The European Journal of Public Health Advance Access originally published online on March 8, 2005
The European Journal of Public Health 2005 15(2):175-184; doi:10.1093/eurpub/cki128
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Public Mental Health |
What happens to British veterans when they leave the armed forces?
Amy Iversen, Vasilis Nikolaou, Neil Greenberg, Catherin Unwin, Lisa Hull, Mathew Hotopf, Christopher Dandeker, John Ross and Simon Wessely*Kings Centre for Military Health Research, Guy's, King's and St Thomas’ School of Medicine, King's College, London, UK
Correspondence: Dr Amy Iversen, Department of Psychological Medicine, Weston Education Centre, London SE5 8RJ, UK, tel. +44 20 7848 0796, fax +44 20 7848 0408, Email: a.iversen{at}iop.kcl.ac.uk
 |
Abstract |
|---|
 Top Abstract METHODSResultsDiscussionReferences |
|---|
Background: Little is known about the factors associated with leaving the armed forces, or what predicts subsequent employment success for veterans. It is likely that there is a complex interaction of adverse social outcomes and mental health status in this group. Method: Analysis of existing data from the King's Military Cohort, a large, randomly selected, longitudinal cohort of service personnel, many of whom have now left the armed forces. The sample consisted of 8195 service personnel who served in the armed forces in 1991; a third deployed to the Gulf (1990–91), a third deployed to Bosnia (1992–97) and the final third an ‘Era’ control group in the Armed Forces in 1991 but not deployed. Results: The majority of service leavers do well after leaving and are in full-time employment. Those with poor mental health during service were more likely to leave and had a greater chance of becoming unemployed after leaving. Mental health problems appear to remain static for veterans after leaving. Veterans of the Gulf War enjoyed more favourable employment outcomes, provided that they came home well. Conclusions: Only a minority of veterans fare badly after service, even amongst those with active tours of duty behind them. Veterans with mental health problems during service seem to be at higher risk of social exclusion after leaving and therefore these individuals represent an especially vulnerable group of the veteran population.
Keywords: veteran, employment, mental health
Each year in the UK 18 000 men and women leave the armed forces and enter civilian life. They leave for a variety of reasons and in a variety of different circumstances. Some have served lengthy terms of service. Others have not even completed their basic training, and we know little about the fate of either group.
There has been a resurgence of interest recently in what happens to ‘veterans’. Media coverage has focused on the minority who fare badly and drift down into social exclusion such as those who become homeless and/or have severe mental health problems.1 In addition, there is increasing international recognition, both within the military and within civilian society, of ‘post-conflict dysfunction’2 including (but not limited to) Post Traumatic Stress Disorder (PTSD) in ex-service personnel, particularly amongst those who have served in the post-Cold War years.3–5
We know from the American literature that military service, for most people, has a positive effect on the life trajectory.6 Studies of men who served during the Second World War demonstrated that, overall, veteran status afforded these individuals greater educational opportunities, better qualifications, higher employment rates and bigger wage packets.7,8 In Germany, although those who fought in the Second World War initially had unfavourable occupational outcomes, these effects diminished rapidly over time as the German economy started to recover.9
Vietnam returnees did not fare so well. Most of the US literature suggests that people returning from the Vietnam War did worse than those who stayed at home in terms of earnings, jobs and educational attainment.10–12 This disadvantage was most acute in those with mental health problems.13 Leavers who suffered from PTSD often ended up with lower salaries,14 and more enduring drug and alcohol problems, as well as more chance of ending up in prison.15
The US literature of returnees from the Gulf War has been no more encouraging, particularly concerning symptomatic individuals exposed to combat.16 Individuals who experienced combat were more likely to end up unemployed or fired from their job,17 within the prison system or alcohol dependent.18 This is despite the fact that combat action during The Gulf War was not comparable in scale, duration or intensity with the other conflicts upon which the main body of literature is based.
In the military population in general (theatre-specific studies aside) those with mental health problems are more likely to leave service prematurely,19 experience lost work days16 and also more likely to end up socially excluded (e.g. homeless).20,21
Whilst much has been written about veterans of the US armed forces, very little is known about their UK counterparts. There has never been any systemized follow-up of ex-service personnel in the UK. Whilst there is a substantial body of American literature, it must be interpreted with considerable caution for the UK, not least because of the existence of the Veterans Administration (VA) in the USA providing bespoke specialist services for veterans, which is not the case in the UK.
In this paper, the King's military cohort was used; a large random sample of the armed forces upon whom we have previously published.22,23 We have recently published the results of our follow-up of the cohort,24 focusing on the health outcomes associated with serving in the Persian Gulf. The present study has a rather different focus, looking at social outcomes, and is interested in UK service-leavers in general, rather than Gulf Veterans in particular. The questions addressed were as follows:
- What are the characteristics of leavers? What factors are associated with leaving?
- Are people who leave early (<4 years service) different from others?
- What factors are associated with being employed after people leave?
- What happens to people's symptoms once they leave the services?
- Does outcome differ for early leavers?
- Mental health will improve for those who leave the military early.
- Those who serve longest in the military are more likely to have mental health problems on leaving due to institutionalization.
|
METHODS |
|---|
|
TopAbstractMETHODSResultsDiscussionReferences |
|---|
Initial survey (phase 1)
Our analyses were conducted on a subset of the King's UK military cohort who completed the phase 1 Health Survey of Military Personnel (n=8195) in 1997.22 The initial survey in 1997 consisted of a cross-sectional postal survey of three groups: those personnel who served in the Gulf region between 1 September 1990 and 30 June 1991 (Gulf cohort), personnel who had served in Bosnia between 1 April 1992 and 6 February 1997 (Bosnia cohort) and personnel who were serving in the Armed Forces on 1 January 1991 but who were not deployed to the Gulf conflict (‘Era’ cohort). Special Forces were excluded for security reasons. The final sample was a random stratified sample of 4250 who served in the Gulf. For comparison cohorts, 4250 service personnel deployed in Bosnia and 4246 non-serving but ‘Era’ service personnel were selected.
At the end of phase 1, we performed a series of detailed nested case controlled studies involving direct clinical investigations of sick and well veterans (see, for example, Sharief et al.)25 . These studies are not reported here, but to avoid confusion with the rest of our publications, we will continue to refer to the follow-up study that is reported in this paper as ‘phase 3’.
Follow-up 2001 (phase 3)
For follow-up, two stratification variables were used, fatigue and gender. All female veterans who completed the phase 1 questionnaire (n=648) were contacted, as females were over sampled in the original cohort. In addition, this allowed us to examine any gender differences in follow-up variables.
The initial research aims of the Gulf study focused on examining the health of symptomatic Gulf veterans. Several studies confirm that fatigue is the one symptom that consistently links the often otherwise disparate health complaints in Gulf War Veterans, and it is also the commonest. As a result, at follow-up, fatigue was utilized as a core outcome measure and a proxy measure for being symptomatic, and the sample stratified according to degree of fatigue reported. In order to ensure that the most severely ill were well represented, all male veterans with a fatigue score of 9 or more were included. A 1:2 sample of male Gulf veterans with mid range fatigue scores of 4–8, along with all Bosnia and Era veterans scoring in this range were selected. Finally, an 
1:8 sample of veterans with fatigue score <4 was selected in order to represent unsymptomatic individuals (250 in each group). The total sample size at follow-up was 3322.
Measures (phases 1 and 3)
Full details of the questionnaire, the methods of data collection, efforts made to trace and the detailed analysis of responders and non-responders are contained in the full reports of the phase 1 study.22
To summarize, taking into account undelivered questionnaires, the phase 1 survey had an effective response rate of 70.6%. [At phase 1, response rates differed slightly between the cohorts: Gulf (70.4%), Bosnia (61.9%) and Era (62.9%).] The commonest reason for non-response was failure to identify a final valid address for participants, which introduces the small but appreciable bias of the possibility that these individuals were homeless. Full details of all outcome measures are contained in the final study reports.22,23
At phase 3, the questionnaire was modified and tailored according to whether the participant was still serving. Full details of the follow-up methodology are presented elsewhere.24 Again, exhaustive attempts were made to keep response rates high with three waves of mailing, telephone tracing of non-responders and use of the electoral register and Department of Social Security (DSS) records to verify correct address information. The response rate at phase 3 was 71.6%. [Again, response rate varied between cohorts: Gulf (73.8%), Bosnia (70.2%) and Era (69.5%).]
Design
The nature of the data set has allowed us to conduct two sorts of analyses:
- A cross-sectional analysis of the associations of leaving and employment from the phase 1 study.
- A more powerful prospective longitudinal design using phase 1 variables to predict outcomes at phase 3.
|
Statistical analysis
The data was analysed used SPSS (version 11.0) and STATA (version 6.0). Several variables of interest were selected from the original data set. These were: demographic details (age, sex, education, marital status), military details (service status, rank, cohort, number of deployments), alcohol intake and cigarette smoking.
For the purposes of our analysis, we generated two new variables: ‘employment’ and ‘education’.
A composite variable labelled ‘Post Traumatic Stress Reaction’ (PTSR) was created for questions asked at phase 1 and phase 3 (figure 2). This consisted of seven symptoms of PTSD taken from the Mississippi Scale.26
In order to validate this measure, Pearsons Correlation Coefficient was calculated on a small subset of the original sample (n=66) for summed PTSR score, and summed score on another extensively validated measure, the Post Traumatic Stress Disorder Checklist, Military Version.27
The measures were highly correlated (0.62, P
0.01). We chose deliberately to construct a de novo composite measure for post traumatic stress symptoms rather than using a full version of an existing PTSD schedule as historically/politically it was crucial at the time that this study was set up that the questions were not seen to be exclusively about mental health. Thus, we deliberately embedded the measure within other more general health measures.
|
A cut-off of 3 or more was used to decree ‘psychiatric caseness’ on the 12 item General Health Questionnaire (GHQ).28 We carried out a logistic regression to investigate the factors associated with leaving the military and those associated with unemployment after leaving. Odds ratios (ORs) along with their 95% confidence intervals (CIs) were reported after adjusting for all the potential confounders. In any analyses which involved phase 3 data, we adjusted for those confounders that were the most significant from phase 1. Those were determined by using analysis of deviance. To take account of potential sampling bias, that is the probability of being selected at phase 3 depending on the fatigue score at phase 1, we generated probability weights (pweights). These are inversely proportional to the sampling fraction, which is the probability of a given observation in phase 1 being chosen for phase 3 (see table 7). Furthermore, analysis of covariance (ANCOVA) was carried out to assess formally a change in scores of ill-health outcomes from phase 1 to phase 3.
|
|
Results |
|---|
|
TopAbstractMETHODSResultsDiscussionReferences |
|---|
We investigated the factors associated with leaving the military and those associated with unemployment after leaving by calculating adjusted ORs for each dependent variable, adjusting for other potential confounders.
Results are presented as OR and adjusted OR (AOR) with 95% CIs. Any analyses which involved phase 3 data have been adjusted for the sample bias at phase 3 caused by stratification by using pweights (details above). In the data tables this is referred to as ‘adjusted for phase 3 sampling frame’.
Characteristics of leavers at phase 1 (table 1)
There were 2908 leavers by phase 1 (36%). When compared with ‘non-leavers’, by far the most significant predictors of leaving were cohort and rank. Those who deployed to Bosnia were less likely to be leavers than their counterparts in other cohorts (AOR 0.18, CI 0.15–0.23), even after adjusting for age and length of service. Officers (AOR 0.11, CI 0.08–0.15) and non-commissioned officers (NCOs) (AOR 0.20, CI 0.17 –0.24) were less likely to leave than junior soldiers. Predictably, men were less likely to leave than women (AOR 0.61, CI 0.48–0.77). Military service arm was relevant with respondents who served in the Navy (AOR 0.77, CI 0.62–0.97) and RAF (AOR 0.53, CI 0.44–0.64) less likely to leave than Army participants. PTSR caseness was associated with being a leaver (AOR 1.75, CI 1.38–2.22).
|
Early leavers (those who had served
4 years) were compared with those who had served for longer (data available from authors). [A cut off rate of 4 years was used as those who leave before this time historically were not eligible for any formal resettlement help.] Men were much less likely to leave early than women (AOR 0.21, CI 0.13–0.31). Higher levels of education tended to predict leaving (AOR 1.55, CI 1.04–2.30). Those in the RAF and Navy seemed to be less likely to leave early than their Army counterparts, but the absolute numbers here were extremely small. Finally it seems that those who served in the Gulf were more likely to have a shorter length of service (AOR 1.43, CI 1.04–1.96) than the Bosnia or Era cohort, even after adjusting for poor psychological health. This may simply reflect the fact that the Gulf War was earlier (in time) than the Bosnia conflict and therefore more soldiers had had the chance to leave by the time the sample was taken.
Predictors of leaving (table 2)
A longitudinal analysis was undertaken of those who left the military between phase 1 and phase 3 (n=392), compared to those who stayed in (n=1029). Poor mental health at phase 1 predicted leaving by phase 3 (AOR 1.67, CI 1.18–2.36). Again, NCO status (AOR 0.47, CI 0.30–0.73) and officer status (AOR 0.33, CI 0.18–0.61) was predictive of retention compared to private soldiers.
|
Associations and predictors of employment (tables 3 and 4)
Analysis of the phase 1 leavers for whom we had employment data (n=2792) revealed that 11.9% (n=333) of the leavers were unemployed. Male gender (AOR 2.44, CI 1.50–3.98), being married (AOR 2.0, CI 1.42–2.83) and NCO status (2.21, CI 1.53–3.19) were associated with employment by phase 3. Interestingly, being deployed to the Gulf was also associated with getting a job (AOR 2.23, CI 1.56–3.18). Those who fulfilled GHQ caseness (AOR 0.43, CI 0.30–0.60) and those who are more symptomatic (AOR 0.96, CI 0.94–0.98) are less likely to be employed. None of these factors except symptoms score (AOR 0.94, CI 0.89–0.98) emerged as significantly predictive when phase 1 data was analysed for leavers between phases 1 and 3 (table 5).
|
|
|
For early leavers, the protective effect of marriage was even more evident (AOR 2.95, CI 1.38–6.31). Also, the negative effects of a high GHQ (AOR 0.19, CI 0.08–0.43) and PTSR score (AOR 0.31, CI 0.11–0.94) appear to have been accentuated (table 3).
What happens to symptoms after leaving? (tables 5–7)
In order to see what happens to people's symptoms after they leave, we looked at the phase 3 data for those who had already left at phase 1 (n=926) (see figure 1). After adjusting for potential confounders (length of service, age, sex, military status, rank and cohort) and correcting for the phase 3 sampling frame (table 5), there was no significant change in mean PTSR score, or mean symptom score; but there was evidence for a slight improvement in GHQ [adjusted mean difference 0.37, CI 0.11–0.63 (corrected for sample stratification at phase 3)]. The pattern in ‘unwell individuals’ (those who were GHQ cases at phase 1) demonstrated much the same pattern (data available from authors).
When the leavers were further subdivided into early versus other (table 6), there was some evidence that early leavers fared slightly worse than others, in the domains of mean PTSR [mean increase 0.28 (CI –0.57, 0.01) versus 0.12 (CI 0.19, –0.05)] and mean symptom score [mean increase 0.76 (CI –2.2, 0.67) versus 0.64 (CI –0.92, –0.34)].
|
|
Discussion |
|---|
|
TopAbstractMETHODSResultsDiscussionReferences |
|---|
Two key findings emerge from the above results. The first is consistent with the findings of a majority of the international literature on veterans: most people do well when they leave the armed forces. The vast majority of our cohort who left between phases 1 and 3 were in full time employment (87.5%). It was a minority who fared badly. Although we did not look specifically at combat experiences, using Gulf service as a proxy measure there was also no robust evidence that deployment to conflict per se disadvantaged people
Individuals who served in the Gulf or in Bosnia were no more likely to leave the forces early, and deployment (including to the Gulf) did not adversely affect people's chances of being in employment after leaving. Indeed, the adjusted data shows that serving in the Gulf was, in fact, predictive of employment. This association with Gulf service and employment needs explaining. Going to the Gulf was associated with worse symptoms and poorer psychological health and therefore, overall, no significant advantage to gaining employment. However, when adjustment is made for the negative confounding effect of poor psychological health, a positive association is revealed. This is because there is ‘a healthy worker effect’ for active duty service,29 but also greater risk of ill health: a textbook example of negative confounding. It may also be that there is a ‘medal effect’ of a tour of duty; employers are impressed by combat military experience provided one comes home well. This is born out by the final analysis (table 8) which shows that well Gulf veterans were more likely to be employed than well veterans from the Era cohort who had not seen combat.
|
The second important finding is that psychological health is one indicator of whether a person is likely to stay in the military and, if they do not stay in, whether they will be in full time employment as a civilian. People who had left the military by phase 1 had higher GHQ scores than those who stayed in. In addition, a higher GHQ score at phase 1 was predictive of leaving at phase 3 follow-up. There was also evidence that higher GHQ scores and symptom burden were associated with unemployment after leaving. These findings are in agreement with the US literature; a recent paper19 demonstrated that 47% of individuals hospitalized with mental health problems left military service within 6 months
We hypothesized that if some people were unhappy or unsuited for military life, their mental health might improve when they leave. Overall, this was not the case—if anything those who left the forces early faired slightly worse. This is in keeping with our previous findings in Gulf veterans who have persistent symptomology.24
Alternatively, we wondered if those who had served for longer in the military would find it difficult to cope on separation, something anecdotally called the ‘The Military Retirement Syndrome’30 Again we found no evidence for this frequently advanced hypothesis. Instead, we found that symptomatic mental health remains fairly static after leaving. Those who are well, remain well, those who are symptomatic, remain symptomatic. This is very much in keeping with the literature, albeit largely American, on the chronic nature of occupational ill health such as PTSD.31 Rosenheck has repeatedly described clear associations of poor mental health with social exclusion (including homelessness) in US veterans.21 In the context of developing UK veterans policy, this data suggests that UK servicemen whose mental health is poor are likely to be the most vulnerable to social exclusion or hardship such as unemployment, and it seems that this risk factor may be a chronic one extending into a veteran's life, with little evidence of remittance of ill health after leaving.
Limitations of this study
This study is a cohort study of those who were serving in the armed forces at one point of time in 1991. As such it captures a ‘snapshot’ of the military at that time. The research findings may not be applicable to earlier or later military cohorts.
By necessity, any study which relies so heavily on retrospective report must be vulnerable to a degree of recall bias. In addition, the measures of employment are based entirely on self report; we have no means of independent corroboration of people's employment status.
Although the response rates were good for the study and we know that, overall, non-responders were similar to responders, it is less likely that the severely social excluded members (e.g. homeless) of the cohort would have made it to follow up at phase 3.
Key points
|
|
Acknowledgments |
|---|
The phase 1 study was funded by the US Department of Defence. The phase 3 study was funded by the UK Medical Research Council, with additional support from the UK Ministry of Defence. We thank Nick Blatchley and Simon Satchell from the Gulf Veterans Illness Unit of the Ministry of Defence and Nick Harris from DASA for assistance in tracking participants, and the participants for their patience in once again completing lengthy questionnaires.
|
References |
|---|
|
TopAbstractMETHODSResultsDiscussionReferences |
|---|
1 Arkell, A. Dying inside from wounds that you cannot see. London: The Times (Supplement) 12 November (2002) 3-8. Available at www.timesonline.co.uk/printFriendly/0,1-7-477293,00.html.
2 Gabriel R, et al. Gulf war syndrome may be post-conflict dysfunction. BMJ 2002;324:914.
3 Orner R, Lynch T, Seed P. Long-term traumatic stress reactions in British Falklands War veterans. Br J Clin Psychol 1993;32:457–9.[Medline]
4 Solomon Z. Immediate and long-term effects of traumatic combat stress among Israeli veterans of the Lebanon War. In: Wilson JP, Raphael B, editors. International Handbook of Traumatic Stress Syndromes. New York: Plenum Press, 1993:321–32.
5 Kulka, R. et al. Trauma and the Vietnam War Generation: Report of Findings From the National Vietnam Veterans Readjustment Study. Brunner-Mazel Psychosocial Stress Series, No. 18. New York: Brunner/Mazel, (1990).
6 Elder GH, Shananhan MJ, Clipp E. When War comes to men's lives: life-course patterns in family, work, and health. Psychology Aging 1994;9:5–16.[CrossRef]
7 Elder G, Clipp E. Wartime losses and social bonding: influences across 40 years in men's lives. Psychiatry 1988;51:177–98.[ISI][Medline]
8 Elder GH, Clipp EC. Combat experience and emotional health: impairment and resilience in later life. J Personality 1989;57:311–41.[Medline]
9 Maas I, Settersen R Jr. Military Service During Wartime. Effects on Men's Occupational Trajectories and Later Economic Wellbeing. Eur Sociolog Rev 1999;15:213–32.
10 Angrist J, Using the Draft Lottery to Measure the Effects of Military Service on Civilian Earnings. Res Labour Econ 1989, volume 10, edited by Ronald Ahrenberg, Greenwich, CT:JAI.
11 Angrist J, Krueger A. Lifetime Earnings and the Vietnam Era Draft Lottery: Evidence from Social Security Administrative Records. Am Econ Rev 1990;80:313–6.
12 Rosen S, Taubman P. Changes in Life Cycle Earnings: What do Social Security Data Show? J Human Res 1982;17:321–38.[CrossRef][ISI]
13 Anderson K, Mitchell J. Effects of Military Experience on Mental Health Problems and Work Behaviour. Med Care 1992;30:554–63.[CrossRef][ISI][Medline]
14 Savoca E, Rosenheck R. The civilian labor market experiences of Vietnam-era veterans: the influence of psychiatric disorders. J Ment Health Policy Econ 2000;3:199–207.[CrossRef][Medline]
15 Yager T, Laufer R, Gallops M. Some problems associated with war experience in men of the Vietnam generation. Arch Gen Psychiatry 1984;41:327–33.[Abstract]
16 Engel CC Jr, et al. Psychological conditions diagnosed among veterans seeking Department of Defense Care for Gulf War-related health concerns. J Occup Environ Med 1999;41:384–92.[CrossRef][ISI][Medline]
17 Prigerson HG, Maciejewski PK, Rosenheck RA. Combat trauma: trauma with highest risk of delayed onset and unresolved posttraumatic stress disorder symptoms, unemployment, and abuse among men. J Nerv Ment Dis 2001;189:99–108.[ISI][Medline]
18 Black DW, Carney CP, et al. Incarceration and Gulf War Veterans. Military Medicine (in press), (2004).
19 Hoge CW, et al. Mental disorders among U.S. military personnel in the 1990s: association with high levels of health care utilization and early military attrition. Am J Psychiatry 2002;159:1576–83.
20 Rosenheck R, et al. Initial assessment data from a 43-site program for homeless chronic mentally ill veterans. Hosp Community Psychiatry 1989;40:937–42.
21 Rosenheck R, Frisman L, Chung AM. The proportion of veterans among homeless men. Am J Public Health 1994;84:466–9.
22 Unwin C, et al. The health of United Kingdom Servicemen who served in the Persian Gulf War. Lancet 1999;353:169–78.[CrossRef][ISI][Medline]
23 Ismail K, et al. The mental health of UK Gulf war veterans: phase 2 of a two phase cohort study. BMJ 2002;325:576.
24 Hotopf M, et al. Gulf War Illness-better, worse, or just the same? BMJ 2003;327:1370–2.
25 Sharief MK, et al. Neurophysiologic analysis of neuromuscular symptoms in UK Gulf War veterans: a controlled study. Neurology 2002;59:1518–25.
26 Keane T, Caddell J, Taylor K. Mississippi Scale for combat-related posttraumatic stress disorder: three studies in reliability and validity. J Consult Clin Psychol 1988;56:85–90.[CrossRef][ISI][Medline]
27 Davidson JR, et al. Assessment of a new self-rating scale for post-traumatic stress disorder. Psychol Med 1997;27:153–60.[CrossRef][ISI][Medline]
28 Goldberg D. The detection of psychiatric illness by questionnaire. London: OUP, 1972.
29 Haley R. Point: Bias from the "Healthy Warrior Effect" and Unequal Follow up in Three Government Studies of Health Effects of the Gulf War. Am J Epidemiol 1998;148:315–23.
30 McNeil J, Gitten M. Military retirement: the retirement syndrome. Am J Psychiatry 1967;123:848–54.
31 Bremner J, et al. Dissociation and posttraumatic stress disorder in Vietnam combat veterans. Am J Psychiatry 1992;149:328–32.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||