schwartz, M.
Anderka, and K. Finison (1984)
J. Metcoff, P. Costitoe. W.
Crosby, S. Outta, II Sandstead,
D. Milne. C. Bodwell. and
S. Majors (1985)
Pregnant WIC applicants not
certified because program had no
slots or those who applied and
were certified postpartum
Pregnant women at non-WIC
health facilities
Non-WIC pregnant women
matched on demographics to WIC
participants
Randomly assigned comparison
group
/hernia and other measures of
nutritional status
Birthweight
Birthweight
Infant mortality
Gestational age
Use of prenatal care
Birthweight
Maternal nutritional status
Increased birthweight (3.273 vs.
3,136 grams) thai also increased
with the number of WIC
vouchers received
Decrease in percent low
birthweight (6.9 vs. 8.7%)
Nonstatistically significant
increase in birthweight (3.281 vs.
3,260 grams)
Increased WIC participation
associated with larger impacts
Decreased infant mortality
Improvement in use of prenatal
care
Increase in birthweight (3,254 vs.
3,163 grams)
Medical and nutrition records for
1.307 live births (897 to WIC
participants, 410 to non WIC) at 9
sites in Massachusetts between 1973
and 1978
Birth and death certificates and WIC
data for 8,252 WIC and non WIC
births in Massachusetts in 1978
Clinical data for 824 WIC-eligible
pregnant women attending
Oklahoma prenatal clinics
r
TABLE II. 1 (continued)
Authors (Year Published) Comparison Group(s)
<VI« <-d Outcomes
Examined
Summary of Reported Impacts
for Outcomes Relevant
to Current Study Data Source
W. Schramm (1985, 1986,
1989)
Medicaid-covered births to WIC
nonparticipants
Medicaid costs within 30 days
of birth f< !9M data; within 45
days for 1982 and 1985-86
Birthweight
NIC'U admissions
For 1980 Medicaid births, $.83
reduction in Medicaid costs for
each dollar spent on the
prenatal WIC program; $.49 and
$.79 in 1982 and 1985-86.
respectively
7,628 Medicaid births in Missouri in
1980; 9.086 Medicaid births in 1982;
and 17,944 Medicaid births in 1985
and 1986.
B. Devaney, L. Bilheimer, and
J. Schore (1990, 1991)
Medicaid-covered births to WIC
nonparticipants
J. Stockbauer (1986, 1987) WIC nonparticipants
Medicaid costs from birth to 60
days after birth
Newborn birthweight
Gestalional age
Birthweight
D. Rush (1986) Low-income, first-time, non-WIC
registrants at prenatal clinics
Birthweight
Fetal mortality
Savings in maternal and
newborn Medicaid costs per
dollar spent on WIC ranged
from $1.77 to $3.13.
Increase in birthweight ranged
from 51 grams to 117 grams
In 1980, mixed effects on
birthweight (depending on
method of comparison), with
consistently more favorable
outcomes among black WIC
participants; 1982 found small
but consistently favorable effects
Both studies found at least 7
months of participation required
to observe improved birthweight
No effect on birthweight
Nonstatistically significant
decrease in fetal deaths
More than 105.000 Medicaid births
in 1987 from 5 slates: Florida,
Minnesota. North Carolina, South
Carolina, and Texas (January-June
1988 in Texas)
1986 study us<-d 1980 data on 6,732
births lo prenatal WIC participants
in Missouri and 5.574 to 6.657 non-
WIC births; 1987 study used 9,411
WIC and 9,411 non-WIC 1982 births
5,205 prenatal WIC participants and
1,358 non-WIC registrants from 174
WIC sites and 55 clinics across the
county
IO ?
The earliest evaluation cited in Table II. 1. Edozien et al.
(1979), was a national effort that involved more than 50.000
women, infants and children at 19 WIC projects in 14 states.
Outcomes from clinical examinations and laboratory samples
collected between 1973 (just a year after the inception of
WIC) and 1976 for WIC participants were compared with
similar outcomes for new WIC enrollees collected at the time
of their enrollment. WIC participation resulted in an increase
in birthweight.
Kennedy et al. (1982) compared medical and nutrition records
collected between 1973 and 1978 for the births of 897 WIC
participants with those of 410 pregnant women on WIC
waiting lists or receiving health services at non-WIC facilities
at nine sites in Massachusetts. WIC participants had infants
with higher average birthweights than did nonparlicipants
(3,273 grams and 3,136 grams, respectively).
Kotelchuck et al. (1984) examined 4,126 pairs of births for
WIC participants and nonparticipants matched on maternal
age, race, parity, education and marital status. Data for the
sample were obtained from 1978 birth and death certificates
and WIC Program records in Massachusetts. A small,
nonsignificant increase in birthweight (from 3,260 to 3,281
grams) was estimated, as was a statistically significant decrease
in the percent of low birthweight babies (from 8.7 to 6.9
percent), a decrease in infant mortality, and an improvement
in the use of prenatal care. The estimated WIC impacts
increased with the length of WIC participation.
Metcoff et al. (1985) randomly assigned half of a sample of
824 WlC-eligible pregnant women attending Oklahoma
prenatal clinics to a WIC treatment group and to a control
group. This was the first use of random assignment in a major
evaluation, although the universe of prenatal clinic enrollees
used for sampling was intrinsically restricted to women with a
commitment to the use of prenatal care. WIC participants had
higher average birthweights than controls (3,254 grams versus
3,163 grams).
Schramm (1985, 1986, and 1989) examined the effect of WIC
participation on Medicaid costs after birth in Missouri at three
points in time-1980, 1982, and 1985-86. For 1980, Schramm
estimated a savings of $.83 in newborn Medicaid
reimbursements within 30 days after birth for each dollar spent
10
on the prenatal component of the WIC Program; in 1982 and
1985-86, the estimated Medicaid savings for services received
within 45 days after birth were $.49 and $.79, respectively.
Mean birthweight was 6 grams greater for WIC participants
than for nonparticipants in 1980. compared with differences of
31 grams and 25 grams in 1982 and 1985-86, respectively.
Devaney et al. (1990 and 1991) also examined the effects of
prenatal WIC participation on Medicaid costs after birth in
five states-Florida, Minnesota, North Carolina. South
Carolina, and Texas. The study period was 1987 for Florida.
Minnesota, North Carolina, and South Carolina; and January
through June 1988 for Texas. The estimated savings in
Medicaid costs within 60 days after birth for each dollar spent
on the prenatal WIC Program ranged from $1.77 to $3.13.
Average newborn birthweight was higher for WIC participants
than for nonparticipants, ranging from an increase of 51 grams
to 117 grams. Prenatal WIC participation was also associated
with a lower incidence of low birthweight, longer gestational
age, and a lower incidence of a preterm birth.
Stockbauer (1986 and 1987) compared 1980 and 1982 Missouri
birth records for WIC participants with those of women not
participating in WIC. The first study compared 6.732 WIC
births with three comparison samples of between 5,574 and
6,657 births drawn from non-WIC births in Missouri vital
records; although income information was not available,
statistical techniques were used to control for other differences
in characteristics. The second study compared 9,411 pairs of
WIC and non-WIC births matched on key maternal
characteristics. The 1980 study estimated mixed overall effects
on birthweight depending on the comparison sample used, but
it consistently estimated favorable outcomes for black WIC
participants. The 1982 study found that prenatal WIC
participation was associated with a reduction in the percentage
of infants with low birthweight and a reduction in the
percentage of women having inadequate prenatal care. Both
studies found that at least 7 months of WIC participation were
required to observe improved birthweight outcomes.
Rush (1986) compared longitudinal data on 5,205 prenatal
WIC participants and 1,358 non-WIC registrants at prenatal
clinics selected from 174 WIC sites and 55 clinics across the
country. The primary Findings concerning the effects of
prenatal WIC participation were no statistically significant
11
effect on newborn birthweight; increased infant head
circumference; increased birthweight and head circumference
with better WIC Program quality; lower incidence of fetal
death and low birthweight of appreciable but not significant
magnitude; and increased intake of protein, iron, calcium, and
vitamin C (four of the five targeted WIC nutrients).1
These evaluations shared a number of features. Each examined the ability
of WIC participation to increase birthweight. There are two primary
reasons for looking at birthweight as an outcome. The first is that low
birthweight predicts subsequent short- and long-term health problems in
newborns, such as respiratory difficulties and developmental disabilities
(Institute of Medicine, 1985). The second is that birthweight is a
relatively reliable quantitative measure that is routinely available on birth
certificates, a major data source for these studies.
Each evaluation also identified a comparison group against which to
compare outcomes, such as birthweight, for WIC prenatal participants.
Ideally, the goal in selecting a comparison group is to identify a sample of
women who are identical to WIC prenatal participants except for their
participation in the program in order to see what would have happened
to the WIC participants in the absence of the WIC Program. Identifying
such a group is difficult. As a result, researchers are confronted with the
problem of interpreting differences in outcomes for WIC participants and
nonparticipant comparison groups in light of the measured and
unmeasured differences that might have existed between the two groups.
For example, most of the evaluations reported that prenatal WIC
participation favorably affected birthweight, and that the effects varied in
size and level of statistical significance. However, a critical question, and
one that seems to dominate critiques of these evaluations, is the extent to
which a significant increase in birthweight (or conversely, the lack of a
significant increase) is an artifact of the comparison group and not a
function of program participation.
'The historical study by Rush (1986) used aggregate county-level data for
the years 1972-1980 to relate WIC penetration rates (estimates of the
proportion of eligible pregnant women enrolled in WIC in the county) to
average birth outcomes, fetal death rates, and infant death rates in each
county. This study found effects on birth outcomes in the expected
direction, but most were not statistically significant. Because this study
raises very different methodological issues than other WIC studies, it is
not discussed in detail here.
12
A number of the studies also investigated a so-called "dose-response
effect" for WIC participation. The hypothesis of the dose-response effect
is that the greater the extent of a woman's participation in the WIC
Program, the greater the size of the program's effects. The extent of
participation is frequently measured in terms of months of WIC
participation or the number of WIC vouchers received or redeemed. A
corollary to the dose-response hypothesis is that there is a threshold below
which WIC participation can be expected to have no statistically
significant effect on birth outcomes. The dose-response issue is of interest
to researchers and policymakers for a number of reasons including its
implications for targeting WIC outreach efforts according to the
gestational age of the fetus. However, as discussed in the following
sections, estimation of a dose-response effect for WIC participation is not
straightforward.
Evaluation WIC evaluations have been scrutinized with unusual care. In response
Reviews to a request from the chairman of the Senate Committee on
Agriculture, Nutrition and Forestry, the General Accounting Office
(1984) produced a review of existing evaluations of WIC. The review
focused on three claims of previous studies:
1. That WIC participation decreases the rate of miscarriage,
stillbirth, and neonatal death, and improves maternal nutrition
2. That WIC participation is related to improved pregnancy
outcomes for high-risk mothers and that improved outcomes
are directly related to length of participation in the WIC
Program
3. That WIC participation results in increased birthweight and
reduces the chances for anemia and mental retardation in
infants and children
The General Accounting Office (GAO) reviewed the findings of the
evaluations as well as the quality of the methodology. GAO summarized
its review of the studies by saying, The information is insufficient for
making any general or conclusive judgments about whether the WIC
Program is effective or ineffective overall. However, in a limited way, the
information indicates the likelihood that WIC has modestly positive effects
in some areas" (GAO 1984, p. ii.).
13
In particular, GAO found that many studies examined the effect of WIC
participation on birthweight and judged several of these studies to be of
high or medium quality. The studies supported, but did not give
conclusive evidence for, the ability of the WIC Program to increase
birthweight. The studies cited a decrease in the rate of low birthweight
(that is, birthweight below 2,500 grams) from 9.5 percent for
nonparticipants to 7.9 percent for WIC participants and an increase in
mean birthweight of 30 to 50 grams. GAO found evidence supporting
claims that the program produced more favorable effects on birthweight
of newborns for teenagers, blacks, and those with several health- and
nutrition-related risks. However. GAO found only inconclusive evidence
for the claim that longer WIC participation increased its effectiveness
regarding birthweight. GAO found substantially less data to support
claims of decreased fetal and neonatal deaths than it found for claims
related to birthweight. (The former two outcomes are particularly difficult
to study because they occur relatively infrequently.)
David Rush has critiqued many of the WIC evaluations in terms of their
methodological limitations. In his review (1982) of the study by Edozien
et al. he noted three primary limitations: (1) it was performed too early
in the life of the WIC Program to judge definitively its effectiveness; (2)
the comparison group was not valid because it assumed comparability
between those participating at different times in pregnancy or in the life
cycle, as well as between those arriving early in the program with those
arriving later; and (3) the authors failed to note in their claim of a dose-response
effect for WIC participation that duration of prenatal WIC
participation was confounded by duration of gestation.
In his review of the Kotelchuck et al. study, Rush (1984) found
shortcomings both in the study design and the analysis. He noted that the
Kotelchuck study excluded 353 women who were terminated from the
WIC Program for any reason; it is possible that this group of 353 included
women who had experienced premature delivery, which would potentially
bias the study's results by eliminating some women from the WIC group
who had low birthweight babies or whose newborns died as a result of
prematurity. In addition. Rush noted that the WIC and non-WIC samples
were matched on age, race, parity, education, and marital status of the
woman (as recorded on birth certificates), but not on income and the
additional health and nutritional risk factors that are the eligibility criteria
for WIC. As a result, the WIC group could have been at higher risk of
poor perinatal outcomes than the comparison group, which would bias the
result in the opposite direction to the bias associated with excluding the
353 terminees. Rush also took issue with Kotelchuck's estimates of the
dose-response effect of WIC. As he pointed out regarding the Edozien
14
study, the number of months of WIC participation is confounded with the
duration of pregnancy.
Rush (1985) viewed Schramm's 1985 investigation of the effect of WIC
on Medicaid costs for newborns as an important contribution to WIC
Program evaluation literature. However, he raised issues with some
aspects of Schramm's findings. In particular, only 21 percent of the
Medicaid births that formed the base of Schramm's sample could be
identified with WIC records. Given the fact that most pregnant women
on Medicaid are likely to be eligible for WIC as well, one could expect
the WIC participation rate to be higher. Rush was concerned that this
low match rate was a result of a flaw in the analysis file creation process.
However, it is also possible that it reflects shortcomings in WIC outreach
efforts to Medicaid recipients, some other problem concerning access to
WIC clinics, or perhaps the attitudes of Medicaid recipients in Missouri
to the WIC Program.
STUDIES OF
THE EFFECTS
OF WIC ON
INFANTS AND
CHILDREN
A large body of literature exists on the impact of WIC on pregnant
women and such birth outcomes as birthweight. Many fewer studies
have examined the impact of WIC on infants and children, and most of
these studies were performed on small samples of children in a local
area. Rush (1986) outline four reasons for the lack of such
evaluations. First, the impact of WIC may not be evident until a
number of years after the child has enrolled in WIC. For example,
most tests of psychological development do not have great predictive
validity for later cognitive performance until the child is at least 4 years
old. Second, it is extremely difficult to find a comparison group that is
truly similar to the treatment group and does not differ from the
treatment group by unmeasured factors related to participation in WIC.
Third, no consensus exists on the correct indicators of the success of
the program. For example, each study of the impact of WIC on
behavioral and cognitive development uses different measures of
development. Fourth, there are few pre-existing data sources on the
health status of children.
Effects of WIC on
the Utilization of
Health Care
Services
This study compares health status and health care utilization of WIC
infants and nonparticipants. Among previous studies, only Rush (1986)
and Paige (1983) examined the impact of WIC participation on the
utilization of health care services.
Paige (1983) collected data on 145 WIC infants in three
counties on the Eastern Shore of Maryland who were enrolled
15
in WIC during the first 3 months after their birth. These data
were compared to data on 213 WIC-eligible infants attending
public health departments in two contiguous counties that did
not serve WIC. Paige examined physical measurements, blood
iron content, and the children's immunization records.
The National WIC Evaluation, Rush (1986). was a large
nationwide study of the impact of WIC on pregnant women,
infants, and children. One component of the evaluation was
the Study of Infants and Children. This study compared cross-sectional
data on 1,459 infants and children up to 4 years of
age who were either current or past WIC recipients with data
on 683 infants and children of the same ages who were neither
current nor past WIC recipients. All infants and children in
both the treatment and comparison groups were children of
women participating in the Longitudinal Study of Pregnant
Women, another component of the evaluation. The Study of
Infants and Children examined a wide variety of outcomes
including physical measurements, psychological development,
nutrient intake, and use of health care services.
These studies have some design problems that have limited their ability to
determine unambiguously the impact of the WIC Program on infants and
children. First, the Paige study used only a small sample of infants and
children from one local area. It is therefore difficult to generalize the
findings of this study to the impact of a national WIC Program.
Second, both studies used a comparison group approach to assess the
impacts of WIC participation. If there were important differences
between the unmeasured characteristics of infants and children in the
WIC group and infants and children in the non-WIC comparison group,
the true impact of the program cannot be distinguished from the impact
of the differences in these characteristics.
Rush (1986) used as their comparison group infants and
children who were income-eligible for WIC but who did not
participate in WIC. But while all the infants and children in
the treatment group must have been determined to be at
nutritional risk in order to be WIC-eligible, some children in
the comparison group may never have been at nutritional risk.
Hence, if the infants and children in the comparison group
were on average more healthy than the infants and children in
16
the treatment group, the results of this comparison could lead
to an underestimate of the program impact.
Paige (1983) used as a comparison group infants who were
fully eligible for WIC but lived in a county not served by WIC.
However, while the parents of all the infants in the treatment
group decided that their infants should participate in WIC. this
choice was never available to parents of the infants in the
comparison group. Therefore, the infants in the two groups
may have differed in a wide variety of ways not measured in
the data but related to program participation, such as their
access to health care and their parents' attitudes and
preferences.
Rush et al. examined the impact of WIC on six measures of the use of
health care by infants and children:
1. Whether the child had a regular source of health care. The
mother was asked where she usually took the child if he or she
was sick. A response of "I don't know" or "the hospital
emergency room" was coded as "no regular source of health
care."
2. Whether the child had received preventive health care
(defined as a regular checkup or immunization) within the past
year.
3. Whether the mother had a record of the child's immunization.
4. Whether the child had received a measles vaccination.
5. Whether the child had received a diphtheria/pertussis/tetanus
(DPT) vaccination.
6. Whether the child had received a polio vaccination.
Paige examined the impact of WIC participation on the likelihood that a
child had received a DPT or a polio vaccination.
Rush et al. found that children who received WIC benefits were
significantly more likely to have a regular source of health care than non-
WIC children. However, since the Study of Infants and Children collected
17
cross-sectional data, we cannot determine whether WIC participation
increased the use of health care services or whether the use of health care
services encouraged participation in WIC. No significant relationship was
found between WIC participation and the use of preventive health care
by infants or children.2
Rush et al. found a positive relationship between WIC participation and
the proportion of infants and children who had received some
immunizations, but the relationship was only significant for some
subsamples of children stratified according to when they first received
WIC benefits. Children who received WIC benefits after their first
birthday were 11 percent more likely to have an immunization card and
11 percent more likely to have had a measles vaccination than non-WIC
children. Children who received WIC benefits were also more likely to
have received a DPT vaccination, but this relationship was only significant
for infants. Children whose mothers received WIC benefits while
pregnant were more likely to have received a polio vaccination.
In contrast, Paige found no significant relationship between WIC
participation and polio vaccination, and a negative relationship between
WIC participation and DPT vaccination. Paige speculated that a higher
proportion of WIC infants may have obtained immunizations from other
sources; WIC infants also had more missing data.
2In the Longitudinal Study of Pregnant women, Rush et al. also found no
relationship between WIC participation and the frequency of prenatal
visits by pregnant women. However, in the Historical Study, there was a
significant positive association between WIC participation and first
trimester registration of prenatal care and a significant negative
association between WIC participation and inadequate prenatal care.
18
III. DATA AND METHODOLOGY
This .,iapicr provides an overview of the data and methodology used in
this study. The data are from the 1988 National Maternal and Infant
Health Survey (NMIHS), a large, nationally representative survey of
mothers who experienced a birth or fetal death in 1988. The NMIHS is
the best available data source with which to assess the characteristics of
WIC participants and nonparticipants, and the effects of WIC
participation on birth and infant outcomes. The analyses of these data in
this report include descriptive tabulations and multivariate analyses. For
both types of analyses, important methodological issues include the choice
of a comparison group, the choice of the sample to be analyzed, and the
correct methods for estimation of means, proportions, and standard errors
given the stratified sample design.
THE DATA The 1988 NMIHS was sponsored by the National Center for Health
Statistics (NCHS) to provide a database for the analysis of factors
associated with pregnancy outcomes, child growth and development,
and the use and cost of health care services. The 1988 NMIHS
database contains data from three sources: (1) randomly selected
nationwide samples of birth, fetal death, and infant death certificates
(vital records); (2) mothers identified from these records who
responded to questionnaires that were mailed to them or telephone
follow-ups (referred to as the maternal survey); and (3) hospitals at
which the mothers were admitted for delivery, and individuals and
institutions who provided prenatal care to mothers and health care to
mothers or infants up to 6 months postpartum (referred to as the
provider survey).1 This study uses data from the vital records and the
1988 maternal survey.
Data from the vital records include the following information. For the
live-birth sample, the birth certificate contains the gender, birthweight,
and gestational age of the newborn; the age, education, race, and state of
residence of the mother; the state in which the delivery occurred; the
number of prenatal visits and the month of pregnancy in which prenatal
care began; and the number and outcomes of any previous pregnancies.
Similar information is available for the fetal-death sample from the report
of the fetal death. For the infant-death sample, the death certificate
contains demographic characteristics of the infant and parents; and the
date, place, and cause of death.
'The 1991 Longitudinal Follow-up reinterviewed the mothers in the
NMIHS sample approximately two-and-one-half years after the initial
survey was conducted. These data are not yet available.
19
The maternal survey collected data from the samples of women identified
in the vital records: (1) 9,953 mothers who experienced a live birth in
1988, (2) 3,309 mothers who experienced a fetal death of at least 28
weeks gestation in 1988, and (3) 5,332 mothers who experienced an infant
death (children up to a year old) in 1988.2 Infants with low birthweight
were oversampled in the live-birth sample, and blacks were oversampled
in all three samples. The final NMIHS data are representative of 48 states
plus the District of Columbia.3
The NMHIS maternal survey provides the most recent nationally
representative data on the characteristics and experiences of pregnant
women including: the use and source of prenatal care; participation in
WIC and other assistance programs; smoking habits, alcohol consumption,
and illegal drug use before and during pregnancy; socioeconomic and
demographic characteristics; employment before and after delivery;
maternal and infant health and use of medical care during the first 6
months postpartum; infant feeding practices; and other pregnancies. The
questions about the WIC Program included whether mothers participated
in and the duration of participation in WIC during pregnancy, and
whether the mother and infant participated in WIC and the duration of
participation up to 6 months postpartum. Information on WIC
participation was obtained both for the target pregnancy and for other
pregnancies.
NCHS contracted with the Bureau of the Census to collect the data.
Sampled mothers were mailed a 35-page questionnaire, a brochure
describing the objectives and importance of the survey, and a prepaid
return envelope. Mothers who did not respond to the initial mailing of
the questionnaire received a second mailing, followed by a postcard
2Data were also collected on a supplemental sample of American-Indian
women who had a live birth in 1988 and on a supplemental sample of
Hispanic women in Texas who had a live birth, fetal death, or infant death
in 1988. However, these supplemental samples were not used in the
analysis.
'The final sample does not include any cases from Montana because state
officials refused to allow NCHS access to their vital records. In addition,
the sample for South Dakota is incomplete because state officials provided
vital records for events (i.e., live births, fetal deaths, and infant deaths)
that occurred through July 1988 and then refused to participate further.
The sample from South Dakota is therefore not representative of all
events that occurred throughout 1988. The sample does not include
observations from Puerto Rico or other U.S. territories.
20
reminder, and if necessary, they were contacted for a telephone or
personal interview. The final response rates for the three national
samples were 74 percent for mothers with live births. 69 percent for
mothers with fetal deaths, and 65 percent for mothers experiencing infant
deaths. Mothers completed the survey between 6 months and 30 months
after the birth, with a median interval of 16 months. Only about 20
percent of the maternal surveys were completed within a year of the birth.
As discussed in Chapter II, most previous studies of WIC participants
have relied on birth certificate data and/or program or clinical records.
The NMIHS is unique in that the birth certificate data are supplemented
by the rich detail collected in the maternal survey. It is also the only
nationally representative database including both WIC participants and
nonparticipants.
METHODOLOG- The analysis of data from the 1988 NMIHS has two parts: (1) a
ICAL ISSUES descriptive analysis that compares the characteristics and outcomes of
WIC participants with income-eligible nonparticipants and higher-income
nonparticipants, and (2) a multivariate analysis of the effects of
WIC participation on birth outcomes. In this section, methodological
issues relevant to both parts of the analysis are discussed:
Selecting a comparison group
Determining the sample to be employed in the analysis
Weighting sample observations and computing standard errors
under the stratified sample design
21
Selecting a An appropriate comparison group is essential to the analysis of the
Comparison effects of WIC participation on maternal and infant outcomes.
Group However, selecting an appropriate comparison group for an ongoing
program like WIC is extremely prohlematic. Ideally, the comparison
group should consist of individuals who, as a group, are identical to
WIC participants in all ways except for participation in the WIC
Program. Our basic approach in this study is to define the comparison
group as income-eligible nonparticipants. It must be recognized,
however, that income-eligible nonparticipants may differ from WIC
participants in multiple ways that may affect the outcomes of interest.
For example, even among pregnant women who are income-eligible for
WIC, some nonparticipants may not be at nutritional risk and.
therefore, not qualify for the program. Alternatively, some pregnant
nonparticipants may be at higher risk than participants for a poor
pregnancy outcome because they do not believe in the efficacy of
prenatal care and have therefore not applied for WIC benefits, or
because access to the program is a problem for them. Similar issues
arise in comparing postpartum women and infant participants to
income-eligible nonparticipants.
Identifying WIC participants from the 1988 NMIHS data is a
straightforward process, since the data reflect each woman's answer to two
questions: (1) whether she received WIC food during pregnancy, and (2)
whether she and her infant received WIC food during the 6 months
following delivery. If a respondent answered "yes" to the first question,
she is identified as a prenatal WIC participant; if she answered "yes" to
the second question, she or her infant can be identified as a WIC
participant after the birth. Additional questions on the number of months
receiving WIC for the mother and baby in the postpartum period allow us
to distinguish postpartum and infant WIC participants.
Income-eligible nonparticipants are identified from the data by comparing
the incomes of nonparticipants to the WIC income-eligibility standard of
185 percent of the federal poverty level. However, some approximations
are required in making this comparison, because the only income data
collected in the 1988 survey reflect the pretax income of the household
during the 12 months prior to delivery. Information was not collected on
income received during the 6 months following delivery or on the rate of
income receipt. Thus, the survey does not contain an ideal set of income
data for determining WIC income-eligibility for nonparticipating women
and infants throughout the entire period covered by the survey.
Nonparticipants who arc income-eligible for WIC are identified, both
during pregnancy and during the 6 months following delivery, by using the
reported total household income on the file as a proxy for the household
22
income that would be used in the actual eligibility determination at any
given point in the period of observation.
More specifically, WIC nonparticipants are classified as income-eligible for
WIC during pregnancy if one or more of the following conditions arc
satisfied:
• Reported household income for the 12 months prior to
delivery was less than or equal to 185 percent of the poverty
level.
• The woman reported Medicaid as a source of payment for her
prenatal care.
• The household received AFDC or food stamp benefits during
the 12 months prior to delivery.
For births that occurred before July 1,1988, the federal poverty guidelines
effective July 1, 1987 through June 30, 1988 arc used to determine
income-eligibility for WIC. For births that occurred on or after July 1,
1988, the federal poverty guidelines effective July 1, 1988 through June
30. 1989 arc used.
Income eligibility for WIC also depends on household size, since this
determines the appropriate poverty threshold for the household. The
1988 NMIHS questioned women about the size of their household (1)
during most of their pregnancy and (2) at the time of the interview.
Household size reported during pregnancy is used to determine income
eligibility during the woman's pregnancy, and household size reported at
the time of the interview is used to determine income eligibility during the
6 months following delivery.
The descriptive analysis in Chapter IV assesses the similarity of WIC
participants, income-eligible nonparticipants, and higher-income
nonparticipants in terms of a wide range of characteristics available from
the survey and the birth certificate, including demographic and
socioeconomic characteristics; use and source of prenatal care; and
behavioral risk factors such as smoking, alcohol consumption, and illegal
drug use during pregnancy. These comparisons do not show the effects
of the WIC Program, but are used to help interpret the differences in
observed outcomes.
23
Multivariate analysis techniques arc used in Chapter V to control for
differences in the observed characteristics of WIC participants and
income-eligible nonparticipants. However, these estimates may not
control for differences in unobserved characteristics that affect key
outcomes. Attempts to apply statistical methods to control for unobserved
differences (selection bias) were not successful, as discussed more fully in
Chapter V. Thus, even after controlling for observed differences between
participants and income-eligible nonparticipants, differences in outcomes
may be due to either the effects of WIC or to unobserved factors
associated with WIC participation.
Determining the
Analysis Sample
The NMIHS maternal survey consists of three samples: the live-birth
sample, the fetal-death sample, and the infant-death sample. All of the
analyses in this report use the live-birth sample. Infant mortality is
examined using the live-birth sample because infant mortality is defined
as the number of infant deaths per 1,000 live births. Because the live
birth sample includes a representative sample of infant deaths, it was
not necessary to include the supplemental infant death sample. In
principle, one could combine the fetal-death and live-birth samples to
produce descriptive tables on the entire population of women who
experienced a pregnancy in 1988. In reality, however, fetal death is
such a rare event that when weighted, the inclusion of the fetal-death
sample would have a negligible effect on the descriptive statistics. In
addition, combining the two samples would increase the design effect,
and thus reduce the power to detect statistically significant differences
between WIC participants and income-eligible nonparticipants.
Therefore, all estimates in this report are computed for the live-birth
sample only.
Weighting Sample
Observations and
Computing
Standard Errors
The stratified sample design of the NMIHS must be considered in an
analysis of the data. The live-birth sample was stratified by race,
birthweight, and state; and the infant-death and fetal-death samples
were stratified by race and state. Blacks were oversampled in all
samples, and infants of very low birthweight (less than 1,500 grams) and
moderately low birthweight (1,500 to 2,499 grams) were oversampled in
the live-birth sample. For the live-birth sample, the sampling rates are
as follows:
Black
- Less than 1,500 grams: 1/14
- 1,500 to 2,499 grams: 1/55
- 2,500 grams and over: 1/113
24
White and all other
- Less than 1.500 grams: 1/29
- 1,500 to 2.499 grams: 1/160
- 2.5(H) grams and over: 1/720
Thus. 1 of every 14 live births among blacks that resulted in a newborn
weighing less than 1.500 grams was selected into the sample, while only
1 of every 720 live births among "nonblacks" that resulted in a newborn
of at least 2.5(H) grams was selected into the sample. These sampling rates
were applied individually to each state rather than to the national
population. Thus, within a given stratum defined by race and birthweight,
the percentage distribution across states of the original sample
(respondents plus nonrespondents) matches that of the national
population.4
All analyses for this report use sample weights constructed by NCHS to
ensure that (1) statistics generated for the total population or population
subgroups arc not biased toward the overrepresented strata and (2)
standard errors are correctly computed. The NMIHS sample weights
designed by NCHS correct for the stratified sample design as well as for
unit nonresponse. The sample weights are derived as the product of two
factors: (1) a factor to account for the different sampling rates across
strata and (2) a factor to account for different rates of unit nonresponse.
For each sample member, the first factor is the inverse of the sampling
rate for the stratum to which the individual was assigned, and the second
factor is the inverse of the response rate for the "response category" to
which the individual was assigned. Response categories were defined by
classifying women in each stratum on the basis of marital status. Thus, the
correction for unit nonresponse allows for variation in response rates by
marital status, race, and for the live-birth sample, birthweight.
To account for the stratified sample design, the software program
SUDAAN-Professional Software for Survey Data Analysis for Multi-stage
Sample Designs, Release 6.30 (Shah et al. 1991,1992)-was used to
compute weighted means, proportions, and regression coefficients, and to
compute standard errors for these statistics. SUDAAN computes standard
errors using the Taylor series linearization method. The estimation of the
standard errors accounted for the stratification by state as well as by race
and birthweight.
4The stratification by state is not mentioned in the documentation for the
NMIHS public use tape, but it was discovered through conversations with
NMIHS staff.
25
BLANK PAGE
U
IV. CHARACTERISTICS OF WIC PARTICIPANTS AND
NONPARTICIPANTS
This chapter presents findings from a descriptive analysis of the
characteristics, behavior, and experience of WIC participants, income-eligible
nonparticipants, and higher-income nonparticipants. Comparisons
of WIC participants and income-eligible nonparticipants arc used to
delineate who participates in WIC and to highlight differences in
outcomes that may be worthy of further analysis. Higher-income
n( >n pa rl it i pan ts -non participants not income-eligible for WIC-are included
in the analysis to show how low-income women and infants differ from
higher-income women and infants. None of the comparisons in this
chapter-including comparisons of birth outcomes-should be interpreted
as indicative of the effects of the WIC Program because they do not
control for other differences in the characteristics of these groups.
The first section of this chapter describes the prevalence of WIC
participation among income-eligible pregnant women, infants, and
postpartum women, and the program-related experiences of prenatal
participants. The second section examines the characteristics and
experience of prenatal WIC participants, and compares them with the
characteristics and experience of income-eligible pregnant women who did
not participate in the WIC Program during pregnancy and with those of
higher-income pregnant women. The third section presents descriptive
comparisons in which the WIC participant sample is defined on the basis
of mothers' and infants' WIC participation during the first six months
following delivery. Two sets of comparisons arc presented: postpartum
WIC participants are compared with postpartum nonparticipants, and
infant WIC participants arc compared with infant nonparticipants.
Throughout this chapter, the statistical significance of the differences
between WIC participants and income-eligible nonparticipants, and
between WIC participants and higher-income nonparticipants is presented.
For categorical variables, the distributions are compared using a chi-squarc
test, which tests whether the distribution of the variable is independent of
WIC participation status. For continuous variables, a two-tailed t-test is
used to determine whether the means for the two groups are significantly
different. Differences that are statistically significant at the .05 level or
better are noted as such in the tables.
PARTICIPATION
AND
EXPERIENCES IN
THE WIC
PROGRAM
This section describes the level and duration of participation in WIC
by pregnant women, infants, and postpartum women, and the
experiences of prenatal participants in the WIC program.
27
WIC Participation Based on the weighted NMIHS data, about 1,151,000 of the women
(Table IV.l) who gave birth in 1988 were prenatal WIC participants, about 941,000
were postpartum WIC participants, and about 1348,000 of the infants
born in 1988 were infant WIC participants. The proportion of income-eligible
persons participating in WIC was highest for infants (70
percent), next highest for pregnant women (62 percent), and lowest for
postpartum women (51 percent).1
Based on the mother's retrospective reports, prenatal WIC participants
participated for an average of 5.5 months during pregnancy. Most
prenatal participants (51 percent) entered WIC during the first trimester.
35 percent entered during the second trimester, and 14 percent entered
during the third trimester. The mothers were asked in which month of
pregnancy they began receiving WIC benefits. The mothers' reports
suggest earlier enrollment in WIC than is shown in program data. In
particular, the 1988 WIC Participant and Program Characteristics data
indicate that only 24 percent of prenatal participants entered during the
first trimester, while 48 percent entered during the second trimester, and
28 percent during the third trimester (Williams 1991). It may be that
mothers interpret the "first month" as the first month after they learn of
the pregnancy, or as the period after they have been pregnant for one
month, which is in fact the second month.
Postpartum mothers reported the total number of months that both they
and their infants participated in the program. About 12 percent of
mothers and 57 percent of infants were still participating in WIC at the
time of the maternal survey.2 (Mothers were surveyed from 6 months to
30 months after the birth, with a median interval of 16 months. Only
about 20 percent of the maternal surveys were completed within a year of
birth.) Because many were still participating, the mean duration of WIC
participation reported in Table IV.l understates the total duration of
'The estimated participation rates are, of course, sensitive to the method
used to estimate income-eligibility. Our definition of income-eligibility is
discussed in Chapter III.
2For those still participating at the time of the survey, the number of
months of participation was not recorded. We assumed these mothers
and infants had been participating since the birth, which may lead to some
overstatement of durations. Even when the number of months
participating was recorded, we do not know if participation started at
birth. In most cases, however, it is reasonable to assume participation
started with birth, since most infant and postpartum participants also were
prenatal participants.
28
TABLE IV. 1
WIC PARTICIPATION AND DURATION OF PARTICIPATION
BY PREGNANT WOMEN. INFANTS, AND POSTPARTUM WOMEN
Pregnant
Women Infanis
Postparlum
Women
Number of WIC Participants (Weighted)
Percent of Income-Eligible. Who Participated in WIC
Mean Duration of Participation (Months)
Month of Pregnancy Began Receiving WIC
(Percent Distribution)
First
Second
Third
Fourth
Fifth
Sixth
Seventh
Eighth
Ninth
Duration of Participation after Birth
(Percent Distribution)"
1-3 months
4-6 months
7-9 months
10-12 months
More than 12 months
Prenatal Participants Who Also Received WIC Postparlum
(for This Pregnancy) (Percent)
Prenatal Participants Whose Infants Received WIC
(for This Pregnancy) (Percent)
Prenatal WIC Participants (Percent)
1.151,181 1,347.960 941.244
61.7 70.1 50.9
5.5 12.8a 6.6a
14.3
16.3
20.5
14.8
11.3
8.4
6.8
3.7
3.9
n.a.
n.a.
77.1
99.3
n.a.
n.a.
n.a.
76.0
n.a.
5.3 27.2
10.3 46.7
9.8 7.3
31.2 9.7
43.4 9.1
n.a.
n.a.
84.3
Sample Size (Unweighted) 3,868 4,500 3,003
SOURCE: 1988 National Maternal and Infant Health Survey.
NOTE: All means and percent distributions are based on weighted data and are calculated using SUDAAN.
"Duration of participation between birth and interview. Duration of participation will be understated for those
still participating at the time of the interview. Approximately 57 percent of infants and 12 percent of mothers
were still participating at the time of the interview; infants may have been recertified as children and
postparlum mothers may have become pregnant again. Interviews occurred from 6 months to 30 months after
birth.
n.a. = not applicable.
29
participation, especially for infants (who may continue to participate as
children up to the age of 5) and for those surveyed early. Nonetheless,
the data indicate that 84 percent of infant WIC participants were in the
program for more than 6 months, and at least 43 percent participated for
more than a year. The average duration of participation was slightly more
than a year. Thus, participation generally lasted through much of infancy.
Most postpartum mothers (74 percent) participated in WIC for 6 months
or less. Only 17 percent participated from 7 to 12 months, and only 9
percent of postpartum WIC participants participated for more than a year
after the child's birth. (Although the program serves postpartum mothers
for at most 1 ycar--l year if breastfeeding. 6 months if not-some of the
mothers reporting longer periods of participation may have become
pregnant again.) The average duration of participation for postpartum
mothers was nearly 7 months.
More than three-quarters of prenatal WIC participants were postpartum
participants, and more than 99 percent of their infants were infant WIC
participants. Examined from the opposite perspective, roughly three-quarters
of infant WIC participants had mothers who were prenatal WIC
participants, and 84 percent of postpartum WIC participants had been
prenatal WIC participants. Thus, infant WIC participants included
essentially all infants of prenatal WIC participants plus other low-income
infants. The population of postpartum WIC participants overlapped for
the most part the population of prenatal WIC participants, but there were
a number of women in each group who were not in the other group.
Program
Experiences of
Prenatal WIC
Participants
(Table PV.2)
The NMIHS survey collected especially detailed data on WIC Program
experiences for prenatal WIC participants. Prenatal WIC participants
usually learned of the WIC Program from friends or relatives (43
percent), doctors or other health care providers (36 percent), or
previous experience with WIC (28 percent). While almost all prenatal
participants reported receiving advice concerning nutrition, the major
focus of the WIC Program, 60 percent or more reported receiving
other types of health advice in addition to nutritional advice.
Information was usually obtained through individual counseling sessions
(71 percent) or through pamphlets (52 percent). Participants less
frequently received information through classes or group sessions (37
percent) or through films and videos (32 percent).
Mothers' self-reports concerning information received from the WIC
Program are not necessarily accurate portrayals of information that was
available from the WIC Program for two reasons: (1) mothers were not
30
TABLE IV.2
PRENATAL WIC PARTICIPANTS" EXPERIENCE WITH THE WIC PROGRAM
All Prenatal
WIC Participants
Classified By Race
White Black
Mean Number of Months Received
WIC during Pregnancy 5.50
Participants Who Reported Receiving
the Following Advice at the WIC
Office (Percent)
Eat proper foods
How to use WIC fcxxls
How much weight to gain
Breastfeed baby
Feed infant formula
How to buy food
Get prenatal care
Reduce/stop smoking
Reduce/stop drinking
Avoid illegal drugs
Participants Who Reported Receiving
Information at the WIC Office through
(Percent)
5.48 5.48
Trimester of Pregnancy Began
Receiving WIC (Percent Distribution)
First 51.0 49.8 52.4 ••
Second 34.6 33.9 36.4
Third 14.3 16.2 11.2
94.1 94.4 93.5
87.2 87.5 86.3
67.3 67.6 66.2
63.7 66.2 56.2
60.2 57.7 64.7
60.9 61.0 60.8
65.8 67.3 63.5
65.3 67.8 60.6
59.7 61.2 56.8
61.7 62.0 61.5
Classes or group sessions 36.6 37.2 35.2
Individual sessions 71.2 73.2 68.4 ••
Pamphlets 51.7 53.4 48.4**
Films or videos 32.3 34.2 27.3 »•
Participants Who Reported Learning
About the WIC Program from
(Percent)
Friend or relative 43.3 46.6 36.7 •*
Doctor or other health care provider 36.3 3.3.4 41.0 ••
Social worker 7.3 7.0 8.2
Advertisement or poster 3.9 4.5 2.9
Previous experience with WIC 27.5 25.9 31.9 ••
Other 3.7 4.3 2.8
Sample Size" 3,868 1,107 2,671
31
TABLE IV.2 {continued)
SOURCE:: 1988 National Maternal and Infant Health Survey.
NOTE: All means, percent distributions, and tests of statistical significance arc based on weighted
data and are calculated using SUDAAN. Tests of statistical significance arc based on t-statistics
for the difference in means of the continuous variables and chi-square statistics
for the difference in the percent distributions of the categorical variables.
aA small number of participants (90) arc cither Asian or Native American. They are included in the
total column but not in the columns for white and black.
* (**): The difference between whites and blacks is statistically significant at the .05 (.01) level.
32
required to participate in nutrition education and related counseling
sessions; and (2) even among those who received a particular type of
advice or service, those mothers who later followed the advice may be
more likely to remember and report it. For example, mothers who later
breastfed their babies may be more likely to report receiving advice
concerning the benefits of breastfeeding.
Differences in the
Experiences of
Black and White
Prenatal WIC
Participants
(Table IV.2)
Blacks and whites differed in the timing of their entry into the WIC
Program, in the types of advice they reported receiving from the
program, and in how they learned of the program.3 Blacks entered
the WIC Program earlier in pregnancy than whites--in particular, blacks
were more likely than whites to enter in the first two trimesters, while
whites were more likely than blacks to enter in the third trimester.
Whites were more likely to report receiving advice to breastfeed their
babies, while blacks were more likely to report receiving advice to give
formula. In addition, whites were more likely to report being advised to
stop or reduce smoking, and to stop or reduce drinking. However, these
differences were not large, and may reflect differences in recall and
reporting by the mothers rather than in the advice actually offered by the
program.
Black mothers reported receiving information from the WIC Program
through fewer sources than white mothers. Black mothers were more
likely than white mothers to have learned of the WIC Program through
their health care providers or through previous experience with WIC, and
less likely to have learned of the program through word-of-mouth.
PRENATAL WIC
PARTICIPANTS
AND
NONPARTICI-PANTS
This section identifies the key characteristics of prenatal WIC
participants and compares them with those of income-eligible pregnant
nonparticipants and higher-income pregnant nonparticipants. These
characteristics include demographic and socioeconomic characteristics,
prenatal care, behavioral risk factors (alcohol use, cigarette smoking,
and drug use), previous pregnancies, and pregnancy outcomes for the
target pregnancy.
3All differences discussed in this section are statistically significant.
33
Demographic and
Socioeconomic
Characteristics
(Table IV3)
The WIC Program serves individuals who are disadvantaged on
average, even relative to the rest of the income-eligible population.
Prenatal WIC participants were more likely to be teenagers, less likely
to have a high school diploma, more likely to be black or Hispanic, and
less likely to be married or living with the baby's father than were
income-eligible nonparticipants or higher-income nonparticipants. In
addition, WIC participants were less likely to have been employed in
the previous year, and were more likely to depend on Mcdicaid and to
lack private health insurance than were nonparticipants. Nonetheless,
fully 54 percent of WIC participants had worked in the 12 months prior
to delivery (compared with 62 percent of income-eligible nonpartici-pants
and 79 percent of higher-income pregnant women).
The characteristics of the fathers showed similar patterns: the fathers of
the WIC participants' babies were younger, less educated, more likely to
be black or Hispanic, and less likely to have been employed than were the
fathers of the babies of income-eligible nonparticipants and higher-income
nonparticipants.
Prenatal WIC participants and income-eligible nonparticipants did not
differ significantly in terms of household size or mean household income,
but WIC participants were more likely to depend on public assistance
income. In addition, a larger proportion of WIC participants had incomes
below the poverty level (56 percent versus 45 percent), although WIC
participants were also more likely than income-eligible nonparticipants to
have household incomes above 185 percent of the poverty level. WIC
participants were less likely than income-eligible nonparticipants to live in
metropolitan counties, suggesting that WIC reaches a larger proportion
of the income-eligible population in rural areas.
'The latter contrast reflects the definition of income-eligible nonpartici-pants.
By definition, income-eligible nonparticipants either have house-hold
incomes below 185 percent of the poverty level or they received
AFDC, Medicaid, or food stamps. Higher-income nonparticipants have
household incomes above 185 percent of the poverty level. However,
WIC participants may have incomes exceeding 185 percent of the poverty
level, because the data on income reflect total household income during
the 12 months prior to delivery, not total household income at the time
of application for WIC.
34
TABLE IV.3
DEMOGRAPHIC AND SOCIOECONOMIC CHARACTERISTICS OF PPENATAL WIC
PARTICIPANTS. INCOME-ELIGIBLE NONPARTICIPANTS.
AND HIGHER-INCOME NONPARTICIPANTS
Prenatal WIC Income-Eligible Higher-Income
Participants Nonparticipants Nonparticipants
Mother's Characteristics
Mean Age (Years) 23.4 25.0 ** 28.4 **
Age (Percent Distribution)
Younger than 18 10.8 5.3 •• 1.0 **
18- 19 14.6 12.1 2.1
20 - 24 38.9 33.3 18.8
25-29 22.2 27.9 38.5
30 - 34 9.5 15.7 28.5
35 and older 4.0 5.7 11.2
Education (Percent Distribution)
8 years or less 8.7 1.1** 0.9 **
9-11 years 30.9 19.2 3.8
High school graduate 43.3 45.0 34.8
Some college 14.7 20.6 30.7
College graduate 2.5 7.6 29.8
Race (Percent Distribution)
White 64.0 73.3 ** 88.9**
Black 31.5 22.1 5.7
Asian or Pacific Islander 2.0 3.3 4.7
Native American 2.5 1.3 0.6
Hispanic (Percent) 20.3 15.7 •• 7.7**
Married (Percent) 48.0 64.3 •• 93.0 *•
Lived with the Baby's Father during
Most of the Pregnancy (Percent) 59.6 73.2 95.3
Employed at Any Time during 12
Months prior to Delivery (Percent) 54.0 61.9 78.8
Covered by Medicaid for Prenatal Care
and/or Delivery (Percent) 55.1 29.2 0.5
No Private Health Insurance during
Pregnancy (Percent) 60.7 44.3 ** 8.6
35
TABLE IV.3 (continued)
Prenatal W1C Income-Eligible Higher-Income
Participants Nonparticipants Nonparticipants
Father's Characteristics
Mean Age (Years) 27.9 29.0 ** 32.0 **
Age (Percent Distribution)
Younger than IS 0.4 0.2** 0.0 **
18- 19 4.0 3.8 0.4
20 - 24 29.9 22.6 6.5
25 - 29 33.1 31.3 27.4
30 - 34 18.6 23.3 36.3
35 and older 14.1 18.8 29.4
Education (Percent Distribution)
8 years or less 9.5 8.4** 0.9**
9-11 years 24.0 16.7 4.6
High sch(K)l graduate 50.7 46.1 34.2
Some college 12.3 18.7 22.9
College graduate 3.5 10.2 37.4
Race (Percent Distribution)
White 60.9 72.2 ** 89.1 **
Black 33.6 23.0 6.1
Asian or Pacific Islander 2.0 2.9 4.2
Native American 3.5 2.0 0.7
Hispanic (Percent)
Employed at Any Time during 12
Months prior to Delivery (Percent)
21.9
84.3
16.1
88.9 • »
6.7
98.4
**
Household Characteristics
Mean Household Size
Nonmetropolitan County (Percent)
Currently Receiving AFDC (Percent)
Currently Receiving Food Stamps
(Percent)
Mean Annual Pretax Income (Dollars)
4.0 4.0 3.0 **
32.6 23.7 ** 17.8 **
37.0 21.9 ** 0.9**
45.0 24.2 ** 1.0**
12,564 13,266 40,029 **
36
TABLE IV.3 (continued)
Prenatal WIC Income-Eligible Higher-Income
Participants Nonparticipants Nonparticipants
Mean Annual Pretax Income per
Household Member (Dollars) 3,858 3,851 15.015
Pretax Income as a Percentage of the
Poverty Level (Percent Distribution)
100 or less 56.2 45.2 ** 0.0 "
101 - 150 17.2 25.6 0.0 *
151 - 185 8.5 21.0 0.0 a
186 - 250 8.7 2.8 15.8
More than 250 9.6 5.5 84.2
74.8 85.4 ** 99.7**
34.1 21.4** 0.08
38.0 21.8** 0.0 a
8.6 5.9** 0.3 **
8.6 7.5 0.8 **
4.8 5.8 2.1 **
1.5 3.0 ** 0.5**
Any Income or Assistance from the
Following Sources during 12 Months
prior to Delivery (Percent)
Wages, salaries, interest, or dividends
AFDC
Food stamps
Housing assistance or public housing
Social security or SSI
Unemployment insurance
Veteran's benefits
Child support/alimony from absent
parent 5J 68 2.0**
Sample Size (Unweighted) 3,868 2,302 3,783
SOURCE: 1988 National Maternal and Infant Health Survey.
NOTE: All means, percent distributions, and tests of statistical significance are based on weighted
data and are calculated using SUDAAN. Tests of statistical significance are based on t-statistics
for the difference in means of the continuous variables and chi-square statistics
for the difference in the percent distributions of the categorical variables.
aBy definition, higher-income nonparticipants had incomes exceeding 185 percent of poverty level and
did not receive AFDC or food stamps.
*(**): The difference between WIC participants and nonparticipants is statistically significant at the
.05 (.01) level.
37
Prenatal Care While fully 82 percent of higher-income nonparticipants received
(Table IV.4) adequate prenatal care, only slightly more than half of both W1C
participants and income-eligible nonparticipants received adequa'.c
prenatal care, based on the Kessner Index. The Ressner Index is one
of the most commonly used measures of the adequacy of prenatal care.
The index combines information on the month in which prenatal care
started, the number of prenatal care visits recorded, and pregnancy
gestation to define the adequacy of prenatal care (Kessner et al. 1973).
For a full-term pregnancy, adequate prenatal care is defined as nine or
more visits, with the first visit occurring during the first trimester of
pregnancy, and inadequate care is defined as four or fewer visits.
Intermediate care for a full-term pregnancy is defined as all levels of
prenatal care between adequate and inadequate care. For prcterm
births (births before 37 weeks gestation), the number of prenatal care
visits required for care to be classified as adequate is adjusted
downward, based on the shorter length of gestation. It should be kept
in mind, however, that the Kessner Index characterizes the utilization
of prenatal care, but provides no information on the quality of care
received. Furthermore, the index is based on recommended visits in a
normal pregnancy; those with high-risk pregnancies may need more
visits. And lastly, the Kessner Index tends to overstate the adequacy of
care for women with full-term or postterm births (Kotelchuck 1987).
WIC participants were less likely to receive inadequate prenatal care than
were income-eligible nonparticipants in several respects: participants were
less likely to receive no prenatal care (1.3 percent versus 4.9 percent for
nonparticipants), and participants who received prenatal care were less
likely than income-eligible nonparticipants to receive inadequate care (as
measured by the Kessner Index) and more likely to receive an
intermediate level of care. The month in which care began for
participants did not differ significantly from the month in which care
began for income-eligible nonparticipants.
WIC participants were more likely than both groups of nonparticipants to
have received their prenatal care from county or city health departments,
community health centers, or hospital clinics, and were less likely to have
used private doctors or HMOs. This pattern reflects the fact that the
former group of care locations are often local sites for the WIC Program.
More than 90 percent of women in all groups reported receiving advice
concerning nutrition during prenatal visits. WIC participants were more
likely to report having received advice concerning other behavior-such as
advice to breastfeed or warnings to avoid smoking and alcohol-during
prenatal visits than were income-eligible nonparticipants. This difference
38
TABLE IV.4
THE USE AND SOURCE OF PRENATAL CARE
(Percentages)
Prenatal WIC Income-Eligible Higher-Income
Participants Nonparticipants Nonparticipants
Women with No Prenatal Care 1.3 4.9** 0.6
Month of Pregnancy in Which
Prenatal Care Began
First 10.5 13.1 21.5 **
Second 31.2 31.7 50.1
Third 22.7 23.0 203
Fourth 13.0 12.7 4.2
Fifth 10.6 9.1 1.9
Sixth 5.7 4.2 0.7
Seventh 3.8 3.0 0.8
Eighth 2.1 2.2 0.4
Ninth 0.4 1.0 0.2
Kessner Index of the Adequacy of
Prenatal Care
Inadequate 6.8 9.7** 1.5**
Intermediate 37.3 30.7 13.6
Adequate 51.8 54.4 82.2
Missing*1 4.0 5.3 2.7
Primary Source of Prenatal Care
Private doctor's office 40.8 59.9 ** 78.8 **
County or city health dept. 18.4 7.0 0.7
Community health center 12.7 7.6 2.8
HMO 1.7 3.8 7.7
Clinic at work or school 0.5 0.4 0.1
Clinic in a hospital 21.7 16.7 6.2
Hospital emergency room 0.3 0.2 0.0
Other 3.9 4.4 3.7
Advice Received during Prenatal
Visits
Vitamin/mineral supplements 96.3 94.9 98.2**
Proper nutrition 91.9 90.8 94.0*
Breastfeed baby 57.0 47.2 ** 52.3 **
Avoid alcohol 64.6 58.7 •* 72.0 **
Avoid smoking 71.7 63.1 ** 68.7*
Avoid illegal drugs 69.3 59.8 ** 64.3 **
39
TABLE IV.4 (continued)
Prenatal WIC Income-Eligible Higher-Income
Participants Nonparticipants Nonparticipants
Sources of Payment for Prenatal
Care
Own or husband's income 18.6 34.5 •• 43.1 **
Parents, other relatives, or
boyfriend 3.0 4.5 1.5**
Private insurance 21.4 42.5 ** 84.6**
Mcdicaid 50.8 26.2 ** 0.0"
Other government assistance 14.0 9.2 ** 0.7 **
Women Reporting Difficulty in
Obtaining Prenatal Care 22.3 19.9 6.7**
Problems with money or
insurance 11.5 10.8 2.4**
Problems with appointments.
work, or transportation 10.1 7.3 •• 2.4**
Problems with health care
providers 7.0 4.3 ** 2.5**
Other problems 3.1 4.5 1.0**
Women Hospitalized During 18.2 13.6 •• 12.0 **
Pregnancy
Sample Size (Unweighted) 3,868 2302 3,783
SOURCE: 1988 National Maternal and Infant Health Survey.
NOTE: All means, percent distributions, and tests of statistical significance are based on weighted
data and are calculated using SUDAAN. Tests of statistical significance arc based on t-statistics
for the difference in means of the continuous variables and chi-squarc statistics
for the difference in the percent distributions of the categorical variables.
"By definition, higher-income nonparticipants did not receive Medicaid payments for prenatal care.
'The Kessner Index could not be computed for observations with a missing value for gcstational age.
*(**): The difference between WIC participants and nonparticipants is statistically significant at the
.05 (.01) level.
40
is likely to reflect efforts of the WIC Program to make health information
availahle to participants. On some issues-breastfeeding, and avoiding
smoking and drugs--WIC participants reported receiving advice more often
than did higher-income nonparticipants. However, WIC participants
reported that they received advice to avoid alcohol less often than higher-income
nonparticipants. As noted above, mothers' reports concerning
health advice they received may be biased in that they may be more likely
to recall advice they later followed; thus, differences in reporting of advice
between WIC participants and nonparticipants may not reflect actual
differences in advice received.
WIC participants were much more dependent than income-eligible
nonparticipants on Medicaid or other government assistance for payment
for prenatal care: 51 percent of WIC participants used Medicaid as
compared with 26 percent of nonparticipants. Income-eligible
nonparticipants were more likely than participants to pay for care with
their own funds (35 percent compared to 19 percent, respectively) or
through private insurance (43 percent versus 21 percent).''
There were few differences between WIC participants and income-eligible
nonparticipants in the barriers to obtaining prenatal care, although both
groups were more likely than higher-income nonparticipants to face
barriers. WIC participants were somewhat more likely to report
difficulties with logistics and with their health care providers.
Finally, 18 percent of WIC participants were hospitalized during
pregnancy as compared with 14 percent of income-eligible nonparticipants
(and 12 percent of higher-income nonparticipants). It is likely that this
difference is related to the fact that the WIC Program is targeted at
women with high-risk pregnancies among the income-eligible group.
5All nonparticipants covered by Medicaid were classified as income-eligible
nonparticipants. Thus, by definition, higher-income nonparticipants paid
for care entirely through private sources.
41
Behavioral Risk
Factors--Alcohol
Consumption,
Cigarette
Smoking, and
Drug Use (Table
IV.5)
The reported prevalence of maternal behaviors that create risk for fetal
development-alcohol consumption, cigarette smoking, and drug use-varies
somewhat among the subgroups of women. (Because these
tabulations are based on mothers' reports, they may understate actual
prevalence.) In general. WIC participants did not differ significantly
from income-eligible nonparticipants in these behaviors except that
they were slightly less likely to report alcohol use before pregnancy.
However, both low-income groups were less likely to report alcohol use
(especially in moderation) and were more likely to report smoking and
drug use. both before and during pregnancy, than were higher-income
women. While 31 percent of WIC participants. .18 percent of income-eligible
nonparticipants. and 56 percent of higher-income
nonparticipants reported using alcohol in the year before the birth,
almost all the mothers reported reduced consumption during
pregnancy. Fully 86 percent of WIC participants. 82 percent of
income-eligible nonparticipants. and 75 percent of higher-income
mothers reported that they did not consume alcohol after they knew
they were pregnant, and more than 95 percent of all three groups
reported less than 1 drink per week. Furthermore, the proportions
reporting high levels of alcohol consumption dropped considerably in
all three groups after the onset of pregnancy.
Before pregnancy. 37 percent of WIC participants smoked cigarettes, but
only 29 percent smoked during pregnancy (implying about one-fifth of
smokers had quit). Those who did not quit reduced the number of
cigarettes they smoked on average by about a quarter, from 16.4 cigarettes
per day to 12.5. Income-eligible nonparticipants were essentially
indistinguishable from WIC participants in their tendency to smoke both
before and during pregnancy. Higher-income nonparticipants were less
likely than WIC participants to smoke before pregnancy (only 24 percent
smoked) and were more likely to quit during pregnancy (only 16 percent
smoked after learning they were pregnant).
Small numbers of WIC participants reported use of illegal drugs: 7.4
percent reported using marijuana, and 2.1 percent reported using cocaine
during the 3 months before learning of the pregnancy; 3.4 percent
reported using marijuana, and 1 percent reported using cocaine while
pregnant. Reported drug use among income-eligible nonparticipants did
not differ significantly from reported drug use among participants, while
higher-income nonparticipants were much less likely to report drug use.
42
TABLE IV.5
ALCOHOL CONSUMPTION. CIGARETTE SMOKING. AND
DRUG USE BEFORE AND DURING PREGNANCY
Prenatal WIC Income-Eligible Higher-Income
Participants Nonparticipants Nonparticipants
1.4 1.2'* 1.1
3.3 3.4 3.8
3.8 5.0 7.8
5.6 6.4 11.4
16.4 21.6 31.7
69.5 62.4 44.1
Alcohol Consumption
Women Who Drank Any Alcoholic Beverage
during the 12 Months before Delivery (Percent) 31.0 37.9 ** 56.0
Reported Frequency of Alcohol Consumption
during the 3 Months before the Woman Found
Out She Was Pregnant (Percent)
14 or more drinks per week
6-13 drinks per week
3 - 5 drinks per week
1 - 2 drinks per week
Less than 1 drink per week
Did not drink
Reported Frequency of Alcohol Consumption
after the Woman Found Out She Was Pregnant
(Percent)
14 or more drinks per week
6-13 drinks per week
3 - 5 drinks per week
1 - 2 drinks per week
Less than 1 drink per week
Did not drink
Drinkers Who Reported Reducing Alcohol
Consumption during Pregnancy (Percent) 89.5 87.7 93.0
Cigarette Smoking
Women Who Reported Smoking Cigarettes
(Percent)
During 3 months before the woman found out
she was pregnant 37.2 35.3 23.9
After she found out she was pregnant 29.3 28.8 15.9
Cigarettes Smoked Per Day (Mean Number)
During 3 months before the woman found out
she was pregnant 16.4 16.8 15.6
After she found out she was pregnant 12.5 12.4 11.5
0.3 0.4 0.1
0.5 0.6 0.2
0.8 1.3 0.7
2.1 1.9 3.1
10.3 13.9 21.4
86.0 82.0 74.5
43
TABLE IV.5 (continued)
Prenatal W1C Income-Eligible Higher-Income
Participants Nonparticipants Nonparticipants
Reported Frequency of Cigarette Smoking
during 3 Months before the Woman Found Out
She Was Pregnant
15 or more per day 20.7 20.8 14.0 **
6 - 14 per day 11.2 10.2 5.8
1 - 5 per day 5.3 4.3 4.1
Did not smoke 62.8 64.7 76.1
Reported Frequency of Cigarette Smoking after
the Woman Found Out She Was Pregnant
15 or more per day 10.1 10.3 5.5**
6 - 14 per day 11.5 10.4 5.4
1 - 5 per day 7.8 8.0 5.1
Did not smoke 70.7 71.2 84.1
Smokers Who Quit Smoking for at Least a
Week during Pregnancy (Percent) 48.3 50.8 62.6
Drug Use
Women Who Reported Smoking Marijuana or
Hashish
During the 3 months before the woman found
out she was pregnant
After she found out she was pregnant
Women Who Reported Using Cocaine or Crack
During the 3 months before the woman found
out she was pregnant
After she found out she was pregnant
7.4
3.4
2.1
1.0
6.1
2.3
2.7
1.2
4.2**
1.8**
1.1 **
0.1 **
Sample Size (Unweighted) 3,868 2^02 3,783
SOURCE: 1988 National Maternal and Infant Health Survey.
NOTE: All means, percent distributions, and tests of statistical significance are based on weighted data and
are calculated using SUDAAN. Tests of statistical significance are based on t-statistics for the
difference in means of the continuous variables and chi-squarc statistics for the difference in the
percent distributions of the categorical variables.
*(**): The difference between WIC participants and nonparticipants is statistically significant at the .05 (.01)
level.
44
Previous
Pregnancies and
Previous WIC
Participation
(Table IV.6)
Data on previous pregnancies were collected both in the maternal
survey and from the birth certificate. The maternal survey provides the
most detailed information on the history of each previous pregnancy,
including previous WIC participation. However, there arc many more
cases for which data on previous pregnancies arc missing in the
maternal survey (nearly 30 percent) than on the birth certificate (15
percent).6 For this reason, we present birth certificate data whenever
they arc available, and maternal survey data on variables not available
from the birth certificate.
a. Number and Timing of Previous Pregnancies and Births
Fewer WIC participants than income-eligible nonparticipants had previous
pregnancies (59 percent and 66 percent, respectively) or live births (54
percent and 61 percent, respectively). Among low-income women with
previous pregnancies, WIC participants and income-eligible
nonparticipants had similar numbers of pregnancies, but there was a
shorter interval between the preceding pregnancy and the most recent
pregnancy for WIC participants. The median number of past pregnancies
was two for both groups, with the mean slightly more than two. Although
the difference between the distributions is not statistically significant. WIC
participants were more likely than income-eligible nonparticipants to have
had their most recent pregnancy in the past 2 years, and they were less
likely to have had the pregnancy 4 years ago or more. Furthermore,
differences in the percentages of women with very long intervals between
pregnancies lead to a statistically significant difference in the mean
interval between pregnancies of about 5 months.7
WIC participants were about as likely as higher-income nonparticipants
to have ever been pregnant or to have ever given birth, but WIC
participants with past pregnancies had been pregnant more often on
6In particular, 28 percent of the sample had missing data on the number
of previous pregnancies (combining two maternal survey questions), while
8 percent had missing data on the interval since the last live birth (from
the birth certificate), and 15 percent had missing data on the interval since
the last pregnancy (from the birth certificate). Cases with multiple births
(about 2 percent) were treated as missing on birth certificate variables,
since in these cases the interval since the last birth or pregnancy was
coded as zero months.
'Similar results hold for the interval between live births, for low-income
mothers with previous live births.
45
TABLE IV.6
PREVIOUS PREGNANCIES AND PREVIOUS WIC PARTICIPATION
Prenatal WIC Income-Eligible Higher-Income
Participants Nonparticipants Nonparticipants
Women with Previous Pregnancy
(Percent)"
Women with Previous Live Birth
(Percent)'
Among Women with Previous Pregnancy
Previous Pregnancies (Mean Number)*1
58.8
54.3
2.3
66.4
613
2.2
59.1
52.6
1.8 »*
Number of Previous Pregnancies
(Percent Distribution)h
1 38.7 41.8 51.6 **
2 27.0 29.1 28.6
3 16.0 15.6 12.1
4 or more 18.3 13.5 7.8
Mean Interval since Last Pregnancy
(Months)* 36.7 41.3 '* 40.4 **
Interval since Last Pregnancy
(Percent Distribution)8
11 months or less 5.5 4.6 3.7 **
12-23 months 36.7 31.8 27.7
24-35 months 20.9 21.5 25.6
36-47 months 13.5 12.7 15.6
48-71 months 12.2 14.4 14.9
72 months and more 11.3 15.0 12.6
Prior Prenatal WIC Participation
(Percent)" 66.8 22.5 ** 4.8**
46
TABLE IV.6 (continued)
Prenatal W1C Income-Eligible Higher-Income
Participants Nonparticipants Nonparticipants
3.0 2.1 0.5 ••
30.9 27.2 21.6
21.9 21.0 27.5
14.7 13.7 18.1
15.1 16.5 17.4
14.3 19.6 15.0
59.9 21.3 ** 4.2**
79.9 38.7 ** 7.9**
Among Women with Previous Live Birth
Mean Interval since Last Live Birth
(Months)8 41.6 46.7 ** 45.1 **
Interval since Last Live Birth
(Percent Distribution)8
11 months or less
12-23 months
24-35 months
36-47 months
48-71 months
72 months and more
Prior Postpartum WIC Participation
(Percent)b
Prior Infant WIC Participation (Percent)
Sample Size (Unweighted) 3,868 2302 3,783
SOURCF.: 1988 National Maternal and Infant Health Survey.
Noil: All means, percent distributions, and tests of statistical significance are based on weighted
data and are calculated using SUDAAiV. Tests of statistical significance are based on t-statistics
SVT the difference in means of the continuous variables and chi-square statistics for
the difference in the percent distributions of the categorical variables. Missing data have
been excluded from the tabulations.
"Birth certificate data. Data on previous pregnancies were missing for about 15 percent of the
sample.
'"Maternal survey data. Data on previous pregnancies were missing for about 30 percent of the
sample.
*(**): The difference between WIC participants and nonparticipants is statistically significant at the
.05 (.01) level.
47
average than higher-income nonparticipants. Most higher-income
nonparticipants had only one previous pregnancy. The intervals hetween
the last pregnancy (or birth) and the most recent pregnancy (or birth)
were shorter on average for WIC participants than for higher-income
nonparticipants.
b. Previous WIC Participation
Not surprisingly, current WIC participants were much more likely than
nonparticipants to have participated in WIC in the past. For prenatal
participants with previous pregnancies and births, two-thirds had been
prenatal WIC participants in a previous pregnancy, while 80 percent had
had infants who had received WIC. and 60 percent had participated
postpartum. These figures suggest substantial continuity in the population
served by the program over time. However, nearly one-quarter (23
percent) of income-eligible nonparticipants with previous pregnancies had
been prenatal WIC participants in the past. Furthermore. 39 percent of
those with previous births had enrolled their infants and 21 percent had
been enrolled as postpartum mothers. Small proportions (under 10
percent) of nonparticipants who are not currently income-eligible for WIC
had also participated in WIC in the past.8
"The data on past WIC participation are consistent with the data on
current participation in the first section of this chapter in that both
sources suggest that the WIC Program serves a higher proportion of
income-eligible infants than of income-eligible pregnant or postpartum
women.
48
Pregnancy
Outcomes (Table
IV.7)
WIC participants exhibited slightly more positive pregnancy outcomes
than did income-eligible nonparticipants. although the differences were
not always statistically significant. In particular, although WIC
participants' newborns did not differ significantly from income-eligible
nonparticipants* newborns in mean birthweight. WIC participants were
significantly less likely to have a low-birthweight (less than 2.5(H) grams)
or a very low-birthweight baby (less than 1.500 grams). WIC
participants' newborns also had a significantly longer mean gcstational
age although the mean difference was only 0.3 weeks. WIC
participants had lower neonatal and infant mortality rates than did
income-eligible nonparticipants. but these differences were not
statistically significant. It is important to note that because WIC
participants and income-eligible nonparticipants differed in respects
other than WIC participation, these observed differences in outcome
are not necessarily the effects of the WIC Program. The multivariatc
analysis in Chapter V controls for differences in observed
characteristics other than program participation that may affect
outcomes.
Higher-income nonparticipants had significantly more positive birth
outcomes in most respects than did WIC participants; mean birthweights
were higher by 144 grams, and the incidence of low birthweight and of
preterm births was roughly half that found for WIC participants. The
infant mortality rate during the 6 months after birth was significantly lower
for higher-income women, but the difference in neonatal mortality rates,
while of similar magnitude, was not statistically significant.
POSTPARTUM
AND INFANT
WIC
PARTICIPANTS
AND
NONPARTICI-PANTS
The preceding section focused on the women served by the WIC
Program during pregnancy. As discussed in the first section of this
chapter, the WIC Program also provided supplemental food to many of
these mothers and their infants after birth and served some mothers
and infants who were not served prenatally. This section describes the
characteristics and experiences of postpartum and infant WIC
participants. We first compare the demographic and sociocconomic
characteristics of postpartum WIC participants with the characteristics
of nonparticipants. Next, we examine demographic and sociocconomic
characteristics of infant WIC participants, nonparticipants, and their
parents. We then compare infant WIC participants and nonparticipants
in terms of health status and health care utilization in the 6 months
after birth. Last, we examine infant feeding practices in the 6 months
after birth.
49
TABLE IV.7
PREGNANCY OUTCOMES FOR PRENATAL WIC PARTICIPANTS.
INCOME-ELIGIBLE NONPARTICIPANTS. AND
HIGHER-INCOME NONPARTICIPANTS
Prenatal WIC Income-Eligible Higher-Income
Participants Nonparticipants Nonparticipants
Mean Birthwcight (Grams) 3.284 3.265 3.428 **
Birth weight (Percent Distribution)
< 1.5(10 grams 1.3 2.1 ** 0.9 **
1,500- 2.499 grams 6.9 8.2 4.1
2 2.500 grams 91.8 89.7 95.0
Mean Gcstational Age (Weeks) 39.2 38.9 ** 39.5 **
Gcstational Age at Delivery < 37
Weeks (Percent) 11.3 12.8 6.2**
Gcstational Age (Percent Distribution)
Less than 28 weeks 0.8 1.4* 0.4 **
28 - 30 weeks 1.0 1.0 0.5
31 - 33 weeks 2.0 2.9 1.0
34 - 36 weeks 7.5 7.6 4.2
37 - 39 weeks 39.9 41.4 42.0
40 weeks and more 48.8 45.8 51.8
Number of Infant Deaths within 28
Days of Birth per 1.000 Live Births
Number of Infant Deaths within 6
Months of Birth per 1,000 Live Births
Maternal Weight Gain during
Pregnancy (Pounds)
Prcprcgnancy Weight (Pounds)
5.5
7.7
33
136
6.6
9.4
32
134
2.9
4.0
33
135
SOURCE: 1988 National Maternal and Infant Health Survey.
NOTE: All means, percent distributions, and tests of statistical significance are based on weighted
data and are calculated using SUDAAN. Tests of statistical significance arc based on t-statistics
for the difference in means of the continuous variables and chi-square statistics for
the difference in the percent distributions of the categorical variables.
*(**): The difference between prenatal WIC participants and nonparticipants is statistically
significant at the .05 (.01) level.
50
It is again important to keep in mind that differences in the characteristics
of participants and nonparticipants often are preprogram differences, not
differences that reflect the effects of WIC Program participation (both
before and after birth). Some of the preexisting differences may reflect
the fact that the WIC Program has appropriately targeted infants and
mothers most likely to be at nutritional risk.
Demographic and
Socioeconomic
Characteristics of
Post part urn WIC
Participants and
Nonparticipants
(Tables IV.8 and
IV.9)
Most postpartum WIC participants (84 percent) were also prenatal
WIC participants (refer back to Table IV. 1). Thus, tabulations
comparing the characteristics of postpartum WIC participants and
nonparticipants are very similar to those comparing the characteristics
of prenatal WIC participants and nonparticipants.
Postpartum WIC participants were from more disadvantaged backgrounds
than were income-eligible nonparticipants, and from much more
disadvantaged backgrounds than were higher-income nonparticipants
(Table IV.8). Postpartum participants were younger, less educated, more
likely to be black, less likely to be married or to have been employed
before the birth, more likely to receive Medicaid. and less likely to have
private health insurance than were income-eligible nonparticipants, and,
to a greater degree, higher-income nonparticipants. Their babies' fathers
were similarly younger, less educated, and more likely to be black than
were the fathers of nonparticipants' babies. Postpartum WIC participants
were also more likely to live in rural areas and to depend on public
assistance than were income-eligible nonparticipants.
Average incomes were similar for postpartum WIC participants and
income-eligible nonparticipants. but WIC participants were both more
likely than income-eligible nonparticipants to have incomes below the
poverty level and above 185 percent of the poverty level.9
9The measure of income used is from the year before the birth, and is
thus somewhat inexact. Families with incomes above 185 percent of the
poverty level were classified as income-eligible only if they received
AFDC, food stamps, or Medicaid at the time of the interview.
51
TABLE IV.8
DEMOGRAPHIC AND SOCIOECONOMIC CHARACTERISTICS OF POSTPARTUM WIC
PARTICIPANTS, INCOME-ELIGIBLE NONPARTICIPANTS.
AND HIGHER-INCOME NONPARTICIPANTS
Postpartum WIC
Participants
Income-Eligible
Nonparticipants
Higher-Income
Nonparticipants
Mother's Characteristics
Mean Age (Years) 23.7 24.6 28.4
Age (Percent Distribution)
Younger than 18 10.2 6.4** 0.9 **
18- 19 13.3 12.4 2.4
20-24 39.3 35.0 17.9
25-29 22.7 27.4 38.5
30 - 34 10.7 14.0 28.7
35 and older 3.8 4.8 11.7
Education (Percent Distribution)
8 years or less 8.8 7.7** 0.7 **
9-11 years 28.3 22.4 3.6
High school graduate 44.2 45.5 33.8
Some college 16.0 18.6 30.9
College graduate 2.7 5.8 30.9
Race (Percent Distribution)
White 66.0 72.5 ** 88.9**
Black 29.3 23.3 5.6
Asian or Pacific Islander 2.1 2.7 4.9
Native American 2.6 1.5 0.6
Hispanic (Percent) 18.7 17.1 7.8**
Married (Percent) 51.6 62.0 ** 92.8**
Lived with the Baby's Father during
Most of the Pregnancy (Percent) 61.6 71.5 95.5 ••
Employed at Any Time during 12
Months prior to Delivery (Percent) 55.0 61.3 78.7
Covered by Mcdicaid for Prenatal
Care and/or Delivery (Percent) 51.1 33.3 ** 1.0
No Private Health Insurance during
Pregnancy (Percent) 56.2 47.7 •• 8.1 • *
52
TABLE IV.8 (continued)
Postpartum WIC Income-Eligible Higher-Income
Participants Nonparticipants Nonparticipants
Father's Characteristics
Mean Age (Years) 28.1 28.7* 32.1 **
Age (Percent Distribution)
Younger than 18 0.4 0.2** 0.0 **
18- 19 4.0 3.2 0.5
20-24 28.8 24.7 6.1
25-29 33.5 32.3 26.6
30 - 34 18.6 23.2 36.5
35 and older 14.9 16.5 30.3
Education (Percent Distribution)
8 years or less 9.6 8.3 ** 0.7**
9-11 years 22.8 17.8 4.7
High school graduate 50.2 49.3 33.0
Some college 13.3 16.5 22.9
College graduate 4.0 8.2 38.6
Race (Percent Distribution)
White 63.9 69.6 ** 89.0 **
Black 30.8 24.9 6.1
Asian or Pacific Islander 2.0 2.8 4.2
Native American 3.3 2.4 0.7
Hispanic (Percent) 21.2 16.8 ** 6.8**
Employed at Any Time during 12
Months prior to Delivery (Percent) 85.5 87.9 98.4
Household Characteristics
Mean Household Size 4.3 4.4 3.6 **
Nonmetropolitan County (Percent) 36.2 22.7 ** 17.3 **
Currently Receiving AFDC
(Percent) 36.2 24.3 ** 0.0 ■
Currently Receiving Food Stamps
(Percent) 43.1 28.7** o.o ■
53
TABLE IV.8 (continued)
Postpartum WIC Income-Eligible
Participants Nonparticipants
Higher-Income
Nonparticipants
Mean Annual Pretax Income
(Dollars)
Mean Annual Pretax Income per
Household Member (Dollars)
12,89.1
3.558
13.024
3.238
40.986
12.611
Pretax Income as a Percentage of
the Poverty Level (Percent)
100 or less 59.0 49.8 ** 0.0 a
101 - 150 15.7 27.6 0.0 "
151 - 185 8.3 16.6 0.0 *
186 - 250 8.8 1.7 23.1
More than 250 8.3 4.3 76.9
Any Income or Assistance from the
Following Sources during 12
Months prior to Delivery (Percent)
Wages, salaries, interest, or
dividends 76.3 84.1 ** 99.7 **
AFDC 31.8 21.0** 1.0**
Food stamps 36.0 22.3 ** 1.0**
Housing assistance or public
housing 8.2 4.6** 0.9**
Social security or SSI 7.9 6.0* 1.6**
Unemployment insurance 4.7 4.0 2.8**
Veteran's benefits 1.4 1.2 1.0
Child support/alimony from
absent parent 5.4 4.8 2.5 **
Sample Size (Unweighted) 3,003 2,660 3,451
SOURCE: 1988 National Maternal and Infant Health Survey.
NOTE: All means, percent distributions, and tests of statistical significance are based on weighted
data and arc calculated using SUDAAN. Tests of statistical significance are based on t-statistics
for the difference in means of the continuous variables and chi-square statistics
for the difference in the percent distributions of the categorical variables.
"By definition, higher-income nonparticipants have incomes exceeding 185 percent of the poverty
level and were not receiving AFDC or food stamps at the time of the survey.
*(**): The difference between WIC participants and nonparticipants is statistically significant at the
.05 (.01) level.
54
Nonparticipants who were not income-eligible had average incomes
roughly three times as high as postpartum WIC participants.
Among postpartum participants, those who were not prenatal participants
were generally less disadvantaged than those who were prenatal
participants (Table IV.9). In particular, women who only enrolled in WIC
postpartum were more educated, more likely to be married, and less often
black or Hispanic. These women were also more likely to be employed
or to have had an employed spouse during pregnancy, and were less likely
to participate in AFDC, food stamps, or Medicaid. Furthermore,
"postpartum only" participants reported much higher incomes in the year
before the birth; more than one-quarter reported incomes exceeding 185
percent of the poverty level. The reasons for these differences are not
clear. "Postpartum only" participants may have had incomes that fell after
the birth, they may have had complications after the birth but not have
been at nutritional risk during pregnancy, or they may not have been
aware of WIC during pregnancy.
Demographic and
Socioeconomic
Characteristics of
Infant WIC
Participants and
Nonparticipants
(Tables IV. 10 and
rv.ii)
a. Characteristics of the Infant at Birth
Infant WIC participants (who include both prenatal participants and
infants who entered WIC after