Introduction

As part of their routine job, police officers are frequently exposed to critical and potentially traumatizing incidents. These police-specific critical incidents (CI) include dangerous and life-threatening situations as well as distressing events with victims of accidents, abuse, violence, murder, or suicide (Carlier et al. 1997; Hartley et al. 2013; Jørgensen and Elklit 2022; Violanti et al. 2017). Cumulative exposure to traumatic incidents in policing is associated with an increased risk of developing mental health problems, in particular post-traumatic stress disorder (PTSD) (Brewin et al. 2020; Syed et al. 2020). Repeated or chronic trauma is further considered a risk factor for complex PTSD (Cloitre 2020), which is associated with higher levels of psychiatric burden (Brewin 2020). Identifying CI in police work is therefore of uttermost importance to prevention initiatives in police organizations.

In this study, we investigate situations that can be considered as critical incidents in police work. We present the development of general categories of police-specific CI based on content analysis of 2960 qualitative descriptions of incidents perceived to cause long-term psychological reactions among 1659 Danish police officers. Thus, this study provides insights into the character of CI in police work with important implications related to organizational and clinical prevention initiatives.

Critical Incidents in Police Work

The concept of CI and the effort to identify these within a specific job are not new. In 1954, John Flanagan developed “the critical incident technique”, which was a method to identify particular aspects of a job that was critical for success or failure (Flanagan 1954). The technique came from studies carried out in the Aviation Psychology Program for the US Army Air Forces, in which data were collected as numerous incident descriptions and analyzed using content analysis techniques to identify common themes and dimensions of CI and appertaining critical behaviors (Bartram 2008; Flanagan 1954). Drawing on this tradition and initial analysis of job-related CI, many police studies now involve a particular interest in the psychological circumstances and potential negative health consequences related to police officers’ experiences with CI (e.g., Brewin et al. 2020; Carlier et al. 1997; Chopko et al. 2018; Hartley et al. 2013; Weiss et al. 2010). However, the definitions and understandings of police-specific CI vary extensively with different conceptualizations.

Concepts, such as “critical incidents” (Liberman et al. 2002; Skogstad et al. 2013; Weiss et al. 2010), “police traumatic experiences” (Chopko et al. 2018; Miller et al. 2022), and “police-specific traumatic events” (Hartley et al. 2013), primarily focus on traumatic events in police work, whereas the term “police work stressors” and “police stress” also includes administrative stressors and lack of organizational support (Berg et al. 2005; McCreary and Thompson 2006; Spielberger et al. 1981; Violanti et al. 2016). This diversity is also exemplified by the development and use of self-developed “trauma questions” in many police studies (e.g., Brewin et al. 2020; Kyron et al. 2022) as well as different police-specific tools, e.g., the 60-item Spielberger Police Stress Survey (PSS) (Spielberger et al. 1981), the 20-item Operational- and 19-item Organizational Police Stress Questionnaires (PSQ-op and PSQ-org) (McCreary and Thompson 2006), the 9-item Police Incident Survey (PIS) (Hartley et al. 2013; Violanti and Gehrke 2004), the 68-item Work Environment Inventory (WEI) (Liberman et al. 2002), the 34-item Critical Incident History Questionnaire (CIHQ) (Weiss et al. 2010), or the most recent Police Traumatic Events Checklist (PTEC) (Miller et al. 2022). These scales are developed in police studies from the USA, Canada, and the UK. However, there can be assumed to be national differences in incident types and frequency due to differences in, e.g., police education, training, work functions, specific police tasks, and societal conditions (e.g., access to firearms is considerably easier in the USA than in Denmark).

Three European studies have described broad categories of traumatic events in police work. Karlsson and Christianson (2003) studied 162 Swedish police officers’ worst experiences and identified nine types of traumatic events (armed threats, traffic accidents, murder, threats, accidents, investigations, suicide, notification, taking children into custody, and miscellaneous). They found that most of the traumatic experiences occurred early on in the police officers’ careers and often remained in memories in the form of visual, tactile, and olfactory sensations. A modest sample size, however, makes generalizations difficult. Carlier et al. (2000) identified 37 types of police incidents and created a list of 15 police critical incidents that were most frequently reported by 82 police officers in the Netherlands with or without PTSD. This study did, however, not describe their coding methodology in detail, and the sample size is limited.

In a Danish police context, Ibsen (1997) identified four main categories of distressing event types in police work (active physical threat, children, strong sensorial impressions, and a mixed category), which were further defined by 36 subtypes of events (see online supplementary Table 1). This study was based on a content analysis of qualitative data collected in the Danish Police Study 1993 (DPS-1993) among all Danish, Greenlandic, and Faroese police officers (9035 participated with a response rate of 89.3%). The results showed that the negative health consequences of exposure to distressing events in police work were the most profound work environment problem (Ibsen 1997; Ibsen and Nigard 2008).

Objective

Considering the diversity in conceptualization, operationalization, and measurement of CI or “traumatic events” in police work, there is a need for an in-depth clarification of what can currently be characterized as CI in Danish police work, based on a larger body of data material and a thorough qualitative method. Therefore, the present study aims to enhance our understanding of the core characteristics of work-related CI that are perceived as most burdensome among police officers. More specifically, the study will explore Danish police officers’ descriptions of perceived work-related incidents that still affect them emotionally and inductively develop general categories of police CI using a qualitative approach of content analysis.

To clarify, we define a CI in police work in accordance with Weiss et al. (2010) and Liberman et al. (2002) as a potentially traumatic incident related to the unique content of police work, e.g., to face life-threatening situations, witness severely injured or dead people, or handle cases with abused children. Incidents related to the collegial work environment (e.g., bullying) or organizational circumstances are perceived as general job-related stressors that may be distressing or even traumatic (Dobry et al. 2013; Jahnke et al. 2019; Matthiesen and Einarsen 2004) but which apply to any area of work and not only police work.

Method

Study Population and Data Collection

Participants for the present study were part of the Are You All right (AYA) project, a prospective cohort study investigating associations between exposure to potentially traumatic CI, work environment and mental health, and sickness absence among Danish, Greenlandic, and Faroes police officers (Hansen et al. 2022). The AYA project is conducted between 2021 and 2024 and consists of 12 electronic surveys that are sent out every third month. The qualitative data used in the present study were collected at baseline.

All police officers in the permanent staff of the Danish Police Force (including police officers on Greenland and Faroese Islands) were invited to participate in the AYA project (n = 11,363) in March 2021. In total, 6040 (53.2%) police officers participated in the baseline survey, and 5380 (47.3%) completed the whole survey. In total, 1659 wrote a qualitative description of distressing events experienced during their careers.

Questions on CI

Two questions were included for analysis. First, respondents were asked about emotionally distressing incidents experienced in police work (TQ2a): During the time you have been employed by the police, have you experienced incidents that still affect you emotionally? (yes/no). If the officers answered “yes” to this item, a text box then became visible with the question (TQ2b): Any comments or elaborations? This allowed the officers to make qualitative descriptions of incidents that still affected them. These questions were replicated from DPS-1993 to allow for comparison.

Data Management and Analysis

Sample Characteristics

Simple descriptive statistics were calculated concerning the answers on TQ2a, TQ2b, and the officers’ gender and age (gender was measured as “woman,” “man,” or “other” and age in whole years) (Table 1). All answers on age and gender were assumed valid except one (reported age > 100.000 years), which was deleted. All answers on TQ2a and TQ2b were assumed valid and included in the analyses.

Development of Categories of CI Using Content Analysis

In total, 2672 police officers (46.5%) answered “yes” on TQ2a. Of these, 1659 elaborated on their answer and wrote a comment on TQ2b. To investigate these comments, we used the Hsieh and Shannon (2005) method of conventional content analysis. This qualitative research method was used as a strategy for interpretation of the written comments through systematic classification, coding, and identification of themes in the data set. In this analysis, the researchers allow the themes to flow from the data without preconceived categories. This approach can also be described as inductive category development (Hsieh and Shannon 2005). The first and second authors conducted the content analysis under guidance of the remaining authors. Only the first author was familiar with the main CI categories identified in Danish police work in 1993, but this knowledge was not initially discussed with the second author to ensure the inductive analytical approach.

First, all 1659 comments were read by both authors to get familiarized with the material. During the reading, impression notes were made, and text was highlighted when appearing to capture thematically different key elements of perceived CI. It became evident that many of the written comments described several thematically different events. The comments were divided so that each qualitative description pertained to a unique situation resulting in a total of 2960 descriptions of emotionally taxing events. During this reading and division process, impression notes of thematic key elements were continuously discussed among the authors, and a total number of 6 main categories and 40 subcategories were initially developed.

The first 50 descriptions were then reviewed jointly and coded by both authors to ensure agreement on the scope and coding usability of subcategories. Next, the full data set was reread by both authors and coded independently according to the developed subcategories. Labels and definitions of inconsistent subcategories were also discussed and refined to better fit the data.

After the first independent coding of the full data material, we used Cohen’s kappa coefficient to assess inter-rater agreement (Cohen 1960), which is a widely used statistical measure of inter-rater agreement for qualitative items (Lombard et al. 2002). The first inter-rater agreement was 0.88, which is excellent according to Fleiss’ guidelines (Fleiss et al. 2003). However, a subcategory of non-categorized descriptions took up 14.7% and 14.3%, respectively, of the authors’ codings. A rereading of these descriptions led to the development of two new subcategories (one consisting of descriptions of officers’ psychological reactions to CI without specification of any CI and one of descriptions of cases with children, not further specified). Then, another round of independent coding of these descriptions followed. A few inconsistent and small subcategories, with less than 20 incident descriptions in each, were also merged and refined as they appeared to overlap with other subcategories in thematic content (e.g., two subcategories of “Poor duty scheduling” and “High work pressure” were grouped under the subcategory “Poor working environment”). Following these refinements of the coding and categories, the Cohen’s kappa inter-agreement was 0.96.

Finally, the main categories and grouping of subcategories were refined, and two main categories (“Violence and assaults” and “Death and Homicides”) were merged into one called “Death and distressing crimes” due to related content in subcategories. The remaining inconsistent categories and codings were then discussed between the two authors until mutual agreement on the codings of all 2960 descriptions was reached.

The final developed categories consisted of 5 main categories and 38 subcategories (Table 2), and only 2.4% of the dataset was defined as “non-categorized” (this mainly involved comments on the job, the survey, or very rare kinds of incidents only described one or two times in total and not similar in content with any other subcategory, e.g., a description of maltreatment of a dog or the experience of a bad treatment from a defense attorney in court). Category labels and definitions were translated into English by the first author including pieces of text exemplifying each category (see online supplementary Table 2 for detailed definitions of each subcategory, coding guidelines, and exemplifying quotations).

Researcher Characteristics and Reflexivity

According to the Standards for Reporting Qualitative Research (SRQR) described by O’Brien et al. (2014), we find it relevant to mention that the first and second authors are both former police officers from the Danish police force. The first author is also a psychologist, and the second author is presently a senior student in her master’s in psychology.

Ethical Considerations

The AYA project was presented to the National Ethics Committee in Denmark (reference no.: 20202000–216), but according to Danish legislation, studies based on surveys do not require approval by official Danish ethical or scientific committees. Participation was voluntary and based on informed consent, and data was handled in agreement with the Danish data legislation (journal no.: 20/41457). The study was conducted in agreement with the Danish Code of Conduct for Research Integrity and the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Findings

Descriptive Characteristics of the Sample

Of the 6040 police officers in the AYA cohort, 18.5% were women, and the mean age was 42.5 (Table 1). In the final study sample of 1659 officers, who wrote a comment in TQ2b, the gender distribution was not significantly different from the AYA cohort, while the mean age was 45.5 years. Data on gender and age of non-responders in the AYA project were not available, and no comparisons were thus made with the study population.

Table 1 Descriptive statistics of gender, age, and results on TQ2a and TQ2b in the AYA project
Full size table

Categories of CI in Danish Police Work

Of the identified 38 thematically different subcategories, 28 were identified as CI, 5 as general work stressors, and 5 as mixed experiences (see Table 2). The categories are further described below with exemplifying quotations. Personally identifiable details may have been altered or deleted in the quotes, and reference to the officers’ age and gender has been omitted to preserve anonymity.

Table 2 Categories of critical incidents in Danish police work 2021 identified by content analysis with percentages (%) and number of respondents (n) describing the incident
Full size table

Danger or Threat

In total, 23.5% of the CI descriptions contained an element of danger or threat. These descriptions were divided into 9 CI subcategories (no. 1–9 in Table 2). Descriptions of these incidents included:

“Attacked with a knife through an open door at a house call for domestic violence. I never stand in front of an open door when I ring the doorbell/knock”; “A violent intrusion searching for a dangerous perpetrator back in the early 90 s is still vivid in my mind but doesn’t bother me any longer”; “Injuries like whiplash, injury in the shoulder after an arrest, I could keep going”; and “After firing shots towards a bank robber, I can sometimes still get flashbacks about this case.”

Being faced with violent demonstrations/groupings was the most occurring type of CI under this main category (4.8%) and included incidents such as “(…) I was involved in 27 days of consecutive service in 1986 in the “Battle of Ryesgade” (The Squatting Movement), where there were many serious street fights and severe injuries among colleagues.”

Accidents

The main category “Accidents” (23.6%) was characterized by 6 CI subcategories (no. 10–15 in Table 2). Descriptions in these subcategories included:

“A child had drowned in my hometown. Luckily my colleague and I didn’t have children of our own yet at the time”; “(…)a mother falls out of the window of a hotel and we are standing there with a family in crisis (…)”; “An infant died in the home of a regular family. As a father this sometimes resurfaces from the back of my mind – however in a ‘sensible’ way.” The most occurring CI under this main category was severe car accidents involving dead or injured adults (6%), which included CI descriptions such as “Serious traffic accident where the person was hit so hard that his body was dismembered into several pieces and where his partner and small child arrived at the scene shortly after.”

Death and Distressing Crimes

The main category “Death and distressing crimes” was the largest (33.3%) and included 13 CI subcategories (no. 16–28 in Table 2). Here, the descriptions included:

“(…) I have found a 16-year-old girl who had hung herself. It made a strong impression since my own girls were not that old at the time”; “Been at a crime scene where a woman was killed, and the sight of an almost severed throat and bloody crime scene comes back to me once in a while”; “Nightmares about a mother who in a puerperal psychosis killed her own child”; “Crude neglect of a child that had fled away from home on a cold night. When I drive by the address, I think about how the child is doing and if it has been removed yet (…)”; “Editing IS-videos where I had to blur recordings with executions, killings, and “football playing” with chopped off heads and with many dead bodies in different versions of mutilations.”

Death notifications were the most occurring CI under this main category representing 6.8% of the descriptions. For example, an officer described the following:

“An incident, in which a man was on the way to the hospital with his wife who had gone into labor. Accident on an icy road, husband and unborn child died. Both the husband’s and the wife’s parents had to be notified. The couple had another child (…) and and and and ........................ and that was only the beginning of that notification.”

Most Frequently Reported CI

As seen in Table 2, there was some variation in the frequency of descriptions under each category. Among the identified 28 CI subcategories, the five most reported were “Death notifications” (6.8%), “Traffic accidents with one or more dead or injured adults” (6.0%), “Suicide — adults” (6.4%), “Other accidents than traffic accidents with dead or injured children/teens” (5.5%), and “Violent demonstrations/groupings” (4.8%).

Variation in the Degree of Detail and Described Psychological Reactions

There was a large variation in the degree of detail in the qualitative descriptions. Some officers elaborated and described CI with clear specifications of the year, place, circumstances, and subsequent psychological reactions, while others only wrote a few words such as “Death notifications” or “18th of May 1993” (referring to severe street riots in Copenhagen, in which police officers ended up shooting toward violent crowds). Some descriptions involved CI that were more than 30 years old, while other officers described CI experienced within the last year. The psychological reactions experienced in relation to the described CI also varied in intensity and overall experience of personal impact. Some described experiences where they had come to some sort of terms with the incidents, e.g., “not severe psychological impact, but however experiences that have become part of the baggage and which occasionally appear on the retina, without me being able to say, that it is a burden. It has simply become part of my story,” while others described experiences resembling PTSD re-experiencing symptoms, e.g.:

“I have tried times when I can only smell blood and feel brain matter between my fingers. I have had flashbacks about killed persons or incidents. I have experienced great uneasiness and excessively beating heart when I am in certain geographical areas. I have wet my bed because of incidents. Have sometimes been able to smell fires.”

Aggravating Circumstances

The content analysis in the present study was primarily focused on the identification of CI themes. However, during the reading and coding process, different aggravating circumstances appeared across the descriptions of CI. Incidents with dead or injured people (e.g., suicide, homicides, traffic accidents) appeared emotionally extra stressful when they involved very intense sensory impressions such as destroyed, smelly decaying bodies or body parts, or screaming victims. For example, one officer wrote about experiences with “(…) several suicides, in every imaginable way. Often done so that the person was severely mutilated, and sometimes unrecognizable.” Several officers also described some CI as particularly difficult when the circumstances or the people involved reminded them of aspects of their private life or members of their family — this was especially evident when the CI involved children, and the officer him/herself also had children, e.g., “An incident where a 4-year-old died in a traffic accident, who at the time was the same age as my child. It’s been 14 years.” Feelings of powerlessness, insufficiency, meaninglessness, and “what if”-thoughts were also described as emotionally burdening across CI categories. Furthermore, it was described as very stressful if the officer had to handle a CI (e.g., a lethal traffic accident) and at the same time (or shortly after) had to handle a death notification to bereaved relatives or the relatives’ sometimes very unexpected and intense reactions. As two officers’ described, “I will never forget the screams/despair of the bereaved relatives” and “It is hard to distance oneself.”

Finally, the experience of lacking support from the management or lack of competence, equipment, personnel, or assistance also appeared aggravating, as well as very unexpected or chaotic CI with a lack of time for mental preparation, or when officers experienced several CI within short periods of time.

Reportings on Factors Other Than CI

Police Stressors and Mixed Experiences

The fourth and smallest main category, “Police stressors” (5.5%), differed from the above-mentioned, as it did not relate to the unique content of police work. Instead, the 5 subcategories (no. 29–33 in Table 2) included descriptions of, e.g., bad working environments and inadequate management, such as “Transferred, where personnel policies and senior policies were not complied with, to an operative section with treble shifts after 20 years in investigation.”

The final main category “Mixed experiences” (14.1%) included 5 subcategories (no. 34–38 in Table 2) that differed from both each other and the above-mentioned in their thematic content, and the main part did not specify any CI. A substantial part (8.8%) involved comments describing the officers’ psychological reactions to CI without further specification of any CI, e.g., “Fear for the health and safety of family members in relation to completely normal everyday tasks like getting around in traffic.”

Discussion

The present study aimed to investigate core characteristics of situations that could be considered as CI in Danish police work. Based on content analysis of 2960 qualitative descriptions of incidents perceived to cause emotional reactions over prolonged periods of time, the current study found 3 main categories of CI which could be conceptualized in a total of 28 CI categories. Noteworthy of the present study was the great diversity of incidents that appear to affect officers over long periods of time. CI were not only consisting of serious threats or witnessing gruesome deaths. CI also consist of common and more routine police tasks, e.g., traffic accidents, handling of mentally unstable individuals, handling sudden deaths, handling cases of violence and sexual assaults, and making death notifications. We call these incidents common and “routine” since in many police functions they can more or less be expected to occur as part of the daily work, in contrast to, e.g., terrorist attacks or the experience of being threatened with life. Taken together, these findings underline the importance of focusing prevention initiatives not only on the aftermaths of more extreme CI but also on the daily exposure to emotionally stressful CI and thus on the potential accumulation of psychological strain due to various CI experienced over time.

Critical Incidents in Police Work

Comparison with International Literature

The identification of three main categories and 28 different subcategories of CI in Danish police work provides new and up-to-date insights into particular aspects of the police job that are critical in regard to police officers’ psychological well-being. The identified main and subcategories in the present study are consistent with category themes and checklist items in the international police literature on trauma exposure (see online supplementary Table 3, for a detailed comparison of the 28 CI categories with international trauma checklist items and categories).

Incidents with an element of danger or threat such as life-threatening situations, shootings, injuries sustained while on duty, and major incidents such as terrorism are similar to trauma event categories identified by Karlsson and Christianson (2003) and Carlier et al. (2000), as well as checklist items on CIHQ, PIS, PSS, and PTEC (Hartley et al. 2013; Miller et al. 2022; Spielberger et al. 1981; Violanti and Gehrke 2004; Weiss et al. 2010). For example, CI no. 1 (Violent demonstrations/groupings) are similar to item no. 25 on Spielberger et al.’s (1981) PSS (Confrontations with aggressive crowds) and Carlier et al.’s (2000) no. 5 (Riot squad under siege), and CI no. 3 (Colleague’s death/severe injury/danger or suicide) is similar to several items on the CIHQ (1, 2, 4, 6, and 18, all involving a colleague’s death or injury), no. 1 on PIS (Shooting of another officer), and no. 26 on PSS (Fellow officer killed in the line of duty). No. 6 on PTEC (Serious injury to public, self, or colleagues) and no. 8 (Physical assault including being the subject) are also comparable in content to CI in this main category, but the PTEC items are more broadly scoping and match several of the 28 CI categories across main categories (see online supplementary Table 3).

CI categories involving traffic accidents and other accidents are also similar to traumatic events identified by Karlsson and Christianson (2003) (no. 2 “Traffic Accident” and 5 “Accidents”) and Carlier et al. (2000), as well as item no. 4 on PIS (Seeing victims of a serious traffic accident) and no. 3 on PTEC (Road traffic collisions and rail incidents (e.g., suicides)). In Spielberger et al.’s (1981) PSS, two items also include victims of accidents (no. 11 “Exposure to death of civilians” and no. 57 “Exposure to adults in pain”), but as with PTEC, these items are more broadly formulated and therefore match several of the 28 CI categories across main categories.

Incidents including deaths and distressing crimes such as death notifications, physical or sexual abuse, suicides, and homicides are also similar to international checklist items and traumatic event categories. CI no. 16 (Death notifications) is similar to notification-specific items in both CIHQ (no. 26), PSS (no. 16), and PTEC (no. 5) and is identified as a traumatic event (no. 6) by Karlsson and Christianson (2003). In regard to CI category nos. 17–28 in the present study, these are consistent with global themes in the checklists and event categories, but these CI categories are more delimited and event-specific in the present study (e.g., no. 17 and 22 are specific only to incidents of suicide either involving adults or children). For example, Spielberger et al.’s (1981) PSS, PTEC, and Carlier et al. (2000) operate with a more broadly scoping “children” category covering all severe incidents involving children, which is different from the division of CI in the present study (see online supplementary Table 3).

Only a few themes involving animal neglect or racial conflicts (e.g., item 33 on CIHQ and 40 on PSS) are not similar to the content of the 28 CI categories found in the present study. Further, the CIHQ has no items involving accidents, and two-thirds of the CIHQ (23 items) relates to life-threatening or dangerous situations (Weiss et al. 2010). This weighing of event themes in the CIHQ differs from the pattern found in the 28 CI categories in the present study. Here, over half (56.9%) of the officer descriptions involved either tragic accidents, deaths, or distressing crimes, and four of the five most frequently described CI were accidents or incidents related to death or distressing situations. The frequency levels of the 28 CI indicate how often the CI occur in the data material and represent CI that are experienced as burdensome by a large part of police officers in the present sample. It illustrates that emotionally burdening CI are not only related to life-threatening situations, major threats, or lethal catastrophes but also less dramatic non-threatening events. This finding is consistent with the results from Miller et al. (2022). They found that different tragic and distressing events covered five of the most frequently described, and thereof most common, events experienced by English police officers (no. 1–5) in the PTEC.

The same pattern of categories was identified by Karlsson and Christianson (2003) in which distressing (and not threatening) event types covered at least 61.8% of the described events including traffic accidents, murder, accidents, investigations, suicide, death notifications, and taking children into custody. Similar incidents were also described as having “stuck” and caused distress by 17 out of 19 English police officers in a small qualitative study by Evans et al. (2013) investigating police officers’ experiences of supportive and unsupportive social interactions following traumatic incidents. Thus, the burden of being witness to tragic accidents or distressing actions of others can be perceived as very emotionally demanding and possibly also traumatizing among many police officers. Together, these results consistently highlight the importance of recognizing “routine” and common police assignments as potentially distressing and emotionally burdensome.

Comparison to the Danish Police Study 1993

The trauma questions (TQ2a and TQ2b) used in the present study were replicated from DPS-1993. This allows us to compare the identified 28 CI categories with the categories of distressing police events that were identified by Ibsen (1997) and give unique insights into potential historical changes in police-specific CI over a 30-year period. The relative number of officers, who had experienced CI that still affected them emotionally, had increased from 25.9 in DPS-1993 to 46.5% in the present study. At first sight, this could indicate that CI experienced in this context might have changed in character and become more emotionally strainful. However, when comparing the 28 categories of CI with the types of distressing events identified in Danish police work in 1993, we find a substantial overlap in CI themes and content. This indicates that police officers’ experiences of CI have had a continuous and general quality over decades (see online supplementary Table 1 for the types of distressing events identified in police work 1993).

Considering incidents involving danger or threat, event type nos. 1–9 (the whole main category of “Acute physical threat” in the 1993 dataset) and nos. 29–31 involving threats against the respondent and/or his/her family are very similar to or defining content of the CI category nos. 1–7 and 9 in the present study. For example, no. 9 in the 1993 dataset “Threatened with a weapon” was similar to category no. 4 “Threatened with a weapon/been shot at” and descriptions of “Risk of HIV infection” (event type no. 6 in the 1993 dataset) was included under no. 2 “Threat to own life or health — other (not involving a weapon)” in the present study (see online supplementary Table 2 for detailed definitions of the 28 CI categories). Only the CI category no. 8 “Terror or massacres,” which included 1.2% (n = 36) CI descriptions in the present study, was not similar to event categories in the 1993 dataset.

The main categories of “Accidents” and “Death and distressing crimes” identified in the present study were different in their labels and division of subcategories from the main categories of “Children” and “Strong sensorial impressions” identified in the 1993 dataset. However, if these main categories were merged in each dataset and thereafter compared, the global themes and the different CI categories and event types were very much overlapping in content and definitions with summed frequency levels being 56.9% in the present study and 51.6% in the 1993 dataset. Under these main categories, all 18 event types (nos. 10–27) identified in the 1993 dataset were either very similar in content to CI categories in the present study (nos. 10, 11, 12–16, 17–21, 23–24, 26–27, 32) or overlapping in overall thematic (nos. 22, 26).

Instead of reflecting a change in police CI characteristics, the increase from 25.9 in DPS-1993 to 46.5% in officers recognizing being psychologically affected by CI experienced during their career may reflect a change in police culture and identity. Ideals and values of resiliency, bravery, and heroism may have been more dominant in the police culture 30 years ago. This could reduce the reporting of strong emotional reactions due to the fear of being stigmatized and considered weak, vulnerable, and unfit for police work (Andersen and Papazoglou 2014; Evans et al. 2013; Ibsen 1997; Papazoglou and Tuttle 2018; Royle et al. 2009). Since 1993, the societal, political, and organizational acknowledgment of the police job as a “high-risk job” has increased both internationally and nationally alongside the publication of numerous research studies reporting a heightened prevalence of mental and physical health problems among police officers (e.g., Syed et al. 2020; Brewin et al. 2020). It may now be more permissible to acknowledge being affected by CI compared to earlier. However, further research into the Danish police culture, attitudes toward mental health, and help-seeking behaviors is needed.

Implications

Clinical and Preventive Perspectives

The identification of many overlapping CI categories with trauma events and themes in both the international studies and the Danish study of Ibsen (1997) indicates that the findings in the present study can be considered as describing general and core characteristics of CI in police work. The results highlight the importance of recognizing “routine” and common police assignments as potentially distressing and emotionally burdensome. The findings also underline the importance of focusing prevention initiatives not only on the aftermaths of severe CI such as terror attacks, large disasters, or life-threatening situations but also on the accumulation of psychological strain following various common and routine CI experienced over time.

Hence, it is important to educate and mentally prepare police officers for the emotional burden that CI experienced in the line of duty might cause. By consistently normalizing emotional reactions following exposure to routine (and not only dangerous) CI on all levels of the police educational and preventive program, it might become easier for the individual police officer to acknowledge potential trauma-related reactions and seek support. The identified 28 CI categories in the present study could be used in educational programs in the training and coping of these emotionally demanding CI situations and in psychoeducation and recognition of possible emotional reactions of grief or post-trauma symptoms.

That nearly 50% experience prolonged psychological reactions following exposure to CI shows that a wide range of both dangerous and distressing CI most likely will have a psychological impact on a large quantity of officers during their police career. Identifying situations where support might be needed is therefore important in occupational prevention strategies. Also, recognition of the risk of late-onset PTSD seems to be of particular relevance since the CI described are dating back 1–30 years while still having an emotional impact today. Studies have indicated that late-onset PTSD is highly prevalent among first responders (Utzon-Frank et al. 2014) and that stressful events later in life are a risk factor for late-onset PTSD (Bonde et al. 2022). The study findings that many officers are experiencing a psychological impact of past CI but continue to work in policing, where they are in continued exposure to CI and other stressors, may therefore put them in an increased risk of delayed onset PTSD.

The identified 28 CI categories in the present study could also form the basis for a systematic screening tool of CI exposure in police work. Systematic screening could help identify and initiate help to the officers who experience excessive exposure to CI. Considering cumulative strain, it is important to be able to focus prevention initiatives on highly exposed groups of police personnel. Using a screening tool, focusing on CI exposure could be a valuable supplement to existing screening of police personnel well-being, where underreporting of mental health problems may be a problem within this occupational group (Berg et al. 2006). Further, screening of CI exposure could supplement the procedures that refer personnel to existing preventive initiatives such as debriefing and psychological crisis intervention. Finally, a CI screening tool may assist in the qualification and documentation of trauma exposure in the course of police duty and thereby assist in the assessment of the trauma criterion of the PTSD diagnosis. This may be particularly important in a Danish context, where ICD-10 (and soon ICD-11) is used as the primary diagnostic tool, as it does not, in the same specific way as DSM-5, recognize the burden of repeated critical work tasks as being potentially traumatic (APA 2013, see Criterion A4). Instead, the formulation of the trauma criterion for the PTSD diagnosis in ICD-11 is more general and particularly focused on exposure to an event or series of events of an extremely threatening or horrific nature (WHO 2019, pp. 6B40). Considering the results of the present study, the findings of Miller et al. (2022), and the meta-study by Sørensen et al. (2022), the evidence is however growing that PTSD symptomatology among police officers does not only develop following exposure to an extremely threatening or horrific event or series of events but may also follow as a result of daily police work.

Research Perspectives

The present study was situated in a Danish police context and provides in-depth knowledge of CI outside the context of the US and UK. The identified 28 CI categories in the present study include the most frequently experienced CI among Danish, Faroese, and Greenlandic police officers and depict many routine police assignments that are common and recognizable in the broad Danish police force. Thereof, the list of 28 CI can help improve the adjustments and use of foreign-developed police trauma scales in future police studies. As mentioned above, the identified CI categories can further form the basis for the development of a police-specific CI scale that can be used and tested as a screening tool not only in Denmark but also in other police contexts. This would permit future studies to assess potential cross-national differences in CI exposure.

Given that the same types of CI within one subcategory can be perceived very differently and not always as emotionally stressful (e.g., one suicide gets “stuck” in memory while another does not), it would also be relevant to explore the potential aggravating circumstances that make some CI more burdensome and traumatic compared to others. Potential aggravating circumstances have mainly been identified as side products of analyses as seen in the present study and the studies of Evans et al. (2013), Karlsson and Christianson (2003), and Miller et al. (2022). Further research is needed to increase our understanding of these aggravating circumstances and their potential role in the process leading from exposure to CI to the development of strong emotional reactions. Furthermore, it could be relevant to conduct studies focusing specifically on the prolonged psychological reactions following CI. Results in the present study showed that police officers described CI dating back to the mid-1980s as still affecting them emotionally and that descriptions of prolonged emotional reactions varied from the experience of severe trauma-related symptoms such as flashbacks, hyperarousal, or a PTSD diagnosis to “just” having clear memories of CI.

Several officers described how these clear memories of CI often resurfaced without the officers’ actively trying to recall the incident. There seemed to be an acceptance of these memories and “involuntary learning points” (as one officer called it), and they were not necessarily perceived as having a negative psychological impact. However, the line between “just” having thoughts and memories of CI and having actual trauma-related and emotionally demanding flashbacks is subtle and maybe the early identification of these “accepted” but potentially burdensome involuntary memories could help prevent them from becoming problematic and more strainful symptoms of mental health problems later in life. This point is in line with research findings regarding PTSD symptom development and PTSD symptom connections. According to Wright et al. (2021), reducing re-experiencing phenomena and intrusive symptoms can be useful not only in its own right but may also have positive downstream effects since they are hypothesized to be a mediator of other PTSD symptoms (Armour et al. 2017; Greene et al. 2018). Furthermore, Bryant et al. (2017) conclude that intrusive memories and associated reactivity are centrally associated with other PTSD symptoms in the acute phase following trauma exposure, potentially pointing to the utility of addressing trauma memories in early intervention and prevention strategies.

Methodological Considerations

Participation in the AYA project was voluntary, and selection bias might therefore be a problem. Particularly, the risk of selection bias should be considered in regard to TQ2b and the content analysis, as this was based on comments from police officers that voluntarily and deliberately chose to elaborate on their TQ2a answer. However, the descriptive statistics of the AYA cohort on mean age (42.5) and gender distribution (18.5% women) fully correspond with online available statistics of the Danish police personnel in 2021 (18.3% women and a mean age of 41.6) (Rigspolitiet 2022). And as seen in Table 1, the gender distribution and mean age of the group of officers writing a comment on TQ2b are comparable with the AYA cohort. Also, the amount of qualitative descriptions of CI situations (2960), as well as the profound similarity in identified CI categories across decades and across CI described in studies based on samples from other countries, indicates that the study results are probably more general and describing core factors of police-specific CI and not mainly a result of systematic errors caused by the selection of respondents.

Another limitation is the open formulation of question TQ2b: Any comments or elaborations? Had this question been more directive, it could have given a more uniform understanding of what had to be answered. The high proportion of answers that did not describe any CI indicates that some respondents could have misunderstood the question intention (e.g., 8.8% elaborated on how they experienced being psychologically affected by CI without describing actual CI). However, a more directive item would entail the risk of excluding important data. Future studies exploring qualitative descriptions of CI exposure among police officers will benefit from testing open-ended survey items using cognitive validation (Brod et al. 2009).

Finally, the two authors’ background as police officers might have limited the inductive analytical process in the content analysis. However, the comments on TQ2b consisted of both short sentences and page-long comments, which could include police-specific terms, keywords, and abbreviations. Therefore, the authors’ experiences with police work, police culture, and police-specific CI were in combination with their understanding of psychological processes and scientific research practices considered an advantage in the analysis and interpretation of the comments and CI descriptions.

Conclusion

Exposure to critical and potentially traumatizing incidents appears to be an unavoidable condition of being a police officer, and it carries an increased risk of developing mental health problems. Definitions and understandings of police-specific CI, however, vary with different conceptualizations, measures, event checklists, and tools, with only few of them originating from a Danish or similar Scandinavian police context. Based on a thorough exploration of 2960 qualitative descriptions of incidents perceived to have caused long-lasting emotional impact, the present study identifies a diversity of 28 CI categories, of which the main part involves common police assignments of a tragic or distressing, but not threatening nature. These experiences of CI in Danish police work appear to have a continuing and general quality over decades and are comparable to CI identified in other countries. Altogether, this highlights the importance of recognizing routine police assignments as potentially distressing and emotionally burdensome. The 28 CI categories show potential as categories for future screening of CI exposure in police work. Furthermore, the 28 CI categories could be useful in police training programs, in psychoeducation, and as a supplement to existing preventive screening initiatives in the police organization.