Association of Child Abuse Exposure With Suicidal Ideation, Suicide Plans, and Suicide Attempts in Military Personnel and the General Population in Canada

Importance  Recent evidence indicates a high prevalence of child abuse exposure in modern US veterans, which may explain in part their higher likelihood of suicide relative to civilians. However, the relationship between child abuse exposure and suicide-related outcomes in military personnel relative to civilians is unknown. Furthermore, the associations among deployment-related trauma, child abuse exposure, and suicide-related outcomes in military personnel have not been examined.

Objectives  To determine whether child abuse exposure is more prevalent in Canadian Armed Forces (CAF) personnel compared with the Canadian general population (CGP); to compare the association between child abuse exposure and suicidal ideation, suicide plans, and suicide attempts among the CAF and CGP; and to determine whether child abuse exposure has an additive or interaction effect on the association of deployment-related trauma and past-year suicidal ideation and suicide plans among Regular Forces personnel.

Data, Setting, and Participants  Data were collected from the following 2 nationally representative data sets: the 2013 Canadian Forces Mental Health Survey (CFMHS) for the CAF (8161 respondents; response rate, 79.8%) and the 2012 Canadian Community Health Survey–Mental Health (CCHS-MH) for the CGP (23 395 respondents; response rate, 68.9% [of these, 15 981 age-matched participants were drawn]). Data were collected from April 15 to August 31, 2013, for the CFMHS and January 2 to December 31, 2012, for the CCHS-MH. Data were analyzed from October 2014 to October 22, 2015. Statistical weights were applied to both data sets.

Main Outcomes and Measures  Child abuse exposure, including physical abuse, sexual abuse, and exposure to intimate partner violence, and deployment-related trauma were assessed in relation to suicide-related outcomes.

Results  Data were analyzed from 24 142 respondents aged 18 to 60 years (Regular Forces, 86.1% male and 13.9% female; Reserve Forces, 90.6% male and 8.9% female; and CGP, 49.9% male and 50.1% female). Any child abuse exposure was higher in the Regular Forces (47.7%; 95% CI, 46.4%-49.1%) and Afghanistan mission–deployed Reserve Forces (49.4%; 95% CI, 46.3%-51.5%) compared with the CGP (33.1%; 95% CI, 31.8%-34.4%). All types of child abuse exposures were associated with increased odds of suicidal ideation, suicide plans, and suicide attempts in the CGP (range of adjusted odds ratios [AORs], 3.0 [95% CI, 2.3-3.9] to 7.7 [95% CI, 5.7-10.3]; P < .05) and CAF (range of AORs, 1.7 [95% CI, 1.0-2.9] to 6.3 [95% CI, 4.2-9.5]; P < .05), with many associations significantly weaker in military personnel relative to civilians. Additive effects for past-year suicide ideation (AOR, 2.7; 95% CI, 1.8-4.2) and past-year suicide plans (AOR, 4.6; 95% CI, 2.3-9.2) but not interactive effects for past-year suicide ideation (AOR, 1.2; 95% CI, 0.7-2.2) and past-year suicide plans (AOR, 0.8; 95% CI, 0.3-2.2) were noted between deployment-related trauma and child abuse exposure among Regular Forces personnel.

Conclusions and Relevance  Individuals with a child abuse history may be more likely to enter the military, and child abuse exposure may increase the likelihood of suicide-related outcomes. Prevention efforts targeting child abuse may reduce suicide-related outcomes.

Suicide is an important public health problem in military and civilian populations. In 2002, in a representative Canadian Armed Forces (CAF) sample, approximately 4% of serving personnel reported past-year suicidal ideation.^1,2 In 2013, the past-year and lifetime prevalences of suicidal ideation were 4.3% and 15.4%, respectively, in the CAF and 3.6% and 13.3%, respectively, in the Canadian general population (CGP) (J.S., T.O.A., T.T., et al, unpublished data, August 2015). Recent data from the United States indicates that the lifetime prevalences of suicidal ideation, suicide plans, and suicide attempts among soldiers were 13.9%, 5.3%, and 2.4%, respectively.³ Despite a large volume of literature about suicide, accurately anticipating which individual will think about, plan, and attempt suicide remains a difficult task in military and civilian populations.

Traumatic experiences may occur during military service, deployment in peacekeeping missions,⁴ and combat.¹ Many studies have examined occupational trauma as an explanation for suicide among military personnel. For example, deployment-related trauma, such as witnessing atrocities or massacres, is associated with increased odds of experiencing past-year suicidal ideation among military personnel.¹ However, increased suicides have been noted among military personnel who have not deployed,⁵ showing that deployment-related trauma does not completely explain the occurrence of suicide among this population. Other studies^6-8 found that deployment was not related to suicide risk. Military and veteran suicide prevention efforts thus hinge on understanding what factors other than deployment drive suicide.

In addition to occupational trauma, military personnel may also experience nonoccupational trauma. Child abuse exposure has been linked to suicidal ideation and suicide attempts mainly in civilian populations.^9-14 In representative Canadian and US general population samples, an increasing number of childhood adversities corresponds with a stepwise increase in the odds of suicidal ideation and suicide attempts.^10,15 Trauma occurring in childhood before military service may be cumulative with deployment-related trauma, meaning that an increased number of lifetime traumatic events may be associated with a greater likelihood of suicidal ideation and suicide attempts. Alternatively, low exposure to trauma may result in development of resilience for some, with an increased ability to manage new trauma.¹⁶ The effect of child abuse exposure and deployment-related trauma on the likelihood of suicide-related outcomes among military personnel remains unknown.

From clinical and nonrepresentative military samples, child abuse exposure is known to be highly prevalent among military personnel, ranging from 17% to 50%.^17-19 In a representative 2002 CAF sample, 1% to 15% of respondents reported being badly beaten, witnessing their mother being beaten, and/or being sexually abused in childhood.²⁰ In addition, being badly beaten as a child and witnessing severe fighting in the home were associated with an increased odds of suicide attempts.²¹ Although informative, these studies are limited owing to subjective assessment of child abuse (ie, how were badly beaten or severe fighting in the home interpreted?), and they predate the mission in Afghanistan. Recent US studies have examined predictors and correlates of suicide among soldiers, but none of these studies has investigated the potential role that child abuse exposure may play in suicide-related outcomes.^3,5,22-25

Inquiry into military samples would be novel because some evidence indicates that individuals who experience childhood adversity may be more likely to select themselves into the military, possibly to escape adversity at home.^26,27 However, this statement has been criticized because it perpetuates the stereotype of the disadvantaged background of military personnel.²⁸ In addition, no representative military and civilian comparisons from any country examine possible differences in the prevalence of child abuse exposure and potential differences of effect size in the relationship between child abuse exposure and suicide-related outcomes. For example, differences between the military and civilian mental health systems might facilitate or inhibit help seeking by those with child abuse exposure, thereby decreasing or increasing the risk for eventual deterioration into suicidal behavior.²⁹ Comparing these populations would provide a unique opportunity to situate military suicide–related outcomes in a larger context.

We undertook the present study with 3 overall objectives. First, we sought to determine whether child abuse exposure is more prevalent in the CAF compared with the CGP. Second, we compared the strength of the association between child abuse exposure and suicide-related behavior among the CAF and the CGP. Third, we examined whether child abuse exposure has an additive or interaction effect on the association between deployment-related trauma and suicide-related outcomes among Regular Forces personnel.

Box Section Ref ID

Data were drawn from 2 Canadian data sets. The Canadian Forces Mental Health Survey (CFMHS) was collected in 2013 (8161 respondents; response rate, 79.8%),³⁰ and the Canadian Community Health Survey–Mental Health (CCHS-MH) was collected in 2012 (23 395 respondents; response rate, 68.9%).³¹ Both data sets are cross-sectional and nationally representative and were collected by Statistics Canada using the same training and quality control procedures, face-to-face interview technology, and survey methods and assessments.^30,31 Anonymity of all respondents was protected by Canada’s Statistic Act,^32,33 which may be especially important among military personnel.³⁴ The CFMHS sample was representative of the Regular Forces personnel (n = 6692), approximately 46% of whom had deployed in support of the mission in Afghanistan, and Reserve Forces personnel (n = 1469) who deployed in support of the mission in Afghanistan. Reservists who had not deployed in support of the mission in Afghanistan were not included in the survey. For accurate comparison, the CCHS-MH sample was restricted to respondents aged 18 to 60 years (n = 15 981) to match the age in the CFMHS. Both surveys were voluntary, and the privacy and confidentiality of the respondents are ensured under the Statistics Act.^35,36 Data were collected from April 15 to August 31, 2013, for the CFMHS and January 2 to December 31, 2012, for the CCHS-MH. Ethical approval for both surveys was provided by the relevance policy committees at Statistics Canada. Respondents provided written informed consent.

Physical abuse and exposure to intimate partner violence were assessed using items from the Childhood Experiences of Violence Questionnaire.³⁷ Respondents 18 years or older were asked about childhood abuse that occurred before their 16th birthday. Binary classifications of child abuse exposure (present or absent) were made according to the Childhood Experiences of Violence Questionnaire guidelines as follows: physical abuse was defined as present if 1 or more of 3 variables met the following threshold criteria: (1) being slapped on the face, head, or ears or hit or spanked with something hard 3 or more times; (2) being pushed, grabbed, shoved, or having something thrown at the respondent to hurt them 3 or more times; and (3) being kicked, bit, punched, choked, burned, or physically attacked 1 or more times. Exposure to intimate partner violence was defined as present if the respondent indicated having seen or heard parents, stepparents, or guardians hitting each other or another adult in the home 3 or more times.³⁷ Sexual abuse was assessed using 2 items and categorized as present if either of the following occurred 1 or more times: (1) attempted or being forced into unwanted sexual activity by being threatened, held down, or hurt in some way and/or (2) sexual touching, meaning unwanted touching or grabbing, kissing, or fondling against the respondent’s will. Any child abuse exposure included physical abuse, sexual abuse, and/or exposure to intimate partner violence.

Deployment-related traumatic events were assessed only in the CFMHS data and included (1) knowing someone who was seriously injured or killed; (2) finding yourself in a threatening situation and unable to respond because of rules of engagement; (3) ever being injured; (4) ever seeing ill or injured women or children whom the respondent was unable to help; (5) ever feeling responsible for the death of Canadian or allied personnel; and (6) ever having a close call, for example, being shot or hit but saved by protective gear. Personnel indicating that they had never been deployed were coded as no to all of these experiences. A variable was computed to assess whether any deployment-related traumatic event had been experienced (yes or no). The eMethods in the Supplement provides additional deployment-related traumatic events.

Military rank in CAF personnel included junior noncommissioned member, senior noncommissioned member, and officer. Sociodemographic covariates included age, sex, educational level, household income, and marital status.

Each respondent was asked whether he or she had ever thought seriously about committing suicide or taking his or her own life (yes or no) and whether this experience had happened in the past year (yes or no). Each respondent was asked whether he or she had ever made a plan to commit suicide (yes or no) and whether this experience had occurred in the past year (yes or no). Each respondent was asked whether he or she had ever attempted suicide or tried to take his or her own life (yes or no). Past-year suicide attempts were not included owing to low occurrence.

Data were analyzed from October 2014 to October 22, 2015. Statistical weights were applied for population representation in the CGP, Regular Forces, and Reserve Forces. We performed bootstrapping to account for the complex survey design. First, descriptive statistics for each population were computed. Second, overall and sex-specific prevalence estimates were computed for child abuse exposure across the study populations. Third, a series of multivariable logistic regression analyses adjusted for sociodemographic covariates were conducted to determine the likelihood of child abuse exposure among Regular Forces compared with the CGP, Afghanistan mission–deployed Reserve Forces compared with the CGP, and Regular Forces compared with Afghanistan mission–deployed Reserve Forces. Ad hoc analyses were conducted for the sexual abuse models, which demonstrated that limited variability across income created unstable results. Therefore, the sexual abuse models were further stratified by income level of less than $80 000 and $80 000 or more (US $58 593).

Fourth, a series of multivariable logistic regression models (ie, adjusted for sociodemographic covariates) was computed to determine whether child abuse exposure increased the likelihood of suicide-related outcomes in each population (findings examining all child abuse exposure types concurrently are provided in eTable 1 in the Supplement). Differences across study populations were examined using an interaction term for child abuse exposure by study population. We calculated population-attributable fractions (PAFs), which represent an estimate of the proportion of the suicide-related outcome that might be decreased if the child abuse exposure had not occurred.³⁸ The PAFs provide informative estimates of the magnitude of the associations, although the causality assumption is not met with these data.

Fifth, the additive effects of child abuse exposure and deployment-related trauma on past-year suicide-related outcomes among Regular Forces personnel were examined using multivariable logistic regression analyses (ie, adjusted for sociodemographic covariates and military rank). Finally, we computed a series of main effects and interaction effects logistic regression models (ie, adjusted for sociodemographic covariates and military rank) to examine whether child abuse exposure and deployment-related trauma had an interactive effect on past-year suicide-related outcomes among Regular Forces personnel. Additional deployment-related trauma analyses are provided in eTable 2 in the Supplement.

Table 1 provides sociodemographic information. The prevalence of any child abuse exposure was 33.1% (95% CI, 31.8%-34.4%) in the CGP, 47.7% (95% CI, 46.4%-49.1%) in the Regular Forces, and 49.4% (95% CI, 46.3%-51.5%) in the Reserve Forces (Table 2). Prevalence of child abuse exposure in the CGP was slightly different compared with prevalence reported in a 2014 publication (33.1% vs 32.1%)¹⁵ because the present investigation was restricted to respondents aged 18 to 60 years to match the military personnel sample. Regular Forces personnel were more likely than the CGP to have experienced all types of child abuse exposure (range of adjusted odds ratios [AORs], 1.6 [95% CI, 1.4-1.9] to 2.8 [95% CI, 2.1-3.5]; P < .001), with the exception of sexual abuse among males. Reserve Forces personnel were more likely than the CGP to have experienced any physical abuse and exposure to intimate partner violence (with the exception of females for the exposure to intimate partner violence only); sexual abuse was more likely in the Reserve Forces compared with the CGP among females in the higher income category only. All types of child abuse exposure were equally likely among Regular Forces and Reserve Forces personnel.

All child abuse exposure types were associated with increased odds of suicidal ideation, suicide plans, and suicide attempts in the CGP (range of AORs, 3.0 [95% CI, 2.3-3.9] to 7.7 [95% CI, 5.7-10.3]; P < .05) and CAF (range of AORs, 1.7 [95% CI, 1.0-2.9] to 6.3 [95% CI, 4.2-9.5]; P < .05), with a few exceptions among the Reserve Forces personnel likely owing to underpowered models (Table 3 and Table 4). When we compared the populations (using interaction terms), the associations between child abuse exposure and many suicide-related outcomes were significantly weaker for many outcomes in the CAF relative to the CGP (Tables 3 and 4). The PAF estimates indicate that child abuse exposure is associated with a substantial proportion of suicide-related outcomes, including 14.9% to 62.6% in the CGP, 10.1% to 47.0% in the Regular Forces personnel, and 10.2% to 52.4% among Reserve Forces personnel.

Among Regular Forces personnel, 43.8% had experienced 1 or more deployment-related trauma. Deployment-related trauma occurring without a history of child abuse exposure was not associated with past-year suicidal ideation (AOR, 1.2 [95% CI, 0.7-1.8]) or suicide plans (AOR, 1.9 [95% CI, 0.8-4.3]). Child abuse exposure without deployment-related trauma and child abuse exposure and deployment-related trauma together were associated with increased odds of past-year suicidal ideation (AORs, 2.0 [95% CI, 1.3-3.0] and 2.7 [95% CI, 1.8-4.2], respectively; P < .05) and suicide plans (AORs, 2.9 [95% CI, 1.4-6.0] and 4.6 [95% CI, 2.3-9.2], respectively; P < .05). Experiencing child abuse exposure only relative to deployment-related trauma only was associated with increased odds of past-year suicidal ideation but not suicide plans. Additive effects were noted because having experienced child abuse exposure and deployment-related trauma significantly increased the odds of past-year suicidal ideation and suicide plans compared with exposure to deployment-related trauma only. In the series of models examining interaction effects, deployment-related trauma in the main-effects models was associated with past-year suicidal ideation (AOR, 1.4; 95% CI, 1.0-1.8; P < .05) and suicide plans (AOR, 1.7; 95% CI, 1.1-2.7; P < .05) (Table 5). However, only the suicide plan models remained significant when adjusting for child abuse exposure (AOR, 1.7; 95% CI, 1.1-2.6; P < .05). In contrast, child abuse exposure was associated with past-year suicidal ideation (AOR, 2.2; 95% CI, 1.6-2.9; P < .05) and suicide plans (AOR, 2.7; 95% CI, 1.7-4.3; P < .05) and remained significant after adjusting for deployment-related trauma (AORs, 2.1 [95% CI, 1.6-2.9] and 2.6 [95% CI, 1.6-4.2], respectively; P < .05). No interaction effects between child abuse exposure and deployment-related trauma were noted.

This study advances knowledge with the following novel findings. First, in Canada, child abuse exposure is more likely among military personnel compared with the general population. Second, although child abuse exposure is associated with increased odds of suicidal ideation, suicide plans, and suicide attempts among the CGP, Regular Forces, and Reserve Forces, the effect of child abuse exposure on suicide-related outcomes is generally stronger among the CGP compared with military personnel. Third, deployment-related trauma was associated with past-year suicidal ideation and suicide plans; however, in comparison, child abuse exposure was more strongly and more consistently associated with suicide-related outcomes.

Almost half of all military personnel in Canada have a history of child abuse exposure, which is a higher prevalence than in the general population. We are unable to determine why this is the case, but escaping from child abuse exposure at home²⁶ or otherwise improving life circumstances with career and education opportunities available through the military may be the cause. Although child abuse exposure was associated with suicide-related outcomes among civilians and military personnel, the effects were often greater among civilians. Fortunately, not everyone with a history of child abuse exposure will experience poor mental health outcomes. Military screening and selection procedures at the time of recruitment and during continued military service may select more resilient individuals and exclude the less resilient. In addition, positive aspects of the military, such as escaping a negative home environment or community and/or having meaningful and secure income and employment, may attenuate the negative effects of child abuse exposure on suicide-related outcomes. However, we are unable to determine whether this attenuation is the case with our data. Further research is necessary to determine what factors may be attenuating the negative effects of childhood abuse among military personnel.

When examining deployment-related trauma and child abuse exposure in separate models, both experiences increased the likelihood of suicide-related outcomes. Although deployment-related trauma is often examined in relation to suicide in military samples,^1,5-8 only a few studies have examined the importance of child abuse exposure.²¹ The present study finds that the relationship between child abuse exposure and suicide-related outcomes was robust in all models and a stronger indicator of suicide-related outcomes than deployment-related trauma.

Strengths of this study include the large and nationally representative data of the CGP and CAF collected at a similar time point using the same sampling designs and measures. However, several limitations apply. First, the data are cross-sectional, self-reported, and retrospective, which preclude causal inferences. Although the data are retrospective, evidence suggests the accuracy of recall of self-reported adverse childhood events.^39-41 Second, although including Reserve Forces personnel adds originality, the sample was not representative because it only included Reserve Forces who had been deployed in support of the mission in Afghanistan. Also, because the Reserve Forces sample was small, they could not be included in all analyses, and some models may have been underpowered. Third, measures of neglect and emotional abuse were not included in either survey. Fourth, the focus of this work was child abuse, but other nonoccupational trauma may be important. Fifth, we cannot determine precisely the temporality of onset of trauma (in childhood and on deployment) and suicide-related outcomes. Finally, selection effects related to stringent occupational fitness standards may preclude the recruitment of some suicide-vulnerable individuals and may deplete suicide-vulnerable individuals from the military population as they develop during service. Although these selection effects certainly occur, our findings demonstrate that they do not extend to child abuse exposure, for which military personnel are at increased risk.

Approximately half of military personnel in Canada begin their service with a history of child abuse exposure, a proportion that is higher than among civilians. However, the association between child abuse exposure and suicide-related outcomes was often significantly weaker in military personnel relative to civilians. The higher prevalence and the broad negative effects of child abuse exposure make this finding an important public health concern in the military, as in civilians. Deployment-related trauma was associated with past-year suicide-related outcomes, but the association was attenuated when adjusting for child abuse exposure. In addition, when deployment-related trauma was experienced with child abuse exposure, the effect on past-year suicidal ideation and suicide plans were additive among Regular Forces personnel. Child abuse exposure, however, was associated with all suicide-related outcomes, even when adjusting for deployment-related trauma. Therefore, prevention efforts targeting child abuse exposure or mediators in the relationship between child abuse exposure and suicide-related outcomes may help reduce suicide-related outcomes.

Corresponding Author: Tracie O. Afifi, PhD, Department of Community Health Sciences, University of Manitoba, S113-750 Bannatyne Ave, Winnipeg, MB R3E 0W5, Canada (tracie.afifi@umanitoba.ca).

Submitted for Publication: August 17, 2015; final revision received October 30, 2015; accepted October 31, 2015.

Published Online: January 27, 2016. doi:10.1001/jamapsychiatry.2015.2732.

Author Contributions: Dr Afifi had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Afifi, Taillieu, Zamorski, Sareen.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: All authors.

Critical revision of the manuscript for important intellectual content: Afifi, Taillieu, Zamorski, Sareen.

Statistical analysis: Afifi, Taillieu, Zamorski, Turner, Cheung.

Obtained funding: Afifi, Zamorski, Sareen.

Administrative, technical, or material support: Sareen.

Study supervision: Afifi, Sareen.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by a New Investigator award from the Canadian Institutes of Health Research (Dr Afifi), an establishment grant from the Manitoba Health Research Council (Dr Afifi), a chair award from the Manitoba Health Research Council (Dr Sareen), and the Department of National Defense (DND) for data collection.

Role of the Funder/Sponsor: The DND reviewed the manuscript but did not provide any comments or ask for changes. The other funding sources did not have any role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: Statistics Canada collected and provided the data for academic purposes, but the analyses are the sole responsibility of the authors. The opinions expressed do not represent the views of Statistics Canada.