It is widely accepted that working in the emergency services is challenging (Wild et al, 2016; Petrie et al, 2018); ambulance personnel belong to this occupational group. Ambulance workers are often exposed to unpredictable and potentially traumatic situations (Wild et al, 2016), alongside organisational issues and increasing pressures (Wankhade et al, 2019).
Such circumstances have negative effects on emotional and physical health (Courtney et al, 2010; Donnelly, 2012; Wild et al, 2016), though the mechanisms that influence these outcomes are unclear. This paper aims to explore to what extent individual factors (resilience, satisfaction with life and perceived stress) affect ambulance personnel mental health, with the authors considering anxiety, depression and post-traumatic stress symptoms as mental health outcomes as determined by the data collection scales used (Weiss, 2007; Zigmond et al, 1983).
Exposure to chronic workplace stressors has been related to low job satisfaction, poor physical health, fatigue, burnout and post-traumatic stress symptomatology in ambulance personnel (Donnelly, 2012). Stress is a physiological and emotional response to an actual or perceived threat; it is a subjective experience (Selye, 1956).
It is well documented that stress can be managed in different ways and varies between individuals; population-level research shows individual differences influence health behaviours and responses to stress (Hannigan et al, 2004). Similarly, ambulance personnel adopt different ways of coping (Mildenhall, 2012).
Epidemiological evidence suggests there is a 27% prevalence rate of general psychological distress in this occupational group (Petrie et al, 2018), which indicates a significant proportion of them do not report this experience. Individual differences may explain the variance in mental health issues in the ambulance worker population. While there is strength in this evidence given that data from over 30 000 ambulance personnel were reported on, there could be some issues with generalisability to the population studied in this paper as data were international (Petrie et al, 2018).
The role of lifestyle in mental health has been explored, where engagement in positive health behaviours has been shown to have protective effects on mental health (Reiner et al, 2013; Dale et al, 2014). These findings are echoed in research for this occupational group (Betlehem et al, 2014), although overall findings are mixed (Hutchinson et al, 2021) and specific lifestyle-related data for ambulance personnel are scarce (Hutchinson et al, 2020).
Furthermore, research has explored the role of personality in the mental health of ambulance personnel. A recent review of 27 articles related to personality, identified high conscientiousness and low neuroticism as protective factors against post-traumatic stress disorder (PTSD); however, it also highlighted that conscientiousness was a risk factor for burnout (Mirhaghi et al, 2016). More recently, research focusing specifically on neuroticism found that higher scores for neuroticism correlated with poorer subjective health (Mutambudzi et al, 2020). With regards to coping, Mildenhall (2012) explored informal coping strategies used by ambulance personnel, and surmised that cognitive coping styles that include avoidance are associated with PTSD and burnout.
Resilience is an individual internal resource that can be fostered by developing flexibility and adaptation to adversity; it is a concept that has been researched extensively in the past decade (Froutan et al, 2018). Consistent with the evidence above on personality, research has shown that lower scores in neuroticism are associated with higher levels of resilience (Froutan et al, 2018).
Furthermore, evidence points to the relationship between satisfaction with life (SWL) and health and wellbeing, showing that those who are more satisfied have positive health outcomes, both physically and mentally (Diener and Chan, 2011; Diener et al, 2018).
While all these individual factors, perceived stress, resilience and SWL have been explored in isolation, their collective impact on mental health in this population has not yet been investigated.
This study aimed to explore individual differences with a focus on the role of perceived stress, SWL and resilience. It was hypothesised that perceived stress will be positively related to all mental health outcomes and that SWL and resilience will be negatively related to all mental health outcomes; those with higher SWL and resilience will have better mental health outcomes and those who perceive higher levels of stress will have poorer mental health outcomes.
Therefore, the authors collected data associated with individual factors and explored the impact and role of these on the mental health of ambulance personnel in the UK.
Methods
Four NHS ambulance trusts in England facilitated the recruitment of a purposive sample of ambulance personnel. The researchers were limited in access to sites and relied on the communication team at each trust to promote the study via a poster in stations and trust email communications over a period of 4 months. The poster signposted people to a web address that held the participant information sheet and further links to the survey.
A screening questionnaire was used to ensure individuals were eligible for participation. Inclusion criteria were that participants had to be currently employed at an NHS ambulance trust and carried out frontline duties, for example as paramedics, emergency medical technicians or emergency care assistants.
NHS and ethical approvals were obtained and participants provided their consent. In addition, a debrief sheet was provided, which included signposting to mental health support resources (reference:18/NSP/058). A small focus group (n=7), consisting of consenting trainee paramedics, was used to assess the average time it took to complete the survey and to discuss participant experience. All trainee paramedic data were deleted and not used in any final analysis.
A cross-sectional design was adopted and the survey used to collect data contained a mixture of publicly available validated measures and demographic questions. Within this study, there were three predictor variables: resilience; perceived stress; and SWL. There were three outcome variables that encompassed mental health: depression; anxiety; and PTSD symptoms.
First, demographic information (age and sex) was collected to profile the participant sample. To capture data on all variables, six validated measures were used: the Brief Resilience Scale (BRS) (Smith et al, 2008); the Satisfaction with Life Scale (SWLS) (Diener et al, 1985); the Perceived Stress Scale (PSS) (Cohen et al, 1983); the Impact of Events Scale revised (IES-r) (Weiss, 2007); and the Hospital Anxiety and Depression Scale (HADS) (Zigmond et al, 1983).
Data on resilience were collected using the BRS (Smith et al, 2008); this comprises six items on a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree), and items were reverse scored as appropriate. Overall scores ranged from 6 to 30 with a higher score indicating greater levels of resilience; however, mean scores were compared in this paper.
SWLS was used to gain an understanding of subjective wellbeing/satisfaction with life (Diener et al, 1985); this is a five-item questionnaire scored on a seven-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree), with a higher mean score indicative of higher life satisfaction.
To measure stress, the PSS was employed (Cohen et al, 1983); this scale has 10 items with a five-point Likert scale, with scores ranging from 1 (never) to 5 (very often). Where necessary, reverse scoring was applied; a higher score indicated higher levels of perceived stress.
The IES-r scale (Weiss, 2007) has 22 questions with a five-point Likert scale to assess the impact of stress/trauma with the scale ranging from 1 (not at all) to 5 (extremely). There are three subscales in the scale—avoidance, hyperarousal and intrusion—however, for the purpose of this study, the overall mean scores were calculated and used, with the higher scores being more indicative of stress/trauma impact.
The HADS (Zigmond et al, 1983) comprises 14 questions with two subscales, anxiety and depression; for the purposes of this study, the subscales were used and an overall, mean score was calculated for each, with higher scores indicating presence of anxiety or depression symptoms.
All questionnaires employed (BRS, SWLS, PSS, IES-r and HADS anxiety and depression subscales) demonstrated a high level of internal consistency as supported by a Cronbach's alpha of 0.90, 0.88, 0.88, 0.96, 0.88 and 0.84 respectively (Table 2).
The data were analysed using SPSS statistical software (v. 26). Once the data had been scored and totalled, the assumptions were checked; there was adequate homoscedasticity and normality of the residuals with all scores for skewness and kurtosis <1.96 (Tabachnick et al, 2007).
Two participants were identified as outliers, one with a mean impact of events score of 4.55 plus another with a mean depression score of 3.29. They were retained in the final analysis as conducting analysis without these participants did not yield any significant changes that would misrepresent the target population. Frequencies and descriptives were calculated for the entire sample and all relevant variables for this paper.
Bivariate analysis was undertaken to explore the relationships between all variables; those that emerged were robust. Multivariate statistics were explored and several linear regression models were produced.
Results
A total of 272 individuals took part in the study; however, 112 did not complete the full battery of questionnaires and were therefore omitted. In total, 160 participants were included in the final analysis.
Of the entire participant sample, 52.2% were men, 46.5% were women and 1.3% chose not to disclose their sex (Table 1). In addition, 159 reported their age; while this may have affected the presented means and standard deviations, it did not affect inferential analysis, so ages were retained. Ages ranged from 20–64 years (χ=40.65; SD=10.62).
| Demographics | Entire sample (n=160) | Entire sample (n=160) mean (SD) frequencies % | |
|---|---|---|---|
| Sex | Male | 52.2 | |
| Female | 46.5 | ||
| Prefer not to say | 1.3 | ||
| Age* | 40.65 (10.62) | ||
The mean scores on anxiety, depression and post-traumatic stress symptoms measured by IES-r indicate that participants generally scored in the positive/adaptive aspect of each scale where a higher mean score indicates more of a presence of associated symptoms. However, it is noteworthy that exploration of raw total score frequencies indicated that 30.4% of participants would qualify as clinical cases of depression, having yielded a mild/moderate/severe score, and 50.8% scoring within the mild/moderate/severe range for anxiety.
For SWL and resilience, a higher score indicates a positive response and it can be observed that generally higher scores were noted as demonstrated by the mean score (χ=3.43 and 4.28; SD=0.82 and 1.45). However, the standard deviations suggest that there are some individual differences where scores in the negative parameters are present for each variable, being highest in SWL (SD=1.45), though less so for depression, with the lowest SD of 0.57. The correlation coefficients for each variable are presented in Table 2; relationships were present in expected directions.
| 1 | 2 | 3 | 4 | 5 | 6 | |
|---|---|---|---|---|---|---|
| 1. Anxiety | - | |||||
| 2. Depression | 0.71* | - | ||||
| 3. Trauma | 0.70* | 0.60* | - | |||
| 4. Perceived stress | 0.71* | 0.73* | 0.67* | - | ||
| 5. Resilience | –0.52* | –0.53* | –0.52* | –0.63* | - | |
| 6. SWLS | –0.57* | –0.70* | –0.63* | –0.62* | 0.37* | - |
| Mean | 2.11 | 1.74 | 2.18 | 2.86 | 3.43 | 4.28 |
| SD | 0.69 | 0.57 | 0.93 | 0.71 | 0.82 | 1.45 |
| Cronbach's alpha | 0.88 | 0.84 | 0.96 | 0.88 | 0.90 | 0.88 |
| Skewness | 0.36 | 0.61 | 0.68 | –0.06 | –0.42 | –0.29 |
| Kurtosis | –0.54 | –0.29 | –0.20 | –0.38 | –0.25 | –0.74 |
SWLS: Satisfaction with Life score SD: standard deviation
As noted, measures of resilience and SWL concern positive aspects of mental health so a significantly negative correlation can be observed and this indicates moderate to strong relationships with anxiety, depression, trauma and perceived stress (rs=–0.52 to –0.70; P<0.01).
Furthermore, as expected, anxiety and depression correlated positively and strongly with each other (r=0.71; P<0.01); stress related strongly with all mental health outcomes (rs=0.67– 0.73; P<0.01). There were no issues with multicollinearity (Pallant, 2013); there is satisfactory evidence to suggest independence of the predictor variables, indicating that linear regression is applicable to consider the unique contribution of personal factors (stress, resilience and SWL) on the mental health of ambulance personnel.
Table 3 displays the linear regressions that were run to understand the impact of individual factors (resilience, perceived stress and SWL) on the mental health outcomes (post-traumatic stress, anxiety and depression symptoms). To assess linearity, a scatterplot of each predictor variable against each outcome variable with superimposed regression lines were plotted. Visual inspection of these plots indicated a linear relationship between the variables. Resilience, perceived stress and SWL statistically significantly predicted post-traumatic stress symptoms (F(3,156)=46.10; P<0.001), anxiety symptoms (F(3,156)=58.98; P<0.001) and depression symptoms (F(3,156)=91.37; P<0.001). They accounted for: 47% of the variance in post-traumatic stress symptoms, with adjusted R2=46; 53% of the variance in anxiety symptoms, with adjusted R2=52; and 64% of variance in depression symptoms, with an adjusted R2=63. In two of the three models, perceived stress primarily contributes to trauma and anxiety, with the highest beta scores (βs=0.53 and 0.48), for depression. SWL is the primary contributor (β=—0.41); but all models show a large effect size.
| Variables | β | SE β | β | R2 | Adj. R2 | F (df) |
|---|---|---|---|---|---|---|
| Resilience | –0.17 | .09 | –0.15* | 0.47 | 0.46 | 46.10 (3156) |
| SWLS | –0.05 | 0.05 | –0.08 | |||
| Perceived stress Dependent variable: lES-r | 0.69 | 0.12 | 0.53* | |||
| Resilience | –0.11 | 0.06 | –0.13 | 0.53 | 0.52 | 58.98 (3156) |
| SWLS | –0.11 | 0.03 | –0.22* | |||
| Perceived stress Dependent variable: Anxiety | 0.47 | 0.08 | 0.48* | |||
| Resilience | –0.09 | 0.04 | –0.13 | 0.64 | 0.63 | 91.37 (3156) |
| SWLS | –0.16 | 0.02 | –0.41* | |||
| Perceived stress |
0.31 | 0.06 | 0.39* |
SWL: satisfaction with life; SWLS: Satisfaction With Life Scale; lES-r: Impact of Events Scale revised
Discussion
Individual factors, specifically perceived stress, resilience and SWL, have a significant impact on the mental health of ambulance personnel and these factors explain nearly half of the variance in each mental health outcome measured (anxiety, depression, post-traumatic stress symptoms).
The highest level of variance was accounted for in the model that included depression as the outcome measure; in this paper, more than half of the variance (63%) was explained by perceived stress, SWL and resilience. It can be observed that the primary contributor in this model was SWL. This makes conceptual sense given that key symptoms of depression are correlated with SWL (Samaranayake and Fernando, 2011).
This was followed closely by perceived stress in this model; evidence indicates robust and causal associations between stressful life events that impact SWL and depressive episodes (Hammen, 2005; 2015). Therefore, this model is consistent with existing research. Furthermore, as outlined, a significant proportion of the participants in the study scored within the realm to count as cases of anxiety and depression (Zigmond et al, 1983) in line with existing epidemiological evidence (McManus et al, 2014; Petrie et al, 2018).
Though resilience contributed collectively to the variance in each model, it was superseded in the depression and anxiety models, significantly contributing only in the model for PTSD symptoms. This finding was interesting as it partially refutes existing evidence that outlines resilience's protective and predictive roles in mental health (Kermott et al, 2019), since it plays a limited role in the anxiety and depression models.
However, in relation to the PTSD symptoms regression model, both resilience and perceived stress were seen to be significant contributors.
In research, post-traumatic growth is a salutogenic concept that refers to a positive outlook following trauma/stress (Kang et al, 2018); some characteristics of this are echoed in the definitions of resilience. Resilience has been shown to be inversely associated with post-traumatic growth (Levine et al, 2009). Thus, taking this into account would support and explain the role of both stress and resilience in overriding SWL in this regression model. Though taken together, the individual factors account for a significant level of variance in PTSD symptoms; by its nature, trauma is complex (Kessler et al, 2017), and there is research that suggests psychopathology associated with PTSD does not always present in people exposed to trauma (Levine et al, 2009; Kessler et al, 2017).
Furthermore, there is a complex interplay between reaction to trauma and past life experiences as explored in research on adverse childhood events, trauma and the impact on ambulance personnel wellbeing (Maunder et al, 2012); however, this evidence is tentative because of limitations associated with sample size. Nevertheless, while the individual factors measured in this study explain a significant proportion of the variance in mental health, it is important to consider what other factors might account for the remaining variance.
Perceived stress was the primary contributor in each model, having both an impact on individual factors of mental health, as well as on all factors collectively. This is logical, as a wealth of research indicates that stress is a common characteristic in many mental health disorders (Hammen, 2005) and is a risk factor in both mental and physical health concerns (Cohen et al, 2007). For example, those who experience work-related stress (from high demands and low cowntrol in their role) are more likely to experience major depressive or anxiety disorders (Wang et al, 2008). Furthermore, stress has been shown to be associated with poorer health behaviours and the adoption of unhelpful coping strategies which, in turn, are implicated in poorer mental health (Velten et al, 2014). Whether this is of a causal or perpetuating nature is unknown. However, it offers some explanation of the relationship that has been reported in this paper.
Each model has its own nuances and explanations and, overall, the findings would suggest that the significant level of variance in the mental health of an ambulance worker can be explained by individual factors, more specifically perceived stress, SWL and resilience. However, this should not necessitate an individual-blaming approach; research has illustrated the role of organisational pressures and work-related stress (Selye, 1956; Donnelly, 2012; Wankhade et al, 2019), which may explain the remaining variance in the models.
Limitations
There is scope for future refinement of this study. First, the sampling method employed relied on trusts’ in-house communications teams for distribution and email lists. Internal communications suggested that the trust would advertise to all staff; however, given time frame of data collection, this was limited to those only who were on shift; ambulance personnel who were on leave or absent from work may not have been captured in the data set.
Furthermore, because of anonymity, the researchers were unable to obtain specific data for each trust so there could be between-trust differences that have not been captured in the data.
The term ‘individual factors’ is subjective and there are whole other realms of research that have defined these in terms of personality (Mutambudzi et al, 2020), lifestyle (Betlehem et al, 2014) and past adverse life events (Maunder et al, 2012), all of which have been shown to affect mental health. Therefore, there are wider variables that could have been explored.
Furthermore, collecting data on mental health in a quantitative way poses significant limitations. A person's lived experience can differ from that of another. Future research could ameliorate this by using mixed methods or specific qualitative research in which interesting findings might be captured (Clompus and Albarran, 2016).
Finally, as noted above, there was a high degree of drop-out (112 incomplete measures), which could impact the generalisability of the study. Given the sensitive and personal nature of the measures, the incomplete data could be reflect withdrawal from the research or withdrawal for pragmatic reasons (e.g. time pressures). Future research could benefit from improvement in engagement strategies such as on-site visits for data collectors and the presence of data collectors, although this would need to be considered carefully to mitigate observation and experimenter influence or effects (Rosenbaum, 1989).
Recommendations
The authors recommend applying these findings at an organisational level, such as by promoting existing wellbeing strategies and support to help improve positive aspects of individual factors i.e. resilience and SWL.
However, in terms of stress, and how to relieve the impacts, further research is warranted to understand the sources and pathways of distress.
Conclusion
This paper has established that individual factors, perceived stress, resilience and SWL play a significant role in the mental health of ambulance personnel, though the methodological limitations associated with cross-sectional data preclude definitive statements on causation between the variables measured.
It is beyond the scope of this paper to explore and attribute the causes of stress or what explains the perception of stress. However, future lines of enquiry would warrant exploration of the sources of stress within the ambulance service. This has the potential to open up various opportunities for intervention and support, for example, if stress is related to organisational issues, management advice and intervention could be provided or, if related to personal factors, then application and use of existing employee wellbeing resources could be promoted and used where appropriate.