Alaska Indigenous · Web viewTo provide mental health services, the THO has Alaska Native village counselors, who live and work in the villages, and non-Native Master’s-level mental

Running Head: FACTORS ASSOCIATED WITH ALASKA NATIVE SUICIDAL BEHAVIORS

Factors Associated with Alaska Native Fatal and Nonfatal Suicidal Behaviors 2001-2009: Trends and

Implications for Prevention

Lisa Wexler

Community Health Education, Department of Public Health

School of Health and Health Sciences

University of Massachusetts, Amherst

Marushka L. Silveira, Elizabeth Bertone-Johnson

Epidemiology, Department of Public Health

School of Health and Health Sciences

University of Massachusetts, Amherst

*Corresponding Author: Lisa Wexler

Assistant Professor

313 Arnold House, 715 North Pleasant St

Amherst, MA 01003

Telephone: (413) 545-2248

Fax: (413) 545-1645

Email: [email protected]

Word count: 5,231

Keywords: American Indian/Alaska Native, suicide, suicidal behavior, community-level trends

1


Abstract

Suicide rates among American Indian and Alaska Native (AI/AN) young people are significantly

higher than other ethnic groups in the United States. Not only are there great differences when comparing

AI/AN rates and those of other Americans, some tribal groups have very low rates of suicide while other

Native communities have much higher rates. Despite this obvious variability, there is little research to

help understand the factors associated with these differences. The current study considers the correlates of

suicidal behavior in one rural Alaska Native region that suffers disproportionately from suicide. The

analysis describes suicide behavior between the years 2001-2009, and considers the characteristics

associated with both suicide deaths and nonfatal suicidal behavior. In multivariate analyses we identified

gender, method of suicide and history of previous attempt as significant predictors of fatal suicide

behavior, similar to results obtained from analyses on the same community’s data from the previous

decade. This descriptive study can offer some insights to shape prevention efforts in this and other rural,

tribal communities.

2


Factors Associated with Alaska Native Fatal and Nonfatal Suicidal Behaviors 2001-2009: Trends and

Implications for Prevention

American Indian/Alaska Natives (AI/AN) have the highest rates of suicide in the United States

(Olson & Wahab, 2006; Rhoades, 2003). Suicide disproportionately affects Alaska Natives when

compared with other ethnic groups in Alaska and the United States (Gessner, 1997; . Between 1999 and

2004, 263 Alaska Natives died from suicide, resulting in an age-adjusted suicide rate of 38.5 per 100,000

people (Centers for Disease Control and Prevention, 2005). This is nearly twice the rate for all Alaskans

(19.8) and more than three times that of the nation as a whole (10.7). This disparity is worse among

Alaska Native youth, particularly young men (Day, Provost & Lanier, 2006). From 1999 to 2003, suicide

was the leading cause of death among Alaska Natives between the ages of 15 and 24 (Day et al., 2006).

The suicide rate among Alaska Native males aged 15-24 was eight times higher (142.9 per 100,000) than

among U.S. White males of the same age group (17.4).

Population-level data hide the variability of suicide rates between indigenous populations in

Alaska and elsewhere. For instance, Alaska Natives in the Aleutian Islands have suicide rates below the

national average, and few incidences of nonfatal suicide behavior (Allen & Butler, 2009). In contrast,

Alaska Natives in Northwest Alaska have the highest suicide rate in the state at 90.9 per 100,000 (Alaska

Injury Prevention Center, & American Association of Suicidology, 2007). The region also had the highest

rate of hospitalization in the state due to suicide attempts in the years 1991-2003. Although the primary

author has delved into the study region’s experience with colonialism and explored the linkages with its

high suicide rates elsewhere (Wexler, 2011; 2009a; 2006), there is little epidemiological information of

correlated to better understand this variability, despite being found in other indigenous communities (US

Department of Health and Human Services, 2003; Chandler & Lalonde, 1998). Community-level

surveillance data related to suicide morbidity and mortality can lead to a better understanding of dynamic

factors that drive suicidal behavior in different tribal communities. Understanding the characteristics of

3


suicidal behavior in a rural Alaska Native community can provide detailed insight to better understand the

interactions among risk and protective factors, and help develop suicide prevention strategies.

Tribes in Northwest Alaska have been collecting data about the characteristics, personal history

and trigger events associated with suicidal behavior for two decades. In this paper, we describe the trends

in fatal and nonfatal suicidal behavior for this region from 2001-2009 and identify predictors of suicide

deaths. The regional reporting of the suicidal behavior correlates for this tribal group can provide insight

into important culturally-specific factors to consider for prevention for this population, and perhaps other

AI/AN communities. To provide a more longitudinal perspective, these findings are discussed in the

context of a prior analysis that looked at surveillance data from the same community one decade earlier.

This decade-to-decade comparison provides a viewpoint to discern how suicide correlates have changed

or remained the same over time, and are considered in the context of the community’s prevention efforts.

Methods

Setting

The region of Alaska represented in this study includes an area over 35,898 square miles with

approximately 8000 predominantly Alaska Native (84%) residents (United States Census Bureau, 2005).

It is a rural region with many people living a subsistence lifestyle, which includes hunting. Thus, virtually

every household has access to guns and ammunition. Roads do not connect the dozen small villages;

instead, airplanes, boats and snow machines are the primary means of transportation within the region.

Northwest Alaska’s villages range in population from approximately 100 to 1000 people. Evenso, each

village has a elementary, middle and high school, with only the smallest village sharing a secondary

school with the nearest village (in this case less than 10 miles away). The Tribal Health Organization

(THO) is the main health care and mental health provider for the region. To provide mental health

services, the THO has Alaska Native village counselors, who live and work in the villages, and non-

Native Master’s-level mental health clinicians who travel to villages intermittently. The mental health

clinicians have a rotating on-call duty where they take and respond to 24-hour crisis calls. Each clinician

4


is assigned 1-3 villages (depending on population) where they offer mental health counseling to village

members and support the village-based counselor. Overall, residents in the area are young with over

40% of the population under the age of 18 (United States Census Bureau, 2005).

Data Collection

In this descriptive study, we analyzed data collected from Suicide Reporting Forms (SRF) from

2001-2009 maintained by the Tribal Health Organization (THO), which provides all the health and social

services to the region (to Native and non-Native residents). The forms are completed by Master’s level

mental health clinicians (hereby referred to as clinicians) who are employed by the THO. THO clinicians

offer the only mental health services in the region, and when someone in the region is suicidal, are either

contacted directly through an on-call system or are referred to by healthcare, law enforcement or family

service workers. Fatal suicide behaviors refer to self-injurious acts that result in death. Fatal events that

are ambiguous (e.g. a death resulting from risky behavior such as walking on thin ice) are not counted in

these fatality data. Nonfatal suicide behaviors (attempts) are documented for self-injurious behaviors that

are accompanied by stated suicide intent. Clinicians are notified more reliably in cases of suicide death

than attempts as part of the THO hospital protocol (usually within 24 hours) so that clinicians can be

available to family or village members. Within a week after notification, clinicians talk to the decedent’s

family, friends, or involved professionals to find out more about the suicide, and then complete the SRF.

In nonfatal suicide behavior, clinicians are made aware of suicide attempts through a variety of

formal and informal channels. If the suicidal behavior requires medical attention, the attending medical

staff is required to inform Counseling Services. The on-call clinician will then complete the SRF.

Sometimes a concerned community or family member will contact a clinician. If a person is suicidal at

school, the principal or school counselor will notify Counseling Services. If someone is considered a

danger to themselves by the police, s/he refers to a clinician directly or indirectly (through a healthcare

provider). After hours, a counseling crisis line is available, and the on-call therapist is responsible for

completing a SRF. Of course, if treatment of any sort—mental health, substance abuse, human serives or

5


law enforcement—is not sought, a suicide attempt can remain unreported. In nonfatal cases, the clinician

files out the SRF after being in contact with the suicidal person.

The 19-item SRFs record the demographics of the person exhibiting suicidal behavior as well as

the date, method, location and community of the suicidal act. Items on the SRF are based on research

linking these factors to suicidal behaviors in both Indigenous and other populations (Allen, 1987;

Borowsky, Resnick, Ireland & Blum, 1999; Borowsky, Ireland & Resnick, 2001; Department of Health

and Human Services, 1999; Hendin, Maltsberger, Lipschitz, Haas & Kyle, 2001; Joiner & Rudd, 1998;

Kirmayer, Boothroyd & Hodkins, 1998; National Institute of Mental Health, 2006; Packman, Marlitt,

Bongar & Pennuto, 2004; Wingate, Joiner, Walker, Rudd & Jobes, 2004). These include history of

previous attempts, family history of suicide, method used, location of act, history of mental illness or

contact with mental health service provider, previous warning signs, other contributing factors (with

open-ended responses), and substance abuse history. Questions include a menu of choices along with

open response categories where clinicians can record factors that are not listed. When written responses

can be categorized under an existingcategory, such as “child death” might be subsumed under “death of a

loved one,” percentages are reported as the larger, combined category.

Analysis

Demographic data from the 2000 U.S. Census for the Northwest region of Alaska were used to

compare characteristics of individuals engaging in suicidal behavior to those of the general population in

the region. Bivariate analyses were conducted using Chi-squared tests for categorical variables. Fisher's

Exact Test was used if expected cell counts were less than 5. Potential predictors for fatal versus non-fatal

suicide behavior were identified using chi-squared tests and logistic regression models controlling for age

and sex. Multivariable logistic regression analysis was used, because many of the factors associated with

suicide death were themselves intercorrelated. It was thus important to test the association between

suicide outcome (nonfatal versus fatal) and each of the factors determined to be associated while

simultaneously controlling for every other factor. Backward elimination was used to identify important

6


predictors (p<0.15). Adjusted odds ratios (OR) were reported with 95 percent confidence intervals (CI).

All analyses were conducted using the Statistical Analysis System (SAS Institute, Cary, North Carolina

version 9.2). Two-sided p-values ≤0.05 were considered statistically significant.

Results

Between 2001 and 2009, there were 38 recorded suicide deaths and 510 suicide attempts from

among a population of 7,965. The annual rate for all fatal and nonfatal suicidal behavior was 863 per

100,000 person-years, with an annual rate 60 per 100,000 person-years for fatal suicides (Figure 1).

Individuals 15-19 years had an annual suicide death rate of 124 cases per 100,000 person-years, while

rates for 20-24 and 25-34 year olds were 209 and 126 cases per 100,000 person-years, respectively. Cases

consisted of males (48.5%) and females (51.6 %), with males being over-represented in fatal suicides.

Individuals exhibiting fatal and nonfatal suicidal behavior were significantly different from the

general population in Northwest Alaska, with respect to demographic characteristics (Table 1). A

significant number of individuals represented in the suicide cases were between 15-34 years of age as

compared to the general population (77.2% vs. 31%; p<0.0001). Three-quarters (76.3%) of individuals

exhibiting suicidal behavior were single and not cohabitating, whereas only 42% of the general

population fits this description (p<0.0001). In addition, over half of the suicide cases were unemployed

(55.5%), which is significantly higher than the unemployment found in the general population (11%;

p<0.0001). Likewise, those exhibiting suicidal behavior were less educated with only half (49.8%) having

an education of high school or greater, as compared to the general population (72%; p<0.0001).

When comparing those who died by suicide versus those who attempted it (Table 2), there were

several significant differences. In terms of historical factors, a larger proportion of individuals with non-

fatal suicide behavior reported previous suicide attempts (51.5% vs. 15.8%; p <0.0001). Family history of

suicide was associated more with non-fatal suicide behaviors than with fatal incidents (44.8% vs. 23.7%;

p=0.01). People who died by suicide (26.3%) were more likely to have a family history of substance

abuse (17.6%). Abuse, including family violence, child abuse and neglect and sexual abuse, were

7


recorded in approximately 12% of all cases with no differences observed between fatal and nonfatal

events. Other forms of victimization such as bullying accounted for an additional 5% of all cases.

The SRF data also indicate that referrals to mental health services often occur in a crisis. Of the

510 suicide attempts recorded, 75% resulted in an inpatient admission. This means that clinicians were

contacted when the suicidal individual was too high risk to remain in community-based care.

Firearms were more frequently used as a method in fatal cases (39.47% vs. 3.58%, p<.0001). The

vast majority of suicidal behaviors took place in the individual’s own home or immediate vicinity (75.2%

non-fatal vs. 65.8% fatal), as opposed to 15% of all suicidal behavior in public venues. In addition, most

individuals exhibiting suicidal behavior had a history of alcohol abuse (59.3% non-fatal vs. 64.9% fatal;

p=0.02).

A greater proportion of individuals with suicide attempts were recorded as showing warning signs

as compared to those who died by suicide (32.6% vs. 13.2%, p=0.01). Signs of depression—long sadness,

eating and sleeping disturbances, physical isolation—being the most common of these, accounted for

31.2% of the total. Approximately 19% of cases reported that the victim “told someone they were

suicidal,” and 45.7% exhibited no warning signs. Romantic/other relationship conflicts represented a

higher percentage of the non-fatal events (36.2% vs. 13.2%; p<0.01). Grief was attributed to

approximately 20% of the suicide behavior. Over half of the people represented in the suicide reporting

did not ever receive behavioral health treatment (42.9% non-fatal and 67.6% fatal suicides; p<0.01).

In the multivariate logistic regression analysis (Table 3), we identified three important factors that

were significantly associated with fatal suicide behavior. Males were 3.9 times more likely to die from

suicides as compared to females (OR=3.9; 95% CI: 1.5, 10.0). Individuals with a history of previous

suicide attempts were 75% less likely to exhibit fatal suicide behavior compared to those with no history

(OR=0.3 ; 95% CI: 0.1, 0.7). Finally firearms were associated with a twelvefold increase in the likelihood

of suicide death as compared to other methods of suicide (OR=12.1; 95% CI: 4.6, 31.6).

Discussion

8


In our study, the suicide rate for the Northwest region of Alaska is 60 per 100,000 which is more

than five times (11.3) that of the United States in general and six-times greater than that targeted by the

Healthy People 2020 (10.0) (National vital statistics system, CDC, NCHS,2007). These findings are

slightly higher than the rates found in the region during the previous decade (51 per 100,000) (Wexler,

Hill, Bertone-Johnson & Fenaughty, 2008).

The annual suicide death rate was 18 times higher for Alaska Native youth ages 15-19 at 124

versus 6.9 per 100,000 for American youth. Likewise, the rate for 20-24 year old Alaska Native youth

(209 per 100,000) was much greater than the national rate of 12.7 for this age group. Suicide cases in the

study, similar to other AI/AN groups across North America were younger than the general population

(Hu, Wilcox, Wissow & Baker, 2008; LeMaster et al., 2004; Mullany et al., 2009). This pattern markedly

differs from the national trend in which older Americans are at highest risk for suicide (Hu et al., 2008;

National Institute of Mental Health, 2006). In the study population, females are only slightly more likely

than males to attempt suicide (54% versus 46%), whereas the national trend reflects a tendency for

females to engage in nonfatal suicide behavior 2-3 times as often as males (Centers for Disease Control

and Prevention, 2005). Data from the Northwest Alaska indicate that more than half of those who

attempted suicide, had previously made a suicide attempt. A previous attempt elevates the risk for suicidal

behavior in the general US population.

Individuals exhibiting suicide behavior differed significantly from the general population in

Northwest Alaska with respect to age, gender, ethnicity, marital status, education and employment status.

People represented in the suicide data were not only younger on average than the local population; they

were more likely to be Native. Although both males and females exhibited suicidal behavior, males were

over-represented in the fatal events, accounting for 84% of the suicide deaths. When compared to census

information, the regional suicide data suggests that being single and not co-habitating is associated with

suicidal behavior. Suicide cases were less likely to be employed and were less educated than the general

9


population of the region. This suggests that having a job and completing at least high school may be

protective or might indicate better functional capacity.

The factors associated with fatal and nonfatal cases provide insights into the personal (and family)

histories and behavioral patterns that correlate with suicide outcomes. Abuse is a significant factor, being

recorded in 12% of the cases. These data also suggest that people who attempt suicide may be a different

population than suicide decedents. Nonfatal cases were more likely to show warning signs (e.g. sadness,

increased alcohol use, isolation), to have attempted suicide previously, and to have had suicide deaths in

their family. People who die by suicide are therefore harder to identify before a crisis.

The method, contributing factors and location of the reported suicide behavior provide additional

information that can be used for prevention. Although hanging is recorded most often across fatal and

nonfatal events, firearms being most lethal are strongly associated with suicide death. The majority of

suicidal behavior was exhibited in conjunction with alcohol use, and the majority of people represented

by the suicide data had a history of alcohol misuse. The data also show that most suicidal behavior

occurred in or around people’s homes. Together, this information suggests that prevention efforts must

involve family members, tackle alcohol misuse, and include means restriction such as gunlocks and gun

cabinets.

Despite the prevalence of alcohol misuse (54.9%), interpersonal abuse (12.5%), signs of

depression and stress (31.2%) and grief (25.6%), very few of the suicide decedents sought out or received

behavioral health services (18.4%). Considering both attempters and decedents, over half of those

receiving mental health treatment discontinued care. This suggests that these services are not only

underutilized, but were not working effectively for many people who need them.

The data from years 2001-2009 are remarkably similar to that reported for 1990-2000 (Wexler et

al., 2008). The rates, correlates, methods and characteristics of those exhibiting suicidal behavior are

consistent for both periods, with no substantial variation. These trends have proven difficult to affect with

the current funding climate. The data reported on here (in contrast to that from the previous decade) were

10


generated while the THO focused attention on the issue, and community-based wellness efforts were

initiated and implemented. More specifically, local community taskforces were organized to understand

the problem at a local level and the THO developed several theoretically-supported and regional

prevention initiatives focused on primary and secondary youth suicide prevention. With several two to

three year federal grants, there were gatekeeper trainings offered in all the villages (Wyman, Brown,

Inman, Cross, Schmeelk-Cone, Guo, Pena, 2008), short trainings given to medical staff in the regional

hospital to increase their knowledge of suicide risk (Rihmer, Rutz, Pihlgren, 1995; Szanto, Kalmar,

Hendin, Rihmer, Mann, 2007), education classes focused on suicide prevention and wellness for middle

and high school students, and other multilevel, community-based programming that has been efficacious

in other Native communties (May, Serna, Hurt, DeBruyn, 2005). Because of travel logistics, these

programs occurred intermittently in each community.

These suicide prevention efforts were significantly modified every couple years due to funding

sustainability. The only federal funding for suicide in the United States requires that programs create

new, competitive grant proposals every two to three years. This means that much time over the last

decade has been spent on orienting staff and communities to the newly-funded programs, rather than

sustaining and strengthening approaches that are well-received at a community level and supported by

some research from other American Indian/Alaska Native communities (May, Sems, Hurt, & DeBruyn,

2005). To be more effective, multilevel prevention efforts need to be done consistently over an extended

period of time (Knox, Litts, Talcott, Feig, & Caine, 2003), and should adjust for rural areas and

culturally-specific approaches that require more time building relationships and trust in order to be

implemented. Because of small breaks in funding and uncertainty in the longevity of staff positions, each

of the efforts in the study region has involved hiring new staff for each new grant. This has hampered

efforts greatly and apparently has not affected the targeted outcomes: suicidal behavior has not been

significantly reduced over the past decade. More consistent, long-term and intensive prevention efforts

are needed. As our data suggest, suicide prevention will require a sustained commitment to achieve.

11


Given the local data, prevention efforts should target youth, in general, and young men in

particular. Suicide prevention initiatives must include alcohol prevention and treatment, means

restriction, and provide culturally appropriate, mental health care. The latter is vitally important given the

prevalence of mental health issues reflected in the suicide data and the lack of service utilization. Mental

health services should incorporate formal and informal supports found in village communities, since

clinicians are often not notified until a suicide crisis. Family members or others in the person’s social

network are likely to notice warning signs or be told about the person’s suicidal feelings, yet tend to call

mental health services only when the situation is dire. This is evidenced by the fact that, according to SRF

data on case disposition,75% of the nonfatal suicide behavior resulted in an inpatient admission,

indicating imminent risk. Previous research has demonstrated that educating primary care physicians

about depression can increase antidepressant use, which reduces suicide in regions with high rates

(Henriksson, & Isacsson, 2006, Rihmer, Rutz, Pihlgren, 1995; Szanto, Kalmar, Hendin, Rihmer, Mann,

2007). Using a similar model, education about depression and perhaps alcohol misuse with

paraprofessional health aids who serve the region’s villages can be a first step in linking those in need to

medication and treatment. In addition, learning about grief processes and coping skills, anger

management and how to build and maintain healthy romantic relationships can contribute to prevention.

Lastly, youth employment and high school graduation seem to be protective, thus prevention initiatives

might do well to include support for these outcomes.

Limitations

The study has several limitations that relate to data collection. The SRFs were completed by

different personnel over 8 years because of the high turn-over of clinicians working for the THO. Each

person who collects the information may have different interpretations of the items (e.g. previous warning

signs, substance abuse history, etc.) and varying degree of knowledge about and/or rapport with

informants. The latter can affect accuracy in reporting of sensitive issues such as child abuse, suicidal

behavior, and interpersonal violence. Secondary sources were relied on to give accurate descriptions of

12


suicide decedent's past behavior and "possible contributing factors." This information is likely to reflect

recall bias since the information was collected after an outcome occurred. Caution should also be taken

when considering the comparisons between the suicide decedents and attempters in relation to mental

health care. We know that due to THO procedures, all attempters that were recorded in this data were at

least offered mental health services whereas the decedents had variable access to mental health services.

History of abuse (physical and sexual abuse, bullying, other forms of victimization) is likely to be under-

reported in the data due to the associated stigma and privacy issues (Wexler, 2005).

No data are collected on unreported suicide attempts. Reporting on nonfatal suicide behavior may

be unreliable due to the variable communication between clinicians and community members (Wexler,

2005). Communication tends to improve with the longevity of the clinician’s residency, yet clinicians on

average only maintained employment for less than 2 years during this period. Nonfatal suicide behavior

are reported by community service providers and community members. The latter source is highly

variable in reporting to mental health services, therefore suicide attempts are probably underreported in

these data.

Ideally the study also would have epidemiological information about psychosocial variables in

the general population of Northwest Alaska so that chi squared and logistical regressions could be

performed. This would enable us to determine if the correlates of suicidal behaviors such as a history of

substance abuse or mental illness, family history of suicide, mental health service utilization rates and

other contributing factors might actually be correlates of the community itself. Instead, our analysis only

examined fatal versus nonfatal cases.

Many additional factors such as cultural identity, community engagement and belongingness that

could be important for suicide prevention are not documented on the SRFs. By describing the

characteristics and correlates of suicidal behavior identified on the reporting forms, this paper offers a

focused perspective about potential ways to prevent suicide. Other research from this region suggests that

13


these culturally-based factors need to be incorporated into suicide prevention efforts Wexler, 2009;

Wexler, 2011; Wexler, In press).

Conclusions

Community-based surveillance offers critical insight into the associated factors and characteristics of

suicide mortality and morbidity of high-risk groups. The unique risk and protective factors associated

with suicidal behavior can help structure prevention efforts for the study population and perhaps other

AI/AN communities. Several important areas to focus prevention efforts include targeting young people,

alcohol misuse and reformulating mental health services so that they are better utilized and more

effective. This could involve education aimed at the primary healthcare providers in the village health

clinics: village-based health aids so they are better able to identify people who show signs of depression

and alcohol disorders. Village-based health aids could enable more people to get the depression (or other

psychiatric) or alcohol treatment they need. Although these recommendations require specific research on

how to implement them effectively, the need to involve family and other local people is clear. This is

especially important when considering the remoteness of many Native communities, the cultural

differences between Native residents and most of the mental health service providers, and the current

underutilization of mental health services in indigenous communities (Freedenthal & Stiffman, 2007;

Oetzel et al., 2006). Additionally, interventions that restrict access to firearms (or ammunition) are

warranted. Health promotion initiatives have the potential to reduce suicide behavior if they increase

employment and school retention. More research is needed to identify how this can be done in a culturally

relevant and meaningful manner for Alaska Native and for other Native communities. Lastly, funding of

suicide prevention efforts should consider how to sustain long-term and consistent, multilevel

interventions, particularly in communities that have extremely high rates of suicide and are under-served

by the current systems of care.

14


15

Documents

Alaska Indigenous · Web viewTo provide mental health services, the THO has Alaska Native village counselors, who live and work in the villages, and non-Native Master’s-level mental