Infant Mortality Trends and Differences Between American Indian/Alaska Native Infants and White Infants in the United States, 1989–1991 and 1998–2000

Abstract

Objectives. To describe changes in infant mortality rates, including birthweight-specific rates and rates by age at death and cause.

Methods. We analyzed US linked birth/infant-death data for 1989–1991 and 1998–2000 for American Indians/Alaska Native (AIAN) and White singleton infants at ≥20 weeks’ gestation born to US residents. We calculated birthweight-specific infant mortality rates (deaths in each birthweight category per 1000 live births in that category), and overall and cause-specific infant mortality rates (deaths per 100000 live births) in infancy (0–364 days) and in the neonatal (0–27 days) and postneonatal (28–364 days) periods.

Results. Birthweight-specific infant mortality rates declined among AIAN and White infants across all birthweight categories, but AIAN infants generally had higher birthweight-specific infant mortality rates. Infant mortality rates declined for both groups, yet in 1998–2000, AIAN infants were still 1.7 times more likely to die than White infants. Most of the disparity was because of elevated post-neonatal mortality, especially from sudden infant death syndrome, accidents, and pneumonia and influenza.

Conclusions. Although birthweight-specific infant mortality rates and infant mortality rates declined among both AIAN and White infants, disparities in infant mortality persist. Preventable causes of infant mortality identified in this analysis should be targeted to reduce excess deaths among AIAN communities.

The infant mortality rate (infant mortality rate), defined as deaths among infants less than 1 year of age per 1000 births, is generally accepted as one of the important indicators of a nation’s health and social well-being. The United States ranked 26th among developed countries in terms of infant mortality rate in 2000,1 and recent data suggest that the rate of decline in the national infant mortality rate has stabilized.2^,3 In addition, racial and ethnic disparities in infant mortality persist and have, in some instances, widened, and risk factors such as rates of low-birthweight have increased.1^,4^,5

Historically, American Indians/Alaska Natives have had high infant mortality rates, especially during the postneonatal period.6 Disease patterns and health disparities among this heterogeneous population have been associated with poverty, limited access to health care services, and cultural dislocation.7^–10 Much of the research on infant mortality among American Indians/Alaska Natives has focused on their greater risk of sudden infant death syndrome (SIDS);11^,12 little is known about other causes of death among (Ameri-can Indians/Alaska Native) AIAN infants.

To develop effective interventions to reduce infant mortality among AIAN communities, we must understand trends in the infant mortality rate among the most vulnerable, low-birthweight infants, whose survival may depend on timely access to quality obstetric and neonatal care,7^,8^,13^–15 and how the infant mortality rate varies by age at death and cause over time. To address these issues, we analyzed changes in birthweight-specific infant mortality rate and in infant mortality rate by cause and timing of death from 1989–1991 to 1998–2000 among AIAN and White singleton infants in the United States.

METHODS

Database Description

We analyzed birth and infant death certificate data for 1989–1991 and 1998–2000 using the Centers for Disease Control and Prevention’s National Center for Health Statistics (NCHS) national linked birth/infant death data file. The 2000 data were the most recent linked birth cohort data available. The linked birth cohort file contains a numerator file that consists of all deaths among infants born in a particular year whether the death occurred in that year or the following year. The denominator file consists of all births occurring in the particular year of interest.

Study Population

The study population consisted of all live-born singleton infants ≥20 weeks’ gestation born to women who were US residents. The race of the infant was defined by the mother’s reported race on the birth certificate. Records with missing maternal race data were deleted. In addition, records with implausible, inconsistent, or missing values for birthweight and/ or gestational age were deleted according to procedures outlined elsewhere.16 Using these criteria, 6.9% of all AIAN births and 6.9% of White births were excluded from the analysis.

Analysis Plan

For each time period, we calculated birthweight-specific infant mortality rates by dividing the number of infants (aged 0–364 days) within a specific birthweight category who were born in that time period and who died by the total number of infants within that category who were born in that time period. Birthweight categories were defined in grams as follows: high birthweight, 4500–7999 g; normal birthweight, 2500–4499 g; moderately low birthweight, 1500–2499 g; very low birthweight, 1000–1499 g; and extremely low birthweight, 500–999 g. We directly standardized the birthweight-specific infant mortality rate using the 1998–2000 gestational age distribution and stratified by maternal race. Infants with birthweights <500 g or >7999 g were not included in our analysis of birthweight-specific infant mortality rate. Birthweight-specific infant mortality rates were not reported if they were on the basis of fewer than 20 deaths in the numerator.

We compared cause-specific infant mortality rates, defined as infant deaths per 100000 live births, by leading causes and age at death between AIAN and White infants. We derived cause-of-death data from the underlying cause of death on the death certificate and as defined by the International Classification of Diseases, Ninth and Tenth Revisions (ICD-9 and ICD-10) groupings.17^,18 In 1999, the ICD-10 groupings replaced the ICD-9 groupings in the national linked birth/ infant death data file. We followed NCHS’s guidance on how to compare ICD-9 to ICD-10 groups,19 and we ranked the leading causes of death using the most recent ranking list for both time periods. To calculate overall and cause-specific neonatal mortality rates and postneonatal mortality rates, we defined the neonatal period as the time from birth to 27 completed days and the postneonatal period as 28–364 days after birth. Cause-specific neonatal mortality rates were calculated by dividing the number of infants born in a particular year who died from a particular cause in the neonatal period by the live births for that year. Cause-specific postneonatal mortality rates were calculated by dividing infants born in a particular year who died from a particular cause in the postneonatal period by the live births in that year minus those who died in the neonatal period. Cause-specific infant mortality rates were not reported if they were on the basis of fewer than 20 deaths.

All statistical analyses were performed using SAS version 8.0 software (SAS Institute Inc., Cary, NC).20 We constructed confidence intervals for all rates and performed the z-test using NCHS methods.21 We compared confidence intervals and z-scores to assess if there was a significant change in rates from 1989–1991 to 1998–2000 for each race and to determine if there was a significant difference in rates between White and AIAN infants in 1998–2000.

RESULTS

Birthweight-Specific Infant Mortality Rates

In 1989–1991, 11804678 singleton infants ≥20 weeks’ gestation were born to US resident mothers; in 1998–2000, there were 11414 331 such births. One percent (0.9% in 1989–1991 and 1.0% in 1998–2000) of these infants were American Indians/Alaska Natives. Birthweight-specific infant mortality rates among AIAN infants were higher across all birthweight categories and in both time periods than among White infants (Table 1 ▶). However, in 1998–2000, the AIAN and White birthweight-specific infant mortality rate in the 2 lowest birthweight categories were not statistically different. Even so, Amer-ican Indians/Alaska Natives had a significant decline in birthweight-specific infant mortality rate in all birthweight categories during the study period. The rate did not decline as much among normal-birthweight AIAN infants as among normal-birthweight White infants, however, leading to a slight widening of the disparity in this category.

TABLE 1—

Birthweight-Specific Infant Deaths and Mortality Rates Among Singleton Infants at ‡20 Weeks’ Gestation, by Race: United States, 1989–1991 and 1998–2000

		White			AIAN
Perioda	Birthweight range, gb	No.	Rate	% Change in Ratec	No.	Rate	% Change in Ratec
1989–1991
	500–999	11 625	443.4		191	515.0
	1000–1499	4202	113.0		79	160.2
	1500–2499	9447	25.6		199	37.3
	2500–4499	27 923	3.2		780	6.3
	4500–7999	461	2.4		17	. . .
1998–2000
	500–999	9331	320.8	–28d	169	343.6	–33d
	1000–1499	2606	67.6	–40d	54	90.7	–43d
	1500–2499	6901	18.4	–28d	126	22.4e	–40d
	2500–4499	18 707	2.2	–31d	494	4.6e	–27d
	4500–7999	280	1.8	–25d	27	9.2e

Note.AIAN = American Indians/Alaska Natives; . . . = fewer than 20 deaths in the numerator and estimate is considered unreliable.

^aData are from the Centers for Disease Control and Prevention, National Center for Health Statistics linked birth/infant death file.

^bBirthweight-specific infant mortality rates per 1000 live births were adjusted by gestational age using the 1998–2000 distribution for total births.

^cPercentage change in rate from 1989–1991 to 1998–2000, by race.

^dSignificant change in rates from 1989–1991 to 1998–2000 from z-test using NCHS methods.

^eSignificant difference between AIAN and White infants in 1998–2000 rates from z-test using NCHS methods.

Infant Mortality Rates by Cause of Death

In 1989–1991 and 1998–2000, the infant mortality rate was greater among AIAN than White infants (Table 2 ▶). Both groups of infants had a significant decline in the overall infant mortality rate and rates of SIDS and congenital anomalies. White infants also had a significant decline in infant mortality rate for respiratory distress syndrome. AIAN infants had a decline in infant mortality rate for accidents and an increase in disorders relating to short gestation and unspecified low birthweight, but these changes were not statistically significant. In 1998–2000, AIAN infants had significantly higher rates than White infants for the 4 leading causes of infant mortality.

TABLE 2—

Infant Deaths and Mortality Rates Among Singleton Infants at ‡20 Weeks’ Gestation for the 5 Leading Causes of Infant Death, by Race: United States, 1989–1991 and 1998–2000

	1989–1991						1998–2000
	White			AIAN			White			AIAN
Cause of Deatha	Rank	No.	Rateb	Rank	No.	Rateb	Rank	No.	Rateb	Rank	No.	Rateb
All causes		58 698	627.6		1329	1183.2		43 006	476.6c		932	797.4c,d
Congenital anomalies	1	16 217	173.4	2	272	242.2	1	11 650	129.1c	1	180	154.0c,d
Sudden infant death syndrome	2	10 100	108.0	1	350	311.6	3	4748	52.6c	2	157	134.3c,d
Disorders relating to short gestation	3	4119	44.0	5	56	49.9	2	4210	46.7	3	88	75.3d
and unspecified low birthweight
Respiratory distress syndrome	4	3951	42.2	4	64	57.0	5	1438	15.9c
Diseases of circulatory system	5	1631	17.4
Accidents (unintentional injury)				3	71	63.2	4	1568	17.4	4	58	49.6d
Influenza and pneumonia										5	28	24.0

Note. AIAN = American Indian/Alaska Natives. There were 112 321 American Indian and Alaska Native and 9 352 745 White live births in 1989–1991 and 116 879 and 9 023 102 in 1998–2000, respectively.

^aClassification of cause of death for 1989–1991 and 1998 was based on the International Classification of Diseases (ICD), Ninth Revision.17 Classification of cause of death was based on the ICD, Tenth Revision.18

^bMortality rates per 100 000 live births calculated using data from the Centers for Disease Control and Prevention, National Center for Health Statistics linked birth/infant death data file.

^cSignificant change in rates from 1989–1991 to 1998–2000 for each race from z-test using NCHS methods.

^dSignificant difference between AIAN and White infants in 1998–2000 rates from z-test using NCHS methods.

The rankings by cause for the 2 groups differed somewhat in both time periods (Table 2 ▶). In 1989–1991, the 2 leading causes of death among AIAN infants were SIDS and congenital anomalies; this ranking was reversed for White infants. These 2 causes of death accounted for 47% and 45% of all deaths among AIAN and White infants, respectively, from 1989–1991. From 1998–2000, AIAN and White infants had the same 4 leading causes of infant mortality: congenital anomalies, SIDS, disorders relating to short gestation and unspecified low birth-weight, and accidents, respectively. However the second and third leading causes were reversed for White infants.

Neonatal Mortality Rates by Cause of Death

The neonatal mortality rate was greater among AIAN than White infants in both time periods, and both groups of infants had a significant decline in the neonatal mortality rate (Table 3 ▶). The proportion of all infant deaths that occurred during the neonatal period increased slightly for both groups from the first to second time period (44%–49% for AIAN and 59%–62% for White infants). Among leading causes of death, cause-specific neonatal mortality rates were greater among AIAN than White infants, but the difference did not reach statistical significance. For those conditions with sufficient numbers of deaths to allow reporting of cause-specific neonatal mortality rates in both time periods, both groups had a significant decline in the neonatal mortality rate for congenital anomalies. In addition, White infants had a significant decline in respiratory distress syndrome.

TABLE 3—

Neonatal Deaths and Mortality Rates Among Singleton Infants at ‡20 Weeks’ Gestation for the 5 Leading Causes of Infant Death, by Race: United States, 1989–1991 and 1998–2000

	1989–1991						1998–2000
	White			AIAN			White			AIAN
Cause of Deatha	Rank	No.	Rateb	Rank	No.	Rateb	Rank	No.	Rateb	Rank	No.	Rateb
All causes		34 345	367.2		586	521.7		26 681	295.7c		456	390.1c,d
Congenital anomalies	1	11 830	126.5	1	189	168.3	1	8445	93.6c	1	131	112.1c
Disorders relating to short gestation and unspecified low birth weight	2	4066	43.5	3	56	49.9	2	4127	45.7	2	85	72.7
Respiratory distress syndrome	3	3699	39.5	2	57	50.7	3	1338	14.8c	5	. . .	. . .
Newborn affected by complications of placenta/cord/membrane	4	1301	13.9	5	25	22.3	5	1241	13.8	3	22	18.8
Bacterial sepsis of newborn	5	1233	13.2	4	26	23.1
Newborn affected by maternal							4	1292	14.3	4	. . .	. . .
complications of pregnancy

Note. AIAN = American Indians/Alaska Natives; . . . = fewer than 20 deaths in the numerator and estimates are considered unreliable. There were 112 321 AIAN and 9 352 745 White live births in 1989–1991 and 116 879 AIAN and 9 023 102 White live births in 1998–2000.

^aFigures for 1989–1991 and 1998 were based on the International Classification of Diseases, Ninth Revision.17 Figures for 1999–2000 based on the International Classification of Diseases, Tenth Revision.18

^bMortality rates per 100 000 live births calculated using data from the Centers for Disease Control and Prevention, National Center for Health Statistics linked birth/infant death data file.

^cSignificant change in rates from 1989–1991 to 1998–2000 for each race from z-test using NCHS methods.

^dSignificant difference between AIAN and White infants in 1998–2000 rates from z-test using NCHS methods.

The 5 leading causes of neonatal deaths in 1989–1991 were the same for White and for AIAN infants, except that the second and third causes and the fourth and fifth causes were reversed (Table 3 ▶). By 1998–2000, bacterial sepsis of the newborn (fourth among AIAN infants and fifth among White infants in 1989–1991) was no longer in the 5 leading causes of neonatal death, and AIAN infants had only 3 ranked causes of neonatal death because of the small numbers of neonatal deaths. The 2 leading causes of neonatal mortality for AIAN and White infants in 1998–2000 were congenital anomalies and disorders relating to short gestation and unspecified low-birthweight, which together accounted for 47% and 47% of all neonatal deaths, respectively.

Postneonatal Mortality Rates and Cause of Death

The postneonatal mortality rate was greater among AIAN than White infants during both time periods of interest (Table 4 ▶). The postneonatal mortality rate significantly declined among both groups, but this rate declined more among AIAN infants (39% and 30%). Both groups of infants had significant declines in postneonatal mortality rate for SIDS and congenital anomalies. AIAN infants had a greater decline in the postneonatal mortality rate for SIDS (58% vs 52%), congenital anomalies (43% vs 24%), and accidents (22% vs 3%). Even with the greater decline for AIAN infants, 1998–2000 rates for postneonatal mortality were 2.6 times as high among AIAN as among White infants for SIDS, 3.0 times as high for accidents, and 4.0 times as high for pneumonia and influenza (data for White infants not shown).

TABLE 4—

Postneonatal Deaths and Mortality Rates Among Singleton Infants at ‡20 Weeks’ Gestation for the 5 Leading Causes of Infant Death, by Race: United States, 1989–1991 and 1998–2000

	1989–1991						1998–2000
	White			AIAN			White			AIAN
Cause of Deatha	Rank	No.	Rateb	Rank	No.	Rateb	Rank	No.	Rateb	Rank	No.	Rateb
All causes		24 353	261.3		743	665.0		16 325	181.9c		476	408.9c,d
Sudden infant death syndrome	1	9 462	101.5	1	330	295.3	1	4 384	48.8c	1	146	125.4c,d
Congenital anomalies	2	4 387	47.1	2	83	74.3	2	3 205	35.7c	3	49	42.1c
Accidents (unintentional injury)	3	1 488	16.0	3	66	59.1	3	1 393	15.5	2	54	46.4d
Diseases of circulatory system	4	1 044	11.2	5	33	29.5	4	783	8.7c
Influenza and pneumonia	5	825	8.9	4	41	36.7				4	25	21.5
Homicide							5	478	5.3	5	. . .	. . .

Note. AIAN = American Indians/Alaska Natives; . . . = fewer than 20 deaths in the numerator and estimates are considered unreliable.

^aFigures for 1989–1991 and 1998 were based on the International Classification of Diseases, Ninth Revision; figures for 1999–2000 were based on the International Classification of Diseases, Tenth Revision.

^bMortality rates per 100 000 live births were calculated using data from the Centers for Disease Control and Prevention, National Center for Health Statistics linked birth/infant death data file. The denominator used to calculate rates was 111 735 for AIAN and 9 318 400 for White infants in 1989–1991 and 116 423 for AIAN and 8 976 421 for White infants in 1998–2000.

^cSignificant change in rates from 1989–1991 to 1998–2000 for each race from z-test using NCHS methods.

^dSignificant difference between AIAN and White infants in 1998–2000 rates from z-test using NCHS methods.

In 1989–1991, the 5 leading causes of postneonatal mortality were identical for the 2 population groups, except that the fourth and fifth causes were reversed (Table 4 ▶). The 2 leading causes, SIDS and congenital anomalies, accounted for 56% and 57% of all postneonatal deaths for AIAN and White infants, respectively. In 1998–2000, the ranking did not change for White infants, except that homicide became the fifth leading cause of death. Among AIAN infants, accidents surpassed congenital anomalies to become the second leading cause, and diseases of the circulatory system was no longer in the 5 leading causes of postneonatal mortality.

DISCUSSION

Racial and ethnic disparities in infant mortality have been a long-standing public health issue in the United States.5 In addition, reducing such disparities has been identified as a national goal, and interventions targeting American Indians/Alaska Natives have been recently introduced.22 In our study, we found that, although birthweight-specific infant mortality rates declined among AIAN and White infants and across all birthweight categories, AIAN infants still had higher birthweight-specific infant mortality rates than White infants in the second time period. Similarly, the infant mortality rate significantly declined among both infant groups over the study period, and yet, in 1998–2000, AIAN infants were still 1.7 times more likely than White infants to die before their first birthday, and AIAN infants had significantly higher cause-specific infant mortality rates for the 4 leading causes of death than White infants. These findings suggest that racial disparities in infant mortality persist.

Despite the racial disparity in birthweight-specific infant mortality rates, AIAN infants had the greater declines in rates up through 1500–2499 g. Birthweight-specific infant mortality rates, particularly among very low-birthweight infants, are thought to be influenced by access to quality obstetric and neonatal care and are often used as a measure of disparity in access to quality care.13^–15^,23 The decline in birthweight-specific infant mortality rates among very low-birthweight AIAN infants may reflect an improvement over the past decade in the quality of health services, greater access to health care, or more appropriate referral of high-risk pregnancies to facilities with the capacity to care for these small, often premature, infants. Obtaining access to quality health care has been a major public health concern among AIAN communities, and differences in access have been reported by residence (urban vs rural) and by geographic location.7^–8^,24 The remoteness of many AIAN communities and difficulties in recruiting and retaining professional staff to serve in isolated Indian Health Service areas are obstacles to providing quality delivery services.25 To assure that these declines in birth-weight-specific infant mortality rates among AIAN infants continue, current successful systems and strategies need to be identified and emulated. As birthweight-specific infant mortality rates among high-birthweight AIAN infants were 5 times as great as rates among White infants in 1998–2000, new culturally appropriate interventions are needed to assure that all high-risk AIAN pregnancies are identified early and delivered in appropriate health care settings.

The risk factors for infant mortality and its causes vary by age at death. Neonatal deaths are often associated with events surrounding pregnancy, delivery, and neonatal care following birth, whereas postneonatal deaths are more likely to be associated with conditions or events that arise following discharge to home, and reflect environmental factors and primary health care issues.26^,27 Because the interventions needed to address these causes are quite different, understanding of the timing and pattern of deaths within a population is essential to designing appropriate mortality-reduction strategies. We found that, unlike the success among low-birthweight infants, normal-birthweight AIAN infants experienced slightly less of a decline in birth-weight-specific infant mortality rate than did White infants, leading to an increase in health disparity. Most of the increase was because of elevated rates of postneonatal mortality, especially because of SIDS, accidents, and pneumonia and influenza. Accidents and pneumonia and influenza are potentially preventable causes of infant death, and the risk factors for SIDS are well defined. Disparities in preventable causes of infant death should compel us to improve the quality of primary preventive and curative health care and of the infant’s environment. For example, interventions to reduce parental smoking and environmental tobacco smoke may help reduce deaths because of SIDS and lower respiratory infections, as infant exposure to environmental tobacco smoke has been associated with increased rates of SIDS and hospitalizations for lower respiratory infections.28^–30 Although the disparities persist, AIAN infants had a significant decrease in rates for SIDS and congenital anomalies. These declines may be because of the success of local, state, and national campaigns promoting safe infant sleep environments and home safety, access to quality healthcare, and the use of folic acid supplements among women of reproductive age and pregnant women. Despite these declines, more work is needed to reduce racial disparities in infant mortality. Attention should be given to disparities, not only between AIAN infants and infants of other racial groups, but also among American Indians/Alaska Natives themselves. There are over 500 AIAN tribes in the United States, representing a variety of cultures, languages, and traditions. The heterogeneity of the AIAN population generates diverse types of risk behaviors, and infant mortality rates vary between AIAN communities.31 Although studies on intertribal disparities in infant mortality would be instructive, the practical limitations, such as the lack of adequate and reliable data and small number of deaths in the subpopula-tions, may prohibit such studies. Prevention efforts and interventions addressing known risk factors should be designed within the AIAN community using the most recent research findings and local knowledge of cultural beliefs and child-rearing practices.

Our study may be subject to a few limitations. First, there are relatively few AIAN births and infant deaths each year. When the number of events is small, the relative standard error may be large. Because of this, we performed statistical tests to determine the significance of the changes in the infant mortality rate over time and the difference in cause-specific rates between White and AIAN infants in 1998–2000. Second, the change in coding from ICD-9 to ICD-10 results in a few shifts in disease groupings. For example, there is nearly a 10% decrease in deaths attributable to congenital anomalies if ICD-10 is used rather than ICD-9, in large part because of a shift of hypoplasia and dysplasia of lung from congenital anomalies to primary atelactasis of newborn.19 Correspondingly, the decrease in congenital anomalies may appear greater than it truly is, but this finding would likely occur in all infant groups under study. Third, the quality of gestational age data on birth certificates in the United States has been studied extensively,32^–33 and studies have found that reporting of gestational age is oftentimes incomplete and inaccurate, especially when the infants are preterm. We attempted to address this issue by excluding cases where gestational age and birthweight were inconsistent, employing methods that have been widely used.16 Last, we should point out that the birthweight-specific infant mortality rates in this paper were standardized by gestational age using the 1998–2000 gestational age distribution for all births in the United States, and were not adjusted to control for differences in maternal characteristics and risk factors between AIAN and White infants.

We found a significant reduction in infant mortality rates, neonatal mortality rates, and postneonatal mortality rates among AIAN infants during the past decade. Although these achievements are commendable, we need to strengthen efforts to reduce racial disparities in infant mortality among American Indians/ Alaska Natives. Public health professionals should continue to monitor changes in infant mortality in the United States and strive to find additional risk factors for infant death that are preventable. Potentially preventable causes of infant mortality, such as accidents and pneumonia and influenza, should be addressed to reduce infant mortality. Organizations and agencies need to work with AIAN communities to design and implement integrated, culturally sensitive interventions to address the multiple risk factors for infant death and target those at greatest risk of death.