Introduction

Greece is in the frontline of the European refugee crisis. During the first quarter of 2016, more than 145,000 refugees disembarked at the Greek islands. The majority of them reached at the small island of Lesvos that shoulders around 70% of the total refugee arrivals in Greece. Furthermore, the estimated number of refugees and migrants that entered Greece in 2015 exceeded the 855,000 [1]. Most of them travel by boats and dinghies from Turkey to the Greek islands of the Aegean Sea (Lesvos, Chios, Samos, Leros, Simi, Kos and Tylos), aiming to traverse the mainland [2] in their effort to reach northern and central European countries. The vast amount (49% from the Syrian Arab Republic, 26% from Afghanistan and 16% from Iran) [1] fled war and persecution to seek a better life. Nevertheless, many times the journey is equally horrifying and dangerous with what they left behind: shipwrecks, drownings, dehydration, exhaustion, starvation and serious injuries that can result in death in the sea or ashore are lurking. Those are common phenomena faced by refugees and migrants, as well as the rescue workers who aid individuals in danger and help bring them in safely [3].

The terms “rescue workers” or “rescuers” refer to individuals that on professional or voluntary basis engage in stressful activities targeted at providing assistance to people in emergency circumstances. Numerous rescue workers are currently operating in Lesvos sharing a common mission; to aid refugees arriving by sea and to offer search, rescue, and first aid services. As a consequence of their duties, they are exposed to several physical and emotional traumatic events on an almost daily basis [4, 5]. Rescue workers witness terrifying scenes of dead, dying or severely wounded people of all ages, screams of people reaching for help and the mourning or agony of people whose relatives or beloved persons are under danger [6]. The repeated exposure to these stressful and/or traumatic experiences can result in lasting negative health effects (such as depression, burnout or substance use), while can also impact on developing Post-Traumatic Stress Disorder (PTSD) [7].

PTSD and rescue workers in the literature

Although some studies have shown that the majority of rescuers cope well emotionally after the exposure to potentially traumatic events [8–10], PTSD is considered a relatively common mental condition among them. Rescue workers who are involved in high-risk missions and interventions, such as wars or natural and human-made disasters, are exposed to numerous chronic and traumatic stressors [11, 12]. This often increases the risk of PTSD [13].

In addition, the risk of developing PTSD increases proportionally with the number of traumatic events. Therefore, rescuers are considered a high risk group for PTSD [14]. According to the literature, the prevalence of PTSD is much higher in rescuers comparing to that of the general population (4%) [11, 15, 16]. A recent systematic review and meta-analysis showed that the pooled current worldwide prevalence of PTSD among different types of rescue workers is 10% [17]. In addition, significant variation in PTSD development between professional and volunteer rescuers has been observed in the literature [18].

The aim of this study was to assess the prevalence of PTSD among the rescue workers operating in the island of Lesvos, Greece during the European refugee crisis. Secondary objectives were to explore the potential differences in PTSD prevalence between the different categories of rescue workers (e.g., professionals vs. volunteer rescuers, Greek vs. international professional rescuers) and to identify the variables that could explain possible observed variations.

Methods

Setting and participants

This cross-sectional study was conducted in Lesvos, Greece from 20th February to 20th April 2016. The survey was performed among the rescue workers that were operating in the island at the time. The following inclusion criteria were adopted: (a) professional or volunteer rescuers operating in Lesvos during the under study period, (b) operating in the island for at least four weeks before the survey, (c) being affiliated as official members of a Greek or an international foundation or committee of rescuers, (d) offering search and rescue services to boat refugees coming ashore and/or first aid services to those rescued and may be in urgent need (e.g., traumatized, dehydrated, etc.) or assisting the vulnerable groups (e.g., physically disabled, children, etc.). Offering other types of services, such as provision of food or cloth, was considered as major exclusion criterion.

Five different foundations were officially vying in Lesvos. Among them were: the OXFAM America (http://www.oxfamamerica.org), the Boat Refugee Foundation (BRF) (http://bootvluchteling.nl), the Hellenic Red Cross (HRC) (http://www.redcross.gr), the Greek Rescue Team (GRT) (http://www.hrt.org.gr) and the United Nations High Commissioner for Refugees (UNHCR) (http://www.unhcr.org). All rescuers operating in Lesvos during the study period were asked to participate in the study. The response rates were: (a) OXFAM: 88% (N = 40), (b) BRF: 90% (N = 45), (c) HRC: 95% (N = 48), (d) GRT: 90% (N = 43), (e) UNHCR: 83% (N = 41). Summarizing, a total of 217 participants were enrolled in the study. Participants were grouped in three main categories according to their affiliation: “Greek Professionals Rescuers/GPR” (HRC, GRT), “International Professionals Rescuers/IPR” (OXFAM, BRF), and “Volunteer Rescuers/VR” including both Greek and international volunteers (UNHCR).

Procedures and tools

After literature review, the Post-traumatic Stress Disorder Checklist-Civilian Version (PCL-C) [19, 20] was selected to assess the development of probable PTSD in the selected rescue workers. The PCL-C is considered to be a validated tool, with good psychometric properties that is easy to administer. It is commonly used when a clinical interview is not feasible to take place [21]. This tool uses the DSM-IV criteria [22].

More specific, the PCLC-C is a self-reported 17-item tool that uses a 5-point likert scale (1. Not at all, 2. A little bit, 3. Moderately, 4. Quite a bit, 5. Extremely). The scores range from 17 to 85, while participants with a score of 50 or greater are classified as having probable PTSD. This cut-off provides a sensitivity of approximately 80% [23]. PCLC-C assesses the full domain of PTSD symptoms categorized in three groups: Group B—intrusive and re-experiencing, Group C—numbing and avoidance, and Group D—hyper-arousal.

In our study, each symptom was assessed as event-specific (“as a result of operating in Lesvos during the refugee crisis”) and current (“within the last 21–30 days”). According to the suggestion of North and Pfefferbaum [24], which was also followed by another similar study [11], the outcome is mentioned as probable PTSD to acknowledge the fact that symptoms derived via the use of a screening tool such as PCL-C do not necessarily indicate psychopathology.

Permission was obtained by the original developers for research purposes only. Both the English and the Greek version of the questionnaire were provided, to address all rescue workers in their native language or language of understanding. Both versions were already validated [19, 25], nevertheless were additionally tested by the authors. To achieve that, a small focus group of ten rescuers (five Greek and five international) was formed to identify any further cultural particularities in this population group. No comments or suggested changes were reported by the participants.

Subsequently, the PCL-C was administered to the rescue workers along with a biographical cover sheet in relation to demographic information such as age, gender, nationality, family status, parenting, and educational level. The occupational parameters covered included occupation/volunteer affiliation, previous experience in rescue interventions during emerging events, operation period in Lesvos, duration of shifts, participation in collecting dead refugees and children from the sea, and the provision of psychological support. Furthermore, behavioral factors were explored regarding smoking habits and alcohol consumption.

Statistical analysis

The analysis was conducted in the SPSS 21, while all tests were two-tailed and performed at a confidence level of 95%. The frequency of different PTSD symptoms (Groups B, C, D) per rescuers category (GPR, IPR, VR) and the total prevalence of probable PTSD were estimated. The Kolmogorov–Smirnov normalization test was conducted to test normality of data. Mann–Whitney and Kruskal–Wallis tests were used to explore any association between post-traumatic stress and occupation/volunteer affiliation. Additionally, a multivariate logistic regression model was performed to assess the risk of high probable PTSD in volunteers and professionals.

The probable PTSD was used as the dependent variable in the regression model, while the independent variables were identified after testing for fitness to the model by performing both univariate and multivariate analysis (a < 0.05).

Ethical standards

The study was approved by the Board of Trustees of the 2nd Health Region of Piraeus and Aegean Islands (Protocol: 7947) and the Scientific Committees of Greek Rescue Team (protocol: 3216). Information letters were sent to all five foundations and upon receiving their approval, participants were informed of the aim of this study and signed an individual’s consent form.

Results

Socio-demographic characteristics and selected parameters regarding the working conditions of the enrolled rescuers are presented in Table 1. The majority of the participants were males (N = 180; 82.9%), with mean age ranging from 35 to 45 years old depending on rescuers’ category (GPR, IPR, VR). The 39.6% of the GPR, the 51.8% of the IPR, and the 61% of the VR were married and had two children on average. Most of the rescuers had participated in more than nine interventions in the past (GPR: median = 13, IPR: median = 15, VR: median = 9). GPR presented the highest values as regards the operation period time in Lesvos (median = 29 days; SE = 2.4), the duration of shifts (median = 9 h/day; SE = 2.8), and the number of dead refugees collected per rescue intervention (median = 12 bodies; SE = 4.2). Furthermore, GPR’s smoking status and alcohol consumption was found higher than the other groups (78% smokers and 14.3% alcohol consumers).

Table 1 Participants’ characteristics and behavioral factors per rescuers’ category
Full size table

Table 2 presents the proportion of probable PTSD per participants’ characteristics for each rescuer’s category. A total of 37 (17.1%) participants reported symptoms consistent with probable PTSD. The average probable PTSD proportion per rescuer’s category varied significantly; 23.1% in GPR, 11.8% in IPR, and 14.6% in VR (p = 0.02). Females, participants over 38 years old, single/divorced/widowed individuals, and those who had no or one child presented significantly higher probable PTSD proportions among all rescuer’s categories (p < 0.05).

Table 2 Proportion of probable PTSD per participants’ characteristics
Full size table

Frequency of previous experience in similar operations was a significant parameter for probable PTSD in both IPR and VR (≤4; >4 events: 19.3%, 10.7%; p = 0.03 and ≤4; >4 events: 19.8, 12.3%, respectively; p = 0.04). In contrary, GPR presented similar probable PTSD scores regardless of their previous experiences (≤4; >4 events: 25.2, 26.3%; p = 0.371). An operation period of more than 14 days and the collection of more than one dead child per rescue contributed significantly to higher probable PTSD proportions (p = 0.01). Furthermore, daily shifts of more than 4 h and the collection of more than six dead refugees per rescue resulted in higher probability of probable PTSD in the GPR and the VR categories (p = 0.03 and p = 0.02, respectively).

The frequency of PTSD symptoms among rescuers that were diagnosed with probable PTSD is presented in Table 3. Percentages of positive symptoms varied significantly among the different rescuers categories (p < 0.03). Overall, higher percentages were met in Group B and D (intrusive re-experiencing and hyper-arousal, respectively). The highest item for Group B regarded the “intense psychological distress to cues” (GPR = 80.9%, IPR = 60%, VR = 83.3%). As regards items of Group C, the highest percentages were observed in “sense of a foreshortened future” (GPR = 57.1%, IPR = 50%, VR = 50%) and “avoidance of thoughts, feelings, and conversation” (GPR = 52.4%, IPR = 50.0%, VR = 50.0%). Furthermore, Group D presented more intense variation among both items and rescuers categories. The highest item for GPR and IPR was the “exaggerated startle response” (61.9 and 60%, respectively), while it was the lowest among VR (0%).

Table 3 Frequency of PTSD symptoms among rescuers with probable PTSD
Full size table

Interaction between variables and their impact on probable PTSD was tested through a multivariate logistic regression model and is presented in Table 4. GPR demonstrated the highest risk for probable PTSD diagnosis (OR 3.4, 95% CI 1.9–4.8). Females had approximately two times higher chance for probable PTSD diagnosis (95% CI 1.1–3.4). The risk also increased proportionally to age (OR 3.8, 95% CI 2.5–5.1). Other significant risk factors were the marital status (single/divorced/widower OR 3.5, 95% CI 2.3–4.7) and the number of children (≤1 OR 1.6, 95% CI 10–2.1). Low frequency of previous similar experience increased the risk of PTSD diagnosis (OR 1.8, 95% CI 1.1–2.5), and longer operation period (OR 2.3, 95% CI 1.4–3.2), longer shift hours (OR 3.9, 95% CI 3.1–4.7), collection of dead refugees (OR 3.4, 95% CI 2.3–4.5) and collection of dead children (OR 3.2, 95% CI 1.9–4.4) were also considered as major risk factors.

Table 4 Risk factors for probable PTSD among rescuers in Lesvos
Full size table

In addition to the above results, it is worth mentioning that the significant majority of GPR, IPR, and VR diagnosed with probable PTSD (97%, p = 0.02) stated that they felt the need for psychological support during their interventions in Lesvos. The majority of the IPR (95%) were offered psychological support (individually or as group therapy) within the context of their foundation (p = 0.03). In contrary, GPR were not aware of (60%) or were not offered (38%) such services from their foundation (p = 0.03).

Discussion

The majority of the studies on PTSD development after traumatic events are focusing on victims, whereas limited research has been conducted on rescue workers. The current study is the first to assess PTSD prevalence among rescue workers that are operating in high-pressure spots during the European refugee crisis.

The overall probable PTSD prevalence found was 17.1% and ranged from 11.8% (IPR) to 23.1% (GPR). The rates observed were higher than the worldwide pooled prevalence of PTSD in rescue workers (10%) [17]. Nevertheless, the observed high frequency of probable PTSD in GPR is similar to the results of a previous study in a group of search and rescue workers in Turkeys’ earthquake, where PTSD estimation was 25% [6].

Results revealed that Greek rescuers presented higher risk for probable PTSD diagnosis, most likely due to their exhausting working conditions (longer operation period and longer duration of shifts or double shifts) and lack of continuous psychological support. IPR seemed to have a more organized, comprehensive, and continuous care provided by their foundation. These results come in accordance with previous studies [11, 18] identifying longer operation period and lack of mental health services as risk factors that are highly associated with PTSD development.

In all rescue workers categories, increasing age was associated with higher probable PTSD diagnosis. Similarly to the present study, a systematic review [17] highlighted increasing age as a variable that can affect PTSD prevalence among rescuers. In our study, female rescue workers presented significantly higher probability of PTSD development. Although this finding is consistent with studies reporting PTSD prevalence in the general population [26–28] and a limited number of studies among rescue workers [29, 30], it is counterintuitive. Many studies on rescue workers have been unable to demonstrate this relationship [17, 31–33]. The lack of association could be influenced by the fact that males represent approximately 85% of the participants in the 3/4 of similar studies [17]. In the literature, it is supported that gender differences in developing PTSD is a result of differences in peri-traumatic emotion, which influence subsequent PTSD [34]. Our findings highlight the need to pay more attention to women rescuers and create targeted interventions to reduce their psychological burden.

Exposure to dead refugees and children were found to affect significantly probable PTSD levels among the enrolled rescue workers. Literature supports that rescue workers who encounter deceased victims are at greater risk of both acute and chronic PTSD [35, 36]. In addition, handling dead children bodies is considered a particularly traumatic event for rescue workers that relates to negative psychological effects and can trigger PTSD reactions [7].

The study findings revealed that living with no partner or having no or one child was identified as significant risk factor for probable PTSD. According to literature, social support from colleagues and family plays an important protective role in the development of PTSD. More specifically, lack or low levels of perceived social support in people who have been exposed to traumatic events was found to be one of the strongest predictors of PTSD [7, 37]. Although many studies have highlighted the importance and substantial role of occupational social support (including sense of community, collaboration and cooperation) in combating PTSD [38–41], there is limited evidence on the effect of family support. A previous study on rescue workers operating at the World Trade Center site revealed that greater family support sources (including spouse/partner, children and parents) acted as protective factors to PTSD [42]. Likewise, other studies have concluded that increased support from family, perceived social support, and marital satisfaction were associated with less post-traumatic stress symptoms [43, 44]. Similar to our study, unmarried status among ambulance personnel have been associated with increased risk for PTSD [45].

In addition, lack of previous experience in similar rescue interventions, which is mainly encountered among volunteers, was significantly associated with higher risk of probable PTSD diagnosis. This result comes in accordance with previous studies indicating that volunteers with limited prior disaster training or experience are at greatest risk of PTSD [11, 18, 36, 46]. Although the present results agree partially with the literature (VR presented higher probable PTSD prevalence compared to IPR), it is of notice that GPR demonstrated the highest frequency of probable PTSD.

PTSD prevention in rescue workers

Due to the nature of their occupation, rescue workers often acquire some important psychological resources and protective mechanisms that provide a buffer against potential stressors. This has a positive impact on rescuers who usually present high levels of occupational satisfaction, deriving from the knowledge of doing something useful to help other people [47]. In addition, individual characteristics such as occupational satisfaction, optimism and hardiness, as well as organizational protective factors including social support, post-traumatic growth, and adaptive flexibility impact on keeping post-traumatic risk under control [7, 48].

Nevertheless, even if rescue workers dispose some individual and organizational protective factors, it is important to determine whether they need psychological support. The main focus should be given in what can be done to boost effective performance of rescue workers, while at the same time reducing vulnerability to secondary trauma and increasing resilience to stressors.

Shift rotations that enable a shorter duration of service at the site, as well as disaster preparedness training could help reduce PTSD among professional and volunteer rescuers in the future on-going operations. There is also great need for enhancing pre-employment strategies to identify the most resilient individuals for rescue interventions, to implement comprehensive preventive measures and to raise awareness toward PTSD.

Special focus should be given on implementing continuous, accessible, and effective preventive measures for PTSD, prior to a traumatic event [49]. Toward this direction, incorporation of collaborative care along with cognitive behavioral therapy in the context of work environment is strongly suggested. Another essential intervention, especially among GPR, is to organize educational campaigns about PTSD to help rescuers become more aware of this disorder and make them more comfortable to talk about it and seek treatment.

Strengths and limitations

To the best of our knowledge, this is the first study that attempted to explore factors that influence PTSD among three different categories of rescuers handling refugees and migrants in danger. The major strength of this study was that it achieved high response rates per rescuers’ category and revealed probable PTSD prevalence rates that are strongly supported by the literature. This provided confidence in the prevalence estimates and enabled comparisons within the under study groups, as well as across similar studies. In addition, all interviews were conducted by a specialized sociologist in quantitative research, while efforts were made to not affect rescuers’ daily tasks.

Nevertheless, results should be interpreted within the context of some limitations. The current study did not perform clinical interviews rather utilized a validated and worldwide used tool (PCL-C) to identify probable PTSD. This was an unavoidable limitation since clinical interviews were not feasible due to the participants’ workload. Although this method is often used in the literature [50], it is considered not to have optimal specificity, leading to overestimations.

Additionally, the findings may not be applicable to all disaster scenarios, due to the fact that the present study was conducted in a single geographical area that suffers from a specific event. Nevertheless, no major deviations are expected to be observed in events similar to the refugee crisis, since rescue workers at Lesvos have handled a vast number of victims.

Conclusions

The present study supports that rescue workers who operate at the island of Lesvos, providing substantial aid to the refugees and migrants, experience significant psychological distress. The excess burden of probable PTSD among rescue workers and the identified risk factors indicate the urgent need for targeted interventions, especially among GPR. The results are expected to contribute to the development and implementation of appropriate interventions that will reduce the psychological burden. Further studies are necessary to address long-term effects of the refugee crisis to rescue workers, as well as other relevant issues such as the prevalence of partial PTSD and burnout, and explore the most effective measures to prevent PTSD among this population.