Annual Summary of Vital Statistics—2003
http://www.100md.com
《小儿科》
Division of Vital Statistics, National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, Maryland
Department of Population and Family Health Sciences, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland
ABSTRACT
The crude birth rate rose slightly in 2003 to 14.1 births per 1000 population, from 13.9 in 2002. The 2002 rate was the lowest ever reported for the United States. The total number of births and the fertility rate (66.1) also increased. The birth rate for teenaged mothers dropped 3% to another record low in 2003, to 41.7 per 1000 females aged 15 to 19 years. The teenage birth rate has fallen by one third since 1991. The birth rate declined for women 20 to 24 years old but rose for women aged 25 to 44 years. The number, rate, and proportion of births to unmarried women all increased in 2003.
Smoking during pregnancy declined to 11%, down from 19.5% in 1989. Prenatal care utilization improved slightly for 2003; 84.1% of women began care in the first trimester of pregnancy. The cesarean delivery rate jumped 6% to 27.6% for another US high. The primary cesarean rate rose 6%, and the rate of vaginal birth after a previous cesarean delivery plummeted 16% from 2002 to 2003. The percent of infants delivered preterm continued to rise (12.3% in 2003). The preterm birth rate is up 16% since 1990. The percentage of children born at low birth weight rose slightly in 2003 to the highest level reported since 1970 (7.9%). The twinning rate increased, but the rate for triplet/+ births declined slightly between 2001 and 2002. Multiple births accounted for 3.3% of all births in 2002.
The infant mortality rate rose to 7.0/1000 live births in 2002 from 6.8 in 2001, marking the first increase in this rate in >4 decades. Increases were distributed fairly widely across age, racial/ethnic groups, and geographic areas. The rise in infant mortality was attributed to increases in <750-g births in both singleton and multiple deliveries. Although the downward trend in infant mortality rates in many developed nations may have stabilized, the United States still ranked 27th among these nations in 2001.
Expectation of life at birth reached a record high of 77.3 years for all gender and race groups combined in 2002. Death rates in the United States continue to decline. Between 2001 and 2002, death rates declined for the 3 leading causes of death: diseases of heart, malignant neoplasms, and cerebrovascular diseases. Death rates for children 1 to 19 years old decreased by 8% for suicide; the death rate for chronic lower respiratory diseases increased by 33% in 2002. Rates for unintentional injuries and homicide did not change significantly for children aged 1 to 19 years. A large proportion of childhood deaths continues to occur as a result of preventable injuries.
Key Words: birth birth weight–specific mortality death infant mortality low birth weight mortality multiple births vital statistics revised populations
Abbreviations: NCHS, National Center for Health Statistics IMR, infant mortality rate NMR, neonatal mortality rate PNMR, postneonatal mortality rate PMR, perinatal mortality rate FMR, fetal mortality rate OMB, Office of Management and Budget LBW, low birth weight VBAC, vaginal birth after previous cesarean ART, assisted reproductive technology VLBW, very low birth weight SIDS, sudden infant death syndrome
This article is a long-standing feature in Pediatrics and had appeared in each December issue since 1950. Until 1995, the article was written by Myron Wegman, a pediatrician and public health leader who died in 2004 at the age of 95. Since 1995, the article has been written by colleagues from the Johns Hopkins Bloomberg School of Public Health and the National Center for Health Statistics (NCHS). Publication of the "Annual Summary of Vital Statistics—2003" was delayed several months as the result of sweeping changes occurring in the vital statistics system1 (see "The Transformation of Vital Statistics" at the end of this article).
In this article we provide a summary of vital statistics data through 2003. We expanded the section on infant mortality and added a brief section on perinatal mortality to give larger context to the increase in infant mortality that occurred between 2001 and 2002. Finally, we include a special feature explaining the fundamental changes occurring in the vital statistic system and the potential short- and long-term implications of the change.
METHODS
The data presented in this report were obtained from vital statistics records: birth certificates, fetal death reports, and death certificates for residents of the US. Data for 2002 and earlier years are final and include all records. Birth data for 2003 are preliminary but are based on nearly 99% of births reported to the Centers for Disease Control and Prevention's NCHS. Mortality data for 2003 are provisional and are based on counts of death certificates reported to NCHS by state health departments. More complete descriptions of vital statistics data systems are available elsewhere.2–5 Preliminary birth estimates for 2003 may differ from the final data for 2003 that will include all records, but differences are usually small.
Current vital statistics patterns and recent trends through 2003 are presented in this report by age, race, and Hispanic origin as well as other birth and death characteristics. More detailed data are available in the final birth and death files for 2002 than in the preliminary and provisional files for 2003, and thus some of the detailed analyses of birth and death patterns focus on the 2002 data. Data on infant deaths from the linked birth/infant death data set are final data for 2002.
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 include data separately for non-Hispanic white, non-Hispanic black, and Hispanic women. Data for Native American and Asian or Pacific Islander women are not shown separately by Hispanic origin, because the vast majority of these women are not Hispanic.
Mother's marital status was reported directly on the birth certificates or through the electronic birth-registration process in all but 2 states (Michigan and New York) in 2002 and 2003. Details about the reporting of marital status in those 2 states and methods of edits and imputations applied to other items on the birth certificate are presented in NCHS publications.2,4,6
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 which initiated the train of morbid events leading directly to death, or (b) the circumstances of the accident or violence which produced the fatal injury."7(p30) From 1999 to the present, cause-of-death data in the United States are classified according to the International Classification of Diseases, 10th Revision.7 Ranking for leading causes of death is based on number of deaths.8
Infant mortality refers to the death of an infant <1 year old. Infant mortality rates (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.3,5,9 Neonatal mortality rates (NMRs) are shown for infants dying between 0 and 27 days old, and postneonatal mortality rates (PNMRs) are shown for infants dying between 28 days and 1 year old. Perinatal mortality rates (PMRs) include fetal deaths at 28 weeks of gestation and infant deaths at <7 days of age. Fetal mortality rates (FMRs) are shown for fetal deaths at 20 weeks of gestation. FMRs and PMRs were computed by dividing the number of fetal or perinatal deaths by the number of live births plus fetal deaths. IMR, NMR, PNMR, PMR, and FMR are all shown per 1000 live births.
The latest infant mortality statistics by race and Hispanic origin were obtained from the 2002 period linked birth/infant death data set.9 In this data set, the death certificate was linked with the corresponding birth certificate for each infant who died in the United States in 2002. 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 1% of infant deaths for whom the matching birth certificate could not be identified.9 The weighting procedure results in the same overall IMR as that based on unlinked mortality data; however, small differences may exist because of geographic coverage differences, additional quality control, and weighting.9
Two States, Pennsylvania and Washington, implemented the 2003 revision of the US Standard Certificate of Live Birth for data year 2003.4 This article includes 2003 birth data for the 2 revised states and for the remaining 48 states and the District of Columbia, which reported birth data based on the 1989 revision of the US Standard Certificate of live birth. The prenatal care item on the 2003 revision of the US Standard Certificate of Live Birth, however, differs from that of the 1989 certificate.10,11 The revised question is more specific in that it asks for the "date of the first prenatal visit" compared with the former: "month that prenatal care began." The source of the revised and unrevised data also may differ as new worksheets are used to encourage the use of the proper sources. As a result, preliminary 2003 data on prenatal care for states that implemented the revised certificates (Pennsylvania and Washington) are not compatible with those from the states that are based on the 1989 revision; therefore, for this article, data on prenatal care exclude the 2 states that implemented revised certificates.4
Population denominators for the calculation of birth, death, and fertility rates are estimates based on the population enumerated by the US Census Bureau as of April 1, 2000. Estimates for 2000-2003 and the revised estimates for the intercensal period 1991-1999 were produced under a collaborative arrangement between the US Census Bureau and the NCHS. Reflecting the new guidelines issued in 1997 by the Office of Management and Budget (OMB), the 2000 census included an option for individuals to report >1 race as appropriate for themselves and household members.12 The 1997 OMB guidelines also provided for the reporting of Asian persons separately from Native Hawaiians or other Pacific Islanders. Under the prior OMB standards issued in 1977, data for Asian or Pacific Islander persons were collected as a single group.13 Birth and death certificates for most states currently collect only 1 race for mother and decedent in the same categories as specified in the 19774 OMB guidelines and do not report Asians separately from Native Hawaiians or other Pacific Islanders. Thus, birth and death certificate data by race (the numerators for birth and death rates) are currently incompatible with the population data collected in the 2000 census (the denominators for the rates).
To produce birth and death rates for 2000-2003 and revised intercensal rates for the 1991-1999 period, it was necessary to "bridge" the reported population data for multiple-race persons back to single-race categories. In addition, the 2000 census counts were modified to be consistent with the 1977 OMB race categories, that is, to report the data for Asians and Native Hawaiians or other Pacific Islanders as a combined category of Asian or Pacific Islanders.14 The procedures used to produce the bridged populations are described in separate publications.15,16 Rates based on bridged population data may differ from previously published rates.2 Several states began reporting multiple-race birth and death data for 2003 events (California, Hawaii, Ohio, Pennsylvania, Utah, and Washington for births; California, Hawaii, Idaho, Maine, Montana, New York, and Wisconsin for deaths). As national vital statistics data based on the 1997 OMB guidelines become available, the use of bridged populations can be discontinued.
Data for the international comparisons of births, birth rates, and IMRs were obtained from the 2001 United Nations Demographic Yearbook.17
NATURAL INCREASE
Almost 1.7 million persons were added to the US population in 2003, the result of natural increase (the excess of births over deaths).4,5 The rate of natural increase rose from 5.4 to 5.8 persons per 1000 population between 2002 and 2003 (Table 1).
BIRTHS
Preliminary data indicate that the number of US births rose 2% for 2003, to 4091063 births (Table 1).4 The 2003 birth rate also rose slightly, from 13.9 to 14.1 births per 1000 population. The 2002 birth rate was the lowest reported for the US since national data became available.2 The fertility rate (number of births per 1000 women aged 15–44 years) rose 2% to 66.1 in 2003 from 64.8 in 2002. Since 1994, the fertility rate has ranged from 63.6 to 66.1. The total fertility rate estimates 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. The total fertility rate also rose 2% for the current year to 2044. The 2003 total fertility rate is only 2% lower than the most recent high reported for 1990 (2081).4 In 2003, the only group with a total fertility level >2100 or "replacement" (the rate at which a given generation exactly replaces itself) was Hispanic women (2785); levels for all other groups were below replacement.
Racial and Ethnic Composition
The general fertility rate rose among non-Hispanic white (2%), Hispanic (3%), and Asian or Pacific Islander women (3%) between 2002 and 2003, declined slightly among non-Hispanic black women (<1%), and was essentially unchanged for Native American women.4 Fertility rates had generally declined for all race and Hispanic-origin groups between 1990 and 2002.
Fertility rates by race and ethnicity often differ widely (Table 2). The 2003 fertility rate for Hispanic women, 96.9 births per 1000 aged 15 to 44 years in preliminary 2003 data,4 was more than one third higher than those of other groups: non-Hispanic black (67.1), Asian or Pacific Islander (66.3), Native American (57.9), and non-Hispanic white (58.5) women. In 2003, >1 of every 5 births (22%) was delivered to a woman of Hispanic origin, compared with 15% of all births in 1990.
Trends in Age-Specific Birth Rates
Teenage Childbearing
The more than decade-long decline in teenage childbearing continued in 2003 (Table 3). The birth rate for teenagers dropped to 41.7 births per 1000 women 15 to 19 years old in 2003, 3% lower than in 2002 and one third lower than the recent peak in 1991 (61.8 per 1000). The 2003 rate is another record low for the nation, the lowest in the >6 decades for which comparable data have been available.4,18 Birth rates have fallen for teenagers of all ages; however, the birth rate for the youngest mothers aged 10 to 14 years has declined the most rapidly. The rate for teenagers 10 to 14 years old declined to 0.6 in 2003, less than half of the 1991 rate of 1.4 per 1000 and the lowest rate recorded since 1946.19 Infants born to the youngest mothers are at 3 times the risk of early death compared with infants born to mothers 20 to 24 years old.19 The birth rate for teenagers 15 to 17 years old fell another 3% to 22.4 per 1000 for 2003 and is down 42% from 1991; the rate for teenagers 18 to 19 years old was also down 3% for the current year to 70.8 and has declined 25% since the 1991 peak.
According to preliminary 2003 data, teenage birth rates fell for all race and ethnic groups (Table 3).4 The largest overall decline was for non-Hispanic black teenagers, down 5% between 2002 and 2003; the birth rate for black teenagers has plunged 45% since 1991. For the current year, teenage childbearing declined 4% among non-Hispanic white and Asian/Pacific Islander teenagers and 1% and 2% for Hispanic and Native American teenagers, respectively. Marked declines in teenage childbearing have been observed for all groups since 1991 (see Fig 1 and Table 3).4 Despite these reductions, however, US teenage birth rates remain among the highest among industrialized nations.17
Not all pregnancies end in live birth. Among teenagers, an estimated 56% of pregnancies ended in live birth, 29% in induced abortion, and 15% in fetal loss in 2000, the most recent year for which pregnancy data are available.20 From 1990 to 2000, the teenage pregnancy rate fell 27% from 116.3 to 84.5 per 1000 females 15 to 19 years old. The rate for 2000 is the lowest reported since 1976.20 The declines in teenage pregnancy have been much steeper for younger than for older teenagers: the rate for teenagers 15 to 17 years old fell one third during 1990-2000, whereas the rate for older teenagers dropped one fifth. The pregnancy rate for the youngest teenagers, <15 years old, has fallen as well from 3.5 to 2.1 per 1000. Regardless of age or race/ethnicity group, rates have fallen for all teenage pregnancy outcomes, with larger reductions for abortion than for birth rates.
Recently published data from cycle 6 of the National Survey of Family Growth, conducted in 2002, suggest some factors that may account for the steady and substantial declines in teenage pregnancy and birth rates.21 Teenagers are delaying sex until older ages. The proportions of teenaged females and males who had ever had sexual intercourse declined between 1995 and 2002. Teenage females are more likely to have used contraception at first intercourse (approximately three fourths) and the proportions of sexually active teenagers using the pill, condoms, and new hormonal methods all increased. There were also increases in the proportions using dual methods.21,22
Childbearing for Women 20 Years Old
The birth rate for women 20 to 24 years old declined 1% between 2002 (103.6 per 1000) and 2003 (102.6 per 1000). Rates for women of this age group have generally trended downward since the early 1990s. The preliminary birth rate for women 25 to 29 years old increased 2% to 115.7, the highest level reported since 1992.2,4
Birth rates for women 30 to 44 years old rose to the highest levels in 3 decades.2,4 The rate increased 4% among women 30 to 34 years old (to 95.2 births per 1000 in 2003), 6% for women 35 to 39 years old (43.8 per 1000), and 5% for women 40 to 44 years old (8.7 per 1000). The birth rate for women 40 to 44 years old has risen 58% since 1990 (from 5.5 per 1000). The rate for women 45 to 54 years old, which had doubled in the 1990s, was stable at 0.5 per 1000. Delayed childbearing and more frequent use of fertility therapies are 2 factors that have been linked to rising birth rates among women 35 years old.2,23
The age distribution of births in the United States has shifted substantially in recent years. Although women in their 20s still contribute the bulk of all births, just more than half for 2003, this level is down from nearly two thirds of all births in 1980. In contrast, the proportion of births to women 30 years old has almost doubled over this period, increasing from 19% to 38%.4,24
Unmarried Mothers
The number of births to unmarried women increased 4% in 2003 (preliminary data).4 This rise was primarily due to an increase in the birth rate among unmarried women, which rose 3% (from 43.7 to 44.9 births per 1000 unmarried women), and to a lesser extent to an increase in the population of unmarried women 15 to 44 years old (up 1%). The birth rate for unmarried women is still lower than the 1994 peak (46.9 per 1000).4 In 2003, 34.6% of all births were to unmarried women, up from 34.0% in 2002. This level has risen very slowly from 32.8 in 1998.2,25 Nonmarital childbearing increased for non-Hispanic white (23.5%) and Hispanic (45.0%) women but was essentially unchanged for non-Hispanic black women (68.5%) between 2002 and 2003.
Smoking During Pregnancy
The proportion of women who smoked during pregnancy declined 4% from 2002 to 2003 (Table 4). This level has declined steadily since at least 1989, when this information first was reported on the birth certificate.4,26 In 2003, 11.0% of women reported smoking during pregnancy, down from 19.5% in 1989. Among those women who still smoke, the amount of tobacco usage also seems to have declined substantially.2 The harmful consequences of tobacco use during pregnancy are well documented; it is a risk factor for poor birth outcome (low birth weight [LBW], intrauterine growth retardation, and infant mortality) and can negatively affect the long-term health of the child.27,28
Declines in prenatal tobacco use were reported among women of all age groups and for each of the largest racial/ethnic groups. For 2003 as for earlier years, levels were lowest among Asian/Pacific Islander (2.3% smokers) and Hispanic (2.8%) women, intermediate for non-Hispanic black women (8.5%), and highest for American Indian (18.3%) and non-Hispanic white (14.5%) women in 2003.
Although smoking among pregnant teenagers declined for 2003 according to preliminary data, teenage smoking rates remain high at 16%. More than 1 of every 4 non-Hispanic white teenagers and 1 of every 5 American Indian teenagers was reported to have smoked during pregnancy in 2003.4
Prenatal Care
For this article, data on prenatal care excludes the 2 states that implemented revised certificates for 2003.4,10,11 As a result, prenatal care levels for years 2002 and 1990 presented in Table 4 by race and Hispanic origin may differ from those published previously.2 The percent of women who began prenatal care in their first trimester of pregnancy edged upward for 2003, increasing to 84.1% from 83.7% in 2002 (Table 4). Timely initiation of prenatal care has been on a slow but steady rise since 1990 (75.6%).4 The proportion of women beginning care in the last trimester of pregnancy, or with no care at all, was 3.5% in 2003 compared with 3.6% in the previous year. The proportion of women with late or no care has dropped from 6.1% since 1990. Recent gains in prenatal care utilization may be attributed in part to the expansion of Medicaid for pregnant women in the late 1980s.29 Appropriate prenatal care can improve pregnancy outcome by performing risk assessment and offering health care advice, psychosocial support, education on childbirth, and care coordination.30,31
Small increases in first-trimester care from 2002 to 2003 were found for non-Hispanic white (89% for 2003), non-Hispanic black (76.0%), and Hispanic (77.4%) women. Although fairly wide disparities in prenatal care utilization persist by race and Hispanic origin, considerable progress in timely receipt of care has been observed especially among groups with less advantageous levels; since 1990 the proportions of non-Hispanic black and Hispanic women with first-trimester care has risen 25% to 29%.2,4
Cesarean Delivery
Preliminary data for 2003 indicate a 6% jump in the rate of cesarean delivery to 27.6% of all births for another record US high (see Table 4 and Fig 2). After declines for 1989–1996, the cesarean delivery rate has risen each year; the current level is one third higher than that for 1996 (20.7%).4,2 Consistent with the recent trend, the primary cesarean rate (percent of cesareans among women with no previous cesarean delivery) rose 6% from 2002 to 2003, from 18.0% to 19.1%, and the rate of vaginal birth after previous cesarean (VBAC) fell 16%, from 12.6% to 10.6%. The primary rate has climbed 31%, and the VBAC rate has dropped 63% since only 1996.
The increase in the rate of primary cesareans may be related to changes in nonmedical risk factors such as maternal age, physician practice patterns, and maternal choice.32–34 Increased use of medical technologies such as induction of labor may also have contributed to the rise; induction before 41 weeks of gestation has been linked to an increased risk of cesarean delivery.33 The precipitous decline in the VBAC rate may be related to reports on the risks associated with VBACs, more conservative practice guidelines, and legal pressures.33,35–37
The primary cesarean rate has risen and the VBAC rate has fallen among low-risk women (ie, women with term, singleton deliveries with vertex presentations) at a pace similar to that for all women.38,39 Primary cesareans rose and VBAC rates declined among all 5 racial/ethnic groups studied from 2002–2003; rates ranged from 24.2% among American Indian women to 29.3% for non-Hispanic black women.4
Multiple Births
The birth rate for twins increased another 3% in 2002, rising to 31.1 twin births per 1000 total births (Table 4) (data for 2002 are the latest available). The twin birth rate has climbed 38% since 1990 (from 22.6 per 1000) and 65% since 1980 (18.9 per 1000).2 The birth rate for triplets and other higher-order multiples (triplet/+) was down slightly for 2002, however, to 184.0 per 100000 live births, from 185.6 in 2001 (Fig 3). Between 1980 and 1998, the triplet/+ birth rate soared from 37.0 to 193.5 per 100000, with an average annual increase of 13% for 1990–1998. This rate has trended slightly downward since, however, suggesting that the remarkable surge in triplet/+ births during the last 2 decades may have subsided. The 2002 rate is 5% lower than the 1998 peak (193.5).2 The 125134 twins, 6898 triplets, 434 quadruplets, and 69 quintuplet and higher-order multiple births accounted for 3.3% of all 2002 births.2
The rise in multiple births has been especially steep among births to women in the oldest childbearing ages; for example, the number of twin births to women 45 to 49 years old jumped from only 39 to 991 per 1000 total births between 1990 and 2002; the twin birth rate climbed from 23.8 to 189.7 per 1000.2
The increase in multiple births, especially higher-order multiples, has been associated with 2 related trends: older age at childbearing and increased use of fertility therapies (assisted reproductive technologies [ARTs] such as in vitro fertilization and non-ART therapies such as ovulation-inducing drugs).40,41 A recent study found that 43% of triplets resulted from ART, 40% were likely the result of non-ART therapies, and 18% were conceived naturally.42
Multiple births tend to be high-risk events. The average twin is delivered 3 weeks and the average triplet >6 weeks earlier than the typical singleton. Accordingly, the risk of early death is substantially elevated; IMRs for twins are 5 times and triplets 12 times those of singletons.2,9 Recognizing their increased risk of poor outcome, the American College of Obstetricians and Gynecologists and the American Society of Reproductive Medicine issued recommendations in 1999 intended to prevent triplet/+ pregnancies.43,44 These recommendations, plus refinements to ARTs, may be contributing to the recent small downturn in the higher-order multiple-birth trend.
Preterm Births
Approximately 500000 infants or 12.3% of all births were delivered preterm in 2003. The preterm birth rate (percent of infants delivered at <37 completed weeks of gestation) increased from 12.0% to 12.3% between 2002 and 2003 and has risen 16% since 1990 (from 10.6%) (Table 4).2,4 Three factors potentially related to this rise are higher rates of multiple births, more obstetric intervention, and greater use of ARTs45
Between 1990 and 2002, the preterm rate for singletons rose 7% (compared with a 14% rise for all births); all of the increase for singletons was among those born moderately preterm or between 32 and 36 weeks of gestation; the proportion of singletons delivered at <32 weeks declined slightly over this time period (from 1.69 to 1.57%).
Although outcomes for infants delivered before term have improved notably in recent years, preterm newborns remain at considerably heightened risk of death or lifelong disability. The IMR for infants delivered at <32 weeks of gestation is nearly 75 times greater than that of term infants, and the rate for those born at 32 to 36 weeks of gestation is >3 times that of those delivered at longer gestations.9
The overall rise in preterm rates is largely driven by the increase among non-Hispanic white births, up by one third between 1990 and 2003 (11.3% for 2003). The preterm rate has increased 8% among Hispanic births (11.9%) and is down by 6% for non-Hispanic black births (17.8%) (Table 4). The etiologies of preterm delivery are not fully understood, and until progress is made in this regard, substantial reduction in the preterm birth rate is unlikely.45,46
LBW
LBW (<2500 g) births rose to 7.9% in 2003 from 7.8% in 2002 (Table 4) and has increased 18% from the 1984 low (6.7%); the current level is the highest reported since 1970.2,4 Very low birth weight (VLBW; births weighing <1500 g) births were essentially unchanged at 1.4% in 2003. VLBW has risen from 1.15% in 1980 but has been essentially stable since 1998. LBW and especially VLBW are major predictors of infant morbidity and mortality. The risk of early death for infants born at moderately LBW (1500–2499 g) is 5 times higher than that of heavier infants; the risk for VLBW infants is >100 times that of infants born at 2500 g.9
Between 2002 and 2003 LBW rates rose or were unchanged among each of the major racial/ethnic groups (see Table 4). Since 1990, LBW levels have climbed 25% for non-Hispanic white births (to 7.0% in 2003), 10% among Hispanic births (6.7%), and 2% among non-Hispanic black births (13.5%). Some of this increase, particularly among non-Hispanic white births, can be attributed to increases in the rate of multiple births, which tend to be born much smaller than singletons. In 2002, almost 1 of every 4 LBW infants was born in a multiple delivery.2 Among singleton births only, the increase in LBW has been more modest. Between 1990 and 2002 the LBW rate rose from 5.90% to 6.12%; the VLBW level rose from 1.05 to 1.11%.2
INFANT MORTALITY
The IMR rose in 2002 to 7.0 deaths per 1000 live births from 6.8 per 1000 in 2001. The historic trend in the US IMR has been one of steady decline; 2002 marks the first increase in this rate in >40 years3,9,47,48 (see Table 5 and Fig 4 for the 1990–2002 trend). The NMR rose from 4.5 to 4.7 per 1000 for 2001–2002, whereas the PNMR was unchanged at 2.3 per 1000 (Table 5). Provisional data suggest that the IMR will not increase again in 2003.5
Increases in infant mortality between 2001 and 2002 were distributed fairly widely across age groups, racial/ethnic groups, and geographic areas.49 A recently released study based on data from the 2002 linked birth/infant death file found that the 2001–2002 increase in infant mortality was driven primarily by a shift in the birth weight distribution toward infants weighing <750 g and particularly toward those weighing <500 g.49 The change in the birth weight distribution accounted for essentially all (98%) of the increase in the 2002 IMR. Multiple births contributed disproportionately (25%) to the overall increase, but 75% of the rise was due to increases among singletons.49 The rise in <750-g births was predominately among women 20 to 34 years old and was not attributed to increases in high-risk births to older mothers. The reasons for the rise in the numbers of these extremely small infants are not clear but may be related to changes in the reporting of vital events, changes in the risk profiles of mothers giving birth, changes in the management of labor and delivery, and the increased use of ARTs.49
Small increases in IMRs were reported for most racial and ethnic groups between 2001 and 2002, but the changes were not statistically significant.9,49 The relative difference in rates between non-Hispanic black (13.9 per 1000) and non-Hispanic white (5.8 per 1000) expressed as a ratio of black/white IMRs was 2.4 in 2002, unchanged for more than a decade (Table 5). The 2002 IMR of 5.6 per 1000 for Hispanic births was slightly lower than that for non-Hispanic white births. Rates among the Hispanic subgroups differ widely, however, from a low of 3.7 for Cuban to a high of 8.2 among births to Puerto Rican infants.
Birth Weight–Specific Infant Mortality
2002 also marks a change in the long-term trend in declining birth weight–specific mortality rates for VLBW infants. For 2001–2002 the birth weight–specific IMR for VLBW (<1500 g) infants increased significantly from 244.4 to 250.0 infant deaths per 1000 live births (Table 6). 9,49 This marks the first increase in mortality for VLBW infants since statistics on this measure became available.49 Although the increase in birth weight–specific mortality for very small infants is cause for concern, as noted above, it was the shift in the birth weight distribution that explained the bulk of the rise in overall infant mortality for 2001–2002.
The 2002 IMR for infants with birth weights of 2500 g was unchanged from the previous year at 2.4 per 1000, whereas the rate for moderately LBW (1500–2499 g) infants was 15.2 per 1000 in 2001, compared with 15.1 per 1000 in 2002 (Table 6).
Between 1985 and 2001, the IMR declined by more than one third (from 10.6 to 6.8 per 1000),3 but the rate of LBW rose 13% (from 6.8% to 7.7%) and the VLBW rate rose from 1.21% to 1.44%.2,9 Thus, all of the decline in the IMR since the mid-1980s was due to declines in birth weight–specific IMRs, declines that occurred among both smaller and heavier infants. These changes were attributed primarily to improvements in obstetric and neonatal care.50
Perinatal and Fetal Mortality
The PMR more fully describes the risk of death at late stages of pregnancy and soon after birth, because it includes both late fetal (28 weeks of gestation) and early neonatal (<7 days) deaths. The PMR has been declining fairly consistently for more than half a century.51 For 2001–2002, a sizable decline in late fetal deaths slightly outpaced the increase in early neonatal deaths, and as a result, the PMR was unchanged at 6.9 deaths per 1000 live births plus fetal deaths.
The FMR (fetal deaths of at least 20 weeks of gestation) was 6.4 per 1000 live births plus fetal deaths for 2002 compared with 6.5 per 1000 for 2001. The FMR has declined substantially over the last 50 years and is down 15% since 1990.52 All of the recent decline in the FMR, however, has been among late fetal deaths; the late FMR declined 23% between 1990 and 2001. In contrast, the early FMR (fetal deaths of 20–27 weeks) has not improved over this period.52 For 2001–2002, the late FMR declined another 3% from 3.3 to 3.2 per 1000; the early FMR was unchanged.
Geographic Variation
Three states, Massachusetts, Maine, and Vermont, reported IMRs <5.0 per 1000 for 2002. In contrast, Louisiana and Mississippi reported rates >10.0. Whereas IMRs increased for more than half of the states between 2001 and 2002, statistically significant increases in rates were reported in 4 states: Kentucky, Missouri, New Jersey, and Texas.
Variations by state in LBW and IMR reflect compositional differences by race, ethnicity, and socioeconomic status in the population in addition to other factors (prenatal, quality-of-care, and postnatal influences on infants) that are associated with LBW or IMR.
Leading Causes of Infant Death
More than half of all infant deaths in 2002 were attributable to 5 leading causes: congenital malformations (20%), disorders relating to short gestation and unspecified LBW (17%), sudden infant death syndrome (SIDS) (8%), newborns affected by maternal complications of pregnancy (6%), and newborns affected by complications of placenta, cord, and membranes (4%) (Table 8). 3 The order of the first 5 leading causes of death was unchanged for 2001–2002.
Cause-specific IMRs increased significantly between 2001 and 2002 for LBW and maternal complications; it is important to note, however, that part of the increase in the latter cause of death is a result of a change in coding rules.3,9 Rates were also higher for congenital malformations and SIDS, but these changes were not statistically significant. For 2000–2001, the death rate for SIDS dropped 11%.53
The rank order of leading causes of death can vary by race or Hispanic origin. For example, in 2002 as in earlier years, LBW (not congenital malformations) was the leading cause of death among non-Hispanic black and Puerto Rican infants.
INTERNATIONAL COMPARISONS
For all 3 years, the IMR is lower for all countries in the table than the US rate with 1 exception; the IMR in the US was lower than the rate in Cuba in 2000. There has been a downward trend in the IMR in less than half of the countries, suggesting a possible trend toward stabilization in the rates in recent years. Nevertheless, the US IMR in 2001 was at least twice as great as the rate for the 6 countries with the lowest levels.
One reason for the high IMR in the US is the high rate of VLBW births in the country relative to other developed countries. It is unlikely that VLBW births will decrease much even among singleton births, because current trends show no declines in recent years. There may be some reporting differences across countries related to distinguishing between live births and fetal deaths,54–57 but the magnitude of these differences is unlikely to be large enough to account for the large disparities between US rates and the rest of the developed world.
DEATHS
There were 2423000 deaths (provisional data) in the United States in 2003 (Table 1), 20387 less than the 2443387 deaths reported in 2002.3,5 The death rate for 2002 was 847.3 deaths per 100000 population, no different from the 2001 rate of 848.5 per 100000. 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.3 The age-adjusted death rate for 2002 was 845.3 deaths per 100000 US standard population. This rate was 1.1% lower than the 2001 age-adjusted death rate of 854.5 100000 and was a record low for the United States.
Expectation of Life
The estimated expectation of life at birth for a given year represents the average number of years that a cohort 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. In 2002, the expectation of life at birth reached a new record high of 77.3 years, an increase of 0.1 year from the previous year. Between 2001 and 2002, life expectancy increased by 0.2 years for black men (68.8) and by 0.1 year for white men (75.1), white women (80.3), and black women (75.6). All 4 groups attained record high levels.
Causes of Death
The 15 leading causes of death in 2002 accounted for >83% of all US deaths (Table 10). 3 Between 2001 and 2002, age-adjusted death rates declined significantly for 4 of the 15 leading causes: diseases of heart by 2.8%, malignant neoplasms (cancer) by 1.3%, cerebrovascular diseases (stroke) by 2.9%, and assault (homicide) by 14.1%.3 Age-adjusted death rates increased significantly for 7 of the 15 leading causes of death: accidents (unintentional injuries) by 3.4%, influenza and pneumonia by 2.7%, Alzheimer's disease by 5.8%, nephritis, nephrotic syndrome and nephrosis (kidney disease) by 1.4%, septicemia by 2.6%, intentional self-harm (suicide) by 1.9%, and essential (primary) hypertension and hypertensive renal disease (hypertension) by 2.9%. The apparently dramatic decline in the homicide rate in 2002 was primarily a result of the September 11, 2001, terrorist attacks, which added 2926 certified resident deaths to this category in 2001. Without the additional deaths resulting from the terrorist attacks in 2001, the homicide rate would have remained the same between 2001 and 2002.
Deaths Among Children
In 2002, 25820 children and adolescents between the ages of 1 and 19 years died in the United States (Table 11). 3 The death rate for this age group was 33.5 per 100000 population, which is approximately the same as the rate in 2001 (33.6 per 100000). From 2001 to 2002, the death rate for children aged 1 to 4 years decreased by 6%. The death rates for children aged 5 to 9, 10 to 14, and 15 to 19 years did not show any significant changes between 2001 and 2002.
For all children aged 1 to 19 years, the first and second leading causes of death in 2002 were accidents (unintentional injuries) and assault (homicide). Unintentional injuries accounted for 44.5% of all deaths, and homicide accounted for 10.3%. The rate for unintentional injuries increased by 2% from the rate in 2001 and homicide increased by 3%, but the changes were not statistically significant. The rate for suicide (the fourth leading cause of death) decreased by 8%, and the rate for chronic lower respiratory diseases increased by 33%. For each of 4 age groups, 1 to 4, 5 to 9, 10 to 14, and 15 to 19 years old, Table 11 provides detailed information on the 5 leading causes of death by rank in 2002 as well as the number of deaths, the percentage accounted for by each of the leading causes, and the cause-specific death rate for each for the years 2001 and 2002. Although some causes increased or decreased in specific age groups, the overall patterns of these death rates have not changed significantly since last year.
THE TRANSFORMATION OF VITAL STATISTICS
Vital statistics in the United States are in the midst of a sweeping transformation.1,10,11,58,59 The experience of reporting and explaining the rise in the IMR for 2001–2002 underscores the urgent need for change. The 2002 increase in infant mortality could not be reported until 200447 and it is not until 2005 that a thorough exploration of the changes that led to the rise will be published.49 Delays of this magnitude are currently largely unavoidable. The long-term result of the fundamental changes in data collection and processing will be vast improvements in the quality and timeliness of vital statistics data and improved flexibility of electronic systems to allow for more immediate response to emerging trends. In the short term, however, increased delays in data release (as evidenced by the delay in publication of this annual article) and the potential loss of national data for some very basic measures of public health are anticipated.
The hundred-plus-year-old vital statistics system of the United States is a state-based, decentralized system.60 These state systems, with only a few exceptions, are based on outmoded registration practices and systems. To rebuild these aging systems more effectively, the National Association of Public Health Statistics and Information Systems, NCHS, and the Social Security Administration are working together to develop national standards and guidelines for electronic vital statistics systems.10,59 The new systems will incorporate standardized data-collection instruments, improved methods for capturing data, immediate query of suspect data, query and edit guidelines, and detailed item definitions. These systems will use the revised 2003 US Standard Certificates of Birth and Death and Report of Fetal Death that introduce a wealth of new and enhanced data items. The new systems also will be designed to integrate with other health information systems and will be easily modifiable to capture new data items as needed and drop those that no longer are.
The reengineered system(s) will potentially transform vital statistics, resulting in key enhancements in data flow and, ultimately, the release of these data within months (not years) of the event. The new systems will automatically link birth and infant death records. This linkage will allow these data to be available simultaneously and will facilitate more timely and comprehensive analysis of, for example, changes in infant mortality. The new and modified data items on the 2003 revised certificates of birth, death, and fetal death will greatly expand our ability to describe and explain such trends. For example, with information from the new birth certificate, we will be able to directly estimate the impact of the use of infertility therapies on perinatal outcome. Information on whether a "trial of labor" was attempted before cesarean delivery will help us to better assess the impact of changes in management of labor and delivery. Finally, better-quality data resulting from more immediate querying of data at the source while the records and respondent are still available, plus the use of detailed and standardized specifications and definitions, will help, for example, to better describe maternal medical-risk profiles such as changes in the prevalence of hypertensive disorders on pregnancy outcome.
Full implementation of reengineered vital statistics systems across the US jurisdictions will take some years. At the close of 2004, 8 states had reengineered their birth systems, and 2 states had partially reengineered their deaths systems. Much has been accomplished, but much work is left to be done to ensure that vital statistics improves its key role in monitoring the nation's maternal and infant health.
ACKNOWLEDGMENTS
We thank Martha Munson, Stephanie Ventura, and Brady Hamilton for important contributions to the manuscript and Sharon Kirmeyer, T.J. Mathews, Brady Hamilton, and Fay Menacker for content review.
FOOTNOTES
Accepted Dec 9, 2004.
No conflict of interest declared.
REFERENCES
Rothwell CL, Sondik EJ, Guyer B. A delay in publication of the "Annual Summary of Vital Statistics" and the need for new vital registration and statistics systems for the United States [commentary]. Pediatrics. 2004;114 :1671 –1672
Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Munson ML. Births: final data for 2002. Natl Vital Stat Rep. 2003;52(10) :1 –113
Kochanek KD, Murphy SL. Deaths: final data for 2002. Natl Vital Stat Rep. 2004;53(5) :1 –115
Hamilton BE, Martin JA, Sutton PD. Births: preliminary data for 2003. Natl Vital Stat Rep. 2004;53(9) :1 –17
Munson ML, Sutton PD. Births, marriages, divorces and deaths: provisional data for 2003. Natl Vital Stat Rep. 2004;52(22) :1 –7
World Health Organization. International Statistical Classification of Diseases and Related Health Problems, Tenth Revision. Geneva, Switzerland: World Health Organization; 1992
Anderson RN, Smith BL. Deaths: leading causes for 2001. Natl Vital Stat Rep. 2003;52(9) :1 –85
Mathews TJ, Menacker F, MacDorman MF. Infant mortality statistics from the 2002 period linked birth/infant death data set. Natl Vital Stat Rep. 2004;53(1) :1 –29
Office of Management and Budget. Race and ethnic standards for Federal statistics and administrative reporting. Statistical Policy Directive 15. 1977
Ingram DD, Parker JD, Schenker N, et al. United States Census 2000 with bridged race categories. Vital Health Stat 2. 2003;(135) :1 –55
Schenker N, Parker JD. From single-race reporting to multiple-race reporting: using imputation methods to bridge the transition. Stat Med. 2003;22 :1571 –1587
Hamilton BE, Sutton PD, Ventura SJ. Revised birth and fertility rates for the 1990s and new rates for Hispanic populations, 2000 and 2001: United States. Natl Vital Stat Rep. 2003;51(12) :1 –94
United Nations. United Nations Demographic Yearbook. New York, NY: United Nations; 2003
Ventura SJ, Mathews TJ, Hamilton BE. Births to teenagers in the United States, 1940–2000. Natl Vital Stat Rep. 2001;49(10) :1 –23
Menacker F, Martin JA, MacDorman MF, Ventura SJ. Births to 10–14 year-old mothers, 1990–2002: trends and health outcomes. Natl Vital Stat Rep. 2004;53(7) :1 –18
Ventura SJ, Abma JC, Mosher WD, Henshaw S. 2004. Estimated pregnancy rates for the United States, 1990–2000: an update. Natl Vital Stat Rep. 2004;52(23) :1 –9
Abma JC, Martinez GM, Mosher WD, Dawson BS. Teenagers in the United States: Sexual Activity, Contraceptive Use, and Childbearing. Vital and Health Statistics Series 23, No. 24. Hyattsville, MD: National Center for Health Statistics; 2004
Mosher WD, Martinez GM, Chandra A, Abma JC, Wilson SJ. Use of Contraception and Use of Family Planning Services in the United States: 1982–2002. Advance Data No. 350. Hyattsville, MD: National Center for Health Statistics; 2004
Ventura SJ. Trends and Variations in First Births to Older Women, 1970–86. Vital and Health Statistics Series 21, No. 47. Hyattsville, MD: National Center for Health Statistics; 1989
National Center for Health Statistics. Advance Report of Final Natality Statistics, 1980. Monthly Vital Statistics Report, Vol. 31, No. 8. Suppl. DHHS Pub.No. (PHS)83–1120. Hyattsville, MD: Public Health Service: 1982
Ventura SJ, Bachrach CA. Nonmarital childbearing in the United States, 1940–99. Natl Vital Stat Rep. 2000;48(16) :1 –40
Mathews TJ. Smoking during pregnancy in the 1990s. Natl Vital Stat Rep. 2001;49(7) :1 –14
Office of the Surgeon General, Public Health Service, US Department of Health and Human Services. Women and Smoking: A Report of the Surgeon General. Washington, DC: US Department of Health and Human Services; 2001
Ventura SJ, Hamilton BE, Mathews TJ, Chandra A. Trends and variations in smoking during pregnancy and low birth weight: evidence from the birth certificate, 1990–2000. Pediatrics. 2003;111 :1176 –1180
Howell EM. The impact of Medicaid expansions for pregnant women: a synthesis of the evidence. Med Care Res Rev. 2001;58 :3 –30
Fiscella K. Does prenatal care improve birth outcomes A critical review. Obstet Gynecol. 1995;85 :468 –479
US Public Health Service. Caring for Our Future: The Content of Prenatal Care. Washington, DC: US Department of Health and Human Services; 1989
Zinberg, S. Vaginal delivery after previous cesarean delivery: a continuing controversy. Clin Obstet Gynecol. 2001;44 :561 –570
American College of Obstetricians and Gynecologists. Evaluation of Cesarean Delivery. ACOG Guidelines. Washington, DC: American College of Obstetricians and Gynecologists; 2000
Minkoff H, Chervenak FA. Elective primary cesarean delivery. N Engl J Med. 2003;348 :946 –950
McMahon MJ, Luther ER, Bowes WA, Olshan AF. Comparison of a trial of labor with an elective second cesarean section. N Engl J Med. 1996;335 :689 –695
Lydon-Rochelle M, Holt VL, Easterling TR, Martin DP. Risk of uterine rupture during labor among women with a prior cesarean delivery. N Engl J Med. 2001;345 :3 –8
Guise JM, McDonagh M, Hashima J, et al. Vaginal Births After Cesarean (VBAC). Evidence Report/Technology Assessment No. 71 (Prepared by the Oregon Health and Science University Evidence-Based Practice Center under contract no. 209-97-0018). Agency for Healthcare Research and Quality Publication No. 03-E018. Rockville, MD: Agency for Healthcare Research and Quality; 2003
Declercq E, Menacker F, MacDorman MF. Rise in "no indicated risk" primary caesareans in the United States: cross sectional analysis. BMJ. doi:10.1136/bmj.38279.705336.OB
US Department of Health and Human Services. Tracking Healthy People 2010. B16-20. Washington, DC; US Government Printing Office; 2000
Martin JA, Park MM. Trends in twin and triplet births: 1980–97. Natl Vital Stat Rep. 1999;47(24) :1 –16
Reynolds MA, Schieve LA, Jeng G, Peterson HB, Wilcox LS. Risk of multiple birth associated with in vitro fertilization using donor eggs. Am J Epidemiol. 2001;154 :1043 –1050
Reynolds MA, Schieve LA, Martin JA, Jeng G, Macaluso M. Trends in multiple births conceived using assisted reproductive technology, United States, 1997–2000. Pediatrics. 2003;111 :1159 –1166
American College of Obstetricians and Gynecologists. Nonselective Embryo Reduction: Ethical Guidance for the Obstetrician-Gynecologist. ACOG Committee Opinion 215. Washington, DC: American College of Obstetricians and Gynecologists; 1999
American Society for Reproductive Medicine. Guidelines on Number of Embryos Transferred. A Practice Committee Report—A Committee Opinion. Birmingham, AL: American Society for Reproductive Medicine; 1999
Tucker J, McGuire W. Epidemiology of preterm birth. BMJ. 2004;329 :675 –678
Johnson RB, Williams MA, Hogue CJR, Mattison DR. Overview: new perspectives on the stubborn challenge of preterm birth. Paediatr Perinat Epidemiol. 2001;15(suppl 2) :3 –6
Kochanek KD, Smith BL. Deaths: preliminary data for 2002. Natl Vital Stat Rep. 2004;52(13) :1 –47
MacDorman MF, Martin JA, Mathews TJ, Hoyert DL, Ventura SJ. Explaining the 2001–2002 infant mortality increase: data from the linked birth/infant death data set. Natl Vital Stat Rep. 2005;53(12) :1 –22
Allen MC, Alexander GR, Tompkins ME, Hulsey TC. Racial differences in temporal changes in newborn viability and survival by gestational age. Paediatr Perinat Epidemiol. 2000;14 :152 –158
Fried VM, Prager K, MacKay AP, Zia H. Chartbook on Trends in the Health of Americans. Health, United States, 2003. Hyattsville, Maryland: National Center for Health Statistics; 2003
Barfield W, Martin JA, Hoyert DL. Racial/ethnic trends in fetal mortality—United States, 1990–2000. MMWR Morb Mortal Wkly Rep. 2004;53 :529 –532
Mathews TJ, Menacker F, MacDorman MF. Infant mortality statistics from the 2001 period linked birth/infant death data set. Natl Vital Stat Rep. 2003;52(2) :1 –28
Liu K, Moon M, Sulvetta M, Chawala J. International infant mortality rankings: a look behind the numbers. Health Care Financ Rev. 1992;13 :105 –118
Sepkowitz S. International rankings of infant mortality and the United States' vital statistics natality data collecting system—failure and success. Int J Epidemiol. 1995;24 :583 –588
Haub C, Yanagishita M. Infant mortality: who's number one Popul Today. 1991;19(3) :6 –8
Kramer MS, Platt RW, Yang H, Haglund B, Cnattingius S, Berjso P. Registration artifacts in international comparisons of infant mortality. Paediatr Perinat Epidemiol. 2002;16 :16 –22
Rothwell CJ. Reengineering vital registration and statistics systems for the united states. Prev Chronic Dis [serial online]. 2004;1(4) :1 –2
Hetzel AM. History and Organization of the Vital Statistics System. Hyattsville, MD: National Center for Health Statistics; 1997
Department of Population and Family Health Sciences, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland
ABSTRACT
The crude birth rate rose slightly in 2003 to 14.1 births per 1000 population, from 13.9 in 2002. The 2002 rate was the lowest ever reported for the United States. The total number of births and the fertility rate (66.1) also increased. The birth rate for teenaged mothers dropped 3% to another record low in 2003, to 41.7 per 1000 females aged 15 to 19 years. The teenage birth rate has fallen by one third since 1991. The birth rate declined for women 20 to 24 years old but rose for women aged 25 to 44 years. The number, rate, and proportion of births to unmarried women all increased in 2003.
Smoking during pregnancy declined to 11%, down from 19.5% in 1989. Prenatal care utilization improved slightly for 2003; 84.1% of women began care in the first trimester of pregnancy. The cesarean delivery rate jumped 6% to 27.6% for another US high. The primary cesarean rate rose 6%, and the rate of vaginal birth after a previous cesarean delivery plummeted 16% from 2002 to 2003. The percent of infants delivered preterm continued to rise (12.3% in 2003). The preterm birth rate is up 16% since 1990. The percentage of children born at low birth weight rose slightly in 2003 to the highest level reported since 1970 (7.9%). The twinning rate increased, but the rate for triplet/+ births declined slightly between 2001 and 2002. Multiple births accounted for 3.3% of all births in 2002.
The infant mortality rate rose to 7.0/1000 live births in 2002 from 6.8 in 2001, marking the first increase in this rate in >4 decades. Increases were distributed fairly widely across age, racial/ethnic groups, and geographic areas. The rise in infant mortality was attributed to increases in <750-g births in both singleton and multiple deliveries. Although the downward trend in infant mortality rates in many developed nations may have stabilized, the United States still ranked 27th among these nations in 2001.
Expectation of life at birth reached a record high of 77.3 years for all gender and race groups combined in 2002. Death rates in the United States continue to decline. Between 2001 and 2002, death rates declined for the 3 leading causes of death: diseases of heart, malignant neoplasms, and cerebrovascular diseases. Death rates for children 1 to 19 years old decreased by 8% for suicide; the death rate for chronic lower respiratory diseases increased by 33% in 2002. Rates for unintentional injuries and homicide did not change significantly for children aged 1 to 19 years. A large proportion of childhood deaths continues to occur as a result of preventable injuries.
Key Words: birth birth weight–specific mortality death infant mortality low birth weight mortality multiple births vital statistics revised populations
Abbreviations: NCHS, National Center for Health Statistics IMR, infant mortality rate NMR, neonatal mortality rate PNMR, postneonatal mortality rate PMR, perinatal mortality rate FMR, fetal mortality rate OMB, Office of Management and Budget LBW, low birth weight VBAC, vaginal birth after previous cesarean ART, assisted reproductive technology VLBW, very low birth weight SIDS, sudden infant death syndrome
This article is a long-standing feature in Pediatrics and had appeared in each December issue since 1950. Until 1995, the article was written by Myron Wegman, a pediatrician and public health leader who died in 2004 at the age of 95. Since 1995, the article has been written by colleagues from the Johns Hopkins Bloomberg School of Public Health and the National Center for Health Statistics (NCHS). Publication of the "Annual Summary of Vital Statistics—2003" was delayed several months as the result of sweeping changes occurring in the vital statistics system1 (see "The Transformation of Vital Statistics" at the end of this article).
In this article we provide a summary of vital statistics data through 2003. We expanded the section on infant mortality and added a brief section on perinatal mortality to give larger context to the increase in infant mortality that occurred between 2001 and 2002. Finally, we include a special feature explaining the fundamental changes occurring in the vital statistic system and the potential short- and long-term implications of the change.
METHODS
The data presented in this report were obtained from vital statistics records: birth certificates, fetal death reports, and death certificates for residents of the US. Data for 2002 and earlier years are final and include all records. Birth data for 2003 are preliminary but are based on nearly 99% of births reported to the Centers for Disease Control and Prevention's NCHS. Mortality data for 2003 are provisional and are based on counts of death certificates reported to NCHS by state health departments. More complete descriptions of vital statistics data systems are available elsewhere.2–5 Preliminary birth estimates for 2003 may differ from the final data for 2003 that will include all records, but differences are usually small.
Current vital statistics patterns and recent trends through 2003 are presented in this report by age, race, and Hispanic origin as well as other birth and death characteristics. More detailed data are available in the final birth and death files for 2002 than in the preliminary and provisional files for 2003, and thus some of the detailed analyses of birth and death patterns focus on the 2002 data. Data on infant deaths from the linked birth/infant death data set are final data for 2002.
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 include data separately for non-Hispanic white, non-Hispanic black, and Hispanic women. Data for Native American and Asian or Pacific Islander women are not shown separately by Hispanic origin, because the vast majority of these women are not Hispanic.
Mother's marital status was reported directly on the birth certificates or through the electronic birth-registration process in all but 2 states (Michigan and New York) in 2002 and 2003. Details about the reporting of marital status in those 2 states and methods of edits and imputations applied to other items on the birth certificate are presented in NCHS publications.2,4,6
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 which initiated the train of morbid events leading directly to death, or (b) the circumstances of the accident or violence which produced the fatal injury."7(p30) From 1999 to the present, cause-of-death data in the United States are classified according to the International Classification of Diseases, 10th Revision.7 Ranking for leading causes of death is based on number of deaths.8
Infant mortality refers to the death of an infant <1 year old. Infant mortality rates (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.3,5,9 Neonatal mortality rates (NMRs) are shown for infants dying between 0 and 27 days old, and postneonatal mortality rates (PNMRs) are shown for infants dying between 28 days and 1 year old. Perinatal mortality rates (PMRs) include fetal deaths at 28 weeks of gestation and infant deaths at <7 days of age. Fetal mortality rates (FMRs) are shown for fetal deaths at 20 weeks of gestation. FMRs and PMRs were computed by dividing the number of fetal or perinatal deaths by the number of live births plus fetal deaths. IMR, NMR, PNMR, PMR, and FMR are all shown per 1000 live births.
The latest infant mortality statistics by race and Hispanic origin were obtained from the 2002 period linked birth/infant death data set.9 In this data set, the death certificate was linked with the corresponding birth certificate for each infant who died in the United States in 2002. 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 1% of infant deaths for whom the matching birth certificate could not be identified.9 The weighting procedure results in the same overall IMR as that based on unlinked mortality data; however, small differences may exist because of geographic coverage differences, additional quality control, and weighting.9
Two States, Pennsylvania and Washington, implemented the 2003 revision of the US Standard Certificate of Live Birth for data year 2003.4 This article includes 2003 birth data for the 2 revised states and for the remaining 48 states and the District of Columbia, which reported birth data based on the 1989 revision of the US Standard Certificate of live birth. The prenatal care item on the 2003 revision of the US Standard Certificate of Live Birth, however, differs from that of the 1989 certificate.10,11 The revised question is more specific in that it asks for the "date of the first prenatal visit" compared with the former: "month that prenatal care began." The source of the revised and unrevised data also may differ as new worksheets are used to encourage the use of the proper sources. As a result, preliminary 2003 data on prenatal care for states that implemented the revised certificates (Pennsylvania and Washington) are not compatible with those from the states that are based on the 1989 revision; therefore, for this article, data on prenatal care exclude the 2 states that implemented revised certificates.4
Population denominators for the calculation of birth, death, and fertility rates are estimates based on the population enumerated by the US Census Bureau as of April 1, 2000. Estimates for 2000-2003 and the revised estimates for the intercensal period 1991-1999 were produced under a collaborative arrangement between the US Census Bureau and the NCHS. Reflecting the new guidelines issued in 1997 by the Office of Management and Budget (OMB), the 2000 census included an option for individuals to report >1 race as appropriate for themselves and household members.12 The 1997 OMB guidelines also provided for the reporting of Asian persons separately from Native Hawaiians or other Pacific Islanders. Under the prior OMB standards issued in 1977, data for Asian or Pacific Islander persons were collected as a single group.13 Birth and death certificates for most states currently collect only 1 race for mother and decedent in the same categories as specified in the 19774 OMB guidelines and do not report Asians separately from Native Hawaiians or other Pacific Islanders. Thus, birth and death certificate data by race (the numerators for birth and death rates) are currently incompatible with the population data collected in the 2000 census (the denominators for the rates).
To produce birth and death rates for 2000-2003 and revised intercensal rates for the 1991-1999 period, it was necessary to "bridge" the reported population data for multiple-race persons back to single-race categories. In addition, the 2000 census counts were modified to be consistent with the 1977 OMB race categories, that is, to report the data for Asians and Native Hawaiians or other Pacific Islanders as a combined category of Asian or Pacific Islanders.14 The procedures used to produce the bridged populations are described in separate publications.15,16 Rates based on bridged population data may differ from previously published rates.2 Several states began reporting multiple-race birth and death data for 2003 events (California, Hawaii, Ohio, Pennsylvania, Utah, and Washington for births; California, Hawaii, Idaho, Maine, Montana, New York, and Wisconsin for deaths). As national vital statistics data based on the 1997 OMB guidelines become available, the use of bridged populations can be discontinued.
Data for the international comparisons of births, birth rates, and IMRs were obtained from the 2001 United Nations Demographic Yearbook.17
NATURAL INCREASE
Almost 1.7 million persons were added to the US population in 2003, the result of natural increase (the excess of births over deaths).4,5 The rate of natural increase rose from 5.4 to 5.8 persons per 1000 population between 2002 and 2003 (Table 1).
BIRTHS
Preliminary data indicate that the number of US births rose 2% for 2003, to 4091063 births (Table 1).4 The 2003 birth rate also rose slightly, from 13.9 to 14.1 births per 1000 population. The 2002 birth rate was the lowest reported for the US since national data became available.2 The fertility rate (number of births per 1000 women aged 15–44 years) rose 2% to 66.1 in 2003 from 64.8 in 2002. Since 1994, the fertility rate has ranged from 63.6 to 66.1. The total fertility rate estimates 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. The total fertility rate also rose 2% for the current year to 2044. The 2003 total fertility rate is only 2% lower than the most recent high reported for 1990 (2081).4 In 2003, the only group with a total fertility level >2100 or "replacement" (the rate at which a given generation exactly replaces itself) was Hispanic women (2785); levels for all other groups were below replacement.
Racial and Ethnic Composition
The general fertility rate rose among non-Hispanic white (2%), Hispanic (3%), and Asian or Pacific Islander women (3%) between 2002 and 2003, declined slightly among non-Hispanic black women (<1%), and was essentially unchanged for Native American women.4 Fertility rates had generally declined for all race and Hispanic-origin groups between 1990 and 2002.
Fertility rates by race and ethnicity often differ widely (Table 2). The 2003 fertility rate for Hispanic women, 96.9 births per 1000 aged 15 to 44 years in preliminary 2003 data,4 was more than one third higher than those of other groups: non-Hispanic black (67.1), Asian or Pacific Islander (66.3), Native American (57.9), and non-Hispanic white (58.5) women. In 2003, >1 of every 5 births (22%) was delivered to a woman of Hispanic origin, compared with 15% of all births in 1990.
Trends in Age-Specific Birth Rates
Teenage Childbearing
The more than decade-long decline in teenage childbearing continued in 2003 (Table 3). The birth rate for teenagers dropped to 41.7 births per 1000 women 15 to 19 years old in 2003, 3% lower than in 2002 and one third lower than the recent peak in 1991 (61.8 per 1000). The 2003 rate is another record low for the nation, the lowest in the >6 decades for which comparable data have been available.4,18 Birth rates have fallen for teenagers of all ages; however, the birth rate for the youngest mothers aged 10 to 14 years has declined the most rapidly. The rate for teenagers 10 to 14 years old declined to 0.6 in 2003, less than half of the 1991 rate of 1.4 per 1000 and the lowest rate recorded since 1946.19 Infants born to the youngest mothers are at 3 times the risk of early death compared with infants born to mothers 20 to 24 years old.19 The birth rate for teenagers 15 to 17 years old fell another 3% to 22.4 per 1000 for 2003 and is down 42% from 1991; the rate for teenagers 18 to 19 years old was also down 3% for the current year to 70.8 and has declined 25% since the 1991 peak.
According to preliminary 2003 data, teenage birth rates fell for all race and ethnic groups (Table 3).4 The largest overall decline was for non-Hispanic black teenagers, down 5% between 2002 and 2003; the birth rate for black teenagers has plunged 45% since 1991. For the current year, teenage childbearing declined 4% among non-Hispanic white and Asian/Pacific Islander teenagers and 1% and 2% for Hispanic and Native American teenagers, respectively. Marked declines in teenage childbearing have been observed for all groups since 1991 (see Fig 1 and Table 3).4 Despite these reductions, however, US teenage birth rates remain among the highest among industrialized nations.17
Not all pregnancies end in live birth. Among teenagers, an estimated 56% of pregnancies ended in live birth, 29% in induced abortion, and 15% in fetal loss in 2000, the most recent year for which pregnancy data are available.20 From 1990 to 2000, the teenage pregnancy rate fell 27% from 116.3 to 84.5 per 1000 females 15 to 19 years old. The rate for 2000 is the lowest reported since 1976.20 The declines in teenage pregnancy have been much steeper for younger than for older teenagers: the rate for teenagers 15 to 17 years old fell one third during 1990-2000, whereas the rate for older teenagers dropped one fifth. The pregnancy rate for the youngest teenagers, <15 years old, has fallen as well from 3.5 to 2.1 per 1000. Regardless of age or race/ethnicity group, rates have fallen for all teenage pregnancy outcomes, with larger reductions for abortion than for birth rates.
Recently published data from cycle 6 of the National Survey of Family Growth, conducted in 2002, suggest some factors that may account for the steady and substantial declines in teenage pregnancy and birth rates.21 Teenagers are delaying sex until older ages. The proportions of teenaged females and males who had ever had sexual intercourse declined between 1995 and 2002. Teenage females are more likely to have used contraception at first intercourse (approximately three fourths) and the proportions of sexually active teenagers using the pill, condoms, and new hormonal methods all increased. There were also increases in the proportions using dual methods.21,22
Childbearing for Women 20 Years Old
The birth rate for women 20 to 24 years old declined 1% between 2002 (103.6 per 1000) and 2003 (102.6 per 1000). Rates for women of this age group have generally trended downward since the early 1990s. The preliminary birth rate for women 25 to 29 years old increased 2% to 115.7, the highest level reported since 1992.2,4
Birth rates for women 30 to 44 years old rose to the highest levels in 3 decades.2,4 The rate increased 4% among women 30 to 34 years old (to 95.2 births per 1000 in 2003), 6% for women 35 to 39 years old (43.8 per 1000), and 5% for women 40 to 44 years old (8.7 per 1000). The birth rate for women 40 to 44 years old has risen 58% since 1990 (from 5.5 per 1000). The rate for women 45 to 54 years old, which had doubled in the 1990s, was stable at 0.5 per 1000. Delayed childbearing and more frequent use of fertility therapies are 2 factors that have been linked to rising birth rates among women 35 years old.2,23
The age distribution of births in the United States has shifted substantially in recent years. Although women in their 20s still contribute the bulk of all births, just more than half for 2003, this level is down from nearly two thirds of all births in 1980. In contrast, the proportion of births to women 30 years old has almost doubled over this period, increasing from 19% to 38%.4,24
Unmarried Mothers
The number of births to unmarried women increased 4% in 2003 (preliminary data).4 This rise was primarily due to an increase in the birth rate among unmarried women, which rose 3% (from 43.7 to 44.9 births per 1000 unmarried women), and to a lesser extent to an increase in the population of unmarried women 15 to 44 years old (up 1%). The birth rate for unmarried women is still lower than the 1994 peak (46.9 per 1000).4 In 2003, 34.6% of all births were to unmarried women, up from 34.0% in 2002. This level has risen very slowly from 32.8 in 1998.2,25 Nonmarital childbearing increased for non-Hispanic white (23.5%) and Hispanic (45.0%) women but was essentially unchanged for non-Hispanic black women (68.5%) between 2002 and 2003.
Smoking During Pregnancy
The proportion of women who smoked during pregnancy declined 4% from 2002 to 2003 (Table 4). This level has declined steadily since at least 1989, when this information first was reported on the birth certificate.4,26 In 2003, 11.0% of women reported smoking during pregnancy, down from 19.5% in 1989. Among those women who still smoke, the amount of tobacco usage also seems to have declined substantially.2 The harmful consequences of tobacco use during pregnancy are well documented; it is a risk factor for poor birth outcome (low birth weight [LBW], intrauterine growth retardation, and infant mortality) and can negatively affect the long-term health of the child.27,28
Declines in prenatal tobacco use were reported among women of all age groups and for each of the largest racial/ethnic groups. For 2003 as for earlier years, levels were lowest among Asian/Pacific Islander (2.3% smokers) and Hispanic (2.8%) women, intermediate for non-Hispanic black women (8.5%), and highest for American Indian (18.3%) and non-Hispanic white (14.5%) women in 2003.
Although smoking among pregnant teenagers declined for 2003 according to preliminary data, teenage smoking rates remain high at 16%. More than 1 of every 4 non-Hispanic white teenagers and 1 of every 5 American Indian teenagers was reported to have smoked during pregnancy in 2003.4
Prenatal Care
For this article, data on prenatal care excludes the 2 states that implemented revised certificates for 2003.4,10,11 As a result, prenatal care levels for years 2002 and 1990 presented in Table 4 by race and Hispanic origin may differ from those published previously.2 The percent of women who began prenatal care in their first trimester of pregnancy edged upward for 2003, increasing to 84.1% from 83.7% in 2002 (Table 4). Timely initiation of prenatal care has been on a slow but steady rise since 1990 (75.6%).4 The proportion of women beginning care in the last trimester of pregnancy, or with no care at all, was 3.5% in 2003 compared with 3.6% in the previous year. The proportion of women with late or no care has dropped from 6.1% since 1990. Recent gains in prenatal care utilization may be attributed in part to the expansion of Medicaid for pregnant women in the late 1980s.29 Appropriate prenatal care can improve pregnancy outcome by performing risk assessment and offering health care advice, psychosocial support, education on childbirth, and care coordination.30,31
Small increases in first-trimester care from 2002 to 2003 were found for non-Hispanic white (89% for 2003), non-Hispanic black (76.0%), and Hispanic (77.4%) women. Although fairly wide disparities in prenatal care utilization persist by race and Hispanic origin, considerable progress in timely receipt of care has been observed especially among groups with less advantageous levels; since 1990 the proportions of non-Hispanic black and Hispanic women with first-trimester care has risen 25% to 29%.2,4
Cesarean Delivery
Preliminary data for 2003 indicate a 6% jump in the rate of cesarean delivery to 27.6% of all births for another record US high (see Table 4 and Fig 2). After declines for 1989–1996, the cesarean delivery rate has risen each year; the current level is one third higher than that for 1996 (20.7%).4,2 Consistent with the recent trend, the primary cesarean rate (percent of cesareans among women with no previous cesarean delivery) rose 6% from 2002 to 2003, from 18.0% to 19.1%, and the rate of vaginal birth after previous cesarean (VBAC) fell 16%, from 12.6% to 10.6%. The primary rate has climbed 31%, and the VBAC rate has dropped 63% since only 1996.
The increase in the rate of primary cesareans may be related to changes in nonmedical risk factors such as maternal age, physician practice patterns, and maternal choice.32–34 Increased use of medical technologies such as induction of labor may also have contributed to the rise; induction before 41 weeks of gestation has been linked to an increased risk of cesarean delivery.33 The precipitous decline in the VBAC rate may be related to reports on the risks associated with VBACs, more conservative practice guidelines, and legal pressures.33,35–37
The primary cesarean rate has risen and the VBAC rate has fallen among low-risk women (ie, women with term, singleton deliveries with vertex presentations) at a pace similar to that for all women.38,39 Primary cesareans rose and VBAC rates declined among all 5 racial/ethnic groups studied from 2002–2003; rates ranged from 24.2% among American Indian women to 29.3% for non-Hispanic black women.4
Multiple Births
The birth rate for twins increased another 3% in 2002, rising to 31.1 twin births per 1000 total births (Table 4) (data for 2002 are the latest available). The twin birth rate has climbed 38% since 1990 (from 22.6 per 1000) and 65% since 1980 (18.9 per 1000).2 The birth rate for triplets and other higher-order multiples (triplet/+) was down slightly for 2002, however, to 184.0 per 100000 live births, from 185.6 in 2001 (Fig 3). Between 1980 and 1998, the triplet/+ birth rate soared from 37.0 to 193.5 per 100000, with an average annual increase of 13% for 1990–1998. This rate has trended slightly downward since, however, suggesting that the remarkable surge in triplet/+ births during the last 2 decades may have subsided. The 2002 rate is 5% lower than the 1998 peak (193.5).2 The 125134 twins, 6898 triplets, 434 quadruplets, and 69 quintuplet and higher-order multiple births accounted for 3.3% of all 2002 births.2
The rise in multiple births has been especially steep among births to women in the oldest childbearing ages; for example, the number of twin births to women 45 to 49 years old jumped from only 39 to 991 per 1000 total births between 1990 and 2002; the twin birth rate climbed from 23.8 to 189.7 per 1000.2
The increase in multiple births, especially higher-order multiples, has been associated with 2 related trends: older age at childbearing and increased use of fertility therapies (assisted reproductive technologies [ARTs] such as in vitro fertilization and non-ART therapies such as ovulation-inducing drugs).40,41 A recent study found that 43% of triplets resulted from ART, 40% were likely the result of non-ART therapies, and 18% were conceived naturally.42
Multiple births tend to be high-risk events. The average twin is delivered 3 weeks and the average triplet >6 weeks earlier than the typical singleton. Accordingly, the risk of early death is substantially elevated; IMRs for twins are 5 times and triplets 12 times those of singletons.2,9 Recognizing their increased risk of poor outcome, the American College of Obstetricians and Gynecologists and the American Society of Reproductive Medicine issued recommendations in 1999 intended to prevent triplet/+ pregnancies.43,44 These recommendations, plus refinements to ARTs, may be contributing to the recent small downturn in the higher-order multiple-birth trend.
Preterm Births
Approximately 500000 infants or 12.3% of all births were delivered preterm in 2003. The preterm birth rate (percent of infants delivered at <37 completed weeks of gestation) increased from 12.0% to 12.3% between 2002 and 2003 and has risen 16% since 1990 (from 10.6%) (Table 4).2,4 Three factors potentially related to this rise are higher rates of multiple births, more obstetric intervention, and greater use of ARTs45
Between 1990 and 2002, the preterm rate for singletons rose 7% (compared with a 14% rise for all births); all of the increase for singletons was among those born moderately preterm or between 32 and 36 weeks of gestation; the proportion of singletons delivered at <32 weeks declined slightly over this time period (from 1.69 to 1.57%).
Although outcomes for infants delivered before term have improved notably in recent years, preterm newborns remain at considerably heightened risk of death or lifelong disability. The IMR for infants delivered at <32 weeks of gestation is nearly 75 times greater than that of term infants, and the rate for those born at 32 to 36 weeks of gestation is >3 times that of those delivered at longer gestations.9
The overall rise in preterm rates is largely driven by the increase among non-Hispanic white births, up by one third between 1990 and 2003 (11.3% for 2003). The preterm rate has increased 8% among Hispanic births (11.9%) and is down by 6% for non-Hispanic black births (17.8%) (Table 4). The etiologies of preterm delivery are not fully understood, and until progress is made in this regard, substantial reduction in the preterm birth rate is unlikely.45,46
LBW
LBW (<2500 g) births rose to 7.9% in 2003 from 7.8% in 2002 (Table 4) and has increased 18% from the 1984 low (6.7%); the current level is the highest reported since 1970.2,4 Very low birth weight (VLBW; births weighing <1500 g) births were essentially unchanged at 1.4% in 2003. VLBW has risen from 1.15% in 1980 but has been essentially stable since 1998. LBW and especially VLBW are major predictors of infant morbidity and mortality. The risk of early death for infants born at moderately LBW (1500–2499 g) is 5 times higher than that of heavier infants; the risk for VLBW infants is >100 times that of infants born at 2500 g.9
Between 2002 and 2003 LBW rates rose or were unchanged among each of the major racial/ethnic groups (see Table 4). Since 1990, LBW levels have climbed 25% for non-Hispanic white births (to 7.0% in 2003), 10% among Hispanic births (6.7%), and 2% among non-Hispanic black births (13.5%). Some of this increase, particularly among non-Hispanic white births, can be attributed to increases in the rate of multiple births, which tend to be born much smaller than singletons. In 2002, almost 1 of every 4 LBW infants was born in a multiple delivery.2 Among singleton births only, the increase in LBW has been more modest. Between 1990 and 2002 the LBW rate rose from 5.90% to 6.12%; the VLBW level rose from 1.05 to 1.11%.2
INFANT MORTALITY
The IMR rose in 2002 to 7.0 deaths per 1000 live births from 6.8 per 1000 in 2001. The historic trend in the US IMR has been one of steady decline; 2002 marks the first increase in this rate in >40 years3,9,47,48 (see Table 5 and Fig 4 for the 1990–2002 trend). The NMR rose from 4.5 to 4.7 per 1000 for 2001–2002, whereas the PNMR was unchanged at 2.3 per 1000 (Table 5). Provisional data suggest that the IMR will not increase again in 2003.5
Increases in infant mortality between 2001 and 2002 were distributed fairly widely across age groups, racial/ethnic groups, and geographic areas.49 A recently released study based on data from the 2002 linked birth/infant death file found that the 2001–2002 increase in infant mortality was driven primarily by a shift in the birth weight distribution toward infants weighing <750 g and particularly toward those weighing <500 g.49 The change in the birth weight distribution accounted for essentially all (98%) of the increase in the 2002 IMR. Multiple births contributed disproportionately (25%) to the overall increase, but 75% of the rise was due to increases among singletons.49 The rise in <750-g births was predominately among women 20 to 34 years old and was not attributed to increases in high-risk births to older mothers. The reasons for the rise in the numbers of these extremely small infants are not clear but may be related to changes in the reporting of vital events, changes in the risk profiles of mothers giving birth, changes in the management of labor and delivery, and the increased use of ARTs.49
Small increases in IMRs were reported for most racial and ethnic groups between 2001 and 2002, but the changes were not statistically significant.9,49 The relative difference in rates between non-Hispanic black (13.9 per 1000) and non-Hispanic white (5.8 per 1000) expressed as a ratio of black/white IMRs was 2.4 in 2002, unchanged for more than a decade (Table 5). The 2002 IMR of 5.6 per 1000 for Hispanic births was slightly lower than that for non-Hispanic white births. Rates among the Hispanic subgroups differ widely, however, from a low of 3.7 for Cuban to a high of 8.2 among births to Puerto Rican infants.
Birth Weight–Specific Infant Mortality
2002 also marks a change in the long-term trend in declining birth weight–specific mortality rates for VLBW infants. For 2001–2002 the birth weight–specific IMR for VLBW (<1500 g) infants increased significantly from 244.4 to 250.0 infant deaths per 1000 live births (Table 6). 9,49 This marks the first increase in mortality for VLBW infants since statistics on this measure became available.49 Although the increase in birth weight–specific mortality for very small infants is cause for concern, as noted above, it was the shift in the birth weight distribution that explained the bulk of the rise in overall infant mortality for 2001–2002.
The 2002 IMR for infants with birth weights of 2500 g was unchanged from the previous year at 2.4 per 1000, whereas the rate for moderately LBW (1500–2499 g) infants was 15.2 per 1000 in 2001, compared with 15.1 per 1000 in 2002 (Table 6).
Between 1985 and 2001, the IMR declined by more than one third (from 10.6 to 6.8 per 1000),3 but the rate of LBW rose 13% (from 6.8% to 7.7%) and the VLBW rate rose from 1.21% to 1.44%.2,9 Thus, all of the decline in the IMR since the mid-1980s was due to declines in birth weight–specific IMRs, declines that occurred among both smaller and heavier infants. These changes were attributed primarily to improvements in obstetric and neonatal care.50
Perinatal and Fetal Mortality
The PMR more fully describes the risk of death at late stages of pregnancy and soon after birth, because it includes both late fetal (28 weeks of gestation) and early neonatal (<7 days) deaths. The PMR has been declining fairly consistently for more than half a century.51 For 2001–2002, a sizable decline in late fetal deaths slightly outpaced the increase in early neonatal deaths, and as a result, the PMR was unchanged at 6.9 deaths per 1000 live births plus fetal deaths.
The FMR (fetal deaths of at least 20 weeks of gestation) was 6.4 per 1000 live births plus fetal deaths for 2002 compared with 6.5 per 1000 for 2001. The FMR has declined substantially over the last 50 years and is down 15% since 1990.52 All of the recent decline in the FMR, however, has been among late fetal deaths; the late FMR declined 23% between 1990 and 2001. In contrast, the early FMR (fetal deaths of 20–27 weeks) has not improved over this period.52 For 2001–2002, the late FMR declined another 3% from 3.3 to 3.2 per 1000; the early FMR was unchanged.
Geographic Variation
Three states, Massachusetts, Maine, and Vermont, reported IMRs <5.0 per 1000 for 2002. In contrast, Louisiana and Mississippi reported rates >10.0. Whereas IMRs increased for more than half of the states between 2001 and 2002, statistically significant increases in rates were reported in 4 states: Kentucky, Missouri, New Jersey, and Texas.
Variations by state in LBW and IMR reflect compositional differences by race, ethnicity, and socioeconomic status in the population in addition to other factors (prenatal, quality-of-care, and postnatal influences on infants) that are associated with LBW or IMR.
Leading Causes of Infant Death
More than half of all infant deaths in 2002 were attributable to 5 leading causes: congenital malformations (20%), disorders relating to short gestation and unspecified LBW (17%), sudden infant death syndrome (SIDS) (8%), newborns affected by maternal complications of pregnancy (6%), and newborns affected by complications of placenta, cord, and membranes (4%) (Table 8). 3 The order of the first 5 leading causes of death was unchanged for 2001–2002.
Cause-specific IMRs increased significantly between 2001 and 2002 for LBW and maternal complications; it is important to note, however, that part of the increase in the latter cause of death is a result of a change in coding rules.3,9 Rates were also higher for congenital malformations and SIDS, but these changes were not statistically significant. For 2000–2001, the death rate for SIDS dropped 11%.53
The rank order of leading causes of death can vary by race or Hispanic origin. For example, in 2002 as in earlier years, LBW (not congenital malformations) was the leading cause of death among non-Hispanic black and Puerto Rican infants.
INTERNATIONAL COMPARISONS
For all 3 years, the IMR is lower for all countries in the table than the US rate with 1 exception; the IMR in the US was lower than the rate in Cuba in 2000. There has been a downward trend in the IMR in less than half of the countries, suggesting a possible trend toward stabilization in the rates in recent years. Nevertheless, the US IMR in 2001 was at least twice as great as the rate for the 6 countries with the lowest levels.
One reason for the high IMR in the US is the high rate of VLBW births in the country relative to other developed countries. It is unlikely that VLBW births will decrease much even among singleton births, because current trends show no declines in recent years. There may be some reporting differences across countries related to distinguishing between live births and fetal deaths,54–57 but the magnitude of these differences is unlikely to be large enough to account for the large disparities between US rates and the rest of the developed world.
DEATHS
There were 2423000 deaths (provisional data) in the United States in 2003 (Table 1), 20387 less than the 2443387 deaths reported in 2002.3,5 The death rate for 2002 was 847.3 deaths per 100000 population, no different from the 2001 rate of 848.5 per 100000. 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.3 The age-adjusted death rate for 2002 was 845.3 deaths per 100000 US standard population. This rate was 1.1% lower than the 2001 age-adjusted death rate of 854.5 100000 and was a record low for the United States.
Expectation of Life
The estimated expectation of life at birth for a given year represents the average number of years that a cohort 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. In 2002, the expectation of life at birth reached a new record high of 77.3 years, an increase of 0.1 year from the previous year. Between 2001 and 2002, life expectancy increased by 0.2 years for black men (68.8) and by 0.1 year for white men (75.1), white women (80.3), and black women (75.6). All 4 groups attained record high levels.
Causes of Death
The 15 leading causes of death in 2002 accounted for >83% of all US deaths (Table 10). 3 Between 2001 and 2002, age-adjusted death rates declined significantly for 4 of the 15 leading causes: diseases of heart by 2.8%, malignant neoplasms (cancer) by 1.3%, cerebrovascular diseases (stroke) by 2.9%, and assault (homicide) by 14.1%.3 Age-adjusted death rates increased significantly for 7 of the 15 leading causes of death: accidents (unintentional injuries) by 3.4%, influenza and pneumonia by 2.7%, Alzheimer's disease by 5.8%, nephritis, nephrotic syndrome and nephrosis (kidney disease) by 1.4%, septicemia by 2.6%, intentional self-harm (suicide) by 1.9%, and essential (primary) hypertension and hypertensive renal disease (hypertension) by 2.9%. The apparently dramatic decline in the homicide rate in 2002 was primarily a result of the September 11, 2001, terrorist attacks, which added 2926 certified resident deaths to this category in 2001. Without the additional deaths resulting from the terrorist attacks in 2001, the homicide rate would have remained the same between 2001 and 2002.
Deaths Among Children
In 2002, 25820 children and adolescents between the ages of 1 and 19 years died in the United States (Table 11). 3 The death rate for this age group was 33.5 per 100000 population, which is approximately the same as the rate in 2001 (33.6 per 100000). From 2001 to 2002, the death rate for children aged 1 to 4 years decreased by 6%. The death rates for children aged 5 to 9, 10 to 14, and 15 to 19 years did not show any significant changes between 2001 and 2002.
For all children aged 1 to 19 years, the first and second leading causes of death in 2002 were accidents (unintentional injuries) and assault (homicide). Unintentional injuries accounted for 44.5% of all deaths, and homicide accounted for 10.3%. The rate for unintentional injuries increased by 2% from the rate in 2001 and homicide increased by 3%, but the changes were not statistically significant. The rate for suicide (the fourth leading cause of death) decreased by 8%, and the rate for chronic lower respiratory diseases increased by 33%. For each of 4 age groups, 1 to 4, 5 to 9, 10 to 14, and 15 to 19 years old, Table 11 provides detailed information on the 5 leading causes of death by rank in 2002 as well as the number of deaths, the percentage accounted for by each of the leading causes, and the cause-specific death rate for each for the years 2001 and 2002. Although some causes increased or decreased in specific age groups, the overall patterns of these death rates have not changed significantly since last year.
THE TRANSFORMATION OF VITAL STATISTICS
Vital statistics in the United States are in the midst of a sweeping transformation.1,10,11,58,59 The experience of reporting and explaining the rise in the IMR for 2001–2002 underscores the urgent need for change. The 2002 increase in infant mortality could not be reported until 200447 and it is not until 2005 that a thorough exploration of the changes that led to the rise will be published.49 Delays of this magnitude are currently largely unavoidable. The long-term result of the fundamental changes in data collection and processing will be vast improvements in the quality and timeliness of vital statistics data and improved flexibility of electronic systems to allow for more immediate response to emerging trends. In the short term, however, increased delays in data release (as evidenced by the delay in publication of this annual article) and the potential loss of national data for some very basic measures of public health are anticipated.
The hundred-plus-year-old vital statistics system of the United States is a state-based, decentralized system.60 These state systems, with only a few exceptions, are based on outmoded registration practices and systems. To rebuild these aging systems more effectively, the National Association of Public Health Statistics and Information Systems, NCHS, and the Social Security Administration are working together to develop national standards and guidelines for electronic vital statistics systems.10,59 The new systems will incorporate standardized data-collection instruments, improved methods for capturing data, immediate query of suspect data, query and edit guidelines, and detailed item definitions. These systems will use the revised 2003 US Standard Certificates of Birth and Death and Report of Fetal Death that introduce a wealth of new and enhanced data items. The new systems also will be designed to integrate with other health information systems and will be easily modifiable to capture new data items as needed and drop those that no longer are.
The reengineered system(s) will potentially transform vital statistics, resulting in key enhancements in data flow and, ultimately, the release of these data within months (not years) of the event. The new systems will automatically link birth and infant death records. This linkage will allow these data to be available simultaneously and will facilitate more timely and comprehensive analysis of, for example, changes in infant mortality. The new and modified data items on the 2003 revised certificates of birth, death, and fetal death will greatly expand our ability to describe and explain such trends. For example, with information from the new birth certificate, we will be able to directly estimate the impact of the use of infertility therapies on perinatal outcome. Information on whether a "trial of labor" was attempted before cesarean delivery will help us to better assess the impact of changes in management of labor and delivery. Finally, better-quality data resulting from more immediate querying of data at the source while the records and respondent are still available, plus the use of detailed and standardized specifications and definitions, will help, for example, to better describe maternal medical-risk profiles such as changes in the prevalence of hypertensive disorders on pregnancy outcome.
Full implementation of reengineered vital statistics systems across the US jurisdictions will take some years. At the close of 2004, 8 states had reengineered their birth systems, and 2 states had partially reengineered their deaths systems. Much has been accomplished, but much work is left to be done to ensure that vital statistics improves its key role in monitoring the nation's maternal and infant health.
ACKNOWLEDGMENTS
We thank Martha Munson, Stephanie Ventura, and Brady Hamilton for important contributions to the manuscript and Sharon Kirmeyer, T.J. Mathews, Brady Hamilton, and Fay Menacker for content review.
FOOTNOTES
Accepted Dec 9, 2004.
No conflict of interest declared.
REFERENCES
Rothwell CL, Sondik EJ, Guyer B. A delay in publication of the "Annual Summary of Vital Statistics" and the need for new vital registration and statistics systems for the United States [commentary]. Pediatrics. 2004;114 :1671 –1672
Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Munson ML. Births: final data for 2002. Natl Vital Stat Rep. 2003;52(10) :1 –113
Kochanek KD, Murphy SL. Deaths: final data for 2002. Natl Vital Stat Rep. 2004;53(5) :1 –115
Hamilton BE, Martin JA, Sutton PD. Births: preliminary data for 2003. Natl Vital Stat Rep. 2004;53(9) :1 –17
Munson ML, Sutton PD. Births, marriages, divorces and deaths: provisional data for 2003. Natl Vital Stat Rep. 2004;52(22) :1 –7
World Health Organization. International Statistical Classification of Diseases and Related Health Problems, Tenth Revision. Geneva, Switzerland: World Health Organization; 1992
Anderson RN, Smith BL. Deaths: leading causes for 2001. Natl Vital Stat Rep. 2003;52(9) :1 –85
Mathews TJ, Menacker F, MacDorman MF. Infant mortality statistics from the 2002 period linked birth/infant death data set. Natl Vital Stat Rep. 2004;53(1) :1 –29
Office of Management and Budget. Race and ethnic standards for Federal statistics and administrative reporting. Statistical Policy Directive 15. 1977
Ingram DD, Parker JD, Schenker N, et al. United States Census 2000 with bridged race categories. Vital Health Stat 2. 2003;(135) :1 –55
Schenker N, Parker JD. From single-race reporting to multiple-race reporting: using imputation methods to bridge the transition. Stat Med. 2003;22 :1571 –1587
Hamilton BE, Sutton PD, Ventura SJ. Revised birth and fertility rates for the 1990s and new rates for Hispanic populations, 2000 and 2001: United States. Natl Vital Stat Rep. 2003;51(12) :1 –94
United Nations. United Nations Demographic Yearbook. New York, NY: United Nations; 2003
Ventura SJ, Mathews TJ, Hamilton BE. Births to teenagers in the United States, 1940–2000. Natl Vital Stat Rep. 2001;49(10) :1 –23
Menacker F, Martin JA, MacDorman MF, Ventura SJ. Births to 10–14 year-old mothers, 1990–2002: trends and health outcomes. Natl Vital Stat Rep. 2004;53(7) :1 –18
Ventura SJ, Abma JC, Mosher WD, Henshaw S. 2004. Estimated pregnancy rates for the United States, 1990–2000: an update. Natl Vital Stat Rep. 2004;52(23) :1 –9
Abma JC, Martinez GM, Mosher WD, Dawson BS. Teenagers in the United States: Sexual Activity, Contraceptive Use, and Childbearing. Vital and Health Statistics Series 23, No. 24. Hyattsville, MD: National Center for Health Statistics; 2004
Mosher WD, Martinez GM, Chandra A, Abma JC, Wilson SJ. Use of Contraception and Use of Family Planning Services in the United States: 1982–2002. Advance Data No. 350. Hyattsville, MD: National Center for Health Statistics; 2004
Ventura SJ. Trends and Variations in First Births to Older Women, 1970–86. Vital and Health Statistics Series 21, No. 47. Hyattsville, MD: National Center for Health Statistics; 1989
National Center for Health Statistics. Advance Report of Final Natality Statistics, 1980. Monthly Vital Statistics Report, Vol. 31, No. 8. Suppl. DHHS Pub.No. (PHS)83–1120. Hyattsville, MD: Public Health Service: 1982
Ventura SJ, Bachrach CA. Nonmarital childbearing in the United States, 1940–99. Natl Vital Stat Rep. 2000;48(16) :1 –40
Mathews TJ. Smoking during pregnancy in the 1990s. Natl Vital Stat Rep. 2001;49(7) :1 –14
Office of the Surgeon General, Public Health Service, US Department of Health and Human Services. Women and Smoking: A Report of the Surgeon General. Washington, DC: US Department of Health and Human Services; 2001
Ventura SJ, Hamilton BE, Mathews TJ, Chandra A. Trends and variations in smoking during pregnancy and low birth weight: evidence from the birth certificate, 1990–2000. Pediatrics. 2003;111 :1176 –1180
Howell EM. The impact of Medicaid expansions for pregnant women: a synthesis of the evidence. Med Care Res Rev. 2001;58 :3 –30
Fiscella K. Does prenatal care improve birth outcomes A critical review. Obstet Gynecol. 1995;85 :468 –479
US Public Health Service. Caring for Our Future: The Content of Prenatal Care. Washington, DC: US Department of Health and Human Services; 1989
Zinberg, S. Vaginal delivery after previous cesarean delivery: a continuing controversy. Clin Obstet Gynecol. 2001;44 :561 –570
American College of Obstetricians and Gynecologists. Evaluation of Cesarean Delivery. ACOG Guidelines. Washington, DC: American College of Obstetricians and Gynecologists; 2000
Minkoff H, Chervenak FA. Elective primary cesarean delivery. N Engl J Med. 2003;348 :946 –950
McMahon MJ, Luther ER, Bowes WA, Olshan AF. Comparison of a trial of labor with an elective second cesarean section. N Engl J Med. 1996;335 :689 –695
Lydon-Rochelle M, Holt VL, Easterling TR, Martin DP. Risk of uterine rupture during labor among women with a prior cesarean delivery. N Engl J Med. 2001;345 :3 –8
Guise JM, McDonagh M, Hashima J, et al. Vaginal Births After Cesarean (VBAC). Evidence Report/Technology Assessment No. 71 (Prepared by the Oregon Health and Science University Evidence-Based Practice Center under contract no. 209-97-0018). Agency for Healthcare Research and Quality Publication No. 03-E018. Rockville, MD: Agency for Healthcare Research and Quality; 2003
Declercq E, Menacker F, MacDorman MF. Rise in "no indicated risk" primary caesareans in the United States: cross sectional analysis. BMJ. doi:10.1136/bmj.38279.705336.OB
US Department of Health and Human Services. Tracking Healthy People 2010. B16-20. Washington, DC; US Government Printing Office; 2000
Martin JA, Park MM. Trends in twin and triplet births: 1980–97. Natl Vital Stat Rep. 1999;47(24) :1 –16
Reynolds MA, Schieve LA, Jeng G, Peterson HB, Wilcox LS. Risk of multiple birth associated with in vitro fertilization using donor eggs. Am J Epidemiol. 2001;154 :1043 –1050
Reynolds MA, Schieve LA, Martin JA, Jeng G, Macaluso M. Trends in multiple births conceived using assisted reproductive technology, United States, 1997–2000. Pediatrics. 2003;111 :1159 –1166
American College of Obstetricians and Gynecologists. Nonselective Embryo Reduction: Ethical Guidance for the Obstetrician-Gynecologist. ACOG Committee Opinion 215. Washington, DC: American College of Obstetricians and Gynecologists; 1999
American Society for Reproductive Medicine. Guidelines on Number of Embryos Transferred. A Practice Committee Report—A Committee Opinion. Birmingham, AL: American Society for Reproductive Medicine; 1999
Tucker J, McGuire W. Epidemiology of preterm birth. BMJ. 2004;329 :675 –678
Johnson RB, Williams MA, Hogue CJR, Mattison DR. Overview: new perspectives on the stubborn challenge of preterm birth. Paediatr Perinat Epidemiol. 2001;15(suppl 2) :3 –6
Kochanek KD, Smith BL. Deaths: preliminary data for 2002. Natl Vital Stat Rep. 2004;52(13) :1 –47
MacDorman MF, Martin JA, Mathews TJ, Hoyert DL, Ventura SJ. Explaining the 2001–2002 infant mortality increase: data from the linked birth/infant death data set. Natl Vital Stat Rep. 2005;53(12) :1 –22
Allen MC, Alexander GR, Tompkins ME, Hulsey TC. Racial differences in temporal changes in newborn viability and survival by gestational age. Paediatr Perinat Epidemiol. 2000;14 :152 –158
Fried VM, Prager K, MacKay AP, Zia H. Chartbook on Trends in the Health of Americans. Health, United States, 2003. Hyattsville, Maryland: National Center for Health Statistics; 2003
Barfield W, Martin JA, Hoyert DL. Racial/ethnic trends in fetal mortality—United States, 1990–2000. MMWR Morb Mortal Wkly Rep. 2004;53 :529 –532
Mathews TJ, Menacker F, MacDorman MF. Infant mortality statistics from the 2001 period linked birth/infant death data set. Natl Vital Stat Rep. 2003;52(2) :1 –28
Liu K, Moon M, Sulvetta M, Chawala J. International infant mortality rankings: a look behind the numbers. Health Care Financ Rev. 1992;13 :105 –118
Sepkowitz S. International rankings of infant mortality and the United States' vital statistics natality data collecting system—failure and success. Int J Epidemiol. 1995;24 :583 –588
Haub C, Yanagishita M. Infant mortality: who's number one Popul Today. 1991;19(3) :6 –8
Kramer MS, Platt RW, Yang H, Haglund B, Cnattingius S, Berjso P. Registration artifacts in international comparisons of infant mortality. Paediatr Perinat Epidemiol. 2002;16 :16 –22
Rothwell CJ. Reengineering vital registration and statistics systems for the united states. Prev Chronic Dis [serial online]. 2004;1(4) :1 –2
Hetzel AM. History and Organization of the Vital Statistics System. Hyattsville, MD: National Center for Health Statistics; 1997