Introduction
Definitions
Terrorist attacks can be defined by their intent to intimidate a larger population more so than the immediate victims. The action must be outside the context of legitimate warfare activities and directed toward achieving a political, economic, religious or social goal.1 According to the Global Terrorism Database, over the past 10 years, explosives have been the most commonly used method of terrorist attack, followed by firearms, incendiary weapons and melee weapons.1 The type of attack also has an impact on the tasks of the emergency services on site. This is relevant for mission preparation, as well as for their own personal safety and risk assessment.Reference Regehr, Carey, Wagner, Alden, Buys and Corneil2
The professional groups deployed usually include firefighters, police officers and rescue services, which are supplemented by other security forces, depending on the type of attack and the country affected.Reference Benedek, Fullerton and Ursano3 The psychological consequences of terrorist attacks are more severe than those of natural disasters for emergency services personnel and the general population.Reference Bromet, Atwoli, Kawakami, Navarro-Mateu, Piotrowski and King4 In addition to the personal involvement and event itself, evaluation processes play an important role in mental processing. This explains why the civilian population directly affected is the most heavily burdened, followed by the deployed emergency services personnel and then the general population.Reference Galea, Nandi and Vlahov5
Occupational and gender differences
Various studies have shown that emergency services personnel from different professions also react different emotionally to a major attack.Reference Bromet, Hobbs, Clouston, Gonzalez, Kotov and Luft6–Reference Rosiello, Desideri, Vinci and Zelinotti8 A recently published systematic review shows that police officers have a higher resilience to post-traumatic stress disorder (PTSD) and post-traumatic stress symptoms (PTSS) because of their training.Reference DiMaggio and Galea9
Studies on gender differences indicate a higher burden on female emergency services personnel.Reference Darves-Bornoz, Alonso, de Girolamo, de Graaf, Haro and Kovess-Masfety10–Reference Wesemann, Mahnke, Polk and Willmund13 Regional differences were also identified. A meta-analysis reported higher prevalence rates of PTSD in Western Europe compared with North America.Reference DiMaggio and Galea9 In contrast, another systematic review found that studies from Asia have a higher prevalence of PTSD than studies from Europe, but a lower prevalence than studies from North America.Reference Berger, Coutinho, Figueira, Marques-Portella, Luz and Neylan14 These differences and contradictions support the above-mentioned hypothesis that there are other relevant factors that play a role in mental processing besides the terrorist event.
Risk factors and crisis intervention
Resilience, psychiatric history and current mental health are some of these identified risk factors.Reference Maercker, Ben-Ezra, Esparza and Augsburger15–Reference Baubet and Rezzoug19 The physical proximity to the critical event has also been shown to influence later symptoms, such as mood, anxiety or sleep disorders.Reference Grenon, Consigny, Lemey, Simson and Coulon20,Reference Motreff, Baubet, Pirard, Rabet, Petitclerc and Stene21 This underscores that the spectrum of symptoms is much broader than that of PTSD, ranging from sleep disorders to cognitive, emotional, physical and behavioural problems, to substance use disorders.Reference Danker-Hopfe, Sauter, Kowalski, Kropp, Ströhle and Wesemann22–Reference Trautmann, Reineboth, Trikojat, Richter, Hagenaars and Kanske28
After the terrorist attack at Berlin's Breidtscheidpatz, it became clear that there was an urgent need for new crisis intervention measures. These contemporary measures must also be able to be carried out by semi-professional providers and must not contain psychotherapeutic or pharmacotherapeutic elements exclusively.Reference Wesemann, Mahnke, Polk, Bühler and Willmund29 To develop specific prevention tools, it is important to record and assess the range of mental health symptoms following critical incidents.Reference Benedek, Fullerton and Ursano3 Risk and resilience factors must also be taken into account.
Purpose
The purpose of this review is to provide information about the psychological impact on emergency services personnel involved in intentionally induced high-stress and traumatic situations. The following questions are of special interest:
(a) What psychological effects can be expected on rescue workers after terrorist attacks?
(b) Are there gender or occupational differences in mental health outcomes after terrorist attacks?
(c) How do rescue workers progress psychologically after a terrorist attack?
After reviewing the search results, the following question was added:
(a) Which risk and resilience factors affect mental health in this context?
Method
This review was designed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol for systematic reviewsReference Moher, Shamseer, Clarke, Ghersi, Liberati and Petticrew30 and registered with the International Prospective Register of Systematic Reviews (PROSPERO; identification number CRD42021241230). Ethics approval and informed consent were not required for this systematic review.
Search strategy
A PubMed and PubPsych search was carried out with the following search terms up to 7 February 2021: ((((terrorist attack) OR (amok[Title/Abstract]))) AND (((mental health[Title/Abstract]) OR (ptsd[Title/Abstract])) OR (stress[Title/Abstract]))) AND (((((police[Title/Abstract]) OR (firefighter[Title/Abstract])) OR (first line[Title/Abstract])) OR (emergency responder[Title/Abstract])) OR (service personnel[Title/Abstract])). A second search took place on 27 August 2021. The publication date was unlimited, and only publications in English were taken into account.
Eligibility
The titles and abstracts of the identified articles were then checked for inclusion criteria. Only peer-reviewed studies with a real terroristic or amok incident were included, as well as those with an examination of the mental health of the emergency responders. Trials or simulations of terroristic events were excluded. Both cross-sectional and longitudinal studies with their own data collection were included; reviews and meta-analyses without their own data collection were excluded. The positively screened articles were then included in the qualitative analysis.
Data extraction
The data synthesis was divided into four different groups: emergency service personnel, gender, symptoms and time after the event. The literature was evaluated according to the PICO scheme (patient/population, intervention, comparison and outcomes) for reviews. The professional groups (P) included police officers, firefighters, rescue services, non-profit organisations (NGOs) and providers of psychosocial services. The exposure group (I) were professional first responders deployed during or after a terrorist attack or rampage. Since these incidents are examined almost exclusively within a naturalistic study design, the comparison group (C) was defined as emergency personnel from the same occupational groups that were not used for the incident. Mental health effects of the emergency service personnel were defined as the result (O). The effect measures of the main results are prevalence rates (in percent) and odds ratios, if applicable. One researcher (U.W.) conducted the literature search and data extraction. A second researcher (H.H.) conducted the literature search independently, and disagreements were resolved with the consent of both. All emergency services personnel related mental health outcomes were extracted. Data compared with control groups other than those specified above were not taken into account in the results. No special software was used for the review. The result data were descriptively collected in Microsoft Excel for Windows 10. Table 1 gives an overview of the literature research and the integration into the PICO scheme.
Table 1 Characteristics of the included studies that report on the mental health of emergency responders after terrorist attacks
All studies have a naturalistic design (terrorist attack). WTC, World Trade Center; SF-8 Health Survey, Short-Form health-related quality of life; THQ, Trauma History Questionnaire; PDEQ, Peritraumatic Dissociative Experiences Questionnaire; ASD, acute stress disorder; PCL, PTSD Checklist; PTSD, post-traumatic stress disorder; PTSS, post-traumatic stress smptoms; PHQ-8, Patient Health Questionnaire-8 depression scale; GAD-7, Generalised Anxiety Disorder scale; PCL-S, PTSD Checklist-Specific Stressor version; DIS, Diagnostic Interview Schedule; PHQ-9, Patient Health Questionnaire-9 depression scale; CAGE, cut-annoyed-guilty-eye (alcohol); SDS, Sheehan Disability Scale; DTS, Davidson Trauma Scale; MINI, Mini International Neuropsychiatric Interview; GAD, generalised anxiety disorder; MSS, Mississippi Scale; CAPS, Clinically Administered PTSD Scale; WEI, Work Environment Inventory; SOS, Sources of Support Scale; PDI, Peritraumatic Distress Inventory; LSC-R, Life Stressor Checklist-Revised; TAS-20, Toronto Alexithymia Scale; PCL-C Posttraumatic Stress Checklist-Civilian Version; NGO, nongovernmental organisation; GHQ, General Health Questionnaire; MSSS, Modified Social Support Survey; STRS, Shortness of Breath, Tremulousness, Racing heart and Sweating scale; CGI, Clinical Global Impression questionnaire; BSI, Brief Symptom Inventory; WHOQOL-BREF, World Health Organisation Quality of Life.
Risk-of-bias assessment
A major selection bias in these naturalistic studies is the involvement of the participants. To control for this, we checked the literature for recruitment strategies (e.g. participants who have come to an aid programme of any kind, evaluation of intervention strategies, etc.) or whether all persons in the total sample had the same chance of being included. This information was taken into account in the qualitative evaluation of the results. In addition, we checked whether the same personnel were included in different studies. When evaluating the results, possible overlaps of personnel participating in multiple studies were taken into account. The risk assessment was carried out by one author according to the Quality Assessment Tool for Quantitative Studies, from the Effective Public Health Practice Project Quality Assessment Tool,Reference Armijo-Olivo, Stiles, Hagen, Biondo and Cummings62 and assessed by a second author. Disagreements were resolved across the study group. The ‘study design’ component was reassessed, as almost all studies were rated as ‘weak’ because of the naturalistic nature of these studies.
This assessment was done in accordance with the suggestions of the Evidence-based Practice Center of the Agency for Healthcare Research and Quality (AHRQ). It states that the assessment should take into account various sources of design bias rather than assessing individual studies for the study design itself. This would put studies at a disadvantage if they were downgraded by the risk of bias solely based on the study design.Reference Reeves, Deeks, Higgins, Shea, Tugwell, Wells, JPT, J, J, M, T and MJ63 This method was also chosen based on the Newcastle-Ottawa Scale to assess the quality of non-randomised studies.Reference Wells, Shea, O'Connel, Peterson, Welch and Losos64 The Newcastle-Ottawa Scale is a not yet established assessment tool for cohort studies. For this reason, we have retained the ‘Quality Assessment Tool for Quantitative Studies’ criteria for this adaptation. According to the AHRQ guideline, we upgraded the study design from weak to moderate if they had large, consistent effects.
Of the 33 included studies, 11 were rated as strong, 19 as moderate and three as weak, as shown in Table 2.
Table 2 Quality assessment of the included studies according to the Quality Assessment Tool for Quantitative Studies
A, selection bias; B, study design; U/D, upgrade/downgrade study design; C, confounders; D, blinding (not applicable); E, data collection; F, drop-out; ./., no data available or not applicable due to study design; G, global assessment; GN, global assessment after upgrade/downgrade (1, strong; 2, moderate; 3, weak); GN: 1 = no weak rating; 2 = one weak rating; 3 = two or more weak ratings.
Strategy for data synthesis
Because of the small number of included studies, no minimum was specified for defining the synthesis. A bias in the selection of recruitment (as described above) was taken into account for the final synthesis. Otherwise, all prevalence data on mental disorders were given for the individual occupational groups. The synthesis was carried out according to the ‘Guidance on the Conduct of Narrative Synthesis in Systematic Reviews’.Reference Popay, Roberts, Sowden, Petticrew, Britten and Arai65 Finally, all first authors with more than two of the included and analysed studies were consulted to verify the validity of the synthesis, which was modified accordingly.
Results
Data collection
The initial and further literature search resulted in 56 papers. Eleven were excluded because they were not related to rescue workers, three did not report their own data, two had no focus on mental health or were a meta-analysis, one was excluded because it was not in the English language and one was excluded because the results were not broken down by professional groups. As a result, 33 full-text articles were included for primary analysis.
Terrorist attacks
The most comprehensive work concerned the terrorist attacks on the World Trade Center in New York, USA on 11 September 2001 (hereafter referred to as 9/11 WTC; n = 25 publications); followed by the terrorist attack on the Christmas market at Breitscheidplatz in Berlin, Germany on 19 December 2016 (n = 3); the terrorist attacks in Paris, France in 2015 (n = 2); the terrorist attack in Madrid, Spain on 11 March 2016 (n = 1) and the terrorist attack on Nasiriyah, Iraq in April 2006 (n = 1). The inclusion process is presented in Fig. 1.
Fig. 1 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram.
Descriptive statistics of the studies
The publication dates range from 2006 to 2021, with mean and median years between 2011 and 2015. Almost all studies included police officers (n = 30), followed by firefighters (n = 6), paramedics and NGOs (each n = 5), psychosocial workers (n = 3) and search and rescue workers (n = 2). There is a wide range of participants of each included study, ranging from 28 to 28 232 participants. The data was collected between 2 weeks and 20 years after the incidents. The exact time of data collection was not given in two studies. Almost half of the studies were cross-sectional (n = 17) or longitudinal (n = 16). Only three studies included a control group from the same occupational group that was not used in the incident.
Prevalence rates within the first 3 months after the terrorist attack
An acute stress disorder (ASD) was identified in 14.6% of the emergency services (including 90 medical personnel, police officers and firefighters) 2–3 weeks after the 9/11 WTC attacks. Depression (results from questionnaires) was reported in 25.8% of the cases, high psychological impairment in 33.3% and at least one peritraumatic dissociation symptom in 83.3%. The comorbidity of ASD and depression was 10.9%. The prevalence rate of ASD was highest among tobacco users, at 33.3–37.9%.Reference Bowler, Han, Gocheva, Nakagawa, Alper and DiGrande32 At 5–12 weeks after the 11 March 2004 terrorist attacks in Madrid, the prevalence of any mental disorder was 3.9% for 153 police officers, including 1.3% with PTSD (with intrusive ideas as the most common symptom), 1.3% with major depression, 0.7% with agoraphobia, 0.7% with generalised anxiety disorder and 0.7% with panic disorders. Furthermore, 51% of the police officers with a mental disorder suffered from another comorbid mental disorder.Reference Gabriel, Ferrando, Cortón, Mingote, García-Camba and Liria42
Prevalence rates within the first year after the terrorist attack
At 6–9 months after the January 2015 Paris terrorist attacks, the PTSD prevalence among police officers was between 3% (interview) and 8% (questionnaire). Other police officers reported at least one anxiety disorder (14%), an inability to work after the attacks (6%) and an increase in alcohol, tobacco or cannabis use (9%).Reference Vandentorren, Pirard, Sanna, Aubert, Motreff and Dantchev54 At 8–12 months after the attacks, the PTSD and PTSS rate was 3.4 and 15.7% in firefighters (n = 203), 4.4 and 10.4% in health professionals (n = 229), 4.5 and 19.4% in personnel of NGOs (n = 134) and 9.5 and 23.2% in police officers (n = 95). The spatial proximity to the unsecured crime scenes increased the risk of having PTSD (odds ratio 7.26).Reference Motreff, Baubet, Pirard, Rabet, Petitclerc and Stene21
Prevalence rates during the first and second years after the terrorist attack
At 1–2 years after the 9/11 WTC attacks, 28 232 notes from consultations of police officers were evaluated. The notes evaluated were from unstructured interviews done by paraprofessionals. The prevalence of behavioural problems (e.g. hypervigilance, isolation, drug use or violent behaviour) was found in 34% of police officers, emotional problems (e.g. anger, sadness, anxiety or hopelessness) in 56%, physical problems (e.g. fatigue, headaches, stomach problems or chronic agitation) in 51% and cognitive problems (e.g. concentration, intrusive thoughts, decision-making or preoccupation with death) in 27%.Reference Feder, Mota, Salim, Rodriguez, Singh and Schaffer41
Prevalence rates during the second and third years after the terrorist attack
At 2–3 years after the 9/11 WTC attacks, the prevalence rate of PTSD (results from questionnaires) in 4017 police officers was 7.8%.Reference Bowler, Han, Gocheva, Nakagawa, Alper and DiGrande32 Emotional numbness was found in 6.7%.Reference Bowler, Harris, Li, Gocheva, Stellman and Wilson33 Another study referring to the same event found PTSD prevalence rates (results from questionnaires) for 3925 police officers (7.2%), 3232 firefighters (14.3%), 1741 emergency medical service (14.1%) and 5438 personnel from NGOs (8.4%).Reference Perrin, DiGrande, Wheeler, Thorpe, Farfel and Brackbill48
Prevalence rates during the fourth and fifth years after the terrorist attack
At 4 years after the 9/11 WTC attacks, 8508 police officers were screened for PTSD, and the prevalence (results from questionnaires) was 5.9%.Reference Luft, Schechter, Kotov, Broihier, Reissman and Guerrera45 In the second wave of the study,Reference Bowler, Harris, Li, Gocheva, Stellman and Wilson33 5–6 years after the 9/11 WTC attacks, 2940 police officers were screened and the prevalence rate climbed significantly, to 16.5%. Emotional numbness also climbed from 6.7 to 15.6%.Reference Bowler, Kornblith, Li, Adams, Gocheva and Schwarzer34 A total of 8466 police officers were screened an average of 4 years after the 9/11 WTC attacks. The 1-month prevalence (interview) of PTSD was 5.4% and prevalence of PTSS was 15.4%.Reference Pietrzak, Schechter, Bromet, Katz, Reissman and Ozbay51 Among 1915 retired firefighters 4–7 years after the 9/11 WTC attacks, the prevalence rate of PTSD was similarly high, with 6% affected (interview). Retirement was the result of a disability related to the attacks in 64% of respondents.Reference Bowler, Kornblith, Li, Adams, Gocheva and Schwarzer34 Five years after the 9/11 WTC attacks, 6777 police officers were subjected to non-standardised interviews. The prevalence of any kind of mental disorder was 17.4%.Reference DePierro, Lowe, Haugen, Cancelmo, Schaffer and Schechter38
Prevalence rates 6–10 years after the terrorist attack
An average of 6.5 years after the 9/11WTC attacks, 8881 police officers were interviewed and completed questionnaires as part of the first annual health surveillance visit with the WTC Health Program; 20.6% of the participants required mental healthcare.Reference Diab, DePierro, Cancelmo, Schaffer, Schechter and Dasaro39 Nine to 10 years after the attacks, the prevalence rate of PTSD (results from questionnaires) for 2204 police officers was 11.0%.Reference Cone, Li, Kornblith, Gocheva, Stellman and Shaikh37 Another study 9–10 years after the attacks reported a 12.9% PTSD prevalence (results from questionnaires) in police officers. The comorbidity categories of those affected included depression (24.7%), anxiety (5.8%), both (47.7%) and none (21.8%).Reference Bowler, Kornblith, Li, Adams, Gocheva and Schwarzer34
Prevalence rates more than 10 years after the terrorist attack
In a web-based survey 11–13 years after the 9/11 WTC attacks, including 2029 police officers, the incidence rate of PTSD (results from questionnaires) was 9.3% and the incidence rate for PTSS was 17.5%.Reference Chen, Salim, Rodriguez, Singh, Schechter and Dasaro35
Gender and occupational differences
Three months after the terror attack at Berlin's Breitscheidplatz in 2016, there was more reported stress and paranoid ideation in female responders, less environmental and physical quality of life in firefighters and more hostility in police officers.Reference Wesemann, Zimmermann, Mahnke, Butler, Polk and Willmund56 A confirmation study 3–4 months and 2 years after the terror attack replicated more paranoid ideation in female responders at both time points. Paranoid ideation increased significantly over time in exposed female personnel compared with controls.Reference Wesemann, Mahnke, Polk and Willmund13 Also, more hostility was displayed in police officers at both time points, as well as less environmental quality of life in firefighters at both time points, compared with controls. Physical quality of life of the firefighters was no longer impaired 2 years after the incident.Reference Wesemann, Bühler, Mahnke, Polk and Willmund55 At 2–3 years after the 9/11 WTC attacks, gender differences in the prevalence rates of PTSD (results from questionnaires) in 4017 police officers were found, with 13.9% in women and 7.4% in men.Reference Bowler, Han, Gocheva, Nakagawa, Alper and DiGrande32 Four years after the 9/11 WTC attacks, no significant correlation between gender and PTSD (r = 0.05) was found in 8508 police officers.Reference Luft, Schechter, Kotov, Broihier, Reissman and Guerrera45 In the second wave of this study, 2940 police officers were surveyed 5–6 years after the 9/11 WTC attacks. No gender-specific differences were found in the PTSD prevalence. Nevertheless, there were more intrusive symptoms in women.Reference Bowler, Harris, Li, Gocheva, Stellman and Wilson33 At 9–10 years after the 9/11 WTC attacks, in the remaining 2204 police officers, the gender differences in PTSD became prevalent again, with 15.5% in women and 10.3% in men.Reference Cone, Li, Kornblith, Gocheva, Stellman and Shaikh37
Risk factors for mental disorders
Risk factors for PTSD were found 8–12 months after the November 2015 terrorist attacks in Paris. These risk factors include female gender (odds ratio 1.3), no training on psychological risks of this type of event (odds ratio 4.88), critical life events 1 year before the incident (odds ratio 1.7) and social isolation (odds ratio 8.4).Reference Motreff, Baubet, Pirard, Rabet, Petitclerc and Stene21 Shortly after the 9/11 WTC attacks, there was a correlation between alexithymia and PTSD (r = 0.35; P < 0.001) in 166 exposed police officers. Additionally, psychological distress 3 years after the 9/11 WTC attacks predicted gastrointestinal symptoms 6 years after the incident.Reference Litcher-Kelly, Lam, Broihier, Brand, Banker and Kotov44
For those with PTSS, the best predictors were lack of social desirability, alexithymia, work environment stress, lack of social support, peritraumatic distress and peritraumatic dissociation.Reference McCaslin, Metzler, Best, Liberman, Weiss and Fagan47 General risk factors in 4017 police officers 2–3 years after the 9/11 WTC attacks were event-related injury and older age.
There were several gender-related risk factors identified. Special risk factors for men were proximity to the event and Hispanic ethnicity. In women, risk factors included witnessing the horrific event and an education level below a college degree.Reference Bowler, Han, Gocheva, Nakagawa, Alper and DiGrande32
Additionally, risk factors were identified and split by personnel. Maintaining an injury was the top risk factor for PTSD among firefighters, emergency services and NGO workers 2–3 years after the 9/11 WTC attacks.Reference Perrin, DiGrande, Wheeler, Thorpe, Farfel and Brackbill48 In 1915 retired firefighters 4–7 years after the 9/11 WTC attacks, there was a 50% decrease in the likelihood of retirement because of a disability with each increase in age of 10 years.Reference Chiu, Webber, Zeig-Owens, Gustave, Lee and Kelly36
In 8466 police officers an average of 4 years after the 9/11 WTC attacks, the risk factors for PTSD and PTSS were older age, the total number of exposures, exposure to human remains, caught in a dust cloud, involved in the early stages of search/rescue, known someone who had died in the attacks, known someone who suffered an injury in the attacks, and the number of stressors before and after the incident.Reference Pietrzak, Schechter, Bromet, Katz, Reissman and Ozbay51
In a web-based survey of 2029 police officers 11–13 years after the 9/11 WTC attacks, risk factors for PTSS were post-event medical comorbidities and other traumatic events.Reference Chen, Salim, Rodriguez, Singh, Schechter and Dasaro35 An average of 6.5 years after the 9/11 WTC attacks, risk factors for any perceived mental health service need were assessed in 8881 police officers. Risk factors included old age; female gender; a diagnosis of depression, anxiety and/or PTSD before the attacks; positive screens for PTSD related to the attacks; alcohol misuse; several somatic diagnoses; functional impairment in work and more life stressors since the attacks.Reference Diab, DePierro, Cancelmo, Schaffer, Schechter and Dasaro39
In a longitudinal study with measurements 3, 6 and 8 years after the 9/11 WTC attacks, 4035 police officers were assessed. Risk factors for the severe chronic trajectory were exposure to human remains, on-site somatic injuries/illnesses, acquaintances with injuries and the traumatic death of a colleague, friend or family member.Reference Pietrzak, Feder, Singh, Schechter, Bromet and Katz49 At 9–10 years after the 9/11 WTC attacks, low social support was a risk factor for PTSD in 2204 police officers. The prevalence of PTSD increased with the number of life stressors that occurred within the past 12 months.Reference Cone, Li, Kornblith, Gocheva, Stellman and Shaikh37
Trajectories and temporal course
In a longitudinal study with measurements 3, 6 and 8 years after the 9/11 WTC attacks, 4035 police officers were included. Four trajectories for PTSD symptoms were found, including no/low symptoms (77.8%), worsening symptoms (8.5%), symptom improvement (8.4%) and a chronic (5.3%) trajectory.Reference Pietrzak, Feder, Singh, Schechter, Bromet and Katz49 The same study with the remaining 1874 police officers and an additional assessment 12 years after the 9/11 WTC attacks was performed. Four trajectories for PTSD symptoms were found, which included no/low symptoms (76.1%), worsening symptoms (12.1%), symptom improvement (7.5%), and a chronic (4.4%) trajectory.Reference Feder, Mota, Salim, Rodriguez, Singh and Schaffer41 Another longitudinal study 1–2, 6–7 and 9–10 years after the 9/11 WTC attacks included 2204 police officers. The exposure event was correlated with PTSD and stress symptoms over time. The peak symptom trajectory was found in wave 2. The symptoms were negatively related to emotional support in wave 3.Reference Schwarzer, Cone, Li and Bowler53 In a yearly study from year 1 to year 9 after the 9/11 WTC attacks, 12 273 protective service or military personnel were included. The cumulative incidence of depression was 7.0%, cumulative incidence of PTSD was 9.3% and cumulative incidence of panic disorders was 8.4%. The study found a steady yearly increase in the incidence rates from year 1 to year 9. Depression rose from 1.7 to 7.0%, PTSD rose from 2.5 to 9.3% and panic disorders rose from 2.3 to 8.4%.Reference Wisnivesky, Teitelbaum, Todd, Boffetta, Crane and Crowley57 Another longitudinal study with 12 398 rescue workers at 2–3, 5–6 and 9–10 years after the 9/11 WTC attacks showed the opposite results. The PTSD prevalence (results from questionnaires) decreased from wave 1 (19.6%) to wave 2 (18.2%) and in wave 3 (16.7%). The sample included police officers, firefighters, paramedics, technicians, plumbers and volunteers. However, there is no differentiation between these subgroups.Reference Wyka, Friedman and Jordan58 Another year 1 to year 9 longitudinal study after the 9/11 WTC attacks included 8466 police officers. A greater initial disaster exposure led to more PTSS and a decrease in overall functioning over time. Post-disaster life stressors amplified these negative effects. These effects were not found in non-traditional WTC-responders.Reference Zvolensky, Kotov, Schechter, Gonzalez, Vujanovic and Pietrzak59 In this study, PTSD, depressive symptoms and overall functioning were stable over the follow-up period.Reference Zvolensky, Farris, Kotov, Schechter, Bromet and Gonzalez61 In the same study, a subgroup analysis was done for 763 police officers who were ex-smokers. Higher initial PTSS were associated with a decreased likelihood of smoking abstinence and decreased smoking reduction. Hyperarousal also indicated a decreased likelihood of abstinence over time.Reference Zvolensky, Farris, Kotov, Schechter, Bromet and Gonzalez60
Kotov et al found an increase in the 1-month prevalence rate of PTSD (results from questionnaires) in 8466 police officers from 6.3% of responders an average of 4 years after the 9/11 WTC attacks. This prevalence rate increased to 8.2% an average of 2.5 years after the first assessment.Reference Kotov, Bromet, Schechter, Broihier, Feder and Friedman-Jimenez43
A longitudinal study investigated 2204 police officers 1–2, 6–7 and 9–10 years after the 9/11 WTC attacks. Exposure was correlated with symptoms of PTSD and stress over time.Reference Schwarzer, Cone, Li and Bowler53
Resilience factors and helpful support
After the 9/11 WTC attacks, there was evidence that peer officers can be used effectively to assist police officers with post-disaster stress.Reference Dowling, Moynihan, Genet and Lewis40 A similar result emerged after the Paris terrorist attacks in 2015. More than half of the rescue workers affected (51%) benefited from their institution's psychological help.Reference Vandentorren, Pirard, Sanna, Aubert, Motreff and Dantchev54 In a longitudinal study with measurements 3, 6, 8 and 12 years after the 9/11 WTC attacks, 1874 police officers were sampled. Positive emotional coping and current treatment were associated with the no/low symptom trajectory for PTSD.Reference Feder, Mota, Salim, Rodriguez, Singh and Schaffer41 In a longitudinal study with 2943 police officers 1–2 and 6–7 years after the 9/11 WTC attacks, social integration was identified as a resilience factor to PTSS.Reference Schwarzer, Bowler and Cone52
Stigma
About 5 years after the 9/11 WTC attacks, 6777 police officers were subjected to non-standardised interviews. Results showed there was less stigma surrounding psychiatric disorders in police officers compared with non-traditional responders (24.8 v. 41.2%). Nevertheless, it was found that police officers with psychiatric disorders have greater self-stigma than non-traditional responders with psychiatric disorders (17.9 v. 9.1%). ‘Trying to deal with it yourself’ was the most common reason for the reluctance to seek help.Reference DePierro, Lowe, Haugen, Cancelmo, Schaffer and Schechter38
Problems in daily life after the events
In the course of the next 9 years after the 9/11 WTC attacks, 8466 police officers were asked about critical post-disaster life events: job loss, layoff or substantial loss of income was reported in 11.3%; changes in living situation were reported in 17.3%; legal problems were reported in 4.0% and arrest was reported in 0.6%.Reference Zvolensky, Farris, Kotov, Schechter, Bromet and Gonzalez61 Military police officers with PTSD after a terrorist attack in Iraq showed significantly less emotional recognition and empathy, and problems interpreting other people's mental states and expressed emotions.Reference Mazza, Giusti, Albanese, Mariano, Pino and Roncone46
Test instruments
In a longitudinal study with measurements 3, 6 and 8 years after the 9/11 WTC attacks, 4035 police officers were included. In both the interviews and questionnaires, five stable symptom clusters best represented the PTSD symptom dimensionality in police officers. The symptom clusters included re-experiencing (e.g. intrusive thoughts), avoidance (e.g. avoidance of thoughts), numbing (e.g. trauma-related amnesia), dysphoric arousal (e.g. irritability) and anxious arousal (e.g. overly alert).Reference Pietrzak, Feder, Schechter, Singh, Cancelmo and Bromet50
Most studies used questionnaires rather than interviews. The most common instrument was the PTSD Checklist (PCL) with its variants. Comparing the PCL-S (specific version) and PCL-C (civilian version) between the same 8466 police officers, the prevalence rate for PTSD ranged from 5.4Reference Pietrzak, Schechter, Bromet, Katz, Reissman and Ozbay51 to 6.3%.Reference Kotov, Bromet, Schechter, Broihier, Feder and Friedman-Jimenez43
Discussion
The prevalence rates of mental disorders after a terrorist attack vary widely. For PTSD, the range is 1.3–16.5%, for major depression, it is 1.3–25.8% and for specific anxiety disorders, it is 0.7–14%.Reference Bowler, Harris, Li, Gocheva, Stellman and Wilson33,Reference Gabriel, Ferrando, Cortón, Mingote, García-Camba and Liria42,Reference Vandentorren, Pirard, Sanna, Aubert, Motreff and Dantchev54 This also applies to the prevalence rate in the same occupational group after the same incident. The largest discrepancy exists between questionnaires and clinical interviews, with higher rates of reported mental disorders in the questionnaires.Reference Vandentorren, Pirard, Sanna, Aubert, Motreff and Dantchev54 Since a psychiatric interview is the gold standard for diagnosis, the questionnaire results are believed to overestimate prevalence rates. This is attributed to the specificity. The lower the specificity of the questionnaires, the higher the overestimation. In the case of questionnaire results being utilised in future research, this should always be described in the limitations and, if possible, the positive predictive value should be given.
From an academic point of view, it would be desirable to have a special set of instruments for evaluating such extraordinary events. This would make the studies more comparable. For this purpose, we recommend any type of evaluated and standardised clinical interview that relates to the current versions of the ICD-11 or DSM-5. The interviews should be conducted by trained and specialised staff, and should cover the most common mental disorders. At a minimum, it should include depressive episodes, anxiety disorders, PTSD, psychosomatic disorders and substance misuse disorders. If such an interview cannot be conducted, validated questionnaires with high sensitivity and specificity are recommended.
The highest prevalence rates were found after the 9/11 WTC attacks. This is attributed to the particular nature of the terrorist attack: it comprised multiple incidents and there was a significant length of time after the attacks until the buildings collapsed. The likelihood of developing dissociative symptoms during this time is higher.Reference Bowler, Han, Gocheva, Nakagawa, Alper and DiGrande32 It is, therefore, not surprising that 2–3 after this attack, 83.3% of the disaster relief workers reported at least one peritraumatic dissociation symptom.Reference Biggs, Fullerton, Reeves, Grieger, Reissman and Ursano31
Given the wide variation of prevalence rates in studies covering the same occupational groups, it seems difficult to compare studies covering different occupational groups. To be able to make conclusive statements, only studies that include different professional groups and differentiated accordingly were considered. Three months after the Berlin terrorist attack in 2016, deployed firefighters, health professionals, NGO personnel and police officers were compared. Firefighters showed the least environmental and physical quality of life, whereas police officers showed more hostility.Reference Wesemann, Zimmermann, Mahnke, Butler, Polk and Willmund56 Aside from the physical quality of life, these results were replicated 2 years after this terrorist attack and were not found in the non-deployed comparison group.Reference Wesemann, Bühler, Mahnke, Polk and Willmund55 In the first year after the January 2015 Paris terror attacks, the PTSD point prevalence rate was 3.4% in firefighters, 4.4% in health professionals, 4.5% in NGO personnel and 9.5% in police officers.Reference Vandentorren, Pirard, Sanna, Aubert, Motreff and Dantchev54 During the second and third years after the 9/11 WTC attacks, this order was reversed. The point prevalence for PTSD (results from questionnaires) was 14.3% for firefighters, 14.1% for health professionals, 8.4% for NGO personnel and 7.2% for police officers.Reference Perrin, DiGrande, Wheeler, Thorpe, Farfel and Brackbill48
These results support the necessity to distinguish between the specific occupational groups. However, they also refer to the importance of the specific task of the emergency services personnel. For example, police officers who searched for perpetrators after the Paris attacks had the highest prevalence rates for PTSD compared with other occupations, whereas police officers who were more likely to conduct evacuations after the 9/11 WTC attacks had the lowest PTSD rates. This task-related hypothesis would also explain the occupational differences identified after the Berlin terrorist attack. Although firefighters concentrated on first aid and risk reduction at the destroyed Christmas market, police officers secured the crime scene from further attacks and searched for the perpetrator. A focus on destruction would then explain a reduction in environmental quality of life, whereas a focus on perpetrators would explain an increase in hostility.
A more consistent picture emerges when it comes to gender differences. If differences are found, these are at the expense of female emergency services personnel with more mental health symptomsReference Wesemann, Mahnke, Polk and Willmund13,Reference Bowler, Harris, Li, Gocheva, Stellman and Wilson33,Reference Wesemann, Zimmermann, Mahnke, Butler, Polk and Willmund56 or higher prevalence rates of mental disorders.Reference Bowler, Han, Gocheva, Nakagawa, Alper and DiGrande32,Reference Cone, Li, Kornblith, Gocheva, Stellman and Shaikh37 The more frequent harassment of minorities in male-dominated professions was given as an explanation for these results.Reference Wesemann, Mahnke, Polk and Willmund13 However, since such differences also occur among non-underrepresented female employees, other factors seem to play a more important role. Risk factors identified included lower education and witnessing horror.Reference Bowler, Han, Gocheva, Nakagawa, Alper and DiGrande32 A combination of situational factors (e.g. horror experiences) and personal factors (e.g. lower education) may account for these differences. In addition to age, ethnicity and level of education, it is therefore advisable to include marital status, relationships with colleagues and supervisors, and general job satisfaction in future studies. In addition to the duration and proximity to the event, the situation-specific aspects also include personal risk perception, preparation for the event, confidence in action, clarity of tasks on-site, quality of equipment, professional experience and media reports on the organisation's emergency services personnel (positive/negative).
However, screening for mental health symptoms (after a terroristic event) remains necessary to prevent or treat mental disorders. This could be done in a two-stage strategy, with an initial questionnaire and an additional interview for at-risk employees. Screening should always be voluntary to avoid simulation and dissimulation. The screening should also be repeated at regular intervals. The data from this review is inconsistent as to whether mental health problems increase or decrease over time. This is likely because of the different trajectories found in the longitudinal studies.Reference Feder, Mota, Salim, Rodriguez, Singh and Schaffer41,Reference Pietrzak, Feder, Singh, Schechter, Bromet and Katz49
This screening strategy could help reduce the high prevalence rate of mental disorders. The focus of the identified studies was clearly on PTSD. However, PTSD is just one of several mental disorders that can be exacerbated after a terroristic event. Depression, anxiety disorders, substance use disorders and psychosomatic disorders should also be considered. Most of the studies in which measurements were taken to detect these disorders showed increased prevalence rates.Reference Bowler, Kornblith, Li, Adams, Gocheva and Schwarzer34 Screening for life events before and after the incidents also appeared to be helpful.Reference Motreff, Baubet, Pirard, Rabet, Petitclerc and Stene21,Reference Litcher-Kelly, Lam, Broihier, Brand, Banker and Kotov44 These catastrophic events tend to increase vulnerability rather than resilience or post-traumatic growth.Reference Pietrzak, Schechter, Bromet, Katz, Reissman and Ozbay51,Reference Zvolensky, Kotov, Schechter, Gonzalez, Vujanovic and Pietrzak59 Another critical incident following the first exposure could more quickly lead to mental health problems.Reference Cone, Li, Kornblith, Gocheva, Stellman and Shaikh37
The main risk factors identified in the studies were female gender,Reference Motreff, Baubet, Pirard, Rabet, Petitclerc and Stene21,Reference Diab, DePierro, Cancelmo, Schaffer, Schechter and Dasaro39 no training on psychological risks of this type of event,Reference Motreff, Baubet, Pirard, Rabet, Petitclerc and Stene21,Reference Litcher-Kelly, Lam, Broihier, Brand, Banker and Kotov44 social isolation,Reference Cone, Li, Kornblith, Gocheva, Stellman and Shaikh37,Reference McCaslin, Metzler, Best, Liberman, Weiss and Fagan47 work environment stress,Reference Bowler, Han, Gocheva, Nakagawa, Alper and DiGrande32,Reference McCaslin, Metzler, Best, Liberman, Weiss and Fagan47 peritraumatic distress and peritraumatic dissociation,Reference Diab, DePierro, Cancelmo, Schaffer, Schechter and Dasaro39,Reference McCaslin, Metzler, Best, Liberman, Weiss and Fagan47 losing someone or maintaining an injury,Reference Perrin, DiGrande, Wheeler, Thorpe, Farfel and Brackbill48,Reference Pietrzak, Feder, Singh, Schechter, Bromet and Katz49 and older age.Reference Bowler, Han, Gocheva, Nakagawa, Alper and DiGrande32,Reference Chiu, Webber, Zeig-Owens, Gustave, Lee and Kelly36 This knowledge could be used in special training courses for emergency personnel dealing with catastrophic events. Each occupational group could implement prevention strategies for these risk factors. Failure to prepare for this type of emergency response resulted in an odds ratio of 4.88 for mental disorders.Reference Motreff, Baubet, Pirard, Rabet, Petitclerc and Stene21 The same applies to post-deployment strategies. Semi-professional psychological crisis interventions should focus on all types of mental symptoms. Knowing about specific risk factors could help to make them more target-group-specific. Institutional peer support has been shown to be effective after the 9/11 WTC attacks,Reference Dowling, Moynihan, Genet and Lewis40 and is probably the most appropriate way to deal with such huge incidents. In addition, destigmatisation programmes showed evidence of a reduction in psychological stressReference Rüsch, Corrigan, Waldmann, Staiger, Bahemann and Oexle66,Reference Schomerus, Stolzenburg, Bauch, Speerforck, Janowitz and Angermeyer67 and the need to implement them.Reference Mak, Chan, Wong, Lau, Tang and Tang68 This could be another important tool for mental health prevention, as the perceived stigma of the emergency workers involved is comparatively high.Reference DePierro, Lowe, Haugen, Cancelmo, Schaffer and Schechter38
There are some study limitations to report. Results may be limited because only two databases were used, only English-language literature was considered, results for female emergency services are limited because of the high proportion of males in emergency services personnel jobs, and the search strategy may have missed other events. Finally, the sample size for police officers was predominant out of all of the occupational groups identified.
In conclusion, given the trajectories and time course of mental disorders,Reference Feder, Mota, Salim, Rodriguez, Singh and Schaffer41,Reference Pietrzak, Feder, Singh, Schechter, Bromet and Katz49,Reference Schwarzer, Cone, Li and Bowler53 voluntary routine screening over an extended period of at least 7 years after the incident is strongly recommended. Otherwise, a notable part of the emergency services personnel could easily be missed in the ‘deterioration curve’ if mental symptoms appear later. If this screening is carried out with questionnaires, a clinical interview could confirm the emergency services personnel with a mental disorder diagnosis. This can minimise the rate of false-positive results.
Additionally, research using only questionnaires should always report the positive predictive value of the questionnaires. For questionnaires with low specificity that search for rare diseases, the rate of false positives can be many times higher than the rate of true positives.
The stigma associated with mental disorders remains a major barrier for emergency services personnel in seeking professional help.Reference DePierro, Lowe, Haugen, Cancelmo, Schaffer and Schechter38 Destigmatisation programmes should therefore be implemented or further expanded in all relevant institutions. There are already a number of established options.Reference Rüsch, Corrigan, Waldmann, Staiger, Bahemann and Oexle66–Reference Mak, Chan, Wong, Lau, Tang and Tang68 If mental damage caused by deployment is recognised as an accident at work and mental disorders are assessed by the employer as an occupational risk, this could have a positive effect on recovery.
The identified risk factors can be taken more into consideration when coordinating primary and secondary prevention. As a result, the programmes will be tailored more specifically to the needs of the organisations and those affected. Very general programmes, or even the lack thereof, are associated with poorer mental health outcomes.Reference Dowling, Moynihan, Genet and Lewis40,Reference Feder, Mota, Salim, Rodriguez, Singh and Schaffer41,Reference Vandentorren, Pirard, Sanna, Aubert, Motreff and Dantchev54
Further research on this topic is needed to better understand the interaction of risk factors for the development of mental disorders. A distinction must be made between situational factors such as proximity to the event, personal factors such as gender, a combination of both situational and personal factors such as personal risk perception, and occupational factors such as specific tasks and organisational factors such as training. Additionally, a distinguishment must be made between intentionally caused major events such as terrorist attacks and unintentionally caused events such as natural disasters, including the severity and duration of the incidents to effectively compare the impact of both types of events.
Data availability
Data availability is not applicable to this article as no new data were created or analysed in this study.
Acknowledgements
We thank Britta Jürgens from the Bundeswehr Hospital Berlin for supporting our work.
Author contributions
U.W. and H.H. designed and guided the study, performed the article screening, data extraction, analysis and interpretation of the results, and prepared the manuscript. B.A. designed and guided the study, analysed and interpreted the results and revised the manuscript. All authors discussed and agreed on the final submitted version of the manuscript.
Funding
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Declaration of interest
None.
Open access
eLetters
No eLetters have been published for this article.