Posttraumatic stress disorder in Belgian police officers: prevalence and the effects of exposure to traumatic events

ABSTRACT

Background: Police officers are at considerable risk of developing posttraumatic symptoms because they frequently encounter violent or emotionally disturbing incidents. We investigate experiences with potentially traumatic events (PTE), traumatic exposure, and the prevalence of probable posttraumatic stress disorder (PTSD), complex PTSD and subclinical PTSD in a sample of Belgian police officers.

Methods: In total, 1,465 police officers from 15 Belgian Local Police zones participated in a web-based survey, consisting of three segments: evaluating experiences with a list of 29 PTE, assessing if any of these PTE accounted for traumatic exposure, and evaluating 1-month probable PTSD, complex PTSD and subclinical PTSD prevalence using the International Trauma Questionnaire (ITQ).

Results: Police officers frequently experience a wide range of PTE. A large majority of 93.0% reports traumatic exposure. Assessment with ITQ shows a 1-month prevalence of 5.87% for probable PTSD and 1.50% for probable complex PTSD, while an additional 7.58% report subclinical PTSD. No demographic variables influenced PTSD prevalence. Cumulative PTE experiences in itself did not predict PTSD, while the characteristics of certain PTE did entail a higher prevalence of probable PTSD and subclinical PTSD.

Discussion: This study is the first to evaluate experiences with PTE, traumatic exposure and 1-month prevalence of probable PTSD, complex PTSD, and subclinical PTSD in Belgian police officers. Police officers are frequently confronted with a broad variety of PTE, and a large majority reports traumatic exposure. The 1-month prevalence of probable PTSD is significantly higher compared to previous international research in the general population, but lower than in similar international research involving police officers. In this study, cumulative PTE experiences in itself did not reliably predict PTSD, while the characteristics of certain PTE did. Posttraumatic symptoms are an important mental health challenge in Belgian police.

HIGHLIGHTS

• This study is the first to evaluate experiences with PTE, traumatic exposure and 1-month prevalence of probable PTSD, complex PTSD, and subclinical PTSD in Belgian police officers.

• Results show that police officers are frequently confronted with a broad variety of PTE, and a large majority reports traumatic exposure.

• The 1-month prevalence of probable PTSD is significantly higher compared to previous international research in the general population, but lower than in similar international research involving police officers.

• In this study, cumulative PTE experiences in itself did not reliably predict PTSD, while the characteristics of certain PTE did.

• Posttraumatic symptoms are an important mental health challenge in Belgian police.

Antecedentes: Los oficiales de policía se encuentran en riesgo considerable de desarrollar síntomas de estrés traumático debido a que, con frecuencia, se enfrentan a incidentes violentos o perturbadores. Se investigaron el experimentar eventos potencialmente traumáticos (EPT), la exposición a trauma, y la prevalencia de un probable trastorno de estrés postraumático (TEPT), de TEPT complejo y de TEPT subclínico en oficiales de policía belgas.

Métodos: Participaron en total 1.465 oficiales de policía de 15 zonas policiales belgas en una encuesta en línea consistente en tres secciones: evaluación de experiencias a través de una lista de 29 EPT, valoración relacionada a si alguno de los EPT identificados fueron experimentados como exposición a trauma, y la evaluación de un probable TEPT al mes, TEPT complejo o TEPT subclínico a través del Cuestionario Internacional sobre Estrés Traumático (ITQ por sus siglas en inglés).

Resultados: Los oficiales de policía experimentaron una amplia variedad de EPT de forma frecuente. Una gran mayoría (93%) reportó exposición a trauma. La evaluación mediante el ITQ encontró una prevalencia al mes de un 5,87% para un probable TEPT y de un 1,5% para un probable TEPT complejo, mientras que reportó un 7,58% adicional como TEPT subclínico. Las variables demográficas no influenciaron la prevalencia del TEPT. La experiencia acumulativa de EPT no predecía el TEPT por sí misma, mientras que determinadas características de ciertas EPT sí se vincularon a una prevalencia más alta de un probable TEPT o TEPT subclínico.

Discusión: Este estudio es el primero en evaluar el experimentar EPT, la exposición a trauma y la prevalencia de un probable TEPT, TEPT complejo o TEPT subclínico en oficiales de policía belgas en el último mes. Con frecuencia, los oficiales de policía enfrentan una gran variedad de EPT y la gran mayoría reporta exposición a trauma. La prevalencia en el último mes de un probable TEPT es significativamente más alta que en investigaciones internacionales previas realizadas en la población general, pero más baja que en investigaciones internacionales realizadas en oficiales de policía. En este estudio, la experiencia acumulativa de EPT por sí misma no predecía con fiabilidad el desarrollo del TEPT, mientras que determinadas características de ciertas EPT sí lo hacía. Los síntomas postraumáticos son un desafío para la salud mental de la policía belga.

背景：警察出现创伤后症状的风险相当大，因为他们经常遇到暴力或情绪不安的事件。我们调查了比利时警察样本中潜在创伤事件 (PTE)、创伤暴露以及可能的创伤后应激障碍 (PTSD)、复杂性 PTSD 和亚临床 PTSD 的流行率。

方法：总共有来自 15个比利时地方警区的 1.465 名警察参与了一项由三个部分组成的网络调查：使用 29个 PTE 列表评估经历，评估这些 PTE 中是否有创伤暴露，以及使用国际创伤问卷 (ITQ) 评估可能的 PTSD、复杂性 PTSD 和亚临床 PTSD 的1个月流行率。

结果：警察频繁经历各种PTE。 93.0% 的大多数人报告有创伤经历。 ITQ 评估显示可能的 PTSD 1个月流行率为 5.87%，可能的复杂性 PTSD为 1.50%，而另外有 7.58%的人报告亚临床 PTSD。没有人口统计学变量影响 PTSD 流行率。累积的 PTE 经历本身并不能预测 PTSD，而某些 PTE 的特征确实导致可能的 PTSD 和亚临床 PTSD 的流行率更高。

讨论：本研究首次评估了比利时警察的 PTE、创伤暴露和可能的 PTSD、复杂性 PTSD 和亚临床 PTSD 的 1个月流行率。警察经常面临各种各样的 PTE，大多数人报告有创伤暴露。可能的 PTSD 的 1个月流行率显著高于以往在一般人群中进行的国际研究，但低于了涉及警察的类似国际研究。在本研究中，累积的 PTE 经历本身并不能可靠地预测 PTSD，而某些 PTE 的特征却可以。创伤后症状是比利时警察面临的重要心理健康挑战。

KEYWORDS:

• Psychological trauma
• posttraumatic stress disorder
• PTSD
• police
• law enforcement

PALABRAS CLAVE:

• Trauma psicológico
• trastorno de estrés postraumático
• TEPT
• policía
• actuación policial

关键词:

• 心理创伤爆炸 、创伤后应激障碍炸、PTSD炸、警察炸、执法

1. Introduction

Police officers encounter potentially traumatic events (PTE), such as incidents of a violent or emotionally upsetting nature, more frequently than the general population. This puts them at a higher risk of developing posttraumatic stress disorder (PTSD) (Jorgensen & Elklit, 2021).

The latest edition of the International Classification of Diseases or ICD-11 (World Health Organisation, 2018) defines PTSD as a condition occurring after exposure to an extremely threatening or horrific event, causing symptoms grouped into three clusters: the reexperiencing of the traumatic event, avoidance of thoughts, emotions, activities and situations linked to the event, and persistent hypervigilance. These symptoms persist for several weeks and cause significant impairment.

In ICD-11, complex PTSD was added as a novel diagnostic category, a condition similarly deriving from exposure to an extremely threatening or horrific event, causing the main symptoms of PTSD as cited above, in addition to severe and persistent difficulties in affect regulation, a negative self-image and difficulties in interpersonal relationships. Research suggests that complex PTSD identifies a clinical category distinct from PTSD, involving a group that has more frequently experienced multiple traumatic events and has greater functional impairment than those suffering from PTSD (Brewin et al., 2017). Given that police officers are often persistently and repeatedly exposed to traumatic events, they could be at greater risk of developing complex PTSD.

In addition to those that meet criteria for PTSD and complex PTSD following a traumatic event, research shows that there is a distinct group that experiences so-called ‘subclinical’ or ‘subthreshold’ posttraumatic symptoms, negatively impacting their daily functioning (Mylle & Maes, 2004; Zlotnick et al., 2002). McLaughlin et al. (2015) concluded that no clear definition of subclinical PTSD exists, and proposed that complying with at least two of any of the DSM-5 criteria B to E identifies the group with the highest negative outcome. Including measurement of subclinical PTSD in research aimed at police officers is particularly interesting since studies have shown that they tend to underreport posttraumatic symptoms (Marshall et al., 2021).

In their systematic review, Wagner et al. (2020) found that the reported point prevalence of PTSD in samples of police officers differs considerably between studies, with a range of 7.8% to 16.5% in European and North American samples of police officers. The reviewers primarily attribute this to the wide variety of measures, diagnostic criteria, and samples used. In another recent systematic review and meta-analysis, Syed et al. (2020) evaluated the global prevalence of PTSD in police personnel, and found an overall pooled point prevalence of 14.2%, but similarly concluded that significant differences in reported prevalence exist. Even with such sizable differences in outcome between studies, both reviewers conclude that there is strong evidence that the prevalence of PTSD in police officers is significantly higher than in the general population, reported at a point prevalence of 3.6% for high income countries in the World Mental Health surveys (Koenen et al., 2017) and 4.4% in the UK Adult Psychiatric Morbidity Survey (McManus et al., 2014).

Only two studies have evaluated the point prevalence of complex PTSD in police officers to date, in both cases in UK police, and demonstrated large differences in prevalence. Brewin et al. (2022) found a point-prevalence for probable complex PTSD of 12.6%, while Steel et al. (2021) only report 2.0%. In the latter study, this difference is attributed by the authors to their use of routine screening in a health surveillance programme for high-risk profiles, rather than in a general, participation-based survey. Only the prevalence found in the former study is considerably higher than in international research in the general population, where a point prevalence of 0.5% to 7.7% was found (Ben-Ezra et al., 2018; Cloitre et al., 2019; Hyland et al., 2021; Kvederaite et al., 2021; Maercker et al., 2018).

In their evaluation of the WHO Mental Health Surveys and using their aforementioned proposed definition, McLaughlin et al. (2015) found a point prevalence of 3.6% for subclinical PTSD, compared to 3.0% for ‘full’ PTSD. Grubaugh et al. (2005) demonstrated that subclinical PTSD is often present in veterans, but little research exists regarding police officers. Chen et al. (2020) defined subclinical PTSD as at least meeting DSM-criterium B, in addition to (at least one symptom of) criteria C and D, and report that in addition to 9.3% PTSD prevalence, no less than 17.5% of police officers implicated in the 9/11 terror attacks fit criteria for subclinical PTSD.

Certain risk factors have shown to influence the likelihood of development of PTSD, both in the general population and in police officers.

In the general population, women have a higher risk of developing PTSD, even when confronted with the same traumatic incident (Kessler et al., 2017). The gender effect in the prevalence of PTSD in police officers has been studied at length, with contrasting results. One systematic review by Sherwood et al. (2019) did not find any effect of gender. Another, by Syed et al. (2020) found a slightly higher risk of PTSD for female police officers, while the aforementioned study by Brewin et al. (2022) found a significantly higher prevalence in male police officers. Wagner et al. (2020) suggest that selection and training might flatten out any gender effect on the development of PTSD and therefore, in general, gender does not play a key role in predicting the onset of PTSD in police officers.

For the development of PTSD in the general public, age has shown to be less of a predictive factor than the exposure to traumatic incidents, the risk of which evidently increases with age (Ozer et al., 2003). Similarly, age as a distinct factor only has limited predictive value on the development of PTSD in police officers, while seniority does play a significant role, as confrontation with PTE clearly increases (Syed et al., 2020; Wagner et al., 2020).

Both the characteristics of the traumatic event and the amount of exposure to traumatic incidents are key predictors of PTSD in the general population (Shalev et al., 2017). The intensity of a traumatic event, with exposure to death or severe injury, along with the abruptness, and the experience of uncontrollability or inescapability are strongly related to the risk of subsequent PTSD (Shalev et al., 2017). Developing PTSD is also significantly more common after experiencing four or more traumatic incidents (Ozer et al., 2003). Similar effects related to the characteristics of the traumatic event have been found in police officers, where the degree of perceived threat (McCaslin et al., 2006), the necessity of using force (Prati & Pietrantoni, 2010), the feeling of helplessness and lack of control (Koch, 2010), or being humiliated (Brewin et al., 2022) entail a higher risk of developing PTSD. In terms of the association between the frequency and accumulation of traumatic events and the development of PTSD in police officers, however, there is contrasting evidence. Some studies could find no effect at all (Chopko & Schwartz, 2012) or only a small association (Weiss et al., 2010), while another study indicated that it was the variety of traumatic events rather than the number of experienced potentially traumatic events in itself that influence subsequent PTSD (van der Meer et al., 2017).

In this study, we assessed experiences with PTE and traumatic exposure for a sample of Belgian police officers and determined the 1-month prevalence of probable¹ PTSD, complex PTSD, and subclinical PTSD. We also evaluated the effect of demographic factors (i.e. gender, age, seniority, police zone size and police service), cumulative PTE experiences and PTE characteristics on the prevalence of PTSD, complex PTSD, and subclinical PTSD.

Given the similarity of our design to the one used by Brewin et al. (2022), we expected to find comparable results in our study, namely an important level of traumatic exposure, and a 1-month prevalence of 8.0% for PTSD and 12.6% for probable complex PTSD. In addition, we aimed to identify factors that have a significant effect on PTSD symptoms. In line with literature, we did not expect significant differences between police zones, nor an effect of gender, or of age when controlled for seniority. We did however expect a higher prevalence in staff working in more incident-prone services such as intervention and criminal investigation, in participants with higher seniority, and in those that encountered a larger number of shocking or emotionally disturbing events.

This study and its predictions were pre-registered using AsPredicted® reference number 77352.

2. Methods

2.1. Participants and procedure

A total of 18 Local Police zones in Belgium were initially selected for participation based upon their size and location, to ensure variety in both criteria.

In Belgian Local Police zones, operational staff (i.e. police officers with police authority) work alongside civilian staff that have an administrative or supporting role. Both staff categories were invited to take part, but only results of the operational staff were considered in this article.

Every police zone provides seven police services, known as ‘basic functionalities,’ i.e. reception at the police station (answering questions, documenting complaints), intervention (patrolling and responding to emergency calls), criminal investigation, public order management (at demonstrations and events), traffic and road safety, community policing and crime victim assistance. We added an eighth service, namely management and administration, to include all staff.

Police zone chief commissioners were contacted by the researchers to invite them to participate ‘in research aimed at determining the prevalence of posttraumatic symptoms in Belgian police officers’. The study was described in the context of the Athena research project, ‘a research programme aimed at improving knowledge on the prevalence, development and treatment of PTSD in Belgian police officers.’

A total of 15 Local Police zones agreed to participate, which accounts for 7,116 potential participants, civilian and operational staff combined.

In October-November 2021, all staff members were invited by e-mail to take part in the study on a Qualtrics® web platform, which ran for 23 days in every zone. To minimise selection bias, only one reference to psychological trauma was made in the invitation, namely when explaining the broader Athena research project. The survey itself was presented as ‘a survey into police officers’ experiences with shocking or emotionally upsetting incidents, and their impact.’ Ample consideration went into stressing the anonymity of participants, and the confidentiality with which their answers would be treated.

Following general instructions and agreeing to the informed consent form, participants were asked to fill out demographical information and were then presented a list of potentially traumatic events. They were asked to indicate whether they had ever experienced any of these incidents during their operational career, and if so, how many times.

Afterwards, to determine traumatic exposure, they were asked if any of these incidents had been shocking, very emotionally upsetting and had a profound impact, and for this incident, to estimate how long ago this occurred. When they responded positively to this question, there were considered to be trauma-exposed.

In the next part of the survey, only participants that were considered to be trauma-exposed, continued to fill out the International Trauma Questionnaire (ITQ) (Cloitre et al., 2018) in order to evaluate the 1-month prevalence of probable PTSD, complex PTSD and subclinical PTSD.

Finally, the PTSD Checklist for DSM-5 (PCL-5) (Blevins et al., 2015), and the Patient Health Questionnaire-9 (PHQ-9) (Kroenke & Spitzer, 2002) were administered in the context of research questions that are beyond the scope of the present study. The results of these measures will be included wherever they contribute to answering the research questions of this study.

Since taking part in a study that enquires about shocking or emotionally upsetting events and ensuing posttraumatic or depressive symptoms might cause distress in participants, contact information of support services of both Local and Federal Police were mentioned in the invitation, and at the start and beginning of the survey.

The study was approved by the Social and Societal Ethics Committee of the KU Leuven under reference number G-2021-3782.

2.2. Measures

2.2.1. Demographics

Participants were asked to fill out personal information, namely their gender, age, seniority, staff category (civilian or operational staff), police zone, and police zone service.

2.2.2. PTE experiences and traumatic exposure

With permission, we based the list of PTE (cf. Supplementary Materials (SM)) on the Police Traumatic Events Checklist (PTEC) developed by Miller et al. (2021). We contacted five experienced police officers to make sure the list was adapted to Belgian police context. Limited changes in wording and layout were made.

In the PTE survey, participants were asked to indicate whether they had ever experienced the event, and if so, how many times, on a four-point Likert scale ranging from ‘never’ to ‘more than five times.’

Subsequently, to determine traumatic exposure, participants were asked if any of the events that were experienced had been shocking and emotionally disturbing and had a profound and long-lasting impact. For this incident participants were asked to estimate how long ago this occurred on a five-point scale, ranging from ‘less than 6 months ago’ to ‘more than 20 years ago’.

Since not all PTE included in the list of PTE would account for traumatic exposure, for example delivering bad news of the death of a loved one, we analysed the trauma-exposed group to ascertain that there had been sufficient exposure to a number of PTE that could be considered as exposure to a traumatic incident, such as being the victim of physical aggression or being involved in a shooting incident.

2.2.3. The international trauma questionnaire (ITQ)

The ITQ is a psychometrically robust screening measure that allows for the assessment of the presence of both probable PTSD and probable complex PTSD in line with ICD-11 diagnostic criteria (Cloitre et al., 2018), and it has shown good internal reliability (α = .89) (Cloitre et al., 2021). The ITQ has repeatedly been used in research with police officers, e.g. by Brewin et al. (2022). The Dutch version, used here, was translated, and validated according to WHO standards (Eidhof et al., 2018). The ITQ shows high agreement with DSM-5 PTSD diagnostics using PCL-5 (κ = .690), but there are consistent findings that ICD-11 criteria lead to a slightly lower PTSD prevalence than when using DSM-5 criteria (Hyland et al., 2017).

After establishing traumatic exposure, by asking participants if they ever experienced a shocking, very emotionally upsetting event that had a profound and long-term impact on them, the ITQ first consists of six items assessing symptoms of PTSD in the three clusters of reexperiencing, avoidance and hypervigilance, and a measure of functional impairment of these symptoms. It is followed by an assessment of complex PTSD by the three ICD-11 diagnostic clusters of affective dysregulation, negative self-concept, and disturbed relationships, similarly followed by a measure of functional impairment. Items in the ITQ are answered on a four-point Likert scale, from 0 or ‘not at all’ to 4, ‘extremely’. According to clinical scoring guidelines, a symptom is considered present when it is answered with a score of at least 2 (i.e. ‘moderately’), and a PTSD or complex PTSD diagnosis can be established if there is at least one symptom present in each symptom cluster, along with at least one functional impairment. It is important to mention that ICD-11 stipulates PTSD and complex PTSD diagnoses cannot co-occur.

Finally, we used the ITQ to estimate the point prevalence of subclinical PTSD. Since the proposed definition by McLaughlin et al. (2015) involved DSM-criteria, as did the recent comparable study by Chen et al. (2020), we opted to define subclinical PTSD conservatively, as fitting any 3 out of 4 ICD-11 PTSD criteria on ITQ.

2.2.4. Secondary measures

For research questions outside of the scope of the present study, two additional measures were administered at the end of the survey, namely the Dutch versions of PCL-5 (Boeschoten et al., 2014) and PHQ-9 (Kroenke & Spitzer, 2002).

PCL-5 assesses posttraumatic symptoms in the past month according to DSM-5 criteria, in four symptom clusters: re-experiencing (cluster B), avoidance (cluster C), negative alterations in cognitions and mood (cluster D) and hyperarousal (cluster E). The twenty items in PCL-5 are answered on a four-point Likert scale, from 0 or ‘not at all’ to 4, ‘extremely’, resulting in a total score ranging from 0 to 80. According to clinical scoring guidelines, a symptom is considered to be present when the score is 2, ‘moderately’, or higher. A probable PTSD diagnosis is established in two ways, either by using the recommended cut-off score of 33 when adding up the total score (Blevins et al., 2015), or by applying the DSM-5 clinical scoring criteria. The latter method has been used in this study. PCL-5 has demonstrated good internal reliability (α = .96) in previous research (Bovin et al., Psychometric properties of the PTSD Checklist for Diagnostic and Statistical Manual of Mental Disorders-Fifth Edition (PCL-5) in veterans, 2016), and the Dutch version, used here, has been applied in samples of police officers (van der Meer, et al., 2017).

PHQ-9 is a widely used nine-item screening measure for depression, determining the presence of depressive symptoms on a four-point Likert scale, ranging from 0 or ‘not at all’ to 3, ‘nearly every day’. The Dutch version, used here, has shown good internal reliability (α = .88) (Zuithoff, et al., 2010). A recent meta-analysis has indicated that a cut-off score greater than or equal to 10 attains optimal sensitivity (86%) and specificity (85%) and has therefore been used in this study (Levis, Benedetti, & Thombs, 2019).

2.3. Statistical analysis

The validity of collected records was evaluated first. No duplicate entries were found. There were some straight-line responses (i.e. answering identically on every question) in ITQ, however, this can be expected on a clinical measure for those without any symptoms. One record was removed because the small disparity between age and seniority that was entered by the participant was factually impossible. We only included records that completed at least the first part of the survey, namely the experiences with PTE, and records were considered for a measure when that part of the survey was entirely completed.

To determine the representativeness of our sample with respect to gender, we gathered information on the gender distribution in operational staff in all 15 Local Police zones, indicating that 26.0% of all staff in the Local Police was female. In our sample, 28.9% of respondents were female. This difference was not significant, when evaluated using binomial testing, p = .13. We also applied weighting for the distribution of gender per police zone to evaluate whether this substantially changed our results. This only changed the outcomes of the ITQ very slightly, lowering PSTD prevalence in trauma-exposed operational staff from 6.37% to 6.29% and raising complex PTSD prevalence from 1.63% to 1.68%. Since we could not correct for age, seniority, and police service, and given the very small changes in prevalence, we opted to continue with the unweighted sample and report it as such.

Experiences with potentially traumatic events were evaluated using descriptive statistics. To attain a very conservative estimate of the cumulative burden of experienced PTE for each participant, responses were coded as ‘no experiences’ for ‘never,’ ‘1 experience’ for ‘1 time’, ‘2 experiences’ for ‘2–5 times’ and ‘6 experiences’ for ‘more than 5 times’. Since there were 29 PTE in total, this yields a theoretical range of 0–174 experiences. The effect of demographic variables on PTE experiences was assessed individually using χ² tests.

Subsequently, traumatic exposure, whether a participant indicated one of the PTE was shocking, very emotionally disturbing and had a profound and long-lasting effect, was evaluated using descriptive statistics. The effect of demographic variables on traumatic exposure was assessed individually using χ² tests. Using simultaneous logistic regression, demographic variables were entered into a model to determine which factors influenced traumatic exposure.

For all measures, the aforementioned scoring procedures were applied. The effect of demographic variables on ITQ results was evaluated per variable using χ² tests, and the effect of cumulative PTE experiences with ANOVA. Finally, multinomial logistic regressions using a stepwise, backward selection method were used for cumulative PTE experiences and for all potentially traumatic events, to determine which of them were uniquely associated with PTSD, complex PTSD, and subclinical PTSD, and could discern between diagnoses. However, since there were only a small number of participants that fit complex PTSD criteria, multinomial logistic regression was only possible for the differentiation between no PTSD vs. subclinical PTSD vs. PTSD, and not to predict complex PTSD.

3. Results

3.1. Demographics

A total of 1.896 Belgian police officers responded to the invitation, which corresponds to a response rate of 26.6%. Of them, 1.701 participants filled out both the demographic information and completed the list of experiences with PTE. There were 236 civilian staff members that filled out the survey, and their results will not be further evaluated in this study. Of the 1.465 participating police officers, there were 1.042 (71.1%) male participants. Average age was 40.68 (SE 9.62) and average seniority was 16.28 years (SE 9.54).

Subsequently, 20 (1.4%) respondents dropped out and 102 (7.0%) police officers indicated they had never experienced any PTE as shocking, very emotionally disturbing and having a profound and long-lasting impact, thus not fitting criteria for traumatic exposure. This means that a total of 1.351 police officers were considered to be trauma-exposed and completed the ITQ, of which 960 (71.1%) respondents were male. The average age of ITQ participants was 40.64 (SE 9.67) and the average seniority was 16.63 years (SE 9.56).

3.2. Exposure to potentially traumatic events

Detailed results for experiences with PTE by incident can be found in the Supplementary Materials (SM) Table 1. In almost all PTE, male police officers reported experiencing more PTE compared to female staff, so results are directly displayed by gender. In being the victim of sexual aggression, however, female staff were significantly more frequently implicated than male police officers, χ²(3) = 167.373, p < .001.

In SM Table 2, descriptive statistics are presented for experiences with PTE. At 99.72% (n = 1.461), almost all operational staff reported experiencing at least one PTE. The average cumulative number of PTE was 45 incidents. The interquartile range for cumulative PTE experiences was 32–119 for male police officers, and 20–107 for females. Male police officers registered significantly more experiences with PTE compared to female staff, χ²(105) = 151.975, p = .002. Older police officers reported a higher number of PTE, χ²(420) = 499.623, p = .004, as did more senior staff, χ²(315) = 584.995, p < .001. There was no effect of police zone size, χ²(315) = 340.764, p = .152, but a significant effect of police zone service, with police officers working in criminal investigation and crime victim assistance reported higher numbers of PTE experiences, χ²(735) = 823.313, p = .007.

3.3. Traumatic exposure

Table 1 includes results for traumatic exposure: a large majority of 93.0% (n = 1.362) of operational staff identified at least one PTE as shocking or emotionally disturbing, having a profound and long-lasting impact. Every participant that was considered to be trauma-exposed had reported at least one direct exposure to a shocking and emotionally disturbing event, such as being the victim of physical violence or being involved in a shooting incident.

Table 1. Descriptive statistics for traumatic exposure (TE) (n = 1465).

Demographic variables		n	TE (%)	Pearson χ^2 for TE
All participants		1465	1363 (93.0%)
Gender	Male	1042	970 (93.1%)	χ^2(1) = 0.015
	Female	423	393 (92.9%)	p = .901
Age	≤ 24	44	33 (75.0%)	χ^2(4) = 32.231
	25–34	436	395 (90.6%)	p < .001
	35–44	462	440 (95.2%)
	45–54	394	371 (94.2%)
	≥ 55	129	124 (96.1%)
Seniority	≤ 4	178	133 (74.7%)	χ^2(3) = 109.090
	5–9	253	235 (92.9%)	p < .001
	10–19	483	468 (96.9%)
	≥ 20	551	527 (95.6%)
Police zone size	Small (<100)	91	82 (90.1%)	χ^2(3) = 3.354
	Average (100–250)	288	263 (91.3%)	p = .340
	Large (250–500)	503	472 (93.8%)
	Very large (>500)	583	546 (93.7%)
Police zone service	Management	121	111 (91.7%)	χ^2(7) = 16.703
	Intervention	636	585 (92.0%)	p = .019
	Reception	37	34 (91.9%)
	Public order	82	72 (87.8%)
	Criminal investigation	237	233 (98.3%)
	Crime victim assistance	15	15 (100.0%)
	Traffic and road safety	93	85 (91.4%)
	Community policing	244	228 (93.4%)

In contrast with mere PTE experiences, there was no gender effect in reported traumatic exposure, χ²(1) = 0.015, p = .901. Older staff, χ²(4) = 32.231, p < .001, and more senior police officers, χ²(3) = 109.090, p < .001, more frequently reported traumatic exposure. There was no effect of police zone size, X²(3) = 3.354, p = .340, but staff in certain police services, such as intervention, criminal investigation, and crime victim assistance more often, or in the case of crime victim assistance, always indicated traumatic exposure, X²(7) = 16.703, p = .019.

To further understand the effect of these demographic variables and whether they predicted traumatic exposure, all variables were entered in simultaneous logistic regression. Age as a factor, however, was removed due to high collinearity with seniority, r (1463) = .874, p < .001. The resulting model, in Table 2, was significant, χ² (1450) = 101.010, p < .001, Nagelkerke R² = 0.17. There was no significant effect of gender or police zone size. Traumatic exposure was significantly more likely for personnel with higher seniority, and highest in staff with more than 20 years seniority (OR 12.12, 95% CI [6.10–24.09]). When compared to management and administrative support services, the risk of traumatic exposure was significantly higher in intervention (OR 3.52, 95% CI [1.51–8.21]), in criminal investigation (OR 6.11, 95% CI [1.83–20.34]) and in crime victim assistance, where all participants had been trauma exposed.

Table 2. Results of simultaneous logistic regression for traumatic exposure (n = 1465).

Demographic variables		Odds ratio (95% CI)	Significance
Gender (reference = Female)	Male	0.88 (0.55–1.43)	p = .612
Seniority (reference = ≤ 4)	5–9	4.32 (2.37–7.87)	p < .001
	10–19	12.58 (6.35–24.95)	p < .001
	≥ 20	12.12 (6.10–24.09)	p < .001
Police zone size	Average (100–250)	0.94 (0.41–2.18)	p = .890
(reference = Small (<100))	Large (250–500)	1.43 (0.63–3.23)	p = .390
	Very large (>500)	1.78 (0.78–4.09)	p = .173
Police zone service	Intervention	3.52 (1.51–8.21)	p = .004
(reference = Management)	Reception	1.66 (0.40–6.89)	p = .488
	Public order	1.66 (0.56–4.93)	p = .358
	Criminal investigation	6.11 (1.83–20.34)	p = .003
	Crime victim assistance	∞ [0.00-∞]	p = .981
	Traffic and road safety	1.28 (0.47–3.52)	p = .632
	Community policing	1.56 (0.67–3.64)	p = .308

Note: The model is significant, χ²(1450) = 101.01, p < .001, Nagelkerke R² = 0.17.

3.4. Posttraumatic symptoms

Results of the ITQ measure are displayed in Table 3. In trauma-exposed operational staff, 1-month prevalence for probable PTSD was 6.36% (n = 86), and 1.63% (n = 22) fit criteria for probable complex PTSD, while we found a prevalence of probable subclinical PTSD of 8.22% (n = 111). Consequently, in the entire sample of operational staff, including non-trauma-exposed staff, PTSD prevalence was 5.87%, complex PTSD prevalence was 1.50%, and subclinical PTSD prevalence was 7.58%.

Table 3. Results of the ITQ measure for trauma-exposed participants (n = 1351).

Variable		ITQ				Pearson χ^2
		Subclinical PTSD	PTSD	Complex PTSD	No PTSD
Gender	Female	30 (7.7%)	22 (5.6%)	5 (1.3%)	334 (85.4%)	χ^2(3) = 1.258
	Male	81 (8.4%)	64 (6.7%)	17 (1.8%)	798 (83.1%)	p = .739
Age by category	24 or younger	5 (15.6%)	0 (0.0%)	0 (0.0%)	27 (84.4%)	χ^2(12) = 10.351
	25–34	35 (9.0%)	27 (6.9%)	4 (1.0%)	324 (83.1%)	p = .585
	35–44	31 (7.1%)	26 (6.0%)	7 (1.6%)	372 (85.3%)
	45–54	29 (7.8%)	25 (6.8%)	10 (2.7%)	306 (82.7%)
	55 or older	11 (8.9%)	8 (6.5%)	1 (0.8%)	103 (83.7%)
Seniority by category	0–4	7 (5.3%)	5 (3.8%)	0 (0.0%)	119 (90.8%)	χ^2(9) = 10.133
	5–9	18 (7.8%)	11 (4.8%)	4 (1.7%)	198 (85.7%)	p = .340
	10–19	44 (9.5%)	36 (7.8%)	7 (1.5%)	377 (81.3%)
	More than 20	42 (8.0%)	34 (6.5%)	11 (2.1%)	438 (83.4%)
Police zone size	Small	7 (8.5%)	4 (4.9%)	1 (1.2%)	70 (85.4%)	χ^2(9) = 5.734
	Medium	22 (8.4%)	21 (8.0%)	6 (2.3%)	212 (81.2%)	p = .766
	Large	38 (8.1%)	23 (4.9%)	5 (1.1%)	401 (85.9%)
	Very large	44 (8.1%)	38 (7.0%)	10 (1.8%)	449 (83.0%)
Police zone service	Management	10 (9.0%)	9 (8.1%)	0 (0.0%)	92 (82.9%)	χ^2(21) = 27.105
	Intervention	41 (7.1%)	35 (6.1%)	11 (1.9%)	491 (84.9%)	p = .167
	Reception	3 (8.8%)	3 (8.8%)	2 (5.9%)	26 (76.5%)
	Public order	5 (7.1%)	4 (5.7%)	0 (0.0%)	61 (87.1%)
	Criminal investigation	25 (10.8%)	14 (6.1%)	3 (1.3%)	189 (81.8%)
	Crime victim assistance	0 (0.0%)	0 (0.0%)	2 (13.3%)	13 (86.7%)
	Community policing	19 (8.4%)	16 (7.0%)	3 (1.3%)	189 (83.3%)
	Traffic and road safety	8 (9.4%)	5 (5.9%)	1 (1.2%)	71 (83.5%)
Time elapsed since traumatic exposure	1–6 months	17 (15.3%)	12 (14.0%)	5 (22.7%)	147 (13.0%)	χ^2(15) = 18.800
	6–12 months	6 (5.4%)	8 (9.3%)	5 (22.7%)	88 (7.8%)	p = .223
	1–5 years	41 (36.9%)	35 (40.7%)	10 (45.5%)	423 (37.4%)
	5–10 years	21 (18.9%)	17 (19.8%)	1 (4.5%)	216 (19.1%)
	10–20 years	24 (21.6%)	11 (12.8%)	1 (4.5%)	213 (18.8%)
	More than 20 years	2 (1.8%)	3 (3.5%)	0 (0.0%)	45 (4.0%)
Total		111 (8.2%)	86 (6.4%)	22 (1.6%)	1132 (83.8%)

When comparing demographic variables in terms of ITQ results, no effects were found. In SM Table 4, the results of a one-way ANOVA for the cumulative amount of PTE in trauma-exposed police officers are presented, and it shows significant differences between the subclinical PTSD (M 52.11, SD 23.82), PTSD (M 57.62, SD 20.93), complex PTSD (M 55.41, SD 23.16) and no-PTSD (M 45.86, SD 21.32) groups, F(3) = 12.945, p < .001, but not between the PTSD and complex PTSD groups, F(106) = .869, p = 353.

Multinomial logistic regression was used to determine whether cumulative PTE and the different PTE were uniquely associated with subclinical PTSD, PTSD, and complex PTSD, and could discern between these groups. However, there were few participants that fit complex PTSD criteria, and therefore regression analyses could only be done to differentiate between no PTSD, subclinical PTSD, and PTSD groups, and could not include complex PTSD.

We entered the cumulative amount of PTE, and exposure to the different PTE into a multinomial logistic regression model, using a stepwise, backward selection method. We controlled for depressive symptoms by adding the PHQ-9 results to the model. The found model, in Table 4, was highly significant, χ²(12) = 177.623, p < .001, Nagelkerke R² = 0.20. Despite differentiating between the different diagnostic groups, the amount of cumulative PTE could not reliably predict PTSD or subclinical PTSD. While subclinical PTSD was only predicted by PTE with an emotional impact (OR 2.96, 95% CI [1.85–4.74]), PTSD was more likely for police officers exposed to situations where they feared for their life (OR 1.86, 95% CI [1.09–3.18]), were involved in a shooting incident where they were fired upon (OR 2.40, 95% CI [1.09–3.18]), in PTE with an emotional impact (OR 2.89, 95% CI [1.63–5.13]) and when they were the victim of physical aggression with a weapon (OR 2.97, 95% CI [1.41–6.24]).

Table 4. Results of multinomial logistic regression for no PTSD vs. subclinical PTSD vs. PTSD (n = 1351).

Variables		Subclinical PTSD		PTSD
		Odds ratio (95% CI)	Significance	Odds ratio (95% CI)	Significance
Cumulative PTE		1.00 (0.99–1.02)	p = .494	1.00 (.99–1.01)	p = .992
(reference = 1)
PHQ-9 score		1.16 (1.11–1.21)	p < .001	1.17 (1.11–1.23)	p < .001
(reference = 0)
PTE exposure	Physical aggression with a weapon – Victim	1.30 (0.76–2.22)	p = .337	2.97 (1.41–6.24)	p = .004
(reference = not exposed)	Shooting incident – Fired upon	1.39 (0.63–3.07)	p = .415	2.40 (1.15–4.98)	p = .019
	Other – Feared for my life	0.91 (0.57–1.44)	p = .687	1.86 (1.09–3.18)	p = .022
	Other – Emotional impact	2.96 (1.85–4.74)	p < .001	2.89 (1.63–5.13)	p < .001

Note: The model is significant, χ²(12) = 177,623, p < .001, Nagelkerke R² = 0.20.

3.5. Secondary measures

PCL-5 prevalence of probable PTSD using diagnostic rules was nearly identical to ITQ, at 6.58% (n = 87). However, ITQ was administered first, and 27 participants had dropped out before filling out PCL-5, of which 3 participants screened positive for probable PTSD on ITQ.

Using the proposed cut-off score of 10, the point prevalence of probable moderate to severe depression as assessed by PHQ-9 in our sample was 9.2%. An additional 18 participants had dropped out after filling out PCL-5. However, none of these participants screened positive for probable PTSD following ITQ or PCL-5 assessment.

There were significant correlations between moderate to severe depression as per PHQ-9 and probable subclinical PTSD, r (1304) = .13, p < .001, probable PTSD, r (1304) = .26, p < .001, and probable complex PTSD, r (1304) = .35, p < .001, following assessment by the ITQ.

4. Discussion

Just about every police officer in this broad sample of Belgian police officers (99.72%) reports having experienced at least one PTE. Exposure to PTE is highly frequent, with a conservatively estimated average of 45 incidents. These results are in line with research by Rudofossi (2007), indicating that dealing with PTE as a police officer is common, and estimating that the total number of PTE in the average police officers’ career ranges between 10 and 900 incidents. Our results show that male, older and more senior staff, and those working in police services such as criminal investigation and crime victim assistance, report more frequent confrontation with PTE, while police zone size does not have a significant effect.

A large majority of police officers (93.0%) report that at least one of these events was shocking or emotionally disturbing to them and had a profound and long-lasting impact. Such high numbers of traumatic exposure are to be expected in police officers (Jorgensen & Elklit, 2021). In Brewin et al. (2022), 84.9% of respondents indicated traumatic exposure. There is no effect of gender or police zone size in the occurrence of traumatic exposure, although older and senior staff report traumatic exposure more frequently, as did police officers working in intervention, criminal investigation, and crime victim assistance.

In trauma-exposed police officers, 1-month prevalence for probable PTSD is 6.36%. This corresponds to a 1-month PTSD prevalence in the entire sample of 5.87%. In comparison with the range of point prevalence for PTSD found by Wagner et al. (2020), from 7.80% to 16.50%, the 1-month prevalence in this sample of Belgian police officers is relatively low. One possible explanation, in which most studies included in this systematic review used measures following DSM-5 criteria, and since assessment with ITQ, using ICD-11 criteria, tends to generate a slightly lower prevalence of PTSD (Hyland et al., 2017), can already be discarded given the very similar PCL-5 results (6.58%). However, underreporting could have lowered the found prevalence as the invitation to participate partially originated from within the Belgian police (Marshall et al., 2021).

In our sample of trauma-exposed police officers, 1.63% fit criteria for probable complex PTSD, corresponding to 1.50% of the entire sample. Brewin et al. (2022) have determined complex PTSD prevalence using ITQ in a sample of trauma-exposed police officers in a similar design, therefore allowing for meaningful comparison. While the prevalence of probable PTSD remains comparable, i.e. 6.36% vs. 7.99%, we find a substantially lower prevalence of complex PTSD, at 1.63% vs. 12.59%. These results indicate that the prevalence of complex PTSD in this sample of Belgian police officers is not elevated in comparison to previous international studies in the general population, while this was the case for British police officers in the aforementioned study.

Finally, 8.22% of the trauma-exposed sample, 7.58% of the total sample, show subclinical PTSD. Very few studies have reported on the prevalence of subthreshold PTSD symptoms in police officers, but our results are in line with earlier findings by Chen et al. (2020).

In this study, the prevalence of PTSD, complex PTSD and subclinical PTSD is not significantly influenced by demographic factors. Similar to what systematic reviews by Syed et al. (2020) and Wagner et al. (2020) have shown, and in line with our predictions, there is no effect of gender, age or police zone size. However, contrary to most studies included in these reviews and our expectations, seniority does not predict PTSD in our sample, and neither does police service. We did expect PTSD prevalence to be higher in more senior police officers, and those working in more incident-prone services. This outcome indicates that posttraumatic symptoms in our sample are not related to one specific subgroup, but rather form a mental health risk for police officers in general. For the lack of effect of police service, we could hypothesise that the widely used so-called ‘mobility’ system in Belgian police, where police officers can change police zone and police service at least every five years, could have possibly flattened out this effect.

Although the reported cumulative exposure to PTE was significantly higher for PTSD, complex PTSD, and subclinical PTSD groups than in the non-PTSD group, in our sample, the accumulation of PTE in itself did not predict prevalence of posttraumatic conditions. Some PTE do have a strong positive association with clinical outcomes, such as being the victim of physical aggression with a weapon, being involved in a shooting incident, following instances where police officers feared for their life or where they endured a substantial emotional impact. Interestingly, there were differences in which PTE predicted subclinical PTSD or clinical PTSD. While situations with a profound emotional impact could predict both subclinical PTSD and PTSD, situations that involved physical aggression with a weapon or firearm, or experiencing a feeling of life danger, directly predicted clinical levels of PTSD. These findings support the hypothesis that in police officers, characteristics of certain PTE rather than cumulative confrontation with PTE as such have an impact on the development of PTSD, and that there is a difference in PTE that entail subclinical PTSD instead of clinical PTSD.

The present study addressed a number of gaps in knowledge and thus made some important contributions to the field.

This study is the first to evaluate PTSD, complex PTSD, and subclinical PTSD in a sample of Belgian police officers, extending our knowledge on this important issue. In addition, including subclinical PTSD in our study showed that subclinical posttraumatic symptoms are prevalent in a large group of police officers. Our results in general underline that posttraumatic symptoms in Belgian police pose an important mental health challenge, allowing for the development of corresponding policy.

The large sample size and evaluation of experiences with PTE allowed for the construction of a regression model predicting the development of PTSD and subclinical PTSD, and we found that characteristics of the PTE seem to be more important than the mere number of encountered PTE.

Our study has several limitations.

The rather limited response rate and the use of self-report measures using a web-based survey could have negatively influenced the representativeness of our sample and the results of our study. However, selection bias was minimised by limiting references to PTSD or trauma, and we did not see any effects when comparing early or late responders. Additionally, low response rates were mostly found in larger police zones; half of the 15 Local Police zones had a response rate well over 35%, some reaching 50%.

A second possible limitation is that the invitation to participate in this research came from within the police organisation. Even though anonymity was stressed, research in police officers shows that participants in an employer-sent survey tend to underreport, especially those with higher symptoms (Marshall et al., 2021). One could therefore argue that, if anything, our results would be an underrepresentation of the actual prevalence.

In addition, since the proposed definition by McLaughlin et al. (2015) for subclinical PTSD uses DSM-5 criteria and this study evaluated posttraumatic symptoms using ICD-11, our conservative estimate by means of ITQ limits the way the results in the present study can be compared to other research.

The way we evaluated participants’ experiences with PTE could have influenced effects in two ways. Firstly, prior traumatic incidents in the respondents’ personal life that were not assessed, such as adverse childhood experiences, could have influenced answers on the clinical measures. Secondly, we used a list of predetermined PTE and set categories for the number of experiences, instead of asking participants to describe the specific PTE that they experienced, and estimate the times they were confronted with it. This could cause some participants to not report PTE that could have had an even more significant impact on the development of PTSD or having to underreport the amount of confrontation. However, precisely determining the characteristics and amount of PTE experienced by police officers is particularly challenging, as it is reliant on participant recollection. Future research, especially in a longitudinal design, can better describe the effects of (cumulative) exposure to PTE, and the effect of certain PTE in particular, on the development of PTSD. The authors will therefore conduct a study aimed at identifying risk factors for the development of PTSD in a longitudinal design.

Finally, any prevalence reported in this article has been established using a screening measure that, although highly reliable in predicting a subsequent clinical diagnosis, only allows for the assessment of the presence of probable PTSD and probable complex PTSD.

5. Conclusion

This study is the first to evaluate experiences with PTE, traumatic exposure, and the prevalence of probable PTSD, complex PTSD, and subclinical PTSD in Belgian police officers. We find that police officers are frequently exposed to a substantial variety of PTE, and a large majority also report traumatic exposure. Prevalence of PTSD, but not complex PTSD, is significantly higher when compared to the general population. A substantial percentage of police officers have subclinical PTSD. The absence of an effect of demographic factors indicates that most police officers are equally at risk for the development of posttraumatic stress disorder. While cumulative PTE experiences in itself are not a predictor of PTSD in this study, the characteristics of some PTE, in particular those that are life-threatening or have an important emotional impact, have predictive value for the development of PTSD and subclinical PTSD. These results clearly indicate that posttraumatic symptoms are a significant mental health problem in Belgian police officers.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 In this article, prevalence of PTSD and complex PTSD as assessed by means of ITQ or PCL-5 is always considered to be ‘probable’. To improve readability, this is sometimes omitted.