Protective and risk factors of suicide in Native American youth: cross-sectional findings from a mixed-methods study

Summary

Background

Native American (American Indian/Alaska Native) youth have the highest rates of suicide and suicide clusters in the United States, and appropriate responses are lacking. This study examined protective and risk factors for suicide in a remote Northern Plains reservation with a recent cluster.

Methods

A mixed-methods study was conducted on the Fort Belknap Indian Reservation, Montana, in 2022 using a community-engaged approach. Quantitative results from the 251-item questionnaire are presented, which were derived from a culturally adapted socio-ecological model. Aaniiih and Nakoda youth 14–24 years of age, living on/near Fort Belknap were recruited.

Findings

In total, 197 youth completed the questionnaire; average age was 16.6 years (SD = 2.60). One hundred (51%) youth were female and 90 (46%) male; 6.1% of data points were missing. Positive family relationships, communal mastery, benevolent childhood experiences, and emotional intelligence were protective, decreasing odds of suicide ideation. Childhood and family risk factors correlated with increased odds of suicidal ideation, including verbal abuse, low community support, sexual abuse, and witnessing community violence. Utilizing multivariable modeling, emotional intelligence (protective factor), and early initiation of substance use, verbal abuse, PTSD, and historical losses (risk factors) were significant factors for suicide ideation, when controlling for all other factors.

Interpretation

Adverse childhood experiences, early initiation of substance use, and PTSD increase suicide ideation and attempts. When youth experience positive and supportive family relationships, odds of suicide ideation are lower, which highlight pivotal areas for intervention and response.

Funding

William T. Grant and Doris Duke Family Foundations.

Research in context.

Evidence before this study

• -The US Government report, Regional Differences in Indian Health—2012 Edition provided foundational evidence for the present study, which highlights differences in suicide and poverty among Native American communities in the US.
• -Our team conducted a similar study on the neighboring Fort Peck Reservation in 2011. Study methods informed the present study, particularly regarding measure selection.
• -The Fort Belknap Reservation (study setting) experienced a youth suicide cluster in 2019, which was the driving force behind the present study. Additionally, the Principal Investigator is an enrolled member of the A'aniniiin Nation and has the relationships and support of community members to carry out this research.

Added value of this study

• -This study adds to the growing literature around suicide and mental health disparities via research done with a sample of youth from a reservation-based community with one of the highest poverty rates in the US and a recent suicide cluster.

Implications of all the available evidence

• -Given the high rate of suicide, targeted policies and programs should prioritize populations with the highest need. By implementing these strategies, we can better address the mental health needs of Native American youth and communities, ultimately reducing the high rate of suicide.
• -Expanded education and training opportunities could include scholarships and loans for Native American students pursuing tailored mental health training; targeted outreach at Tribal Colleges and Universities on mental health-related topics; and grant writing opportunities in every reservation area via tribal-academic partnerships.
• -Tailored funding opportunities could include offering technical assistance for grant applications, particularly for mechanisms that prioritize Native youths' health and wellbeing.
• -Enhanced service provision may include providing up-to-date reporting on reservation conditions; enhanced telehealth services for increased mental health provision; and prioritizing job recruitment programs for Native American communities.

Introduction

Suicide is a global public health problem. More than 700,000 people die annually by suicide, with 77% of suicides occurring in low to middle income countries.¹ Suicide rates among Native Americans (NA) (American Indian/Alaska Native) significantly increased between 2015 and 2020, from 20 per 100,000 persons to 23.9, compared to less than a 1% increase among the overall United States (U.S.) population.² In general, suicide rates vary by geographic location: the Alaska Area rate (highest at 41.5) is 5.2 times the Nashville Area (lowest at 8.0), across the 12 Indian Health Service geographical areas.³ NA youth, specifically, experience the highest burden of suicide of any racial/ethnic group in the U.S., and many of them live in more remote areas.⁴ Remote NA communities ultimately bear the burden of suicide, which is also where suicide point clusters are more likely to occur—i.e., a greater number of suicides or suicide attempts in the same timeframe and geographical area.⁵

Little is known about why suicide clusters form and at times—persist. However, research has shown that suicide clusters are more prevalent in rural reservations and/or villages,⁶ which have limited access to mental health services and mental and behavioral health providers.⁷ In general, NA youth who have spent at least two-thirds (66%) of their lives on a reservation are at an increased risk for suicide ideation (SI) and suicide attempts (SA) compared to their peers who live in urban areas.⁸ Additionally, 15–19-year-old youths who are exposed to suicide(s) are at a 2–4 times higher risk for SI and SA than any other age group.⁹

Reservation communities established by treaty (1850–1887) have been particularly shaped by both historic and current policies that create disadvantages and have significant implications for life expectancy, morbidity, and mortality. These reservations typically represent census tracts characterized by enduring and concentrated poverty,¹⁰ which is a key social determinant of health—shaping the conditions in which people are born, grow, work, live, and age. These factors account for 30–55% of health outcomes¹¹ and are evidence of a pervasive health crisis linked to systematic disadvantage for over a century. Concentrated poverty areas are defined by census tracks with an overall poverty rate at 30% or more.¹² Considering this root cause of disadvantage—and its sustained impact on youth mental and behavioral health—is critical to affecting NA youth suicide outcomes.

Protective and risk factors

The high incidence of NA youth suicide has been well documented for over 40 years.¹³ However, there is limited understanding of the communities with the highest rates, and what differences exist between those with higher versus lower incidence rates of suicide.¹⁴ Protective factors have been elucidated through prior research and are significantly associated with a reduction in suicide risk for NA youth. These factors include tribal identity, school engagement, communal mastery, positive family relationships, and family and community connectedness.15, 16, 17 Risk factors associated with a higher likelihood of experiencing SI/SA include post-traumatic stress disorder (PTSD), depression, childhood trauma, substance use, discrimination, low self-esteem, and impulsivity.^15,17,18

Like other NA reservations established during the reservation era, the Fort Belknap Indian Reservation in Montana, U.S.A., is an area of concentrated poverty and deep disadvantage. In July 2019, after three youth suicides occurred in one month, the Fort Belknap Community Council declared a State of Emergency regarding youth suicide.^19,20 Therefore, the purpose of this study was to investigate protective and risk factors for SI and SA among Aaniiih and Nakoda youth 14–24 years of age, in a rural, reservation community with a recent suicide cluster.

Methods

Study design and participants

Results presented include quantitative findings from a larger mixed-methods study. We administered a cross-sectional survey to a convenience sample of Aaniiih and Nakoda youth. An Indigenous-led research team recruited participants from all four reservation communities in November and December 2022. A Tribal Advisory Board was established, comprised of spiritual leaders, those who work or have worked in suicide prevention efforts, teachers, and behavioral health experts. They provided invaluable guidance and recommendations throughout the research process, including approving the survey questionnaire and interview guides and recommending safety measures. The study and all related activities were guided by an adapted socio-ecological model developed through formative work previously conducted in the Fort Belknap community.²¹ Importantly, the study followed tribal protocols, including the research team participating in a ciibáakʔi/Įnį́bi (sweat lodge), before and after data collection, in which team and community members prayed for the safety of participants and data collection team. As part of debrief, participants were offered ʔiikyɔ́ɔ́ɔ́hʔɔ́ɔʔ/Wizítgiya (smudge) upon survey completion. Smudging is a traditional activity that removes negative thoughts, feelings, and fosters positive energy by burning plant medicines (e.g., sage, sweetgrass, cedar, and sweetpine) to create smoke that cleanses the air and those within it.²² Additionally, team check-ins and prioritization of self-care were strongly recommended for all research team members, further providing additional team support when collecting sensitive data. Additionally, our team included two Psychiatric and Mental Health Nurse Practitioners that supported screening, debriefing, and referral to local mental health services.

Study setting

The Fort Belknap Indian Reservation was established by treaty in 1888 and is governed by the elected Fort Belknap Community Council. Located in northcentral Montana, U.S.A., Fort Belknap spans 675,147 acres and is home to approximately 8654 enrolled Aaniiih and Nakoda members.^23,24 There are four distinct communities in Fort Belknap (Hays, Lodgepole, Agency/Harlem, and Dodson). Fort Belknap is extremely rural with an average annual unemployment rate of 64%²⁴; the average unemployment rate among NAs in the U.S. is 10.5%.²⁵ Forced relocation and government control of resources has contributed to pervasive poverty and unemployment, which characterize reservations in the U.S. Today, 36% of Fort Belknap residents live below the federal poverty level, and it increases to 50% for children under 18 years of age.²⁴ In the Billings Area, where Fort Belknap is located, the Indian Health Service provides care to over 67,000 NAs. A 2013 report from the Montana Department of Health and Human Services found that NA men (56 years), live 19 fewer years than non-NA men (75 years), and NA women (62 years) live 20 years fewer than non-NA women (82 years).²⁶

Recruitment

In spring 2022, community engagement forums were held in each of the communities, allowing community members to engage in the research process, and ask questions while sharing a meal with study staff, a demonstration of tribal etiquette. The onsite Research Coordinator, along with “community champions” led these efforts. Recruitment efforts for study participants included word-of-mouth, flyers, and radio public service announcements. Youth were eligible to participate if they were 14–24 years of age, identified as a member of the Fort Belknap Community, lived within 1-h of the Fort Belknap Indian Health Service Unit, and had parental consent if younger than 18 years. Written parental consent and assent were achieved for participants under 18 years; written consent was achieved for participants 19–24 years.

Ethical considerations

Tribal Resolution (#182–2020) by the governing Fort Belknap Community Council authorized this study. Human subjects approval was provided by Johns Hopkins School of Medicine (00305879) and Fort Belknap's Aaniiih Nakoda College Institutional Review Boards. The Tribal Advisory Board guided the study, informing and approving concepts and measures for inclusion in the study, and then reviewed and approved this report.

Measures and outcomes

Measures were selected based upon our conceptual model, which is published elsewhere²¹ and details factors by socio-ecological level. Strong psychometric properties and prior utilization in NA communities were also factored into measure selection. SI (with or without attempt) was the primary outcome variable based on a single item, “Have you ever had serious thoughts that you would be better off dead or of hurting yourself in some way?” SA was a secondary outcome measured with a single item “Have you ever tried to kill yourself?” Both questions are from the Columbia Suicide Severity Rating Scale (C-SSRS),²⁷ and any “yes” response warrants clinical referral. All measures were self-reported and are detailed below by protective versus risk factors. Cronbach's alpha values reported below are for the current study sample.

Protective factors

• 1.Tribal Identity was assessed using the mean of the 6-item Orthogonal Cultural Identification Scale with a 4-point response scale (0-None to 3-A lot).²⁸ Cronbach's alpha was 0.9, and it was included as a continuous variable.
• 2.Emotional Intelligence was assessed via mean score from the 33-item Schutte Self Report Emotional Intelligence Test (SSEIT) with a 5-point response scale (1-Strongly disagree to 5-Strongly agree).²⁹ Cronbach's alpha was 0.92, and it was included as a continuous variable.
• 3.Communal Mastery, or group efficacy, was assessed using the mean score from the 10-item Communal Mastery Scale with a 4-point response scale (0-Strongly disagree to 3-Strongly agree).³⁰ Cronbach's alpha was 0.77, and it was included as a continuous variable.
• 4.Positive Family Relationships were measured using the mean from the 25-item Brief Family Relationship Scale with a 3-point response scale (1-Not at all to 3-A lot).³¹ Cronbach's alpha was 0.95, and it was included as a continuous variable.
• 5.Benevolent Childhood Experiences (BCEs) were measured by the Benevolent Childhood Experiences Scale which assesses feelings of safety, support, and predictability.³² The score is the total count of 10 possible positive childhood experiences that a participant endorses prior to the age of 18 years. Cronbach's alpha was 0.72, and it was included as a continuous variable.

Risk factors

• 1.Depression was assessed using the summed score of the 10-item Revised Center for Epidemiologic Studies Depression Scale (CESD-R).³³ Scores of 10 or above indicate clinical levels of depression requiring additional clinical support. Cronbach's alpha was 0.84, and it was included as a continuous variable.
• 2.Post-Traumatic Stress Disorder was assessed using the mean score from the 4-item post-traumatic stress disorder checklist (PCL)³⁴ with a 5-point response scale (0-Not at all to 4-Extremely). Scores of 8 or higher indicate moderate active trauma symptoms, which may require additional clinical support. Cronbach's alpha was 0.83, and it was included as a continuous variable.
• 3.Traumatic Grief was assessed via the mean frequency of respondents' traumatic grief experiences as reported on the Traumatic Grief Inventory-Self Report (TGI-SR)³⁵ The subscale consisted of 18 items with a 5-point response scale (1-Never to 5-Always); Cronbach's alpha was 0.94, and it was included as a continuous variable.
• 4.Cyberbullying Victimization including visual, verbal, and social cyberbullying was measured by the cyberbullying victimization 27-item scale with a 5-point response scale (1-Not at all to 5-Very often) as mean scores.³⁶ Cronbach's alpha was 0.96, and it was included as a continuous variable.
• 5.Racial Discrimination was measured using the 11-item Racial Discrimination Scale with a 3-point response scale (0-Never to 2-Always).³⁷ The scale was originally developed to assess discrimination experienced by NAs. The mean of the items was used, and Cronbach's alpha was 0.82, and it was included as a continuous variable.
• 6.Historical Losses was measured via the mean frequency of historical losses that directly impacted the respondents. Four items from the Historical Loss Scale were included that assess losses from the effects of alcoholism, the loss of respect by children and grandchildren for Elders, the loss of people through early death, and the loss of respect by children for traditional ways and included a 6-point response scale (0-Several times a day to 5-Never).³⁸ Cronbach's alpha was 0.92, and it was included as a continuous variable.
• 7.Adverse Childhood Experiences (ACEs) were collected based on the Philadelphia ACE Questionnaire.³⁹ Mean frequencies (all of the time, most of the time, some of the time, none of the time) were measured for lack of community support, hearing community violence, experiencing physical abuse, experiencing sexual violence, experiencing verbal abuse, experiencing challenging family situations, witnessing physical abuse, and witnessing verbal abuse during childhood. This measure was included as a continuous variable.
• 8.Binge drinking, defined as four or more alcoholic drinks during a single occasion for females and five or more alcoholic drinks during a single occasion for males based on guidelines from the Centers for Disease Control and Prevention, was included as a binary variable. This was categorized as low risk, fewer than four or more alcoholic drinks during a single occasion and high risk, higher frequency of four or more drinks during a single occasion.⁴⁰
• 9.Early initiation of substance use, defined as any substance use (smoking, vaping, marijuana, prescription pain medicine without a doctor's prescription, inhalants, methamphetamines, hallucinogenic drugs) at age 13 or younger, was included as a binary variable.⁴¹ This was categorized as low risk, 14 or older, and high risk, 13 or younger.
• 10.Polysubstance use, or the concurrent use of substances over the past 12 months, was included as a categorical variable (i.e., “none”, “any”, or “always”), and assessed using an adapted version of Midanik et al. Polysubstance Use Scale.⁴²

Data collection

Prior to implementation, the full survey was piloted with 11 participants to assess survey acceptability, usability, and cultural relevance. Detailed study procedures are further outlined in Brockie et al., 2023.²¹ Data collection was conducted in community centers and/or school computer labs. To promote confidentiality, participants were seated at every other computer and cellular phone usage was not allowed. REDCap, a secure web application for building and managing online surveys, was utilized to collect and store data.⁴³ The survey took approximately 1 h to complete. Study team members were available to complete consent/assent procedures, provide technological assistance as needed, and support debriefing after survey completion. Smudge, a culturally appropriate healing practice, a mental health resource card, and a referral, if warranted, was used as part of the debrief process. Participants received a $30 cash incentive for their knowledge and time contributions.

Statistical analyses

Frequencies were calculated for all categorical variables. Continuous predictors were standardized to have a mean of zero and a standard deviation of one prior to conducting the regression analyses, for improved interpretation and easier comparison among predictor variables keyed to different scales. Distributions including means, standard deviation, and skewness were assessed prior to standardizing the continuous variables. All protective and risk factors were tested as predictors of SI using bivariate logistic regression. Those found to be significant in bivariate logistic regressions were included in the final multivariable model. Area under the curve was used to assess the goodness-of-fit of the multivariable model. Among those with SI, we examined the predictors of SA versus SI without an attempt. Multiple imputation using Fully Conditional Specification, implemented by the MICE algorithm as described in Van Buuren and Groothuis-Oudshoorn,⁴⁴ was used to handle missing data with 10 imputed datasets. R version 4.3.0 (2023-04-21 ucrt) and STATA version 16.0 were used for all analyses in this study.⁴⁴ The final sample size for analyses was N = 197, therefore the study can detect moderate effect sizes (OR ≥ 2.20) in the bivariate regressions with power of 0.80 and alpha of 0.05.

Role of the funding source

The study sponsors had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Results

Two hundred participants consented, and 197 participants completed the survey (98.5%); 100 (51%) were female and 90 (46%) were male, respectively (Table 1). Since participants were not required to respond to all items, there was minimal missing data at the item level (6.1% of the data elements). The average age of participants was 16.6 years, and ages ranged from 14 to 24 years. Most participants, 172 (87%) reported their sexual orientation as Straight and 25 (13%) self-reported as Lesbian/Gay/Bisexual/Two-spirit (LGBT2S) or other sexual orientation. About 84% of participants reported being enrolled in school at the time of study. The average depression score for participants was 10.94 out of a possible range of 0–30, and 53% of participants had significant depressive symptoms (defined as participants with scores of 10 or above); 39% (n = 76) screened positive for PTSD. In total, 38 participants had responses on the CESD-R, PCL and/or Adapted CSSR-S scales that met the clinical threshold for risk and were referred for mental health services.

Table 1.

Demographic and clinical characteristics of participants.

	Total sample N = 197a	No suicide ideation or attempt N = 100a	Suicide ideation or attempt N = 97a	p-valueb
Age, mean (SD)	16.60 (2.60)	16.49 (2.48)	16.72 (2.73)	0.545
Enrolled in school	166 (84%)	85 (86%)	81 (84%)	0.771
School grade				0.764
Less than high school	157 (80%)	80 (86.7%)	77 (88%)
High school and above	23 (12%)	12 (13.3%)	11 (12%)
Missing	17 (8.6%)	8 (8.0%)	9 (9%)
Sex				<0.001
Girl/woman	100 (51%)	39 (39%)	61 (63%)
Boy/man	90 (46%)	60 (60%)	30 (31%)
Other/missing	7 (3%)	1 (1%)	6 (6%)
Sexual orientation				<0.001
Gay/bisexual/two-spirit/other	25 (13%)	5 (5.0%)	20 (21%)
Straight	172 (87%)	95 (95%)	77 (79%)
Depression, mean (SD)	10.94 (6.32)	7.49 (4.09)	14.47 (6.25)	<0.001
Significant depressive symptoms	105 (53.5%)	29 (29%)	76 (78%)	<0.001
PTSD, mean (SD)	6.51 (4.29)	4.33 (3.69)	8.75 (3.68)	<0.001
Screened positive for PTSD	76 (39%)	18 (18%)	58 (60%)	<0.001

Bold values indicate p = < 0.001.

^an (%); Mean (SD).

^bPearson's Chi-squared test; T-test. Sex (biological factors) and gender identity (psychosocial/cultural factors) are reported in accordance with the Sex and Gender Equity in Research (SAGER) guidelines.⁴⁵

Prevalence of SI and SA

Of all participants, 49% (n = 97) had SI with or without an attempt; 27% (n = 53) had SI without attempt; and 22% (n = 44) had at least one SA. Of those who had SI, 11% (n = 10) reported thoughts at the time of the survey, 31% (n = 28) reported having thoughts of suicide within the last 3 months, 34% (n = 30) had thoughts in the last 4–12 months, and 24% (n = 21) had thoughts of suicide more than a year ago from the time they completed the survey and eight participants did not respond to this question.

Protective factors for SI with or without an attempt

Fig. 1 summarizes the ORs from the bivariate analyses of protective and risk factors for SI.

Fig. 1.

Odds ratios for risk and protective factors of SI with or without an attempt.

Protective factors

A one-standard-deviation increase in: 1) positive family relationships decreased odds of SI by 67% (OR = 0.33, p < 0.001); 2) communal mastery decreased odds by 60% (OR = 0.40, p < 0.001); 3) BCEs decreased odds by 60% (OR = 0.40, p < 0.001); and 4) emotional intelligence decreased odds of SI by 53% (OR = 0.47, p < 0.001).

Risk factors for SI with or without an attempt

Demographic factors

Female participants had much higher odds (OR = 3.13, p < 0.001) of SI compared to male participants; 63% of females had SI and only 39% of males reported having it. Compared to those who reported Straight sexual orientation, those who reported being LGBT2S had greater odds of SI (OR = 4.94, p = 0.003), with 80% reporting SI or SA, compared to 45% of Straight participants (Fig. 1).

Childhood and family risk factors

Several types of ACEs were associated with increased risk of SI. Greater experiences of verbal abuse had the strongest association (OR = 3.74, p < 0.001), followed by experiences of physical abuse (OR = 3.32, p < 0.001). Reports of witnessing verbal abuse (OR = 2.41, p < 0.001), lack of community support (OR = 2.37, p < 0.001), experiencing sexual abuse (OR = 1.94, p < 0.001), and hearing community violence (OR = 1.81, p < 0.001) were all associated with increased odds of SI. Having a more challenging family situation (OR = 1.76, p < 0.001) and witnessing physical abuse (OR = 1.74, p = 0.001) were also associated with SI.

Other risk factors

PTSD was the strongest risk factor for SI (OR = 3.67, p < 0.001), followed by traumatic grief (OR = 2.39, p < 0.001). For each one standard deviation increase in PTSD, the odds of SI increased by 267%, while a one standard deviation increase in traumatic grief increased odds by 139%. Cyberbullying was associated with a 68% (OR = 1.68, p = 0.003) increase in the odds of SI while racial discrimination was associated with a 54% (OR = 1.54, p = 0.006) increase in the odds of SI. Historical losses were associated with 42% increased odds of SI (OR = 1.42, p = 0.018).

High risk substance use

Participants with high risk of early initiation of substances had 371% greater odds (OR = 4.71, p < 0.001) of SI. Those that reported always using polysubstance had higher odds of SI (OR = 2.82, p = 0.002) compared to those with no polysubstance use. High risk of binge drinking increased odds of SI by 114% (OR = 2.14, p = 0.026) compared to those with low risk of binge drinking.

Multivariable model

The multivariable model for suicide ideation included all the protective and risk factors identified as significant in the bivariate models (22 variables) (Supplement 1). The area under the curve for the multivariable model was 0.77, signifying a good fit of the model. As expected, the protective and risk factors were correlated with VIFs ranging from 1.26 to 2.76. Therefore, not all variables remained significant due to multicollinearity of the predictor variables. Variables that remained significant in the multivariable model included: early initiation of substance use (OR = 3.24, p = 0.031), experiences of verbal abuse as a child (OR = 2.84, p = 0.003), PTSD (OR = 2.21, p = 0.016), historical losses (OR = 1.92, p = 0.030), and emotional intelligence (OR = 0.44, p = 0.020).

Protective and risk factors for SA among those with SI (subgroup analysis)

Fig. 2 presents a forest plot summarizing the exploratory bivariate analyses of factors associated with a SA among the subset of participants who reported SI (N = 97). A one-standard-deviation increase in emotional intelligence decreased the odds of SA by 42% (OR = 0.58, p = 0.009), and a one standard deviation increase in positive family relationships decreased the odds of SA by 54% (OR = 0.46, p = 0.003) among those with SI. Witnessing physical abuse increased the odds by 121% (OR = 2.21, p = 0.001) while having challenging family situations increased the odds of SA by 71% (OR = 1.71, p = 0.001). Experiencing sexual abuse increased the odds of SA by 62% (OR = 1.62, p = 0.010). Like SI, a one-standard-deviation increase in BCEs decreased the odds of SA by 39% (OR = 0.61, p = 0.017) among those with SI.

Fig. 2.

Odds ratios of protective and risk factors for SA compared to SI without attempt.

Discussion

The current study identified factors that protect against and elevate risk for SI and SA in one reservation community that was established by treaty and recently experienced a youth suicide cluster, which builds upon findings from Brockie et al., 2015,¹⁵ which was conducted with the neighboring Fort Peck Reservation. We analyzed survey results from 197 Aaniiih and Nakoda youth ages 14–24 years. We included protective factors for suicide and associated factors identified as significant in other studies of NAs, such as Tribal identity and communal mastery,⁴⁶ and included additional protective factors such as emotional intelligence, positive family relationships, and BCEs to counter the overemphasis on risk factors in research of NA suicides.⁴⁷ Almost half the sample reported SI (49%), and 22% reported at least one suicide attempt.

Harnessing protective factors may reduce risk

Most importantly, there were myriad factors which were shown to decrease risk of SI and SA. Aaniiih and Nakoda youth with positive family relationships, more BCEs (often referred to as “counter ACEs”), and higher emotional intelligence were found to have lower risk for SI and SA. These findings are consistent with previous research demonstrating that strong family cohesion and connectedness are among the most robust protective factors against suicidal thoughts and behaviors in youth.^48,49 Particularly in Indigenous contexts, culturally grounded family support is essential for resilience and emotional regulation. Communal mastery was protective for SI only in this population. These findings for communal mastery are consistent with other studies utilizing a NA sample, including the Fort Peck Reservation, which found communal mastery to be protective against high-risk substance use.¹⁵ Positive early life experiences and cultural and tribal strengths are critical to the development and protection of NA children. These findings corroborate the growing movement to consider not only adversity but also resilience-building experiences across the lifespan; specifically, emotional intelligence,⁴⁹ being a core competency that has been found to protect youth from suicidal behavior.

Increasing personal, familial, community, and cultural strengths is our anticolonial remedy for alleviating social suffering in our community.

The ‘risks’ we speak of were by environmental design

Passing of the Indian Appropriations Act of 1852 into law initiated the Indian reservation system, establishing areas of concentrated poverty by intentional design—a blatant violation of human rights.⁵⁰ Violent theft of land is one of four forms of violence the International Justice Mission considers when investigating genocide.⁵¹ Similarly, “imposing living conditions intended to destroy the group” is one of the five articles of genocide acknowledged by the United Nations.⁵² Subsequent inequities in power, money and resources have led to inequities in everyday living conditions, including toxic stress.⁵³ This law in and of itself has contributed to expansive, intergenerational effects on poverty and violence exposure, and health, and education outcomes in reservation-based areas, and continued colonial subjugation is hurting our young people, which includes increasing their risk of suicide. In a national sample of youth, witnessing community violence was found to be associated with substance use and suicide risk.⁵⁴

Among our sample of Aaniiih and Nakoda youth, high risk substance use, including early onset of substance use, binge drinking, and polysubstance use significantly increased odds of SI. The age of onset of use and substance use results from a complex interplay of factors, inclusive of youths' environment and peer-to-peer relationships. Other factors including PTSD, traumatic grief, cyberbullying, and racial discrimination increased odds of SI in our sample. Growing literature underscores the impact of experiences of bullying victimization, particularly in adolescents, on emotional regulation—this, in turn, can contribute to greater risk behaviors.⁵⁵ Additionally, Whitbeck et al., 2022,⁵⁶ found that ongoing discrimination contributed to identity-related stress and despair in Indigenous youth. Thus, understanding the extent of these specific risk factors can expand conventional models of suicide risk to better reflect the lived experiences of Native populations.

Familial and household experiences are at the heart of it all

The disruption of families, communities, and entire Tribal nations that occurred by “the encroaching white civilization” has had lasting traumatic effects, particularly for those who endure living on reservations established by treaty. Tribal identity, school engagement, interpersonal safety, and traditional parenting practices have all been negatively affected. These situations and realities are often overlooked yet are often at the very heart of youths' challenges. We found ACEs, including experiencing verbal, physical, and sexual abuse; witnessing physical and verbal abuse; family challenges; lack of community support; and hearing violence occurring are associated with increased odds of SI. Among those who reported SI, experiences of sexual abuse, cyberbullying, and challenging family situations significantly increased odds of SA. Similar relationships between ACEs and suicide risk were reported in a study of rural, reservation-based NA youth.⁴⁶ These findings are also consistent with the Adolescent Brain Cognitive Development study which has demonstrated the compounding effects of early adversity on poor mental health outcomes later in life.⁵⁷ Similarly, in a national sample of youth, witnessing community violence was found to be associated with substance use and suicide risk.⁵⁴

Implications for policy and practice

The impact of the reservation experience on generations of NA children is best addressed by multi-level policy and practices changes, particularly for Tribes that have established treaties with the US Government. Since the release of the 1928 Meriam Report: The Problem of Indian Administration, there has been no report on the poor conditions of reservations established via treaty—a comprehensive update is long overdue. Additionally, Table 2 outlines recommendations for relevant policies and programs to consider in determining how to best support NA communities experiencing the greatest need.

Table 2.

Recommendations for enhancing policy and practice to support Native youth and communities.

1. Expand education and training opportunities

•Offer scholarships and loans for tailored mental health training for Native American students via formal and informal avenues. •Target outreach at Tribal Colleges and Universities on mental health-related topics, including suicide prevention, to reach tribal communities and children in school. •Identify and leverage any existing Department of Education resources that could support expanded education and training efforts for Native American students.

2. Tailor funding opportunities

•Offer technical assistance to apply for grant mechanisms that prioritize Native youths' health and wellbeing across various bureaus (e.g., Maternal and Child Health Bureau, Agency for Healthcare Research and Quality). •Provide grant-writing training in every reservation area, potentially through tribal-academic partnerships.

3. Enhance service access and provision

•Provide up-to-date reporting on the conditions and outcomes of reservations established by treaty. •Increase telehealth services for improved mental health service delivery, particularly in areas with the highest youth suicide rates. •Assess what Job Corps has learned about effective strategies and consider how these could be adapted and implemented in tribal communities. •Target new Job Corps programs towards Native American communities and/or prioritize these communities in existing programs.

Limitations

These results should be contextualized in terms of study limitations. First, the study was limited to a one reservation/two tribe convenience sample, thereby restricting broad generalizations to NA and Indigenous communities globally. It should be noted, however, that the Fort Belknap community is not unlike other rural reservation communities—especially those in the Northern Plains—in many ways (e.g., socioeconomic and disparate health outcomes). Second, causation cannot be inferred from these cross-sectional associations, particularly as all data were self-reported; thus, there is risk of recall and/or selection bias. Many factors were explored given the sample size and multi-collinearity among the predictor variables in the models limited robustness of the multivariable models. Comparisons across gender were based on a small sample size for LGBT2S individuals. Finally, the SA analysis had limited power due to the small subset of individuals who reported SA; therefore, conclusions should be interpreted cautiously.

Conclusion

Although NA youth experience the highest incidence of suicide of any racial or ethnic group in the U.S., disaggregated data reveals a 10-fold difference in suicide across geographies. This underscores the major impact that social determinants of health, like concentrated poverty, have on youth suicide. The current study found that odds of suicide risk decreased or increased across a range of protective and risk factors. Findings underscore the impact that historical violations of human rights—the literal formation of concentrated poverty—continues to have on the lives of reservation-based youth today. This work requires sustained engagement to develop and implement comprehensive strategies for addressing poverty, health, education, and safety—because our children are the future, and they deserve better.

Contributors

TB: Study conceptualization, study design, cultural expertise, data interpretation, and manuscript writing.

JPG: Study design, cultural expertise, data interpretation, and manuscript writing.

KEN: Study design, data collection, data interpretation, and manuscript writing.

ASM: Study conceptualization and data interpretation.

MKJ: Study design, cultural expertise, data collection, and data interpretation.

WO: Accessed and verified the data, Data analysis, data interpretation, tables & figures, manuscript writing.

NP: Accessed and verified the data, study design, data analysis, data interpretation, tables & figures, and manuscript writing.

All authors: Provided critical revisions to the final manuscript proofs.

Data sharing statement

The data supporting this study's findings are protected by tribal law and are available pending approval from the Fort Belknap Tribal Council, upon reasonable request. The data are not publicly available due to the sensitive nature of the research and the author team's efforts to honor and uphold Fort Belknap's data sovereignty and governance.

Generative artificial intelligence use

The authors declare that they have not used any generative artificial intelligence for the preparation or writing of this manuscript.

Declaration of interests

TB has declared financial support from the William T. Grant and Doris Duke Charitable Foundations and grants/contracts from the National Institute of Nursing Research, National Institute of General Medical Sciences, National Institute of Neurological Disorders and Stroke, and the National Institute of Mental Health. TB declared receiving support for attending/traveling to meetings from the William T. Grant Foundation and Betty Irene Moore Fellowship. JPG, KEN, and MKJ declared no competing interests. ASM declared grants/contracts from the William T. Grant Foundation, the National Institute of Mental Health, the American Foundation for Suicide Prevention and the Western Colorado Community Foundation. ASM has declared receiving honoraria/payment for lectures and/or presentations from: Texas Suicide Prevention Coalition, State Chapters of the American Foundation for Suicide Prevention, Indiana Youth Institute, Colorado School Safety Resource Center, Maryland Center for School Safety, Michigan Association for College Admissions Counseling Conference, Suicide Prevention Conference, Utah Valley University Suicide Prevention Conference, Lenape Valley Foundation, Office of the Colorado Attorney General, Colorado Department of Public Health & Environment/Office of Suicide Prevention, Colorado Department of Education, Colorado School Safety and Resources Center, Suicide Prevention Coalition of Colorado, and the Suicide Prevention Commission (in Office of Suicide Prevention at Colorado Department of Public Health and Environment). ASM has also declared advisory experience for: Project AWARE, Indiana; Social Psychology, Personality, and Interpersonal Processes Study Section (SPIP) at National Institutes of Health; American Foundation for Suicide Prevention; and for Project Suicide Cultures: Reimagining Suicide Research for University of Edinburgh's Amy Chandler (funded by the Wellcome Trust). WO and NP declared financial support from the William T. Grant Foundation.