Trends in Perinatal and Infant Health Disparities Between Rural American Indians and Alaska Natives and Rural Whites

Abstract

Objectives. We examined disparities in perinatal care, birth outcomes, and infant health between rural American Indian and Alaska Native (AIAN) persons and rural Whites over time.

Methods. We compared perinatal and infant health measures for 217 064 rural AIAN births and 5 032 533 rural non-Hispanic White births.

Results. Among American Indians and Alaska Natives, unadjusted rates of inadequate prenatal care (1985–1987, 36.3%; 1995–1997, 26.3%) and postneonatal death (1985–1987, 7.1 per 1000; 1995–1997, 4.8 per 1000) improved significantly. However, disparities between American Indians and Alaska Natives and Whites in adjusted odds ratios (AORs) of postneonatal death (1985–1987, AOR = 1.55; 95% confidence interval [CI] = 1.41, 1.71; 1995–1997, AOR = 1.46; 95% CI = 1.31, 1.64) and adjusted risk ratios (ARRs) of inadequate prenatal care (1985–1987, ARR = 1.67; 95% CI = 1.65, 1.69; 1995–1997, ARR = 1.84; 95% CI = 1.81, 1.87) persisted.

Conclusions. Despite significant decreases in inadequate prenatal care and postneonatal death among American Indians and Alaska Natives, additional measures are needed to close persistent health gaps for this group.

Previous studies and reports published by the Indian Health Service have demonstrated dramatic improvements in perinatal and infant health among American Indian and Alaska Native (AIAN) populations over the past 50 years. Infant mortality rates declined substantially from 62.7 per 1000 live births in 1955 to 9.3 per 1000 live births in the years 1994 to 1996.¹ Yet disparities between American Indians and Alaska Natives and Whites have persisted. In 1989–1991, American Indians and Alaska Natives overall had 2.4 times the rate of postneonatal death compared with the White population¹; rural American Indians and Alaska Natives had a postneonatal death rate 2.6 times that of Whites.²

Since the mid-1980s, considerable attention has been paid to improving access to health care services, changing risk behaviors among pregnant women, and modifying provider practices, with the intention of improving birth outcomes and lowering infant mortality rates.^3–5 Among the general population, some of these efforts have been associated with higher rates of early and adequate prenatal care, as well as declining postneonatal mortality rates, especially from sudden infant death syndrome (SIDS).^6–8 However, it is not known how these efforts have influenced the perinatal health status of American Indians and Alaska Natives specifically, especially among rural American Indians and Alaska Natives, many of whom live in remote settings that may be more distant from health services.

Our goal was to determine whether the disparities in perinatal care, birth outcomes, and infant health among rural American Indians and Alaska Natives and rural Whites diminished, remained stable, or increased during a period of policy, funding, and practice changes in maternal and child health care from the mid-1980s through the 1990s. We addressed these questions by examining trends in prenatal care receipt, low-birthweight rates, neonatal and postneonatal death rates, and causes of death among rural American Indians and Alaska Natives and Whites between 1985 and 1997.

METHODS

Study Database

This study used the 1985–1987, 1989–1991, and 1995–1997 sets of the National Linked Birth and Infant Death Data. The 1995–1997 database was the latest available from the National Center for Health Statistics (NCHS) at the time our study began. These databases contain selected information compiled from birth certificates for all 50 states and the District of Columbia on all live births during these 3 time periods. We obtained identifiers for the mother's county of residence for each birth in the database, allowing classification of counties as rural or urban. Death certificate data were linked to these births if the infant died within 1 year of birth.

Study Population

The study population included singleton births of American Indians and Alaska Natives and non-Hispanic Whites to women who were US residents in rural counties during the 3 study time periods. AIAN births were those for which either the mother or father was reported as American Indian or Alaska Native on the birth certificate. This differs from the NCHS practice since 1989 of tabulating births primarily by the race of the mother.⁹ We included births with AIAN fathers regardless of the mothers' racial identity because non-AIAN women giving birth to children with AIAN fathers are likely to be eligible for Indian Health Service (IHS) services. Non-Hispanic White births were identified by the race and ethnicity of the mother only, after excluding those with AIAN fathers. In the 1985–1987 time period, 27 states did not record Hispanic ethnicity. Because these states had low rates of births to Hispanic mothers, all White births were included in the non-Hispanic White comparison group. We chose White births for comparison because this group has consistently demonstrated attainable and more-favorable perinatal health outcomes than have American Indians and Alaska Natives.

Study Variables

Outcome variables.

We created 2 measures of access to prenatal care: initiation of care in the first trimester and receipt of an inadequate pattern of prenatal care based on the adequacy of prenatal care utilization index developed by Milton Kotelchuck.¹⁰ Low birthweight was defined as less than 2500 g. We identified all infant deaths within 1 year of birth and categorized these into neonatal deaths (occurring at less than 28 days of age) and postneonatal deaths (occurring at 28 days of age through 1 year of age). Death rates are presented per 1000 live births. We used both individual and aggregated categories defined by the National Center for Health Statistics (based on the International Classification of Diseases, Ninth Revision) to present causes of death for the neonatal and postneonatal periods separately.^11,12

Independent variable.

AIAN or White race was the independent variable of interest. We created 3 AIAN race categories for some subanalyses: both parents AIAN, AIAN mother only, AIAN father only.

Maternal characteristics.

We defined a birth as rural if the mother's residence county on the birth certificate was nonmetropolitan according to the 1993 federal Office of Management and Budget's definition.¹³ We applied the 1993 Office of Management and Budget's definition to all of the study births to maintain a consistent rural definition across the study periods. Each rural county was further categorized as remote rural or nonremote rural. Designation as remote rural meant that the county was not adjacent to a metropolitan county and did not have a town with a population of 10 000 or more.

We described the following maternal characteristics for the births in all 3 time periods: age (< 18 years, 18–34 years, and ≥ 35 years), educational attainment ( < 12 years, 12 years, some college), marital status (married, unmarried), parity (0, 1–4, ≥ 5). Several other characteristics were available to describe maternal characteristics in the 1989–1991 and 1995–1997 cohorts only: cigarette use (none, < 11 cigarettes per day, ≥ 11 cigarettes per day), alcohol use (none, 1–4 drinks per week, ≥ 5 drinks per week), preexisting medical risk (one or more of the following: maternal cardiac disease, chronic hypertension, gestational or established diabetes), complications of labor or delivery (1 or more of the following: eclampsia, anemia, oligohydramnios, incompetent cervix, uterine bleeding, abruptio placentae, placenta previa, pregnancy-induced hypertension), and history of prior preterm birth or small-for-gestational-age infant.

Analyses

We first used the χ² test to compare maternal characteristics, receipt of prenatal care, low-birthweight rates, infant death rates, and cause of death between rural AIAN births and rural non-Hispanic White births nationally. Unadjusted and adjusted odds ratios were calculated to compare differences between births of rural American Indians and Alaska Natives and rural non-Hispanic Whites on all of these measures except cause of death. Using multiple logistic regression analysis, we calculated adjusted odds ratios after we controlled for maternal characteristics available in all 3 time periods' data, including remote rural residence status. We repeated these regression analyses in 1989–1991 and 1995–1997, and we controlled for the expanded number of variables (e.g., smoking, pregnancy complications). We found comparable results, and we have reported only the original regression results for all 3 time periods. We used published methods to convert odds ratios to risk ratios for those outcomes occurring in more than 10% of the population (inadequate prenatal care, first-trimester care).¹⁴ We used rate ratios to compare causes of death between rural American Indians and Alaska Natives and non-Hispanic Whites, and we used the χ² test to test for differences between these groups.

RESULTS

There were 217 064 rural AIAN singleton births and 4 981 936 rural non-Hispanic White singleton births during the 3 study time periods, or roughly 70 000 AIAN births and 1.6 million non-Hispanic White births in each period. Compared with mothers of White infants, mothers of AIAN infants were more likely to be younger than 18 years, to have completed less than 12 years of education, to be unmarried, to be multiparous, and to live in a remote rural county (Table 1). Over the study period, there was a decreasing proportion of births to families with both an AIAN father and an AIAN mother. There was a small increase in the proportion of both AIAN births and White births to women 35 years and older. There was also a small increase in the proportion of births to adolescents younger than 18 years, more so for American Indians and Alaska Natives than for Whites. There was a decrease in the proportion of births to unmarried mothers among both American Indians and Alaska Natives and Whites, and there was a decrease in the proportion of births to mothers living in remote rural counties among both American Indians and Alaska Natives and Whites.

TABLE 1.

Sociodemographic and Risk Characteristics of Singleton Rural American Indian and Alaska Native (AIAN) and Non-Hispanic White Births: National Linked Birth and Infant Death Data, 1985–1987, 1989–1991, and 1995–1997

	1985–1987a		1989–1991b		1995–1997c
Characteristic	AIAN	White	AIAN	White	AIAN	White
Sample, no.	70 012	1 796 428	75 752	1 633 309	71 300	1 552 199
Race of parents, %
Both parents AIAN	43.7	NA	41.3	NA	37.3	NA
AIAN mother only	42.8	NA	44.2	NA	44.4	NA
AIAN father only	13.5	NA	14.5	NA	18.2	NA
Mother's age, y, %
< 18	7.7***	4.4	7.3***	4.5	8.5***	4.9
18–34	86.4	90.2	85.9	88.8	83.6	86.3
≥ 35	5.9	5.4	6.8	6.7	7.9	8.8
Mother's education, %
< 12 years of school	37.9***	20.2	35.7***	20.7	31.8***	18.6
12 years of school	43.3	45.9	44.2	44.9	43.1	39.9
≥ 1 year of college	18.8	33.9	20.2	34.4	25.0	41.4
Married, %	54.0***	86.5	49.0***	81.6	43.8***	75.4
Parity, %
0	32.3***	40.5	30.9***	40.7	35.9***	42.0
1–4	62.3	58.2	63.3	57.9	58.9	56.6
≥ 5	5.4	1.3	5.8	1.4	5.2	1.5
Remote rural residence county, %	45.8***	26.7	44.7***	25.2	41.8***	24.7
Smoking, %
Nonsmoker	…	…	79.5***	77.7	79.2***	79.7
1–10 cigarettes/day	…	…	14.6	11.7	15.7	12.4
≥ 11 cigarettes/day	…	…	5.9	10.6	5.2	7.9
Drinking, %
Nondrinker	…	…	95.8***	98.3	97.9***	99.2
1–4 drinks/week	…	…	3.3	1.6	1.6	0.7
≥ 5 drinks/week	…	…	0.9	0.2	0.5	0.1
Preexisting medical conditions,d %	…	…	5.1***	3.1	5.7***	3.7
Complications of labor or delivery,e %	…	…	13.2***	8.5	13.6***	9.6
Prior preterm or small-for-gestational-age infant, %	…	…	2.3***	1.7	1.8***	1.6

Note. NA = not applicable. Ellipses indicate no data were available. Column percentages may not total 100% because of rounding. Asterisks indicate statistically significant differences in overall χ² tests between AIAN and White populations within each time period. Within the AIAN population, all but 1 of the variables—proportion with complications of labor or delivery—demonstrate statistically significant differences over time. Within the White population, all of the variables demonstrate statistically significant differences over time.

^aMissing values for 1985–1987: mother's education: AIAN 7.1%, White 8.4%; parity: AIAN 0.2%, White 0.3%.

^bMissing values for 1989–1991: mother's education: AIAN 4.3%, White 5.0%; parity: AIAN 0.2%, White 0.2%; smoking: AIAN 22.0%, White 14.5%; drinking: AIAN 22.1%, White 10.2%; preexisting medical conditions: AIAN 10.2%, White 3.1%; complications of labor or delivery: AIAN 11.3%, White 9.8%; prior preterm or small-for-gestational-age infant: AIAN 10.7%, White 6.5%.

^cMissing values for 1995–1997: mother's education: AIAN 1.3%, White 0.6%, parity: AIAN 0.3%, White 0.2%; smoking: AIAN 13.3%, White 10.9%; drinking: AIAN 12.6%, White 4.2%; preexisting medical conditions: AIAN 2.1%, White 1.2%; complications of labor or delivery: AIAN 3.2%, White 5.2%; prior preterm or small-for-gestational-age infant: AIAN 2.1%, White 1.2%.

^dConditions include maternal cardiac disease, chronic hypertension, and diabetes.

^eComplications include eclampsia, anemia, oligohydramnios, incompetent cervix, uterine bleeding, abruptio placentae, placenta previa, and pregnancy-induced hypertension.

^***P ≤ .001.

Maternal risk factor data were available only in the second and third time periods. Compared with mothers of White infants, mothers of AIAN infants were slightly more likely to smoke, but they smoked fewer cigarettes per day. Mothers of AIAN infants were more likely than were mothers of White infants to drink alcohol and to have had preexisting medical conditions, labor complications, and a history of preterm births. Over the study period, there was a small decrease in the smoking rate for Whites and in the number of cigarettes smoked among American Indians and Alaska Natives. Both American Indians and Alaska Natives and Whites decreased their drinking rates. There were increases in the proportion of both American Indians and Alaska Natives and Whites with preexisting medical conditions. Whites had an increase in their labor or delivery complication rates between the 2 later time periods. American Indians and Alaska Natives demonstrated a meaningful and significant decrease in rates of prior premature or small-for-gestational-age delivery.

Over the study period, prenatal care use increased substantially for both American Indians and Alaska Natives and Whites (Figure 1, Table 2). The greatest improvement in prenatal care use occurred between the 1989–1991 and 1995–1997 time periods, during which time the disparity in the crude rates of inadequate prenatal care and initiation of prenatal care in the first trimester narrowed between American Indians and Alaska Natives and Whites. However, compared with Whites, American Indians' and Alaska Natives' adjusted risk of inadequate prenatal care increased over the study period (risk ratio [RR] = 1.67–1.84; Table 2).

FIGURE 1.

Comparison of rates for rural American Indians and Alaska Natives and rural Whites for (a) inadequate pattern of prenatal care, (b) first-trimester prenatal care, (c) low-birthweight, (d) infant death, (e) neonatal death, and (f) postneonatal death: National Linked Birth and Infant Death Data, 1985–1987, 1989–1991, and 1995–1997.

Note. See Table 2 for exact percentages and confidence intervals (represented here by vertical lines). A number of confidence intervals are too narrow to be visible on this figure.

TABLE 2.

Prenatal Care Receipt, Birth Outcomes, and Infant Health of Singleton Rural American Indian/Alaska Native (AIAN) and White Births: National Linked Birth and Infant Death Data, 1985–1987, 1989–1991, and 1995–1997

	1985–1987a				1989–1991b				1995–1997c
	AIAN, % (95% CI)	White, % (95% CI)	AIAN-to-White Unadjusted OR (95% CI)	AIAN-to-White Adjusted OR (95% CI)	AIAN, % (95% CI)	White, % (95% CI)	AIAN-to-White Unadjusted OR (95% CI)	AIAN-to-White Adjusted OR (95% CI)	AIAN, % (95% CI)	White, % (95% CI)	AIAN-to-White Unadjusted OR (95% CI)	AIAN-to-White Adjusted OR (95% CI)
Received inadequate pattern of prenatal care	36.29*** (35.92, 36.66)	14.62 (14.57, 14.67)	2.48 (2.46, 2.51)	1.67 (1.65, 1.69)	33.86*** (33.52, 34.20)	13.47 (13.42, 13.52)	2.51 (2.49, 2.54)	1.75 (1.73, 1.78)	26.32*** (25.99, 26.65)	9.96 (9.91, 10.01)	2.64 (2.61, 2.68)	1.84 (1.81, 1.87)
Received first-trimester care	57.36*** (56.99, 57.73)	78.29 (78.23, 78.35)	0.733 (0.728, 0.738)	0.89 (0.88, 0.89)	58.69*** (58.34, 59.04)	78.80 (78.74, 78.86)	0.745 (0.740, 0.749)	0.88 (0.88, 0.89)	66.70*** (66.35, 67.05)	84.01 (83.95, 84.07)	0.794 (0.790, 0.798)	0.89 (0.89, 0.90)
Low birthweight (< 2500 g)	5.21*** (5.05, 5.38)	4.79 (4.76, 4.82)	1.09 (1.06, 1.13)	0.89 (0.86, 0.92)	5.16*** (5.00, 5.32)	4.87 (4.84, 4.90)	1.06 (1.03, 1.10)	0.89 (0.86, 0.92)	5.50 (5.33, 5.67)	5.38 (5.34, 5.42)	1.02 (0.99, 1.06)	0.87 (0.85, 0.90)
Mortality (rate/1000)
Neonatal (0–28 days)	5.47* (4.92, 6.02)	4.88 (4.78, 4.98)	1.12 (1.01, 1.24)	0.93 (0.84, 1.03)	5.02*** (4.52, 5.52)	4.19 (4.09, 4.29)	1.20 (1.08, 1.33)	1.04 (0.93, 1.15)	3.80 (3.35, 4.25)	3.70 (3.60, 3.80)	1.03 (0.91, 1.16)	0.85 (0.75, 0.97)
Postneonatal (29 days to 1 year)	7.13*** (6.51, 7.75)	3.27 (3.19, 3.35)	2.19 (2.00, 2.40)	1.55 (1.41, 1.71)	6.69*** (6.11, 7.27)	3.15 (3.06, 3.24)	2.14 (1.95, 2.34)	1.50 (1.36, 1.65)	4.80*** (4.29, 5.31)	2.40 (2.32, 2.48)	2.01 (1.79, 2.24)	1.46 (1.31, 1.64)
Infant death (first year total)	12.60*** (11.77, 13.43)	8.14 (8.01, 8.27)	1.55 (1.45, 1.66)	1.20 (1.12, 1.29)	11.71*** (10.94, 12.48)	7.34 (7.21, 7.47)	1.60 (1.50, 1.72)	1.26 (1.18, 1.35)	8.60*** (7.92, 9.28)	6.10 (5.98, 6.22)	1.41 (1.30, 1.53)	1.11 (1.02, 1.21)
Number of births	70 012	1 796 428	NA	NA	75 752	1 633 309	NA	NA	71 300	1 552 199	NA	NA

Note. OR = odds ratio; CI = confidence interval. NA = not applicable. Using published methods,¹⁴ ORs were converted to risk ratios for inadequate pattern of prenatal care and first-trimester care variables because these outcomes were common in the study population. Multiple logistic regression adjusted for mother's age, age², parity, education, marital status, residence in a remote rural county. For 1985–1987, the total number of rural AIAN and rural White births was n = 1 866 440; for 1989–1991, the total was n = 1 709 061; and for 1995–1997, the total was n = 1 623 499.

^aMissing values for 1985–1987: inadequate pattern of prenatal care: AIAN 7.7%, White 6.4%; first-trimester care: AIAN 2.5%, White 1.7%; low birthweight: AIAN 0.2%, White 0.1%.

^bMissing values for 1989–1991: inadequate pattern of prenatal care: AIAN 3.5%, White 2.7%; first-trimester care: AIAN 1.9%, White 1.1%; low birthweight: AIAN 0.2%, White 0.1%.

^cMissing values for 1995–1997: inadequate pattern of prenatal care: AIAN 3.7%, White 2.7%; first-trimester care: AIAN 2.1%, White 1.2%; low birthweight: AIAN 0.1%, White 0.03%. Asterisks indicate statistically significant differences between AIAN and White populations within each time period.

*P ≤ .05; ***P ≤ .001.

Further investigation demonstrated that this increased disparity in the adjusted risk of inadequate prenatal care related to differing trends in inadequate prenatal care among single and married American Indians and Alaska Natives and Whites. Among single women, the risk of inadequate prenatal care widened between American Indians and Alaska Natives and Whites; among married women, the inadequate prenatal care risk narrowed between the groups. Because American Indians and Alaska Natives are more likely than are Whites to be single in the later time period, the disparity in the adjusted risk of inadequate prenatal care between American Indians and Alaska Natives and Whites widened over time. The adjusted risk of initiation of first trimester prenatal care for American Indians and Alaska Natives compared with Whites did not change throughout the study period (RR = 0.89 in both 1985–1987 and 1995–1997).

Low-birthweight rates remained stable for American Indians and Alaska Natives throughout the study period but increased slightly for Whites in the 1995–1997 time period. The adjusted odds of having a low-birthweight infant were lower for American Indians and Alaska Natives than for Whites throughout the study period.

Neonatal, postneonatal, and overall infant death rates decreased for both American Indians and Alaska Natives and Whites over the study period. The unadjusted neonatal death rate of AIAN infants was higher than that of White infants in 1985–1987 and 1989–1991, but not in 1995–1997. After adjustment, the odds of an AIAN infant dying in the neonatal period were no different from those of a White infant in the 1985–1987 and 1989–1991 time periods, and were lower than those of a White infant in the 1995–1997 time period. The postneonatal death rates of AIAN infants were significantly higher than those of White infants in each of the three time periods. The adjusted odds of postneonatal death among AIAN infants compared with White infants did diminish slightly (but not statistically significantly), from 1.55 in 1985–1987 to 1.46 in 1995–1997.

Neonatal and postneonatal causes of death for American Indians and Alaska Natives and Whites over the study period are presented in Table 3. In the neonatal period, congenital anomalies, respiratory conditions, and short gestation and low birthweight were the most common causes of death for American Indians and Alaska Natives and Whites in all 3 time periods. AIAN infants were more likely than were White infants to have SIDS reported as the cause of death in the first 28 days of life throughout the study period (AIAN-to-White rate ratio = 1.97 in 1985-1987, and 2.52 in 1995–1997).

TABLE 3.

Causes and Rates (per 1000 Live Births) of Neonatal and Postneonatal Death Among American Indian and Alaska Native (AIAN) and White Births: National Linked Birth and Infant Death Data, 1985–1987, 1989–1991, and 1995–1997

	1985–1987			1989–1991			1995–1997
	AIAN Rate	White Rate	AIAN-to-White RR	AIAN Rate	White Rate	AIAN-to-White RR	AIAN Rate	White Rate	AIAN-to-White RR
Neonatal Death
Congenital anomalies	1.63	1.62	1.00	1.62	1.45	1.12	1.15	1.30	0.89
Respiratory distress syndrome	0.59	0.56	1.05	0.42	0.44	0.97	0.29	0.22	1.32
Other respiratory conditions	0.51	0.52	0.99	0.50	0.38	1.33	0.13	0.26	0.49*
Short gestation, unspecified low-birthweight-related disorders	0.60	0.48	1.25	0.45	0.49	0.91	0.56	0.49	1.15
Complications of cord, membrane, placenta	0.17	0.21	0.81	0.24	0.19	1.25	0.36	0.18	1.99**
Intrauterine hypoxia, birth asphyxia	0.24	0.20	1.20	0.24	0.12	1.97**	0.10	0.12	0.81
Infections specific to perinatal period	0.23	0.20	1.17	0.21	0.17	1.21	0.11	0.14	0.80
Complications of pregnancy	0.23	0.15	1.57	0.22	0.16	1.39	0.11	0.20	0.57
Sudden infant death syndrome	0.19	0.09	1.97*	0.16	0.10	1.66	0.15	0.06	2.52**
Infectious diseases	0.04	0.04	1.00	0.15	0.05	3.15***	0.06	0.05	1.02
Maternal conditions	0.03	0.03	1.05	0.03	0.03	0.84	0.08	0.02	3.44*
Unintentional injuries and accidents	0.04	0.02	2.33	0.07	0.02	2.87*	0.01	0.02	0.70
Homicide	0.01	0.004	3.21	…	0.01	…	…	0.01	…
All other causes	0.96	0.75	1.27	0.71	0.59	1.22	0.67	0.62	1.09
Postneonatal Death
Sudden infant death syndrome	2.81	1.27	2.22***	2.73	1.31	2.09***	1.57	0.80	1.95***
Congenital anomalies	0.76	0.51	1.50**	0.82	0.52	1.59***	0.56	0.43	1.30
Infectious disease	1.17	0.35	3.40***	0.94	0.27	3.44***	0.76	0.23	3.35***
Unintentional injuries and accidents	0.60	0.24	2.46***	0.70	0.24	2.87***	0.41	0.23	1.80**
Other respiratory conditions	0.11	0.10	1.14	0.05	0.08	0.70	0.06	0.05	1.09
Respiratory distress syndrome	0.07	0.04	1.83	0.07	0.04	1.83	…	0.02	…
Homicide	0.14	0.03	4.58***	0.12	0.04	2.83**	0.14	0.05	2.26**
Intrauterine hypoxia, birth asphyxia	…	0.01	…	…	0.01	…	…	0.01	…
Infections specific to the perinatal period	0.01	0.01	1.83	…	0.01	…	0.01	0.01	1.67
Short gestation, unspecified low-birthweight disorders	…	0.01	…	…	0.01	…	0.01	0.01	2.18
Complications of cord, membrane, placenta	…	0.001	…	…	0.003	…	…	0.003	…
Complications of pregnancy	…	0.001	…	…	…	…	…	0.001	…
Maternal conditions	0.01	…	…	0.03	0.001	26.0**	0.01	0.001	21.77
All other causes	1.43	0.71	2.03***	1.24	0.62	1.99***	1.26	0.56	2.26***

Note. RR = rate ratio. Ellipses indicate there were no deaths from this cause in this study time period.

*P ≤ .05; **P ≤ .01; ***P ≤ .001, for statistically significant differences in death rates per 1000 live births.

The most common causes of death in the postneonatal period for both American Indians and Alaska Natives and Whites were SIDS, congenital anomalies, infectious disease, and unintentional injuries and accidents. AIAN infants had higher rates than did White infants of postneonatal death from each of these conditions, as well as from homicide, throughout the study period. However, it is encouraging to note that the AIAN-to-White rate ratio of postneonatal death decreased for each of these measures between 1985–1987 and 1995–1997. For SIDS and congenital anomalies, the 1995–1997 AIAN-to-White rate ratios of postneonatal death were about 87% of the 1985–1987 ratios. For unintentional injuries and accidents, and homicide, the 1995–1997 AIAN-to-White rate ratios of postneonatal death were 73% and 49% that of the 1985–1987 rate ratios, respectively.

Despite these improvements, as recently as 1995–1997, there were more than 3 times as many infectious disease deaths in the postneonatal period among rural American Indians and Alaska Natives as among rural Whites (54 AIAN deaths, but only 16 White deaths in an equivalently sized population of 71 300). Similarly, there were nearly twice as many postneonatal SIDS deaths (112 AIAN deaths, 57 White deaths) and deaths caused by unintentional injuries and accidents (29 AIAN deaths, 16 White deaths) among rural American Indians and Alaska Natives as among rural Whites in an equivalently sized population of 71 300.

DISCUSSION

We examined perinatal care, birth outcomes, and infant health among rural American Indians and Alaska Natives and Whites during the 1980s and 1990s, a period when funding for public programs in maternal and child health expanded. We found that the perinatal and infant health of rural White populations improved alongside that of American Indians and Alaska Natives during this time. As a result, disparities between American Indians and Alaska Natives and Whites in postneonatal death rates and prenatal care access either remained stable or deteriorated somewhat.

Though funding for public maternal and child health programs grew during this period, the US Commission on Civil Rights, Noren et al. and Roubideaux document that funding for the Indian Health Service fell far below what was necessary to provide for the health care needs of American Indians and Alaska Natives.^15–17 The significant improvement in access to prenatal care and infant death rates among both rural American Indians and Alaska Natives and rural Whites between the mid-1980s and the late 1990s is encouraging, but the persistent disparities between rural American Indians and Alaska Natives and rural Whites in access to care and infant death are of considerable concern. It is particularly important to pay attention to rural American Indians and Alaska Natives because of the very high proportion of that population living in rural settings—39.5% in the 2000 census—and the more limited medical resources available in rural areas.^18–20

The late 1980s and early 1990s was a time of expanded funding of programs to enroll low-income pregnant women into Medicaid as early as possible to ensure their receipt of timely prenatal care and to provide services such as case management and social support for women with high-risk pregnancies.³ Thus, it is not surprising that we documented improvements among both rural American Indians and Alaska Natives and Whites in use of early and adequate prenatal care or that these improvements were most dramatic directly following the most intensive period of Medicaid expansions for maternity care. However, it is disappointing that the rural AIAN population, whose very high rates of inadequate and late initiation of prenatal care afforded the opportunity for dramatic improvement, demonstrated deterioration in adequacy of prenatal care compared with the White population, as shown in our adjusted analyses. This deterioration is of particular concern because rural American Indians and Alaska Natives have persistently higher rates of preexisting medical conditions, such as cardiac disease, chronic hypertension, and diabetes, that may benefit from early and ongoing intervention during pregnancy. Neither the rural AIAN population nor the rural White population met the Healthy People 2000 goal of 90% of women receiving first-trimester care, and rural American Indians and Alaska Natives continued to fall far below this target, with only 66.7% receiving first-trimester care in 1995–1997.²¹

Our prior analyses of the 1989–1991 National Linked Birth and Infant Death Data had demonstrated the marked disparity in infant death rates between American Indians and Alaska Natives and Whites and we identifed the disparity to be among infants in the postneonatal period.² The current study demonstrated that both the neonatal and postneonatal death rates decreased continuously from 1985–1987 through 1995–1997, with the greatest decline occurring between 1989–1991 and 1995–1997. In these outcomes, rural American Indians and Alaska Natives made some gains relative to Whites in postneonatal mortality. Nonetheless, even in 1995–1997, the odds of postneonatal death among rural AIAN infants were still 1.46 times that of White infants. By the 1995–1997 time period, rural Whites met the Healthy People 2000 objectives of fewer than 7 infant deaths per 1000 live births, 4.5 neonatal deaths per 1000 live births, and 2.5 postneonatal deaths per 1000 live births. The Healthy People 2000 objectives for AIAN infants were less ambitious, with target rates of 8.5 infant deaths, 4.5 neonatal deaths, and 4 postneonatal deaths per 1000 live births. The rural AIAN infants reached or came very close to each of these goals.

Our cause-of-death analysis suggests that the greatest gains in closing the postneonatal death rate disparity between rural American Indians and Alaska Natives and Whites were in the areas of SIDS and congenital anomalies, the 2 most common causes of postneonatal death, as well as in the less common causes of unintentional injuries and accidents, and homicide. There was little improvement among rural American Indians and Alaska Natives relative to rural Whites in deaths from infectious disease, the most common of which was pneumonia. This is consistent with recent literature demonstrating the substantial burden of lower-respiratory-tract disease among American Indians and Alaska Natives.^22,23 These infections are largely treatable and preventable, suggesting that significant improvement in rural AIAN postneonatal mortality could occur with early care for infectious illnesses.

Low birthweight, a frequently used measure of perinatal health, is similar between American Indians and Alaska Natives and Whites, a finding reported elsewhere in the literature.²⁴ It is therefore not surprising that neonatal death rates, which are largely determined by birthweight, are comparable between rural American Indians and Alaska Natives and rural Whites.

Limitations

This study is limited by some inconsistency in the data between the 3 time periods, including lack of measurement of Hispanic ethnicity in 27 states during the 1985–1987 time period, which may have misclassified some individuals of Hispanic ethnicity as White. This would have minimized the differences between rural AIAN and White populations in that time period, although we expect the impact to be small, because most of the states with the largest Hispanic populations were gathering those data at that time. In addition, several of the maternal risk characteristics (e.g., cigarette use, preexisting medical conditions) were not available in the 1985–1987 data. Thus, we were only able to control for a limited number of covariates in our regressions. However, we conducted a subanalysis limited to births in the 1989–1991 and 1995–1997 time periods, controlled for a full set of variables, and did not find meaningful differences in our results.

Another limitation of this study is the age of the data, which may not be representative of the current situation. These data were the most current available when this study was begun, however, and our analysis tracks rural perinatal care, birth outcomes, and infant health during an important period of policy and funding changes in maternal and child health. National Linked Birth and Infant Death Data are now available through 2002, providing future opportunities to monitor continuing trends in perinatal care and birth outcomes among American Indians and Alaska Natives in both rural and urban areas.²⁵

Our identification of mothers' residence location as rural was done at the level of the county. Zip code–based rural classifications are more accurate, but zip code of mother's residence at birth is not available in the Linked Birth and Infant Death Data. Thus, these results do not reflect the outcomes of births in rural zip codes located in urban counties. Also, it is important to note that the data reported here reflect outcomes for women living in rural areas both within and outside IHS areas. However, the vast majority of the AIAN births in this study, 83.0%, were to women living in IHS area counties, so that much of the care that these women received was in IHS-funded facilities.

Conclusions

There have been a number of changes over the past few decades in the organization of health services for American Indians and Alaska Natives, resulting in increased tribal autonomy over their own health systems.^17,26 Tribal control over health services provides an excellent opportunity to improve perinatal and infant outcomes by implementing culturally appropriate interventions that could prevent or modify preexisting risk factors (such as hypertension and diabetes) known to be higher among American Indians and Alaska Natives,^27–30 increase prenatal care use, and decrease the risk of preventable conditions (such as infection) that result in postneonatal death.

However, some groups, especially those from direct IHS and tribal health programs, face challenges because of the sizeable proportion of rural American Indians and Alaska Natives living in remote locations, where it is difficult to recruit both primary care and specialty physicians and where there are fewer health care facilities with more-limited services, longer distances to health care providers, and lack of public transportation options.³¹ Adequate funding is needed to ensure that American Indians and Alaska Natives have access to services and programs that help prevent postneonatal infant death and improve access to prenatal care.³² In this way, American Indians and Alaska Natives will have an opportunity to reach the Healthy People 2010 objectives,³³ in which American Indians and Alaska Natives and Whites are expected to reach the same goals.