INTRODUCTION
Paramedics are exposed to a variety of stressors during their day-to-day work; these may be related to the provision of patient care or more chronic in nature. A significant body of research has examined the stressors related to patient care, known as “critical incident stress.” Critical incident stress includes, but is not limited to, responding to the death of a child,Reference Regehr, Goldberg and Hughes 1 , Reference Alexander and Klein 2 providing care to friends, family or those known to the responder,Reference Alexander and Klein 2 - Reference Jonsson, Segesten and Mattsson 6 and treating acutely ill or seriously injured patientsReference Alexander and Klein 2 , Reference Beaton, Murphy and Johnson 4 , Reference Brough 7 , Reference Regehr, Hill, Goldberg and Hughes 8 . Additional stressors associated with the provision of care include the risks of exposure to blood-borne pathogens,Reference Leiss, Ratcliffe and Lyden 9 - Reference Rischitelli, Lasarev and McCauley 11 verbal or physical violence,Reference Brough 7 , Reference Suserud, Blomquist and Johansson 12 , Reference Bigham, Jensen and Tavares 13 and injury or death from vehicle-related crashesReference Maguire, Hunting and Guidotti 14 - Reference Becker, Zaloshnja and Levick 16 . To a lesser extent, other research efforts have found that paramedics may also experience chronic work-related stressors, including insufficient salaries,Reference Beaton, Murphy and Pike 17 , Reference van der Ploeg and Kleber 18 conflict with administrators,Reference Beaton, Murphy and Pike 17 , Reference Young and Cooper 19 , Reference Regehr and Millar 20 lack of support from or conflict with colleagues,Reference Alexander and Klein 2 , Reference Bennett, Williams and Page 3 , Reference Clohessy and Ehlers 5 , Reference Beaton, Murphy and Pike 17 - Reference Beaton, Murphy and Pike 21 and interference with non-work-related activitiesReference Bennett, Williams and Page 3 , Reference Clohessy and Ehlers 5 , Reference Beaton, Murphy and Pike 17 , Reference Beaton, Murphy and Pike 21 .
Exposure to work-related stressors has been linked to stress reactions, most frequently to posttraumatic stress.Reference Regehr, Goldberg and Hughes 1 , Reference Bennett, Williams and Page 3 , Reference Clohessy and Ehlers 5 , Reference Jonsson, Segesten and Mattsson 6 , Reference van der Ploeg and Kleber 18 , Reference Haslam and Mallon 22 - Reference Marmar, Weiss and Metzler 29 Most research in this area has focused on the link between critical incident stress and posttraumatic stress symptomatology (PTSS). However, recent research has found preliminary evidence that chronic work stressors may significantly contribute to PTSS. A study of United States emergency medical services (EMS) personnel found a significant link between organizational stressors (the stresses associated with the culture of the EMS service), operational stressors (the stress associated with the practice of EMS), and posttraumatic stress.Reference Donnelly 30 These findings suggest that a more holistic view of stress may be required to fully understand the risk for posttraumatic stress associated with paramedic practice. However, there are unique structural differences in the provision of EMS services in the United States, so it is not clear is if this phenomenon is generalizable to the Canadian context. Beyond knowledge gaps regarding the factors involved in work-related stress, there remains a paucity of literature examining where paramedics would prefer to seek out support for issues relating to work-related stress.
It is important to understand all the ways in which paramedics may be at risk for posttraumatic stress; it is equally important to understand how paramedics would like to receive support to manage that stress. The primary objective of this study was to determine if a relationship exists between chronic work-related stress and critical incident stress with the development of PTSS, and to identify variables associated with the development of PTSS among Canadian paramedics. The secondary objective was to determine where paramedics would prefer to receive support for work-related stress.
METHODS
Study design and sampling
All paramedics employed in one county-based EMS service in southwest Ontario (annual call volume approximately 80,000) were contacted via email using the survey protocol recommended by Dillman.Reference Dillman 31 The EMS service was comprised of staff working at both the primary care paramedic (PCP) and advanced care paramedic (ACP) levels. The email contained a link to the survey as well as an option to unsubscribe from the study. Respondents received a total of up to five contacts: a pre-survey informational contact alerting participants to the upcoming study, followed by two invitations to participate and two reminders. The surveying took place during the fall of 2011.
The study received ethics approval by the University of Windsor Research Ethics Board.
Instruments
Posttraumatic stress symptomatology (PTSS) was measured using a standardized tool called the PTSD (posttraumatic stress disorder) Checklist (PCL).Reference Weathers, Litz and Keane 32 , Reference Blanchard, Jones-Alexander and Buckley 33 The PCL is a 17-item scale that provides a continuous measure of PTSS and a threshold cut-off that indicates possible PTSD. Response options are on a 5-point Likert scale and possible scores range from 17 to 85. Scores over 50 are indicative of possible PTSD. The PCL has been successfully used in prior paramedic research.Reference Donnelly 30
Two types of chronic stress were assessed in this study. The EMS Chronic Stress ScalesReference Donnelly, Chonody and Campbell 34 assess both organizational and operational types of chronic workplace stress. Each scale consists of 10 items. Operational stress includes the stress associated with the structural elements of working on an ambulance service, like shift work, risk of being injured, and fatigue. Organizational stress encompasses factors associated with the culture of the organization in which the respondent is working (e.g., conflict with supervisors, changes in policies). Respondents were asked to report levels of stress over the past six months on a 7-point Likert scale, and the responses were summed, resulting in scores that could range from 10 to 70.
Critical incident stress was assessed using the Critical Incident Stress Inventory for EMS,Reference Donnelly and Bennett 35 an inventory that examines both the number of exposures to a select number of critical incidents, as well as levels of stress associated with those exposures. If respondents indicated they had been exposed to a critical incident, they were asked to report on a 7-point Likert scale how much stress that incident had caused them over the past six months. The responses were summed, resulting in scores that could range from 0 to 252.
In order to determine where paramedics would primarily prefer to go for support in dealing with work-related stress, respondents were asked, “If you felt that you were suffering from work-related stress, how likely is it you would go to the following for help?” Respondents were asked to rate the sources of support on a 7-point Likert scale. Response options ranged from 1 (not at all likely) to 4 (neither likely nor unlikely) to 7 (extremely likely). Respondents rated sources of support, including a supervisor, a co-worker, a base hospital educator, a union representative, a partner, a family member or friend, and the employee assistance program or a therapist.
Demographic measures collected in this study included age, gender, marital status, level of certification, length of service in EMS, weekly hours worked, and income.
Statistical analysis
Analyses were conducted using SPSS (v. 22). In order to reduce error due to missing data, respondents who had not completed at least 85% of the survey were removed from the sample.Reference Hertel 36 Descriptive statistics were used to determine demographic breakdown of the survey sample and the prevalence of types of work-related stress and posttraumatic stress. Ordinary least squares (OLS) linear regression was used to determine the relative influence of different stressors on posttraumatic stress. The R 2 coefficient of determination was used to assess the goodness of fit of the model. Repeated measures analysis of variance (ANOVA) strategies were used to assess the relative difference in preferred sources of support.
RESULTS
Of 269 paramedics invited to participate in the study, 162 (60%) responded to the survey. Nine individuals declined to participate and eight respondents were not included, as they had not completed at least 85% of the survey. The final resulting number of usable responses was 145 (a 54% response rate). The scales in this study demonstrated acceptable reliability (operational stress scale α=0.863, organizational stress scale α=0.876, and posttraumatic stress scale (PCL) α=0.915). Demographic characteristics of the study population are presented in Table 1.
Predictors of PTSS
The data were initially examined to determine whether significant bivariate relationships existed, as it would be inappropriate to undertake multivariate analyses in the absence of this. The results of this analysis are provided in Table 2 and indicate that PTSS is significantly correlated with all three types of workplace stress (p<0.001).
*p<0.001
In our bivariate analyses, we determined whether differences existed in critical incident stress, organizational and operational stress, as well as posttraumatic stress, by demographic characteristics (e.g., age, income, level of training). The results were virtually uniformly non-significant, with the exception that advanced care paramedics (ACPs) were found to have significantly higher levels of critical incident stress than primary care paramedics (PCPs, p<0.01).
Given the significant bivariate results, we proceeded to a multivariate regression analysis. Our use of OLS regression allowed for the testing of the relationship between the different types of workplace stress and posttraumatic stress, while controlling for the influence of demographic factors (e.g., level of training or years of experience). In Table 3, the standardized coefficients (beta weights) are provided; standardization converts the beta weights of all variables to a common metric, which allows the magnitude of the coefficients to be meaningfully compared. Model 1 tested PTSS against demographic factors, none of which were found to be significantly associated. We then added other variables in a forward stepwise manner to test their impact on the predictive power of the model. In Model 2, the stress variables were added. Operational stress was found to be significantly associated with PTSS (p<0.001), whereas organizational stress was not (p=0.672). Critical incident stress was also significantly associated with PTSS (p<0.05). In Model 3, an interaction term was applied to determine whether an interaction of operational stress and critical incident stress increased the overall predictive power of the model. We found the introduction of the interaction term resulted in critical incident stress losing significance as an independent predictor of PTSS.
*p<0.01, **p<0.001
In the final model, operational stress retained significance as a predictor. The standardized beta weights indicated that operational stress had the greatest influence on the predictive power of the model, both as an independent predictor and as part of the interaction term. The final model was found to have an R 2 of 0.39, thus explaining almost 40% of the variance in PTSS.
Sources of social support
More than 80% of respondents reported they were likely to go to a friend or family member for assistance and more than 70% reported they would go to a work partner (someone with whom they regularly work on the ambulance) in dealing with work-related stress. Less than half of respondents were likely to seek help from a co-worker, with even fewer being willing to seek assistance from the other sources of support. In order to assess if the observed differences in preferred sources of social support were statistically significant, repeated measures ANOVA were utilized. These revealed significant differences in preferred sources of support (Table 4). Mauchly’s test indicated that the assumption of sphericity had been violated (χ2 (20)=69.3, p<0.001), so the Greenhouse-Heisser estimate of sphericity (ε=0.845) was used to correct the degrees of freedom. After correction, significant differences remained between the preferred sources of social support (p<0.001). Post-hoc analyses, utilizing a Bonferroni correction, reaffirmed our finding that a statistically significant proportion of respondents preferred their sources of support to be a friend, family member, or work partner. The groups that were statistically distinct from each other were, by order of most to least preferred:
-
1) a friend, family member, or work partner
-
2) a co-worker, employee assistance program, or other therapist
-
3) a union representative or supervisor
-
4) a base hospital educator.
Result of repeated measures ANOVA: F(5.01, 710.11)=108.94, p<0.001
DISCUSSION
Our findings suggest that a multiplicity of stressors are significant predictors of PTSS, and moreover that it is critical incident stress combined with operational stress that contributes to the risk of PTSS, rather than critical incident stress alone. We did not find a significant association between demographic factors and posttraumatic stress. However, when controlling for demographic characteristics, we found that operational stress was significantly associated with PTSS. While critical incident stress did not significantly predict posttraumatic stress, critical incident stress interacting with operational stress was found to be a significant predictor of posttraumatic stress.
While unique to Canada, these findings are consistent with the results of research done on US paramedics,Reference Donnelly 30 which also found a significant relationship between operational stress and PTSS and a significant interaction between operational stress and critical incident stress.
Operational stress was found to have a significant association with PTSS in our study population, suggesting the stressors that fall into that domain may be important as a target for intervention. Operational stress encompasses many factors that reflect the structure of service provision in EMS. Specifically, operational stress involves such things as the strain from shift work, missing and working through meals, fatigue, managing a social life outside of work, worry of injury, lack of understanding from friends and family, and feeling like one is always “on the job.” These stressors may not currently be addressed by existing interventions for workplace stress in EMS systems. Given the strength of the relationship that we found between operational stress and PTSS, this may be an area in which there is great opportunity for beneficial intervention.
Current interventions, like Critical Incident Stress Debriefings,Reference Mitchell 37 , Reference Everly, Flannery and Mitchell 38 focus exclusively on critical incident stress, and despite their widespread use, have inconsistent evidence supporting their useReference Rose, Bisson and Churchill 39 . It is possible that having paramedics return to work on a scheduled day off for debriefing could increase their operational stress, especially in a shift work environment. Other common interventions for work-related stress are offered in the form of employee assistance plans, Workplace Safety Insurance Programs, staff psychologists, or other institutional provisions of mental health services. But such interventions are focused on treating pathological responses after they have occurred, rather than on prevention. Further, significant barriers to accessing these services have been identified, including stigma, scheduling challenges, personal beliefs about mental health and mental illness,Reference Vogt 40 , Reference Corrigan 41 and concerns about the professional ramifications of help-seekingReference Cares, Pace and Denious 42 . Given the problematic nature of the current interventions available to paramedics, our findings underscore the need for the development and validation of evidence-based interventions to address the multiplicity of factors that can contribute to the development of stress reactions in paramedics. Further, our findings suggest that interventions for managing workplace stress should be holistic and target both critical incident and chronic workplace stressors.
We identified significant differences between preferred sources of support for managing workplace stress among paramedics. Paramedics indicated a strong preference for receiving support from a work partner or friend or family member. These findings suggest that emerging interventions, like peer support programsReference Scully 43 and programs that equip families with the resources to support paramedic family members, might be beneficial for paramedic health and well-being. Our findings also agree with previous research which found that social support from both work and family sources ameliorated stress and was protective against distress for police officers and firefighters.Reference Graf 44 - Reference Patterson 46
LIMITATIONS
A number of limitations should be kept in mind when interpreting our results. Our study was conducted using a convenience sample, which may limit the ability to generalize our findings. In particular, it is not possible to determine whether the relationships between different stresses and posttraumatic stress we observed would be found in other EMS services. Another limitation lies in the response rate. While we feel our response rate of over 50% was reasonable, our results may be biased by non-responders differing from responders in ways that we were not able to determine. At the time of data collection, it was not possible to access demographic data for the entire service, which precluded us from carrying out an assessment to determine whether those who responded to the survey were representative of the overall service. Future research may avoid this limitation with the use of a non-responder survey to determine whether there are qualitative differences between those who did or did not participate. A survey, while arguably the best modality for addressing the research questions outlined in this study, is by its nature cross-sectional, thus providing a “snapshot” of providers without the ability to assess how provider experience may change over time. Further, self-reporting is vulnerable to social desirability bias, whereby respondents alter their responses to reflect how they think they ought to reply rather than in a way that reflects their true experiences or beliefs. Unfortunately, in order to reduce the response burden in this survey, we were unable to include a measure to assess the degree to which concerns of social desirability might be influencing the findings. Given the complexity and multidimensionality of social science research, capturing every factor that might influence PTSS is impossible. Our study did not assess a multiplicity of factors which have been identified as associated with PTSS (e.g., personal coping style, previous trauma history). However, we feel that our finding that the inclusion of operational stress and the interaction between operational stress and critical incident stress account for nearly 40% of the variation in posttraumatic stress scores remains extremely compelling.
CONCLUSION
Both chronic and critical incident stressors appear to be significant predictors of PTSS. Our findings suggest that holistic health and wellness initiatives that address the impact of both critical incident stress and the chronic stressors associated with day-to-day operations may help mitigate PTSS. Our findings also provide preliminary evidence that interventions may benefit from a focus on peer support and on friends and family members who can support the affected paramedic. Future research efforts should focus on exploring the impact of chronic and critical incident stress on other stress reactions, such as depression and anxiety, as well as exploring how stress might influence safety-related outcomes like injury, errors, adverse events, and safety-compromising behaviours.
Competing Interests: None declared.