Prevalence and Correlates of Mental Health Problems in Canadian Forces Personnel Who Deployed in Support of the Mission in Afghanistan: Findings From Postdeployment Screenings, 2009–2012

Abstract

Objective:

An important minority of military personnel will experience mental health problems after overseas deployments. Our study sought to describe the prevalence and correlates of postdeployment mental health problems in Canadian Forces personnel.

Method:

Subjects were 16 193 personnel who completed postdeployment screening after return from deployment in support of the mission in Afghanistan. Screening involved a detailed questionnaire and a 40-minute, semi-structured interview with a mental health clinician. Mental health problems were assessed using the Patient Health Questionnaire and the Posttraumatic Stress Disorder Checklist—Civilian Version. Logistic regression was used to explore independent risk factors for 1 or more of 6 postdeployment mental health problems.

Results:

Symptoms of 1 or more of 6 mental health problems were seen in 10.2% of people screened; the most prevalent symptoms were those of major depressive disorder (3.2%), minor depression (3.3%), and posttraumatic stress disorder (2.8%). The strongest risk factors for postdeployment mental health problems were past mental health care (adjusted odds ratio [AOR] 2.89) and heavy combat exposure (AOR 2.57 for third tertile, compared with first tertile). These risk groups might be targeted in prevention and control efforts. In contrast to findings from elsewhere, Reservist status, deployment duration, and number of previous deployments had no relation with mental health problems.

Conclusions:

An important minority of personnel will disclose symptoms of mental health problems during postdeployment screening. Differences in risk factors seen in different nations highlight the need for caution in applying the results of research in one population to another.

Canada has deployed more than 40 000 personnel in support of its combat and peace support operation in Afghanistan. Psychiatric injury rates remain high: 13.5% of people deployed in support of the mission in Afghanistan from 2001–2008 were diagnosed by the CF (recently renamed to the Canadian Armed Forces but was Canadian Forces at the time of our study) with a mental disorder attributed at least, in part, to that operation.¹

Military mental health problems are impactful: when treatment requirements or dysfunction preclude future deployments, operational readiness is threatened. If persistent, these can lead to medical release. As well, if civilian data^2–5 and emerging US military findings^6,7 on absenteeism and presenteeism apply to the CF workplace as well, deployment-related mental disorders erode productivity in garrison. Mental health services also represent a large and growing fraction of health care provided by military organizations.⁸ Finally, mental health problems are leading contributors to benefits in today’s veterans.⁹ For all of these reasons, military organizations take an active interest in surveillance, prevention, and control of deployment-related mental health problems.

A broad range of risk factors for postdeployment mental health problems in the military have been identified. Predeployment risk factors include baseline mental health¹⁰ and resilience,¹¹ as well as factors related to operational tempo, such as number of previous deployments,^12,13 total number of months deployed over a period of time,¹⁴ and dwell time in garrison between deployments.^12,15 As a result, military organizations have guidelines¹⁶ in place to regulate operational tempo, in part to keep its impact on mental health to sustainable levels. Intradeployment factors include exposure to traumatic stressors,¹⁷ deployment length,¹⁸ and leadership.¹⁹ Prominent postdeployment factors include social support²⁰ (including homecoming experiences²¹) and current life stressors.²⁰ Other identified risk factors could act across the deployment cycle, including sex,²² age, rank,¹ marital status,²³ component (Regular, compared with Reserve Force),²⁴ and element (Army, Navy, or Air Force).²⁵

Clinical Implications

• Heavy combat exposure and a past history of mental health services use were strong predictors of postdeployment mental health problems; people with these risk factors can be targeted for prevention and control efforts.

• Military risk factors identified in US studies (Reservist status, number of deployments, and deployment length) were not replicated in CF personnel, highlighting the need for caution in applying research from one military organization to another.

Limitations

• Postdeployment screening is confidential but not anonymous; other research suggests that there is substantial underreporting of mental health problems in nonanonymous surveys of military personnel.

• Our study does not include a control group of nonscreened personnel, thus the impact of the screening program on service use and well-being are unclear.

Available services and supports may also act across the deployment cycle to decrease the prevalence of postdeployment mental health problems. For example, predeployment mental health training,²⁶ postdeployment mental health training,²⁷ psychological debriefing,²⁷ and postdeployment psychological decompression²⁸ have all been linked to better postdeployment mental health.

The sum total of these factors (which differ from nation to nation) drives the prevalence of postdeployment mental health problems. While other Canadian studies have described the prevalence of mental disorders in garrison^29,30 and during deployment,³¹ published studies¹ on the prevalence of postdeployment disorders in CF personnel are limited.

Understanding risk factors for postdeployment mental health problems helps identify

1. potential targets for intervention¹¹;

2. risk groups for prevention and control efforts^32,33; and

3. opportunities for strengthening human resources policies (for example, those regulating operational tempo).

For example, US and UK studies have shown higher rates of postdeployment mental health problems in reservists,^23,24 those with longer deployments,¹⁸ and those with multiple deployments.^12,14,15 However, important differences in human resource policies and other factors could alter these risk factors for other countries.

Our paper sought to explore the prevalence and correlates of mental health problems in a large cohort of CF personnel deployed in support of its mission in Afghanistan, using data collected during EPDS for mental health problems.

Methods

Context: The Enhanced Postdeployment Screening

CF policy requires the EPDS process, which takes place between 90 to 180 days after return. The process consists of completion of a confidential mental health screening questionnaire and a 40-minute, semi-structured interview with a mental health professional. Questionnaire data are captured for health surveillance and research purposes. Compliance with the EPDS process is at least 76%.

Study Population

The study population consisted of all CF personnel (N = 16 193) who deployed in support of the mission in Afghanistan and who completed a useable EPDS questionnaire from January 2009 to July 2012.

Questionnaire Content

The EPDS questionnaire contains questions on sociodemographic and military characteristics developed for the process. Key risk factors were selected based on other research showing a relation to mental health problems in military personnel. Factors potentially acting across the deployment cycle included sociodemographic characteristics (age, sex, first official language, marital status, years of military service, rank, component [Regular, compared with Reserve Force], and element [Army, Navy, or Air Force]). Predeployment factors included number of previous operational deployments and past use of mental health care. Intradeployment factors included the duration of the deployment and the extent of combat exposure. Postdeployment factors included current use of mental health care and time since return from deployment.

Current and past mental health care use was assessed using 2 questions adapted from the CCHS 1.2, “Have you ever seen, or talked to on the telephone, a health professional about your emotional or mental health? [Do not include routine pre- or postdeployment screening]” and “Are you currently seeing a health professional about your emotional or mental health?” Current use of psychiatric medications was assessed using a single item from the PHQ: “Are you taking any medication for anxiety, depression, or stress?” Respondents were categorized into 3 groups: never, past care only, and current care.

Combat exposure was assessed using 30 items developed by the US Army,³⁴ which assessed having ever experienced a range of potentially traumatic events during the recent deployment, such as receiving small arms fire, being wounded or injured, and seeing ill or injured women or children who you were unable to help. A score reflecting the simple sum of affirmative responses (potential range, 0 to 30) was calculated, and the score was analyzed in tertiles relative to a larger reference population of CF personnel undergoing screening after a broad range of military operations.

Common mental health problems were assessed using portions of the PHQ³⁵ and the full PCL-C.³⁶ The civilian version was used because the screening process is intended to capture PTSD related to other events. For the PHQ, algorithms developed by the instrument’s authors were used.³⁵ Based on validation against a structured clinical interview,³⁵ standard algorithms were used for PHQ MDD (hereafter symptoms of MDD), PHQ other depressive syndrome (hereafter symptoms of minor depression), and PHQ other anxiety syndrome (hereafter symptoms of GAD). For PHQ panic syndrome (hereafter symptoms of panic disorder), the EPDS questionnaire omitted the symptom checklist for panic attacks; the original algorithm requires 4 or more symptoms during the last panic attack in addition to endorsing a panic attack in the previous 4 weeks, having had other attacks previously, having had at least 1 attack out of the blue, and having been bothered by the attacks or worried about future attacks. For this analysis, the attack symptoms criterion was not imposed, which, based on earlier versions of the EPDS questionnaire that did include those items, results in about 10% higher prevalence. Suicidality was assessed using one of the PHQ depression items, “Thoughts that you would be better off dead, or of hurting yourself in some way,” during the previous 2 weeks. For PTSD (hereafter symptoms of PTSD), a cut-off of 50 or higher on the PCL-C was used.³⁷

Primary Outcomes

The primary outcome for this analysis was the presence of 1 or more of 6 mental health problems identified above.

Statistical Analysis

For people who completed more than 1 EPDS questionnaire during the study period (n = 972), only the first questionnaire was used for the analysis to simplify analysis and interpretation. Analyses were performed with SAS software, version 9.2 (SAS Institute Inc, Cary, NC). Given that the findings represent a near census of the underlying population, confidence intervals for prevalence estimates are not reported. Logistic regression was used to characterize univariate and multivariate associations between risk factors and postdeployment mental health problems. Results are reported as odds ratios and 95% confidence intervals. After filling in missing values using administrative data, missing data accounted for, at most, 2.1% of each individual covariate. List-wise deletion was used for regression modelling, resulting in the elimination of 1218 cases (7.5%).

Ethical Aspects

This analysis is part of the CF’s Mild Traumatic Brain Injury Outcome Study, which was reviewed and approved by an independent Research Ethics Board (Veritas IRB, Dorval, QC).

Results

As shown in Table 1, respondents were largely male, junior noncommissioned members in the Regular Force with substantial military experience. Army personnel predominated. Past and current mental health care were reported by 13.1% and 7.7% of respondents, respectively. Most deployments were 6 to 8 months in length. There was a broad range of combat exposure among respondents (median 6, interquartile range 3 to 12). Most screenings were completed within the period targeted by the EPDS policy (90 to 180 days), but an important minority (27.6%) occurred more than 180 days after return.

Table 1.

Sociodemographic and military characteristics of the study population (overall N = 16 193)

Variable	n	%
Sex
Male	14 678	90.7
Female	1510	9.3
Total	16 188
Age at time of questionnaire, years
≤24	3139	19.4
25 to 35	7085	43.8
35 to 44	4067	25.1
≥45	1889	11.7
Total	16 180
Language
English	11 864	73.5
French	4188	26.0
Other	80	0.5
Total	16 132
Marital status
Married or living with partner	8989	55.5
Single (never married)	5968	36.9
Widowed, divorced, or separated	1231	7.6
Total	16 188
Rank
Officer	2333	14.4
Senior NCMa	2876	17.8
Junior NCMb	10 965	67.8
Total	16 174
Component
Regular Force	13 841	85.5
Reserve Force	2350	14.5
Total	16 191
Element
Army	12 861	79.6
Navy	961	5.9
Air Force	2341	14.5
Total	16 163
Years of military service
≤5	5244	32.4
6 to 15	6255	38.6
≥16	4692	29.0
Total	16 191
Mental health care history
Never	12 644	79.2
Past care only	2095	13.1
Current care	1228	7.7
Total	15 967
Number of previous deployments
0	7530	46.9
1 or 2	5034	31.4
≥3	3481	21.7
Total	16 045
Deployment length, months
<2	324	2.0
2 to 5	1975	12.2
6 to 8	12 651	78.2
≥9	1229	7.6
Total	16 179
Combat exposurec
First	4545	28.4
Second	6409	40.1
Third	5046	31.5
Total	16 000
Lag time between return from deployment and questionnaire completion, days
<90	2130	13.3
90 to 180	9426	59.1
181 to 365	3585	22.4
≥366	822	5.2
Total	15 963

^a≥Sergeant or equivalent

^b≥Master Corporal or equivalent

^cTertiles were determined with respect to a larger reference population, accounting for the unequal distribution.

NCM = noncommissioned member

As shown in Table 2, 10.2% had 1 or more of the 6 mental health problems assessed. Among individual problems, the most prevalent were symptoms of MDD (3.2%), minor depression (3.3%), and PTSD (2.8%). Comorbidity (having more than 1 of the 6 problems) was seen in 38.6% of respondents with 1 or more problems.

Table 2.

Prevalence of mental health problems (overall N = 16 193)

Symptomsa	n/N	%
PTSD	447/16 155	2.7
MDD	517/16 112	3.2
Minor depression	527/16 108	3.3
Suicidal ideation	390/16 100	2.4
Panic disorder	293/16 027	1.8
GAD	303/15 854	1.9
Any of 6 mental health problems	1602/15 672	10.2

^aSee Methods section for case definitions.

GAD = generalized anxiety disorder;

MDD = major depressive disorder

PTSD = posttraumatic stress disorder

Respondent characteristics having a univariate relation with the primary outcome of any mental health problem included the following: female sex; aged 24 years or older; anglophone; Regular Force (compared with Reserve Force); being a junior NCM; and being widowed, divorced, or separated. Protective factors included being single, an officer, and in the Air Force. Total years of military service were not associated with the primary outcome. Among predeployment factors, past mental health care had a strong univariate association with mental health problems, but number of previous operational deployments did not. Among intradeployment factors, combat exposure had a substantial univariate association with postdeployment mental health problems, but deployment duration did not. Finally, among postdeployment factors, respondents screened before 90 days after return had a lower univariate risk of postdeployment mental health problems. Current mental health care was unsurprisingly a strong correlate.

Online eTable 3 also shows the adjusted odds ratios for the variables having a univariate relation with the primary outcome. Female sex had a small, independent association with the primary outcome (AOR 1.48; 95% CI 1.26 to 1.73). Francophones remained at significantly lower risk for postdeployment mental health problems (AOR 0.63; 95% CI 0.54 to 0.74), as did officers (AOR 0.54; 95% CI 0.44 to 0.67). Being widowed, divorced, or separated as a risk factor persisted after adjustment (AOR 1.25; 95% CI 1.03 to 1.52). Air Force personnel remained at lower risk (AOR 0.82; 95% CI 0.67 to 0.99). Adjustment altered the relation between the primary outcome and age: people aged 35 to 44 years and people aged 45 years or older had elevated risk (AOR 1.20; 95% CI 1.02 to 1.40 and AOR 1.53; 95% CI 1.23 to 1.90, respectively). After adjustment, there was no longer any association between Reservist status and the primary outcome. Among predeployment factors, past mental health care remained a strong, independent predictor (AOR 2.89; 95% CI 2.49 to 3.35). Among intradeployment factors, heavy combat exposure remained a strong predictor of postdeployment mental health (AOR for first, compared with third tertile 2.57, 95% CI 2.16 to 3.06). Among postdeployment factors, the strongest correlate was current mental health care (AOR 12.52; 95% CI 10.82 to 14.49). People screened less than 90 days after return from deployment remained at slightly lower risk (AOR 0.75; 95% CI 0.62 to 0.92).

Discussion

Summary of Key Findings

An important minority (10.2%) of CF personnel who deployed in support of the mission in Afghanistan reported 1 or more of 6 common mental health problems; the most common conditions were symptoms of MDD, minor depression, and PTSD. Logistic regression modelling largely confirmed univariate findings on risk factors, but nearly all variables were weak independent predictors, with odds ratios being 1.5 or below. However, past and current mental health care and heavy combat were strong predictors.

Comparison With Other Literature

Prevalence Rates

Methodological and contextual differences (ably reviewed elsewhere^38–40) make it difficult to interpret differences in prevalence rates. For example, the most reliable prevalence rates for CF personnel come from the 2002 CF Supplement to the CCHS 1.2, which reported a 12-month prevalence rate for PTSD of 2.2%,²⁹ compared with the 2.8% point prevalence seen in our study. Key differences between these studies include the use of very different assessment tools, recall periods, and survey modes (a computer-assisted personal interview with Statistics Canada personnel, compared with a paper-and-pencil survey done for clinical screening purposes). Differences in the study populations also erode comparability (for example, in-garrison, compared with postdeployment population, sociodemographic differences in the samples, and differences in the year of the surveys). That said, the fundamental story told by our study coheres with those of many other studies on CF and allied personnel. In the aftermath of a trauma-laden deployment, most personnel are well, but an important minority have mental health problems, including many other than PTSD.

Risk Factors for Mental Health Problems

Risk factors for mental health problems are presumably less sensitive to the methodological problems that complicate prevalence rate comparisons. For example, women are consistently noted to be at increased risk for depression, independent of the recall period, the instrument, and cutoff used, and the subpopulation studied.⁴¹ Our findings on independent risk factors for postdeployment mental health problems largely mirror those of other work showing better health in Air Force personnel and officers.^1,42 Consistent with the larger literature on sex and mood and anxiety disorders, we found a significantly increased risk in female personnel.⁴¹ Other military research has been inconsistent about this finding.⁴³ However, we did not see the significantly increased risk of postdeployment mental health problems in Reserve Force personnel that is consistently seen in the United States²⁴ and the United Kingdom.²³ However, this observation is consistent with recent Canadian findings.^1,31 Our data showed an increased risk in older personnel, which has not been consistently noted in other studies. The magnitude of this association was small, thus our generous sample size and substantial numbers of older personnel may have enhanced our ability to detect such an association. Our finding of better mental health in Francophones stands in contrast to civilian data suggesting, if anything, worse mental health for Francophones in Canada.^44–46 Given that Francophones in the CF are concentrated in a limited number of bases, each with its own culture, operational history, and deployment experiences, this difference may be due to nonlinguistic factors.

Among predeployment factors, our finding of a strong association of past mental health care (a surrogate for predeployment mental health) and postdeployment mental health is consistent with other findings.¹⁰ Our finding of no association with number of previous deployments coheres with recent Canadian findings^1,31 but stands in contrast to largely US intradeployment data showing strong effects, at least for combat deployments,³⁴ and smaller and less consistent effects postdeployment.^12,13,15

Among intradeployment factors, we confirmed the findings of many other intra-³¹ and postdeployment¹⁷ studies showing a strong association of heavy combat exposure and mental health. Our finding of no association with deployment length is consistent with Canadian findings^1,31 but stands in contrast to US and UK findings.^14,18 American studies⁷ and Canadian findings from earlier in the conflict⁴⁷ have shown a marked increase in the apparent prevalence of mental health problems during the first year after return, but we saw only a slightly lower risk for respondents screened before 90 days. We hypothesize that differences in mental health related stigma⁴⁸ or changing attitudes toward mental health problems over time may have facilitated earlier reporting of symptoms.

Limitations

The primary limitation of our study is that the underlying data were collected during a confidential but not anonymous screening process. Data from elsewhere suggest a powerful effect of anonymity on the results of postdeployment screening questionnaires.^49–51 The prevalence rates reported here are thus likely a systematic underestimate of the true prevalence at the time of screening. Prevalence rates reported at a single point in time soon after return from deployment will significantly underestimate the true burden of deployment-related mental disorders.¹

The cross-sectional nature of the data means that the observed associations may not reflect causality. Nevertheless, the most important associations we have reported here have indeed been substantiated in longitudinal research.^10,42 We used past care as a surrogate measure of past mental health, for which it is, at best, an approximation. We did not control for postdeployment life stressors,²⁰ homecoming experiences,²¹ or leadership,¹⁹ which may have led to some residual confounding. Using an aggregate outcome of any mental health problem simplified analysis and interpretation, but it may have obscured important differences in risk factor profiles for the individual conditions. Finally, we could not assess potential differences between people who did and did not complete the screening process.

Strengths of our study include its Canadian context, the large sample size, the ability to control for many potential confounders, and the breadth and precision of the survey instruments used relative to other studies^25,52 using screening data.

Implications

The sizeable fraction willing to report mental health problems during postdeployment screening challenges the criticism that such screening is valueless because no one will be willing to disclose problems owing to stigma.⁵³ While we are certain that some minimization of symptoms occurred, the willingness of 10% of personnel to give voice to their concerns provides an opportunity to help at least that fraction. There is also evidence that even among people who do not report problems or for whom follow-up is not recommended, screening can still increase care-seeking in its immediate aftermath.⁵² Unfortunately, our study design prevented us from drawing any conclusions about the impact of screening on well-being or service use.

Our study did identify risk factors for postdeployment mental health problems, but most were small in magnitude and essentially unmodifiable, hence of limited practical value. The greater risk in NCMs relative to officers argues for increased attention to the needs of this group. Indeed, this group has been a primary target of mental health and resilience training in the CF. The higher risk in widowed, divorced, and separated personnel speaks to the need for incremental efforts for outreach and education that specifically target their needs. However, until very recently, support efforts have instead focused on married personnel. Targeting groups based on heavy combat exposure in particular is appealing given its particularly strong association with mental health problems. The slightly lower risk of mental health problems in people who were screened less than 90 days after return provides support for the recommended timing of screening (90 to 180 days after return).

Past mental health care is a proxy for predeployment mental health, which is known to be a risk factor for postdeployment mental health problems.¹⁰ While the odds ratio was substantial (2.89), many people who had previously sought care were in good postdeployment mental health. Thus selecting people for deployment on this basis alone would result in the unnecessary exclusion of many people who would do well. Conversely, such people might be offered interventions to prevent relapse before, during, or after their deployment. The strong association between current mental health care and postdeployment mental health problems simply demonstrates that people in care are there for a good reason.

Additional research is needed to understand which people with a history of mental health problems or mental health care will do well postdeployment. Research is also needed to determine if the higher risk of postdeployment mental health problems in people with a history of mental health problems in the past is due to an increased vulnerability to deployment-related adversity or if it merely reflects a carrying forward of their underlying risk for mental disorders independent of deployment. Other interactions between adversity and characteristics of personnel (for example, sex) are also of interest. Finally, research on the impact of mental health screening in military personnel is sorely needed.⁵³

The absence of an observed higher risk in CF Reservists (seen, in contrast, in both the United States²⁴ and United Kingdom²³) has no immediate explanation. Potential factors include various differences among the Reservists in different countries (for example, their sociodemographic characteristics, how they are trained, how they are selected for deployment, their deployment experiences, and the mental health services and supports available to them when they return). The absence of a higher risk in people with multiple deployments and longer deployments (again, seen elsewhere^14,18) suggests that the totality of the CF’s policies regulating operational tempo and the mental health services and supports in place are working as they should. However, we do not have sufficient data to explore which factors account for this finding.

Conclusion

With war comes the inevitability of psychiatric injury in an important minority of people who serve. This inevitability is tempered by our findings that potential excess risk in Reservists, those with multiple deployments, and those with longer deployments was not seen in the CF. The divergence between our findings and those in other military populations speaks to the need for caution in applying the findings of one military organization to another when it comes to postdeployment mental health problems.

Abbreviations

CF

Canadian Forces

CCHS 1.2

Canadian Community Health Survey—Mental Health and Well-Being

EPDS

enhanced postdeployment screening

GAD

generalized anxiety disorder

MDD

major depressive disorder

NCM

noncommissioned member

PCL-C

PTSD Checklist—Civilian Version

PHQ

Patient Health Questionnaire

PTSD

posttraumatic stress disorder