Objective. To determine geographic variation in urban American Indian and Alaska Native (AI/AN) rates of infant mortality, low birth weight, prenatal care use, and maternal-child health care service availability.
Methods. This was a retrospective cohort study using data from the 1989 to 1991 birth-death linked database from the National Center for Health Statistics. We examined births from metropolitan areas with a minimum of 300 AI/AN births during the study period. Key outcomes of interest included rates of low birth weight, neonatal mortality, postneonatal mortality, and women receiving inadequate prenatal care using the modified Kessner index. To determine the type of health services tailored to AI/AN mothers residing in these urban areas, we conducted a telephone survey of the 36 urban Indian health programs operating in 1997 using a semistructured survey. Items in the survey included questions about the availability of prenatal and infant health care.
Results. During the 1989 to 1991 study period, there were 72 730 singleton births to AI/AN mothers and/or fathers residing in urban areas, representing 49% of all AI/AN births in the United States. Overall 14.4% of urban AI/AN births were to women who received inadequate care during pregnancy, 5.7% of pregnancies resulted in low birth weight infants, and 11.0 infants died per 1000 live births. Death rates for the neonatal period (5.5 per 1000 births) and postneonatal period (5.4 per 1000 births) were similar. Marked disparity in these indicators exist between pregnancies to AI/AN and white women. Among the 54 metropolitan areas, 46 had a rate ratio (AI/AN: white) for inadequate care of ≥1.5 (range: 0.9–8.5). The mean rate ratios for neonatal and postneonatal mortality were 1.6 (range: 0.3–4.0) and 2.0 (range: 0.5–5.5). There was also considerable geographic variation of AI/AN mortality rates between metropolitan areas in all of the outcomes studied. All of the 20 metropolitan areas with the highest birth counts had some type of direct medical care or outreach services available from an urban clinic targeted toward AI/AN patients.
Conclusions. Considerable variation also exists among rates of AI/ANs between metropolitan areas. Disparity exists in rates of perinatal outcomes between AI/ANs and whites living in the same metropolitan areas Although AI/AN urban health programs exist in most cities with large birth counts, it seems that many have inadequate resources to meet existing needs to improve perinatal outcomes and infant health.
Despite considerable improvements over the past 40 years, the health status of American Indian and Alaska Native (AI/AN) children is lower compared with most children in the United States.1 Health status indicators are collected primarily for AI/ANs who reside within the service boundaries of the Indian Health Service (IHS). The AI/AN population has become increasingly urbanized during the past 2 decades; there is relatively little information that specifically describes the perinatal health status of AI/AN women and infants who reside in urban areas. One study of the AI/AN population in a single metropolitan area found that urban Indian women had considerably lower rates of prenatal care use and higher rates of infant mortality than their Indian reservation counterparts within the same state.2
The health status of urban AI/AN women and infants is of particular importance because, in most cities, these children are not eligible for or do not have access to health services provided directly by the IHS or tribal health programs. The IHS has provided some limited funding for health services to urban AI/ANs under Title V of the 1976 Indian Health Care Improvement Act.3 These funds, which constitute <2% of the entire IHS budget, have been used to support clinical care and outreach programs in >35 large and small cities throughout the United States. A few cities, such as Anchorage, Albuquerque, and Phoenix, have long established IHS or tribally operated clinical care programs and facilities that are funded from regular IHS hospital and clinic budgets. In some cities, such as Albuquerque and Phoenix, Title V urban health programs also operate in the same area as regular IHS Programs. Some tribes (eg, the Puyallup tribe in Tacoma, Washington) also operate clinics in metropolitan areas because of the proximity of the reservation to the city. Access to these tribal clinics varies for AI/AN patients who are not members of the tribe and is determined by the tribe.
The ascertainment of the health status of pregnant AI/AN women and their infants in urban locations can be used to target interventions toward populations that are at potentially high risk of poor outcomes. The purpose of this study was to profile and describe variation in perinatal risk factors and AI/AN infant mortality rates for metropolitan areas within the United States. Specifically, we sought to determine the rates of inadequate receipt of prenatal care, low birth weight, and infant mortality for AI/AN births in US metropolitan areas with >300 births during a 3-year period. A second aim of the study was to survey urban AI/AN health programs to determine the availability of prenatal and pediatric health services targeted to AI/AN women and children who reside in these urban areas.
A full description of the methods used to conduct this study is published elsewhere.4 Briefly reviewed, this study was a retrospective cohort design using data from the 1989 to 1991 National Linked Birth-Death Database from the National Center for Health Statistics.5 This database contains information from birth certificates for all 50 states and the District of Columbia on all live births, and linked information from death certificates for infants who died during the first year of life. This study is limited to singleton births in which either the mother or father was reported as being of AI/AN race. Although this racial classification rule differs from National Center for Health Statistics convention, we chose to include infants with either AI/AN mothers or fathers because both groups were potentially eligible for IHS services.
Study Variable Definitions
Urban births were classified on the basis of the county of residence of the mother. We used the Metropolitan Statistical Area (MSA) designation as defined by the Federal Office of Management and Budget and implemented by the US Department of Agriculture.6 All MSAs were classified as urban, and non-MSA counties were classified as rural. We used Consolidated Metropolitan Statistical Areas that aggregate adjacent metropolitan counties. For example, San Francisco, Oakland, and San Jose were combined into 1 Consolidated Metropolitan Statistical Area.
The following indicators were examined in this study: birth counts and rates (number of births per 100 000 AI/AN population), low birth weight rate (expressed as the proportion of infants with birth weight <2500 g), neonatal death rate (the number of deaths among infants within 27 days of birth per 1000 live births), postneonatal death rate (the number of deaths among infants between 28 days and 364 days of age per 1000 live births), and the proportion of women who received inadequate prenatal care (using the modified Kessner index). The modified Kessner index is based on the month when prenatal care began and the number of prenatal visits adjusted for gestational age at birth.7 The modified Kessner index does not reflect prenatal care quality but rather the pattern of care received. Women who were classified as having an “inadequate” pattern of prenatal care were primarily those who initiated care in the third trimester, regardless of the number of prenatal visits they received.
Urban Health Program Survey
A brief 15-item telephone survey of urban Indian health programs was prepared using closed- and open-ended questions. The survey included questions about program patient volume and characteristics, the availability of prenatal care, and linked obstetric services. We also asked whether children could receive medical care at the clinic. Data were collected by 1 of the authors (B.N.) in telephone interviews with the directors (or designate) of the health program.
The survey sample was obtained from the IHS urban program office. The initial sample included a list of 36 urban Indian clinics and programs identified as operating in 1997 by the IHS urban program office. The study was initially introduced to the clinic director by letter to the executive director and was administered to the director or a designate by mail survey with telephone follow-up during the summer of 1997. Of the 36 programs contacted, 34 completed the survey. Completed surveys were coded and entered into SPSS statistical package (SPSS, Inc, Chicago, IL) for analysis.
There were 148 482 singleton births to AI/AN mothers and/or fathers in the United States during the 3-year period of 1989 to 1991. Of these, 49% (n = 72 730) were to mothers residing in urban counties. The characteristics of these births, the parents, and the pregnancies are described in Table 1. Only approximately 18% of births in urban areas were to parents who were both identified as AI/AN. Almost 7% of births were to mothers under age 18, and approximately 10% experienced some complications from labor.
Overall, 14.4% of urban women received inadequate care during their pregnancy, and 5.7% of births resulted in low birth weight infants. The overall infant mortality rate in this group was 11.0 per 1000 live births. Deaths were approximately equally distributed between the neonatal (5.5 deaths per 1000 births) and postneonatal (5.4 deaths per 1000 births) periods.
The rankings of metropolitan areas by birth count and the availability of different types of urban Indian health programs are shown in Table 2. There were 52 metropolitan areas with 300 or more AI/AN births in the 3-year study period. These areas accounted for 80% of all urban AI/AN births during this time period. The 5 metropolitan areas with the highest AI/AN birth counts included Los Angeles-Orange County, California (n = 5283); Phoenix-Mesa, Arizona (n = 4930); Tulsa, Oklahoma (n = 3919); Seattle-Tacoma-Bremerton, Washington (n = 3116); and Albuquerque, New Mexico (n = 2557). Only 1 of these areas, Los Angeles, was also in the top 5 urban areas for births of all races in the United States.
The use of prenatal care is frequently used as a marker for the health access and health status of populations. Rates of inadequate prenatal care (ie, modified Kessner Index) among AI/ANs were highest in Minneapolis-St Paul, Minnesota (29.8%); Yakima, Washington (28.2%); Billings, Montana (27.1%); Yuma, Arizona (24.8%); and Bellingham, Washington (23.7%; Table 3). It is interesting that only 1 of these cities, Yuma, Arizona, also appeared among the top 5 areas for inadequate care among white mothers. The areas with the greatest disparity in prenatal care use, as indicated by high rate ratios for AI/AN and whites, were Minneapolis, Minnesota (rate ratio [RR]: 8.5); Duluth, Minnesota (RR: 4.2); Billings, Montana (RR: 4.1); Great Falls, Montana (RR: 4.3); and Salt Lake City, Utah (RR: 4.2). In only 4 areas (San Diego, Fresno, Houston, and Yuma), the rates of inadequate care were lower for AI/ANs than for whites. The urban areas with the highest number of births with inadequate prenatal care use included Phoenix (n = 1095), Los Angeles (n = 634), Minneapolis (n = 605), and Albuquerque (n = 512).
The urban areas with the highest proportion of low birth weight infants were mostly in the eastern United States and included New York City (9.2% of births); Boston area (8.4%); Philadelphia-Wilmington-Atlantic City area (8.2%); Great Falls, Montana (8.0%); and Washington, DC-Baltimore (8.0%; Table 4). In contrast, the top 5 areas for white low birth weight included Albuquerque, New Mexico (6.7%); Denver-Boulder, Colorado (6.2%); and Reno, Nevada (5.6%). The disparity between whites and AI/AN infants for low birth weight was not as great as for prenatal care. The areas with the greatest low birth weight disparity had AI/AN rates that were approximately 1.5 to 2 times higher than for white rates. Areas with the highest numbers of low birth weight infants were Los Angeles (n = 303), Phoenix (n = 257), Tulsa (n = 181), and Seattle (n = 180). None of the areas had rates of low birth weight that were less than the white population in the same area.
Infant mortality rates are summarized and stratified into neonatal mortality and postneonatal mortality rates in Table 4. Neonatal death rates among AI/ANs in IHS service areas were reported to be approximately 21% higher than rates among whites in 1993.1 However, in this study, some urban areas were noted to have local rates that were considerably higher. (Because of the small counts of deaths, these rates are somewhat unstable with large confidence intervals.) The 5 metropolitan areas with the highest AI/AN neonatal death rates included Buffalo-Niagara Falls, New York (12.54 per 1000 live births); Providence, Rhode Island-Fall River-Warwick, Massachusetts (11/1000); Yakima, Washington (10/1000); Bellingham, Washington (10/1000); and Milwaukee-Racine, Wisconsin (9/1000) (Table 4). These metropolitan areas also exhibited considerable disparity between races for neonatal mortality rates with rate ratios (AI death rates/white death rates for these areas ranging from 2.7 to 4.0). The urban areas with the highest number of neonatal deaths included Los Angeles (n = 34), Phoenix (n = 33), Seattle (n = 19), and Oklahoma City (n = 15). Altogether, there were 31 metropolitan areas in which the AI/AN neonatal death rate was higher than the highest white rate in Wichita, indicating considerable disparity within most metropolitan areas.
The postneonatal death rate among AI/ANs was highest in Yakima, Washington (16.3 per 1000 live births); Minneapolis-St Paul, Minnesota (12/1000); Redding, California (12/1000); Bakersfield, California (11/1000); and Reno, Nevada (11/1000; Table 4). These areas also had rates that were 3- to 5-fold higher than rates among whites in the same metropolitan areas. There were 22 metropolitan areas in the sample in which the AI/AN postneonatal mortality rate was at least twice as high as the mortality rate among white infants. Disparities between AI/AN and white postneonatal mortality rates were highest in Yakima, Washington (RR: 8.9); Redding, California (RR: 5.2); Minneapolis-St Paul, Minnesota (RR: 5.2); Fayetteville, North Carolina (RR: 5.0); and Reno, Nevada (RR: 4.4). Thirty-two of the 50 metropolitan areas we examined had AI/AN postneonatal mortality rates greater than the highest white rate experienced by Reno, Nevada (4.37 per 1000 births). The urban areas with the highest counts of deaths occurring during the postneonatal period included Minneapolis (n = 29), Seattle (n = 27), Phoenix (n = 26), and Los Angeles (n = 18).
Urban Indian Health Programs
All of the most populous 20 urban areas have some type of health services specifically targeted to the local AI/AN population (Table 2). Most of these services are Title V urban health programs with direct medical service care (Urban Health Center), but several (eg, Los Angeles, New York, Denver) offer only outreach and referral services and do not provide direct medical care. Some of the urban areas, such as Albuquerque and Tucson, have IHS direct care services available. Nearly all (87%) of these Title V urban health programs providing direct medical care offered prenatal care, but only a quarter of these had physicians who attended the clinic’s deliveries. Many of the clinics reported a referral relationship with an obstetric service program in the community.
This national study demonstrates considerable disparity between AI/AN and white infants who reside in the same metropolitan areas of the United States in the prevalence of risk factors associated with morbidity and mortality. This disparity is most likely associated with the higher rates of poverty, levels of maternal education, and other pregnancy risk factors among AI/AN women compared with whites.
Our study also found that some form of direct medical care or outreach services is available in the metropolitan areas with the highest AI/AN birth counts. However, in some cases, the services may not be well matched to need. One example is the Los Angeles-Orange metropolitan area of California, the area with the highest number of AI/AN births. This area has a single urban Indian health center that provides outreach and referral services but not direct medical or prenatal care. In contrast, the Seattle-Bremerton, Washington, area has a large multispecialty urban clinic that provides prenatal care and has linkage with obstetric providers. This lack of uniformity with regard to access to maternal and child health care contrasts sharply with direct IHS care, where access to basic health care and prenatal services is generally similar across reservations and states.
We are unaware of other studies that have examined variations in perinatal outcomes among specific metropolitan areas for the AI/AN population and examined disparities between AI/ANs and other ethnic or racial groups. Our findings are consistent with several other studies of urban AI/AN populations, documenting high rates of inadequate care and high-risk births.2,8
The use of secondary data for this study is associated with limitations. First, these data are now dated, although they represented the latest available data at the time the study was initiated in 1997. It is conceivable that changes have occurred in the relative rankings since that time. For example, recent reports by Robertson et al9 indicate a significant recent drop in AI/AN infant mortality in the Pacific Northwest. However, it is unknown whether the metropolitan areas have also experienced this drop and whether this trend has also occurred among whites who live in the same areas. Second, the rankings for mortality rates are based on small numbers of deaths and have relatively wide confidence intervals. The exact rank position is of less importance than the overall consistency of these data indicating the rates are considerably higher than for the white populations in the same geographic areas. Finally, no association between the presence or absence of a health care facility and morbidity or mortality rates can be inferred from these data. Our own survey results indicate that many of the urban health centers are caring for a relatively small fraction of the pregnant AI/AN women and infants in their metropolitan area (data not shown). This may occur for many reasons, including that many of these urban counties cover large geographic areas, making geographic access a potential barrier to obtaining care at these facilities. Of interest, some tribes have attempted to address the shortfall of services for their members who reside in adjacent urban areas by providing urban satellite clinics. Because linked birth-death data were used, misclassification of infant race in our death data, reported as a problem in past studies of infant death, was minimized.10 Finally, misclassification may exist in the counting of urban births. Some pregnant women return to their reservation homes to deliver to access reservation health care and to ensure tribal enrollment of their offspring.
How can policy makers use these data? First, this study clearly indicates that a large fraction of the US AI/AN population giving birth is living in metropolitan areas, most of which are not served by traditional IHS full-service medical and public health programs. The urban health programs clearly are positioned to play a larger role in efforts to reduce AI/AN infant mortality. Urban AI/AN programs have been increasingly recognized by the IHS as important constituents of the IHS mission. Although they are often not able to meet all clinical needs, they are well-situated to work with local, state, and federal agencies to address the disparities outlined in this study. Second, policy makers should target resources and efforts to those communities with the greatest disparity in rates between AI/ANs and whites.
Finally, there is an increasing recognition that the IHS clinical care program is not adequately funded to address the public and personal health needs of AI/ANs who reside in urban areas. Funding of the IHS has not kept pace with the growth of the population or health care inflation rates, making it difficult, if not impossible, for it to address adequately the needs of these urban populations.11 Until Congress provides adequate funding for the health care of all AI/ANs, local urban Indian health centers, as some have done, may need to partner with public and private agencies and provide leadership in efforts to address the maternal and child health care needs of all urban AI/AN residents, not just those who currently use clinical services.
We thank the National Center for Health Statistics for their help with the linked birth and death certificate data. We also thank Bill Freeman, Aaron Handler, and Linda Querec at the Indian Health Service for assistance with this project. Jonathan Sugarman provided additional guidance on the project, and George Brenneman provided valuable feedback on an earlier draft of this manuscript.
- ↵Indian Health Service. Trends in Indian Health 1997. Rockville, MD: Indian Health Service;1997
- ↵Baldwin LM GD, Casey S, Hollow W, Sugarman JR, Freeman WL, Hart LG. Perinatal and infant health among rural and urban American Indians/Alaska Natives. Am J Public Health.2002. In press
- ↵US Department of Health and Human Services. Linked Birth/Infant Death Data Set: 1989 Birth Cohort. Hyattsville, MD: USDHHS;1995
- ↵US Department of Agriculture. Electronic 1993 Department of Agriculture Update of Rural-Urban Continuum Codes. Washington, DC: US Department of Agriculture;1993
- ↵Kessner DM, Kalk CW. Infant Death: An Analysis by Maternal Risk and Health Care. Contrasts in Health Status. Vol. 1. Washington, DC: Institute of Medicine, National Academy of Sciences;1973:50– 95
- ↵US Department of Health and Human Services. LNF Workgroup. Level of Need Funded Cost Model. Rockville, MD: Indian Health Service;1999
- Copyright © 2002 by the American Academy of Pediatrics