Abstract
Objectives
Over 600 UK Armed Forces personnel and civilians were seriously injured in the conflict in Afghanistan. We examined whether combat injury, and limb loss specifically, reduced the likelihood of employment after leaving the UK Armed Forces and to what extent this was mediated by pain and mobility.
Methods
Combat-injured participants who were aeromedically evacuated to a UK hospital while on deployment to Afghanistan and subsequently left the UK Armed Forces (n=406) and a comparison group of uninjured personnel who had left the military (n=107) were drawn from an existing cohort (the ADVANCE study). Current employment was determined by self-report questionnaire, and pain and mobility mediating variables were taken from the EuroQol EQ-5D 5-Level measure.
Results
21.2% of the injured group were not in paid employment, compared with 14.3% of the uninjured comparison group; this difference was not statistically significant. Unmediated analyses showed that although those with amputation injuries had an increased risk of being unemployed compared with the comparison group, this was not statistically significant. For those who were injured without limb loss, compared with the comparison group, there were indirect effects on employment through both mobility (adjusted risk ratio (aRR) 1.32, 95% CI 1.11 to 1.69) and pain (aRR 1.10, 95% CI 1.01 to 1.31).
Conclusions
In this severely injured cohort, combat injury does not result in a significantly reduced rate of post-military employment. Any potential relationship between injury and employment for those injured without limb loss compared with the uninjured is due in part to the mediating effects of mobility and pain.
Keywords: Veterans, Military Personnel, Wounds and Injuries
WHAT IS ALREADY KNOWN ON THIS TOPIC
Advances in combat casualty medicine have resulted in more surviving serious combat injuries (including some leading to limb loss). Many will have to leave the military following injury; however, the consequences to their employment after leaving are under-researched in a UK context. This is of particular concern due to impacts on employment observed in general populations linked to pain and mobility restrictions.
WHAT THIS STUDY ADDS
This study examines how post-service employment is affected by combat injury. Additionally, it examines the impact of pain and mobility via mediation analysis of this main effect.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This study identified that those injured without limb loss, as opposed to those whose injuries result in amputation, may be at particular risk of unemployment related to ongoing pain and mobility issues and hence require additional support for the sequelae of their injuries.
Background
During the recent conflict in Afghanistan (2002–2014), a number of casualties in the two most serious categories defined by the UK military (‘seriously injured’ and ‘very seriously injured’) occurred; there were 616 (very) serious casualties among UK Armed Forces and civilian personnel;1 545 (very) seriously injured service personnel were aeromedically evacuated back to the UK for treatment from Afghanistan.2 Due to advances in combat casualty medicine, many survived injuries which, at any previous time, would have been fatal;3 some of these injuries resulted in limb amputation.4 As a result of these injuries, many personnel were not able to continue their military careers and consequently transitioned to civilian life.5 While there has been some research on UK personnel leaving the Armed Forces with health conditions,6 evidence regarding the transition outcomes of those who leave service following serious combat injury is sparse.7
In the general population, limb amputation is a risk factor for job loss; for example, in a Scottish civilian sample, employment rates for those with amputations were 75% prior to amputation but 44% afterward.8 Those who do return to work often have a change in duties, involving less physically demanding work9 and lower socioeconomic occupational classification;8 others retire while still relatively young.10 Among those who are not employed, the majority attribute their unemployment to their limb loss.9 11
This injury- and amputation-related impact on employment could be a consequence of several factors. One potential cause is pain; there is evidence in the general population that chronic pain can affect occupational status, giving rise to job loss,12 reducing hours of paid work and increasing the likelihood of early retirement.13 Pain is itself a frequent consequence of limb loss; a study of those with amputations following the Vietnam war found that they reported higher levels of bodily pain than those without amputations;14 by contrast, the sample of UK Armed Forces personnel with combat injuries used in this study found higher levels of pain in those injured without limb loss compared with those with limb loss.15 Around 17% of those with lower-limb traumatic amputations in a US military sample reported pain sufficient to interfere with daily activities;16 in the sample used in this study, 19% of those with amputations reported moderate to severe pain.15 Evidence regarding the relationship between pain and employment for those with amputations within military populations is mixed; one study found that pain was significantly negatively associated with employment,17 while another found that phantom pain negatively affected likelihood of being in employment although other sources of pain did not.8
Mobility also potentially impacts employment. Physical functioning following major limb trauma predicted subsequent return to work in a civilian population,18 and shorter walking distance and restrictions in mobility were associated with unsuccessful employment reintegration among those with amputations.19
The Armed Services Trauma Rehabilitation Outcome (ADVANCE) study comprises a cohort of current and former UK Armed Forces personnel who were injured in combat and aeromedically evacuated to the UK and a frequency-matched comparison group of demographically similar personnel who sustained no such injuries.20 The injured group contains both those with limb loss and those without. Using this cohort, this analysis will investigate post-service employment and how associations between combat injury (with or without limb loss) and post-service employment are mediated by pain and mobility. This will be fulfilled utilising counterfactual mediation analysis to identify potential mediators and help understand the mechanisms of the exposure-outcome relationship, that is, how much of the relationship between injury and unemployment is through pain/mobility.
This paper aims:
To report rates of employment after leaving the UK Armed Forces for those with combat injuries from the recent conflict in Afghanistan and compare these rates to a demographically similar uninjured group.
To determine whether pain or mobility mediates any relationship between combat injury, limb loss and post-service employment.
Specific hypotheses to be tested are:
Sustaining a combat injury, and in particular limb loss, will be associated with lower rates of post-service employment compared with uninjured personnel.
The association between combat injury and post-service employment will be mediated by pain or mobility.
Methods
Participants
Injured personnel were recruited from UK Armed Forces Servicemen who were deployed to Afghanistan between 2003 and 2014 and suffered physical combat injuries20 of sufficient severity to require aeromedical evacuation to a UK hospital. Participants were identified from records provided by the Ministry of Defence, Defence Statistics (UK). Those with a prior history of cardiovascular, liver or renal disease were excluded (as the initial aim of the primary study was to investigate cardiovascular consequences of combat injury). All participants were male due to small numbers of serious physical injuries among female service personnel.
The comparison group consisted of male UK Armed Forces personnel who were frequency-matched to the injured cohort based on their age, rank, regiment, role on deployment and deployment period. There were only a very small number of female UK military combat casualties in Afghanistan, primarily because they were unable to take on front line duties until 2014, after the main conflict phase in Afghanistan. This small number, with the physiological differences between males and females, would confound the primary hypotheses of the study; hence, only male UK injured personnel were eligible for this cohort study. Eligibility criteria for the uninjured group were: deployed to Afghanistan in a combat role but did not sustain serious physical combat injuries while on deployment and no history of cardiovascular, liver or renal disease prior to deployment.
Participants were recruited through postal, email and telephone invitations. For those who had left the military, efforts to trace them were made through electoral roll data. 2329 participants (1163 injured and 1166 uninjured) were invited to attend study recruitment appointments. Adjusted response rates (excluding those who had died or had no available contact details) were 59.6% and 56.3% for the injured and the uninjured groups, respectively.21 579 injured participants and 566 uninjured participants were successfully recruited. A clinical interview with a research nurse was part of data collection, which included sociodemographic information and self-completed participant questionnaires enquiring into pain, musculoskeletal functioning and occupational history.
For these analyses, participants had to have left the military in order to be eligible for employment in the civilian sector, thus excluding 610 of the original sample who were still serving in the UK Armed Forces. A further 22 were removed from these analyses as they indicated they were retired as they were no longer in the labour market, and one individual was excluded as they had sustained a serious injury outside of military service but were in the uninjured comparison group. 39 individuals were excluded as they had served as reservists rather than regulars and hence had been part of the civilian work force before leaving the Armed Forces. A single participant was excluded as their current employment status could not be ascertained. Consequently, the sample comprised 472 of the 1145 individuals from the original sample (41.3%), 385/579 of the overall injured group (66.5%) and 87/565 of the uninjured group (15.4%). The injured group included 128 of 161 who sustained an amputation injury in the original sample (79.5%) and 257 of 418 who sustained a non-amputation injury (61.5%).
Measures
Demographics
Sociodemographic information and military histories were collected via self-report questionnaires and clinical interviews. Data collected from participants was supplemented by information provided by Defence Statistics (UK). Rank was determined from data at time of sampling, coded as a proxy for socioeconomic status and categorised as Junior Non-Commissioned Officer/Other rank (NATO OR2-OR4), Senior Non-Commissioned Officer rank (NATO OR5-OR9) and Officer rank (NATO OF1-OF6).22 The service arm was divided into Naval Services (including Royal Marines), Army and Royal Air Force.
Combat injury
‘Limb loss’ was defined as having an amputation above, below or through the knee for one or more lower limbs or above or below the elbow for one or more upper limbs. Isolated partial amputations (eg, partial foot, partial hand, finger or toe) were not included in the limb loss group, but instead placed in the injured without limb loss group.
Pain/mobility
Measures of pain and mobility were both taken from the five-level EuroQol EQ-5D questionnaire (EQ-5D-5L), a standardised measure of health status,23 24 which asks respondents to endorse the response which best described their health that day.
Pain responses were ‘I have no pain or discomfort’, ‘I have slight pain or discomfort’, ‘I have moderate pain or discomfort’, ‘I have severe pain or discomfort’ and ‘I have extreme pain or discomfort’. These were combined into two categories: no/slight discomfort and moderate-extreme discomfort. Mobility responses were ‘I have no problems in walking about’, ‘I have slight problems in walking about’, ‘I have moderate problems in walking about’, ‘I have severe problems walking about’ and ‘I am unable to walk about’. These were combined into two categories: no problems and slight problems—unable to walk about. Both variables were reduced to two categories due to small numbers in the more affected categories (n <10). Due to the distribution of these variables in this sample, we used ‘no problems’ compared with any problems for mobility but no/light pain compared with moderate or worse pain when categorising the pain measure due to small numbers.
Employment
Paid employment was the main outcome of interest examined. Employment status was self-reported at time of assessment: participants were asked the stem question ‘Are you currently:’, with possible responses ‘still serving, full time in the armed forces’, ‘still serving, medically downgraded’, ‘still serving, not active due to ill-health’, ‘working full/part time in civilian paid employment’, ‘self-employed’, ‘volunteering’, ‘not working due to ill health’, ‘not working by choice’, ‘not working, seeking employment’ and ‘retired’. Those who were still serving or endorsing ‘retired’ were excluded from the study, as stated previously. Participants were also asked for their employment history since leaving the military, which was used to fill any missing data. To most closely replicate the approach of other occupational health studies regarding return-to-work post-injury, those who were in paid work were categorised as ‘employed’; and all other responses (seeking work, unable to work due to ill health, in full-time education, etc) were categorised as ‘not economically active’.
Analysis
Data were analysed using the statistical software package Stata version 17.0.25 Missing values for variables included in these analyses were uncommon and less than 1.5% for any variable (n=0 for injury status, n=1 (0.2%) for employment status, n=6 (1.3%) for mobility and n=4 (0.8%) for pain); hence, the main analysis was performed on a complete case basis. Statistical significance for differences in employment levels between groups was determined by binomial regression with a log link and risk ratios (RR) are reported. Statistical significance was defined as a p value less than 0.05.
Counterfactual mediation methods were used to assess the mediating effects of pain and mobility on the association between combat injury and not being in paid employment.26 The theoretical model can be seen in figure 1. Combat injury was modelled as injury without limb loss and injury with limb loss separately with uninjured as the reference group.
Figure 1.
Theoretical mediation model. Reference group is uninjured comparison group. Confounder sets are identical for the exposure-mediator and mediator-outcome relationships.
In a logit model when the outcome is common, decomposing the indirect effects using the product method can produce biased coefficients;27 therefore, the indirect effect was decomposed using the paramed 28 command in Stata with a binomial distribution and a log link for the outcome model and logit link for the mediation model. An exposure-mediator interaction was included.26 The natural indirect effect relates to the portion of the total effect that is mediated by pain or mobility, and the natural direct effect is the portion not explained by the mediator. Conditional Direct Effects (CDE) are also reported, defined as the effect of the exposure on an outcome that is observed if the mediator were fixed at a certain level and is the effect due to neither mediation nor interaction. CDE, total, natural direct and natural indirect effects are reported as RR and their 95% CIs. Bias-corrected bootstrap confidence intervals using 1000 replications are reported. All analyses were a priori adjusted for rank and age at time of injury/deployment. A sensitivity analysis was performed to assess the effect of any unmeasured mediator-outcome confounding. Vanderwheele et al proposed an E-value which is a measure of the minimum strength of association that an unmeasured confounder would need to have to fully explain the exposure-outcome association. This E-value is reported on the RR scale.29 30
Results
The sample was all male and was primarily aged 25–34 years at time of assessment (63%). (Time of assessment was used rather than time of injury, as the uninjured group has no time of injury.) At the time of their discharge, the majority were of lower rank (OR2-OR4) (79%). Most were in paid employment (79.2%) (table 1). For a description of the sample by injury group, please see online supplemental material.
Table 1.
Descriptive statistics for the total sample and by employment status, number (n) and percentages (%) are reported. Total sample column has column percentages; employed and not in employment are row percentages
| Variable | Category | Total sample, n=472 (n (%)) |
Employed, n=374 (79.2%) (n (%)) |
Not in employment, n=98 (20.8%) (n (%)) |
| Age at assessment (years) (mean (SD)) | 33.8 (5.3) | 34.0 (5.4) | 33.2 (4.8) | |
| Time since injury (injured group only) (years) (mean (SD)) | 8.4 (2.2) | 8.4 (2.3) | 8.2 (1.8) | |
| Rank; n (%) | Junior Non-Commissioned Officer/Other rank | 372 (78.8) | 284 (76.3) | 88 (23.7) |
| Senior Non-Commissioned Officer | 57 (12.1) | 50 (87.7) | 7 (12.3) | |
| Officer | 43 (9.1) | (>90)* | <10 (<10)* | |
| EQ-5D-5L mobility; n (%) | No problems | 271 (58.2) | 230 (84.9) | 41 (15.1) |
| Slight problems-unable to walk about | 195 (41.9) | 142 (72.8) | 53 (27.2) | |
| EQ-5D-5L pain/discomfort; n (%) | No/slight pain or discomfort | 332 (70.9) | 269 (81.0) | 63 (19.0) |
| Moderate-extreme pain or discomfort | 136 (29.1) | 102 (75.0) | 34 (25.0) | |
| Injury group; n (%) | Comparison group | 87 (18.4) | 74 (85.1) | 13 (14.9) |
| Injured group | 385 (81.6) | 300 (77.9) | 85 (22.1) | |
| Injured with limb loss | 128 (27.1) | 93 (72.7) | 35 (27.3) | |
| Injured without limb loss | 257 (54.5) | 207 (80.5) | 50 (19.5) |
*Data obscured due to risk of identifiability.
EQ-5D-5L, five-level EuroQol EQ-5D questionnaire.
oemed-82-11-s001.pdf (91KB, pdf)
Overall, the estimated risk of unemployment was higher among those men with a combat injury but the association was not statistically significant (adjusted RR (aRR) 1.37, 95% CI 0.81 to 2.33). In the subgroup analysis, the association was stronger in those with limb loss compared with the comparison group (aRR 1.69, 95% CI 0.95 to 3.00) than in those injured without limb loss (aRR ratio 1.19, 95% CI 0.68 to 2.07), but the association was not statistically significant in either subgroup.
Mobility mediated the effect of injury on employment in those injured without limb loss compared with the comparison group (aRR 1.32 (1.11 to 1.69), as did pain (RR 1.10, 95%CI 1.01 to 1.31) (table 2). In participants without limb loss (setting the mobility mediator to 0), those who had no mobility problems had a CDE of 0.54 (0.26, 1.13) compared with the comparison group. For pain, those with no/slight discomfort had a CDE of 1.06 (0.52, 2.15), compared with the comparison group. In the group with limb loss, there was no indirect effect through mobility or pain.
Table 2.
The effect of combat injury on unemployment overall (total), through mediators (natural indirect effect) and independent of mediators (natural direct effect)
| N (%) | Total effect aRR (95% CI) | Natural direct effect aRR (95% CI) | Natural indirect effect aRR (95% CI) | |
| Injured without limb loss (reference group: uninjured comparison group) | ||||
| Mobility mediator | 340 | 1.16 (0.64 to 2.26) | 0.88 (0.48 to 1.86) | 1.32 (1.11 to 1.69) |
| Pain mediator | 340 | 1.18 (0.51 to 2.38) | 1.08 (0.45 to 2.16) | 1.10 (1.01 to 1.31) |
| Injured with limb loss (reference group: uninjured comparison group) | ||||
| Mobility mediator | 213 | 1.67 (0.86 to 3.43) | 1.68 (0.81 to 3.92) | 0.99 (0.78 to 1.31) |
| Pain mediator | 213 | 1.71 (0.56 to 3.89) | 1.72 (0.57 to 3.86) | 0.99 (0.91 to 1.03) |
aRR; all models controlled for age at assessment and rank (as a proxy for socioeconomic status) at time of injury/deployment. Indirect effect: the portion of the total effect that is mediated by pain or mobility. Direct effect: the portion not explained by the mediator.
aRR, adjusted risk ratio.
Sensitivity analysis to determine what effect size would be required for a confounder to fully explain the indirect effect observed in those injured without limb loss found that it would require a confounder to have a RR of 1.97 (95%CI 1.45 to 2.77) on both not working and the mobility mediator to explain the observed indirect effect. For the pain mediator, the RR of such a confounder would have to be 1.43 (95% CI 1.11 to 1.95).
Discussion
Armed Forces personnel may leave service for a number of reasons, including end of service term, voluntary release or medical discharge. Personnel who are seriously injured in service will often have to leave as a consequence of their injuries, either because the injury itself prevents continuance of service or because ongoing sequelae of injury result in medical downgrading. In the majority of cases, former service personnel will expect to embark on a civilian career subsequent to leaving the Armed Forces. We examined whether combat injury resulted in a lower likelihood of being in paid employment and whether or not such a relationship might be mediated by pain or mobility. These analyses show that our first hypothesis was not supported as, overall, sustaining a physical combat injury was not significantly associated with employment outcomes after leaving the UK Armed Forces. Our second hypothesis, that the relationship between combat injury and post-service employment will be mediated by pain or mobility, is partly supported in that there was evidence of mediation by pain and mobility in those injured without limb loss.
Type of injury provides context to this relationship. Similar to general population findings,9 those who sustained limb loss may have a higher likelihood of not being economically active; while this relationship fell short of statistical significance, this finding should be considered in the context of the RR, CI and limited sample size, and hence there may be a true effect that is not statistically significant in this study. This relationship was not mediated by either mobility or pain, suggesting that the reasons for any reduction in employment are the consequence of other factors not examined here. These might include lack of information regarding appropriate employment for those with limb loss or lack of understanding from employers in the civilian sphere. The importance of mobility on finding employment after leaving the Armed Forces following combat injury is a focus for rehabilitation for those with amputations; thus, lack of impact of pain or mobility may be partly due to effective treatment.31 Nonetheless, the effect size was not large; this may be because those factors positively associated with return to work in US industrial workers who had received amputations as a result of injury (including younger age, prosthetic use and vocational training) generally apply to the population included here (who largely had their injuries while young, received high-quality prosthetics and have extensive rehabilitative and resettlement support).32 Thus, this study does not support earlier findings from a study of a US military population that, for those with limb loss, pain is significantly associated with employment status.17
There was no overall association between injury without amputation and not being employed, but nonetheless there were significant indirect effects of both mobility and, to a lesser extent, pain. In most cases where there is no total effect but indirect effects, inconsistent mediation is likely, which may be the case for mobility. However, for pain, the direct and indirect effects are both positive, and therefore a further explanation could be that the analysis lacked statistical power to detect the total effect.33 Sensitivity analysis demonstrated that substantial unmeasured confounders (particularly for mobility) would be required to explain the observed indirect effects. The existence of indirect effects of combat injury through these factors for those without limb loss, but not for those with limb loss, suggests that there may be unmet needs for this subgroup with regard to pain and mobility; hence additional, ongoing health surveillance (ie, active checking and reassessment of health status) is recommended for this subgroup.
Limitations
Participants were drawn from those who were seriously injured in combat and aeromedically evacuated from the conflict in Afghanistan and a matched uninjured comparison group. Therefore, the participants in this study are not necessarily representative of the general injured population of the UK Armed Forces, nor other populations with serious injuries or limb loss.
Due to a small number of participants with upper limb loss, most of whom also had lower-limb loss; these analyses did not discriminate between upper and lower limb loss. The mobility measure used here focuses on ability to walk; limited mobility in upper limbs is unmeasured in our approach, but this is unlikely to affect the findings in this sample as few (<5) individuals had upper limb loss alone; in this study population, men with lower limb loss only were found to have increased odds of having more upper limb disability than the uninjured comparison group.34
Due to the use of the paramed Stata package, mediators could not be modelled together to estimate separate effects; since there was no indirect effect of pain, it is not anticipated that a combined model would lead to different findings regarding the mediators. Mediation analysis was performed irrespective of a lack of direct effect of the exposure on the outcome, following suggestions that a significant association between the exposure and outcome in the absence of mediators is not required to establish mediation.35 It should also be noted that the test of the indirect effect requires stronger assumptions than the total effect.33 One strong assumption of estimating natural indirect and direct effects is that there be no exposure induced mediator-outcome confounding. Adjusting for these variables introduces bias; however, equally, not adjusting for these may also result in some bias and non-identifiability of the indirect and direct estimates. The relationships between some of the measured variables are complex and at times bidirectional. Having considered the theory of potential mediator-outcome confounders, we did not find any relevant variables. There is also the potential for unmeasured confounding of the mediator-exposure relationship. While this may not bias the total effect, it may bias the size of the indirect effect. We hope that we have addressed this limitation with a sensitivity analysis which suggested that any unmeasured confounder would need to have had a twofold effect on the mobility mediator and the outcome to explain away the effect observed. The strength of an unmeasured confounder in the pain-mediated model would need to be much smaller in order for the indirect effect to disappear and therefore the results of this model should be interpreted with caution. Overall, it may be that the complex analysis contributed to power issues, resulting in a number of relationships falling short of significance.
This study only examined whether participants were in paid employment, not the quality or sustainability of such employment. It is possible that studies into other aspects of employment—for example, full- vs part-time, remuneration or job satisfaction—will reveal different relationships with pain, mobility and other consequences of combat injury. In the general population, those with amputations generally demonstrate high job satisfaction,36 37 but this may differ for a military population who transition to civilian employment as a consequence of their injuries. Furthermore, this analysis included only measures of pain, mobility and unemployment at a single point in time, on average 8 years from the time of injury; future longitudinal analysis may help elucidate lagged effects of amputation, pain and mobility on employment and may also help determine whether unemployment itself affects subsequent perception of mobility or pain.
Conclusions
The association between combat injury and post-service employment was not statistically significant overall. A potential factor affecting post-service employment for those who leave the UK Armed Forces after combat injury is mobility, while pain may have a smaller effect. However, these relationships do not apply to those injured with limb loss. Further dimensions of employment, such as sustainability and job strain, should be examined to gain a wider understanding of the impact of combat injury and limb loss.
Acknowledgments
The ADVANCE study is funded through the ADVANCE Charity. Key contributors to this charity are the Headley Court Charity (principal funder); HM Treasury (LIBOR grant); Help for Heroes; Nuffield Trust for the Forces of the Crown; Forces in Mind Trust; National Lottery Community Fund; Blesma, The Limbless Veterans; and the UK Ministry of Defence. We wish to thank all of the research staff at both Headley Court and Stanford Hall who helped with the ADVANCE study, including Eleanor Miller, Stefan Sprinckmoller, Maria-Benedicta Edwards, Helen Blackman, Melanie Chesnokov, Emma Coady, Sarah Evans, Guy Fraser, Severija Juskaite, Meliha Kaya-Barge, Maija Maskuniitty, David Pernet, Helen Prentice, Urszula Pucilowska, Lajli Varsani, Anna Verey, Molly Waldron, Danny Weston, Tass White, Seamus Wilson, Louise Young, etc.
Footnotes
Contributors: Conception and design of underlying study: SS, ABe, CB, ABu, NTF. Literature review: HB. Data analysis and interpretation: HB, SS, DMD. Drafting of the manuscript, guarantor: HB. Drafting of the manuscript (analytical methods): SS. Critical revision: DMD, ABe, CB, ABu, NTF.
Funding: The ADVANCE study is funded through the ADVANCE Charity. Key contributors to this charity are the Headley Court Charity (principal funder); HM Treasury (LIBOR grant); Help for Heroes; Nuffield Trust for the Forces of the Crown; Forces in Mind Trust; National Lottery Community Fund; Blesma, The Limbless Veterans; and the UK Ministry of Defence.
Competing interests: HB’s work is supported by a grant from the ADVANCE Charity and Forces in Mind Trust. SS’s role is funded by the ADVANCE charity. ABu reports founder shares of university spinouts: Smart Surgical Solutions, Biomex and Go Assistive Technology. NTF’s work is supported by a grant from the ADVANCE Charity and Forces in Mind Trust, and salary is part covered by a grant from the UK Ministry of Defence. She is an unpaid trustee of Help for Heroes (a UK-based charity supporting the health and wellbeing of the Armed Forces Community). Travel and accommodation have been covered by Gallipoli Medical Research (based in Brisbane, Australia) to enable her attendance and presentation at a military medical conference which was being held in Brisbane. She also reports grants to support research from the Medical Research Council, Colt Foundation and Scar Free Foundation. The rest of the authors have no competing interest.
Provenance and peer review: Not commissioned; externally peer-reviewed.
Supplemental material: This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
Data availability statement
Data are available upon reasonable request. Data are not publicly available due to high security regarding the participants; however, access to anonymised data may be possible on reasonable request, in which case please contact the main study (https://www.advancestudydmrc.org.uk)
Ethics statements
Patient consent for publication
Not applicable.
Ethics approval
This study involved human participants and was approved by the UK Ministry of Defence Research Ethics Committee (MODREC; protocol no: 357/PPE/12). Participants gave informed consent to participate in the study before taking part.
References
- 1. Dempsey N. Research briefing: Afghanistan statistics: UK deaths, casualties, mission costs and refugees. London, UK: House of Commons Library, 2021. [Google Scholar]
- 2. Ministry of Defence . Annual tracking of Afghanistan very seriously injured/seriously injured UK operational casualties 8 October 2007-30 September 2015. 2007.
- 3. Penn-Barwell JG, Roberts SAG, Midwinter MJ, et al. Improved survival in UK combat casualties from Iraq and Afghanistan: 2003-2012. J Trauma Acute Care Surg 2015;78:1014–20. 10.1097/TA.0000000000000580 [DOI] [PubMed] [Google Scholar]
- 4. Chandler H, MacLeod K, Penn-Barwell JG, et al. Extremity injuries sustained by the UK military in the Iraq and Afghanistan conflicts: 2003-2014. Injury 2017;48:1439–43. 10.1016/j.injury.2017.05.022 [DOI] [PubMed] [Google Scholar]
- 5. Ministry of Defence . Annual medical discharges in the UK regular armed forces: 1 April 2017 to 31 March 2022. London, UK, 2017. [Google Scholar]
- 6. Hynes C, Scullion L, Lawler C, et al. The impact of in-service physical injury or illness on the mental health of military veterans. BMJ Mil Health 2023;169:e51–4. 10.1136/bmjmilitary-2020-001759 [DOI] [PubMed] [Google Scholar]
- 7. Burdett H, Verey A, Fear NT. Overview of literature on transition outcomes of battlefield injuries. King’s Centre for Military Health Research; 2023. [Google Scholar]
- 8. Whyte AS, Carroll LJ. A preliminary examination of the relationship between employment, pain and disability in an amputee population. Disabil Rehabil 2002;24:462–70. 10.1080/09638280110105213 [DOI] [PubMed] [Google Scholar]
- 9. Schoppen T, Boonstra A, Groothoff JW, et al. Employment status, job characteristics, and work-related health experience of people with a lower limb amputation in The Netherlands. Arch Phys Med Rehabil 2001;82:239–45. 10.1053/apmr.2001.18231 [DOI] [PubMed] [Google Scholar]
- 10. Armstrong AJ, Hawley CE, Darter B, et al. Operation Enduring Freedom and Operation Iraqi Freedom Veterans with amputation: An exploration of resilience, employment and individual characteristics. J Vocat Rehabil 2018;48:167–75. 10.3233/jvr-180923 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Sinha R, van den Heuvel WJA, Arokiasamy P. Factors affecting quality of life in lower limb amputees. Prosthet Orthot Int 2011;35:90–6. 10.1177/0309364610397087 [DOI] [PubMed] [Google Scholar]
- 12. Piper A, Blanchflower DG, Bryson A. Is pain associated with subsequent job loss? A panel study for Germany. Kyklos (Oxford) 2023;76:141–58. 10.1111/kykl.12319 [DOI] [Google Scholar]
- 13. Gerdle B, Björk J, Henriksson C, et al. Prevalence of current and chronic pain and their influences upon work and healthcare-seeking: a population study. J Rheumatol 2004;31:1399–406. [PubMed] [Google Scholar]
- 14. Dougherty PJ. Long-term follow-up of unilateral transfemoral amputees from the Vietnam war. J Trauma 2003;54:718–23. 10.1097/01.TA.0000046260.16866.A9 [DOI] [PubMed] [Google Scholar]
- 15. Vollert J, Kumar A, Coady EC, et al. Pain after combat injury in male UK military personnel deployed to Afghanistan. Br J Anaesth 2024;132:1285–92. 10.1016/j.bja.2024.02.019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Frisch HM, Andersen CRC, Mazurek CMT, et al. The Military Extremity Trauma Amputation/Limb Salvage (METALS) study: outcomes of amputation versus limb salvage following major lower-extremity trauma. J Bone Joint Surg Am 2013;95:138–45. 10.2106/JBJS.K.00734 [DOI] [PubMed] [Google Scholar]
- 17. Hawley CE, Armstrong AJ, Darter BJ, et al. Veterans of Operation Enduring Freedom and Operation Iraqi Freedom: Employment Status Following Traumatic Amputation. J Appl Rehabil Couns 2022;53:68–80. 10.1891/jarc-d-20-00027 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. MacKenzie EJ, Bosse MJ, Kellam JF, et al. Early Predictors of Long-Term Work Disability After Major Limb Trauma. J Trauma 2006;61:688–94. 10.1097/01.ta.0000195985.56153.68 [DOI] [PubMed] [Google Scholar]
- 19. Schoppen T, Boonstra A, Groothoff JW, et al. Factors related to successful job reintegration of people with a lower limb amputation. Arch Phys Med Rehabil 2001;82:1425–31. 10.1053/apmr.2001.26074 [DOI] [PubMed] [Google Scholar]
- 20. Bennett AN, Dyball DM, Boos CJ, et al. Study protocol for a prospective, longitudinal cohort study investigating the medical and psychosocial outcomes of UK combat casualties from the Afghanistan war: the ADVANCE Study. BMJ Open 2020;10:e037850. 10.1136/bmjopen-2020-037850 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Boos CJ, Schofield S, Cullinan P, et al. Association between combat-related traumatic injury and cardiovascular risk. Heart 2022;108:367–74. 10.1136/heartjnl-2021-320296 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Yoong SY, Miles D, McKinney PA, et al. A method of assigning socio-economic status classification to British Armed Forces personnel. J R Army Med Corps 1999;145:140–2. 10.1136/jramc-145-03-07 [DOI] [PubMed] [Google Scholar]
- 23. Feng Y-S, Kohlmann T, Janssen MF, et al. Psychometric properties of the EQ-5D-5L: a systematic review of the literature. Qual Life Res 2021;30:647–73. 10.1007/s11136-020-02688-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Herdman M, Gudex C, Lloyd A, et al. Development and preliminary testing of the new five-level version of EQ-5D (EQ-5D-5L). Qual Life Res 2011;20:1727–36. 10.1007/s11136-011-9903-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. StataCorp . Stata statistical software: release 17. College Station, TX: StataCorp LLC, 2021. [Google Scholar]
- 26. Valeri L, Vanderweele TJ. Mediation analysis allowing for exposure-mediator interactions and causal interpretation: theoretical assumptions and implementation with SAS and SPSS macros. Psychol Methods 2013;18:137–50. 10.1037/a0031034 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Rijnhart JJM, Valente MJ, Smyth HL, et al. Statistical Mediation Analysis for Models with a Binary Mediator and a Binary Outcome: the Differences Between Causal and Traditional Mediation Analysis. Prev Sci 2023;24:408–18. 10.1007/s11121-021-01308-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Emsley R, Liu H. PARAMED: stata module to perform causal mediation analysis using parametric regression models. Statistical Software Components: Boston College Department of Economics, 2013. [Google Scholar]
- 29. VanderWeele TJ. Mediation Analysis: A Practitioner’s Guide. Annu Rev Public Health 2016;37:17–32. 10.1146/annurev-publhealth-032315-021402 [DOI] [PubMed] [Google Scholar]
- 30. Ding P, Vanderweele TJ. Sharp sensitivity bounds for mediation under unmeasured mediator-outcome confounding. Biometrika 2016;103:483–90. 10.1093/biomet/asw012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Vascular Society of Great Britain and Ireland . Quality improvement framework for amputation. 2016.
- 32. Millstein S, Bain D, Hunter GA. A review of employment patterns of industrial amputees--factors influencing rehabilitation. Prosthet Orthot Int 1985;9:69–78. 10.3109/03093648509164708 [DOI] [PubMed] [Google Scholar]
- 33. Loeys T, Moerkerke B, Vansteelandt S. A cautionary note on the power of the test for the indirect effect in mediation analysis. Front Psychol 2014;5:1549. 10.3389/fpsyg.2014.01549 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Watson FCE, Kedgley AE, Schofield S, et al. Upper Limb Function in People With Upper and Lower Limb Loss 8 Years Postinjury: The Armed Services Trauma Outcome Study (ADVANCE) Cohort Study. Phys Ther 2024;104:pzae082. 10.1093/ptj/pzae082 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Zhao X, Lynch JG, Chen Q. Reconsidering Baron and Kenny: Myths and Truths about Mediation Analysis. J Consum Res 2010;37:197–206. 10.1086/651257 [DOI] [Google Scholar]
- 36. Schoppen T, Boonstra A, Groothoff JW, et al. Job satisfaction and health experience of people with a lower-limb amputation in comparison with healthy colleagues. Arch Phys Med Rehabil 2002;83:628–34. 10.1053/apmr.2002.32473 [DOI] [PubMed] [Google Scholar]
- 37. van der Sluis CK, Hartman PP, Schoppen T, et al. Job adjustments, job satisfaction and health experience in upper and lower limb amputees. Prosthet Orthot Int 2009;33:41–51. 10.1080/03093640802555917 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
oemed-82-11-s001.pdf (91KB, pdf)
Data Availability Statement
Data are available upon reasonable request. Data are not publicly available due to high security regarding the participants; however, access to anonymised data may be possible on reasonable request, in which case please contact the main study (https://www.advancestudydmrc.org.uk)