Factors associated with well-being among treatment-seeking UK Veterans: A cross-sectional study

Abstract

Introduction:

Comorbidity is the norm, rather than the exception, among UK treatment-seeking armed forces Veterans, who face a greater risk of experiencing mental health difficulties than the general UK population. Veterans may also experience poor overall well-being across various domains, including work-related challenges, loneliness, and instability in relationships. This study aimed to explore associations between mental ill health and mental well-being to add to the literature on the complexity of functioning and mental health presentations in clinical Veteran samples.

Methods:

A total of 428 UK Armed Forces Veterans (mean age 50.5 [SD 10.9] y) who received treatment for mental health difficulties at a UK Veteran mental health charity completed a survey that collected responses on the Warwick-Edinburgh Mental Well-being Scale measure of mental well-being, as well as a range of mental health and functioning outcomes.

Results:

Linear regression analysis revealed that well-being was significantly negatively associated with symptoms of anxiety and depression, symptoms of physical health difficulties, possible problems with anger, moral injury, and symptoms of posttraumatic stress disorder (PTSD) and complex PTSD. Functioning outcomes related to well-being were alcohol misuse, sleep problems, and loneliness. Alcohol misuse, sleep problems, and symptoms of loneliness were associated with lower well-being scores.

Discussion:

Implications of this pattern of relationships are discussed, including the potential benefits of a transdiagnostic and sub-clinical outcome measure for interventions targeting mental ill health, as well as positive mental health, among Veterans.

Lay Summary

Veterans seeking treatment for mental health problems are likely to face more than one problem. Well-being is a concept that does not rely on any mental health diagnosis or difficulty and may offer a broader way to capture mental health. In this study, 428 UK Armed Forces Veterans seeking treatment for mental health difficulties at a UK-based charity completed a survey asking about mental health and well-being. Overall, treatment-seeking Veteran well-being was associated with symptoms of anxiety and depression, physical health, problems with anger, moral injury, posttraumatic stress disorder (PTSD), and complex PTSD. Sleep problems, alcohol misuse, and loneliness were also linked with well-being scores. These associations suggest that well-being may capture Veteran health in a way that is separate from any single diagnosis, meaning outcomes for Veterans with complex needs who require a range of support can be compared. These results also highlight the relevance of positive mental health, as opposed to mental ill health.

Introduction

UK Veterans face an increased risk of mental ill health compared with the civilian population, with a higher prevalence of posttraumatic stress disorder (PTSD; 8% vs. 5%), symptoms of common mental health difficulties, such as anxiety and depression (23% vs. 16%), and alcohol misuse (11% vs. 6%).¹ In recent studies of UK Veterans seeking treatment for mental health difficulties, these rates have been higher, with 71% meeting probable caseness for PTSD and 43% meeting criteria for alcohol misuse.² These difficulties may arise after potentially traumatic experiences during combat or other unique aspects of military culture,³ such as the transition to civilian life after leaving the armed forces, which Veterans have described as difficult.⁴

Positive mental health

Comorbidity is the norm rather than the exception among Veterans,⁵ and approaches to understanding Veteran mental health are increasingly recognizing the complexity of presentations in this group, for example, via network analysis.⁶ As such, it may be clinically useful to capture Veteran mental health in a broader way. Well-being offers a way of comparing interventions across disorders and an alternative metric to symptomatic reduction,⁷ yet it is rarely collected as a routine outcome measure in populations seeking treatment for mental ill health.

Well-being is more than the absence of mental ill health. Several conceptualizations of this complex construct exist, incorporating life domains such as functioning in relationships, work, and hobbies and spiritual concepts such as fulfilment and self-realization.⁸ The World Health Organization defines positive mental health, or well-being, as encompassing coping with the normal stresses of life, productive work, contribution to community, and satisfying relationships.⁹ Although hedonic well-being primarily refers to affect and the subjective absence of negative feelings, eudaimonic well-being includes life satisfaction factors relevant to life fulfilment, such as functioning across life domains.⁸ Recognizing the multifarious nature of the well-being construct, researchers have highlighted the advantages of a multi-domain conceptualization of well-being that incorporates internal and external factors relevant to a Veteran population, such as social integration well-being, financial well-being, health, and employment well-being.¹⁰ The authors emphasize the benefits of the well-being construct in Veteran samples, particularly with respect to the transition to civilian life after service.

Well-being in clinical sample of Veterans

Indeed, Veterans experiencing mental health difficulties often face well-being challenges in one or more of these domains. In a sample of treatment-seeking Veterans, loneliness and perceived social support were frequently endorsed as areas of functional impairment (79.1% and 72.2%, respectively).⁵ Furthermore, Veterans appear to experience greater work-related challenges than civilians,¹¹ and they may be more dissatisfied with work after leaving the armed forces.¹² Unemployment is common among Veterans struggling with mental health difficulties, with nearly three-quarters of treatment-seeking Veterans in a recent study not working, mostly due to ill health.¹³ Another aspect of functioning that may contribute to well-being among Veterans is poor sleep, which research suggests contributes to worse quality of life and well-being among Veterans with mental health problems.¹⁴ Furthermore, Veterans with PTSD were found to have low life satisfaction,¹⁵ and a qualitative study found that Veterans’ lived experience of the disorder was associated with loneliness and social isolation.¹⁶

Previous studies collecting data on Veteran well-being have typically been conducted in the United States and to evaluate programs involving adjunctive aspects of psychological therapy, including horseback riding,¹⁷ horticulture,¹⁸ art therapy,¹⁹ and sport.¹⁶ Although these studies reveal that Veterans may have poorer well-being compared with the general population, a paucity of research has explored the pattern of relationships among mental health, functioning outcomes, and well-being. Such research could further inform understanding of the complexity of functioning and mental health presentations in clinical Veteran populations. Well-being was previously found to be associated with mental ill health,⁷ although the study used a non-Veteran sample.

Current study

This study describes mental well-being in a sample of UK Veterans seeking treatment for mental health difficulties at a UK Veteran mental health charity. It aims to investigate associations between well-being and a number of mental health and psychological functioning outcomes. In particular, it uses a scale of mental well-being (Warwick-Edinburgh Mental Well-being Scale) that measures participants’ appraisals of their mental well-being,⁷ in contrast to primarily external well-being factors, such as financial well-being. It is hypothesized that higher well-being scores will be negatively correlated with scores on mental health and functioning outcomes.

Methods

Ethics

This study was approved by the Combat Stress Research Ethics Committee (ref. pn2020).

Participants

A random 20% sample of Veterans seeking treatment at a UK Veteran mental health charity was selected to complete a questionnaire consisting of demographic and health measures. The sample was drawn from a national sample of treatment-seeking Veterans from a charity offering UK-wide support that receives approximately 2,500 referrals per year. A total of 1,147 patients met the following inclusion criteria: 1) provided an email address, 2) consented to be contacted about research, and 3) were treatment seeking (defined as attending at least one treatment appointment with the charity). Of these Veterans, 989 were contacted (158 had an invalid email address). Overall, 428 Veterans (mean age 50.4 [SD 10.9] y) responded (43.3% response rate).

Procedure

Responses and informed consent were collected online via SurveyMonkey between August and October 2020. Participants were contacted a maximum of five times over a six-week period with the survey link and reminded that participation was voluntary. A full description of the recruitment strategy can be found elsewhere.⁵

Measures

Participants answered basic demographic questions and completed the following measures of mental health problems and psychological functioning.

Primary outcome

The Short Warwick-Edinburgh Mental Well-being Scale (SWEMWBS) is a 7-item measure of mental well-being that uses items 1, 2, 3, 6, 7, 9, and 11 from the longer Warwick-Edinburgh Mental Well-being Scale (WEMWBS).⁷ These items primarily cover feelings about functioning rather than internal cognitions with respect to mental well-being and include “I’ve been dealing with problems well,” “I’ve been feeling optimistic about the future,” and “I’ve been feeling close to other people.” Responses were collected on a 5-point Likert scale ranging from one (none of the time) to five (all of the time), and participants were instructed to reference their experiences in the previous two weeks. The scale has good psychometric properties, including internal consistency (Cronbach’s α = 0.93),²⁰ test-retest reliability,⁷ and convergent validity.²¹ The scale is scored by summing item scores. A score of 19.5 or lower is used as a threshold for low well-being.²¹

Health outcomes

The Dimensions of Anger Reactions-5 (DAR-5) is a five-item measure of anger difficulties, with possible anger difficulties indicated by scores above 12.²²

The International Trauma Questionnaire (ITQ) is an 18-item measure of PTSD and complex PTSD (CPTSD).²³ A score of two or higher on at least one symptom in each of three PTSD clusters, as well as on one of the three functional impairment items, is the criterion for probable PTSD caseness. Diagnostic criteria for CPTSD require meeting the criteria for PTSD and scoring a two or above on at least one symptom in each of three disturbances in self-organisation clusters, as well as on at least one functional impairment item for both PTSD and CPTSD symptoms.

The General Health Questionnaire-12 (GHQ-12) is a 12-item measure of symptoms of common mental health disorders (CMDs),²⁴ such as anxiety and depression, on which a score of four or above indicates potential caseness of CMDs.

The Moral Injury Outcome Scale (MIOS) includes a question about any potentially morally injurious events and 14 items measuring possible symptoms linked to moral injury, with higher scores indicating more severe moral injury.²⁵

The Patient Health Questionnaire-15 (PHQ-15) is a 15-item measure of somatic symptoms,²⁶ with scores between 15 and 30 indicative of high severity of somatic symptoms.

Functioning outcomes

The Alcohol Use Disorder Identification Test is a 10-item measure of alcohol use on which a score higher than 8 indicates hazardous alcohol use and a score greater than 16 represents harmful alcohol use.²⁷

The Sleep Condition Indicator n eight-item measure of sleep quality on which higher scores indicate better quality of sleep and a score less than 16 may indicate a possible insomnia disorder.²⁸

The UCLA Loneliness Scale-3 (UCLA-3) is a three-item measure of loneliness, with higher scores indicating greater loneliness.²⁹

Data analysis

STATA (version 13.0) was used to prepare the fully anonymized data. Descriptive statistics and inferential analysis were computed in IBM SPSS Statistics (version 26). For mental health outcomes, total scores on the measures were calculated and analysed as continuous variables. Data met assumptions for multiple linear regression. Single regression models were fit for each mental health factor with well-being as the outcome variable to identify any associations between demographic and health outcome measures and well-being. A multiple regression model then adjusted for demographic variables and each other mental health factor.

Results

Of the 363 participants who answered the well-being items, 264 (73%) were classified as having low well-being, scoring below the identified 19.5 cut-off.²¹ Descriptive statistics for demographic variables are included in Table 1 for the full sample.

Table 1.

Descriptive statistics for the full sample

Characteristic	n (%)*
Age, y mean (SD)	50.5 (10.9)
Gender (N = 428)
Male	417 (97.4)
Female	11 (2.6)
Ethnicity (n = 400)
White	379 (94.8)
Ethnic minority	21 (5.2)
Relationship status (n = 403)
Single, divorced, separated, widowed	266 (66.2)
In a relationship	137 (34.0)
Employment status (n = 396)
Working or retired	223 (56.3)
Not working	173 (43.7)
Housing status (n = 403)
No fixed address	36 (8.9)
Fixed address	367 (91.1)
Early service leaver† (n = 385)
Yes	17 (4.4)
No	368 (95.6)

Note: N = 428. Because of missing data, some frequencies do not sum to 428.

* Unless otherwise indicated.

^† Left after < 4 years.

Descriptive statistics for mental health and functioning outcomes for the full sample and for lower and higher well-being groups are displayed in Table 2. The mental health outcome for which the highest proportion of Veterans met criteria for probable caseness was symptoms of anxiety and depression (80.7%), followed by loneliness (79.1%), and CPTSD (68.7%).

Table 2.

Descriptive statistics of the full sample and well-being caseness for mental health and functioning outcomes

Outcomes	Full sample, mean (SD)	Well-being group, n (%)
		Lower* (n = 264)	Higher (n = 99)
Well-being	16.83 (4.74)	22.53 (2.87)	14.70 (3.33)
CMDs (anxiety and depression)	7.43 (3.98)	4.30 (3.89)	8.82 (3.22)
Moral injury	34.02 (10.20)	27.54 (9.80)	35.87 (9.14)
Somatisation	4.25 (6.70)	10.11 (5.51)	13.53 (5.14)
Anger difficulties	10.10 (5.84)	7.93 (5.64)	10.89 (5.68)
Sleep problems	12.24 (9.09)	13.56 (7.43)	11.67 (9.49)
Loneliness	7.09 (1.94)	5.85 (2.02)	7.55 (1.69)
Alcohol misuse	10.66 (8.46)	8.99 (6.32)	11.21 (9.01)
PTSD	15.83 (5.89)	12.46 (6.28)	17.05 (5.22)
CPTSD	32.22 (11.10)	24.17 (11.15)	35.26 (9.39)

* Lower well-being is calculated as a score of <19.5.²¹

CMD = common mental health disorder, PTSD = posttraumatic stress disorder, CPTSD = complex PTSD.

Descriptive and military characteristics and well-being

Demographic characteristics significantly associated with well-being were employment, relationship, and housing statuses. Working or being retired, in a relationship, and having a home were positively associated with higher well-being (Table 3). Being an early service leaver was not related to well-being. Further analyses of the sample are reported in Williamson et al.⁵

Table 3.

Standardised beta coefficients for single and multiple linear regression well-being models

Model variable	Standardised β coefficient
	Single model	Multiple predictor
Age	.09	.09
Employment status	−.16	.02
Housing status	−.12	−.04
Relationship status	−.11	−.05
CMDs (anxiety and depression)	−.61	−.22
Moral injury	−.51	−.13
Somatisation	−.39	.001
Anger difficulties	−.42	−.12
Sleep problems	.19	.04
Loneliness	−.52	−.19
Alcohol misuse	−.20	−.02
PTSD	−.47	−.31
CPTSD	−.60	−.55

Notes: Boldface indicates p < 0.05. The reference category for employment is working, for relationship status, being in a relationship, and for housing, having stable living conditions.

Multiple predictor model adjusts for all other variables.

CMD = common mental health disorder, PTSD = posttraumatic stress disorder, CPTSD = complex PTSD.

Health outcomes and well-being

Higher scores on measures of symptoms of anxiety and depression (GHQ-12), physical health (PHQ-15), possible problems with anger (DAR-5), and moral injury (MIOS), and PTSD and CPTSD (ITQ) were significantly associated with lower well-being (Table 3). The strongest association was observed between well-being and CMDs, which include anxiety and depression, followed by symptoms of moral injury and CPTSD.

Functioning outcomes and well-being

Higher loneliness was moderately associated with lower well-being scores (Table 3). Both alcohol use and sleep problems were also associated with worse well-being scores. Higher levels of alcohol consumption and worse symptoms of sleep disturbances were associated with lower well-being. The beta coefficient for well-being was positive because higher scores represent better sleep quality. This value indicates that worse sleep scores were associated with lower well-being scores.

Well-being multiple linear regression model

Each mental health and functioning variable was a significant predictor of well-being in single linear regression models (Table 3). The multiple regression model was a good fit (R² = 0.53, F_{13, 169} = 14.91, p < 0.001). When controlling for all other variables in the multiple predictor linear regression model, GHQ-12, DAR-5, UCLA-3, and PTSD and CPTSD ITQ scores significantly predicted well-being. Comorbidity, defined as the number of mental health and functioning outcomes for which caseness was met, was also significantly associated with worse well-being scores (B = −1.44; 95% CI −1.67 to −1.22, p < 0.001), and the model had good fit (R² = 0.61, F_{1, 268} = 162.02, p < 0.001).

Discussion

This study aimed to describe the pattern of relationships between well-being and mental health and functioning outcomes in a clinical UK Veteran sample. As hypothesized, well-being was significantly associated with worse scores on measures of common mental health problems, such as anxiety and depression, physical ill health, and possible anger difficulties, as well as PTSD, CPTSD, and moral injury. Well-being was also related to functioning outcomes, including sleep difficulties, loneliness, and alcohol misuse. This finding is consistent with a study validating the WEMWBS, which was found to have a correlation of 0.53 with a composite measure of mental health difficulties.⁷ Furthermore, meeting caseness for a greater number of mental health outcomes was associated with worse well-being.

In the current sample, the average Veteran well-being score on the SWEMWBS was lower than the mean score of 23.5, which has been reported in the UK general population.²¹ This is consistent with previous studies with clinical Veteran samples.^18,19 This finding is also in line with findings in non-Veteran clinical samples; for example, well-being was observed to be lower among adults attending therapy for common mental disorders in primary care than in the UK general population.³⁰

Implications

The associations observed indicate that well-being and mental ill health are overlapping but distinct constructs rather than different ends of a single mental health spectrum. This is in line with conceptions of positive and negative mental well-being.³¹ However, these associations suggest that overlap exists between the well-being construct and several different mental health and functioning outcomes. These findings adhere to those of studies reporting improvements in well-being after a wide variety of interventions for Veterans with a range of mental health profiles.^16–19 These studies did not statistically analyze the relationships between well-being and mental health outcomes. However, that well-being improved along with other mental health outcomes in these studies offers tentative support for the findings of the present study. Moreover, in studies with non-Veteran samples experiencing mental health difficulties, the WEMWBS and SWEMWBS have been used to evaluate change and found to strongly correlate with measures of common mental disorders.^30,32,33

Taken together, this evidence suggests the usefulness of the well-being construct as a possible transdiagnostic outcome measure of interventions targeting Veteran mental health. A transdiagnostic outcome measure could be useful when the outcomes of several differing interventions or programs for different disorders need to be compared. A second benefit of well-being as a transdiagnostic outcome measure is that it can provide indications of change in sub-clinical samples because it does not rely on a mental health diagnosis yet is related to a range of mental health disorders. Individuals who are sub-clinical on any other mental health disorder may score low on measures of these disorders, such that changes on these measures over time may be minimal and, therefore, potentially less indicative of improvement in overall functioning.

These benefits are particularly relevant for the clinical Veteran population, which has a high rate of comorbidity and therefore complex mental health presentations.⁵ As such, a transdiagnostic outcome measure may offer a way to compare clinical changes across a variety of complex presentations and interventions to address these presentations. Furthermore, mental health diagnoses in Veterans sometimes persist after treatment, especially PTSD.³⁴ Veterans with comorbid disorders, or who retain a mental health diagnosis after treatment, may require multiple and integrative interventions along with psychological therapy.³⁵ These interventions could be evaluated by including well-being as an outcome, assessing general change across disorders, or in cases in which mental health diagnoses are retained. Finally, even when an individual does not meet the diagnostic threshold for a mental health disorder (sub-clinical), interventions might still address well-being or target positive mental health, and a transdiagnostic outcome measure could evaluate these interventions. Sub-threshold diagnoses could include functioning problems such as poor sleep, loneliness, or work-related problems, which are known to be present in the Veteran population.^11,36,37

Benefits and limitations of this study

The key benefit of this study is that it explored the possible overlap between well-being and mental health and functioning constructs in a clinical Veteran sample, which could inform the usefulness of well-being as a transdiagnostic measure in this population. Another benefit is that the findings indicate the consistency of the well-being construct in the Veteran population. Observed associations between well-being and mental health outcomes are consistent with previous findings and support the conceptualization of well-being as a distinct construct that consists of more than the absence of mental ill health.

However, this study also has limitations. First, data were collected at a single time point, meaning that associations between mental health and well-being across time cannot be analyzed, and the direction of causation cannot be determined. However, the study is the first of its kind to explore these relationships in a clinical Veteran sample, and findings indicate the potential benefits for measuring positive mental health among Veterans.

The second limitation is that data collection took place during the COVID-19 pandemic. The resulting unprecedented changes to daily life may have affected aspects of well-being and mental ill health and therefore responses to the measures. However, longitudinal studies measuring mental health variables among treatment-seeking Veterans and a community sample of Veterans reported that Veterans did not experience significantly worse mental health as a result of the pandemic.^38,39

Another limitation is that the present sample was primarily male, white, and Western. This was the result of the convenience sampling method and the demographic make-up of the UK Veteran population. Although this is indicative of the UK Veteran population,⁴⁰ future research might explore any differences among minority and ethnic Veteran groups.

Finally, all responses were self-reported, and participants may not have been able to effectively map their internal feelings to items listed in the measure or may not have responded truthfully because of the awareness that the answers would be reviewed by a researcher. However, the main measure in this study was well-being, a construct that is inherently subjective in nature; therefore, self-reporting is likely the most effective way to capture it.⁴¹

Future directions

One future direction for research would be to compare the associations between mental health outcomes and well-being over time, which could develop an understanding of any possible direction of causation. Second, future studies might explore the overlap between well-being and mental health outcomes in more depth. Although cross-sectional analysis can reveal the extent of the association between these constructs, a network approach investigates the symptom- or item-level relationships between and within constructs and among symptoms. For example, aspects of impaired functioning are common to several mental health disorders, such as sleep difficulties and interpersonal problems, including PTSD and CPTSD.^9,16 Functioning is also a key concept in eudaimonic well-being, which refers to quality of life and personal fulfilment,⁸ and therefore functioning aspects of both constructs could be found to more strongly connect than other factors by means of a network analysis. This would be consistent with the finding that in the current study CPTSD was more strongly correlated with well-being than PTSD, given the centrality of functional impairment to the former.

A further exploration of the relationship between well-being and mental ill health could have practical and theoretical relevance. Practically, well-being and mental health factors most strongly connected with each other could be targeted more directly by transdiagnostic interventions. A further explanation of the overlap between well-being and mental health disorders would develop theoretical understanding of the constructs as more than two sides of a single mental health spectrum.

Finally, further data on well-being and mental health and functioning outcomes could be collected from non-help-seeking Veterans to understand the pattern of relationships in this group. Understanding the associations between well-being and mental ill health in a potentially sub-clinical sample of Veterans might identify possible risk factors for help seeking, because evidence suggests that Veterans have worse overall well-being than the general population, particularly in the first year after military service.⁴²

Conclusions

This study found that well-being in a sample of help-seeking Veterans was moderately associated with several mental health disorders and functioning outcomes, including sleep problems, loneliness, and alcohol misuse. These findings cohere with conceptualizations of the well-being construct as more than the absence of mental ill health. Results suggest the potential usefulness of including well-being as a transdiagnostic outcome measure in the clinical Veteran population and as part of interventions to support positive mental health. Future research could further explore the overlap between these constructs at the symptom level using a network approach and also in additional Veteran samples.

Biographies

Natasha Biscoe, MSc, is a research assistant at Combat Stress, working on studies exploring psychotherapeutic interventions for moral injury and complex posttraumatic stress disorder.

Dominic Murphy, PhD, DClinPsy, is an academic clinical psychologist and has worked in the field of military mental health since 2003. He is Head of Research at Combat Stress and a senior team member of the King’s Centre for Military Health Research, King’s College London. Murphy is President of the UK Psychological Trauma Society, Director of the European Society for Traumatic Stress Studies, and Director/Trustee at The Forces in Mind Trust. He has published over 180 articles and reports.

Funding Statement

This study received funding from a Mobilizing Insights in Defence and Security Collaborative Network Grant.