Introduction

Combat operations during the Iraq/Afghanistan conflicts involved participation in, and witnessing of, life-threatening incidents [1]. For some military personnel, this included contact with improvised explosive devices and exposure to ambushes [2]. Combat deployment increases exposure to potentially traumatic events. This can lead to increased risk for combat-related health consequences among (ex-)military personnel.Footnote 1 Health consequences can include post-traumatic stress disorder (PTSD) [3], traumatic brain injury (TBI) [4] or major limb trauma [5]. One potential strategy to reduce health and wellbeing complications following combat exposure may be to increase the availability of social support [6].

Social support is defined within the psychological and epidemiological literature as the availability and adequacy of social connections, comprising structural (received) and functional (perceived) support [7]. Structural social support refers to the presence of relationships, which includes objective measures such as the size of social network, whilst functional social support incorporates quality of support, for example, how successfully a relationship fulfils one’s needs [7, 8]. Sources of support can include family, friends, partners and colleagues. Social support can be further divided into emotional, tangible and informational support [9]. These consist of showing concern for one’s feelings, sharing material or financial resources and providing advice or information, respectively. Perceived social support (PSS) encompasses an individual’s beliefs about the scope and suitability of support they experience [10]. Evidence suggests that PSS is more predictive of wellbeing outcomes than structural measures [7, 11, 12]. Therefore, the focus of this review will be on PSS.

A growing body of evidence posits that PSS is one of the key determinants of health and wellbeing across a variety of populations and contexts [7, 13], whilst low PSS has been associated with increased morbidity and mortality [14, 15]. Military samples may have unique experiences of PSS compared to the general population as comradeship and collectivism are central to military culture and are essential for operational success [16]. In a sample of United States (US) (ex-)military personnel who experienced major limb trauma during Iraq/Afghanistan deployment, almost one third (27%) reported low levels of PSS [17].

There is a need to focus research efforts on (ex-)military personnel who served in Iraq/Afghanistan, as survival rates were higher than in previous conflicts due to technological and medical advances [18]. This population may therefore be prone to experiencing psychosocial, occupational and health-related difficulties during deployment and thereafter [19], which may be alleviated by PSS [20]. Studies have investigated PSS during deployment [21], however less is known about PSS following deployment. The post-deployment period may give rise to additional challenges. For example, some personnel were required to leave their military role having sustained a combat-related mental or physical injury [4, 19]. Studies have shown that experiencing medical discharge or involuntary termination is associated with reduced social group engagement [22, 23]. Furthermore, in a sample of (ex-)military personnel who served in Iraq/Afghanistan, loss of brotherhood was reported as the most prevalent challenge to cope with upon returning home [24]. Accordingly, there is a need to understand the social experiences of this population. A thorough investigation may be useful for policy makers and service providers in devising targeted messages and informing psychosocial intervention, to meet the needs of combat-exposed (ex-)military personnel.

To date, no review has investigated PSS in combat-exposed (ex-)military personnel. Therefore, this systematic review and meta-analysis was directed by the following research questions:

  1. 1.

    What levels of PSS are reported in combat-exposed (ex-)military personnel who served in Iraq/Afghanistan?

  2. 2.

    What factors are associated with PSS in combat-exposed (ex-)military personnel who served in Iraq/Afghanistan?

Methods

This review was prospectively registered with PROSPERO (CRD42023389759) and was conducted according to PRISMA guidelines [25].

Search strategy

The following five bibliographic databases were searched in January 2023 and re-searched in August 2023: Embase, Medline, PsycINFO, Social Policy and Practice and Applied Social Sciences Index and Abstracts (ASSIA). The search contained key words relating to: “military personnel” and “combat” and “perceived social support” (Appendix 1). Search terms were developed with reference to published literature and following consultation with a librarian. Medical Subject Headings (MeSH) terms were included where available. Searches were restricted from 2001 onwards (beginning of the Iraq/Afghanistan conflicts). Reference lists from eligible studies were searched manually and forward citation tracking was performed.

Eligibility criteria

Inclusion criteria

  1. (a)

    Measured PSS in (ex-)military personnel who deployed on a combat operation in Iraq/Afghanistan.

  2. (b)

    Peer-reviewed publications, quantitative study design.

  3. (c)

    Used a validated measure of PSS ≥ 6 months post-deployment to cover the period after homecoming and beyond [26].

  4. (d)

    Published in the English language with full-text availability.

Exclusion criteria

  1. (a)

    Non-peer reviewed literature, meta-analyses, literature reviews, experimental studies and qualitative studies.

  2. (b)

    Measured PSS < 6 months post-deployment.

  3. (c)

    Investigated PSS from the perspective of families, carers or healthcare professionals.

  4. (d)

    Measured structural social support, for example size of social network, marital status or living arrangements.

  5. (e)

    Investigated formal or paid support, for example counselling therapy.

Screening and data extraction

L.E.G. performed the search, removed duplicates using Zotero reference manager and completed title and abstract screening followed by full-text screening. C.W. independently reviewed 10% of the records during title and abstract (n = 120) and full-text (n = 36) screening stages. A standardised data extraction form was completed independently by L.E.G., and 10% by C.W. (n = 4). Any disagreements were discussed and resolved. The following data were extracted: study information, PSS measure, mean level of PSS and estimates for all investigated factors including size and direction of effect. Effect estimates primarily included odds ratios and regression coefficients. Estimates were extracted for fully adjusted models where available.

Quality assessment

Methodological quality was assessed using the National Heart, Lung and Blood Institute (NHLBI) quality assessment tool for observational cohort and cross-sectional studies [27]. L.E.G. and C.W. independently performed the quality assessment and discussed any discrepancies. For each included study, 14 questions relating to study quality were answered and articles were defined as “good”, “fair” or “poor” based on the NHLBI quality assessment criteria.

Data analysis and synthesis

To assess the level of PSS, a meta-analysis was conducted using the statistical software package Stata version 17. A pooled mean PSS score was generated for 13 (out of 15) studies using the Deployment Risk and Resilience Inventory-Post-deployment Social Support-1 (DRRI-PDSS-1) measure [28]. Two studies which used the DRRI-PDSS-1 were excluded from the meta-analysis as the mean level of PSS was not reported [29] and another had substantial missing data [30]. Analysis was not possible with the additional PSS measures due to limited data.

Two subgroup analyses were performed. The first assessed the level of PSS among (1) non-clinical and (2) clinical samples. The second further divided the clinical samples, to assess the level of PSS across the following groups: (1) non-clinical, (2) probable PTSD, (3) mild TBI, (4) amputation injury and (5) any other treatment-seeking samples. This includes individuals referred to a health clinic, but the paper does not specify for which health condition. For example, personnel who accessed treatment through Veterans Affairs (VA). If studies consisted of multiple groups, for example a probable PTSD group and a treatment-seeking group [31], these were categorised into separate subgroups. One study recruited a non-clinical sample, however most of the sample (70.4%) reported probable PTSD therefore this was included within the probable PTSD subgroup [32]. In studies where means and standard deviations (SD) were not reported for the total sample, means for reported subgroups were combined prior to analysis using combined mean formula. A random-effects model was generated due to the high heterogeneity of the included studies. This suggests studies were different in nature, in this case a wide range of samples were included with different clinical presentations. The random-effects approach is deemed a sufficient method for dealing with heterogeneity as it assumes that underlying effects follow a normal distribution. Therefore, the pooled estimate would be the mean or average effect. The effect sizes are assumed to represent a random sample of all possible effect sizes, as compared to fixed-effects, which assumes one true effect size underlies all the studies [33]. Results were displayed using forest plots. Funnel plots were generated to visually assess for publication bias. It is also important to quantify publication bias and this was done using Egger’s test [34].

To examine factors associated with PSS, a narrative synthesis was performed. This approach was chosen due methodological and clinical heterogeneity; for example, variation in sample sizes, target populations and clinical outcomes [35].

Results

Study selection

Figure 1 illustrates our search strategy results. Inter-rater agreement was strong at 96% at both the title and abstract stage, and full-text screening stage.

Fig. 1
figure 1

PRISMA flow chart

Full size image

Study characteristics

In total, 35 papers published between 2011 and 2023 were included (Table 1). Thirty-one studies were conducted in the US, and the remaining studies were undertaken in the Netherlands [36, 37], Denmark [38] and Norway [39]. Twenty-three studies employed a cross-sectional methodology with the remaining studies being prospective (11 studies) and retrospective cohort studies (1 study). Sample sizes ranged from 45 [40] to 3465 [39] and the mean sample size was 572.5 (SD 706.6). Most samples were predominantly male (31 studies). Studies were made up of non-clinical samples (n = 12) [36,37,38,39, 41,42,43,44,45,46,47,48,49], probable PTSD (n = 8) [29, 32, 50,51,52,53,54,55], mild TBI (n = 1) [56], physical/amputation injury (n = 3) [17, 40, 57] and all other treatment-seeking samples where health condition is not specified (n = 12) [24, 30, 31, 53,54,55, 58,59,60,61,62,63]. One study investigated specific domains of social support (structural, functional), whilst three investigated type of social support (emotional, tangible).

Table 1 Study characteristics and quality assessment for included studies
Full size table

Quality appraisal

Of the 35 included studies, 24 were rated as having “fair” methodological quality and 11 were rated as “good” (Table 1). Lower rated studies tended to have small sample sizes or did not adequately adjust for confounders. Overall, study quality was satisfactory, however the strength of evidence is reduced in cross-sectional studies as the direction of causality cannot be determined.

Measuring PSS

Seven different instruments were used to examine PSS (see Appendix 2), all of which were validated. The most commonly used measure was the DRRI-PDSS-1 scale [28] (n = 15). Additional measures include the DRRI-PDSS-2 [65] (n = 7), Multidimensional Scale of Perceived Social Support (MSPSS) [66] (n = 4), Interpersonal Support Evaluation List (ISEL) [8] (n = 3), Medical Outcome Study Social Support Survey (MOS-SSS) [67] (n = 4), Oslo Social Support Scale-3 (OSSS-3) [68] (n = 1) and the Provisions of Social Relations Scale [69] (n = 1).

Level of PSS

Most studies reported mean level of PSS along with SD whilst four did not. The pooled mean PSS score was 54.40 (95% CI 51.78 to 57.01). Subgroup analyses revealed a higher pooled mean PSS score in non-clinical samples (57.26) compared to clinical samples (52.29) (Appendix Fig. 3). Further subgroup analyses indicated samples with probable PTSD reported the lowest mean level of PSS (44.40), followed by mTBI (52.92), treatment-seeking (54.44) and non-clinical (57.99), whilst an amputation injured sample reported the highest level of PSS (59.10) (Fig. 2). Substantial heterogeneity was observed among included studies (I2 = 99.15%, p < 0.001) meaning a random-effects model was selected. The funnel plots showed no evidence of publication bias (Appendix Fig. 4) as did the findings from Egger’s test (p = 0.68).

Fig. 2
figure 2

Forest plot showing pooled mean PSS score by sample type. PSS measured using DRRI-PDSS-1 where scores can range from 15 to 75. PTSD post-traumatic stress disorder, TBI traumatic brain injury

Full size image

Associated factors

Table 2 summarizes the breadth and strength of evidence for all investigated factors. Findings are presented in Table 3 in more detail. As indicated in Table 3, we identified which studies controlled for confounders. Studies where adjusted effect sizes were reported illustrate that potential confounders have been considered and accounted for. In total, 13 studies adjusted for confounders whilst 22 did not. The confounders varied depending on the outcome of interest. Potential confounders were identified from existing literature and generally included demographic characteristics (age and sex), psychosocial factors (family functioning and unit support), deployment stress and the number of deployments. For the narrative synthesis, factors have been categorised into the following six domains: mental health, social/behavioural, physical health, military, psychological and demographic factors. Next to each header, n denotes the number of papers that examined these factors.

Table 2 Summary table illustrating breadth and strength of evidence for factors associated/not associated with PSS
Full size table
Table 3 Measure and level of PSS, associated or non-associated factors
Full size table

Mental health (n = 19)

Mental health-related factors were most frequently investigated in relation to PSS. Evidence is strongest for the relationship between PSS and probable PTSD, in terms of breadth and strength. Lower PSS scores were associated with increased self-reported [29, 36, 38, 41, 42, 45, 49, 55] and clinician-assessed PTSD [31, 54, 55, 59], cross-sectionally and at 1, 2, and 10 years post-deployment. Effect sizes ranged from β = −0.81 to −0.25. In one study PSS was not associated with clinician-assessed PTSD [55]. For PSS subscales, lower levels of appraisal, belonging and tangible support were associated with higher PTSD scores [53].

There is strong evidence that lower PSS scores were associated with increased self-reported [24, 37, 42, 49, 51] and clinician-assessed depression [59]. Effect sizes ranged from β = −0.48 to −0.38. However, presence (compared to absence) of comorbid major depressive disorder (MDD) and PTSD was not associated with PSS [51].

Lower PSS scores were cross-sectionally and longitudinally associated with presence of clinician-assessed generalised anxiety disorder [37, 59] and agoraphobia [37]. Additionally, suicidal ideation [63] and behaviour [24], help-seeking intentions for psychological problems [49], probable pathological gambling [32], drug abuse [41] and hazardous drinking [41] were all associated with lower PSS. Hazardous drinking caseness (compared to no hazardous drinking) was associated with distinct types of PSS, specifically, lower tangible and emotional support in both men and women [62]. Conversely, one study found hazardous drinking (defined as AUDIT score ≥ 8) was not associated with PSS [59]. Two studies examined overall mental health status and found higher PSS scores were positively associated with mental health status cross-sectionally [61] and longitudinally [41], where higher scores indicate better mental health.

Social/behavioural (n = 10)

Lower PSS scores were associated with increased relationship concerns during deployment [44], being unmarried [61], self-reported hostile behaviour at 1, 2 and 10 years post-deployment [37] and having a maladaptive childhood family environment [29]. Higher PSS scores were associated with social gambling (compared to non-gambling) [32], post-deployment divorce [52], and academic adjustment in student veterans, defined as an ability to handle the demands of college [46]. One study assessed the relationship between two types of PSS and found a positive association between emotional and tangible support in both males and females [47]. Another study found that PSS was not associated with received social support [48].

Physical health (n = 4)

Only a few studies investigated the association between lower PSS scores and physical health-related outcomes. Having a service-related injury (compared to not) [40], worsened physical functioning [30] and self-reported sleep disturbances [43] were associated with lower PSS scores, whilst erectile dysfunction was not [50]. In addition, higher PSS scores were associated with increased self-reported pain and role limitation [30], as well as disability and dissatisfaction [58].

Military (n = 5)

Military personnel within enlisted ranks (compared to officers) reported lower PSS [61]. Higher combat exposure was associated with lower PSS scores [44]. For subscales, combat exposure was negatively associated with emotional PSS in both men and women and positively associated with tangible PSS in women [47]. Higher PSS scores were associated with increased unit support [30, 44], exposure to wounded people during deployment [61] and combat disclosure, defined as willingness to disclose deployment events [45].

Psychological (n = 6)

Lower PSS scores were associated with increased sexual desire problems [50] and avoidant coping styles [29], whilst higher PSS scores were associated with increased coping self-efficacy [48], vitality [30], satisfaction with life [56] and post-traumatic growth (PTG) [57]. Furthermore, one study found PTG was positively associated with both structural and functional PSS [39].

Demographics (n = 2)

Only two studies investigated demographic factors in relation to PSS. One found (ex-)military personnel within Black ethnic groups reported lower PSS, compared to White personnel [61]. Another study found no association between emotional or tangible PSS and gender [47].

Discussion

This study systematically synthesised the evidence base on the levels of PSS, and associated factors, in combat-exposed (ex-)military personnel who were deployed to the Iraq/Afghanistan conflicts. A pooled mean PSS score of 54.40 (95% CI: 51.78 to 57.01) was generated, with probable PTSD samples reporting the lowest levels and an amputation injured sample reporting the highest. A range of mental health, social/behavioural, physical health, military, psychological and demographic factors were associated with different levels of PSS. The associations most frequently examined were low levels of PSS and probable PTSD, depression or anxiety, as well as high levels of PSS and increased PTG or unit support.

Measuring PSS

Seven different scales were used to examine PSS. Broadly, the scales measure perceptions of the extent to which family, friends, co-workers and the community provide support following deployment. For some scales, types of PSS were examined including emotional, tangible and informational [8, 28, 65, 67, 68]. The DRRI-PDSS-1 was most frequently used and was developed with a sample of Gulf War veterans (1990–1991) [28]. The DRRI-PDSS-2 [65] was created to align with experiences from Iraq/Afghanistan, such as increased risk for insurgency warfare and less emphasis on nuclear exposure [65].

Self-reported level of PSS

The meta-analysis revealed that the mean DRRI-PDSS-1 score ranged from 41.69 to 60.35, with a pooled mean score of 54.40 (95% CI 51.78 to 57.01). Scores were similar across non-clinical samples but less so for clinical samples. Overall, confidence intervals were small indicating a high degree of certainty. Similar levels of PSS have been reported in US Gulf War veterans (mean 56.69, SD 10.52) [28], where demographic characteristics were similar to those of the studies included in this review. The participants were predominantly white, male and aged between 31 and 40 years. As the DRRI-PDSS is heavily centred around military experience, no general population studies have used this measure therefore comparisons were not possible.

Subgroup analyses revealed a greater level of PSS in non-clinical samples than in clinical samples. Further dividing the clinical samples into subgroups revealed that those with probable PTSD reported the lowest level of PSS (pooled mean 44.50, 95% CI 39.10 to 49.70), whilst an amputation injured subgroup reported the highest level of PSS (59.10, 95% CI 58.10 to 60.10). Several explanations exist for this finding. Amputation injuries are visible and involve intensive rehabilitation, and whilst the PSS questionnaires examine support with daily tasks, participants with a physical injury may access more tangible support due to physical limitations. The rehabilitation process may provide injured personnel with opportunities to build a social network, for example by engaging in collaborative sporting events such as the Invictus games foundation, whilst uninjured personnel may be unaware of these possibilities and have less targeted interventions. This finding may also be explained by PTG, which is positively correlated with PSS and is greater among injured personnel than uninjured personnel [66]. Conversely, PTSD is an invisible injury and may go undiagnosed or undetected. Individuals with PTSD might feel undeserving of support due to stigma, particularly serving personnel who may fear negative consequences for their career [70]. Subgroup analyses for gender, nation and serving status were not possible due to limited data. Samples were predominantly US male personnel and included a combination of serving and ex-serving personnel.

Associated factors

Probable PTSD and depression were the most robustly assessed factors associated with PSS and were examined cross-sectionally and longitudinally. Comorbid PTSD and MDD was not associated with PSS in one study [46], however the p-value was close to significance at 0.06. One study also found no association between clinician-rated PTSD and PSS although the small sample size may explain why an effect was not detected. The association between PTSD and PSS appears to be bidirectional. One study used a cross-lagged panel approach [54], meaning that reporting lower PSS predicts the later occurrence of PTSD and that having PTSD predicts a reduction in PSS as perceptions are framed by symptomatology. This relationship may occur directly or indirectly. Individuals with PTSD may drive their support networks away through negative thoughts and behaviours, such as avoidance and mistrust [71]. Alternatively, those with low levels of PSS may later experience PTSD due to a lack of support networks with which to discuss traumatic experiences. The disclosure of combat-related experiences is associated with reduced PTSD [45]. There is also moderate evidence for the association between anxiety and lower levels of PSS, which may be explained by similar mechanisms such as social withdrawal.

Turning attention to physical health-related factors, higher PSS scores were associated with increased pain and role limitations [30], as well as disability and dissatisfaction [58]. Individuals who report high levels of pain, role limitations and disability may, in turn, require more tangible support and assistance with daily tasks. It appears that individuals affected by these physical difficulties can still access support despite these challenges. This highlights the importance of measuring subdomains of PSS, for example, treating tangible and emotional support as distinct entities. These findings contradict those of a non-military sample with rheumatoid arthritis in which lower levels of PSS were linked to greater pain intensity [72]. However, the study included within this review had substantial missing data. This could also be condition specific, as rheumatoid arthritis is chronic and debilitating therefore certain symptom profiles may impact PSS differently. This review also revealed that lower PSS scores were associated with more sleep disturbances [43]. This may be because PSS protects against social isolation and increases feelings of connectedness leading to positive health behaviours. Sleep disturbances are also concomitant with other physical and mental health problems [73].

Alongside mental and physical health-related factors, cross-sectional evidence suggests that PSS is associated with a range of social/behavioural factors. Lower PSS was associated with being unmarried [61], whilst higher PSS was associated with greater post-deployment divorce [52]. Those who have recently experienced a divorce may report greater levels of PSS as support networks are more prevalent during times of stress [74]. Individuals might lean on other sources of support such as friends and family in their partners absence.

This review also found evidence to suggest PSS is associated with some psychological, military and demographic factors. Within these topical domains, unit support and PTG have been the most investigated and are positively associated with PSS [30, 57]. This finding aligns with existing research in the general population, as positive associations have been found between PTG and social support [75].

The findings were mostly in line with other populations. In general, and in clinical populations, PTSD [76], depression [13, 77], anxiety [13], suicidal ideation [78] and poor sleep quality [77, 79] are all negatively associated with PSS. Conversely, gender differences in PSS are established within the general population [80], but only one study in this review examined this topic and found no association [81]. Despite this, the p-value was close to 0.05, albeit not significant, which may be indicative of a type II error defined as failure to reject the null hypothesis.

Overall, higher levels of PSS were related to better health and wellbeing outcomes after deployment. However, PSS may not always be beneficial as one study found that high levels of PSS were associated with reduced professional help-seeking for psychological problems. Individuals may feel sufficiently supported by their social network and refrain from seeking professional support when it may benefit them. One study showed that 86% of a sample of United Kingdom (UK) (ex-)military personnel sought informal sources of support for mental health-related problems, whilst only 55% accessed medical support [82].

Limitations of the evidence base

Most studies were rated as having fair methodological quality; whilst this was deemed sufficient for inclusion within this review, most had a cross-sectional design and longitudinal research was lacking. Longitudinal research would provide an opportunity to examine variables across multiple time points and to determine the direction of the relationship between variables; for example, PSS causing a change in mental health. Therefore, the causal direction of the relationship between PSS and health-related factors cannot be determined in most cases. Second, less than half of the included studies adjusted for confounders showing a potential source of bias. When running unadjusted analyses, it is not clear to what extent other variables may play a role in the relationship between the exposure and the outcome. This is especially important in social epidemiology, where there are numerous variables that are potentially intertwined, and significant collinearity between these variables. Third, few studies have separated PSS into distinct domains, such as source (family and friends) or type of support (emotional and tangible). This is important for gaining an understanding of the differential contributions of subdomains of PSS, as evidence suggests that they may have differing consequences [83]. Fourth, few studies in this review used the updated DRRI-PDSS-2 measure, which was specifically designed for those who deployed to Iraq/Afghanistan; therefore, the meta-analysis was centred around the DRRI-PDSS-1. This should be kept in mind when comparing the findings across studies using different measures and over different time periods. Despite this, only 20% of the questionnaire items from the DRRI-PDSS V1 were revised to create V2, therefore much of it remains the same as before. Fifth, no studies have investigated associations between PSS and injury severity or PTSD symptom profiles. Finally, the data are predominantly derived from US samples (n = 31). The experience of PSS may differ between nations due to differences in social, political and military experiences. Therefore, our reported findings may reflect the specificities of the US military and may not be generalisable to different cultures or contexts.

Strengths and limitations of the review

The review included studies in which a validated measure of PSS was used. A comprehensive list of search terms were used, and we included concepts related to PSS such as comradery. This review focused on functional measures of PSS, that indicated the quality of PSS, as opposed to structural measures, such as number of social contacts. Functional support is recommended within the social support literature as a more useful source for examining support and it is a better predictor of wellbeing outcomes than structural measures [11].

This review is also subject to various limitations. The review did not examine structural dimensions of social support. Although functional PSS measures are beneficial for understanding the quality of support, they may be influenced by individual differences in perception, memory and judgement [12]. What one person considers to be supportive behaviour may differ from another. Stressful events, such as combat deployment, may challenge one’s belief system and decrease their perception of support, despite it being present. This has been illustrated by the deterioration-deterrence model [74]. This model proposes that the impact of traumatic events on mental health occurs both through social network disruption and a decrease in one’s perception of support. The buffering impact may also be impaired by the person’s ability to process support they receive and relate meaningfully to another individual.

In addition, the review did not include grey literature therefore increasing the risk of publication bias. However, to reduce this concern, we generated funnel plots and conducted Egger’s test which showed no evidence of publication bias. We also included non-statistically significant findings from the included studies where p > 0.05. Finally, the review focused only on quantitative, observational data. Qualitative studies may provide more subjective insights into how individuals perceive their social support. As described previously, functional support is contingent upon perception and subjectivity. Therefore, determining how (ex-)military personnel experience their support is important for further enquiry.

Implications

Understanding whether these associations are reflected in UK samples will be an important addition to the field. Future research should investigate PSS in (ex-)military personnel from other countries, as the evidence is heavily weighted toward US samples. The experience of PSS may differ in UK samples due to different deployment experiences and sociocultural trends, such as greater alcohol consumption among UK personnel [84]. Studies should also distinguish between sources of support and examine associations with physical health-related outcomes, such as injury severity, as well as military factors.

Based on the review findings, there is sufficient evidence to make recommendations regarding mental health-related factors for the US population. The factors within other categories were based on single and mostly cross-sectional studies. Yet, the strength and breadth of evidence is robust for the relationship between post-deployment PSS and both PTSD and depression. Policies should highlight the importance of PSS, not just immediately after transition from military to civilian life, but earlier on in service to reduce the negative impact after leaving, and for 10 +  years into the future. In addition, policies should be targeted towards enhancing people’s feelings of being supported during challenging times. Psychosocial interventions should include avenues to access social support and enhance understanding of how to foster high quality relationships. Furthermore, developing more peer-to-peer support programmes and highlighting opportunities for societal integration would be beneficial to the individual and society. Increasing levels of PSS may reduce the strain on military mental health services. The UK Government’s Office for Veterans’ Affairs (OVA) strategy for veteran’s has highlighted the importance of communities and relationships for reducing social isolation. The findings from this review hope to pave the way for more UK research in this area to establish policy and practice recommendations.

Conclusions

This review aimed to synthesise research on the level of PSS, and associated factors, in combat-exposed (ex-)military personnel who deployed to Iraq/Afghanistan. Level of PSS was lowest in samples with probable PTSD and highest in an amputation injured sample. PSS was negatively associated with mental health-related factors, including PTSD, depression and anxiety, and positively associated with psychosocial factors, including PTG and unit support. Findings from this review highlight the importance of maintaining adequate levels of PSS following deployment. The findings suggest that (ex-)military personnel who deployed on a combat operation to Iraq/Afghanistan may benefit from interventions to improve PSS, to drive positive health outcomes and protect against negative outcomes.