Identifying the evidence base of interventions supporting mental health nurses to cope with stressful working environments: A scoping review

Aim: To scope the evidence on interventions used to help mental health nurses cope with stressful working environments. Background: Nursing managers may implement interventions to support mental health nurses cope in their role. However, the evidence supporting these interven tions has not been recently reviewed. Methods: A scoping review was conducted which entailed searching and selecting potential studies, undertaking data extraction and synthesis. Results: Eighteen studies published since 2000 were identified. They employed dif ferent designs, ten used quasi- experimental methods. Interventions involving active learning appeared beneficial, for example stress reduction courses and mindfulness. However, small sample sizes, short follow- up periods and variation in outcome meas ures make it difficult to identify the optimum interventions. No studies have consid ered cost- effectiveness. Conclusion: There is some evidence that mental health nurses benefit from inter ventions to help them cope with stressful working environments. However, higher quality research is needed to establish the effectiveness and cost- effectiveness of different interventions. Implications for should provide and encourage mental health nurses to engage in active learning interventions, for exam ple mindfulness to help them


| INTRODUC TI ON
Mental health nurses face stressful working environments because of their role in delivering support to people experiencing mental health issues. The role can entail significant emotional labour (Edward et al., 2017), aggression from service users (Jalil et al., 2017) and demanding workloads (Yanchus et al., 2017). This can result in workplace stress (Lanctôt & Guay, 2014), reduced well-being (Edward et al., 2017) and burnout (Morse et al., 2012). Furthermore, it can negatively impact on the quality of care delivered to service users (Roche et al., 2011) and result in increased absenteeism and decreased workplace retention (Lamont et al., 2017).
In response, nurse managers are implementing interventions to support mental health nurses to cope with working in stressful environments (Wood et al., 2019). Interventions include mindfulness practice (Munn, 2018a), resilience training (Foster, Shochet, et al., 2018) and communication skills courses (Traynor, 2017). We define coping as mental health nurses themselves or organisations adopting methods to support people working in mental health services (which are considered stressful environments), so coping is a process rather than a dichotomous variable of someone either coping or not coping (Carson & Kuipers, 1998). Often interventions focus on addressing a specific psychological construct including burnout, resilience or well-being but are comparable in terms of seeking to support mental health nurses to cope with stressful working environments (King & Rothstein, 2010). Furthermore, research has found that relevant psychological constructs are inter-related, so interventions which reduce burnout can also result in people experiencing improvements in their levels of stress and resilience (Lee et al., 2019).
Research has evaluated the impact of interventions supporting mental health nurses. These were reviewed by Edwards and Burnard (2003), who identified that interventions including stress management courses, relaxation sessions and training in psychosocial skills appeared effective. A more recent meta-analysis by Dreison et al. (2018) considered burnout interventions for mental health professionals and had similar findings. Foster et al. (2019) undertook a review focused on resilience and identified that resilience training programmes appear beneficial. Despite these two reviews being conducted a decade after Edwards and Burnard (2003), they all reported similar weaknesses with the evidence base. These included studies having small samples, short follow-up periods and studies using different outcome measures, making comparisons difficult. Whilst these three reviews are useful, they have limitations. Edwards and Burnard's (2003) review was conducted almost twenty years ago; Dreison et al. (2018) do not specifically focus on mental health nurses and Foster et al. (2019) only consider resilience interventions. Consequently, there is a need to understand the breath of literature that evaluates interventions to support mental health nurses to cope with stressful working environments, irrespective of the specific psychological construct that they are aimed at. This will help nurse managers understand the evidence on interventions that they may be implementing whilst also identifying priorities for future research.

| Aim of the review
The aim of the review is to scope the literature to identify the nature of evidence evaluating interventions supporting mental health nurses to cope with stressful working environments.

| ME THODS
A scoping review was conducted because we wanted to explore the nature of evidence as well as considering what future research is needed (Armstrong et al., 2011;Grant & Booth, 2009). Scoping reviews can be undertaken iteratively, developing the parameters of the review as literature is identified. We conducted the review in 2020, drawing upon established guidance (Arksey & O'Malley, 2005;Colquhoun et al., 2014) and reporting standards (Tricco et al., 2018) (Appendix S1).

| Stage 1-Identifying the research questions
The research questions were to scope the nature of evidence on interventions that may help mental health nurses cope with stressful working environments and to identify areas of future research.

| Stage 2-Identifying the relevant studies
We iteratively refined the inclusion and exclusion criteria after performing the search. Initially, we were unsure about the extent of research on mental health nurses. Consequently, a search process was designed that also identified literature related to any type of nurse or mental health professional, for example psychologists. This gave us the potential to consider the relevance of studies focused on similar occupational groups if there was a lack of literature on mental health nurses.
Database searches were undertaken in MEDLINE, EMBASE, PsycINFO and CINAHL. Search terms were related to workplace, potential interventions, for example yoga, psychological constructs such as stress and staffing terms including team and staff (Appendix S2 provides an example of the search). The search sought to identify the breadth of literature rather than be exhaustive (Grant & Booth, 2009). Consequently, the search strategy did not undergo the extent of refinement that would be undertaken for a systematic review (Morris et al., 2016). We also conducted reference checking of identified reviews for relevant primary studies. A lack of researcher resource prevented us from utilizing other search techniques.
As the search identified a number of studies involving mental health nurses, we amended the inclusion criteria to focus on them.
Other inclusion criteria were studies including information on an intervention's impact, be published between 2000 and June 2020 and be in English.

| Stage 3-Study selection
AF screened all the titles and abstracts, undertaking full-text review on potentially relevant studies. EW and MC provided support including giving a second opinion about some studies' eligibility. We did not have the resources to have two researchers undertake study selection. The results were reported using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart (Moher et al., 2009).

| Stage 4-Charting the data
AF undertook data extraction using a standardized form to chart the data. We considered study design, population, sample size, setting, conduct, findings and reported limitations. We also extracted information on the interventions including content and delivery method.
We reflected on the quality of studies because it has implications for future research. However, systematic quality appraisal using a specific criteria was not undertaken because it was a scoping rather than systematic review (Munn et al., 2018b).

| Stage 5-Collating, summarizing and reporting the results
Narrative synthesis was used to collate the extracted data (Barnett-Page & Thomas, 2009).
The quantitative studies were of variable quality. The majority had small sample sizes and short follow-up periods. For example, Edwards (2015) and Flarity et al. (2016) had 10 or less participants.
Most of the studies did not consider the long-term impact of interventions, with only Bernburg et al. (2019) measuring outcomes for longer than six months. An additional challenge in comparing findings was that studies used different primary outcome measures including the Maslach Burnout Inventory (Alenezi et al., 2019;Edwards, 2015;Ewers et al., 2002) the Nursing Stress Scale (Yang et al., 2018) and measures designed for the study (Ghazavi et al., 2010;Henshall et al., 2020).
The majority of studies were based in adult inpatient wards and a small number focused on mental health nurses working in specialist services including forensics (Henshall et al., 2020) and dementia care (Edwards, 2015). A small number of studies targeted the interventions at people with higher levels of stress (Yang et al., 2018) or those who had experienced workplace violence (Hsieh et al., 2020;Nhiwatiwa, 2003). Some studies reported recruitment issues because nurses felt stigmatized for accessing support, worrying that nurse managers and colleagues would think that they cannot cope with their role (Henshall et al., 2020).

| The interventions
All of the studies focused on interventions aimed at individual nurses. The majority of studies evaluated active learning interventions, for example assertiveness training and mindfulness, where nurses had the opportunity to practise coping strategies. Nine studies evaluated group-based courses where participants shared their concerns, learnt about specific mental health constructs, received peer support, developed coping strategies and practised relaxation techniques (Abdelaziz et al., 2020;Alenezi et al., 2019;Bernburg et al., 2019;Flarity et al., 2016;Foster, Cuzzillo, et al., 2018;Foster, Shochet, et al., 2018;Henshall et al., 2020;Safarzei et al., 2016;Sailaxmi & Lalitha, 2015). Four studies focused specifically on relaxation techniques including mindfulness (Edwards, 2015;Hsieh et al., 2020;Rush, 2018;Yang et al., 2018). Two studies evaluated supervision-based interventions (Lakeman & Glasgow, 2009;Olofsson, 2005) The studies evaluated interventions that were heterogeneous in their length and nature of delivery (Table 1). Variation in length ranged from six all day workshops (Henshall et al., 2020) to a twohour reflection group (Olofsson, 2005). Fourteen studies focused on in-person group-based interventions whereas others evaluated online interventions (Rush, 2018). One reason for online delivery was that nurses did not always have time to be released from their duties to attend training (Lakeman & Glasgow, 2009). Hsieh et al. (2020) reported no difference in outcomes between in person and online delivery.

| Identified impact
The identified studies generally reported that mental health nurses appeared to benefit from receiving support, with all of the interventions besides the booklet (Nhiwatiwa, 2003) having a positive impact (Table 1). Benefits included improvements in resilience (Foster, Shochet, et al., 2018), assertiveness (Abdelaziz et al. (2020) and reductions in stress (Yang et al., 2018). One of the larger studies: Alenezi et al. (2019) identified that mental health nurses receiving a burnout prevention group programme experienced a statistically significant reduction in their burnout compared to the control group at 1-month post-intervention (p = <.001) and 6 months postintervention (p = .04) (measured by the Maslach Burnout Inventory). Bernburg et al. (2019) also identified that participants experienced a statistically significant improvement in their stress levels when engaging in a mental well-being group at 3 months (p = <.001), 6 months (p = <.001) and 12 months post-intervention (p = .01) (measured by the Perceived Stress Scale). The only intervention which did not appear beneficial was the booklet, the intervention group experienced a worse outcome than the control group (p = .03) (measured using the Impact of Events Scale) (Nhiwatiwa, 2003). Undertaking a metaanalysis would be useful to understand the relative effectiveness of different interventions.
There appears a need for studies which consider the longerterm impact of interventions. This is because there was some evidence that the benefits of an intervention decreased over time.

| D ISCUSS I ON
The aim of the review was to scope the evidence on interventions used to support mental health nurses cope with stressful working environments. Eighteen studies were identified which used different methods to evaluate a range of interventions including stress management courses and mindfulness practice. The studies indicated interventions involving active learning appeared beneficial.
However, many of the quantitative studies had small sample sizes, short follow-up periods and did not consider cost-effectiveness.
These factors indicate that whilst nurse managers may want to implement interventions, there is a need for further research to identify the most effective interventions.
Active learning interventions appeared to result in improvement in burnout, resilience and stress. The need to involve active learning reflects the findings of Edwards and Burnard's (2003) review. Furthermore, other reviews of different occupational groups have also identified the importance of active learning (Askey-Jones, 2018; Dreison et al., 2018;Hamilton-West et al., 2018;Richardson & Rothstein, 2008). However, it is unknown which specific interventions are most effective or the optimum delivery models, for example whether the same benefits are derived from 2 or 8 sessions.
Consequently, there is a need to undertake a meta-analysis so that recommendations can be made to nurse managers on the relative value of different interventions. At present, the identified studies focus on interventions aimed at supporting individual nurses rather than organisational changes such as reduced caseloads. Nurse managers may want to consider the impact of organisational approaches because mental health nurses feel these approaches are important (Itzhaki et al., 2015). Supporting this rationale is evidence that organisational approaches such as different shift patterns are effective for general nurses (Barrientos-Trigo et al., 2018).
Some studies targeted mental health nurses who had higher levels of stress or who had been assaulted at work whereas other studies were aimed at any mental health nurses. Further research is needed to establish which approach has the greatest impact. This is because Dreison et al. (2018) identified that interventions may have greater impact when targeted at mental health professionals experiencing higher levels of burnout. However, Johnson et al. (2018) propose that all mental health nurses need support because the role generally involves high levels of stress and burnout because of the working environment. Furthermore, the studies were generally based on inpatient settings and nurse managers should consider how relevant the findings are for community-based mental health nurses, who may experience specific stressors, for example lone working (Edwards et al., 2001). Issues were identified that mental health nurses were concerned that accessing support attracted stigma. This is consistent with Knaak et al. (2017) in respect of health care professionals generally. Consequently, future studies need to consider the acceptability alongside the effectiveness of interventions.
We identified similar issues with the evidence base as Edwards and Burnard (2003) including small sample sizes and short follow-up periods. This indicates that the quality of evidence has not evolved and there is a need for further quantitative studies that utilize larger sample sizes and collect outcome measures for longer than 6 months. The latter is important because nurse managers want interventions which provide longer-term benefits (Wood et al., 2019).
The variety of outcome measures used by studies including researchers developing measures specifically for their study makes it difficult to compare the findings of studies. This challenge was also identified in Edwards and Burnard (2003) and Dreison et al. (2018).
Consequently, it is recommended that a Core Outcome Set is developed for future studies to utilize (Prinsen et al., 2014). None of the identified studies considered cost-effectiveness. This absence of cost-effectiveness has also been identified in relation to other occupational groups (Pieper et al., 2019) These two studies were evaluating the same intervention.
reviews should do this, so that nurse managers can understand the nature of evidence underpinning potential interventions.

| CON CLUS ION
There have been a number of studies identifying interventions which can support mental health nurses develop their coping mechanisms.
Further primary research along with meta-analyses is needed to establish the most effective interventions including the optimum delivery models, the cost-effectiveness of interventions and whether they have longer-term benefits.

| IMPLI C ATI ON S FOR N UR S ING MANAG EMENT
Emerging evidence indicates that managers should encourage and provide opportunities for mental health nurses to engage in interventions which support nurses develop techniques to help them cope with stressful working environments. However, to date it is not possible to make recommendations on which interventions are most beneficial, the optimum delivery models, whether nurses sustain improvements, the cost-effectiveness of providing support and whether some nurses should be prioritized for support.
Furthermore, there is a gap in the literature evaluating organisational-and management-level interventions. Given the limitations of the current evidence base, nurse managers are encouraged to evaluate any interventions that they deliver.

ACK N OWLED G EM ENTS
We would like to thank Susan Ridgway for supporting the review.

CO N FLI C T O F I NTE R E S T
All of the authors declare that we have no conflicts of interests.

E TH I C A L A PPROVA L
Ethical approval was not required for this paper.

DATA AVA I L A B I L I T Y S TAT E M E N T
As it is a review all the data is drawn from articles available in the public domain.