Most vital statistics indicators of the health of Americans were stable or showed modest improvements between 1997 and 1998. The preliminary birth rate in 1998 was 14.6 births per 1000 population, up slightly from the record low reported for 1997 (14.5). The fertility rate, births per 1000 women aged 15 to 44 years, increased 1% to 65.6 in 1998, compared with 65.0 in 1997. The 1998 increases, although modest, were the first since 1990, halting the steady decline in the number of births and birth and fertility rates in the 1990s.
Fertility rates for total white, non-Hispanic white, and Native American women each increased from 1% to 2% in 1998. The fertility rate for black women declined 19% from 1990 to 1996, but has changed little since 1996. The rate for Hispanic women, which dropped 2%, was lower than in any year for which national data have been available. Birth rates for women 30 years or older continued to increase. The proportion of births to unmarried women remained about the same at one third.
The birth rate for teen mothers declined again for the seventh consecutive year, and the use of timely prenatal care (82.8%) improved for the ninth consecutive year, especially for black (73.3%) and Hispanic (74.3%) mothers.
The number and rate of multiple births continued their dramatic rise; the number of triplet and higher-order multiple births jumped 16% between 1996 and 1997, accounting, in part, for the slight increase in the percentage of low birth weight (LBW) births. LBW continued to increase from 1997 to 1998 to 7.6%.
The infant mortality rate (IMR) was unchanged from 1997 to 1998 (7.2 per 1000 live births). The ratio of the IMR among black infants to that for white infants (2.4) remained the same in 1998 as in 1997. Racial differences in infant mortality remain a major public health concern. In 1997, 65% of all infant deaths occurred to the 7.5% of infants born LBW. Among all of the states, Maine, Massachusetts, and New Hampshire had the lowest IMRs. State-by-state differences in IMR reflect racial composition, the percentage LBW, and birth weight-specific neonatal mortality rate for each state. The United States continues to rank poorly in international comparisons of infant mortality.
Expectation of life at birth increased slightly to 76.7 years for all gender and race groups combined. Death rates in the United States continue to decline, including a drop in mortality from human immunodeficiency virus. The age-adjusted death rate for suicide declined 6% in 1998; homicide declined 14%. Death rates for children from all major causes declined again in 1998. A large proportion of childhood deaths, however, continue to occur as a result of preventable injuries.
- birth weight-specific mortality
- infant mortality
- low birth weight
- multiple births
- vital statistics
Note to the Reader: 1999 is the 50th anniversary of the publication of this annual article in Pediatrics. A series of monthly columns and special articles began in the very first volume of Pediatrics, in 1948, and included a feature called “Public Health, Nursing, and Medical Social Work” authored first by Dr Leona Baumgartner, Chief of the Bureau of Child Hygiene of the New York City Health Department, and later by Myron Wegman.1 In the fourth volume of Pediatrics in 1949, Wegman devoted his first column to the discussion of the nation's vital statistics.2The regular publication of the “Annual Summary” article in the December issue of Pediatrics began in 1957.3
The column has changed and evolved over the years to take eventually the form that most of us associate with Dr Wegman's long stewardship. It continues to evolve. This year we include a special feature—the publication of state-by-state comparisons of infant mortality rates, low birth weight rates, and birth weight-specific mortality rates. These data will allow readers to make more informed inferences about state-by-state differences in infant mortality.
Most vital statistics indicators of the health of Americans were stable or showed modest improvements between 1997 and 1998. The birth rate for teen mothers declined again for the seventh consecutive year, and the use of timely prenatal care improved, especially for Black mothers. Life expectancy at birth increased slightly to 76.7 years. Death rates in the United States continue to decline, including drops in mortality from human immunodeficiency virus (HIV), suicide, homicide, and unintentional injury. However, the infant mortality rate (IMR) was unchanged from 1997 to 1998. The proportion of births to unmarried mothers remained about the same at one third. The number and rate of multiple births continue their dramatic rise, accounting, in part, for the increase in low birth weight (LBW) births.
The data presented in this report were obtained from vital statistics records—birth certificates, fetal death reports, and death certificates—for residents of the United States. Data for 1997 and earlier years are final and include all records. Data for 1998 are preliminary, and are based on 99% of births and at least 85% of deaths reported to the National Center for Health Statistics. More complete descriptions of vital statistics data systems are available elsewhere.4–6 The preliminary estimates may differ from the final data for 1998 that will include all records, but most differences are usually small.
Current vital statistics patterns and recent trends through 1998 are presented in this report by state of residence, age, race and Hispanic origin, as well as other birth and infant death characteristics. More detailed data are available in the final files for 1997 than in the preliminary files for 1998, so some of the detailed analyses of birth and death patterns focus on the 1997 data. Hispanic origin and race are collected as separate items in vital records. Persons of Hispanic origin may be of any race, although most births and infant deaths of Hispanic origin (97%) are to white women. Because there are often important differences in childbearing patterns between non-Hispanic white and Hispanic women, all tables that present data by race have been reformatted to include data separately for non-Hispanic white and Hispanic women.
The mother's marital status for birth data, underlying cause of death for deaths, and birth weight for infant deaths have special considerations for which readers should be aware. Mother's marital status was reported directly on the birth certificates or through the electronic birth registration process in all but 3 states in 1997 and all but 2 states in 1998. Details about the reporting of marital status in those 3 states and methods of edits and imputations applied to other items on the birth certificate are presented in publications of the National Center for Health Statistics.4,,6
Cause of death statistics in this report are based solely on the underlying cause of death. The underlying cause of death is defined as “(a) the disease or injury that initiated the train of events leading directly to death, or (b) the circumstances of the accident or violence that produced the fatal injury.” From 1979 to the present, cause of death data in the United States have been classified according to the International Classification of Diseases, 9th Revision.7
Infant mortality statistics by birth weight in this report were obtained from the 1997 period linked birth-infant death dataset.8 In this dataset, the death certificate is linked with the corresponding birth certificate for each infant who died in 1997 in the United States. The purpose of this linkage is to use additional variables available from the birth certificate, such as birth weight, to better interpret infant mortality patterns. Numbers of infant deaths were weighted to compensate for the 2.1% of infant deaths in 1997 for which the matching birth certificate could not be identified.8 The weighting procedure results in the same overall IMR as that based on unlinked death or mortality data; however, small differences may exist because of geographic coverage differences, additional quality control, and weighting.8
Population denominators for the calculation of birth, death, and fertility rates are estimates of the US population as of July 1 of each year, produced by the US Bureau of the Census.9,,10However, IMRs were computed by dividing the total number of infant deaths in each calendar year by the total number of live births in the same year.5,,8 Fetal and perinatal mortality rates were computed by dividing the number of fetal or perinatal deaths by the number of live births plus fetal deaths.
According to preliminary data, 1 605 976 persons were added to the population in 1998 as a result of natural increase, the excess of births over deaths (Table 1).6 The rate of natural increase has been the same for each year 1996–1998, at 5.9 persons per 1000 population.
The preliminary number of births in the United States increased in 1998 to 3 944 046, up 2% compared with the final total for 1997 (Table 1). The preliminary birth rate in 1998 was 14.6 births per 1000 population, up slightly from the record low reported for 1997 (14.5). The fertility rate, defined as the number of births per 1000 women aged 15 to 44 years, increased 1% to 65.6 in 1998, compared with 65.0 in 1997. The 1998 increases, although modest, were the first since 1990, halting the steady decline in the number of births and birth and fertility rates in the 1990s.
The numbers of births as well as birth and fertility rates increased in the majority of states according to 1998 preliminary data. The number of births increased in 43 states and declined in 7 states and the District of Columbia. Numbers rose 5% to 7% in Colorado, Nevada, and Utah. Birth rates per 1000 total population increased in 34 states, declined in 7 states, the District of Columbia, Puerto Rico, the Virgin Islands, and Guam, and were unchanged in 9 states. Fertility rates per 1000 women aged 15 to 44 years increased in 43 states, declined in 6 states, the District of Columbia, Puerto Rico, and the Virgin Islands, and were unchanged in New York and Guam (Table 2).
Births to Native American and Asian or Pacific Islander women and women of Hispanic origin tend to be highly concentrated geographically, as indicated in Table 2. Nearly half of births to Native American women were to residents of Alaska, Arizona, California, New Mexico, and Oklahoma. Residents of California, Hawaii, and New York accounted for half of births to Asian or Pacific Islander women. Three-fifths of births to women of Hispanic origin were to residents of California, New York, and Texas.
Racial and Ethnic Composition
Fertility rates for total white (64.7 per 1000), non-Hispanic white (57.8), and Native American women (70.5) each increased from 1% to 2% in 1998. The fertility rate for black women (71.0 in 1998) declined 19% from 1990 to 1996, but has changed little since 1996. The rate for Asian or Pacific Islander women (63.5) declined 4% between 1997 and 1998, reaching the lowest level since this rate was first calculated in 1980.4 Similarly, the rate for Hispanic women (101.1), which dropped 2%, was lower than in any year for which national data have been available (1989).4 Among populations of Hispanic origin for which fertility rates can be reliably computed, Mexican-American women continue to have the highest fertility, with a rate of 116.6 per 1000 in 1997 (Table 3), and the highest age-specific birth rates among women under age 30.
Trends in Age-specific Birth Rates
Since 1991, when it reached a 20-year high (62.1 per 1000 aged 15–19), the birth rate for teenagers has fallen 18%, to 51.1 in 1998 (Table 4). The 1998 rate, which is 2% lower than in 1997, is close to the 1986 record low of 50.2.11 The slight increase in the number of births to teenagers in 1998 reflects entirely the 3% increase from 1997 to 1998 in the number of teenage females in the population.10
Birth rates for teenagers in all age groups declined from 1997 to 1998. The rate for the youngest group, aged 10 to 14 years, was 1.0 per 1000, the lowest level since 1969; the number of births in this age group in 1998 (9481) was 27% below its recent high in 1994. The birth rate for teenagers 15 to 17 years declined 5% to 30.4 in 1998, a record low. The 21% decline in this rate since 1991 essentially reverses the 27% rise in the rate between 1986 and 1991. The rate for older teenagers 18 to 19 was 82.0 in 1998, 2% lower than in 1997, and 13% lower than its 1992 high of 94.5.
Birth rates for teenagers differ considerably by race and Hispanic origin but all rates have declined in the 1990s (Table 4).4,,11,12 Between 1997 and 1998, rates per 1000 teenagers aged 15 to 19 fell 2% for all whites (45.4 in 1998) and non-Hispanic whites (35.2), by 3% for black women (85.3) and Asian or Pacific Islander teenagers (23.1), and by 4% for Hispanic teenagers (93.7). The rate for Native American teenagers was unchanged at 71.8. Since 1991, black teenagers have experienced the largest decline of any group. Their rate fell 26% from 115.5 in 1991 to 85.3 in 1998, lower than in any year since 1960 when data for black women first became available.11 The birth rate for Hispanic teenagers has declined steadily since 1994, falling 13% to a lower level than in any year since national rates became available.
In examining the recent declines in teenage childbearing, it is useful to review trends separately for first births and for repeat births to teenagers. Rates of first births for teenagers declined in the 1990s but more slowly than overall rates, with nearly all of the total 10% reduction since 1991 occurring from 1995 (49.2) to 1997 (44.7). In contrast to this modest decline, repeat childbearing—that is, the rate of second births to teenagers who have already had 1 child—fell substantially in the 1990s, dropping 21% from 220.9 per 1000 in 1991 to 173.7 in 1997.4,,11 Although repeat births account for only 22% of all teen births, they are of particular concern; a teenager with 2 or more children is at greater risk for a host of difficulties.13
Childbearing for Women 20 Years of Age and Older
Birth rates for women in their twenties, the principal childbearing ages, increased 1% to 2% in 1998, to 111.2 per 1000 for ages 20 to 24 and 116.0 for ages 25 to 29 years. These rates generally declined from 1990 through 1995, but the year-to-year changes have been small.
Birth rates for women in their thirties continued to increase in 1998, rising to the highest levels in at least 30 years, to 87.5 per 1000 aged 30 to 34 and 37.4 per 1000 aged 35 to 39 years. Birth rates have also increased for women in their forties. The rate for women aged 40 to 44 years in 1998, 7.3 per 1000, has nearly doubled from its low point in 1981 (3.8). The steady upward trend in the rates for women in their thirties reflects in large part the ongoing tendency for many of these women to make up for previously postponed childbearing.14 For example, the rate of firstbirths per 1000 childless women aged 30 to 34 years in 1997, 92.1, is the highest ever recorded, 70% greater than the rate reported in 1975 (53.4) (tabular data not shown).
The total fertility rate (TFR) provides an estimate of the number of births that a hypothetical group of 1000 women would have if they experienced, throughout their childbearing years, the age-specific birth rates observed in a given year. TFR increased 1% to 2059.5 in 1998. This increase reflects the rise in birth rates for women aged 20 and older. TFR rose 1% for black women (2171.5) in 1998 and 2% each for non-Hispanic white (1839.5) and Native American women (2086.0). The rate declined 2% for Hispanic women (2949.0), and 4% for women of Asian or Pacific Island descent (1852.5) (tabular data not shown for 1998).
The number, rate, and percent of births to unmarried women increased in 1998. The number of births increased 3% to 1 292 534. The preliminary birth rate for unmarried women rose to 44.3 births per 1000 unmarried women aged 15 to 44, about 1% higher than in 1997 (44.0), but still nearly 6% lower than its peak level, 46.9, in 1994. The number for 1998 is the highest ever reported in the United States; the increase is due mostly to continued increases in the number of unmarried women of childbearing age (up 6% since 1994).15
The proportion of all births that were to unmarried women increased from 32.4% in 1997 to 32.8% in 1998 (Table 5).6 This proportion has been fairly stable since 1994. Nonmarital birth proportions rose for non-Hispanic white (21.9%) and Hispanic women (41.6%), and declined slightly for black women (69.0%).
The proportion of teen births that were to unmarried women continued to increase in 1998, rising to 78.8%. Birth rates for unmarried teenagers (ie, births per 1000 unmarried teenagers), available through 1997, describe the risk that an unmarried teenager will give birth. This rate declined by 9% overall between 1994 and 1997.
Smoking During Pregnancy
The rate of reported smoking during pregnancy continued to decline in 1997, a trend underway since this information first became available from the birth certificate in 1989.16 In 1997, 13.2% of women reported smoking during pregnancy, down about one third since 1989 (19.5%) (Table 5). Tobacco use during pregnancy is associated with a variety of adverse outcomes, including LBW, intrauterine growth retardation, and infant mortality, as well as negative consequences for child health.17–19
Maternal smoking declined or was unchanged in most racial/ethnic origin groups; rates are highest for non-Hispanic white, Native American, and Hawaiian women and lowest for Hispanic (except Puerto Rican) and Asian or Pacific Islander (except Hawaiian) women.4 Smoking rates fell for women in most age groups, with the particular exception of pregnant teenagers, whose rates have now increased for 3 consecutive years, rising from 16.7% in 1994 to 17.6% in 1997, and are now higher than for any other age group. Although still relatively rare, smoking during pregnancy by black teenagers has risen from 5.0% to 6.6% since 1994.4,,16
Timely prenatal care improved among US women for the 9th consecutive year, rising to 82.8% for 1998 (Table 5). The proportion of women beginning care in the first trimester of pregnancy changed little during the 1980s, but has increased 10% since 1989 (75.5%). For 1997–1998, the percent of women with first trimester care was unchanged for non-Hispanic white women (87.9%), but rose for black (73.3%) and Hispanic women (74.3%). Since 1989, first trimester care has risen 22% and 25%, respectively, among the latter 2 groups.
Although the relationship between prenatal care and pregnancy outcome is difficult to evaluate, timely prenatal care may promote better birth outcomes by providing early risk assessment to prevent or treat medical conditions, and by offering health behavior advice such as smoking cessation and nutrition counseling.20
The proportion of women beginning care late in pregnancy (during the third trimester), or with no care at all, was unchanged for 1997–1998 at 3.9%. Since 1989, this measure has declined from 6.4% for all births; among black women late or no care has dropped from 11.9% to 7.0%, and among Hispanic women from 13.0% to 6.3%.
The trend toward fewer cesarean deliveries may have ended, at least in the short-term. The cesarean delivery rate rose from 20.8 to 21.2 per 100 live births for 1997–1998 (Table 5). The primary cesarean rate (first cesareans per 100 live births to women who had no previous cesarean) increased for the first time since 1989 when these data became available from birth certificates, rising 2%, from 14.6% to 14.9%. The primary cesarean rate had fallen steadily between 1989 and 1996 (from 16.1 to 14.6%) and was unchanged for 1996–1997. Primary cesarean rates rose in 1998 among the 3 major racial/ethnic groups: non-Hispanic white (15.1%), black (16.0%), and Hispanic women (13.6%).
The rate of vaginal birth after previous cesarean delivery (VBAC; number of vaginal births per 100 births to women with a previous cesarean) was down 4%, dropping from 27.4 in 1997 to 26.3 in 1998. The VBAC rate rose 50% between 1989 and 1996, but has fallen 7% in the past 2 years. The recent decline in the VBAC rate may reflect renewed controversy over the safety of VBAC compared with elective repeat cesarean.21,,22
Overall cesarean section rates increase with maternal age; for 1997, mothers 40 years of age and older were more than twice as likely to have a cesarean delivery as teenaged mothers (32.4% and 14.3%, respectively).
The number and rate of multiple births continues to climb. The number of births in twin deliveries rose 3% between 1996 and 1997, and has risen 52% since 1980 (from 68 339 to 104 137).23The number of triplet and other higher-order multiple births jumped 16% between 1996 and 1997, soaring 404% since 1980 (from 1337 to 6737 births).23 Twins, triplets, and other higher-order multiples now account for nearly 3% of all births.
The twin birth rate (the number of twin births per 1000 live births) was up 3% to 26.8 per 1000 for 1997 (Table 5). On the heels of a 20% rise from 1995 to 1996, the higher-order multiple birth rate (the number of triplets, quadruplets, and other higher-order multiple births per 100 000 live births) rose 14% between 1996 and 1997 to 173.6 per 100 000. The higher-order multiple birth rate has doubled since 1991 (81.4) and quadrupled since 1980 (37.0).
The growth in twin and triplet birth rates in recent years among women 30 years of age and older is extraordinary. Between 1980–1982 and 1995–1997 (3 years of data are combined to generate stable rates), the twin birth rate rose 63% for women 40 to 44 years, and soared nearly 1000% for women 45 to 49 years. The higher-order multiple birth rate rose 400% for women in their thirties, and exploded by more than 1000% for women in their forties (Fig 1).23
The increase in multiple births, especially higher-order multiples, has been associated with two related trends—older age at childbearing and increased use of ovulation-inducing drugs and assisted reproductive technologies, such as in vitro fertilization.24
Multiple births, regardless of how conceived, tend to be high-risk births. About half of all twins and the great majority of triplets are born preterm or LBW. This higher risk, coupled with the escalating multiple birth rate, has had a large influence on overall national and state measures of infant health.25,,26
The rate of LBW (<2500 g) was 7.6% in 1998, up from 7.5% in 1997. The percent LBW has been increasing fairly steadily from the low of 6.7% reported in 1984. The rate of very low birth weight (VLBW; births weighing <1500 g) also rose slightly to 1.45% for 1998, from 1.42% in 1997. VLBW has risen moderately during the 1980s and 1990s (from 1.15% in 1980).4 When compared with heavier infants (2500 g or more), the risk of infant death in 1997 was 5 times higher for infants weighing 1500 to 2499 g, and 92 times higher for infants born at under 1500 g.8
Between 1997 and 1998, LBW increased among non-Hispanic white mothers (from 6.5% to 6.6%), but was essentially the same for black (13.0%) and Hispanic mothers (6.4%). LBW among black mothers has declined from a high of 13.6% reported for 1990, but remains higher than levels reported during the early and mid 1980s (12.6% to 12.8%).
The rise in the multiple birth rate has markedly influenced overall rates of LBW. Whereas LBW among all births rose 12% between 1984 and 1997, when only singleton births are examined, the rise in LBW is a comparatively modest, 5%. Multiple births also have an important impact on age-specific LBW rates, especially among older mothers. For example, among all births to women 45 to 54 years of age the LBW rate was 17.4%; when multiple births are excluded, however, the level drops to 9.2%, almost half the rate.
LBW risk varies broadly among Hispanic and Asian and Pacific Islander subgroups. For 1997, levels of LBW among mothers of Puerto Rican origin (9.4%) were 57% higher than those of their Mexican counterparts (6.0%). Among Asian and Pacific Islanders, LBW risk ranged from 5.1% for Chinese mothers (the lowest level reported among any of the racial/ethnic groups) to 8.3% for Filipino mothers.
In 1998, an estimated 28 486 infant deaths were reported in the United States (Table 1). The preliminary IMR was 7.2 per 1000 live births, the same as the final 1997 rate of 7.2, and matches the lowest ever recorded in the United States5,,6 The neonatal mortality rate (NMR; infant deaths under 28 days) was 4.8 in 1997 and in 1998, while the postneonatal mortality rate (PNMR; 28 days through 11 months) declined from 2.5 to 2.4. Between 1997 and 1998, IMRs and NMRs declined for infants of black mothers; however, the decrease was not statistically significant. The respective rates for infants of white mothers remained the same. IMRs were higher for infants whose mothers were teenagers or 40 years of age or older, did not complete high school, were unmarried, began prenatal care after the first trimester of pregnancy, or smoked during pregnancy. IMRs were also higher for male infants, multiple births, and infants born preterm or LBW.8
Infant mortality in the United States has declined by >40% since 1980 (Table 6 and Fig 2). The NMR declined more rapidly during the 1980s, whereas the PNMR has declined more rapidly during the 1990s. The decline in the perinatal mortality rate (number of fetal deaths at 28+ weeks of gestation plus number of infant deaths at <7 days of age per 1000 live births plus fetal deaths) has closely paralleled the decline in the NMR, while the fetal mortality rate (number of fetal deaths at 20+ weeks of gestation per 1000 live births plus fetal deaths) has declined more slowly.
Racial differences in the IMR remain a major national concern. Although all race groups have experienced declines in IMR, the relative difference in rates between black and white newborns, expressed as the ratio of black to white IMRs, increased from 2.0 in 1980 to 2.4 in 1990 (Table 6). The ratio remained unchanged at 2.4, from 1990–1998. The absolute difference in IMRs has actually decreased since 1990, from 10.4 to 8.1 deaths per 1000 live births in 1998. The Hispanic IMR was the same as the non-Hispanic white IMR in 1998. While the recent decline is very encouraging, racial disparities in IMR present continued challenges for researchers and health care providers alike.27,,28 Some states are beginning to address the problem.29
Birth Weight-specific Infant Mortality
Birth weight is one of the most important predictors of infant mortality. The IMR for a given population can be partitioned into 2 key components: the birth weight distribution and birth weight-specific mortality rates (the mortality rate for infants at a given weight). The IMR can decrease when either the percentage of LBW births decreases or birth weight-specific mortality rates decrease. The percentage of LBW births plateaued during the early 1980s, but has generally increased since then (Fig 2). Thus, all of the decline in the IMR since 1980 has been attributable to declines in birth weight-specific IMRs, and not to the prevention of LBW. These declines have been attributed primarily to improvements in obstetric and neonatal care. The United States has been unsuccessful in reducing the number of preterm and LBW deliveries in recent years even though prevention efforts have the potential to save many more infant lives and reduce subsequent morbidity than do further improvements in neonatal care.
In 1997, 65% of all infant deaths occurred to the 7.5% of infants born LBW, and 51% of all infant deaths occurred to the 1.4% of infants born VLBW.8 Almost 9 out of 10 infants with birth weights <500 g die within the first year of life—most within the first few days of life (Table 7). An infant's chances of survival increase rapidly thereafter with increasing birth weight. At birth weights of 1250 to 1499 g, >95 out of 100 infants now survive the first year of life. IMRs are lowest for infants weighing 4000 to 4499 g, with small increases among the heaviest infants.
IMRs are lower for infants of black mothers than for infants of non-Hispanic white mothers for individual birth weight categories under 2500 g, but are higher at birth weights of 2500 g or more. In contrast, IMRs are higher for infants of black mothers than for infants of non-Hispanic white mothers for the broad birth weight categories of <1500 g and <2500 g. The reason for this discrepancy is that the birth weight distribution for infants of black mothers includes a much higher proportion of births at extremely low birth weights, and, as a result, influences the IMRs for these broad birth weight groupings. The largest relative difference, nevertheless, in birth weight-specific IMRs between infants of non-Hispanic white and black mothers is for infants weighing 2500+ g (2.5 and 4.1, respectively). Thus, much of the excess mortality for black infants can be explained by 2 factors: 1) A birth weight distribution with a higher incidence of LBW, VLBW, and preterm births among infants of black mothers; and 2) higher IMRs for black infants weighing 2500+ g.
Birth weight-specific IMRs for infants of Hispanic mothers are quite similar to those for non-Hispanic white mothers, but with slightly higher IMRs for infants born weighing 750 to 1999 g, and slightly lower IMRs for infants born weighing 2000 to 3999 g.
From 1990 to 1997, IMRs declined rapidly (by 34%–39%) for infants weighing 750 to 1499 g at birth, moderately (by 24%–31%) for infants weighing 500 to 749, and 1500 to 3999 g, and more slowly (by 19%–22%) for infants weighing >4000 g. In contrast, mortality rates for infants born at <500 g declined very little (4%) from 1990–1997, reflecting the limited success of intensive efforts made to save these very small infants. The few infants who do survive at these VLBWs may suffer lifetime disabilities such as blindness, mental retardation, and neurologic disorders, necessitating increased levels of medical and parental care.30,,31
Tables 8 and9 present information on state variations in LBW, IMR, and birth weight-specific NMR. The percent LBW relates mostly to prenatal factors such as the proportion of multiple births, preexisting maternal risk factors (eg, smoking, maternal medical conditions), and the quality and content of prenatal care. The NMR for 500 to 1499-g births more closely reflects the quality and availability of neonatal intensive care. Infant mortality can be viewed as a composite measure, reflecting the influences of these, and other factors (for example, postneonatal mortality attributable to sudden infant death syndrome [SIDS] or injuries).
Percent LBW and IMR by state are presented in Table 8 for 1997. Oregon, South Dakota, and Washington State had the lowest percent of LBW births (5.5–5.6), while Mississippi (10.1), Louisiana (10.2), and the District of Columbia (13.4) had the highest. When examining IMRs, Maine, Massachusetts, and New Hampshire had the lowest rates in 1997, and the District of Columbia, Mississippi, West Virginia, South Carolina, Louisiana, and Alabama had the highest. Although the highest rates of both LBW and IMR were for the District of Columbia, it is more appropriate to compare these rates to those for other large US cities because of the high concentrations of high-risk women in these areas. Differences in LBW and IMR by state reflect, in part, differences in the racial, ethnic, and socioeconomic composition of their populations.
NMRs for infants with birth weights of 500 to 1499 g are presented in Table 9. Because of small annual numbers of neonatal deaths at 500 to 1499 g in some states, data are presented for a 3-year time period (1995–1997), and confidence intervals are provided to aid in the interpretation of differences. States with rates that were significantly lower than the national average were Minnesota, Utah, Kentucky, Massachusetts, New Jersey, Texas, and Florida. Those with rates that were significantly higher than the national average were Nebraska, Arizona, Indiana, and Illinois.
It is possible to compare the statistics presented here for a given state to get an idea of how well the state is doing in overall infant mortality, and then to gain further insight into areas of strength or weakness in regard to maternal risk factors and prenatal care (%LBW), neonatal intensive care (NMRs for 500- to 1499-g births), or other factors. For example, the high IMR for the District of Columbia relates primarily to the high percent of LBW births to black mothers, who constitute the majority of mothers in the District of Columbia. In contrast, the NMR for 500- to 1499-g births is not significantly different from the national average. The comparatively low IMR for New Hampshire (4.3) primarily reflects a lower percentage of LBW births; although the NMR for 500- to 1499-g births is also lower than the US average, the difference is not statistically significant. Finally, although Nebraska has the highest NMR for 500- to 1499-g births, the IMR is not significantly different from the US average, attributable to Nebraska's relatively low percentage of LBW births.
Leading Causes of Infant Death
The 10 leading causes of infant death for 1998 are shown in Table 10. About half of all infant deaths were attributable to the 4 leading causes of infant death: congenital anomalies, disorders relating to short gestation and unspecified LBW, SIDS, and respiratory distress syndrome (RDS) and newborns affected by maternal complications of pregnancy. IMRs from congenital anomalies declined by 38% from 1979 to 1998, less than the decline for overall infant mortality. IMRs for RDS declined by 63% since 1989 when new medical treatments for this condition became widely available.32–34 Of concern, however, is the lack of any significant decline since 1979 for disorders related to short gestation and unspecified LBW, the second leading cause of infant death. Following slow declines during the 1980s, SIDS rates fell >40% after 1992 when the American Academy of Pediatrics issued a recommendation to reduce the risk of SIDS by placing infants on their backs or sides to sleep.35–37
Table 11 shows the IMR for countries with a population of at least 2.5 million and for which the IMR was lower than the US rate in 1995, 1996, or 1997. These data were obtained primarily from the United Nations' 1997 Demographic Yearbook38 and confirmed by United Nations personnel (M. Otero, personal communication, September 1999). Hong Kong has been retained in the table for purposes of consistency with previous years despite its change in status. The order of some of the countries has shifted in the table from last year, but generally the countries in the top half, with the lowest IMRs, have remained the same.
As in previous years, the United States is in the unenviable position of having one of the highest IMRs among developed countries, along with the highest birth rate. Some caution should be taken in interpreting these comparisons as the high IMR in the United States compared with other countries may be attributable in part to variations in reporting requirements for infant deaths and for live birth registration.39–41 Nevertheless, the unfavorable birth weight distribution in the United States continues to be one reason for its unenviable IMR, and as noted earlier, the proportion of the smallest and most vulnerable infants has been increasing over the 1990s.
There were an estimated 2 338 070 deaths in the United States in 1998 (Table 1), 23 825 more than the 2 314 245 deaths reported in 1997. The preliminary death rate for 1998 was 865.0 deaths per 100 000 population, essentially the same as the final 1996 rate of 864.7. The preliminary age-adjusted death rate for 1998 was 470.8 deaths per 100 000 US standard population.6 This rate was almost 2% lower than the final 1997 age-adjusted death rate of 479.1 and was a record low for the United States.5,,6 Age-adjusted death rates are better indicators of the risk of mortality over time than crude death rates because they control for variations in the age composition of the population; the aging of the US population results in higher crude death rates, despite lower age-specific rates.
Expectation of Life
The estimated expectation of life at birth for a given year represents the average number of years that a group of infants would be expected to live if, throughout their lifetime, they were to experience the age-specific death rates prevailing during the year of their birth. Based on preliminary data for 1998, the expectation of life at birth reached a new record high of 76.7 years, an increase of 0.2 year from the previous year. Life expectancy increased from the previous year by 0.6 years for black males, 0.3 years for black females, 0.3 years for white males, and was the same for white females, setting record highs for the 3 former groups, and matching the record high set in 1997 for white females. In 1998, life expectancy at birth was 79.9 years for white females, 75.0 years for black females, 74.6 years for white males, and 67.8 years for black males.
Causes of Death
Based on preliminary data, the 10 leading causes of death in 1998 accounted for approximately 80% of all US deaths (Table 12). Between 1997 and 1998, age-adjusted death rates declined for a number of causes of death including: HIV infection by 21%, homicide by 14%, suicide by 6%, accidents and adverse effects (unintentional injuries) by 5%, chronic liver disease and cirrhosis (cirrhosis) by 4%, cerebrovascular diseases by 4%, diseases of heart by 3%, and malignant neoplasms (cancer) by 2%. Among the 10 leading causes of death in 1998, age-adjusted death rates increased for pneumonia and influenza by 5%, nephritis, nephrotic syndrome, and nephrosis (kidney disease) by 2%, and chronic obstructive pulmonary disease by 2%.
HIV and Other Infectious Diseases
Deaths attributable to infectious diseases increased noticeably between 1979 and 1994, but the age-adjusted death rate from infectious diseases overall declined by 45% between 1995 and 1998; they account for a relatively small percentage of all deaths (2.2% in 1998). In 1987, a special cause of death category was created in the United States to uniquely classify deaths attributable to HIV infection.42 Between 1987 and 1994, the age-adjusted death rate for HIV infection rose an average of 16% annually, before leveling off in 1995. After 1995, mortality attributable to HIV infection fell by 71%, and HIV infection dropped out of the 15 leading causes of death in 1998. This downturn has been attributed to improvements in the treatment of the disease.43,,44 HIV, however, remains a leading cause for specific age, race, and sex groups. Other infectious diseases, however, have continued to increase. Although septicemia increased from 1979 to 1998, in most cases physicians should have provided additional information on why septicemia developed. The age-adjusted rate for viral hepatitis more than quadrupled during the same period. In 1998, septicemia accounted for 46% of deaths attributable to infectious and parasitic diseases, while HIV infection accounted for 26% and viral hepatitis, for 9%.
Homicide and Suicide
The age-adjusted death rate for homicide and legal intervention (homicide) declined by 14%, from 8.0 per 100 000 in 1997 to 6.9 in 1998. Since 1979, the age-adjusted homicide rate has fluctuated; however, the 1998 rate is 37% less than the peak of 10.7 in 1991; 1998 is the fifth consecutive year in which homicide rates have decreased. The age-adjusted death rate for suicide continued to decline with a 6% decline from 1997 to 1998, and is 16% less than the peak of 11.9 in 1986.
Deaths Among Children
Based on preliminary data, an estimated 26 522 children and adolescents between the ages of 1 and 19 years died in the United States in 1998 (Table 13). The death rate for children 1 to 4 years old in 1998 was 34.2 per 100 000 population, 4% lower than the rate of 35.8 in 1997. From 1997 to 1998, the respective death rates for children and teenagers aged 5 to 9, 10 to 14, and 15 to 19 declined by 5%, 6%, and 7%. Since 1979, death rates have declined by 47% for children 1 to 4 years old, by 43% for children 5 to 9 years old, by 31% for children 10 to 14 years old, and by 30% for teenagers 15 to 19 years old.
For children 1 to 4 years of age, unintentional injury was the leading cause of death with congenital anomalies and homicide as the second and third causes, respectively. Unintentional injuries accounted for 36% of all deaths in this age group. Death rates for unintentional injuries and congenital anomalies have decreased 53% and 57%, respectively, since 1979. An estimated 355 children 1 to 4 years old died from cancer, making cancer the fourth leading cause of death in this age group. Since 1979, death rates for cancer for this age group have declined by 50%.
For children 5 to 9 years old, unintentional injury, cancer, congenital anomalies, and homicide were the leading causes of death. Unintentional injury accounted for nearly 43% of all deaths in 1998 while the second leading cause, cancer, accounted for 14% of all deaths in this age group. Since 1979, death rates attributable to each of the these leading causes of death have declined by at least 20%.
For children 10 to 14 years of age, unintentional injury was the leading cause of death and accounted for 39% of all deaths in this age group. The second leading cause was cancer, followed by suicide and homicide. The death rate attributable to unintentional injuries and cancer for this age group has declined by 48% and 32%, respectively, since 1979. The death rate for suicide more than doubled from 1979 to 1993 but decreased by 6% from 1993 to 1998. The homicide death rate doubled from 1979 to 1993 but declined 44% between 1993 and 1997.
For teenagers aged 15 to 19 years, the leading cause of death, unintentional injuries, accounted for 46% of all deaths in 1998. The death rate for unintentional injuries has dropped by 46% since 1979. An estimated 2216 teenagers were victims of homicide, the second leading cause, in 1998, accounting for 16% of all deaths. The teen homicide rate peaked in 1993 and decreased 45% between 1993 and 1998. Suicide was the third leading cause of death for this age group, accounting for 13% of all deaths. From 1979 to 1988, suicide rates increased by 32% for this age group and remained fairly stable from 1988 to 1994, but declined by 22% between 1994 and 1998.
Summary of Childhood Deaths
In conclusion, death rates for children and teenagers dropped by 38% between 1979 and 1998. Despite declines, the leading cause of death in 1998 among children and teenagers continued to be unintentional injury. These declines have been attributed, in part, to injury prevention measures such as mandatory car seat requirements, smoke alarms, and sprinkler systems in homes and schools.45 Congenital anomalies was the second leading cause of death at ages 1 to 4 years but dropped in importance for successively older age groups as cancer, homicide, and suicide became more prominent. Cancer accounts for 5% to 14% of deaths for each of the age groups under 19 years of age. Decreases in recent years for homicide and suicide, hopefully, signal the beginning of a long-term reversal in mortality from these causes among children. Yet, a large proportion of childhood deaths continue to occur as a result of preventable injuries.46 American pediatricians must further strengthen their efforts to prevent many of these deaths.
USE OF VITAL STATISTICS: EXAMPLES OF EVALUATION OF REGIONALIZATION OF PERINATAL CARE IN 4 STATES
As noted earlier, the rate of infant mortality in a population is a function of both the distribution of birth weight and birth weight-specific mortality rates in that population. In the past few years, we have expanded the data on birth weight for the United States by presenting estimates of birth weight-specific mortality from linked birth and infant death data files. This year, we go one step further by presenting IMRs by state for the most vulnerable live births, those weighing between 500 and 1499 g. These rates along with the percentage of VLBW infants in the population have a profound effect on overall IMRs. As we close this year's article, we present an example of how the data available from linked certificates have been used to assess the organization of perinatal services in 4 states.
A major systems approach to organizing clinical care for pregnant women and newborns in the past 3 decades has been regionalization of perinatal care. Regionalization evolved as an important concept in the early 1970s because of the limited availability of both neonatologists and perinatologists at the time, coupled with a climate of regional planning and the need to make efficient use of scarce and expensive resources to care for high-risk mothers and newborns. The early stimulus for regionalization was to improve access to intensive care for high-risk mothers and newborns through the development of a system that linked hospitals and providers at different levels of care. These levels were initially designated as primary, secondary, and tertiary (today also called basic, specialty, and subspecialty, respectively) and were differentiated based on the complexity of the patient needs and the capabilities of provider or hospital to meet these needs.47
There is considerable evidence pointing to the success of regionalized systems in the 1970s in concentrating births of high-risk infants at tertiary centers, in reducing death rates of LBW infants in these centers, and in improving the survival of these infants at community hospitals, presumably through infant transport or provider education.48 Reductions in neonatal mortality in states and other geographic areas were also reported after regionalization, particularly for LBW infants, although a cause-and-effect relationship of regionalization to reduced mortality has not been empirically documented.49 Despite considerable progress in the 1970s and 1980s, recent changes have occurred in the organization of perinatal services in a number of states that threaten to erode existing systems. These changes have been documented empirically using vital statistics data and information about the level of perinatal care provided in hospitals throughout several states.
One measure of the extent to which states have been successful in assuring access of high-risk mothers and newborns to risk appropriate care is the proportion of VLBW infants born in subspecialty hospitals. In 2 states, California50 and Washington,51there has been a rise, albeit not large, in the proportion of VLBW infants born in level II (specialty) hospitals. Although this rise may not appear alarming on the face of it, in the above 2 states as well as in South Carolina52 and Missouri,53 NMRs for VLBW infants born in specialty hospitals are similar to those for infants born in community hospitals and higher than the rates for subspecialty hospitals. Moreover, Phibbs et al50 noted both a volume and level of care effect in California; the lowest NMRs were found in large subspecialty hospitals, with all other hospitals, including smaller subspecialty hospitals having similar and consistently higher rates.
The rise in level II, and to a lesser extent self-designated level III hospitals, is thought to be a result of competition for perinatal patients. Development of a full-service hospital is seen as a marketing advantage because the availability of neonatal intensive care unit care is attractive to pregnant women, and women tend to continue to use the same facility for family care after birth. Moreover, the number of neonatologists has increased considerably in recent years, yielding more specialists to staff the newly designated level II and III facilities.50
These analyses point to difficult policy decisions for states regarding the voluntary designation of hospitals and the shifting health care market. States must use such analyses of data from their vital statistics systems to monitor the impact of changing health care policies on populations, highlighting again the importance of these data and the need to ensure their accuracy.
We thank Mary Anne Freedman for contributions to the manuscript and Melissa Park for preparation of the tables. Manuel Otero, Senior Statistician at the United Nations Statistical Office, assisted with the verification of the IMR data for the international comparisons. We wish to give special thanks to Becky Newcomer for her patience with the authors and preparation of the manuscript.
- Received October 4, 1999.
- Accepted October 4, 1999.
Reprint requests to (B.G.) Johns Hopkins School of Hygiene and Public Health, 615 N Wolfe St, Baltimore, MD 21205.
- HIV =
- human immunodeficiency virus •
- IMR =
- infant mortality rate •
- LBW =
- low birth weight •
- TFR =
- total fertility rate •
- VBAC =
- vaginal birth after previous cesarean delivery •
- VLBW =
- very low birth weight •
- NMR =
- neonatal mortality rate •
- PNMR =
- postneonatal mortality rate •
- SIDS =
- sudden infant death syndrome •
- RDS =
- respiratory distress syndrome
- Baumgartner L
- Wegman ME
- Wegman ME
- ↵Ventura SJ, Martin JA, Curtin SC, Mathews TJ. Births: Final Data for 1997. National Vital Statistics Reports. Vol 47, No. 18. Hyattsville, MD: National Center for Health Statistics; 1999
- ↵Hoyert DL, Kochanek KD, Murphy SL. Deaths: Final Data for 1997. National Vital Statistics Reports. Vol 47, No. 19. Hyattsville, MD: National Center for Health Statistics; 1999
- ↵Martin JA, Smith BL, Mathews TJ, Ventura SJ. Births and Deaths: Preliminary Data for 1998. National Vital Statistics Reports. Vol 47, No 24. Hyattsville, MD: National Center for Health Statistics; 1999
- ↵World Health Organization. Manual of the International Statistical Classification of Diseases, Injuries and the Causes of Death, Based on the Recommendations of the Ninth Conference-1975. Geneva, Switzerland: World Health Organization; 1977
- ↵MacDorman MF, Atkinson JO. Infant Mortality Statistics from the 1997 Period Linked Birth/Infant Death Data Set. National Vital Statistics Reports. Vol 47, No 23. Hyattsville, MD: National Center for Health Statistics; 1999
- ↵Hollmann FW, Kuzmeskus LB, Perkins RC, Weber EA, Deardorff KE, Montgomery P. US Population Estimates by Age, Sex, Race, and Hispanic Origin: 1990 to 1997. Washington, DC: US Bureau of the Census, PPL-91; 1998
- ↵US Bureau of the Census. Unpublished census file NESTV98. WK1. Consistent with populations published in: US Bureau of the Census. Population Estimates of the US by Age, Sex, Race, and Hispanic Origin: 1990;-1998. Washington, D. C.: US Bureau of the Census. http://www.census.gov/population/www/estimates/uspop.html Internet release, June 4, 1999
- ↵Ventura SJ, Mathews TJ, Curtin SC. Declines in Teenage Birth Rates, 1991–98: Update of National and State Trends. National Vital Statistics Reports. Vol. 47, No. 26. Hyattsville, MD: National Center for Health Statistics; 1999
- ↵Ventura SJ, Mathews TJ, Curtin SC. Declines in Teenage Birth Rates, 1991–97: National and State Patterns. National Vital Statistics Reports. Vol. 47, No 12. Hyattsville, MD: National Center for Health Statistics; 1998
- ↵Maynard RA, ed. Kids Having Kids: A Robin Hood Foundation Special Report on the Costs of Adolescent Childbearing. New York, NY: The Robin Hood Foundation; 1996
- ↵Ventura SJ. Trends and Variations in First Births to Older Women, 1970–86. Vital and Health Statistics 21(47). Hyattsville, MD: National Center for Health Statistics; 1989
- ↵Lugaila TA. Marital Status and Living Arrangements: March 1998 (Update). US Bureau of the Census. Current Population Reports P20–514. Washington, DC: US Department of Commerce; 1998
- ↵Mathews TJ. Smoking During Pregnancy, 1990–96. National Vital Statistics Reports. Vol 47, No. 10. Hyattsville, MD: National Center for Health Statistics; 1998
- Kleinman JC,
- Madans JH
- Schoendorf KC,
- Kiely JL
- Cunningham J,
- Dockery DW,
- Speizer FE
- Fiscella K
- ↵Curtin SC, Park MM. Trends in Attendant, Place and Timing of Births, and in the Use of Obstetric Interventions. United States: 1989–97. National Vital Statistics Reports; Vol 47, No 27. Hyattsville, MD: National Center for Health Statistics; 1999
- Flamm BL
- ↵Martin JA, Park MM. Trends in Twin and Triplet Births: 1980–97. National vital statistics reports; Vol 47, No 25. Hyattsville, MD: National Center for Health Statistics; 1999
- ↵Maryland Commission on Infant Mortality Prevention. “Closing the Gap;” Addressing the Disparity of Infant Mortality Among African-American and White Infants. Approaches for Clinicians and Policy Makers. Baltimore, MD: Governor's Office of Children, Youth and Families; 1997
- Ericson A,
- Kallen B
- AAP Task Force on Infant Positioning and SIDS
- Willinger M,
- Hoffman HJ,
- Hartford RB
- Gibson E,
- Cullen JA,
- Spinner S,
- Rankin K,
- Spitzer A
- ↵Department of Economic and Social Affairs. 1997 Demographic Yearbook. New York, NY: United Nations; 1999
- Liu K,
- Moom M,
- Sulvetta M,
- Chawala J
- Sepkowitz S
- Fenton AL,
- Field DJ,
- Mason E,
- et al.
- ↵National Center for Health Statistics. Technical Appendix. In: Vital Statistics of the United States, 1987. Vol II. Mortality. Pt A. Washington, DC: Public Health Service; 1987
- ↵Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report. Vol 10, No 1. Atlanta, GA: US Department of Health and Human Services; 1998
- ↵Kaplan JE, Masur H, Holmes KK. 1999 USPHS/IDSA Guidelines for the Prevention of Opportunistic Infections in Persons Infected with Human Immunodeficiency Virus. MMWR Morb Mortal Wkly Rep. 1999;48(RR-10)
- Rivara FP,
- Grossman DC
- American Academy of Pediatrics
- ↵National Committee on Perinatal Health. Toward Improving the Outcome of Pregnancy. White Plains, NY: National Foundation–March of Dimes; 1976
- Copyright © 1999 American Academy of Pediatrics