Stigmatising beliefs about seeking help for mental health problems are commonplace among military personnel [1]. Holding such beliefs may influence the decision to seek help particularly when social position, social connectedness or occupational functioning may be adversely affected [2]. The mechanism by which stigmatisation operates is unclear; it may be a direct impediment to help-seeking, however, research suggests that it may have variable effects or be wholly unrelated to help-seeking behaviours [3]. Some research suggests that military personnel may fail to recognise that sub-threshold mental health symptoms are often linked to occupational impairment and may need to be addressed [4]. In this case, stigma is not the prime driver of non-help seeking. Other research suggests that self-management may be the preferred mental health management option for military personnel [5] and that this should be considered as a component of help-seeking decision making alongside stigma/BTC. Recent research suggests that military personnel often report stigma/BTC in relation to mental health counselling but can effectively override these to access care [6]. Given the substantial rates of military stigmatisation [7], many country’s military, including the UK, have sought to reduce stigma/ BTC through psychoeducation, based around mixed media campaigns and peer support programmes [8]. The effectiveness of such an approach is mixed with some suggesting that measurable and durable positive effects upon stigma-related knowledge may occur [9], while others suggest that rates of stigmatisation remain largely unaffected [10,11]. The available evidence therefore seems to suggest that there is a complex relationship between stigma/BTC, recognition of symptoms and current mental health status. Given the lack of any clear evidence of a direct relationship between help-seeking and stigma/BTC and variable research findings about how stigma/BTC might operate when deciding whether to seek help, we sought to examine the association of stigmatising beliefs, perceived barriers to care, mental health symptoms and mental health help-seeking in UK military personnel and to define the nature of any such relationship. Our research question was, is there a relationship between perceived stigmatisation, barriers to care and help-seeking and if so, what role does current mental health status have in modulating the decision to seek help? We chose to sample military personnel in a garrison setting (some of whom had previously deployed) as this area is under-researched and the bulk of our previous studies have been conducted with deployed personnel as described by Osorio et al. [7].

Study participants were serving members of the British Army (BA) who provided baseline information prior to voluntarily participating in the evaluation of a stigma reduction intervention. The study took place at an army base and the soldiers gave consent to take part in the stigma reduction study after having been given a full briefing by the study co-ordinator. They were randomly selected a priori to receive either a comedy show with an active stigma reduction component or a standard comedy show. They were informed that their responses would be used for further research purposes and they were offered the opportunity to withdraw prior to providing baseline information. At no time were they asked to volunteer for the study. Participants completed a questionnaire, which enquired about demographic characteristics, mental health status, stigmatising beliefs about help-seeking for mental health problems and potential barriers to care (hereafter termed ‘stigma/BTC’), potential discrimination towards other people with mental health problems, the role of the affected person in managing mental health problems, current levels of help-seeking and potential and actual sources of support accessed. The study was approved by the Ministry of Defence Research Ethics Committee.

Mental health indices included current alcohol use, which was assessed using the 3 item Alcohol Use Disorders Identification Test (AUDIT-C), a brief validated screening tool used to identify potential alcohol use disorders [12]. Given the high prevalence of alcohol use disorders in the UK AF [13], a higher AUDIT-C cut off score of ≥ 8 rather than ≥ 4 was used to indicate the presence of potentially harmful alcohol use. Symptoms of common mental disorder (CMD) were assessed using the 12 item General Health Questionnaire (GHQ-12) [14,15]; endorsing four or more symptoms indicated CMD caseness. Symptoms of posttraumatic stress disorder (PTSD) were assessed using the four item Primary Care PTSD scale (PC-PTSD) [16]; endorsing three or more symptoms on this measure indicated the presence of symptoms of probable PTSD. A variable was generated combining probable PTSD or CMD caseness to assess the association of stigma/ BTC and other variables with mental ill-health.

As Stigma/BTC constituted the dependent variable in this study these were directly assessed using an 11 item stigma/BTC scale adapted from a measure used in US military research [17] originally termed ‘perceived barriers to seeking mental health services’ which has subsequently been used with UK military samples. This scale has been reported as having high internal consistency [18]. The adaptation included the deletion of four items, which did not apply to UK samples, including cost of care, arranging transport, scheduling appointments and getting time off work. In the UK military, logistics and work absence are provided for and care is free. Two further items were not included as they were rarely endorsed in our previous surveys and instead we included four items which included the availability of services, thinking less of colleagues, lack of military support for mental health problems and giving responsibility to those with mental health problems, all of which have been used to assess military samples [19,20]. Respondents were asked to rate their strength of agreement with each stigma or BTC item using a four-point Likert scale which yielded scores ranging from 11-44. Item 2, ‘the military supports Service members who have mental health problems’ was reverse scored and transformed so that the scoring format corresponded to other scale items; higher scores indicated greater levels of stigma/BTC. The stigma/BTC scale was then summed and tertiles were generated; lower and middle tertiles were combined to compare the highest levels of stigma/BTC with lower levels. In addition, strongly agree and agree responses were combined for each stigma scale item as were strongly disagree and disagree responses. This produced a binary variable for each stigma/BTC item. A count variable was generated, giving a scale ranging from 0-11. Agreeing with three or more items was used to give an estimate of high levels of stigma/ BTC [21].

As discrimination is often linked to stigma/BTC, we assessed reported intended behaviour toward people with mental health problems using items 5 to 8 of the Reported and Intended Behaviour Scale (RIBS) [22]. This was endorsed using a Likert scale, indicating strength of agreement with four statements: I would live with, work with, live nearby and continue a relationship with someone with a mental health problem. ‘Don’t know’ responses were recoded to neutral. The scale was summed to give a range of scores from 4 to 20 with lower scores reflecting greater levels of potential discrimination. Tertiles were then generated and the middle and upper tertiles were combined so that greater levels of potential discrimination (the lower tertile) could be compared with lower levels (the middle and upper tertiles).

As identified in the introduction, recent research suggests that individuals experiencing mental health problems often report that they wish to manage their symptoms themselves. Participants were therefore asked to consider five statements regarding their view of the individual’s role in managing mental health. These were; ‘mental health support can be helpful for those who need it’; ‘it takes courage or strength to get treatment for a psychological problem’; ‘I would prefer to manage my problems on my own’; ‘strong people can resolve psychological problems by themselves’ and ‘psychological problems tend to work themselves out without help’. Respondents were asked to rate their strength of agreement with each item using a four-point Likert scale which yielded scores ranging from 5–20; lower scores indicated greater agreement with the five statements and therefore a more substantial personal role in managing mental health problems. The scale was first summed, tertiles were generated and the lower and middle tertiles were combined so that those rating the personal role in managing mental health less markedly (the upper tertile), could be compared with those who felt that it was more important (the middle and lower tertiles).

Personnel were asked if they were currently accessing support for a stressful, emotional or family problem and if they were currently interested in receiving support. They were asked to consider nine potential help sources, which represented military medical assistance (mental health professional, regimental medical officer or doctor and other medical staff) and non-medical help (friend or family, unit chain of command, colleague, Trauma Risk Management (TRiM) practitioner, chaplain and Big White Wall therapist support). TRiM practitioners are military peers serving in the same unit trained to provide support following exposure to traumatic events [23]; the Big White Wall is an online early intervention service for people experiencing psychological distress which has been further developed for use by those with a history of military service [24]. Participants were further asked whether they would be willing to use these help sources in the future.

Analysis

The study was powered to detect a change of plus or minus 5% in reporting one or more stigma scale item among 212 personnel with 95% confidence. Osorio et al. [21] reported that stigma at this level has a mean frequency of approximately 60% among deployed military personnel. All analyses were conducted in the Statistical Package for Social Sciences version 20. Categorical data were analysed using Pearson’s chi square test and potential trends in the data were assessed with chi square test for trend. Multivariable logistic regression was used to generate odds ratios with 95% confidence intervals which were adjusted for observed confounding variables selected a priori based upon their reported association with mental health in UK AF personnel [25,26]. These included rank, categorised into three groups (Junior Rank, Senior Rank and Officer), age grouped into five categories, service length grouped into five categories, sex, being in a relationship or not and being deployed in the last year or not. Numbers and percentages of those who returned a valid answer are reported in this paper, whereas missing data are not; percentages and numbers may not sum to sample and sub-sample totals. Statistical significance was p ≤ 0.05.

594 survey questionnaires were distributed, 110 were not returned or returned blank giving an overall response rate of 81.5% (n=484).

Sample description

51.6% (n=248) of respondents were aged 16-23 years and 48.4% (n=233) were aged 24 years and over. Official statistics give the median age of the Army as 29 years [27]; for the study sample, this was 25 years. 51.4% (n=237) had short service lengths of 0-4 years and 48.6% (n=224) had served for ≥ 5 years. The study sample consisted primarily of regular forces rather than reserves (98.3%) and most were male (95.1%); official statistics [27] suggest that regulars should comprise 83.6% of a representative sample and 90.3% should be males. 79.8% (n=383) were junior ranks (Private soldier and Lance Corporal equivalent) whereas 53.3% would normally be expected. 28.5% (n=138) had completed ≥ two operational tours; 50.6% (n=245) had not deployed. There were no statistically significant differences in stigma/BTC levels within the various socio-demographic categories (Table 1).

Mental health outcomes, help-seeking and sources of help

25.2% (n=112) reported symptoms of CMD, 12.4% (n=55) screened positive for probable PTSD and 35.0% (n=155) screened positive for potentially harmful alcohol use. 42.9% (n=48) of the CMD cases had not sought help; of those that had and who reported a help source, nonmedical sources of support were more commonly accessed than military medical. Of the probable PTSD cases, 32.7% (n=18) had not sought help; among PTSD help-seekers, non-medical sources were more commonly accessed than military medical. 70.3% (n=109) of probable harmful alcohol users had not sought help; of the alcohol-related help-seekers, non-medical support was more frequently accessed than military medical support (Table 2). Of those who screened positive for either CMD or PTSD (28.4%, n=137), 57.6% had sought help (n=76 of 132 who reported help seeking status).46 of 460 (n=10.0%) personnel had not sought help but were interested in receiving support for a significant stressful, emotional or family problem. Of these 43.5% (n=20) screened positive for CMD or PTSD.

Overall, 36.9% (n=164) of the sample (n=445) were help-seekers; of these, 20.0% (n=89) sought help from non-medical sources; 8.8% (n=39) from military medical sources, alone or in combination with non-medical sources; 8.1% (n=36) sought help from an unspecified source. Respondents were most willing to engage with friends and family followed by a mental health professional; personnel were least willing to engage with an online therapist and the unit chain of command. Potential sources were ranked in a similar pattern to actual sources of help with the exception of the unit chain of command, which constituted the second commonest source of actual support. 49.3% (n=219 of 444) reported having had contact with other people who were suffering from a mental health condition (Table 3).

Table 1: Socio-Demographic and Operational Characteristics.

Table 2: Mental Health Outcomes and Help-Seeking.

Stigmatisation, Barriers to Care, Potential Discrimination and Personal Role in Mental Health Management.

Approximately three quarters of respondents endorsed three or more of the eleven items on the stigma/BTC scale. The most frequently reported items were; ‘unit leaders might treat me differently’, ‘unit members might have less confidence in me’ and ‘the military does not support Service members who have mental health problems’. A minority of personnel felt that ‘I would be blamed by my leaders’, that ‘help is not available’, that ’I do not know where to access support’ and that ‘I would think less of a colleague if they knew that they were receiving mental health care’. Stigmatising beliefs related to the self were more commonly reported than both stigma related to others and perceived BTC; the exception being the BTC perception ‘The military does not support Service members who have mental health problems’.

Table 3: Source of Mental Health Support.

Of the RIBS items; two thirds of respondents reported that they would continue a relationship with a friend who had a mental health problem; around half would live nearby and work with someone who they knew had a mental health problem. Around one third would be prepared to live with someone with a mental health problem.

The majority of personnel viewed mental health support as helpful and that help-seeking for a mental health problem required courage or strength. Around a quarter of personnel felt that mental health problems tend to work themselves out without help. Over 90% of those who screened positive for a mental health problem felt that mental health treatment was helpful and around 80% felt that help-seeking required courage or strength (Table 4).

Overall, the adjusted odds of reporting greater levels of Stigma/BTC were no higher amongst those who sought help for an emotional, family or stressful problem than those who did not (36.0%, n=58 vs. 29.2%, n=79, AOR 1.21, 95% CI 0.78-1.89). When the sample was stratified by class of help, personnel who sought help from military medical sources were no more likely to report high levels of stigma than those who had not sought help at all, however, those seeking help from non-medical sources had significantly increased odds of reporting stigma compared to non-help-seekers, an effect that remained significant after adjusting for a range of socio demographic factors and deployment in the last year. When the sample was stratified by interest in receiving help, helpseekers or those interested in receiving help had significantly increased odds of reporting Stigma/BTC compared to non-help-seekers who were not interested in receiving help. The greatest adjusted odds of reporting stigma/BTC (Wald test=15.44) were amongst those who were interested in help but had not yet sought it.

The adjusted odds of reporting higher levels of stigma/BTC were greater in those who had experienced contact with another person with a mental health problem. Viewing mental health management as requiring higher levels of personal responsibility was also associated with significantly increased odds of reporting stigmatising beliefs. Those who expressed potentially discriminatory views about people with mental health problems also reported significantly higher levels of Stigma/BTC.

When the sample was stratified by mental health status, the odds of reporting stigmatising beliefs were significantly higher in those reporting symptoms of either CMD and or PTSD but not amongst those drinking alcohol at potentially harmful levels. Personnel who screened positive for a mental health condition but had not sought help, had the greatest adjusted odds of reporting stigma/BTC compared to the reference group (Wald test=18.28). Compared to the negative screening non-help-seekers, higher rates of stigma/BTC were also found amongst help-seekers who screened positive for mental health problems (Wald test=8.85) (Table 5). There was a statistically significant trend for increasing levels of stigma to be reported, from the lowest levels amongst non-help-seekers who screened negative for a mental health problem, through moderate levels among negative screening help-seekers and positive screening help-seekers to the highest levels amongst positive screening non-help-seekers (χ² Test for Trend =25.23, p ≤ 0.0001) (Table 5).

This study describes the association between help-seeking, potential stigmatising beliefs about help-seeking and mental health problems in a randomly chosen sample of non-deployed British Army personnel; there were a number of key findings. Firstly, we found that around a third of those surveyed reported potentially harmful alcohol use and approximately a quarter of personnel screened positive for a probable mental health disorder. Secondly, of those who screened positive for a possible mental health disorder, around 40.0% had not sought any form of help despite the widely held view that mental health support was helpful or necessary and that help-seeking required courage or strength. Seeking support was lower amongst probable harmful alcohol users than potential CMD and PTSD cases; around three quarters of probable alcohol misusers had not sought any form of help. High levels of Stigma/BTC were reported by three quarters of all respondents though, overall, Stigma/BTC was no higher in help-seekers than nonhelp- seekers. We found that stigma/BTC were associated with potential intended discrimination, contact with mental health condition sufferers and screening positive for a possible mental health condition.

In this study, the act of help seeking was not associated with stigma/ BTC; however, harbouring potentially stigmatising beliefs about mental health and help-seeking were associated with mental health status. Despite nearly a quarter of personnel screening positive for a probable mental health condition, just under half of these personnel had not sought or received help from any source, which is in keeping with rates reported in previous surveys of military personnel [28]. Previous studies have reported a complex association between perceived need for support, attitudes about mental health, and severity of symptoms [29]. In our study, around a fifth of personnel who did not have substantial mental health symptoms nevertheless sought help for a family, stressful or relationship problem.

Table 4: Mental Health Stigma/BTC, Discrimination and Personal Responsibility for Mental Health Care.

Table 5: The Association of Stigma, Help Seeking, and Mental Health.

Stigma/BTC did not appear to be associated with help-seeking when mental health symptoms were absent; however, compared to personnel who were psychologically well and not seeking help, help-seekers who screened positive for possible mental health problems reported substantially more stigma/BTC and the highest stigma/ BTC rates were found amongst those who screened positive but had not sought help. It seems therefore, that perceptions of the potential negative consequences of help-seeking in the presence of mental health symptoms may have influenced the decision to seek support and that stigma/BTC are at their peak when personnel are psychologically unwell and recognise the need for help.

When asked to consider their willingness to access help, respondents stated they were least willing to engage with an online therapist and the unit chain of command was also unpopular; distrust of the chain of command corresponds with the most common stigmatising belief that the unit leadership might treat a person differently should they declare a mental health problem and is commensurate with the finding that those reporting more stigma/BTC also reported greater potential future discrimination. It seems however, that thoughts do not always translate into behaviour, as the chain of command was the second most frequently accessed source of support.

Commensurate with previous research [30], help-seeking personnel were more likely to access non-medical rather than military medical sources. Personnel appear to have the greatest concerns about possible adverse reactions from leaders and loss of military credibility with peers, while believing that they will not be supported by the military system. Given that there are potential short-term negative occupational consequences associated with military medical treatment of mental health problems in the form of medically imposed restrictions on deployment and other work activity [31], it is not surprising that personnel may avoid seeking out military medical help. What is more surprising is that personnel were able to overcome their stigma/BTC to seek support from the chain of command, which is very much in a position to adversely affect a military career, but can also be of great psychological benefit [32]. An alternative explanation for this finding is that military personnel are closely supervised and concealing a mental health problem by self-certification, where individuals supply their employer with a reason for short periods of sickness absence without consulting a doctor or by other forms of absenteeism is difficult. In this sense, military personnel may be compelled to interact with the chain of command when behavioural or psychological disturbance is present.

Stigma/BTC, discrimination and viewing mental health management as helpful or necessary and requiring a substantial personal contribution were clustered in those who expressed a desire for help with their mental health problems, but who had not sought support. Previous research has indicated that the perceived utility of mental health services is a fundamental component in the decision to seek help [33]. Amongst our sample population, there was a somewhat conflicted view of mental health where such problems are deemed undesirable, but that they can be effectively treated, though such treatment may be associated with potentially negative outcomes. Personnel appear to seek out help from non-medical sources that offer the possibility of negotiating a desirable occupational outcome or avoiding a potential negative consequence, rather than having an undesirable occupational outcome imposed by military medical practitioners.

Military commanders may therefore wish to consider the inclusion of a message in any stigma reduction programme that seeks to maximise appropriate help-seeking, that effective mental health treatment is available from military sources. Furthermore it could be emphasised that such support facilitates recovery and may help to avoid potential negative occupational outcomes. Such a strategy might also include families as they were an important source of support in this study.

Given the cross sectional nature of our data, we are unable to say whether the form of support influences long-term mental health outcome; however, published outcome data regarding occupational outcomes in UK AF personnel, suggests that around three quarters will return to full occupational fitness if they engage with military medical services [34-36].

In keeping with other published data, alcohol use was substantial in our sample with a full third screening positive for possible harmful alcohol use using a high screening cut off score. Only a quarter of those deemed to be cases had sought help. This figure is commensurate with the rate of help-seeking for alcohol problems reported by Iversen et al. [23,25] which was based upon data obtained during 2006-2007. Similar low rates of help-seeking amongst young people have been reported in international studies [37,38]. It seems that treatment engagement for alcohol misuse is a stubborn and widespread problem that has not improved over a five year period amongst UK AF personnel. Previous UK studies have reported rates of hazardous alcohol use of 67% using a cut off score of ≥ 8 on the 10 items AUDIT [39] and rates of 13% for alcohol use that might cause physical harm using a higher cut off score of ≥ 16. The rate of alcohol caseness using a substantial cut off score on the AUDIT-C in our study suggests that it is higher than previously measured in studies of UK AF personnel. This may be a reflection of the greater than expected number of young soldiers in this study, who are thought to be at greater risk of alcohol misuse [40].

We note that in this study stigma/BTC were lower in hazardous alcohol users than in those reporting mental health problems, so it is more likely, as the literature suggests [41], that personnel do not seek help for their drinking as they do not view it as concerning rather than in response to psychological impediments to care. This may be related to research findings that suggest that functional impairment appears to be concentrated in those reporting very high levels of alcohol use greater than the cut-off used in this study [42]. Some form of public health intervention that helps alcohol misusers to view their drinking as problematic might be helpful in promoting engagement with support.

Strengths and limitations

The strengths of this study included the random selection of personnel and achieving a high response rate. We used multiple assessment tools to assess the impact of both mental health and stigma/ BTC. Although we used an explicit definition of what constituted help-seeking, we were not able to objectively verify actual sources of support and we are unable to say whether those accessing medical forms of support received a definitive evidence-based intervention. Similarly, we cannot comment on the quality of support received from the non-medical support sources including the unit chain of command, chaplain and so on.

This study examined the psychological determinants of help seeking in a randomly chosen sample of British Army personnel. Around 40% of those screening positive for PTSD or CMD had not sought help or support. Stigma/BTC was highest amongst symptomatic personnel who were interested in receiving help for an emotional, stressful or family problem but had not sought help and was also substantial amongst symptomatic help-seekers. Although there appears to be a complex interaction between psychological symptoms, the desire for help and actual help seeking, we suggest that perceptions of the perceived potential negative occupational and social consequences of help-seeking may be more acute in those with poorer mental health and may act to influence the decision to seek support. We tentatively suggest that military stigma reduction strategies should focus upon reassuring personnel that adverse occupational consequences are not inevitable, not dissimilar to seeking help for a physical problem and that help-seeking from any source may be a useful first step to improved mental health. Such a strategy might also include families as they were an important source of support in this study. This strategy is unlikely to be appropriate for alcohol misusers who may in the first instance benefit from a strategy that helps them to view their alcohol use as potentially problematic from the individual’s perspective in terms of social, occupational or health impairment.

Partial funding for the project was received from a People Award grant from the Wellcome Trust; funding for military research personnel was received from the Ministry of Defence.