Participants

This study was based on a sample drawn from phase 1 of the King's Centre for Military Health Research (KCMHR) military health study. Full details of this study and responders can be found in Hotopf et al. ^{Reference Hotopf, Hull, Fear, Browne, Horn and Iversen6} Phase 1 was the first phase of a cohort study of UK military personnel in service at the time of the 2003 Iraq War. In total 4722 regular and reserve personnel who were deployed on the war-fighting phase of the Iraq War and 5550 regular and reserve personnel who were not deployed to Iraq at this time completed a questionnaire about their military and deployment experiences, lifestyle factors and health outcomes. A proportion of the study participants were subsequently deployed on later Iraq deployments, whose mission was counter-insurgency rather than war fighting.

The participant pool for the current study was drawn from the pool of individuals who returned completed questionnaires from phase 1 of the KCMHR military health study and gave signed consent to be followed up. Those who were likely to have psychiatric problems were identified using the 12-item General Health Questionnaire (GHQ) ^{Reference Goldberg11} data from phase 1. A random sample of those who scored above the threshold for ‘GHQ caseness’ (score of ≥4) were selected for interview together with a randomly selected subgroup of the non-GHQ cases. Case participants were oversampled; 70% of the sample were in the GHQ case group, and 30% were in the non-GHQ case group. Within the case and non-case groups, in order to ensure adequate power to make key statistical inferences, we stratified the sample by regular/reserve status (50% regular, 50% reserve), and deployment status (50% deployed on the main fighting period of the Iraq War (TELIC 1), 50% deployed elsewhere or not deployed). In all other respects, group participants were representative of the KCMHR military health study responders with regards to service branch and demographic characteristics (age, rank and ethnicity) and in turn the KCMHR military health study was representative of the UK military in 2003 and those who deployed on the main war fighting phase of the Iraq War. ^{Reference Hotopf, Hull, Fear, Browne, Horn and Iversen6} From the 1083 participants who were eligible, we achieved a response rate of 76% (n = 821).

Interview schedule

The telephone interview schedule included the following sections: deployment experience since 2003; Patient Health Questionnaire (PHQ) ^{Reference Spitzer, Kroenke and Williams12} with an additional measure for the diagnosis of posttraumatic stress disorder (PTSD; the 4-item Primary Care PTSD Screen (PC–PTSD)); ^{Reference Prins, Kimerling, Cameron, Oumiette, Shaw and Thraikill13} perceived needs, health service use (based on a modified version of the Client Services Receipt Inventory), ^{Reference Chisholm, Knapp, Knudsen, Amaddeo, Gaite and van Wijngaarden14} receipt of treatment, and barriers to care (these will be reported in a separate publication). The only demographic information collected was change in marital status and rank since all other relevant information was collected during phase 1. ^{Reference Hotopf, Hull, Fear, Browne, Horn and Iversen6} All interviews were conducted during 2006–7.

Diagnostic assessment

We used the Patient Health Questionnaire (PHQ) as a diagnostic instrument to measure the prevalence of common mental disorders, ^{Reference Spitzer, Kroenke and Williams12} which has been validated for telephone use. ^{Reference Pinto-Meza, Serrano-Blanco, Penarrubia, Blanco and Haro15} Standard diagnostic algorithms were used to score the PHQ. For PTSD symptoms, we used a short measure developed for primary care by the National Center for PTSD (PC–PTSD) ^{Reference Prins, Kimerling, Cameron, Oumiette, Shaw and Thraikill13} and we used a score of 3 or more to define caseness. The screen is used in the Post Deployment Health Assessment (PDHA) ^{Reference Hoge, Auchterlonie and Milliken5} and Post Deployment Reassessment (PDHRA) mandated by the US Department of Defense. ^{Reference Milliken, Auchterlonie and Hoge16} In order to improve the specificity of the measure, we included a lifetime DSM–IV Criterion A1 event-screening questionnaire taken from the National Comorbidity Study. ^{Reference Kessler, Somnnega, Bromet, Nelson, Hughes and Nelson17}

Mental health care utilisation

For those who self-reported an ‘emotional, stress-related, alcohol or family problem’ (hereafter referred to as a mental health problem), we enquired about their past year's contact with a variety of mental healthcare providers in both primary (general practitioners (GPs) and medical officers) and secondary care (community psychiatric nurses, psychologists and psychiatrists). Comorbidity was common but often unrecognised, and it was not possible to ask about individual diagnoses in turn; instead we opted to ask about service utilisation in relation to the most significant problem as defined by the participant. We also asked about their contact in the past year with non-medical professional service providers (such as social workers, welfare officers and chaplains). Finally, we enquired about informal sources of help (for example partners/spouses, internet searches and friends). For each medical provider, we asked about the number of contacts in the past year and for all other sources of help we invited participants to respond simply ‘yes’ or ‘no’.

Treatment receipt

For those who reported a mental health problem, we asked whether they had received any treatment in the past year. If they reported receipt of treatment, we then asked whether they had received: psychotherapy or counselling; medication; and/or cognitive–behavioural therapy (CBT). For those who reported taking medication, we asked whether they had taken it as instructed in order to determine adherence. In order to determine the type of talking treatment that participants had received, and specifically to detect whether they had received CBT, interviewers were trained to ask about a series of cue words and phrases that would identify the psychotherapy as CBT (for example, between-session work or homework, agenda setting with the therapist etc).

Analysis

All statistical analyses were undertaken using the statistical software package STATA (version 10.0 for Windows). The majority of the data reported are descriptive statistics (percentages with their 95% confidence intervals). To compare service use among regulars, reservists and veterans, odds ratios are presented with 95% confidence intervals, using regular still-serving personnel as the reference group. Percentages are weighted to take account of the sampling strategy used for this study. For the purposes of this paper, the majority of the analyses reported are based on those that have a diagnosis on the structured clinical interview, and perceived that they had a problem (n = 281).

Ethical Issues

The study received approval from both the King's College Hospital NHS Research Ethics Committee (ref: 05/Q0703/155) and also from the Ministry of Defence (Navy) Personnel Research Ethics Committee (ref: 0522/22).

Perceived mental health problems

For those who had a diagnosis identified at interview, we examined individuals' perception that they had a ‘stress, emotional, alcohol or family problem’. Three-quarters of regulars with either depressive/anxiety disorders or PTSD perceived that they had a problem, whereas just under half of those with alcohol misuse perceived they had a problem (Table 1). Findings were similar for reservists. Although not significant, veterans were, in general, more likely than regulars and reservists to recognise that they had a diagnosis, with 97% of veterans with PTSD perceiving they had a mental health problem.

Table 1 Percentage of those with a Patient Health Questionnaire (PHQ) diagnosis, depressive/anxiety disorder, alcohol misuse or post-traumatic stress disorder (PTSD) symptoms who perceived that they had a mental health problem^a in the past year, by serving status

	Overall (n = 821)		Regulars (n = 301)		Reserves (n = 302)		Veterans (n = 218)
	Weighted %	95% CI	Weighted %	95% CI	Weighted %	95% CI	Weighted %	95% CI
Any PHQ diagnosis (n = 325)	56.1	43.8–67.7	57.0	42.0–70.8	65.5	48.6–79.3	69.4	47.7–85.0
Depressive/anxiety disorder (n = 205)	74.5	60.9–84.5	75.3	51.4–89.8	75.0	51.3–89.5	72.9	46.9–89.2
Alcohol misuse (n = 196)	62.2	52.3–71.3	46.2	29.1–64.2	56.3	36.1–74.6	77.5	46.5–93.2
PTSD (n = 110)	79.7	57.1–92.1	78.0	35.9–95.7	70.0	36.0–90.6	97.2	90.1–99.2

Help-seeking

Those who perceived they had a mental health problem, were asked about help-seeking (Table 2). Over 80% of those who perceived they had a problem had sought some type of help. Most help-seekers had made use of informal sources of support such as a spouse or friend, rather than seeking professional help. Only 23% of those with alcohol problems had sought professional help compared with 50% of those with depression/anxiety and 64% of those with PTSD. Rates of informal help-seeking for alcohol were also lower (73%) than for the other disorders (depression 82%, PTSD 85%). Regulars did not differ from reservists in terms of help-seeking behaviours. Veterans were less likely to use informal sources of help for depression than regulars, but this difference was no longer significant once confounders were adjusted for (age and deployment status).

Table 2 Prevalence of help-seeking for those with a perceived mental health problem^a in the past year

	n	Weighted %
Perceived mental health problem (n = 508)
Any help	441	82.1
Any professional help	196	31.2
Medical professional helpb	129	18.5
Non-medical professional helpc	108	17.5
Any informal helpd	412	78.2
Perceived mental health problem + depressive/anxiety disorder (n = 173)
Any help	156	88.5
Any professional help	85	49.9
Medical professional helpb	65	30.9
Non-medical professional helpc	46	28.8
Any informal helpd	141	82.2
Perceived mental health problem + alcohol misuse (n = 147)
Any help	130	79.0
Any professional help	58	23.2
Medical professional helpb	38	15.6
Non-medical professional helpc	37	14.7
Any informal helpd	120	72.6
Perceived mental health problem + PTSD symptoms (n = 95)
Any help	90	97.3
Any professional help	48	63.6
Medical professional helpb	34	54.3
Non-medical professional helpc	29	26.9
Any informal helpd	83	84.6

Characteristics of help-seekers

We examined a series of potential demographic associations with help-seeking (further detail available from the authors on request). Univariate analysis showed that for regulars, having no qualifications made it more likely that an individual would seek help. For reservists, females were more likely to seek help. For veterans, being in the Royal Air Force, being medically downgraded and having a diagnosis (on the PHQ or on the PC–PTSD) was associated with help-seeking. There was an association between functional impairment of borderline significance in all groups (test for trend P = 0.003). Of note, still-serving personnel who deployed to Iraq on the main war fighting phase of the war were not more likely to seek help.

Sources of help

Among those who perceived they had a mental health problem and were receiving professional help, the most frequently used source of medical professional help was from a GP or medical officer (this varied from 72% for alcohol misuse to 98% for depressive/anxiety disorder) (online Table DS1). Because of the small numbers of people who had sought help it was difficult to determine whether there were any real differences in sources of help for regulars, reservists or veterans (data not shown but available from the authors).

Interest in receiving help

Those who perceived that they had a mental health problem but were not seeking help were asked whether they were interested in receiving help (online Table DS2). Only a small proportion of personnel in this group were interested in receiving help (this varied from 15% for alcohol misuse to 18% for PTSD). Veterans were significantly more likely to be interested in receiving help than regulars for depressive/anxiety disorder (adjusted odds ratio (OR) = 2.58, 95% CI 1.45–4.59) and alcohol problems (adjusted OR = 2.87, 95% CI 1.46–5.64). There was no difference between regulars and reservists in their interest in receiving help for any diagnosis examined, or for regulars and veterans for PTSD symptoms.

Treatment being received

Of those who perceived that they had a mental health problem, had a diagnosis of depressive/anxiety disorder and were seeking help from medical sources, 76.5% were receiving treatment (Table 3). In total 55.8% were receiving medication, 50.6% were receiving counselling/psychotherapy and 3.2% had received in-patient treatment. For those who had alcohol problems, 84% were receiving treatment, the majority of which was counselling (47%), in-patient treatment (25%) and medication (25%). There was no difference between regulars, reservists and veterans in the type of treatment being received (data not shown but available from the authors), but numbers receiving treatment within the overall sample are small. Few service personnel with a diagnosis of depressive/anxiety disorder were receiving CBT (12.9%). The most common medications prescribed were antidepressants (84%) and benzodiazepines (11%). Adherence with medication was 86% for depression, 40% for alcohol misuse and 77% for PTSD. There were no differences between regulars, reservists and veterans in rates of adherence.

Table 3 Proportion of those with perceived mental health problems and seeking help from medical sources (within the past year) by diagnosis

	n	Weighted %
Depressive/anxiety disorder (n = 65)
Any treatment	44	76.5
Medication	34	55.8
Cognitive–behavioural therapy	7	12.9
Other psychotherapy/counselling	26	50.6
In-patient stay	1	3.2
Alcohol misuse (n = 38)
Any treatment	24	84.3
Medication	16	25.3
Cognitive–behavioural therapy	1	6.3
Other psychotherapy/counselling	14	46.8
In-patient stay	1	25.1
Post-traumatic stress disorder symptoms (n = 34)
Any treatment	27	74.1
Medication	19	29.2
Cognitive–behavioural therapy	7	6.0
Other psychotherapy/counselling	16	47.4
In-patient stay	2	19.6

Key findings

Three key finding emerge from this study. The first is that more than three-quarters of those with a diagnosis chose not to access medical help for their problem, but instead are making use of informal sources of help such as chaplains. This reticence to seek medical help is most evident among those with alcohol problems and depression, compared with those with PTSD. Second, the majority of those with a diagnosis are not receiving treatment. Third, a percentage of interviewees with a diagnosis on structured interview fail to recognise that they suffer from a mental disorder and might benefit from treatment; recognition was lowest for alcohol problems and depression/anxiety and better for PTSD.

Strengths and weaknesses of the study

To our knowledge, this is the first study to date that has systematically examined service utilisation and treatment receipt in a representative sample of the UK military. The strengths of this study are the relatively large sample and high response rate, with no evidence of bias in terms of health between responders and non-responders. ^{Reference Iversen, van Staden, Birtles, Hacker Hughes, Browne and Langston10} The sample is diverse, representing all three services, both those who have deployed and those who have not, and includes reservists and those who have left. The study used a structured diagnostic instrument, and did not rely on questionnaire self-report of symptoms or distress. As data collection took place independently of the military, data quality was unlikely to be unaffected by participants' concerns that their problems would have been reported back to the chain of command. Although our response rate was high, our sample was already based on a 61% response rate. ^{Reference Hotopf, Hull, Fear, Browne, Horn and Iversen6} It is possible that we missed the most vulnerable, unwell or socially excluded members of the still-serving or veteran population such as those in prison or those who are homeless. But importantly, we found no evidence that non-participants were less healthy than participants in the original sample. ^{Reference Tate, Jones, Hull, Fear, Rona and Wessely18}

We used the PHQ and PC–PTSD instruments to detect common mental disorders, and these measures were chosen for their brevity, acceptability, and validity in military populations. ^{Reference Hoge, Auchterlonie and Milliken5,Reference Iversen, Dyson, Smith, Greenberg, Walwyn and Unwin19} However, like all tests for mental disorders, they have limitations. Many argue that the existing measures in use for common disorders and PTSD are unable to sift out those with mental disorders that result in functional impairment and require treatment. ^{Reference Rona, Hyams and Wessely20–Reference Rona, Jones, Iversen, Hull, Greenberg and Fear22} Of relevance to this study, we cannot be sure that the common mental disorders we have detected are disabling and persistent and therefore require treatment, or whether service personnel are in fact making the correct decision not to consult a professional about their symptoms. Even if this were true, a proportion of those scored as cases on our screening measures will include those who are functionally impaired and overall rates of help-seeking and treatment receipt among those with symptoms is low. As a result of the low numbers of help-seekers, we had small numbers when comparing receipt of different services or treatments and thus we may have been underpowered to detect true differences between groups.

We did not obtain objective data on service utilisation, and therefore were unable to ascertain the validity of the self-report methodology we used. Interviewers were carefully trained to probe to ascertain the sort of talking treatment that individuals were receiving; nonetheless, it is possible that some people were receiving CBT but did not identify their treatment as such. Although we relied on self-recall of the past year's service utilisation and treatment receipt, recent studies suggest that this is a relatively accurate method of obtaining information on service use, ^{Reference Beebe, McRae and Barnes23} particularly if the interviews are conducted in a setting in which confidentiality is preserved. ^{Reference Aquilino24,Reference Aquilino, Wright and Supple25} It has been suggested that recall bias may lead to an overestimation of self-report of service use in people who have more distressing disorders. ^{Reference Rhodes, Lin and Mustard26} If this has occurred, our results will have underestimated the extent of unmet need, especially among those with serious disorders.

Recognition of mental disorders

A substantial proportion of those we diagnosed with common mental disorders, especially those with an alcohol problem, did not acknowledge that they had experienced a stress, emotional or alcohol problem. It has been suggested that ‘mental health literacy’ is an important first step in the pathway to care. ^{Reference Jorm, Korten, Jacomb, Christensen, Rodgers and Pollitt27} Subjective perception of a mental disorder has been shown to influence the likelihood of help-seeking and receipt of treatment, especially in primary care. ^{Reference Verhaak, Prins, Spreeuwenberg, Draisma, van Balkom and Bensing28} Poor recognition of mental disorders is not unique to the military; only 39% of the Australian population were able to recognise depression when described in a vignette, ^{Reference Jorm, Korten, Jacomb, Christensen, Rodgers and Pollitt27} and men are less likely to recognise mental illness than women. ^{Reference Wright, Harris, Wiggers, Jorm, Cotton and Harrigan29} The physical symptoms of anxiety and depressive disorders are particularly poorly recognised. ^{Reference Barkow, Heun, Ustun, Berger, Bermejo and Gaebel30,Reference Regier, Hirschfeld, Goodwin, Burke, Lazar and Judd31} Mental health literacy is clearly important for the military as they are at increased risk of occupational psychiatric injury. It would be interesting to compare recognition of mental disorders in the military with similar occupational groups such as firefighters or police.

Help-seeking

In the veteran population we found that objective evidence of poor health (as evidenced by previous medical downgrading and/or the presence of a current diagnosis on the PHQ/PC–PTSD) predicted help-seeking, but we did not demonstrate this for still-serving personnel or reservists. Subjective functional impairment predicted service use in all groups, but the association was of borderline significance. This is of concern and is contrary to what is reported in the literature on mental health service use in the general population, which has consistently suggested that the most unwell are the most likely to receive treatment. ^{Reference Verhaak, Prins, Spreeuwenberg, Draisma, van Balkom and Bensing28,Reference McManus, Meltzer, Brugha and Bebbington32}

Sources of help

The percentage of personnel who are seeking help from medical sources is low, despite access to both primary and secondary care being easily available for those in the still-serving military, and bespoke services provided for reservists (The Reservists Mental Health Programme) ³³ and veterans (for example, by the charity Combat Stress). These findings are similar to those reported in general population studies, ^{Reference Verhaak, Prins, Spreeuwenberg, Draisma, van Balkom and Bensing28,Reference McManus, Meltzer, Brugha and Bebbington32,Reference Wang, Lane, Olfson, Pincus, Wells and Kessler34} and also similar to the rates reported from the USA and Canada for active duty personnel. ^{Reference Hoge, Castro, Messer, McGurk, Cotting and Koffman4,Reference Fikretoglu, Guay, Pedlar and Brunet35,Reference Rhodes, Lin and Mustard26} Hoge et al reported that only 27% of those returning from combat in Iraq with a diagnosis had seen a mental health professional. ^{Reference Hoge, Castro, Messer, McGurk, Cotting and Koffman4} The rate of help-seeking from professionals for alcohol problems in comparison with depression and PTSD is especially low and has been reported previously in military populations. ^{Reference Erbes, Westermeyer, Engdahl and Johnsen37}

The general population literature has consistently demonstrated that, regardless of what services are made available, people prefer to seek help for mental disorders from social networks such as friends, family and the clergy ^{Reference Wang, Berglund and Kessler38} rather than to consult a professional. ^{Reference Jorm, Korten, Jacomb, Christensen, Rodgers and Pollitt27,Reference Parker and Brown39} This may be especially true in the ‘macho’ culture of the military where great emphasis is placed on resilience and self-sufficiency during training. Indeed such attributes are an asset and adaptive during deployment, but become problematic when maintained after homecoming.

The very real career implications of formal presentation with mental health problems while still serving also need the borne in mind and are likely to be a potent driver steering participants to non-professional sources of help. We know that mental health problems do have career implications for certain occupational roles within the military – for example, whether personnel are allowed to carry weapons, or pilot military aircraft. If service personnel do choose to consult their doctor, although confidentiality is a guiding principle, it cannot be absolute. The military chain of command may be informed if a member of their unit is suffering from a health problem, physical or mental, which is deemed to affect their employability. Although the health information that is passed on is general in nature and covers basic employability issues, rather than health details, this is still presumably a deterrent for personnel. We have previously shown that non-medical sources of help are preferable to peacekeepers returning with psychological difficulties. ^{Reference Greenberg, Maingay, Iversen and Wessely40} The implication of this for the military is that peer-led interventions such as Trauma Risk Management (TRiM) may be especially acceptable and useful. ^{Reference Greenberg, Langston and Jones41} It also suggests that programmes designed to educate non-medical providers about the symptoms of common mental illness, and appropriate routes for referral may be fruitful. For example, it may be possible to train non-medical help providers such as clergy in simple interventions that can aid recovery as suggested by Schnell et al. ^{Reference Schell, Marshall, Tanielian and Jaycox42}

For those who have left, it highlights the role that service charities play in helping veterans to maintain the cohesive social network they had in the military, and also to signpost those who are unwell and require treatment.

The majority of those seeking medical help for their mental disorder are dealt with in primary care; the same is true in the general population. ^{Reference McManus, Meltzer, Brugha and Bebbington32} The general population literature suggests that the majority of the population prefer to be seen in primary care for a mental disorder, ^{Reference Jorm, Korten, Jacomb, Christensen, Rodgers and Pollitt27,Reference Priest, Vize, Roberts, Roberts and Tylee43,Reference Wolff, Pathare, Craig and Leff44} and GPs are considered by the general public to be skilled in the assessment and treatment of such conditions. ^{Reference Verhaak, Prins, Spreeuwenberg, Draisma, van Balkom and Bensing28} The implications of this are that primary care providers are ideally placed to assess and begin treatment for common mental disorders seen in military personnel.

Receipt of treatment

Given that there are evidence-based effective treatments available for the disorders we identified in this occupational group (alcohol misuse, depressive disorders and PTSD), most individuals have the potential to return to normal levels of functioning with appropriate care. In this sample most respondents who were in receipt of treatment were receiving medication or talking treatments such as counselling (not CBT), despite CBT being a recognised first-line treatment for PTSD ⁴⁵ and moderate/severe depression. ⁴⁶ The low levels of personnel receiving evidence-based treatments for common mental disorders is a concern. A recent Improving Access to Psychological Therapies (IAPT) initiative that aims to make access to CBT within primary care settings swifter and easier for veterans is timely and relevant (www.iapt.nhs.uk/special-interests/veterans).

Comparisons with other deployed forces

Our results are comparable with rates of help-seeking reported in UK veterans of the 1991 Gulf War. Iversen et al reported that 58% of those with a diagnosis on the Primary Care Evaluation of Mental Disorders (PRIME–MD) were currently seeking help for their problem. ^{Reference Iversen, Dyson, Smith, Greenberg, Walwyn and Unwin19} This suggests that despite the initiatives to facilitate help-seeking and to decrease the stigma of mental disorders implemented since 2003, ^{Reference Gould, Greenberg and Hetherton47} the percentage of people who seek help has not increased in the past 5 years. This is in contrast to US work that has demonstrated a significant increase in the percentage of returnees seeking mental healthcare after conflict in recent years, although this is almost certainly a side-effect of screening, and has yet to be matched with evidence of improved outcomes. ^{Reference Hoge, Auchterlonie and Milliken5,Reference Rosenheck and Fontana48} The reasons for the lack of change in the UK are complex but are in part the result of the glacial pace of behavioural and cultural change especially within large organisations. We also know that prior to the current ‘Time to Change’ programme ( www.time-to-change.org.uk), anti-stigma interventions to date (whether they be military or civilian) have concentrated on awareness raising and attempts to change general attitudes towards mental illness, rather than being directly designed to change behaviour or promote help-seeking, and the evidence that knowledge or attitude interventions alone (or even in combination) change behaviour is weak. ⁴⁹

Our results are also comparable with those reported by Schell et al, in a large telephone-based survey of Iraq War returnees. ^{Reference Schell, Marshall, Tanielian and Jaycox42} The uniformity of unmet need across both countries, despite very different models of service delivery for veterans, suggests that the problem with help-seeking may be intrinsic to military personnel themselves, and not the services available. We explore barriers to care in a subsequent paper, but previous reports have identified stigma as the dominant reason for reticence in seeking care. ^{Reference Hoge, Castro, Messer, McGurk, Cotting and Koffman4,Reference Schell, Marshall, Tanielian and Jaycox42}

Comparisons with the general population

Rates of help-seeking between the general population and our sample are broadly comparable. The recent Adult Psychiatric Morbidity Survey (APMS) conducted in England in 2007 ⁵⁰ revealed that 22% of those screening positive for PTSD had some form of community care within the past year compared with 20% of regulars, 14% of reservists and 24% of veterans screening positive for PTSD in our study. In the APMS, 65% of those with a depressive disorder had spoken with their GP in the past year compared with 93% of regulars, 87% of reservists and 72% of veterans in our study (although in our study we cannot be sure they consulted for their mental health problems).

Rates of help-seeking reported here are also comparable with historical reports of the general population in the UK, ^{Reference Meltzer, Bebbington, Brugha, Farrell, Jenkins and Lewis51} Europe ^{Reference Alonso, Codony, Kovess, Angermeyer, Katz and Haro52} and the USA general population. ^{Reference Wang, Lane, Olfson, Pincus, Wells and Kessler34} It is perhaps surprising that the military population is comparable with the general population in terms of their help-seeking behaviour, given that certain socioeconomic groups at risk of non-consulting are over-represented in the military (e.g. younger men, those from lower socioeconomic groups). However, as this is a group at high risk of occupational psychiatric injury, it could be argued that they should be even more prepared to seek help than the general population when they have symptoms, especially as mental health services can be easily and quickly accessed by those who are still serving.

In this sample, the overall rate of treatment for those in the still-serving population with mental disorders (13%) is even lower than that reported in the general population (26% received treatment in a European study, ^{Reference Wittchen and Jacobi53} 33% received treatment in an American study). ^{Reference Kessler, Demler, Frank, Olfson, Pincus and Walters54}

Implications

Rates of help-seeking from medical sources and receipt of treatment are low in this sample of UK armed forces personnel, especially for alcohol misuse and depression/anxiety. Instead, personnel indicate a clear preference to consult peers, friends and non-medical sources of help such as chaplains. These results are similar to those reported for military personnel in the USA, and to rates in the general population. It has often been assumed that the unmet burden of mental illness in the still-serving and veteran community is as a result of a paucity of appropriate services being available. This study indicates that the main reason that people do not access services is that they make a decision much earlier in the pathway to care that that they do not need help. Therefore, novel ‘system level’ interventions to improve service personnel's recognition of common mental disorders that require treatment, and reduce the stigma of consulting, are important areas for research and evaluation.