Evaluating Return-to-Work Programmes after Critical Incidents: a Review of the Evidence

In most countries, employers have a responsibility to ensure the safety and well-being of their employees, including those who work in hazardous occupations. Policing can be a hazardous occupation and officers can face physical threats to life. Yet in the police service, there appears to be a dichotomy between treating officers involved in critical incidents solely in terms of serious clinical issues (e.g. PTSD) whilst simultaneously setting a strong expectation that these officers will return to work quickly and with little effect on their work performance. We review the research literature on return-to-work programmes to establish how much the effect of critical incidents on officers’ work performance is being considered and to identify future research directions. Findings indicate an emphasis on improving mental health but a lack of research into occupational outcomes. They also reveal distinct promise for peer-led reintegration programmes.


Introduction
Policing is recognised as an extremely stressful occupation which exposes officers to a high number of critical incidents over the course of their careers (Lennie et al. 2020). In a survey of Finnish police officers, over 80% of officers reported their work was very or extremely emotionally demanding and more than 90% recognised a link between their exposure to critical incidents (i.e. traumatic events) and their mental health (Andersen et al. 2015). Only 1% of those surveyed said they had never experienced a critical incident during their normal workdays, with over half saying they were exposed to critical incidents for more than 20% of their work time. Similar results have been found in other countries. For example, 80% of Ghanian officers have experienced a potentially traumatic event, such as traffic accident or assault, during their routine work and a fifth have experienced five or more (Barnett et al. 2021).
Unfortunately, officers report that they often do not know how to support themselves after a critical incident and rarely receive the emotional support they need (Ricketts 2002).
There is strong evidence from surveys and interviews that officers want access to programmes to help them deal with their responses to critical incidents and return to work safely and with confidence (Andersen et al. 2015;Lennie et al. 2020). It is important from an organisational or operational level to provide this kind of support or training to help officers return to work effectively after a critical incident: not just to avoid serious mental health problems, but also because of the ongoing effect that heightened stress and distress can have on performance. Stress is known to negatively affect performance, and this has been specifically demonstrated for emergency services professionals in critical incidents (Leblanc et al. 2012). Police officers' stress levels are associated with negative work-related issues such as burnout and absenteeism (Magnavita and Garbarino 2013;Rothmann 2008) and wider social stressors associated with police work are also known to increase psychological distress and emotional exhaustion, as well as turnover intentions (Adams and Buck 2010).
But what kind of support or training programme is effective? The approaches taken in different countries and even within different police forces within a country vary substantially. Some of these programmes may be founded on research evidence whilst others may be developed based on an individual understanding of best practice and often their effectiveness goes untested. They often include access to psychological services and an increasing number of programmes include a formal element of peer support for its potential benefits. These benefits are suggested to include easier access, building on established relationships and a better understanding of the job (Greenstone 2005).
In this paper, we review the literature on programmes or interventions that have been designed to support officers' return to or continuation with work after critical incidents. Our aim is to provide a summary and narrative integration of the state of the science to provide researchers and practitioners with an evidence base for future programme construction and evaluation, as well as to identify gaps in evidence.
We describe the content of different kinds of interventions and the outcomes that have been tested and identify the programmes or elements of programmes that have demonstrated effectiveness. Our emphasis is on quantitative studies that used psychological or physical measures of outcomes and we highlight specific measures for these outcomes that those evaluating these programmes may wish to use in the future. Although the focus is on police-specific findings, we also draw on related research with other emergency frontline workers who are exposed to critical incidents as a standard element of their jobs: for example, paramedics, firefighters and military personnel.
To achieve these aims, we conducted searches for relevant studies using the ProQuest and Psycinfo databases, as well as Google Scholar and individual publisher search engines. Search terms included those related to the focus of the research (return to work, reintegration, critical incident) combined with terms for types of interventions identified in our initial scan of the research (programmes, stress/trauma/crisis intervention/risk management, debriefing, EAP, mindfulness). We also searched for specific occupations, focusing on police but including search terms such as emergency workers.
In the following sections, we provide a narrative summary of the findings along with suggestions for example measures of key outcomes that can be used in future evaluation research. The summary is organised into two sections which represent the key elements of evaluations of return-towork programmes after critical incidents. The first provides a summary of programme content, distinguishing between secondary and tertiary interventions. The second section discusses how these programmes are evaluated and the outcomes that have been associated with them, from general measures of behaviour or stress levels through to occupationally specific measures.

Programme Approaches
Similar to generic occupational risk management, exposure to CIs is typically managed at three levels: primary, secondary, or tertiary. Primary interventions aim to reduce the risk of experiencing a CI and are outside the scope of this review. Secondary interventions focus on reducing the severity or impact of the event, for example by developing employee resilience to critical events or providing training to help them modify their responses to stress. Tertiary interventions involve treating the ongoing symptoms and effects of a CI, including rehabilitation to work. The programmes identified in this review vary substantially in their content, and we provide a summary of these approaches using these levels of classification.

Secondary Interventions
Secondary interventions are those that provide programmes designed to help workers to build skills and resilience that will buffer against the negative effect of CIs. These programmes are not used in the aftermath of a critical incident, but rather are intended to help employees to develop personal resources, skills and resilience to reduce the extent of psychological harm when they are later exposed to traumatic events. Theoretically, the assumption is that people can learn to buffer themselves from more extreme challenges of the job and develop effective strategies to cope with critical incidents (Arble et al. 2017).
The effectiveness of this type of programme appears to be generally, though not always, positive. The charity 'MIND' conducted a randomised control trial to evaluate a resilience intervention for emergency workers and found no effect or in some cases a reduction in well-being after the intervention, and this was unrelated to exposure to a CI during that time (Wild 2016). In contrast, a mindfulness-based resilience training programme (MBRT) demonstrated improvements in several outcomes including mindfulness, resilience and mental health (Christopher et al. 2016). This immediate positive effect was confirmed by another programme which included both emotional management skills (such as resilience training, coping strategies, mindfulness and emotional awareness) and professional skills training (optimal police tactics) and also demonstrated significant positive effects on officers a year later, including reduced anxiety and alcohol use and improved coping mechanisms (Arble et al. 2017).
An alternative approach to reducing the potential impact of a future CI is imagery training, which involves relaxation exercises, imagining and rehearsing responses to specific scenarios and group discussion (Arble et al. 2017). The aim of imagery training is to enhance participants' adaptive response to specific stressors and thereby reduce negative outcomes and enhance optimum performance. For example, an imagery training programme for police consisted of an educational session followed by 10 2-h small group sessions for developing relaxation skills as well as imagery training to mentally rehearse the skills, such as correct tactics, required to deal with specific incidents (Arnetz et al. 2009). This type of programme was found to have both short-and long-term positive effects. It improved performance in a realistic simulated critical incident (evaluated by an independent police officer), reduced stress and improved mood afterwards, when compared to the control group (Arnetz et al. 2009). Providing this training to junior officers was also associated with significant reductions in anxiety a year later (Arble et al. 2017).

Tertiary Interventions
Programmes that provide treatment or aim to reintegrate workers after exposure to a CI fall into the tertiary intervention domain. There are two main approaches here, expert-led programmes, using input from psychologists, psychiatrists, counsellors or other professionals, or peer-support, with peer mentors providing input or facilitating workshops.

Expert-Led Programmes
Programmes constructed around expert input vary substantially in terms of both their content and their aims. Some are intended to reduce specific symptoms of PTSD, such as aggressive outbursts (Leonard and Alison 1999), whilst others aim to educate personnel in coping with stress reactions more generally (Tuckey and Scott 2014).
Critical Incident Stress Management (CISM) is a multicomponent approach which can include small and large group interventions such as 'defusings' (short meetings straight after an incident which allow expression of initial feelings and offer information and support) or CISD (critical incident stress debriefing, a structured group discussion incorporating verbal expression and catharsis, group support and health education), individual counselling and family interventions and referral to mental health professionals where necessary. A meta-analysis of studies on CISM found that this type of programme had a significant, large effect on reducing PTSD symptoms (Everly et al. 2002).
This comprehensive CISM approach results in greater positive effects on employees than CISD alone (Richards 2001). For example, a group randomised study with US peacekeepers deployed to Kosovo measured PTSD symptoms prior to the intervention and then at 3-and 9-month follow-ups. CISD was compared to a simple stress management class and a control group and found not to have any effect on PTSD symptoms, though it was viewed more positively by participants than the stress management class (Adler et al. 2008). It should be noted that single session CISD was never intended to be used outside the more comprehensive CISM approach (Everly et al. 2002). Similarly, a review of single-session individual debriefing (in which the individual is encouraged to recollect the trauma soon after the event) called for an end to the use of this kind of debriefing as it has no effect on PTSD symptoms, depression or anxiety (Rose et al. 2002).
Because stress is a subjective experience, different people will have different needs in terms of the support they need to return to work. Although several studies controlled for these individual differences in impact of events when evaluating the effects of return to work programmes, only one attempted to provide tailored support. This study allocated paramedics and emergency medical technicians to three groups depending on the severity of the incident as assessed by the researchers (Macnab et al. 2003). The intervention for a mild incident included a trained 'listening ear' over the phone plus information on critical incident stress. The moderate intervention added a referral to debriefing and the severe intervention added small group defusing and subsequent debriefing. Unfortunately, take up of these options was low and, as the study used some interventions that are now known to be ineffective, it was not able to establish distinctions between the levels of intervention. But attempts to tailor the response to the individual and the CI remains an important area for future research.

Peer Support Programmes
Although many of the tertiary interventions provided to those who experience CIs are expert-led, an increasing number of intervention programmes make use of peer support. Peer support programmes have some advantages over expert-led programmes. Peer support can reduce the stigma associated with poor mental health and encourage helpseeking because the peers are already integrated into the work units. It can also take advantage of the social cohesion in already established teams and be delivered at a substantially reduced cost than expert-led interventions (Jones et al. 2003). A study on a peer-led Trauma Risk Management programme in military groups found that, whilst the programme did not improve attitudes towards mental health or have a measurable effect on psychological health, it did have a modest positive effect on occupational functioning (Greenberg et al. 2010).
Peer group practitioners may also be involved identifying psychological risk factors for those who have been through CIs and following up with management and referrals where necessary (Jones et al. 2003). This approach is promoted as avoiding 'unnecessary medicalisation' of normal reactions to traumatic events whilst boosting post-event resilience. Research in this area is still minimal but a doctoral dissertation which involved interviews with emergency responders (police, paramedics and firefighters) noted that the majority of them found debriefing processes to be especially beneficial when run by peers because there was a greater sense of trust and a belief that needing 'external' or professional help would have a negative effect on career prospects (Beaton 2003).
In summary, then, secondary interventions aim to develop mindfulness, resilience or specific skills that can be effective in reducing the impact of future CIs. Tertiary interventions are designed to support and rehabilitate those who have been through CIs. Expert-led programmes can provide research-led interventions whilst peer support programmes can help to reduce hesitation in asking for help as well as identify risks of which management may not be aware.

Programme Evaluations
In the following two sections, we provide a summary of the outcomes that have been measured in reintegration/returnto-work programme evaluations: general and occupational measures. Table 1 provides examples of specific questionnaires that have been used in previous research.

Behaviour
This outcome includes behaviours that are commonly affected by CIs and therefore would be expected to improve with an effective return to work programme. Studies have measured aggressive behaviour, sleep disturbances and maladaptive coping strategies such as increased drug or alcohol use. No effect of intervention on aggressive behaviour has been found (Adler et al. 2008), but a within-group MBRT evaluation showed improvements in sleep and emotion regulation (Christopher et al. 2016). Similarly, experience and expression of anger were reduced after a CISD intervention (Leonard and Alison 1999).
The effect of interventions on maladaptive coping strategies such as drug and alcohol use showed mixed results: not changing in one within-group evaluation of a debriefing programme (Busuttil et al. 1995), showing a decrease in one CISD evaluation study that used a randomised control design (Tuckey and Scott 2014) but an increase in another (Adler et al. 2008) and a decrease in a within group imagery programme (Arble et al. 2017). This lack of consensus is perhaps not surprising given the variety in programme content and approach identified above. Evaluations should ensure they are using measures of behaviours that are known to be both affected by CIs and susceptible to improvement by the specific intervention. For example, insomnia is a common response to the stress of CIs and mindfulness training is known to improve sleep quality.

Stress and Well-being
Outcome measures in this category include perceived stress, psychological symptoms of poor mental health or psychopathology and measures of well-being or quality of life. In studies evaluating officers' stress or anxiety responses to simulated incidents, some researchers used physical measures such as heart rate or salivary cortisol (e.g. LeBlanc et al. 2008), but none of the studies evaluating programmes themselves have reported physiological stress measures. Instead, they have used psychological measures such as those listed in Table 1, from general perceived stress scales through to questionnaires developed specifically for police. MBRT (mindfulness-based resilience training) showed a reduction in both general stress and organisational or operationally specific stress levels (Christopher et al. 2016).
Depression, PTSD and general psychopathology symptoms were reduced by a psychological debriefing intervention which included medication for the treatment of PTSD (Busuttil et al. 1995). But distress, depression and PTSD symptoms were not improved by a CISD (Adler et al. 2008), a single-session debriefing (Prince 2019) or an imagery-based programme focused on rehearsal of appropriate police tactics (Arble et al. 2017) and, similarly, there was no significant effect of CISD or debriefing on distress (Prince 2019; Tuckey and Scott 2014). General anxiety levels were reduced by an imagery-based intervention (Arble et al. 2017) but not by single-session debriefing (Prince 2019).
Although many studies sought to identify how interventions could reduce the negative effects of critical incidents, some consideration has also been given to potential positive outcomes, such as post-traumatic growth. Sattler et al. (2014), for example, found that the majority of firefighters in a cross-sectional study reported some post-traumatic growth and that greater growth was associated with attendance at critical incident stress debriefings and positive occupational experiences such as support and satisfaction, as well as improved coping strategies. The authors did note, however, that post-traumatic growth was most strongly related to individual personal resources and suggested that these resources enable firefighters to interpret events more positively as well as give and receive support. Improvements in self-rated quality of life have been recorded a month after CISD intervention (Tuckey and Scott 2014), but measures of general health showed more improvement following CISM than CISD (Richards 2001). Well-being did not improve in a general resilience intervention (Wild 2016).  (Hart et al. 1993)

Functioning and Support
In this category, we bring together findings on mental and social functioning (such as coping skills or family functioning) and related mindsets that can help to buffer against the negative effects of CIs and improve recovery, including resilience, mindfulness and sources of support for the individual. Emotional and avoidant coping styles are associated with higher levels of trauma symptoms in simulations, whilst task-oriented coping is associated with lower anxiety, but none are associated with performance (LeBlanc et al. 2008). Although many interventions include some education about adaptive coping strategies, only a few studies have evaluated whether participants are able to increase their use of more adaptive approaches. CISD and imagery-based training improved some adaptive coping strategies (Arble et al. 2017;Leonard and Alison 1999) but resilience training did not (Wild 2016).
A comparison of group-based and online-only resilience training programmes found no improvement in resilience compared to a control group (Wild 2016) but an MBRT intervention did increase participants' resilience (Christopher et al. 2016). This latter intervention also increased mindfulness, which is known to improve stress-related occupational outcomes (Jamieson and Tuckey 2017).
Social capital or support has a negative relationship with PTSD symptoms (Sattler et al. 2014) and this has led at least one study to attempt to improve access to social support through training. Unfortunately, it found that MBRT had no effect on family functioning, including affective responsiveness and communication (Christopher et al. 2016), perhaps because these outcomes were so distal to the intervention.

Occupational Measures
As noted earlier, most evaluations of return-to-work programmes focus on their effect on general measures of behaviour, stress and functioning and surprisingly few on specific occupational outcomes. In this section, therefore, we summarise research on these work-specific behaviours, attitudes and experiences, noting findings so far and the implications of this research for future work.

Work-Related Behaviours
There are several work-related behaviours assessed in studies that investigate the effect of stress at work that would be beneficial to include in evaluations of CI programmes. For example, research has demonstrated decreased work performance for emergency workers in simulated high stress as opposed to low stress clinical scenarios (Leblanc et al. 2012) and those who experience CIs (Rizwan 2018), whilst the imagery-based intervention discussed above was found to significantly improve police officer performance in a simulated CI (Arnetz et al. 2009).
Absenteeism represents one of the few 'objective' measures of work-related behaviour and is known to be associated with stress levels (Magnavita and Garbarino 2013) but we found only one evaluation study that had used this measure and unfortunately reported no improvement (Wild 2016). Another objective measure of organisational functioning is the number of disciplinary offences, and a CI peer support programme was found to improve this outcome for military personnel (Greenberg et al. 2010).

Work-Related Attitudes and Emotional Management
The effect of return-to-work programmes on work attitudes has not been directly evaluated, but engagement is known to be negatively related to occupational stress in police officers (Rothmann 2008) whilst burnout is associated with increased work hassles (Kohan and Mazmanian 2003) and PTSD symptoms (Sattler et al. 2014). One of the dimensions of burnout, emotional exhaustion, is higher for police officers who experience greater social stressors such as workplace incivility and results in increased intention to leave their jobs (Adams and Buck 2010). Burnout can be reduced by MBRT intervention (Christopher et al. 2016).
Interestingly, job satisfaction is not necessarily reduced by occupational stress in police, though it is negatively affected by reduced work engagement and increased burnout (Rothmann 2008), suggesting that police do not find immediate stress to be a significant problem but only the longer-term outcomes of raised stress levels.

Work Experiences
Police officers' experience of work hassles (stemming from policies and practices of the organisation or as a part of their duties) has been found to predict elements of burnout, whilst uplifts (positive work experiences) are associated with greater engagement (Kohan and Mazmanian 2003). Only one study investigated the effect of a CI intervention on participants' perceptions of how much the organisation supported them and found that it did not improve (Adler et al. 2008).
Whilst measures of personal and job resources have been used in studies investigating the link between occupational stress and work outcomes, finding that they contribute to post-traumatic growth for example (Sattler et al. 2014), no studies have evaluated the extent to which CI interventions might enhance these resources.

Summary
The evidence map in Fig. 1 presents a visual summary of the evidence reviewed in this paper, where the size of the circles indicates the number of studies. There is evidently a strong focus on stress and well-being outcomes across all types of studies but a dearth of research into the influence of CIs and CI interventions on work performance, attitudes or experiences. This reflects a concern with viewing CIs solely in terms of clinical effects (e.g. PTSD) without considering the effects on work. In terms of types of studies, peer-led tertiary interventions have had relatively little attention and, given their promising findings, are worth investigating further.

Conclusions and Recommendations
We start this section with some cautions. The accumulating evidence seems to indicate that single session debriefing (such as CISD) immediately following a critical incident can have a negative effect on psychological outcomes. Several studies show no effect of these kinds of programmes and there are some which demonstrate an increase in various measures of distress and symptomology. Combined with the finding in one of these studies that many firefighters experience post-traumatic growth following CIs, this may indicate that debriefing alone can interfere with the normal coping response and result in poorer outcomes. Single session debriefing is therefore not recommended for use as a stand-alone intervention following CI.
We can, however, make some clear recommendations for future programme content. Debriefing should be part of a larger programme, such as CISM. Mindfulness and resilience training are showing promising results so far: helping officers to develop their personal resources seems effective as part of both preventative and treatment programmes. Peer-led programmes are received positively by participants, so although there is obviously a clear need and place for expert-led input, effective programmes should include peer-led elements as well. And finally, it is worth considering participants' attitudes towards the programme itself. Valued programmes will encourage greater participation and engagement, whilst a negative reception may indicate that participants regard the content as irrelevant or perhaps indicate deeper problems with organisational culture, such as a stigma around asking for help (Lennie et al. 2020).
This review has also highlighted several areas for future research. First, most of the research focuses on the negative effects of CIs and adopts a clinical or medicalised model of human coping with CIs, with measures focusing on PTSD, distress, stress and so on. The promising findings around post-traumatic growth and quality of life indicate that future research would do well to include more positively framed measures as well.
Second, we need more research on how these programmes can be tailored to individuals. Although some studies in our review controlled for individual responses to CIs in evaluating the programmes, only one study attempted to investigate the utility of matching the programme to the individual's level of distress and unfortunately did not have Fig. 1 Return to work after critical incidents: evidence map enough participants to be able to make conclusive recommendations. Recent work on a low-dose mindfulness training programme in the military, for example, has found that participants' levels of well-being and distress on entering the programme had a significant impact on how well the programme worked (Roemer et al. 2020). Shorter programmes may therefore be more effective for those with less distress and this may underlie some of the mixed or negative findings in evaluations of brief programmes: they were simply not substantial enough to bring about improvements after serious CIs. This review has also found that the more extensive programmes, such as CISM, imagery-based training over 10 weeks and MBRT, had significant positive effects. In designing programmes for those who have experienced significant or repeated traumatic events, it is likely to be necessary to devote greater time and resources.
Third, there is very little research on how CIs affect work outcomes specifically. This is perhaps related to the medicalisation of CIs, so that the focus in most cases is on reducing mental illness. Yet for many organisations, an essential element in providing these programmes is to ensure that emergency workers are returning to work effectively. Absenteeism is a very broad-brush measure because of course not all absence will be related to the CI, but if the aim of these programmes is to enhance workers' effective return to work, it would be beneficial to include absenteeism measures in future studies. Alongside this, we recommend the use of 'presenteeism' measures which will account for workers attending work even when they should be off sick, perhaps due to organisational pressure or a culture which stigmatises mental health problems. Another useful outcome may be organisational citizenship behaviours (actions that a worker may engage in that go beyond the normal requirements of the job and contribute to the organisation as a whole) as this review found they are negatively related to stress levels, but again have not been used directly in CI programme evaluations. It is noteworthy that when measures of normal workplace functioning and well-being are used, the studies indicate that CI interventions can have remarkable positive effects.
Fourth, whilst we know a substantial amount about the negative effects of occupational stress and some of the resources that can buffer against these negative outcomes, few studies have been conducted to identify whether CI interventions can enhance these resources, nor do many programmes give much space to the development of these resources. A notable exception here is the use of mindfulness-based approaches which specifically aim to develop a well-proven personal resource. But future research could measure work resources, such as organisational, peer or managerial support, as well as personal resources that can feasibly be developed in an organisational context.
In conclusion, this review cautions against the use of 'easy-fix' approaches like single-session debriefing but paints a positive picture of the effectiveness of a range of more extensive CI support and return-to-work programmes. There is much that police forces can do to support officers dealing with the serious, complex or repeated CIs they face in their everyday work and ensure a healthy and sustainable return to work.
Funding Open Access funding enabled and organized by CAUL and its Member Institutions.

Declarations
Ethics Approval This article does not contain any studies with human participants or animals performed by any of the authors.

Conflict of Interest
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.