Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 94
CHAPTER 3
Trends in Low Birthweight
This chapter analyzes changes in the rate of low birthweight and in
the composition of the low birthweight population during the past 10 to
15 years. The proportion of low birthweight infants has declined, but
only modestly--from 7.6 percent in 1971 to 6.8 percent in 1981. The
decline has not been substantial, despite several favorable develop-
ments: the proportion of pregnant women starting prenatal care in the
first trimester increased; educational attainment among women rose
markedly; the proportions of pregnancies among teenagers and older
mothers decreased; and the prevalence of cigarette smoking among women
20 to 44 years of age declined. Not all changes, however, were
favorable. For example, the proportion of unmarried pregnant women
increased during the same period.
The relatively small decline in the low birthweight rate contri-
buted in only a minor way to the substantial decline in the overall
infant mortality rate. Far more important were reductions in the
mortality rates of low birthweight infants. Still, the low birth-
weight infant remains at very high risk of mortality.
This chapter attempts to understand the recent past in seeking
means to reduce future low birthweight rates. Three categories of
questions are examined.
1. Have larger reductions in the low birthweight rate begun to
appear in recent years? Has a decline in the low birthweight rate been
masked by a concomitant reduction in fetal mortality?
2. Is the composition of the low birthweight population changing?
In particular, has the decline in low birthweight occurred among very
low birthweight infants (1,500 grams or less), who are especially
vulnerable to serious morbidity and mortality? Or has the decline
occurred among moderately low birthweight infants (1,501-2,500 grams)?
Are the changes concentrated in term deliveries (37 or more weeks
gestation) or in preterm infants?
3. Has the differential in low birthweight rates between high-risk
and low-risk women widened or narrowed? How large is the differential
presently? What might be the effect on low birthweight rates of
reducing or eliminating certain sociodemographic r isk factors?
What might be the impact of improvements in patterns of prenatal care?
94
OCR for page 95
95
To answer these questions, the committee analyzed data for the
United States as a whole and for five states--California,
Massachusetts, Michigan, North Carolina, and Oregon. Although analysis
of data from all reporting states gives the broadest picture of trends
in low birthwe~ght, examination of the detailed experiences of selected
states permits comparison of trends across different populations and
jurisdictions. Accordingly, the states were chosen to reflect a wide
variation in geography and demographic characteristics. The states
also were selected because the vital records data collected in these
states are believed to be of high quality, thus permitting additional
consideration of U.S. data as a whole, which necessarily include data
of widely varying quality.
The analysis of national trends and r isk factors was based on
published data and unpublished analyses performed by the National
Center for Health Statistics. For California, Michigan, and Oregon,
the analyses were derived from cross-tabulatzons performed by organi-
zations within the respective states. For Massachusetts and North
Carolina, the cross-tabulations were performed by the committee from
public use tapes.
Birth records for the United States as a whole were analyzed for
the period 1971-1981 and for individual, selected states by single
year, as available, from 1968 to 1982. The salient results of the
analyses are reported in Tables 3.1 through 3.13 in this chapter. More
detailed tabulations are provided in Tables B.1 through B.13 in
Appendix B.
The committee recognizes that entries on birth certificates, the
original source of both the national and state data presented in this
chapter, vary in their accuracy and completeness. Inferences based on
reported gestat~onal age, in particular, need to be qualified.
Further, while birth records encode the month when prenatal care began
and the tote' number of prenatal visits, the content and quality of
care cannot be determined from such data alone. Nevertheless, the
committee believes that important insights about trends In low
birthweight rates and changes in the composition of the low birthweight
population can be gained from information available on the birth record.
Overall Trends in Low B~rthweight Rates
Table 3.1 shows changes in low birthweight rates in the United
States and the five selected states from 1971 to 1981. Data by
individual year for the United States are given in Appendix Table B.1
and for selected states in Appendix Table B.2. For some states, data
are provided for the years 1968 to 1970 also.
Overall, from 1971 to 1981, the low birthweight rate for the united
States declined by 11 percent. Although low birthweight rates declined
by comparable magnitudes in the individual states studied, the decline
was more marked for Massachusetts and less significant for North
Carolina and Michigan. There is no clear indication, either in Table
3.1 or in the more detailed Tables B.1 and B.2, that low birthweight
rates were declining more rapidly in the earlier or later portions of
OCR for page 96
96
TABLE 3.1 Low-Weight Births (2,500 Grams or Less) per 1,000 Live
Births in the United States and Five States, 1971, 1976, and 1981
~ 971 1976 1981
Percentage
Decrease,
1971-1981a
United States 76 72 68 11
California 66 62 58 12
Massachusetts 71 66 59 17
Michigan 77 75 69 10
North Carolina 87 83 79 9
Oregon 57 54 50 12
NOTE: All data include both singleton and multiple births. For 1971
and 1976, low birthweight includes live births weighing 2,500 grams or
less. Since 1979, the National Center for Health Statistics has
defined low birthweight as less than 2 ,500 grams. Data presented in
this table are der ived from ache sources decor ibed in reference no. 1.
aComputed from rates calculated to three significant f igures .
the 1968-1982 period. (From the data in Table B.2 alone, it is not
possible to determine if the apparent stagnation of low l~irthweight
rates in Michigan and North Carolina from 1980 to 1982 is of longer-
term s ignif icance. J
Table 3.2 and Tables B.3 and B.4 show trends in live births and
late fetal deaths . I f changes in medical pract' ce or bir th cer tif i-
cation have moored increasing numbers of pregnancies from the fetal
death category to the live birth category, then the computed trend in
low birthweight rates among live births may understate the actual rate
of decline in low-weight rates among all pregnancies. Because there is
likely to be substantial undercounting of early fetal deaths, and
because shifts in classification from fetal deaths to live births are
much more 1 i kely in the late fetal death groups, only fetal deaths of
28 or more weeks gestation are included. The data in Table 3.2 suggest
that the assessment of trends in low-weight births among live births
alone may understate slightly the rate of decline in low birthweight
rates. AS Appendix Table B.4 shows, this conclusion is not altered
when all fetal deaths of 20 or more weeks gestation are included.
Hereafter, only data for live births are considered.
The Compos action of Low-Weight Bir ths
Table 3.3 depicts changes in very low birthweight (VT~BW) and
moderately low birthweight (MLBW) rates in the united States and five
states. The data uniformly show that the decline in low birthweight
OCR for page 97
97
TABLE 3.2 Low-Weight Births and Late Fetal Deaths per 1,000 Live
Births in Four States, 1971, 1976, and 1981
1971 1976 1981
Percentage Decline,
1971-198 1
California 71 66 61 14
Michigan 81a 77 72 11
North Carolina 93 87 82 12
Oregon 60 57 51 15
NOTE: For California, late fetal deaths are defined as fetal deaths
weighing 500 grams or more. For other states, late fetal deaths
include all fetal deaths of at least 28 weeks gestation. Data
presented In this table are derived from the sources described in
reference no. 1.
Data
aData are for 1972, the earliest year available.
completely confined to the 1,501-2,500-gram group. With the
of Massachusetts, there may have been a slight increase in
rates was
exception
the VLBW rate from 1971 to 1981.
Table 3.4 shows more detailed trends in low birthweight rates In
relation to gestational age for the selected states. Interpretation of
the results needs to be tempered by the uncertain quality of data on
duration of pregnancy. Most states did not perform extensive editing
to remove or revise implausible gestational age entries. Moreover,
birth records with missing data on gestational age were excluded from
the analysis. Because data on gestational age reported before 1975
were considered to be less reliable, only the experience from 1975 to
1981 is reported.
The results show some var iatic~n from state to state. For
Massachusetts and Oregon, there was a slight decline in the rate of
preterm births (those of less than 37 weeks durations , while there was
no discernable decline in the preterm rate In North Carolina or
Michigan. Concomitantly, Massachusetts and Oregon showed declines in
preterm low-weight births, while Michigan and North Carolina did not.
All states, however, showed a decline in the term low birthweight
group. The larger absolute declines In low birthweight rates in
Massachusetts and Oregon appear to be attributable to declines in
preterm low-weight births.
Because the decline in low birthweight appears to be confined to
the moderately low birthweight group (Table 3.3), the remaining two
rows of Table 3.4 analyze trends in preterm and term moderately
low-weight births. The results suggest that the decline in the
proportion of 1,501-2,500-gram births reflects mostly a decrease in
term infants.
These results are consistent with those reported nationally. For
the United States during the period 1970 to 198O, Ressel et al.2
OCR for page 98
98
TABLE 3.3 Very Low Birthweight (VT~BW) and Moderately Low Birthweight
(=BW) Rates per 1,000 Live Births in the United States and Five
States, 1971, 1976, and 1981
Weight
Categorya 197 1
1976 1981
IJnited States VLBW 11 12 12
MLBW 65 61 56
California VLBW 9 9 10
MLBW 57 52 49
Massachusetts VLBW 10 9 9
MLBW 61 57 SO
Mich igan VLBW 12 13 13
MLBW 64 62 57
Nor th Carol ina VLBW 14 14 15
MLBW 74 69 65
Oregon VLBW 8 8 9
MLBW 49 46 41
NATE: Data presented in this table are derived from the sources
descr ibed in reference no. 1.
every low birthweight defined as 1,500 grams or less; moderately low
birthweight as 1,501 to 2 ,500 grams.
reported that the preterm low birthweight rate declined from 38 per
1,000 to 36 per 1,000. The term low birthweight rate declined from 36
per 1,000 to 28 per 1,000.
Though declines in term ~ ow birthweight
were predominant, the state data presented here suggest that those
areas with more marked declines in low birthweight rates may be
experiencing some additional decline in preterm births. Analysis
_
based on a larger number of states is needed to assess this
relationship further.
Sociodemographic Characteristics
Table 3.5 depicts the relationship between low birthweight rates
and race in the United States and five states from 1971 to 1981. More
detailed data are given in Tables B.1, B.5, and B.6 of Appendix B.
For the United States as a whole, the relative decline in white low
birthweight rates exceeded the corresponding relative decline in black
low birthweight rates. Thus, white low birthweight rates declined by
OCR for page 99
99
TABLE 3.4 Preterm, Low-Weight Preterm, and Low-Weight Term Births
per 1,000 Live Births of Known Gestational Age in Four States, 1975
and 1981
Massachusetts Michigan
North
Carolina Oregon
1975 1981 1975 1981 1975 1981 1975 1981
. 71 67 95 94 92 93 66 61
Preterm LBW 33 30
Term LBW 34 29
42 42 42 42 29 25
33 28 38 32 27 22
Preterm MLBW 24 21 31 29 21 19
Term MLBW 33 28 38 31 25 22
NOTE: Data presented in this table are derived from the sources
described in reference no. 1.
14 percent from 1971 to 1981, while black rates declined by only 6
percent. The absolute declines among whites and blacks, however, were
more comparable. Although quantitative assessment of changes In the
white-black differential depends on the measure used, there is no clear
closing of the gap.
The numbers of black births in Massachusetts and Oregon were rela-
t~vely small. Hence, the reported low birthweight rates In those
states are subject to greater variability. For the remaining states
studied, however, the trends accord with that observed nationally. It
is noteworthy that both white and black low birthweight rates vary
across states. Michigan's relatively high overall low birthweight
rate, for example, appears attributable to the very high low birth-
weight rate among blacks In that state. North Carolina's relatively
high overa1 1 low birthweight rate, by contrast, reflects elevated rates
for both races. The trend in low birthweight among Spanish surnamed
whites in California is comparable to that for other whites in the
state.
For the United States as a whole from 1971 to 1981, Appendix Table
B.7 depicts trends In low-weight births according to age of the mother
for whites and blacks separately. To display more clearly the relative
positions of the various age groups, Table 3.6 computes the relative
risks of low birthweight. The relative risk is the ratio of the low
birthweight rate for a particular category to the low birthweight rate
for all births of the same race in the same year.* Thus, the low
birthwe~ght rate among all white births in 1971 was 65.5 per 1,000 (see
Appendix Table B.71. Among white mothers under 15 years of age, the
rate was 127.8 per 1,000, that is, 1.95 times the overall white rate in
1971 (see Table 3.6~. For selected states, more detailed, absolute
rates are given in Appendix Table B.8. AS Table 3.6 shows, the highest
*This definition of relative risk is specific to this chapter and is not consistent
with the usual use of the term or with its definition elsewhere in the report.
OCR for page 100
100
TABLE 3 . 5 Low-weight Births per 1, 000 Live Births in Relation to Race
in the United States and Five States, 1971, 1976, and 1981
White
1971 1976 1981
Black
1971 1976 1981
United States 66 61 57 133 129 125
Californian 51 46 44 111 108 101
Massachusetts 65 63 55 113 125 111
Michigan 64 61 57 144 140 137
North Carolina 70 64 61 135 130 124
Oregon 56 52 47 118 127 103
NOTE: Data presented in this table are derived from the sources
descr ibed in reference no. 1.
aMultiple live births excluded. Data for white births exclude
those with Spanish surnames. For Spanish surnamed births, rates per
1,000 were 53 in 1971, 47 in 1976, and 45 in 1981.
relative risks are found among births to teenage mothers, particularly
among whites. However, a much greater proportion of black births are
to teens. The gap in low birthweight rates between the high- and
lower isk maternal age groups shifted only slightly from 1971 to 1981.
No particular age group displayed a marked departure from the general
pattern of small reductions in the low birthweight rate dur sing the
decade. Analysis of the individual state data produced a s impalas
conclusion.
Table 3.7 shows trends in the relative risk of low birthweight
according to maternal educational attainment for whites and blacks
separately from 1971 to 1981. For both races, the relative risk of low
birthweight declines sharply among mothers with at least 12 years of
education. Moreover, the increased r Ask among mothers with less than
12 years education is more pronounced for whites. The relationship
between education and low birthweight rates prevails among individual
maternal age groups (Appendix Table B.91.
There is no indication of a narrowing gap In relative r isk among
births to mothers with disparate educational attainment. The trend, in
fact, is toward a widening gap. Because the number of states reporting
educational attainment varied during the 1971-1981 period, such a
f inding requires caution. Generally consistent findings, however, were
obtained from analysis of individual selected states (Appendix Table
B.10~. Given the overall improvement in levels of educational
attainment of mothers from 1971 to 1981, the finding of an increasing
gap suggests that those who have remained poorly educated constitute an
increasingly high-risk group.
OCR for page 101
101
TABLE 3.6 Relative Risk of Low Birthweight by Age of Mother and Race: united States,
1971, 1976, and 1981
Relative Riska
Percent Distribution
White Black Total Live Births (1981)
Maternal Age
(years) 1971 1976 1981 1971 1976 1981 White Black
Less than 15 1.95 1.91 1.84 1.43 1.31 1.33 O.1 0.9
15-19 1.26 1.32 1.36 1.13 1.17 1.12 12.7 24.4
15-17 -- -- 1.54 -- -- 1.17 4.2 10.5
18-19 -- -- 1.27 -- -- 1.09 8.6 13.9
20-24 0.93 0.98 1.02 0.96 0.97 1.00 33.3 35.4
25-29 0.91 0.86 0.88 0.88 0.87 0.91 32.5 23.7
30-34 0.98 0.94 0.88 0.89 0.89 0.90 16.7 11.5
3S-39 1.20 1.14 1.10 1.02 1.01 0.99 4.0 3.4
40 or older 1.56 1.37 1.31 0.97 1.01 1.07 0.6 0.7
NOTE: Data presented in this table are derived from the sources described in reference
no. 1.
aThe base for calculating relative risks is the low b~rthweight rate for the total in
each column.
Table 3.8 reports trends in the relative risk of low birthweight by
marital status, maternal age, and race in the United States from 1976
to 1981. Unmarried mothers have a consistently higher risk of bearing
a low birthweight infant than those who are married. The increased
risk is not explained by age differences among married and unmarried
women. The relative risk differential is more pronounced for whites
than blacks; but the much larger proportion of unmarried mothers among
blacks makes unwary fed status a far more important risk factor for
blacks than for whites.
Of even greater concern is the fact that births to unmarried women
increased throughout the 1970s. Whereas 7 percent of white mothers and
51 percent of black mothers were reported unmarried in 1976, the
proper tions were 12 percent for whites and 56 percent for blacks in
1981.
It is difficult to discern a clear trend in relative risks from
Table 3.8. Such a trend conceivably could be masked by changes in the
number of reporting states and the relatively short period examined.
For two selected states, Table 3.9 presents trends in low
birthweight rates in relation to a simple two-way classification of
soc~odemographic risk factors. ~High-risk. mothers are defined as
those who reported less than 12 years education, were unmarried, or
were young {less than 20 years of age for Massachusetts, less than 18
years of age for North Carolina), or any combination of these risk
factors. "Low-risk n mothers include all others. The data show that
OCR for page 102
1
102
TABLE 3.7 Relative Risk of Low Birthweight by Educational Attainment of Mother and
Race: 1971, 1976, and 1981
Relative Riska
Maternal . Percent Distribution
Educational White Black Total Live Births (1981)
Attainment
(years) 1971 1976 1981 1971 1976 1981
white Black
0-8 1.35 1.39 1.46 1.08 1.14 1.17
9-11 1.32 1.35 1.47 1.10 1.12 1.18
12 0.92 0.94 0.96 0.92 0.92 O.95
13-15 0.80 0.82 0.81 0.81 0.85 0.83
16 or more 0.74 0.73 0.74 0.74 0.72 0.71
3.7 4.7
15.8 30.1
44.0 41.5
18.9 15.5
16.2 6.5
NOTE. Includes 49 states and Washington, D.C., in 1981; 44 states and Washington, D.C.,
in 1976; 38 states and Washington, D.C., in 1971. Data presented in this table are
derived from the sources described in reference no. 1.
aThe base for calculating relative risks is the low birthweight rate for the total in
each column.
declines in low birthweight rates have been concentrated in ache
lower isk groups . Such evidence supper ts the conclus ion that the gap in
low birthweight rates between low-risk and high-risk groups, at least
as defined by maternal sociodemographic characteristics, is not closing.
Pregnancy History
Birth order is associated with smaller differentials in relative
r isk than is age, and here again the relative r isk var ies ~ ess among
blacks than whites (Table 3.10 and Appendix Table B.ll). The changes
in relative r isk have been small. Still, among whites there is a
suggestion that first births are at increased relative r isk for low
bir thweight.
The effect of the interaction between maternal age and birth order
on low birthweight has been well documented in the past and these
relationships persist to the present. Figure 3.1, based on 1981 data,
indicates the high r isk among women aged 15-19 bear ing their second or
later child; the sharp increase in risk among women having their first
child as age advances past 25-29 years; and the lowered r isk at third
or higher birth orders when women are 25-34 years of age. Very similar
patterns are found among wh, te and black women.
Termination of last prior pregnancy with a fetal death elevates
risk for low birthweight in the next pregnancy, but the increase is not
dramatic (Table 3.10~. The relationship is stronger for whites than
OCR for page 103
103
TABLE 3.8 Relative Risk of Low Birthweight by Marital Status, Age of Mother, and Race:
1976 and 1981
Relative Riska
Percent Distr ibution
White Black Total Live Births (1981)
1976 1981 1976 1981 White Black
Married 0.97 0.92 0.87 0.81 88.4 44.1
Unmarried 1.60 1.57 1.14 1.15 11.6 56.0
Marr fed
Less than 15 2.07 1.70 1.55 1.50 b ~
15-19 1.25 1.26 1.06 0.97 8.3 3.4
15-17 - - 1.45 -- 1.07 2.2 0.7
18-19 - - 1.17 -- 0 .94 6 .1 2.7
20-34 0.90 0.88 0.83 0.79 75.7 37.8
35 or older 1.36 1.06 1.02 0.87 4.4 2.9
Unmar r fed
Less than 15 1.89 1.88 1.32 1.33 0.1 0.9
15-19 1.59 1.58 1.16 1.14 4.4 21.0
15-17 -- 1.61 -- 1.18 2.0 9.9
18-19 -- 1.54 -- 1.12 2.4 11.2
20-34 1.66 1.56 1.12 1.14 6.7 32.8
35 or older 1.97 1.82 1.13 1.32 0.3 1.2
NOTE: Includes total united states in 1981, 38 states and Washington, D.C., In 1976.
Data presented in this table are derived from the sources described in reference no. 1.
aThe base for calculating relative r isks is the low birthweight rate for the total
In each column.
bless than 0 .1 percent.
blacks. Differentials in relative risk between births preceded by a
live birth and those preceded by a fetal death have decreased. The
data related to the outcome of the last pregnancy need to be
interpreted cautiously, however, because of likely problems in
reporting accuracy and consistency, and variations in geographic
coverage. Still, the narrowing of the gap observed is one of the few
evidences of positive change in relative risk during the 1971-1981
period.
Special tabulations on 1981 live births, performed by the National
Center for Health Statistics for this report, showed the relationship
of birthweight to interval between termination of ~h`? ~ A=F nr-"n~n^v
and the first day of the last menstrual per con before the current
pregnancy. Analysis was confined to single births in which the prior
pregnancy ended in a live-born infant. Table 3.11 shows a sharply
. . .
~ = _ _ ~ ~ · _,!
· · .
eleValea EelaLlVe ~1SK arena heath rant then Who i"~',=1 i- 1~- -~_
6 months. (Bowever, only 5-10 percent of births fail into this inter-
pregnancy interval.) Risks decrease moderately thereafter to reach an
optimum level at 2 to 3 years after the previous pregnancy in the white
group and 2 to 4 years in the black group. The relative risk increases
for births with longer interpregnancy intervals, especially in whites,
reflecting, perhaps in part, impaired fecundity. Although the data are
not shown here, the pattern of exceptionally high relative risk for a
~ _ = ~ ~ ~~ _~. _~ ~ 44 ~^ or ~ ^— ~ =~— ~1
OCR for page 104
104
TABLE 3 . 9 Trends in Low Birthweight Rates in Relation to Two-Way
Classification of Maternal Risk, Massachusetts and North Carolina,
1971, 1976 , and 1 981
.
1971 1976 1981
Massachusetts Higher isk 93 8 9 8 7
Lower isk 61 57 49
Nor th Carol ina Higher isk 10 9 10 6 10 6
Lower isk 62 58 55
NOTE: ~High-risk. births defined as either maternal education less
than 12 years, mother unmarried, or young maternal age (less than 20
for Massachusetts, less than 18 for North Carolinas, or any
combination of the above. Lower isk. includes all others. Data
presented in this table are der iced from the sources descr ibed in
reference no. 1.
TABLE 3.10 Relative Risk of Low Birthweight by Birth Order, Outcome of Prior
Pregnancy, and Race: United States, 1971, 1976, and 1981
Relative Riska
Percent Distribution
White Black Total Live Births tl981)
1971 1976 1981 197' 1976 1981 White Black
Birth Order
First birth 1.01 1.06 1.08 1.02 0.85 1.00 43.7 39.2
Second birth 0.93 0.89 0.90 1.01 0.98 0.96 32.5 29.3
Third birth 0.98 0.97 0.94 1.00 0.97~ 0.99 14.7 16.6
Fourth birth 1.09 1.10 1.08 0.99 1.02 1.06 5.2 7.6
Fifth and over 1.17 1.14 1.11 1.05 0.98 1.12 3.5 9.9
Outcome of Prior
_ I_ . .
PregnancyD
Live birth O.95 0.94 0.92 0.95 0.94 0.95 75.6 72.6
Fetal death 1.31 1.32 1.21 1.32 1.28 1.11 17.8 17.0
Unknown 1.37 1.38 1.30 1.31 1.22 1.17 6.5 10.S
NOTE: Data presented in this table are derived from the sources described in reference
no. 1.
aThe base for calculating relative risks is the low birthweight rate for the total in
each column.
bData are for second order births or higher. States reporting the information
varied: 49 and Washington, D.C., in 1981; 43 and Washington, D.C., in 1976; 37 and
Washington, D.C., in 1971.
OCR for page 105
OCR for page 107
OCR for page 108
OCR for page 109
OCR for page 110
OCR for page 111
OCR for page 112
Representative terms from entire chapter:
birthweight rates
105
20
18
16
_ 14
- o
o—
'_ 12
At
c' 1 0
OR
us 8
6
4
2
o
20
18
t6
14
o
12
U]
C) 10
8
6
4
2
o
~[
—3rd and 4th
\
—2nd \
\ \ ;
1st ~ ~
Huh and Over
~ — ~
ALL RACES
-
1 1 1 1
1 5-1 9 20-24 25-29 30-34 35+
WE ITE
3rd and 4th
\ 5th and Over
—2nd \
—1st\ \
N~ \
i'
-
-
- 1 1 1 1
J
15-19 20-24 25-29 30-34 35f
20 3rd and 4th
18 In>\ " Sth and Over
\ \ "
~ \ "` ~
~ 1st \ \
_ ~ ~
16
14
12
U]
106
TABLE 3.11 Relative Risk of Low B~rthweight by Interval Between Last
Pregnancy and Current Pregnancy (Single Live Births), 1981
Relative Risks Percentage of All Live Births
Interval
(months) White Black White Black
Less than 6 1.63 1.46 5.4 8.9
6-11 1.04 1.06 12.1 13.2
12-23 0.81 0.92 24.2 18.9
24-35 0.78 0.83 15.1 11.7
36-47 0.86 0.83 8.9 8.4
48 or more 1.07 0.90 17.6 23.0
Unknown 1.24 1.13 16.7 15.9
NOTE: Includes only births in which the last prior pregnancy terminated
in a live birth. Pregnancy measured from first day of last menstrual
period. Reporting areas include 49 states and Washington, D.C. Data
presented in this table are derived from the sources described in ~
reference no. 1.
aThe base for calculating relative risks is the low birthweight rate
for the total in each column.
short interval between pregnancies and of declining relative risk with
increases in the interpregnancy interval is found at every age and
birth order. An independent effect of interpregnancy interval on
birthweight remains apparent even after controlling for the potentially
confounding variables of age, birth order, race, marital status, and
educational attainment.
The relationship between birthwe~ght and interpregnancy interval is
much different for pregnancies where the previous pregnancy terminated
in ~ fetal death. A short interval between pregnancies is not
associated with increased risk of low birthweight. In fact, the
reverse is suggested.
Prenatal Care
The relationship between the risk of low birthweight and the timing
and quantity of prenatal care is reviewed in detail in Chapter 6.
Table 3.12 and Appendix Table B.12 provide information on this issue
drawn from vital statistics. Among whites, there is an increase in
relative risk as the time of the first prenatal visit advances from the
first and second month of pregnancy to the second trimester. AS other
investigators have noted, 3 the risk of low birthweight appears
reduced for mothers who begin care in their third trimester because
such pregnancies already have reached more advanced gestational ages.
Despite this bias, the relative risk remains high among the whites who
107
TABLE 3.12 Low Birthweight Relative Risk by Month Mother Began Prenatal Care and Race:
1971, 1976, and 1981
Relative Riska
- Percent Distr ibution
Start of White Black Total Live Births (1981)
Prenatal Care
(month) 1971 1976 1981 1971 1976 1981 White Black
First-second 0.92 0.90 0.89 0.93 0.91 0.92 55.0 38.6
Third 0.94 0.95 0.95 0.93 0.97 0.94 24.4 23.9
Fourth-sixth 1.12 1.15 1.16 0.96 1.00 1.01 16.3 28.5
Seventh-ninth 1.08 1.10 1.12 0.86 0.93 0.94 3.2 6.2
No prenatal care 2.75 2.69 2.88 2.13 2.22 2.21 1.1 2.8
Seventh month or
later or no care 1.42 1.48 1.56 1.24 1.31 1.35 4.3 9.1
Four th month or
later or no care 1.17 1.20 1.23 1.04 1.06 1.09 20.6 37.6
NOTE: Includes total U.S. in 1981; 44 states and Washington, D.C., in 1976; 39 states and
Washington, O.C., in 1971. Data presented in this table are derived from the sources
descr ibed in reference no. 1.
aThe base for calculating relative r isks is the low bir thweight rate for ache total in
each column.
began care after the sixth month. For the no-care group, by contrast,
many of the pregnancies may have terminated early, before such women
would ordinarily have started prenatal care.
Analysis of trends in low birthweight rates in relation to prenatal
care is complicated by variations over time in the number of states
reporting data on such care. From reporting states (see Appendix Table
B.12) , it appears that low birthweight rates for all races combined
declined in both early- and late-care groups, though the low birthweigh.
rate for mothers with early care improved relatively more than the
other s .
For the individually selected states, analysis of the relationship
between low birthweight rates and the timing of the first prenatal
visit gave s imilar r esults (Appendix Table B.13~. Except in North
Carolina, the great bulk of the decline in low birthweight rates has
occurred among the early-care group.
It is noteworthy that substantial increases occurred during the
1970s in the proportion of mothers reporting a first visit during the
first trimester. Thus, for Massachusetts, mothers with early care
comprised 79 percent of live births in 1970 and 87 percent of live
births in 1980. For Michigan, the corresponding proportions for 1970
and 1980 were 69 percent and 78 percent, respectively. For North
Carolina, the proportions were 67 percent and 78 percent, respectively;
and for Oregon, the proportions were 71 percent in 1971 and 77 percent
in 1980. For reporting states in the United States as a whole, the
proportions with first trimester care were 79 percent of whites and 63
percent of blacks by 1980. The following year there was no decrease
and in 1982, the figures were 79 and 61 percent, respectively,
representing a departure from the steady increases during the decade
1970 to 1980.
108
Analysis of 1981 Single Live Births in the United States*
This section explores in more detail the relationship between low
b~rthweight and certain maternal characteristics. The objective is to
assess what fraction of current low-weight births might be eliminated
by improvements in prenatal care and specific risk factors. This
analysis ignores the finding of previous sections that past reductions
in low birthweight were in some cases more pronounced among low-risk
groups.
The upper part of Table 3.13 shows, for whites and blacks sepa-
rately, the estimated effect on the 1981 low birthwe~ght rate of
eliminating all late prenatal care and the associated excess risk, as
well as the estimated effect of eliminating late prenatal care and
improving selected maternal risk factors (less than 12 years education,
unmarried status, and high age of mather/b~rth order (ABO) risk, as
defined in Table 3.13~. The lower part of the table shows the same
analysis, except that elimination of "nonadequate" care, rather than
late care, is considered. "Late prenatal care" is defined as care
begun after the first trimester of pregnancy or no care. "Nonadequate
prenatal care. means either late start of prenatal care or a first
trimester start with fewer prenatal visits than prescribed for the
reported duration of pregnancy (an index adapted from the one developed
by Ressner, et al .4 ~ .
The estimated effects in Table 3.13 correspond to what epidem~olo-
gists term "attributable risk. reduction; that is, they reflect both
the relative risks of low birthweight among the maternal risk
categories and the proportions of single live births in 1981 among the
risk categories.
As the upper section of the table shows, late prenatal care by
itself is a relatively minor factor in explaining the level of the low
birthweight rate in 1981 for both racial groups; social and pregnancy
history factors are far more important. The quantitative contribution
of prenatal care increases appreciably when the measure of care takes
into account the frequency of prenatal visits as well as the timing of
the first visit (lower part of table). Eliminating excess risks among
births with "nonadequate" care reduces the low birthweight rate by 15
percent among whites and 12 percent among blacks, in contrast to a 3
percent reduction for both races when prenatal care is gauged only by
whether care started in the first trimester. The joint effect of
eliminating nonadequate prenatal care and the other characteristics is
substantial.
It is recognized that the preceding analysis controlled for only a
limited number of risk factors. From the vital record data alone, the
content of prenatal care could not be examined. Still, the results in
Table 3.13 suggest that the pattern of prenatal care, rather than early
initiation of care alone, may play an important role in determining
birthweight.
*Based on special tabulations provided by Dr. Joel Rle~nman, National
Center for Health Statistics.
109
TABLE 3.13 Estimated Effects on Low Birthweight of Improvements in
Prenatal Care and in Selected Maternal Risk Factors (Single Live
Births), 1981
white
Black
Rate Percent Rate Percent
(per 1,000) Reductiona (per 1,000) Reductions
Original base rate 47 -- 112
Estimated rate after
elimination of excess
risk among births with:b
Late prenatal care 46 3 107
__
3
Late prenatal care and
less than high school
education, unmarried
status, or high
ABO risks 38 20 84 24
Estimated rate after
elimination of excess
risk among births with:b
Nonadequate prenatal
care 40 15 97 12
Nonadequate prenatal
care and less than high
school education,
unmarried status, or
high ABO risks 34 29 77 30
NOTE: United States, excluding births in California, New Mexico, Texas, and
Washington, which did not report mother's education or number of prenatal
visits. Data presented in this table are derived from the sources described in
reference no. 1.
aPercent reduction determined from rates calculated to additional
significant digit.
stimated rates are derived by adjusting data on live births
cross-tabulated by the specified variables to eliminate associated excess
risks.
CHigh ABO risk includes births to women under 18 years of age, second or
higher order births to 18- and 19-year-old women, first order births to
women 30 years of age or older, and all other births to those age 40 and
older.
110
Conclusions
The committee investigated trends in the rate of low birthweight
and the composition of low-weight births in the United States as a
whole and in five selected states (California, Massachusetts, Michigan,
North Carolina, and Oregon) during the past 10 to 15 years. For the
United States, the proportion of low-weight births declined from 7.6
percent of live births in 1971 to 6.8 percent of live births in 1981.
Although Massachusetts showed a more marked decline in low birthweight
rates than the other states, the relative declines in all of the states
and the United states were of the same order of magnitude (Table 3. 1~ .
There is no clear indication that low birthweight rates were
declining more rapidly during the earlier or later part of the
1968-1982 period. Moreover, inclusion of fetal deaths in the analysis
of low birthweight trends results in only a slight to moderate increase
in the rate of low birthweight decline (Table 3.2~.
The decline in low birthweight rates was confined to the moderately
low birthweight group (1,501-2,500 grams). No decline, or perhaps a
slight increase, was observed in the very low birthweight group (1,500
grams or less) (Table 3.3~. Although birth certificate data on gesta-
tional age are incomplete and of uncertain quality, the observed
decline in low birthweight was apparently concentrated mostly in the
full-term low birthweight group. However, in Massachusetts and Oregon,
where overall declines in low birthweight rates were larger, a decline
in preterm low birthweight rates also was observed. In each state,
among moderately low-weight births, most of the decline was observed
among term infants (Table 3.41.
Both white and black low birthweight rates have declined, but
blacks remain at increased risk of low birthweight compared to whites.
The data show no closing of the gap, although quantifying the
white-black differential in low birthweight rates depends on the index
used to measure the gap. Both white and black low birthweight rates
vary appreciably across the states studied individually.
Teenage mothers and those aged 35 years or over have higher risks
of low birthweight than mothers in their twenties or early thirties.
Teenage mothers have the highest relative risk of low birthweight,
especially among whites (although this elevated risk probably derives
more from other characteristics of teenage mothers than from young age
itself, as discussed in Chapter 2~. Childbearing among teenagers is
more prevalent among blacks, however. In both races, no per titular age
aroun showed a marked departure from the general pattern of small
_ ~ . _ _ _ ~ ~ , _
reductions in low birthweight from 1971 to 1981 (Table 3.61.
For both races, the risk of low birthweight declines sharply among
mothers with at least 12 years of education. The relationship between
education and low birthweight is independent of maternal age and race.
The gap in low birthweight rates among mothers with disparate
educational attainment is not closing, and may be widening. Because
educational attainment of mothers has increased during the past 10 to
15 years, the finding of a widening gap in low birthweight among
mothers with disparate education suggests that the poorly educated
constitute an increasingly high-risk group (Table 3.7~.
may
111
Unmarried mothers have consistently higher risks of low birth-
weight. The elevated risk is not attributable to differences in age or
race. No clear change in the low birthweight differential between
married and unmarried births has been observed. However, the propor-
tion of unmarried mothers appears to be increasing. Among blacks, 56
percent of mothers were reportedly unmarried in 1981 (Table 3.8~.
A two-way measure of maternal sociodemographic risk was devised to
reflect maternal educational attainment, marital status, and age. An
analysis of trends in Massachusetts and North Carolina showed that the
gap in low birthweight rates between high- and low-risk groups has been
widening (Table 3.9~.
There has been no clear change in the relationship between par ity
and the risk of low-weight birth. High-parity births continue to have
slightly higher risks of low birthweight (Figure 3.1~. Moreover,
termination of the last pregnancy with a fetal death increases the risk
of low birthweight in the subsequent pregnancy. However, the
differential in low birthweight between those with a prior live birth
and those with a prior fetal death appears to have narrowed (Table
3.10~. Among mothers with a previous live birth, an interpregnancy
interval of less than 6 months enhances the subsequent risk of low
birthweight (Table 3 .11 ) .
The r isk of low birthweight is reduced among mothers who initiate
prenatal care dur ing the first 3 months of pregnancy. The proportion
of mothers with early prenatal care increased during the past 10 to 15
years. There was also a decline in low birthweight rates among mothers
In the early care group (Table 3 .12 ~ .
The committee performed a multivariate tabulation of single live
births in the United States during 1981 according to educational at-
tainment, marital status, age/b~rth order category, and the timing and
quantity of prenatal care. With other factors controlled, a change in
only timing of prenatal care starts was associated with a minor
reduction in risk of low birthweight. The estimated contribution of
prenatal care to a reduction in the low birthweight rate was more
marked, however, when care was gauged by an index of "adequacy" that
reflected both the start of care in the first trimester and number of
visits. The tabulation suggested that the analyzed risk factors might
account together for as much as 30 percent of the risk of low birth-
weight in both races. Moreover, the sensitivity or the estimated
contribution of prenatal care to the method of measuring such care
suggests that the pattern of care may be an important influence on the
risk of low birthweight (Table 3.131. Chapter 6 explores this theme
more fully.
In reaching these conclusions, the committee identified several
issues for future analyses of low birthweight using vital statistics.
First, as noted also in Chapter 2, pregnancy outcome measures need to
be defined more precisely--e.g., very low birthweight, moderately low
birthweight and preterm and full-term low birthweight. And differen-
t~al rates among white and black mothers, among those with higher and
lower educational attainment, and among other high-risk and low-risk
groups deserve careful scrutiny.
me,
112
Future sources of information on low birthweight and other preg-
nancy outcomes also need to be more timely. Available data collection
and reporting procedures, as well as terminology, need to be more
uniform across states.
vital record data alone may not gauge adequately the content and
effectiveness of prenatal care; nor do such data character ize important
aspects of maternal behavior (such as cigarette smoking) and medical
history {such as maternal diabetes). Accordingly, high priority ought
to be given to more detailed studies of selected cohorts of pregnant
women. These should include new prospective designs and reanalysis of
existing data. Further study is required on methods to measure more
precisely the content and timing of prenatal care.
Greater attention should be given to analyses of the experiences of
individual states and other jurisdictions. Geographic var. iations in
pregnancy outcome or in the relationship between low birthweight and
other r isk factors could reflect differences in local public health
programs and economic activity, as well as sociodemographic variations.
In making these observations, the committee recognizes that it is
giving only limited attention to important problems In the nature,
quality, and timeliness of the vital statistics data collected and
analyzed in the United States. Limits of time and resources prevented
the formulation of detailed suggestions in this area, aside from those
above, and the committee's charge did not require an ~n-depth review of
data collection issues. Nonetheless, Appendix D describes some
important national vital statistics data sets that can help in studying
the problem of low birthweight. The committee is also aware that
several groups are already concerned with vital statistics issues (such
as the National Committee on Vital and Health Statistics, the National
Association of Vital Registrars, and the ad hoc Standard Terminology
Group), and urges such groups to address the data problems highlighted
in this chapter, Chapter 6, and elsewhere in the report.
References and Notes
1. All data for the United States are from published and unpublished
data provided to the committee by the National Center for Health
Statistics. For California, Michigan, and Oregon, the analyses
were derived from cross-tabulations performed by organizations in
the respective states. For Massachusetts and North Carolina, the
cross-tabulations were performed by the committee from public use
tapes.
Ressel SS, Villas J. Berendes HW, and Nugent RP: The changing
pattern of low birthwe~ght in the United States: 1970-1980. JAMA
251:1978-1982, 1984.
3. Harris JE: Prenatal medical care and infant mortality. In
Economic Aspects of Health, edited by VR Fuchs, pp. 15-52.-
Chicago: University of Chicago Press, 1982.
4. Institute of Medicine: Infant Death: An Analysis by Maternal RiSk
and Health Care. Contrasts in Health Status, edited by DM Ressner,
Vol. . 1. Washington, D.C .: National Academy of Sciences , 1973 .
