Abstract

Background Unit cohesion is recognized as a potentially modifiable factor in the aetiology of mental illness among military personnel.

Aims To examine the association between unit cohesion and probable post-traumatic stress disorder (PTSD), common mental disorder and alcohol misuse, in UK armed forces personnel deployed to Iraq.

Methods A sample of 4901 male UK armed forces personnel who had deployed to Iraq was drawn from a cohort of personnel who participated in a cross-sectional postal questionnaire study between June 2004 and March 2006. Information was collected on socio-demographic and military characteristics, deployment experiences and information on current health.

Results Perceived interest from seniors was associated with less probable PTSD [odds ratio (OR) 0.42, 95% confidence interval (CI) 0.26–0.67] and common mental disorder (OR 0.68, 95% CI 0.53–0.87). Among regular personnel, feeling well informed was associated with less common mental disorder (OR 0.74, 95% CI 0.58–0.95) and comradeship was associated with greater alcohol misuse (OR 1.98, 95% CI 1.19–3.28). Feeling able to talk about personal problems was associated with less alcohol misuse among reserve personnel (OR 0.31, 95% CI 0.16–0.60). The general construct of unit cohesion was predictive of less probable PTSD (OR 0.69, 95% CI 0.58–0.81) and common mental disorder (OR 0.80, 95% CI 0.73–0.87).

Conclusions Unit cohesion had a linear association with less probable PTSD and common mental disorder. Of the individual items, perception of leadership was associated with less probable PTSD and common mental disorder. Comradeship was associated with greater alcohol misuse among regular personnel, while feeling able to talk about personal problems was associated with less alcohol misuse for reserve personnel.

Introduction

Unit cohesion describes the process that, to varying degrees, unites or keeps together members of any group [1]. Despite recognition of unit cohesion as an important phenomenon for organizational functioning, there is considerable divergence in how unit cohesion has been defined and measured [2,3]. The conceptualization of unit cohesion in the military literature varies from the use of individual items, to assess how cohesive or unified a unit is, to multidimensional constructs that assess peer bonding, bonding with leaders, task support and emotional support [2,4,5].

The effect of cohesion on individual and group performance has been well researched in both civilian and military fields, and it has long been considered that cohesion within military units enhances performance and operational effectiveness [2]. Unit cohesion has also been recognized as a protective factor for troops in the aetiology of mental health problems [6]. Wessely [7] describes how the conventional wisdom of the last century has been ‘that soldiers are supported by and support the small group in which they live and fight and that breakdown becomes more likely when they cease to be part of that group’.

Following the long-term commitment of the UK armed forces to conflicts such as those in Iraq and Afghanistan, increasing attention has been paid in recent years to the potential effects on the mental health of servicemen and women [8,9]. Unit cohesion has been recognized as a potentially modifiable factor in the aetiology of mental illness and has been found to be a protective factor of post-traumatic stress disorder (PTSD), common mental disorder and physical ill health in military personnel [3,6,8,10,11]. In contrast, little has been written about the impact of unit cohesion on alcohol misuse in military personnel [12]. Alcohol has long been used by military personnel as a bonding tool, as well as a coping device [13]. Browne et al. [12] have shown that high levels of comradeship and low levels of leadership were linked with heavy drinking.

The association between unit cohesion and mental health outcomes has also been shown to differ depending on how unit cohesion is measured. For instance, Rona et al. [9] found that there was a borderline non-significant association between a multidimensional measure of unit cohesion and PTSD; yet, high levels of comradeship had a significant association with lower levels of PTSD. Furthermore, there is some disparity in the literature as to the type of relationship, whether there is a linear or curvilinear association, between unit cohesion and PTSD.

We have previously reported on factors associated with PTSD and alcohol misuse, including cohesion and comradeship [8,12] and on pre-enlistment and military factors that are predictive of unit cohesion and morale [14]. In this paper, we examine the protective effect of unit cohesion, measured with separate items and as a general construct, on probable PTSD, common mental disorder and alcohol misuse. We also examine whether unit cohesion serves as a moderator between deployment exposures and the mental health outcomes. Finally, we examine whether there is a non-linear trend for the associations between the unit cohesion construct and the mental health outcomes.

Methods

This study utilized data from the first phase of a longitudinal cohort study of UK armed forces personnel undertaken by the King’s Centre for Military Health Research (KCMHR) [15]. The data were collected with self-report questionnaires distributed through a postal survey and visits to military bases. The sample included 4722 personnel who had deployed on Operation TELIC 1 (TELIC is the UK codename for operations in Iraq and TELIC 1 denotes the initial phase of the Iraq war in 2003, defined as 18 January 2003 to 28 April 2003) and 5500 personnel who served in the military at this time but had not deployed to Iraq (the ‘ERA’ group). Of the ERA group, a number were subsequently deployed on later TELIC deployment (TELIC 2–6). The cohort study had a response rate of 59%.

The analytical sample consists of 4901 male regular and reserve (personnel who have enlisted as volunteers) service personnel from the Royal Navy, Royal Marines, British Army and Royal Air Force who were deployed to Iraq on any TELIC operation and who had responded to at least two of the four questions about unit cohesion. Women were excluded due to small numbers.

Participants answered questions about military and deployment factors, lifestyle factors and health outcomes. Variables included socio-demographic and military factors [rank, service and enlistment status (regulars or reserves) and marital status], childhood adversity measured as a composite score of 16 adverse childhood events (including questions such as coming from a close family, playing truant from school or being hit by parents or caregivers regularly), [16] relationship with family while deployed (reporting major problems at home and having no perceived family support), being deployed to a forward area (zone of confrontation), perception of work carried out in theatre (theatre of deployment is the geographical area where armed conflict is conducted as part of military operations), in theatre experience, which was measured as a composite score of eight factors such as discharging weapon in direct combat, seeing personnel wounded or killed, experiencing landmine strikes, coming under mortar or artillery fire or experiencing hostility from civilians; thinking they might be killed and difficulties adjusting to being back home [8,12].

Unit cohesion was measured with four variables: ‘I felt a sense of comradeship between myself and other people in my unit’, ‘I could have gone to most people in my unit if I had a personal problem’, ‘my seniors were interested in what I did or thought’ and ‘I felt well informed about what was going on’. The variables were measured on a five-point scale from strongly disagree to strongly agree. The responses were collapsed into a three-point scale (agree, neither and disagree) to avoid small cell counts for analyses of the mental health outcomes; this also fits with how these items have been used in previous work on this cohort [8,12].

For the purpose of generating the unit cohesion construct the ‘neither’ responses were recoded as missing. This decision was taken as a response of neither is a type of non-committal response and may reflect a valid absence of opinion or belief or an inaccurate reflection of the person’s opinion [17]. Therefore, including a neither category may not result in a truly ordinal scale. Recoding of the central category as missing improved the scale reliability from a Cronbach’s alpha of 0.74 to an alpha of 0.79. We also carried out analyses that included the neither category, which showed that excluding the central category had little impact on the general construct for unit cohesion (data not shown).

The unit cohesion construct was generated through principal component analysis (PCA) of a polychoric correlation matrix with pairwise deletion. The polychoric correlations were moderate to large (0.43–0.71). The PCA resulted in a one-factor solution and the general factor explained 67% of the total variance. All variables loaded on the general factor, with factor loadings ranging between 0.68 and 0.79. The factor loadings were standardized to a mean of 0 and a SD of 1.

PTSD was measured with the Post-traumatic Stress Disorder Checklist (PCL); cases were defined as those scoring ≥50 [18]. Common mental disorder was assessed with the 12 item General Health Questionnaire (GHQ-12), with cases defined as those scoring ≥4 [19]. Alcohol misuse was assessed with the Alcohol Use Disorders Identification Test (AUDIT) with cases defined as those scoring ≥16 [20].

All analyses were carried out in STATA 10 (Stata Corporation, College Station, TX, USA). Frequencies and percentages, or means and standard deviations for continuous variables, were calculated to describe the sample and the unit cohesion items. The associations between both the individual unit cohesion items and the unit cohesion construct and the mental health outcomes (probable PTSD, common mental disorder and alcohol misuse) were examined using logistic regression. Stratified analyses by enlistment status are reported for alcohol misuse but not for PTSD and common mental disorder due to small numbers of reserves and few differences in the associations with unit cohesion. The logistic regression analyses were adjusted for factors previously shown to be associated with the mental health outcomes and variables that may impact on unit cohesion, such as having major problems at home and feeling that one did not receive enough personal support from family during deployment (Table 1) [8,12]. Sample weights were adjusted for with the STATA survey commands (svy) or with the ‘pweight’ option. The moderating effect of unit cohesion was examined with an interaction effect for the general construct of unit cohesion and in theatre experiences. The associations between the unit cohesion construct and the three mental health outcomes were examined for non-linear trends with Box Tidwell regression analyses and graphically with lowess smoothing plots. The Box Tidwell analysis reports the difference in deviance between a continuous-power polynomial model for a predictor and a model with the linear predictor and indicates whether the relationship between the predictor and the outcome is linear.

Table 1.

Sample descriptives and unit cohesion variables (n = 4901), frequency (n) and weighted percentages (%), or weighted mean and standard deviation (SD)

Variable N (%) or mean (SD)a 
Serving status  
    Regular 4194 (92) 
    Reserve 707 (8) 
Service  
    Navy 426 (9) 
    Marines 264 (6) 
    Army 3317 (67) 
    RAF 894 (19) 
Age (years) 32.4 (7.5)a 
Educational statusb  
    None 404 (9) 
    GCSEs 2090 (46) 
    A level 1393 (30) 
    Degree 745 (15) 
Marital status  
    Relationship 3733 (76) 
    Single 870 (18) 
    Ex-relationship 281 (6) 
Childhood adversity  
    0/1 1113 (23) 
    2/3 1587 (32) 
    4/5 996 (20) 
    ≥6 1205 (25) 
Rank  
    Other ranks 999 (21) 
    Junior non-commissioned officer 1743 (35) 
    Senior non-commissioned officer 1330 (28) 
    Officer 790 (16) 
Had major problems at home during deployment 1341 (27) 
Did not receive enough personal support from family 158 (3) 
Time spent in a forward area (zone of confrontation)  
    No time 1925 (40) 
    Up to 1 month 1538 (32) 
    >1 month 1310 (28) 
Work in theatre (area of deployment) was generally outside experience/ability 819 (17) 
In theatre experience  
    0/1 1890 (39) 
    2/3 1253 (25) 
    4/5 839 (17) 
    ≥6 919 (19) 
Thinking one might be killed 2855 (58) 
Found it difficult to adjust to being back home  
    Strongly disagree 1155 (26) 
    Disagree 1975 (42) 
    Agree 1242 (25) 
    Strongly agree 405 (8) 
Felt a sense of comradeship with other people in the unit  
    Disagree 303 (6) 
    Neither agree nor disagree 464 (10) 
    Agree 4128 (85) 
Could have approached most people within the unit with personal problem  
    Disagree 1380 (28) 
    Neither agree nor disagree 1025 (21) 
    Agree 2491 (51) 
Seniors were interested in my actions and thoughts  
    Disagree 1209 (24) 
    Neither agree nor disagree 862 (18) 
    Agree 2822 (58) 
Felt well informed about what was going on  
    Disagree 1339 (27) 
    Neither agree nor disagree 773 (16) 
    Agree 2787 (58) 
General construct of unit cohesion 0.01 (1.0)a 
Variable N (%) or mean (SD)a 
Serving status  
    Regular 4194 (92) 
    Reserve 707 (8) 
Service  
    Navy 426 (9) 
    Marines 264 (6) 
    Army 3317 (67) 
    RAF 894 (19) 
Age (years) 32.4 (7.5)a 
Educational statusb  
    None 404 (9) 
    GCSEs 2090 (46) 
    A level 1393 (30) 
    Degree 745 (15) 
Marital status  
    Relationship 3733 (76) 
    Single 870 (18) 
    Ex-relationship 281 (6) 
Childhood adversity  
    0/1 1113 (23) 
    2/3 1587 (32) 
    4/5 996 (20) 
    ≥6 1205 (25) 
Rank  
    Other ranks 999 (21) 
    Junior non-commissioned officer 1743 (35) 
    Senior non-commissioned officer 1330 (28) 
    Officer 790 (16) 
Had major problems at home during deployment 1341 (27) 
Did not receive enough personal support from family 158 (3) 
Time spent in a forward area (zone of confrontation)  
    No time 1925 (40) 
    Up to 1 month 1538 (32) 
    >1 month 1310 (28) 
Work in theatre (area of deployment) was generally outside experience/ability 819 (17) 
In theatre experience  
    0/1 1890 (39) 
    2/3 1253 (25) 
    4/5 839 (17) 
    ≥6 919 (19) 
Thinking one might be killed 2855 (58) 
Found it difficult to adjust to being back home  
    Strongly disagree 1155 (26) 
    Disagree 1975 (42) 
    Agree 1242 (25) 
    Strongly agree 405 (8) 
Felt a sense of comradeship with other people in the unit  
    Disagree 303 (6) 
    Neither agree nor disagree 464 (10) 
    Agree 4128 (85) 
Could have approached most people within the unit with personal problem  
    Disagree 1380 (28) 
    Neither agree nor disagree 1025 (21) 
    Agree 2491 (51) 
Seniors were interested in my actions and thoughts  
    Disagree 1209 (24) 
    Neither agree nor disagree 862 (18) 
    Agree 2822 (58) 
Felt well informed about what was going on  
    Disagree 1339 (27) 
    Neither agree nor disagree 773 (16) 
    Agree 2787 (58) 
General construct of unit cohesion 0.01 (1.0)a 

Due to missing data, values do not add up to the total.

a

Mean and standard deviation.

b

General Certificate of Secondary Education (GCSEs) are examinations usually taken at age 16 years. A levels are usually taken at age 18 years and are required for entry to university.

Results

The majority of the sample was regular personnel who served in the Army (Table 1). Fifty-five per cent of the sample were educated to General Certificate of Secondary Education standard (examinations usually taken at age 16 years) or had no qualifications and 76% were in a relationship. The level of unit cohesion, as measured with the individual items, was generally high. In particular, perception of comradeship was high, with 85% of the sample agreeing with this statement.

When adjusted for covariates, all individual unit cohesion variables were associated with a lower risk of probable PTSD (Table 2). However, after adjusting for all individual unit cohesion items, only perceived senior interest remained significant. There was no indication of a difference in the patterns of associations between the unit cohesion measures and PTSD for regular and reserve personnel (data available from the authors).

Table 2.

Association between the unit cohesion items and probable PTSD and common mental disorder, ORs and 95% CIs

Measure Probable PTSD OR (95% CI)
 
Common mental disorder OR (95% CI)
 
Unadjusted Adjusteda Adjustedb Unadjusted Adjusteda Adjustedb 
Felt a sense of comradeship with other people in the unitc 
    Neither agree nor disagree 0.44 (0.25–0.78)** 0.53 (0.24–1.15) 0.63 (0.29–1.39) 0.51 (0.36–0.71)*** 0.62 (0.41–0.92)* 0.69 (0.46–1.04) 
    Agree 0.34 (0.23–0.52)*** 0.50 (0.29–0.86)* 0.92 (0.50–1.71) 0.43 (0.33–0.55)*** 0.53 (0.39–0.73)*** 0.71 (0.50–1.01) 
Could have approached most people within the unit with personal problemc 
    Neither agree nor disagree 0.50 (0.34–0.73)*** 0.71 (0.45–1.13) 0.81 (0.49–1.33) 0.70 (0.57–0.86)*** 0.86 (0.68–1.10) 1.02 (0.79–1.31) 
    Agree 0.39 (0.29–0.53)*** 0.47 (0.32–0.69)*** 0.67 (0.43–1.10) 0.60 (0.51–0.71)*** 0.71 (0.58–0.87)** 0.94 (0.75–1.18) 
Seniors were interested in my actions and thoughtsc 
    Neither agree nor disagree 0.53 (0.36–0.77)*** 0.64 (0.40–1.01) 0.70 (0.43–1.15) 0.64 (0.52–0.79)*** 0.75 (0.58–0.96)* 0.85 (0.65–1.11) 
    Agree 0.24 (0.18–0.34)*** 0.34 (0.23–0.50)*** 0.42 (0.26–0.67)*** 0.42 (0.35–0.49)*** 0.55 (0.45–0.67)*** 0.68 (0.53–0.87)** 
Felt well informed about what was going onc 
    Neither agree nor disagree 0.69 (0.46–1.02) 0.83 (0.51–1.36) 1.03 (0.62–1.71) 0.72 (0.58–0.89)** 0.87 (0.67–1.12) 0.98 (0.75–1.28) 
    Agree 0.36 (0.26–0.49)*** 0.55 (0.38–0.80)** 0.84 (0.56–1.27) 0.46 (0.39–0.55)*** 0.62 (0.51–0.76)*** 0.77 (0.62–0.97)* 
General construct of unit cohesion 0.61 (0.53–0.70)*** 0.69 (0.58–0.81)***  0.72 (0.66–0.77)*** 0.80 (0.73–0.87)***  
Measure Probable PTSD OR (95% CI)
 
Common mental disorder OR (95% CI)
 
Unadjusted Adjusteda Adjustedb Unadjusted Adjusteda Adjustedb 
Felt a sense of comradeship with other people in the unitc 
    Neither agree nor disagree 0.44 (0.25–0.78)** 0.53 (0.24–1.15) 0.63 (0.29–1.39) 0.51 (0.36–0.71)*** 0.62 (0.41–0.92)* 0.69 (0.46–1.04) 
    Agree 0.34 (0.23–0.52)*** 0.50 (0.29–0.86)* 0.92 (0.50–1.71) 0.43 (0.33–0.55)*** 0.53 (0.39–0.73)*** 0.71 (0.50–1.01) 
Could have approached most people within the unit with personal problemc 
    Neither agree nor disagree 0.50 (0.34–0.73)*** 0.71 (0.45–1.13) 0.81 (0.49–1.33) 0.70 (0.57–0.86)*** 0.86 (0.68–1.10) 1.02 (0.79–1.31) 
    Agree 0.39 (0.29–0.53)*** 0.47 (0.32–0.69)*** 0.67 (0.43–1.10) 0.60 (0.51–0.71)*** 0.71 (0.58–0.87)** 0.94 (0.75–1.18) 
Seniors were interested in my actions and thoughtsc 
    Neither agree nor disagree 0.53 (0.36–0.77)*** 0.64 (0.40–1.01) 0.70 (0.43–1.15) 0.64 (0.52–0.79)*** 0.75 (0.58–0.96)* 0.85 (0.65–1.11) 
    Agree 0.24 (0.18–0.34)*** 0.34 (0.23–0.50)*** 0.42 (0.26–0.67)*** 0.42 (0.35–0.49)*** 0.55 (0.45–0.67)*** 0.68 (0.53–0.87)** 
Felt well informed about what was going onc 
    Neither agree nor disagree 0.69 (0.46–1.02) 0.83 (0.51–1.36) 1.03 (0.62–1.71) 0.72 (0.58–0.89)** 0.87 (0.67–1.12) 0.98 (0.75–1.28) 
    Agree 0.36 (0.26–0.49)*** 0.55 (0.38–0.80)** 0.84 (0.56–1.27) 0.46 (0.39–0.55)*** 0.62 (0.51–0.76)*** 0.77 (0.62–0.97)* 
General construct of unit cohesion 0.61 (0.53–0.70)*** 0.69 (0.58–0.81)***  0.72 (0.66–0.77)*** 0.80 (0.73–0.87)***  
a

Adjusted for serving status, service, age, rank, education, marital status, childhood adversity, major problems at home, family support, being in a forward area, work in theatre, in theatre experience, thinking one might be killed and found it difficult to adjust to being back home.

b

Adjusted for the variables in a and the other unit cohesion items listed in the table.

c

Disagree is the reference category.

Significance *P < 0.05; **P < 0.01; ***P < 0.001.

When adjusted for socio-demographic factors and deployment experiences, all individual unit cohesion variables were associated with a lower risk of common mental disorder (Table 2). Perceived senior interest and feeling well informed about what was going on remained significant after adjusting for all individual unit cohesion items. Stratified analyses by enlistment status showed that feeling well informed was associated with less common mental disorder among regular but not among reserve personnel [regulars odds ratio (OR_ = 0.60, 95% CI 0.48–0.74; reserves OR = 0.76, 95% CI 0.48–1.20). All other associations were similar between regular and reserve personnel.

Perceived senior interest and feeling well informed about what was going on were associated with lower levels of alcohol misuse for regular personnel (Table 3). After adjusting for all unit cohesion items, there was an association between higher perceived levels of comradeship and alcohol misuse. For reserve personnel, feeling able to approach most people in the unit with a personal problem was associated with lower levels of alcohol misuse (Table 3). The fully adjusted models were not fitted for reserve personnel due to small numbers.

Table 3.

Association between the unit cohesion items and alcohol misuse stratified by regular and reserve personnel, ORs and 95% CI

Measure Alcohol misuse OR (95% CI)
 
Regular personnel
 
Reserve personnel
 
Unadjusted Adjusteda Adjustedb Unadjusted Adjusteda 
Felt a sense of comradeship with other people in the unitc 
    Neither agree nor disagree 1.48 (0.95–2.33) 1.72 (0.98–3.01) 1.84 (1.05–3.22)* 0.77 (0.30–1.97) 0.75 (0.25–2.31) 
    Agree 1.43 (0.97–2.09) 1.61 (0.99–2.61) 1.98 (1.19–3.28)** 0.66 (0.34–1.30) 0.81 (0.35–1.86) 
Could have approached most people within the unit with personal problemc 
    Neither agree nor disagree 1.01 (0.81–1.26) 1.03 (0.78–1.36) 1.00 (0.75–1.33) 0.56 (0.30–1.04) 0.43 (0.20–0.91)* 
    Agree 0.92 (0.77–1.11) 0.84 (0.67–1.06) 0.87 (0.68–1.12) 0.40 (0.24–0.68)** 0.31 (0.16–0.60)*** 
Seniors were interested in my actions and thoughtsc 
    Neither agree nor disagree 0.83 (0.66–1.05) 0.92 (0.70–1.22) 0.91 (0.68–1.22) 0.66 (0.33–1.30) 0.72 (0.31–1.66) 
    Agree 0.59 (0.49–0.71)*** 0.75 (0.60–0.94)* 0.78 (0.60–1.02) 0.64 (0.39–1.05) 0.91(0.48–1.72) 
Felt well informed about what was going onc 
    Neither agree nor disagree 0.89 (0.70–1.12 0.91 (0.68–1.21) 0.95 (0.71–1.28) 0.87 (0.46–1.65) 0.77 (0.34–1.73) 
    Agree 0.65 (0.54–0.78)*** 0.78 (0.63–0.97)* 0.84 (0.66–1.08) 0.61 (0.37–1.02) 0.58 (0.31–1.09) 
General construct of unit cohesion 0.92 (0.85–1.00)* 0.96 (0.87–1.06)  0.77 (0.63–0.95)* 0.80 (0.61–1.04) 
Measure Alcohol misuse OR (95% CI)
 
Regular personnel
 
Reserve personnel
 
Unadjusted Adjusteda Adjustedb Unadjusted Adjusteda 
Felt a sense of comradeship with other people in the unitc 
    Neither agree nor disagree 1.48 (0.95–2.33) 1.72 (0.98–3.01) 1.84 (1.05–3.22)* 0.77 (0.30–1.97) 0.75 (0.25–2.31) 
    Agree 1.43 (0.97–2.09) 1.61 (0.99–2.61) 1.98 (1.19–3.28)** 0.66 (0.34–1.30) 0.81 (0.35–1.86) 
Could have approached most people within the unit with personal problemc 
    Neither agree nor disagree 1.01 (0.81–1.26) 1.03 (0.78–1.36) 1.00 (0.75–1.33) 0.56 (0.30–1.04) 0.43 (0.20–0.91)* 
    Agree 0.92 (0.77–1.11) 0.84 (0.67–1.06) 0.87 (0.68–1.12) 0.40 (0.24–0.68)** 0.31 (0.16–0.60)*** 
Seniors were interested in my actions and thoughtsc 
    Neither agree nor disagree 0.83 (0.66–1.05) 0.92 (0.70–1.22) 0.91 (0.68–1.22) 0.66 (0.33–1.30) 0.72 (0.31–1.66) 
    Agree 0.59 (0.49–0.71)*** 0.75 (0.60–0.94)* 0.78 (0.60–1.02) 0.64 (0.39–1.05) 0.91(0.48–1.72) 
Felt well informed about what was going onc 
    Neither agree nor disagree 0.89 (0.70–1.12 0.91 (0.68–1.21) 0.95 (0.71–1.28) 0.87 (0.46–1.65) 0.77 (0.34–1.73) 
    Agree 0.65 (0.54–0.78)*** 0.78 (0.63–0.97)* 0.84 (0.66–1.08) 0.61 (0.37–1.02) 0.58 (0.31–1.09) 
General construct of unit cohesion 0.92 (0.85–1.00)* 0.96 (0.87–1.06)  0.77 (0.63–0.95)* 0.80 (0.61–1.04) 
a

Adjusted for serving status, service, age, rank, education, marital status, childhood adversity, major problems at home, family support, being in a forward area, work in theatre, in theatre experience, thinking one might be killed and found it difficult to adjust to being back home.

b

Adjusted for the variables in a and the other unit cohesion items listed in the table.

c

Disagree is the reference category.

Significance *P < 0.05; **P < 0.01; ***P < 0.001.

The general construct of unit cohesion was predictive of lower levels of probable PTSD and common mental disorder after adjustment for socio-demographic factors and deployment experiences. However, there was no association between the general construct of unit cohesion and alcohol misuse after adjusting for covariates; childhood adversity explained the majority of this association. These results were comparable for regular and reserve personnel.

In order to examine whether unit cohesion served as a moderator between deployment exposures and the health outcomes, an interaction term was fitted between the general construct of unit cohesion and the in theatre experience variable in the logistic regression models. There was no support for a moderating effect of unit cohesion on any of the health outcomes (probable PTSD F = 1.39, P = NS; common mental disorder F = 0.28, P = NS; alcohol misuse F = P = NS).

The results from the Box Tidwell regression analyses showed that there was no indication of a non-linear association between the unit cohesion construct and the three health outcomes (probable PTSD non-linear deviation = 1.08, P = NS; common mental disorder non-linear deviation = 0.05, P = NS; alcohol misuse non-linear deviation = 1.06, P = NS).

Discussion

Unit cohesion was associated with lower levels of probable PTSD and common mental disorder in UK troops who were deployed to Iraq. Associations between the individual unit cohesion items and PTSD and common mental disorder were similar between regular and reserve personnel, but there were several differences between the groups in the patterns of associations with alcohol misuse. Perceived senior interest was associated with lower levels of probable PTSD and common mental disorder, highlighting the importance of caring leadership as a protective factor. Feeling well informed about what was going on was also a protective factor for common mental disorder among regular personnel.

There were also associations between the general construct of unit cohesion and lower levels of probable PTSD and common mental disorder but not with alcohol misuse. There was no indication of a moderating effect of unit cohesion on the association between in theatre experience and any of the three health outcomes. The results showed that the associations between unit cohesion and the three health outcomes were linear.

This study utilizes a large sample, taken from a cohort representative of the UK military, which had a good response rate (59%). Limitations of the present study include the exclusion of women due to small numbers, and further research is needed to assess the effect of unit cohesion upon mental health outcomes within this group. This study relied on self-report data to assess health outcomes, which represent probable outcomes rather than actual diagnosis. However, all health outcomes were measured with well-validated questionnaires. The general unit cohesion construct has not been validated but it had good face validity and satisfactory reliability. Furthermore, as well as using this general construct, we have examined the effect of the individual unit cohesion items.

Previous research supports our findings of an association between the general construct of unit cohesion and probable PTSD and common mental disorder [3,11]. A number of studies have reported that unit cohesion, measured as a general construct, has a protective effect for the development of PTSD and other mental health problems [6,10,21]. On the other hand, we did not replicate other studies that found individual unit cohesion variables, measuring unit support or morale, were associated with less PTSD and psychological distress [11,22].

We have previously shown that alcohol misuse is associated with high levels of comradeship and low levels of perceived leadership [12]. In this study, the protective effect of perceived leadership did not reach significance after full adjustment in the analyses for regular personnel. However, since our point estimate was similar between the partially adjusted and the fully adjusted model, it is likely that this was due to a lack of power. The association between comradeship and alcohol misuse among regular personnel contrast with a study of peace-keeping forces which found no association between levels of perceived morale during deployment and alcohol misuse [23]. Instead, Maguen et al. [23] showed that post-deployment alcohol misuse was best predicted by pre-deployment alcohol use and total stress symptom severity. There is little other literature regarding unit cohesion and alcohol misuse in the Armed Forces, but research on alcohol misuse in the police force supports an association with unit cohesion. Richmond et al. [24] argued that drinking alcohol aids the gathering of information and socialization among police colleagues, on a background of a hierarchical and male-dominated institution. Middleton Fillmore argues that in situations where colleagues are in a more close-knit team, occupational drinking subcultures are more likely to form [25]. The findings on alcohol misuse among reserve personnel differs from these studies as those who felt able to approach most people in the unit with a personal problem reported lower levels of alcohol misuse. It is possible that this is a reflection of alcohol misuse as a negative coping mechanism among reserve personnel and that those who feel able to talk with their peers about personal problems are able to engage in a more positive coping style.

Whether unit cohesion has a linear moderating effect or not on the association between exposures and mental health problems has been an area of contention. Work by Fontana et al. [26] on psychopathology among Vietnam veterans suggested that there was a non-linear moderating effect of unit cohesion on psychopathology. More recent work by Brailey et al. found support for a linear moderating effect of unit cohesion on life experiences in the aetiology of PTSD [3,26]. Similarly, a recent study of US air force personnel that examined the relationship between unit cohesion, stressors and PTSD found evidence of a linear moderating effect; although the authors suggest that the magnitude of warzone stress and short length of deployment may not have been sufficient to detect a curvilinear response [27]. In the present study, we found evidence of a linear relationship between unit cohesion and mental health outcomes, but there was no support for a moderating effect. We therefore did not further investigate a curvilinear moderating effect.

Unit cohesion is a potentially modifiable factor in the aetiology of various mental health outcomes. We have shown that unit cohesion as a general construct has a significant association with lower levels of probable PTSD and common mental disorder among both regular and reserve personnel. Perceived leadership in particular was shown to be an important factor as it was associated with less probable PTSD and common mental disorder, as well as an indication that leadership was associated with less alcohol misuse for regular personnel. Similar associations were found for regular and reserve personnel between individual unit cohesion items and PTSD and common mental disorder but not with alcohol misuse. In contrast to reserve personnel, where feeling able to approach most people in the unit with a personal problem was associated with lower levels of alcohol misuse, comradeship was associated with higher levels of alcohol misuse among regular personnel. It is thus important that the UK armed forces strive to foster features which contribute to the overall construct of unit cohesion, such as comradeship, leadership and morale, given that so many other factors which have a positive association with higher levels of mental health problems are un-modifiable (for example, family background and exposures on deployment).

Key points

  • A general construct of unit cohesion was found to have a direct linear association with lower levels of probable post-traumatic stress disorder and common mental disorder.

  • Perceptions of leadership appear to be a particularly important factor for regular personnel with associations with lower levels of probable post-traumatic stress disorder and common mental disorder independently of other unit cohesion items such as comradeship.

  • Feeling able to approach most people in the unit with a personal problem was associated with lower levels of alcohol misuse for reserve personnel; this contrasted with regular personnel for whom comradeship was associated with higher levels of alcohol misuse.

Funding

UK Ministry of Defence (MoD contract number R&T/1/0078).

Conflicts of interest

S.W. is Honorary Civilian Consultant Advisor in Psychiatry to the British Army and a Trustee of Combat Stress, a UK charity that provides services and support for veterans with mental health problems. All the other authors declare that they have no conflict of interests.

References

1.
Dion
KL
Group cohesion: from ‘field of forces’ to multidimensional construct
Group Dynam
 , 
2000
, vol. 
4
 (pg. 
7
-
26
)
2.
Ahronson
A
Cameron
JE
The nature and consequences of group cohesion in a military sample
Mil Psychol
 , 
2007
, vol. 
19
 (pg. 
9
-
25
)
3.
Oliver
LW
Harman
J
Hoover
E
Hayes
SM
Pandhi
NA
A quantitative integration of the military cohesion literature
Mil Psychol
 , 
1999
, vol. 
11
 (pg. 
57
-
83
)
4.
Griffith
J
Relating cohesion to stress, strain, disintegration and performance: an organising framework
Mil Psychol
 , 
2002
, vol. 
11
 (pg. 
27
-
55
)
5.
Bartone
PT
Adler
AB
Cohesion over time in a medical task force
Mil Psychol
 , 
1999
, vol. 
11
 (pg. 
85
-
107
)
6.
Brailey
K
Vasterling
JJ
Proctor
SP
Constans
JI
Friedman
MJ
PTSD symptoms, life events and unit cohesion in US soldiers: baseline findings from the neurocognition deployment health study
J Trauma Stress
 , 
2007
, vol. 
20
 (pg. 
495
-
503
)
7.
Wessely
S
Twentieth century theories on combat motivation and breakdown
J Contemp Hist
 , 
2001
, vol. 
41
 (pg. 
268
-
286
)
8.
Iversen
A
Fear
NT
Ehlers
A
, et al.  . 
Risk factors for post-traumatic stress disorder among UK Armed Forces personnel
Psychol Med
 , 
2008
, vol. 
38
 (pg. 
511
-
522
)
9.
Rona
R
Hooper
R
Jones
M
, et al.  . 
The contribution of prior psychological symptoms and combat exposure to post Iraq deployment mental health in the UK military
J Trauma Stress
 , 
2009
, vol. 
22
 (pg. 
11
-
19
)
10.
Pietrzak
RH
Johnson
DC
Goldstein
MB
, et al.  . 
Psychosocial buffers of traumatic stress, depressive symptoms and psychosocial difficulties in veterans of Operations Enduring Freedom and Iraqi Freedom: the role of resilience, unit support and post-deployment social support
J Affect Disord
 , 
2010
, vol. 
120
 (pg. 
188
-
192
)
11.
Manning
F
Fullerton
T
Health and well-being in highly cohesive units of the US army
J Appl Soc Psychol
 , 
1988
, vol. 
18
 (pg. 
503
-
519
)
12.
Browne
TE
Iversen
A
Hull
L
, et al.  . 
How do experiences in Iraq affect alcohol use amongst male UK armed forces personnel?
Occup Environ Med
 , 
2008
, vol. 
65
 (pg. 
628
-
633
)
13.
Jones
E
Fear
NT
Alcohol use and misuse within the military: a review
Int Rev Psychiatry
 , 
2011
, vol. 
23
 (pg. 
166
-
172
)
14.
Murphy
D
Sharp
D
Exploring pre-enlistment and military factors associated with the morale of members of the UK Armed Forces
Mil Med
 , 
2011
, vol. 
176
 (pg. 
13
-
18
)
15.
Hotopf
M
Hull
L
Fear
N
, et al.  . 
The health of UK military personnel who deployed to the 2003 Iraq war: a cohort study
Lancet
 , 
2006
, vol. 
367
 (pg. 
1731
-
1741
)
16.
Iversen
AC
Fear
NT
Simonoff
E
, et al.  . 
Influence of childhood adversity on health among male UK military personnel
Br J Psychiatry
 , 
2007
, vol. 
191
 (pg. 
506
-
511
)
17.
Kent
R
Data Construction and Data Analysis for Survey Research
 , 
2001
New York
Palgrave Publishers Ltd
18.
Blanchard
EB
Jones-Alexander
J
Buckley
TC
Forneris
CA
Psychometric properties of the PTSD Checklist (PCL)
Behav Res Ther
 , 
1996
, vol. 
34
 (pg. 
669
-
673
)
19.
Goldberg
D
Williams
P
A Users’ Guide to the General Health Questionnaire
 , 
1988
Windsor, UK
NFER-Nelson
20.
Babor
TF
Higgins-Biddle
JC
Saunders
JB
Monteiro
MG
AUDIT: The Alcohol Use Disorders Identification Test
 , 
2001
2nd edn
Geneva, Switzerland
World Health Organization
21.
McTeague
LM
McNally
RJ
Litz
BT
Prewar, war-zone and postwar predictors of posttraumatic stress in female Vietnam Veteran Health Care providers
Mil Psychol
 , 
2004
, vol. 
16
 (pg. 
99
-
114
)
22.
Vogt
DS
Samper
RE
King
DW
King
LA
Martin
JA
Deployment Stressors and posttraumatic stress symptomatology: comparing active duty and National Guard/Reserve personnel from Gulf War I
J Trauma Stress
 , 
2008
, vol. 
21
 (pg. 
66
-
74
)
23.
Maugen
S
Litz
BT
Wang
JL
Cook
M
The stressors and demands of peacekeeping in Kosovo: predictors of mental health response
Mil Med
 , 
2004
, vol. 
169
 (pg. 
198
-
206
)
24.
Richmond
RL
Kehoe
L
Hailstone
S
, et al.  . 
Quantitative and qualitative evaluations of brief interventions to change excessive drinking, smoking and stress in the police force
Addiction
 , 
1999
, vol. 
94
 (pg. 
1509
-
1521
)
25.
Middleton Fillmore
K
Roman
PM
Occupational drinking sub-cultures: an exploratory epidemiological study
Alcohol Problem Prevention Intervention in the Workplace. Employee Assistance Programs and Strategic Alternatives
 , 
1990
Westport, CT
Quorum Books
(pg. 
77
-
94
)
26.
Fontana
A
Rosenheck
R
Horvath
T
Social support and psychopathology in the war zone
J Nerv Ment Dis
 , 
1997
, vol. 
185
 (pg. 
675
-
680
)
27.
Dickstein
B
Mclean
C
Mitz
J
, et al.  . 
Unit cohesion and PTSD symptom severity in Air Force medical personnel
Mil Med
 , 
2010
, vol. 
175
 (pg. 
482
-
486
)

Author notes

*
Joint first author.