The maternal health of American Indian and Alaska Native people: a scoping review

Abstract

Indigenous Peoples in the United States experience disadvantage in multiple domains of health. Yet, their maternal health receives limited research attention. With a focus on empirical research findings, we conduct a scoping review to address two questions: 1) what does the literature tell us about the patterns and prevalence of maternal mortality and morbidity of American Indian and Alaska Native (AI/AN) people? and 2) how do existing studies explain these patterns? A search of CINAHL, Embase and Medline yielded 66 eligible English-language articles for close review. Empirical research focused on AI/AN people’s maternal health experiences is limited. AI/AN people experience higher levels of maternal mortality and morbidity than non-Hispanic White people, with estimates that vary substantially across samples and geography. Explanations for the maternal health of AI/AN people focused on individual factors such as poverty, cultural beliefs, and access to health care (e.g. lack of insurance). Studies rarely addressed the varied historical and structural contexts of AI/AN tribal nations, such as harms associated with colonization and economic marginalization. Research for and by Indigenous communities and nations is needed to redress the effective erasure of AI/AN people’s experiences of maternal health and to advance solutions that will promote their health and well-being.

1. Introduction

The United States of America (U.S.) is experiencing a maternal health crisis. Around 800 deaths occur every year from pregnancy-related complications and most of these deaths are preventable (Hoyert 2022; Petersen et. al. 2019). Additionally, maternal health in the U.S. has been declining in recent decades (MacDorman et al. 2016), a situation that appears to have grown worse during the COVID-19 pandemic (Thoma and Declercq 2022). In 2020, the U.S. maternal mortality rate was 23.8 deaths per 100,000 live births (Hoyert 2022), a level more than three times higher than nine other high income countries (Gunja et al. 2022), including nearby Canada (Statistics Canada 2022). Further, considerable research has documented alarming racial and ethnic disparities in maternal health in the U.S. (Braveman et al. 2001; Bryant et al. 2010; Lyndon et al. 2012); birthing people who identify as Black and American Indian and Alaska native (AI/AN) are 3.2 and 2.3 times more likely to die from pregnancy-related causes than people who identify as non-Hispanic White (Petersen et al. 2019).

Amidst this maternal health crisis, attention to the maternal health of AI/AN¹ people is notably limited. This is troubling given evidence that AI/AN people are particularly subject to high rates of maternal mortality and morbidity. For example, according to 2012 data from 13 states collected by the Pregnancy Risk Assessment Monitoring System (PRAMS, a joint state-federal research project on maternal and infant health), almost 18% of AI/AN new mothers reported experiencing postpartum depressive symptoms (Ko et al. 2017), double the percentage of non-Hispanic White people reporting postpartum depressive symptoms in the study. In 2018, 9.9% of pregnant AI/AN people experienced gestational diabetes, a percentage almost 1.5 times greater than the national average (Martin et al. 2019). Finally, AI/AN people are 2.3 times as likely to die from pregnancy-related causes, and roughly twice as likely to experience severe maternal morbidity, compared to non-Hispanic White people (Kozhimannil et al. 2020; Petersen 2019). Though attention to the maternal health of AI/AN people is limited, existing evidence clearly indicates they experience disproportionately poor maternal health outcomes. In 2020, Kozhimannil summarized the crisis and moral imperative for evidence and action, and observed, “addressing our nation’s maternal health crisis requires attention to the experiences of Indigenous people and communities” (p. 2).

In this paper, we seek to contribute to the understanding of the maternal health of AI/AN people in the U.S. by conducting a scoping review of the scientific literature. Specifically, we ask: 1) what does the literature tell us about the patterns and prevalence of maternal mortality and morbidity of AI/AN people? and 2) how do existing studies explain these patterns? After presenting the results of this review, we outline gaps in knowledge and suggest future directions for research. First, we briefly describe the historical background that shapes the health of AI/AN people and situate their maternal health within the larger national context of maternal health.

2. Background

2.1. Historical Oppression and Health

The U.S. Census Bureau refers to American Indians and Alaska Natives (AI/AN) as people with origins in North and South America who maintain community connections or tribal affiliations (Norris et al. 2012). In 2019, the Census Bureau Population Estimates Program estimated that approximately 5.7 million Americans identify as AI/AN (The U.S. Department of Health and Human Services Office of Minority Health n.d). There is substantial variation in tribal affiliation among AI/AN people in the U.S.; as of 2022, the federal government official recognized 574 tribal nations (Office of the Federal Register, National Archives and Records Administration 2020).

The maternal health outcomes of AI/AN peoples occur within a distinct social, political, and historical context. European colonizers arrived in North America in the late 1400s, and under the “doctrine of discovery,” claimed ownership of Indigenous land (Shelton 2004). Centuries of human rights violations, including massacres, forced relocation, broken treaties, the placement of children in Indian boarding schools, and forced assimilation followed. In total, from 1492 to the current day, some scholars estimate that upwards of 5 million Indigenous peoples occupying the present-day U.S. died as a direct effect of European colonization (Thornton 1990).

Contemporary oppressions persist. Systemic structural barriers that limit access to safe housing, stable income, and education constrain the health and well-being of AI/AN people and communities (Committee on Community-Based Solutions to Promote Health Equity in the United States 2017; Gordon 2022). Though the U.S. government ended its formal practice of removing Indian children from their homes and placing them in boarding schools in the late 1960s, today, foster care and state interventions continue to remove children from their homes and Tribes (Kruse et al. 2022). Researchers suggest these many historical losses, and the ongoing group trauma induced by separation and other harms, can reverberate through generations and manifest in challenges such as stress, depression, substance use, and suicidal ideation (John-Henderson et al. 202; Whitbeck 2004). Further, such trauma and stress can have long-term and intergenerational effects on health through epigenetics, a process by which a person’s experiences are embodied in ways that affect gene expression (Brockie, Heinzelmann, and Gill 2013). All of these harms should likewise be expected to affect maternal health (Committee on Community-Based Solutions to Promote Health Equity in the United States 2017).

2.2. Indian Health Services

The system that provides health care to AI/AN people is both separate from, and interwoven with, the broader U.S. healthcare system and policies. In 1954, the U.S. government guaranteed free health care to federally-recognized tribal nations and established the Indian Health Service (IHS) within what is now known as the U.S. Department of Health and Human Services (HHS) (Allison et al. 2007). The domain of the IHS includes public health and health care, and each year, about 2.2 million AI/AN people receive their health care from the IHS (The U.S. Department of Health and Human Services Office of Minority Health n.d). However, problems diminish several dimensions of access² and quality. Only people from federally recognized tribal nations are eligible for health care services, some health centers are small and informal, and some provide less than optimal care (Roubideaux 2002; Zuckerman et al. 2004). Further, some AI/AN people may not be registered, while others may belong to tribal nations that are state recognized, but not federally recognized. Indeed, the IHS is regularly criticized for being inaccessible and underfunded (Roubideaux 2002; National Congress of American Indians 2016; Zuckerman et al 2004).

The organization and structure of the IHS has changed over time and has resulted in a complex network of care that engages the federal government, Tribal governments, and urban Indian groups. While the United States Public Health Service originally owned and operated IHS facilities, in 1975, the U.S. Congress passed the Indian Self-Determination and Education Assistance Act, which allowed IHS and Bureau of Indian Affairs functions to be contracted by federally recognized Tribes and tribal organizations (Allison et al. 2007). In the 1990s, Tribes gained more autonomy and responsibility for the provision of care to Nations through tribal compacting, whereby Tribes could elect to run programs and services to meet local needs (Kruse et al. 2022). However, due to inadequate funding, services in the system and referrals outside the system are limited, greatly diminishing care provision (Frerichs et al. 2022). In 2010, the Affordable Care Act (ACA) permanently authorized funding for the IHS system, and through its insurance provisions, improved health insurance coverage for some AI/AN people, however, these gains were small and substantial disparities remain (Frerichs et al. 2022).

The health of pregnant AI/AN people is affected by this complicated provision of services. The IHS system, which is chronically underfunded, operates facilities that do not provide obstetric care, leading many Indigenous people to give birth outside of this IHS system (Kozhimannil et al. 2020). Further, as observed by the membership organization American College of Obstetricians and Gynecologists (ACOG) in a 2012 statement, the IHS may not cover services if they are outside the beneficiary’s home service area, unless the person has been away for less than 180 days, creating complicated gaps in coverage or unnecessary travel (ACOG 2012). In 2021, ACOG reaffirmed its 2012 statement on healthcare for AI/AN people calling for its Fellows to recognize the need for obstetric care and coverage among AI/AN people in urban areas (ACOG website 2021). Finally, though many of the challenges related to access to obstetric services and maternal care for AI/AN people result from the larger problem of inadequate health care provision to AI/AN people, the maternal health of AI/AN people is also situated in the larger context of a U.S. maternal health crisis.

2.3. Maternal Health Crisis

Despite spending more money on maternal healthcare than any other country, the U.S. has a disproportionately high maternal mortality rate compared to countries with similar gross domestic products (GDPs) (Maternal Health Task Force 2015). Moreover, the maternal mortality rate has increased over time (MacDorman et al. 2016). The rate of severe maternal morbidity (SMM), defined as “unexpected outcomes of labor and delivery that result in significant short- or long-term consequences to a woman’s health” (American College of Obstetricians and Gynecologists 2016) also almost doubled between 1993 and 2014 (CDC 2021). Rates of morbidity and mortality vary along numerous dimensions. In addition to higher rates among AI/AN people, poorly educated, Medicaid-insured, and non-Hispanic Black people are more likely to experience SMM or mortality than highly educated, privately insured, and non-Hispanic White people (Chen et al. 2021; Singh and Lee 2021).

Individual predictors of maternal mortality and SMM include maternal age, pre-existing conditions, multiple births, and a prior cesarean section (Gray et al. 2012). Current research has also identified structural risk factors, such as barriers to early uptake of prenatal care (Bingham, Strauss, and Coeytaux 2011). For example, transportation issues, travel time, and administrative burdens limit Medicaid enrollment and prenatal care uptake (Bingham, Strauss, and Coeytaux 2011; Herd and Moynihan 2018). Furthermore, people of color (including AI/AN people) are often subjected to structural racism during healthcare interactions, which manifests as racial essentialism, medicalization, and social control (Bridges 2011).

The pathways connecting individual and structural factors with maternal health vary across racial, ethnic, socioeconomic, and geographic lines. Indeed, AI/AN people on Medicaid may be disproportionately likely to experience barriers to care, as they are less likely to report that it is “always or usually easy to get need medical care” and more likely to report that “they are never able to see a specialist as soon as needed” than non-Hispanic White Medicaid beneficiaries (Medicaid and CHIP Payment and Access Commission 2021; Frerichs et al. 2019). Given that an estimated 42% of AI/AN people rely on public health insurance (including Medicaid, Medicare, and Veteran Affairs coverage) (U.S. Census Bureau 2019), access to public healthcare is a particularly salient issue for this population.

The maternal health outcomes of AI/AN people are often overlooked in the larger conversation on U.S. maternal health. The American Public Health Association (APHA) and the Maternal Health Task Force at the Harvard Chan School of Public Health call attention to the drastic disparities between White pregnant people and both Black and AI/AN pregnant people (American Public Health Association 2011; Maternal Health Task Force 2015). While scholars have examined the multitude of ways that racism affects maternal health (Davis, 2019; Slaughter-Acey et al., 2019), research on the maternal health of AI/AN people is notably limited.

3. Methods

Given the breadth of our topic, we performed a scoping review rather than a systematic review, which is better suited for narrowly defined research questions (Arksey and O’Malley 2003). Scoping reviews are particularly useful for identifying gaps in the literature, which is still necessary in the field of AI/AN maternal health: a primary contribution of this review. We utilized the scoping review framework outlined by Arksey and O’Malley (2003) and followed the PRISMA extension reporting guidelines for scoping reviews (Tricco et al. 2018). We conducted a search for relevant literature based on specific inclusion criteria outlined below. We coded the data using the software Covidence and summarized the results in a set of descriptive tables.

We conducted our literature search in December 2021 using three databases: Medline, CINAHL, and Embase. We retrieved all articles that contained terms related to maternal health and terms related to populations indigenous to the U.S. The maternal health terms included those related to pregnancy, pregnancy complications, prenatal, maternal morbidity, maternal mortality, maternal health, obstetrics, and maternal health services. We used the search filters on Indigenous peoples of the U.S. created by the University of Alberta for Embase, Medline, and CINAHL (Campbell 2021a; Campbell 2021b; Campbell 2021c), which includes over 200 terms related to individual tribal nations and Indigenous populations. We limited our search to articles published after 2010.

We restricted the search to peer-reviewed, published original research available in the English-language that included qualitative or quantitative data on the maternal health of AI/AN people. In practice, this means we excluded articles if 1) they did not focus on people indigenous to the U.S., 2) they did not include data on maternal health specific to AI/AN people, or 3) they were not empirical research. We also excluded intervention studies and papers exclusively on the postpartum period unless they were about maternal mortality or severe maternal morbidity.

Two authors developed a test extraction sheet to code eligible articles for study design, sample, measures, outcomes, and mechanisms. They each then coded 4 articles independently, reviewed consistency, and refined the extraction sheet. The remaining articles were coded using a refined extraction sheet, with the help of an undergraduate research assistant. Any discrepancies were discussed and resolved.

We used a broad definition of maternal health, including physical health, mental health, and health behaviors (such as smoking and drinking during pregnancy). The inclusion of health behaviors required further discussion among the authors, as it was unclear whether these should be considered maternal morbidities for the purpose of this review. These behaviors are known to complicate pregnancy and childbirth and can have negative effects on a person’s well-being (Centers for Disease Control and Prevention 2022). While severe maternal morbidity (SMM) refers to serious clinically defined complications of labor and delivery, often referred to as “near-miss” complications, not all maternal morbidity is severe. Given the lack of a clear and comprehensive definition, in 2012, the World Health Organization supported an effort to review and broaden the definition, conceptualization, and measurement of maternal morbidity. This effort led to an expert group, the Maternal Morbidity Working Group, which defines maternal morbidity as “any health condition attributed to and/or complicating pregnancy and childbirth that has a negative impact on the woman’s wellbeing and/or functioning” (Chou et al. 2016). The group explicitly stated that maternal health is a social and economic phenomenon, and more than a clinical and biological issue. Guided by this expanded conceptualization and definition, we include health-related behaviors, such as smoking and drinking alcohol during pregnancy as indicators of maternal morbidity, as they reflect and are a result of social and structural phenomena.

4. Results

4.1. Characteristics of the Literature

Out of 4757 articles yielded by the search, the authors screened 4151 abstracts. Of these, 66 met eligibility criteria. Figure 1 summarizes the selection of articles.

Figure 1.

Search strategy for scoping review

Table 1 reports study objectives, samples, and other details on the included articles. Of these articles, only one study used qualitative research methods (Hanson 2012). None of the studies named a specific tribal nation, though several focused on a specific region of the country. Instead, most papers studied “American Indians and Alaska Natives” as a single group.

Table 1.

Summary characteristics of studies included in the review (n=66)

	Study Characteristics				Measurement of Maternal Health*
Reference	Category**	Location of Study	Years of Data Collection	Data Source	Mortality Outcomes	Morbidity Outcomes
Admon 2018	2	National	2012-2015	National Inpatient Sample (NIS)	-	Severe maternal morbidity (SMM)
Alemu 2020	3	North Carolina	2000-2014	State Inpatient Database	-	Opioid use during pregnancy
Azagba 2020	2	National	2010-2017	National Center for Health Statistics public-use natality data	-	Smoking during pregnancy
Aziz 2019	2	National	2012-2014	NIS	-	SMM
Barlow 2010	1	4 rural reservation communities in the Southwest	2006-2008	Primary survey and interview data collection	-	Drug use during pregnancy
Booker 2018	2	National	1998-2014	NIS	Maternal death	SMM
Bronars 2018	1	Anchorage, Alaska	Not specified	Primary survey data collection	-	Perceived risks and reasons for tobacco use during pregnancy
Cabacungan 2012	2	Wisconsin	2005-2007	Healthcare Cost and Utilization Project State Inpatient Dataset	-	Obstetric complications during labor and delivery
Chalouhi 2015	1	Gallup, New Mexico	2009-2012	Rehoboth McKinley Hospital medical records	-	Postpartum hemorrhage
Chang 2014	2	National	2010	National Center for Health Statistics public-use natality data and NIS	-	Pregnancy-associated hypertension, eclampsia, diabetes, tobacco use during pregnancy
Creanga 2014	2	AZ, CA, FL, MI, NJ, NY, NC	2008-2010	Healthcare Cost and Utilization Project’s State Inpatient Database	-	SMM
Curtin 2016	3	46 states and D.C.	2014	Birth certificate data	-	Smoking and quitting smoking during pregnancy
Danielson 2018	1	North Dakota	2007-2012	Birth data from the North Dakota Division of Vital Records	-	Cesarean section and risk factors for poor birth outcomes
deRavello 2015	1	National	2002-2009	Indian Health Services Patient Information Reporting System	-	Ectopic pregnancy
England 2013	1	Western Alaska	1997-2005	Resource and Patient Management Systems of the regional hospital/medical center and the Alaska Native Medical Center	-	Pregnancy-associated hypertension, preeclampsia
Fridman 2014	3	California	1999, 2002, 2005	Vital statistics birth data and hospital discharge cohort data	-	Pregnancy-related hypertension, gestational diabetes mellitus (GDM), mental health, substance use, tobacco use
Gray 2012	2	Washington State	1987-2008	Birth certificate and hospital discharge data from the Comprehensive Hospital Abstract Reporting System	-	SMM
Gyamfi-Bannerman 2018	2	National	2012-2014	NIS	In-hospital death for women with postpartum hemorrhage	SMM
Gyamfi-Bannerman 2020	2	National	2012-2014	NIS	In-hospital death for women with preeclampsia	SMM
Hadley 2021	1	Alaska	2018-2019	Hospital medical records from a Level III maternal referral center	Maternal death	Postpartum hemorrhage, antepartum bleeding, preeclampsia
Hanson 2012	1	A reservation in the Northern Plains	Not specified	Primary interview data collection	-	Prenatal health behaviors
Hebert 2021	1	Alaska, New Mexico, Oklahoma, South Dakota, and Washington	2015-2017	Pregnancy Risk Assessment Monitoring System (PRAMS)	-	Smoking during pregnancy, alcohol use during pregnancy
Henke 2014	3	44 states	2009	Healthcare Cost and Utilization Project State Inpatient Databases linked with national demographic data	-	Cesarean delivery
Hitti 2018	3	Washington State	2013-2017	Delivery records from the University of Washington Medical Center	-	SMM
Hoshiko 2019	2	Southern California	1999-2002	Project Baby’s Breath	-	Smoking and environmental tobacco smoke exposure during pregnancy
Houston-Ludlam 2020	2	Missouri	2010-2017	Individual-level state birth record data linked with census-tract level American Community Survey data	-	Smoking during pregnancy
Hsieh 2020	2	National	2003-2013	NIS	In-hospital mortality	Cesarean section complications and comorbidities
Hunsberger 2010	2	Oregon	2004-2005	PRAMS	-	GDM
Hunt 2011	2	Chicago, Illinois	2003-2005	Birth certificates from Illinois Vital Records	-	Smoking during pregnancy
Jorda 2021	1	One Tribal Nation in the Great Plains Area, North Dakota	2007-2015	Safe Passage Study	-	Smoking and drinking during pregnancy
Kern-Goldberger 2021	2	Not specified	1999-2002	The Cesarean Registry from the Eunice Kennedy Shriver National Institute of Child Health and Human Development Maternal-Fetal Medicine Units	Maternal mortality	Composite of critical care interventions, alcohol and drug use during pregnancy
Khan 2013	1	Anchorage, Alaska	Not specified	Primary survey data collection	-	Alcohol use and smoking during pregnancy
Khanolkar 2020	2	Washington State	2006-2008	Washington State Birth Events Records Database	-	Gestational weight gain, gestational hypertension, preeclampsia/ eclampsia, cesarean section, and length of stay in hospital
Kim 2010	1	Alaska	2000-2003	PRAMS	-	Tobacco use during pregnancy
Kim 2012a	2	Florida	2004-2007	Live birth certificate data linked with Florida Hospital Inpatient Discharge database	-	GDM
Kim 2012b	2	Northern California	1995-2006	Kaiser Permanente Northern California electronic health records	-	Placenta previa
Kim 2013	2	California	2007-2009	Hospital delivery records	-	GDM
Kim 2014	1	Western Alaska	1997-2005	Abstracted medical records	-	GDM, pregnancy-associated hypertension, anemia, cesarean section
Kozhimannil 2019	3	National	2007-2015	NIS	Maternal mortality	SMM excluding transfusion
Kozhimannil 2020	1s	National	2012-2015	NIS	Death during delivery	SMM, cesarean section
Leonard 2019	2	California	1997-2014	Birth certificate records linked to delivery discharge records	-	SMM during delivery hospitalization
Lisonkova 2013	3	Washington State	2003-2008	The Comprehensive Discharge Abstract Database linked with the Birth Events Record Database	-	Early-onset and late-onset preeclampsia
Luke 2021	2	National	2012-2017	NIS	In-hospital mortality	SMM
Melville 2010	3	Not specified	2004-2009	Primary survey data collection	-	Antenatal depressive disorders
Merkt 2021	2	National	2011-2016	Pregnancy Mortality Surveillance System (PMSS) linked with National Vital Statistics System birth data	Pregnancy-related deaths	-
Morriss 2018	3	46 states	2011	NIS linked to the American Hospital Association Annual Survey database	In-hospital deaths	Transfers to an acute care hospital
Mukherjee 2018	2	National	2009-2011	PRAMS	-	Number of maternal morbidities during pregnancy
Okah 2014	3	Kansas City, Missouri	1990-2009	Vital Statistics birth certificate data	-	Tobacco or alcohol use during pregnancy
Palladino 2011	3	SC, GA, NC, VA, NJ, MD, AK, MA, OR, CO, OK, RI, WI, KY, NM, UT	2003-2007	National Violent Death Reporting System (NVDRS)	Pregnancy-associated violent deaths	-
Petersen 2019a	3	National	2011-2015	PMSS	Pregnancy-related mortality	-
Peterson 2019b	2	National	2007-2016	PMSS	Pregnancy-related mortality and cause-specific mortality	-
Schummers 2019	2	National	2004-2013	Vital Statistics Birth Cohort-Linked Birth-Infant Death data files and Fetal Death data files	-	Cesarean delivery
Shah 2020	3	National	2006-2015	NIS	Maternal inpatient mortality	Acute kidney injury during pregnancy-related hospitalization
Singh 2018	2	National	2014-2015	National birth certificate data and natality files	-	Pregnancy-related hypertension
Specker 2018	1	South Dakota	2014	South Dakota Pregnancy Survey data linked with birth certificate data	-	Tobacco and alcohol use during pregnancy and quit status during pregnancy
Stulberg 2014	2	AZ, CA, CO, FL, IL, IN, IA, LA, MA, MI, MN, MS, NY, TX	2004-2008	Medicaid administrative claims data	-	Ectopic pregnancy
Stulberg 2016	2	AZ, CA, CO, FL, IL, IN, IA, LA, MA, MI, MN, MS, NY, TX	2004-2008	Medicaid administrative claims data	Ectopic pregnancy mortality	Ectopic pregnancy and associated complications (blood transfusion, hysterectomy, other sterilizing surgery, or hospitalization length >2 days)
Sullender 2020	2	Wisconsin	2011-2016	Wisconsin Interactive Statistics on Health	-	Smoking during pregnancy
Tabet 2017	2	National	2012	Birth certificate data	-	Smoking during pregnancy, gestational weight gain
Tiwari 2021	2	Washington State	2014-2018	Foundation for Health Care Quality’s Obstetrical Care Outcomes Assessment Program	-	GDM, preeclampsia, cesarean section
Toh 2013	3	National	2001-2007	Medication Exposure in Pregnancy Risk Evaluation Program	-	Antipsychotic use at delivery
Ukah 2019	3	Washington State	2004-2013	Birth Event Record Database linked with Comprehensive Hospital Abstract Reporting System	-	Gestational weight gain
Watt 2012	1	National	2005-2009	National Survey of Drug Use and Health (NSDUH)	-	Cigarette and alcohol use during pregnancy
Wingo 2012	1	National, split into 9 geographic regions (Aberdeen, Alaska, Bemidji, Billings, California, Oklahoma, Portland, Southwest, and Nashville)	1990-2007	National Vital Statistics public use natality data	-	Pregnancy-associated hypertension and cesarean section
Ye 2020	1	Northern Plains region	2006-2017	The Safe Passage Study	-	Alcohol use during pregnancy
Zamora-Kapoor 2016	1	Washington State	2003-2013	Washington State linked birth-hospital discharge records	-	Preeclampsia and smoking during pregnancy

^*Maternal health outcomes included in the study. Results are not presented for AI/AN people for all outcomes

^**1: Specific focus on AI/AN people; 2: Focus on racial/ethnic disparities with AI/AN included; 3: Purpose is unrelated to race/ethnicity but they disaggregate by race

We categorized the articles in the review according to their level of engagement with AI/AN people, assigning each to one group, either 1) specific focus on AI/AN people, 2) general focus on racial/ethnic disparities with AI/AN people included, or 3) purpose unrelated to health of AI/AN people specifically, or race/ethnicity disparities, but data disaggregated to present relevant statistics on AI/AN people. Nineteen studies specifically focused on AI/AN people. Some of these had samples that were exclusively AI/AN people, and some included other populations, but all outcomes and discussion was related to the health of AI/AN people. Thirty-two studies focused on racial/ethnic disparities and included AI/AN as one of the subpopulations they studied. Finally, fifteen studies did not focus on racial or ethnic disparities, but disaggregated their findings, and included AI/AN people as a sub-population.

The studies we reviewed captured substantial geographic (and consequently, tribal) variation in the U.S. Twenty-six studies presented nationwide or nearly nationwide data, and thirty-three presented data from one state or geographic region. The rest either did not specify a location (n=2) or analyzed data from 2 or more states/regions (n=5).

4.2. Maternal Health among American Indian and Alaska Native People

We captured studies that analyzed maternal mortality and morbidity, which includes physical health, mental health, and health behaviors. We describe findings related to each in turn.

4.2.1. Maternal Mortality

Overall, fifteen studies present data on maternal mortality among AI/AN people at a national or subnational level (summarized in Table 2). The majority of the articles focused on maternal mortality at the national level within hospital settings, defining mortality using the CDC’s ICD-9-CM codes. Across articles, estimates of national maternal mortality rates of AI/AN people ranged from 28.8 for 2011-2016 to 32.5 for 2011-2015 (Merkt et al. 2021; Petersen et al. 2019) measured as the number of maternal deaths per 100,000 live births. Estimates vary by rurality; one study found higher pregnancy-related mortality ratios in small cities and rural areas than in large metropolitan areas from 2011-2016 (Merkt et al. 2021).

Table 2.

Data on maternal mortality among AI/AN people

	Mortality Estimates
Study	Mortality Measure	Total sample size and percentage of sample that is AI/AN	Reference Group	Statistically Significant Estimates†$ with 95% CIs or p-values
Booker 2018	Maternal mortality	n=1,724,694, 0.41% AI/AN	NA	No cases reported for AI/AN
Gyamfi-Bannerman 2018	Maternal in-hospital death for women with postpartum hemorrhage	n=360,370, % AI/AN not specified	NA	No cases reported for AI/AN
Gyamfi-Bannerman 2020	Maternal in-hospital death for women with preeclampsia	n=107,741, 0.95% AI/AN	NA	Authors note: Not reported due to small cell size
Hadley 2021	Postpartum death due to postpartum hemorrhage	n=384, 100% AI/AN	NA	No cases reported for AI/AN
Hsiesh 2020	Mortality among mothers undergoing cesarean section with perioperative epidural catheter placement	n=87,076, 1% AI/AN	NA	No cases reported for AI/AN
Kozhimannil 2019	SMM and mortality excluding transfusion-only	Rural Residents: n=987,185, 1.81% AI/AN Urban Residents: n=5,806,147, 0.52% AI/AN	non-Hispanic White people	Adjusted odds ratio of SMM and mortality: 1.61 (1.44, 1.8) (p<0.001)
Kozhimannil 2020	SMM and mortality	n=7,561,729, 1.3% AI/AN	non-Hispanic White people	Risk ratio of SMM and mortality: 1.8 (1.0 - 2.6)
			Urban Indigenous pregnant people	Risk ratio of SMM and mortality: 1.3 (1.0 - 1.6)
Luke 2021	Inpatient SMM and mortality	n=4,494,089, 0.72% AI/AN	non-Hispanic White people	Adjusted odds ratio of SMM with transfusion and mortality: 1.56 (1.44-1.7) Adjusted odds ratio of SMM and mortality excluding blood transfusion: 1.47 (1.3-1.67)
Merkt 2021	Maternal mortality across urban and rural settings	n=3747, 1.73% AI/AN	non-Hispanic White people	Pregnancy Related Mortality Ratio (PRMR) Large Metro: 18.5 (8.9, 34.1) PRMR Medium Metro: 20.3 (9.3, 38.6) PRMR Small Metro: 39.2 (20.2, 68.4) PRMR Micropolitan: 32.3 (18.1, 53.2) PRMR Noncore: 37.5 (22.6, 58.6)
Morriss 2018	Inpatient mortality	n=4,250,230, 0.80% AI/AN	NA	Overall estimate provided, not disaggregated by race.
Palladino 2011	Pregnancy- associated suicide	n=4,539,551 live births, 1.6% AI/AN	NA	Proportion of total pregnancy-associated suicides in sample: 5.1% (n=5)
Petersen 2019a	Maternal mortality	n=3410 deaths, 1.8% AI/AN	NA	PRMR: 32.5 (Pregnancy-related deaths per 100,000 live births)
Petersen 2019b	Maternal mortality	n=6765, 1.73% AI/AN	NA	PRMR: 29.7 (Pregnancy-related deaths per 100,000 live births)
			non-Hispanic White people	PRMR disparity ratio: 2.3
				Prevalence of death due to hemorrhage with chi-squared test for association with race: 19.7% (p<0.05)
				Prevalence of death due to infection with chi-squared test for association with race: 8.5% (p<0.05)
				Prevalence of death due to hypertensive disorders of pregnancy with chi-squared test for association with race: 12.8% (p<0.05)
Shah 2020	Inpatient mortality	n=42,190,790 delivery hospitalizations, 0.7% AI/AN	NA	Overall estimate provided by acute kidney and non-acute kidney infections, not disaggregated by race.
Stulberg 2016	Mortality due to ectopic pregnancy	n=45,201,325 person years of Medicaid enrollment, 1.2% AI/AN	NA	No cases reported for AI/AN

^†statistical significance set at p<0.05

^$in studies that reported results from multiple models, we present results for the model with the full set of covariates.

The relatively small number of studies on mortality may reflect the scarcity of national-level data sources with sufficient sample size for AI/AN women; apart from one case control study that exclusively focused on AI/AN people (Hadley 2021), all of the mortality studies were based on samples where AI/AN people represented, at most, 2% of the sample. Booker and colleagues (2018) analyze risk of maternal death by reported race and ethnicity but note that they are unable to report on this outcome for AI/AN people due to insufficient sample size.

Several studies, using both national and regional data, focused on maternal mortality due to a specific cause of death, such as postpartum hemorrhage, preeclampsia, cesarean sections, interhospital transfers and suicide/homicide (n=8), while other provided estimates of combined mortality and SMM (n=4). In all cases where comparisons are made, AI/AN people experienced a higher risk of mortality (or mortality and SMM) than non-Hispanic White people. The level of difference depended on sample and outcome, but in many cases AI/AN people experienced a risk of death that was at least 1.5 times greater (see Table 2).

4.2.2. Maternal Morbidity

Sixty-two studies provided estimates of maternal morbidity, including SMM and other morbidities, such as pregnancy complications, mental health, and health behaviors.

4.2.2.1. Severe Maternal Morbidity

Table 3 shows estimates of composite measures of SMM. Despite different samples, years, and methods of analysis, the studies generally find that AI/AN people experience high rates of SMM (see Table 3). At a national level, two studies estimate that about 2% of AI/AN people experienced SMM from 2012-2015 (Admon et al. 2018; Kozhimannil et al. 2020). However, such estimates mask considerable heterogeneity across local areas and subpopulations. For example, at a single hospital in Washington State from 2013-2017, an estimated 11.7% of AI/AN delivery hospitalizations were classified as cases of SMM (Hitti et al. 2018). Further, Kozhimannil et al. (2020) estimates that AI/AN people living in rural areas experience a prevalence of SMM that is about 0.5 percentage points higher than AI/AN people in urban areas. These regional variations reflect geographic differences, but also the diversity of tribal nations that inhabit different lands.

Table 3.

Data on severe maternal morbidity (SMM) among AI/AN people

	Severe Maternal Morbidity Estimates
Study	Morbidity Measure	Total sample size and percentage of sample that is AI/AN	Reference Group	Statistically Significant Estimates†$ with 95% CIs or p-values
Admon 2018	SMM	n=2,523,528, 0.81% AI/AN	non-Hispanic White people	Prevalence of SMM: 0.02% Adjusted rate ratio of SMM with blood transfusion incidence per 10,000 hospitalizations: 1.5 (1.3-1.7)
Aziz 2019	SMM	n=207,730, 0.89% AI/AN	non-Hispanic White people	Prevalence of SMM without transfusion with chi-squared test of association with race: 16.7% (p<0.01) Prevalence of SMM with transfusion with chi-squared test of association with race: 21.0% (p<0.01)
Booker 2018	SMM	n=1,724,694, 0.41% AI/AN	non-Hispanic White people	Unadjusted risk ratio of SMM with transfusion: 1.2 (1.03-1.39)
Creanga 2014	SMM	n=3,476,392, 0.6% AI/AN	non-Hispanic White people	Prevalence of SMM with transfusion: 2.2% Rate of SMM with transfusion per 10,000 delivery hospitalizations: 225.47 +/− 10.36 Adjusted risk ratio of SMM with transfusion: 1.7 (1.5-1.8) Prevalence of SMM without transfusion: 0.8% Rate of SMM without transfusion per 10,000 delivery hospitalizations: 75.97 +/− 6.06 Adjusted risk ratio of SMM without transfusion: 1.3 (1.1-1.5)
Gray 2012	SMM	n=9,485, 3.9% AI/AN	non-Hispanic White people	Adjusted odds ratio of SMM: 1.52 (1.32, 1.73)
Gyamfi-Bannerman 2018	SMM	n=360,370, %AI/AN not specified	non-Hispanic White people	No significant results
Gyamfi-Bannerman 2020	SMM among women with pre-eclampsia	n=101,741, 0.95% AI/AN	non-Hispanic White people	Unadjusted risk ratio of SMM with transfusion: 1.38 (1.11-1.71)
Hitti 2018	SMM	n=7,025, 1.58% AI/AN	non-Hispanic White people	Prevalence of SMM: 11.7% Unadjusted odds ratio of SMM: 3.3 (1.8-6.0)
Kozhimannil 2019	SMM and mortality excluding transfusion-only	Rural Residents: n=987,185, 1.81% AI/AN Urban Residents: n=5,806,147, 0.52% AI/AN	non-Hispanic White people	Adjusted odds ratio of SMM and mortality: 1.61 (1.44, 1.8) (p<0.001)
Kozhimannil 2020	SMM	n=7,561,729, 1.3% AI/AN	NA	Prevalence of: SMM: 2.0% SMM among rural pregnant people: 2.3% SMM among urban pregnant people: 1.8% SMM without transfusion: 0.6% SMM without transfusion among rural pregnant people (0.4%) SMM without transfusion among urban pregnant people (0.6%)
			non-Hispanic White people	Adjusted risk ratio of SMM: 1.8 (1.6-2.0)
			Urban Indigenous pregnant people	Adjusted risk ratio of SMM among rural AI/AN pregnant people: 1.3 (1.0-1.6)
			non-Hispanic White people	Adjusted risk ratio of SMM excluding transfusion: 1.9 (1.5-2.3)
Leonard 2019	SMM during delivery hospitalization	n=8,252,025, 0.3% AI/AN	non-Hispanic White people	Prevalence of SMM with transfusion: 1.30% Prevalence of SMM without transfusion: 0.55% Adjusted risk ratio of SMM with transfusion: 1.29 (1.15, 1.44)
Luke 2021	SMM	n=4,494,089, 0.72% AI/AN	non-Hispanic White people	Adjusted odds ratio of SMM with transfusion and mortality: 1.56 (1.44-1.7) Adjusted odds ratio of SMM and mortality excluding blood transfusion: 1.47 (1.3-1.67)

^†statistical significance set at p<0.05

^$in studies that reported results from multiple models, we present results for the model with the full set of covariates.

When presenting SMM in terms of odds ratios or risk ratios, almost every study in this review used non-Hispanic White people as the reference category. Most findings suggest that AI/AN people have higher risk or odds of SMM relative to non-Hispanic White people (see Table 3). Again, estimates vary by region and year. Using national data from 1998-2014, Booker and colleagues (2018) did not find a significant difference in the adjusted risk of SMM between AI/AN and non-Hispanic White people. However, Leonard and colleagues (2019) use data from roughly the same period and estimate that AI/AN people in California have a 1.29 times greater adjusted risk of SMM compared to non-Hispanic White people. Creanga and colleagues (2014) estimate an even higher risk ratio using 2008-2010 data from seven states. In their analysis, AI/AN people have a 1.7 times greater adjusted risk of SMM relative to non-Hispanic White people.

4.2.2.2. Other Morbidities

Given space limitations, we list study-specific outcomes related to other morbidities in Table 1 and present an overview of findings in this section. We do not provide study-specific estimates of each physical or mental health condition described in every article. Many studies measured gestational diabetes mellitus (GDM), preeclampsia, hypertension, and cesarean sections as specific morbidity outcomes (see Table 1). Measures and estimates vary considerably across studies, revealing no clear pattern of the prevalence or risk of these conditions among AI/AN people. For instance, patterns of gestational diabetes mellitus (GDM) vary substantially. Two studies in North Dakota and California find that AI/AN people have a higher likelihood of being diagnosed with GDM than non-Hispanic White people, even after adjusting for sociodemographic characteristics (Danielson et al. 2018, Fridman et al. 2014). However, studies in Washington State and Oregon do not find significant differences in GDM between AI/AN pregnant people and others (Hunsberger, Rosenberg, and Donatelle 2010; Tiwari et al. 2021).

AI/AN people have a similar risk of cesarean section to other pregnant people, and the prevalence of this delivery complication is highly variable across studies. Estimates range from a 5.5% prevalence of cesarean delivery in Western Alaska from 1997 to 2005 (Kim et al. 2014) to a 23% prevalence in Washington State over the more recent time frame of 2006-2008 (Khanolkar et al. 2020). Risk may also vary by insurance type; in a study of 47 states, Henke et al. (2014) found that AI/AN pregnant people using Medicaid to pay for delivery hospitalization had a lower odds of a cesarean section than non-Hispanic White people, but the opposite was true for AI/AN people using private insurance.

Two studies considered mental health outcomes during pregnancy among AI/AN people. A study by Melville et al. (2010) examined antenatal depressive disorders and found that AI/AN people experienced levels similar to that of other groups. The second study documented typical use of antipsychotic medicine in about 1.0% of AI/AN deliveries (Toh et al. 2013). As this measure only captures the use of medications, it likely underestimates the population of pregnant AI/AN people with psychosis. In sum, we know little about the mental health experiences of pregnant AI/AN people.

4.2.2.3. Health Behaviors

Finally, Table 4 summarizes health behaviors (such as alcohol, tobacco, and drug use) among pregnant AI/AN people. Among the studies included in this review, estimates of the national prevalence of tobacco use among AI/AN people during pregnancy range from 16.6% during 2010-2017 (Azagba et al 2020) to 29.6% during 2005-2009 (Watt 2012). In comparison, the same studies estimate that among pregnant non-Hispanic Whites, almost 9% used tobacco from 2010 to 2017 (Azagba et al 2020) and 19.7% smoked from 2005-2009. There is substantial geographic and temporal variation in the estimates of tobacco use among AI/AN pregnant people. While approximately 43% of Alaska Natives in Alaska used tobacco during pregnancy from 2000-2003 (Kim et al. 2010), closer to 20% of pregnant AI/AN people used tobacco in Washington State between 2003-2013 (Zamora-Kapoor 2016). Again, this geographic variation may reflect tribal variation.

Table 4.

Data on health behaviors during pregnancy among AI/AN people

	Health Behavior Estimates*
Study	Morbidity Measure	Total sample size and percentage of sample that is AI/AN	Reference Group	Statistically Significant Estimates†$ with 95% CIs or p-values
Alemu 2020	Opioid use	n=1,937,455; 1.26% AI/AN	non-Hispanic White people	Rate of opioid use: 5.03 (4.14, 5.92) per 1,000 pregnancy-related hospital discharges Adjusted odds ratio for opioid use: 0.74 (0.61, 0.90)
Azagba 2020	Smoking	n=27,912,380; 1.10% AI/AN	NA	Prevalence of smoking (all years) with chi-square test: 16.6% (p<0.001) Trend in prevalence of smoking from 2010 to 2017 with Cochran-Armitage test: ~16% in 2010, ~15% in 2017 (p<0.001)
Barlow 2010	Meth and other drug use	n=322; 100% AI/AN	AI/AN who lived in 1 home in the past year	Adjusted risk ratio of drug use for those who lived in 4+ homes in the past year: 1.98 (1.15, 3.39)
			AI/AN with parents who don’t have a serious drinking problem	Adjusted risk ratio of drug use for those whose parents have serious drinking problem: 1.76 (1.12, 2.75)
			AI/AN with no family member who committed suicide	Adjusted risk ratio of drug use for those with family members who committed suicide: 2.22 (1.23, 3.98)
			AI/AN who do not find traditional values important	Adjusted risk ratio of drug use for those who think traditional values are very important: 0.42 (0.23, 0.76)
			NA	Prevalence of drug use and chi-squared test for association with whether parents/guardians argue a lot: 39.4% (p=0.003)
				Prevalence of drug use and chi-squared test for association with whether the respondent does a lot of things with parents/guardians: 19.2% (p=0.001)
				Prevalence of drug use and chi-squared test for association with whether parents/guardians have frequent arguments with yelling/screaming: 41.8% (p=0.002)
				Prevalence of drug use and chi-squared test for association with whether respondent thinks their parents/guardians like talking/being with them: 23.1% (p=0.037)
				Prevalence of drug use and chi-squared test for association with whether respondent thinks their parents/guardians have a good idea of their interests: 23.1% (p=0.027)
Bronars 2018	Reported risks and reasons for tobacco use	n=118; 100% AI/AN	NA	Prevalence of agreement that tobacco use impact’s baby’s development with chi-squared test: Smokers 58.5%; Non-smokers 77.4% (p=0.029) Prevalence of agreement that addiction is a reason for smoking during pregnancy with chi-squared test: Smokers 63.0%; Non-Smokers 82.8% (p=0.015) Mean (SD) and t-test of difference in number of children for smokers and non-smokers: Smokers 2.5 (3.4); Non-smokers 1.4 (1.5) (p=0.0217) Prevalence of living with another smoker with chi-squared test: Smokers 85.2%; Non-smokers 42.2% (p<0.0001) Mean (SD) and t-test of difference in hours of secondhand smoke exposure per day for smokers and non-smokers: Smokers 5.9 (5.9); Non-smokers 1.6 (2.9) (p<0.0001)
Chang 2014	Tobacco use	n=3,113,164; 1.10% AI/AN	NA	Prevalence of tobacco use with chi-squared test for association with race: 18.6% (p<0.01)
Curtin 2016	Smoking	n=3,819,114; <1% AI/AN	NA	Prevalence of: smoking during pregnancy: 18% smoking during first trimester: 17.50% smoking during second trimester: 14.70% smoking during third trimester: 14% quitting smoking during pregnancy: 22.20%
Danielson 2018	Smoking, drug use, alcohol use	n=51,844; 11% AI/AN	non-Hispanic White people	Prevalence ratio of smoking 2010-2012: 2.7 (p<0.05) Prevalence ratio of illegal drug use 2010-2012: 13.3 (p<0.05) Prevalence ratio of alcohol use 2010-2012: 2.1 (p<0.05)
Fridman 2014	Substance use and tobacco use	n=1,551,017; 0.40% AI/AN	non-Hispanic White people	Adjusted odds ratio of tobacco use: 0.603 (0.516, 0.706)
Hall 2020	Tobacco use	n=167,463; 2.39% AI/AN	NA	Prevalence of tobacco use during pregnancy: 11.6%
Hebert 2021	Smoking	n=14,176; 29.43% AI/AN	non-Hispanic White people	Prevalence of alcohol use during last trimester with chi-squared test of association with race: Did not drink then 88.7%; less than 1 drink/week 7%; 1 or more drinks/week 4.3% (p=0.0049)
			non-Hispanic White people	Prevalence of smoking during pregnancy with chi-squared test of association with race: 13.6% (p<0.0001)
			AI/AN <19 years old	Adjusted odds ratio of drinking during pregnancy by age: 20-24 years 3.38 (1.13, 10.14); 35 and older 4.22 (1.06, 16.88)
			AI/AN who did not experience homelessness	Adjusted odds ratio of drinking during pregnancy among those who experienced homelessness: 2.76 (1.16, 6.55)
Hoshiko 2019	Smoking	n=3014; 7.1% AI/AN	NA	Prevalence of smoking during mid-pregnancy: 12.7% (9.4, 16.9)
Houston-Ludlam 2020	Smoking	n=159,668; 1.47% AI/AN	non-Hispanic White people	Adjusted odds ratio of smoking during pregnancy among those unmarried with acknowledged paternity: 2.00 (1.75, 2.66) Adjusted odds ratio of smoking during pregnancy among those unmarried with unacknowledged paternity: 2.80 (2.03, 3.84) Adjusted odds ratio of smoking during pregnancy among those married with acknowledged paternity: 10.38 (4.09, 26.36)
Hunt 2011	Smoking	n=48,483; 0.289% AI/AN	NA	Prevalence of smoking with pair-wise comparison: Chicago 10.1%; U.S. 24.4% (p<0.001)
Jorda 2021	Smoking and alcohol use	n=421; 100% AI/AN	NA	Prevalence of: quitting smoking during pregnancy among those living alone: 6.6% (p<0.05) quitting smoking during pregnancy among those whose mother has 6-11 years of education: 6.9% (p<0.05) quitting smoking during pregnancy among those whose parents did not complete HS: 1.9% (p<0.05) quitting smoking during pregnancy among those with a yearly income <$500: 7.04% (p<0.05) smoking among those living alone: 79.3% (p<0.05) smoking among those whose mother has 6-11 years of education: 80.6% (p<0.05) smoking among those whose parents did not complete HS: 79.3% (p<0.05) smoking among those unemployed >1 year: 83.6% (p<0.05) smoking among those whose parents are unemployed: 83.5% (p<0.05) smoking among those with a yearly income <$500: 76.1% (p<0.05) smoking among those with depression: 84.5% (p<0.05) smoking among those with high PTSD: 88.2% (p<0.05) smoking and drinking among those living alone: 41.9% (p<0.05) smoking and drinking among those whose mothers have 6-11 years of education: 37.8% (p<0.05() smoking and drinking among those who unemployed >1 year: 35.7% (p<0.05) smoking and drinking among those whose parents are unemployed: 36.2% (p<0.05) smoking and drinking among those with a yearly income<$500: 38.0% (p<0.05) smoking and drinking about those with depression: 46.5% (p<0.05) smoking and drinking among those with high PTSD: 52.9% (p<0.05)
Kern-Goldberger 2021	Illegal drug use	n=73,096; 0.20% AI/AN	non-Hispanic White, non-Hispanic Black, Hispanic, Asian, and Unknown race	Prevalence of illegal drug use with chi-squared test for association with race: 4.3% (p<0.01) Prevalence of alcohol use with chi-squared test for association with race: 1.4% (p<0.01) Prevalence of cigarette use with chi-squared test for association with race: 15.0% (p<0.01)
Khan 2013	Alcohol use and smoking	n=125; 100% AI/AN	NA	Prevalence of binge drinking and chi-squared test for association with smoking during pregnancy or during the month prior to pregnancy: smoking and drinking 36.07%; smoking no drinking 63.93%; drinking no smoking 4.69%; no smoking or drinking 95.31% (p<0.0001)
Kim 2010	Tobacco use and smoking	n=5,458; 28% AI/AN	AI/AN	Adjusted odds ratio of quitting during pregnancy among White people who smoked but did not using chewing tobacco before pregnancy: 2.30 (1.75, 3.03)
			White#	Prevalence of: any tobacco use during pregnancy with chi-squared test for association with race: 43.4% (0.9 SE) (p<0.0001)
			NA	any tobacco use during pregnancy: 43.40%
				smokeless tobacco use during pregnancy: 14.60%
				smoking during pregnancy: 24.90%
				quitting during pregnancy among those who only smoked cigarettes: 35.7% (1.4 SE)
				quitting during pregnancy among those who only used smokeless tobacco: 15.7% (1.7 SE)
				quitting during pregnancy among those who smoked and used chewing tobacco: 24.1% (3.1 SE)
				quitting during pregnancy among those who used any form of tobacco: 28.9% (1.0 SE)
Okah 2014	Health-compromising behaviors (tobacco, alcohol and drug use)	n=137,374; <4% AI/AN	non-Hispanic White	Adjusted odds ratio of health-compromising behaviors during pregnancy: 1.27 (1.08 1.50)
Specker 2018	Tobacco and alcohol use, smoking	n=813; 31.20% AI/AN	White#	Adjusted odds ratio of quitting smoking during pregnancy: 3.60 (1.75, 7.44)
			NA	Prevalence of: smoking during pregnancy: 27.10% quitting smoking during pregnancy: 66.1% (58.7, 73.6) alcohol consumption in the last 3 months of pregnancy: 4.6% (2.2, 7.0)
Sullender 2020	Smoking	n=402,252; 1.10% AI/AN	NA	Prevalence of smoking during pregnancy: 38.6% (36.7, 40.4)
			NA	Prevalence of smoking during pregnancy by education: Less than HS: 51.6% HS degree: 41.2% Some college: 32.7% College or more: 11.2%
			NA	Prevalence of smoking during pregnancy by year: 2001: 37.7% 2011: 40.7% 2016: 36.1%
Tabet 2017	Smoking	n=214,686; 1.14% AI/AN	non-Hispanic White, Black, Asian, Hispanic	Average number of cigarettes per day with t-test for association with race: 0.94 (2.86 SD) (p<0.05)
Watt 2012	Smoking and alcohol use	n=4,720; 2.35% AI/AN	non-Hispanic White, African American, Hispanic, Asian, Other race	Prevalence of cigarette use in the past month with chi-squared test of association with race: 29.6% (p<0.001)
			non-Hispanic White, African American, Hispanic, Asian, Other race	Prevalence of alcohol use in the past month with chi-squared test of association with race: 8.7% (p<0.001)
			non-Hispanic White	Adjusted odds ratio of cigarette use in the past month: 0.489 (0.010 SE) (p<0.001)
			non-Hispanic White	Adjusted odds ratio of alcohol use in the past month: 0.921 (0.015 SE) (p<0.001)
Ye 2020	Alcohol use	n=4,877; 44% AI/AN	White#	Adjusted odds ratio of binge drinking: 1.38 (1.05, 1.83)
			Those with no frequent change in residence	Adjusted odds ratio of drinking among those with frequent change in residence: 1.38 (1.16, 1.64)
			Those with no frequent change in residence	Adjusted odds ratio of binge drinking among those with frequent change in residence: 1.29 (1.08, 1.54)
Zamora-Kapoor 2016	Smoking	n=71,080; 10.12% AI/AN	NA	Prevalence of smoking: 19.70%

^*All estimates refer to behaviors among AI/AN peoples during pregnancy unless otherwise specified

^†statistical significance set at p<0.05

^$in studies that reported results from multiple models, we present results for the model with the full set of covariates.

^#Study either does not specify whether White reference group was Hispanic or non-Hispanic or included both Hispanic and non-Hispanic people in the White reference group.

Fewer studies estimate the prevalence of alcohol use during pregnancy. Among the studies in this review, alcohol use estimates range from 1.4% from 1999-2002 (data from 19 medical centers through the U.S.) (Kern-Goldberger et al. 2021) to 8.7% nationally from 2005-2009 (Watt 2012). One study provides an estimate of the prevalence of illegal drug use during pregnancy among AI/AN people, estimating it to be 4.3% from 1999 to 2002 (Kern-Goldberger et al. 2021).

The importance of understanding the social determinants of substance use among pregnant AI/AN people was emphasized in these studies. Factors such as family history, housing, education, and mental health are shown to be associated with engaging in tobacco, alcohol, or drug use during pregnancy. For example, housing instability and family-related trauma is related to an increased odds of drug use during pregnancy in a study of four rural reservations in the Southwest (Barlow 2010). Similarly, another study focused on American Indians in the Northern Plains, reports that frequently moving houses is associated with a significantly higher odds of alcohol use (Ye 2020).

Smoking during pregnancy varies by education; a descriptive study found that in the state of Wisconsin, the prevalence of smoking during pregnancy was higher for AI/AN people with a high school degree or less than for AI/AN people with at least some college (Sullender and Remington 2020). Intergenerational effects may also be present; in one unnamed tribal nation in the Great Plains region of North Dakota, Jorda (2021) finds significant associations between smoking during pregnancy and parental education and employment. These findings illuminate the importance of examining different risk factors for substance use within American Indian and Alaska Native communities instead of merely comparing the health behaviors of AI/AN people with those of other population groups.

4.3. Mechanisms of AI/AN Maternal Health

How does the existing literature describe the factors that contribute to the maternal health of AI/AN people? In the following sections, we summarize how articles that specifically focused on AI/AN maternal health (n=19) explain their findings. We adopt a simplified version of the social ecological model proposed by Spencer and Grace (2016) and differentiate between individual and structural/institutional explanations for health outcomes. This model proposes that people are “nested in a series of increasingly broader networks and systems, all of which work together to influence health” (Spencer and Grace, 2016, p. 104). The individual factors in this model include patient characteristics such as social class, age, and race, and interpersonal factors including social support, family dynamics, and cultural capital. Structural factors include characteristics of the provision of healthcare, and state and policy dynamics. Although we separate these categories to better synthesize our findings, individual and structural factors are interconnected and cannot be adequately understood in isolation. For instance, while socioeconomic status is measured at the individual level, it stems from historical events and is perpetuated by structural factors (Jones 2000).

4.3.1. Individual Factors

Authors proposed several individual-level mechanisms that contribute to the maternal health of AI/AN people, including socioeconomic status, individual health behaviors, and cultural beliefs. Socioeconomic factors, such as income, employment, insurance, and education were often cited as reasons for high rates of poor maternal health among AI/AN people and AI/AN-White disparities in maternal health. In one example, Watt (2012) suggests that socioeconomic status can account for non-Hispanic White-AI/AN differences in tobacco use during pregnancy, positing that “inclusion of social structural strains completely accounts for the difference” between AI/AN people and non-Hispanic White people in rates of smoking during pregnancy (p. 271). Elaborating further, Watt (2012) writes that “once controls are included, American Indian/Alaska Native women are considerably less likely to smoke during their pregnancy than White women” (p. 271). Several other studies similarly suggest that socioeconomic status plays a role in shaping AI/AN health behaviors during pregnancy (Danielson et al. 2018; Specker et al. 2018).

Many studies pointed to “individual behavior” as the mechanism responsible for maternal health. Several authors suggest that substance use earlier in life may drive patterns of maternal health. For instance, Barlow et al. (2010) find that early marijuana use is correlated with using meth during pregnancy. Hanson (2012) also finds that substance use may exacerbate barriers to prenatal care, particularly among pregnant people in rural areas (p. 34). Other authors propose that behavioral risks are mechanisms of other maternal health outcomes, such as ectopic pregnancy (de Ravello et al. 2015) and preeclampsia (Zamora-Kapoor et al. 2016).

Lastly, a few studies focus on cultural beliefs as a possible mechanism for maternal health, suggesting several pathways. Barlow et al. (2010) finds that culture may be a protective factor among pregnant AI/AN people; they find that strong cultural identification is associated with lower usage of meth and other substances during pregnancy (p. 18). In contrast, Kim et al. (2010) proposes that cultural norms of using smokeless tobacco might be partially responsible for comparatively low rates of quitting smokeless tobacco use during pregnancy among AI/AN people. However, the authors did not directly collect data to support this assertion.

4.3.2. Structural Factors

Studies also highlighted structural barriers to prenatal care and maternal health services, particularly geographic access to care and care quality, suggesting that AI/AN people experience disproportionate barriers to health care (Hadley et al. 2021; Hanson 2012; Jorda 2021; Kozhimannil et al. 2020). Several authors proposed that geographic inequality of obstetric care might inhibit maternal health among AI/AN people in rural areas. For example, Kozhimannil et al. (2020) note that “access to obstetrics services is declining in rural areas” (p. 299). Contextualizing their findings, Hadley et al. (2021) find that rural residence is associated with a higher risk of postpartum hemorrhage in the state of Alaska and point to “variations in prenatal care or differences in accessibility to maternal care in more remote regions in the state, highlighting a need for ongoing assessment of delivery of maternal care to rural Alaska” (p. 293). In a study of the Northern Plains region, evidence of long wait times and transportation issues prevented timely uptake of prenatal care (Hanson 2012).

Though these studies highlight important barriers to obstetric care, low uptake of institutionalized healthcare among AI/AN people does not universally indicate access issues or poor health. Instead, lower obstetric care rates may reflect preferences for in-home or community-centered pregnancy care (Carraher-Kang 2020; Schwarzburg 2013). Discussion of this critical perspective on prenatal care and other forms of obstetric care is largely absent from the literature we reviewed.

Maternal health of AI/AN people may also be affected by low-quality healthcare. A number of studies noted a relative paucity of culturally appropriate care for AI/AN people (Hanson 2012; Kim et al. 2010; Specker et al. 2018). In addition to a lack of consistency in providers, Hanson (2012) finds that prenatal care for AI/AN people in the Northern Plains lacks “holistic, women-centered culturally informed care” (p. 34). Culturally appropriate care may also reduce tobacco use during pregnancy. Kim et al. (2010) write that “culturally appropriate tobacco control approaches particularly targeting AN people may improve cessation during pregnancy and help prevent relapse.” (p. 371). Specker et al. (2018) also suggests that prenatal care offers smoking cessation counseling.

Discussions of how the historical oppression of AI/AN peoples may affect maternal health are largely absent from this literature. Overall, while the studies reviewed raised critical structural issues related to maternal care, most focused on either health service delivery or culturally appropriate care. However, one study calls for attention to tribal culture and history to improve the maternal health of AI/AN people. Kozhimannil et al (2020) write: “many tribes have deep traditions of healing and a broad understanding of both personal and historical trauma, which could helpfully inform strategies for Indigenous women during pregnancy” (p.299).

5. Discussion

5.1. Key Takeaways on AI/AN Maternal Health

Research and data on the maternal health of pregnant American Indian and Alaska Native people is relatively sparse. While existing studies suggest general patterns, the small amount of research on this topic prevents a nuanced analysis of tribal variation or concrete conclusions about the state of maternal health of AI/AN people at the national level. At best, they are a starting point for further research. We summarize key take-aways from the studies reviewed.

Overall, AI/AN people have high rates of maternal morbidity and mortality, particularly compared to non-Hispanic White people. Nationally, estimates of the number of pregnancy-related deaths per 100,000 live births range from 28.8 from 2011-2016 to 32.5 from 2011-2015 (Merkt et al. 2021; Petersen et al. 2019). Estimates of adjusted risk ratios for SMM (comparing AI/AN people to non-Hispanic White people) range from 1.29 in California from 1997-2014 to 1.8 nationally from 2012-2015 (Leonard et al. 2019; Kozhimannil et al. 2020).

Across the studies in this review, relatively high percentages of AI/AN people experience prenatal substance use, gestational hypertension, gestational diabetes, and cesarean sections. However, some articles note that significant differences by race and ethnicity largely disappear after including social and economic controls. While this framing is intended to elucidate mediators, it does not diminish the first-order importance of systemic injustice that disadvantages the maternal health and lives of AI/AN people. These very controls, and their associations, are products of economic marginalization, racism, and other structural factors. If only to underscore this point, AI/AN people tend to have higher rates of maternal mortality and SMM with or without “adjustment” through inclusion of other variables.

Importantly, the studies we capture suggest heterogeneity within the American Indian/Alaska Native population that is perhaps of equal or greater magnitude than disparities between AI/AN people and other population groups. However, given the wide variation in measures and contexts in the existing body of research, it is difficult to thoroughly assess this heterogeneity.

Finally, though multiple studies proposed structural mechanisms that might exacerbate poor maternal health outcomes, these mechanisms were often not evaluated empirically. Rather, most studies focused on, and empirically investigated, individual factors contributing to maternal health, such as socioeconomic status, cultural beliefs, and behavioral risk factors. Yet, transportation infrastructure, insurance provisions, family welfare systems, prenatal care infrastructure, and racism, among others, are critical upstream factors that deserve further attention. Notably, despite the colonialist history in which AI/AN health is situated, few studies discussed the role of historical context in the maternal health of AI/AN women.

5.2. Gaps in the Literature and Future Research

More research is needed to assess the prevalence and trends of AI/AN maternal health outcomes at the national, state, and tribal levels. Nationwide data is important to provide overall levels and trends of health outcomes and health disparities. Such national data is relevant to inform policy making at the federal level. At the same time, due to small sample sizes, national data tend to present American Indians and Alaska Natives as a homogenous group and studies can be underpowered to provide health outcome estimates. This was particularly notable for maternal mortality estimates. For instance, Palladino and colleagues (2011) could not draw conclusions about pregnancy-associated homicide or intimate-partner violence among AI/AN people due to small samples, and Leonard and colleagues (2019) could not report trend data for AI/AN maternal morbidity for the same reason. Future research that oversamples AI/AN people might help resolve this problem.

Though many studies in our review provided estimates of morbidity and mortality for smaller geographic regions (see Table 1), the measures, data sources, locations, and years vary dramatically across studies. Ongoing research focused on particular tribal nations or geographic regions would allow for an understanding of variation within AI/AN subpopulations and over time. AI/AN people live in diverse regions and have diverse languages, cultures, and health needs. Thus, more consideration of heterogeneity through state-level and regional/tribal data, as well as tribal-specific research, would be valuable.

In nearly every study that presented odds or risk ratios of maternal health outcomes, the reference group was non-Hispanic White people. While such information is useful, only one study in our review directly compared or discussed maternal health outcomes between AI/AN people and people identifying with other racialized groups (Singh et al. 2018 uses Chinese people as a reference group for several analyses). Future research that examines the comparative maternal health experiences of, for example, AI/AN people and Black people, or that considers the connections between the racialized experiences of undocumented or migrant pregnant people and AI/AN peoples, may contribute to new understandings of the diverse factors contributing to maternal health in the U.S. Further, though more research that engages racism instead of race is urgently needed, it will be important that such research considers how racialized care paths in maternal health vary across communities. Scholarship on intersectionality, syndemics, and critical race theory would provide important insights to these efforts.

Finally, our analysis of mechanisms suggests a confluence of individual and structural factors potentially relevant to the maternal health of AI/AN people. Some of these mechanisms were empirically supported by the studies at hand, but more research is needed. In particular, many studies employed mechanisms that focused on individual risk factors, such as socioeconomic status (e.g., education, insurance, housing, poverty). Though these are important, the literature would benefit from a more thorough examination of the upstream historical and contemporary structures that shape AI/AN maternal health, many of which are related or are root causes of the individual characteristics captured in these studies (Short and Mollborn 2015). Finally, a few studies noted cultural factors that were potentially beneficial for maternal health. More research on protective factors at the individual and structural levels would bring depth and nuance to the study of AI/AN maternal health experiences. In-depth qualitative research, notably absent in this literature, would be especially useful for exploring the interconnected forces shaping the maternal health experiences of AI/AN people.

Comparative research might also enrich our understanding of these mechanisms. Indigenous communities across North America are subjected to the enduring effects of settler colonialism (Bonds and Inwood 2016; Wolfe 2006) with consequences for health and well-being, including pregnancy-related health (Palacios and Portillo 2009; Paradies 2016). We focus on populations indigenous to the U.S. in this review. However, the literature on Indigenous maternal health would benefit from a comparative study of the U.S. and Canada for several reasons. First, AI/AN populations are not defined exclusively by their relationship to the U.S., and many tribal nations cross the U.S.-Canada border. Second, Indigenous peoples in Canada were also colonized and experience ongoing oppression and health effects stemming from that history, albeit in distinct ways (Burnett et al. 2020; Eni et al. 2021). Third, the provision of healthcare and maternity care is markedly different in the two countries (Ridic, Gleason, and Ridic 2012; United Nations Department of Economic and Social Affairs 2018). Similar to the U.S., in Canada, there is evidence that First Nations, Inuit, and Métis people are at a higher risk of poor birth outcomes than non-Indigenous Canadians (Oliveira et al. 2013; Sharma et al. 2016; Sheppard et al. 2017). Future studies that compare and contrast the patterns and drivers of maternal health of Indigenous peoples in the U.S. and Canada could elucidate otherwise-hidden structural mechanisms related to the maternal health of Indigenous people in both countries.

The legacy of colonialism is relevant to the maternal health of Indigenous people in other parts of the world as well, such as Australia and New Zealand. Indigenous populations in these areas also experience disproportionately high rates of maternal mortality and morbidity (Australian Institute of Health and Welfare; Dawson et al. 2022) and research suggests that factors such as the racist legacies of colonialism may act as barriers to equitable maternal health outcomes (Dawson et al. 2019; Gordon et al. 2022; Sharma et al. 2016). In response to evidence of the lasting health effects of colonization, the World Health Organization calls for attention to the unique structural contexts that shape the health of Indigenous peoples globally (Commission on Social Determinants of Health 2008).

In sum, to understand more fully the maternal health of AI/AN people, we need more research on specific AI/AN tribal nations in combination with robust evidence on AI/AN health at the national level, with special attention to the drivers of maternal health outcomes. To address this data gap, scholars in sociology and public health have begun important work on the data needs of Indigenous populations. Carroll, Rodriguez-Lonebear, and Martinez (2019) call for Indigenous data sovereignty and governance, positioning AI/AN people at the helm of data collection. The authors argue that data sovereignty is necessary to “protect [Indigenous] cultural knowledge and sustain their way of life for past, present, and future generations” (Carroll, Rodriguez-Lonebear, and Martinez 2019, p. 6). Similarly, maternal health research would benefit from greater incorporation of Indigenous research principles which are generally lacking in this sphere, such as Indigenous conceptualizations of health and consideration of colonialism as a determinant of health (Patterson et al. 2022). One example of indigenous data sovereignty includes the First Nations regional health survey in Canada (First Nations Information Governance Centre 2022). As another route to improve data collection on AI/AN maternal health, national maternal mortality review committees might collect data on Indigenous identity and ensure Indigenous expertise is represented in all phases of review.

5.4. Limitations of this Review

This review was based on a replicable, broad search which resulted in a diverse sample of articles. We note that this search did not include gray literature, state-published data, or tribal literature. Additionally, the systematic literature search was restricted to English language articles, which would exclude any peer-reviewed literature in indigenous or other languages. Still, this review identifies a critical gap in knowledge about the maternal health of AI/AN people and provides a starting point for future work.

6. Conclusion

American Indian and Alaska Native people are largely underrepresented in research on maternal health, despite preliminary evidence that they experience high rates of severe maternal morbidity and mortality compared to non-Hispanic White people. Our findings suggest that their maternal health varies substantially by region, Tribe, and health measure. Multiple individual, structural, and historical factors are proposed as drivers of these maternal health outcomes. We conclude that research on the maternal health of American Indians and Alaska Natives at the national and tribal levels is urgently needed, including research that collects data on the use of health services in non-IHS settings. Research that adopts diverse methodologies and incorporates the perspectives and knowledge of AI/AN pregnant people and communities is essential. Qualitative research methods, including ethnography, in-depth interviews, and participant observation would greatly benefit our understanding of the mechanisms underlying patterns of maternal health outcomes. Additionally, research that moves past individual risk and incorporates how structural factors and historical contexts have contributed to Indigenous maternal health in the U.S. is vital.

• The maternal health of AI/AN people is understudied

• AI/AN people have high rates of maternal mortality and severe maternal morbidity

• Most studies focused on individual-level predictors of maternal health outcomes

• Few studies examined the structural or historical determinants of maternal health

• More research should pay attention to the harmful legacies of colonization