Understanding the factors that affect retention within the mental health nursing workforce: a systematic review and thematic synthesis

: There are over 41 000 vacant nursing posts across the United Kingdom’s National Health Service (NHS), with more people leaving the profession that joining it. Despite mental health being acknowledged as a priority area, some of the most signiﬁcant staff shortages are occurring within mental health services. Urgent action is needed to retain the mental health nurses (MHNs) currently in post to ensure the profession is ﬁt for purpose and aid future recruitment efforts. This review set out to identify the individual factors that affect the retention of MHNs. A systematic search of six databases was conducted (CINAHL, PsychINFO, MEDLINE, Web of Science (Core collection), EMBASE and the British Nursing Index). Studies were systematically screened for inclusion based on predetermined eligibility criteria. The studies were quality assessed using the Mixed Methods Appraisal Tool (MMAT). Findings were synthesized using Thematic Synthesis. A total of 23 studies consisting of a range of study designs were included in the review. Four key themes emerged from the synthesis: Individual characteristics, Working within mental health services, Training and skills and Work environment . The ﬁndings from this review suggest that MHNs encounter some factors unique to working in mental health services, which suggests that retention strategies should be speciﬁc to each nursing speciality. Beyond nursing speciality, the factors identiﬁed vary between clinical settings in mental health due to the differences in work environments and services they provide. Future studies should now set out to explore what factors exist in which clinical settings to inform better tailored retention strategies to generate better outcomes.


BACKGROUND
Nursing shortages represent a global concern, with the World Health Organization (2013) estimating a 12.9 million deficit in skilled health professionals made up of physicians, nurses and midwives by 2035. The most significant shortages are currently occurring in Africa and South-East Asia (World Health Organization 2020), whilst future nursing workforces in the UK, United States, Australia and Portugal are vulnerable to significant projected shortages from an ageing workforce and limited workforce planning policies (Buchan et al. 2015). Nurses make up 50% of the global healthcare workforce (World Health Organization 2020) and account for over a quarter of the UK's NHS staff (Beech et al. 2019). With increasing demand, insufficient resources and deteriorating job satisfaction, the NHS nursing workforce is overstretched and depleting rapidly, now accounting for 40% of staff shortages in hospital, community and mental health services (Beech et al. 2019). There are currently over 41,000 vacant nursing posts within the NHS (Buchan et al. 2019), with more people currently leaving the profession than joining it (Nursing & Midwifery Council 2017). Despite efforts to increase the number of nurses being trained, there was a decline in the numbers of nursing applications for nurse training in England in 2017 and 2018 (Buchan et al. 2019).
Mental health is a priority area identified within the 'NHS long-term plan', 'interim NHS people plan' and 'five-year forward view', yet the most significant staff shortages are occurring within the NHS mental health nursing workforce. Years of underinvestment and funding cuts to mental health services, nurses pay and training bursaries have given rise to these staff shortages (Unison 2017a), evidenced by a drastic 12% fall in MHNs between 2010 and 2017 in the NHS (Care Quality Commission 2017). Even with increased efforts to recruit and retain MHNs, their numbers rose by <0.5% in 2018 (Buchan et al. 2019). Due to poor retention and high turnover, a National Retention Programme (NRP) was launched in 2017 by the NHS. Under this scheme, clinical mental health staff turnover has decreased from 14.3% to 13.4%, but there is still a long way to go to resolve this current crisis (NHS Improvement 2019a). Poor retention reduces the ability of MHNs to meet service user's needs leading to dangerous consequences for service user care and safety (Unison 2017a) and is likely to deter people from wanting to join the profession.
Developing strategies focused on encouraging recruitment may be a futile occupation if retention is not addressed. Whilst nurse turnover can be beneficial by bringing in new perspectives, ideas and experience, high turnover rates often come with high economic and non-economic costs (Dewanto & Wardhani 2018). Instability of nursing staff leads to unstable nursing care, which increases the risk of care and treatment errors that can have grave effects of patient care and safety (O'Brien-Pallas et al. 2010), whereas relying on bank and agency staff, continuous recruitment and orientation and training processes constitutes a large economic burden (Jones & Gates 2007). In 2018, the NHS had reportedly spent £353 million on agency and bank staff to meet safe staffing requirements (Buchan et al. 2019). Urgent action is required to reverse this trend and reduce the use of bank and agency nurses (Royal College of Nursing 2015), before patient safety is put at greater risk (Unison 2017b). More immediate emphasis should be placed on retaining the staff already in post. By doing this, we can help to ensure the current profession is fit for purpose, and one that others aspire to join to aid retention and boost recruitment (Unison 2017b).
Job dissatisfaction appears to be strongly associated with poor retention (Happell 2008;Health Education England 2014;Ward 2011). Job satisfaction for MHNs is reportedly the lowest it has ever been, with many nurses emotionally and physically exhausted, leading them to consider leaving their profession (Unison 2017b). This results in a negative cycle where the consequences of poor job satisfaction and turnover, for example increased workload, pressure and working with temporary or inadequate numbers of MHNs, lead to even more turnover (Jones & Gates 2007). Recent events have seen us delve into the depths of a global pandemic which has now placed even more pressure on the mental health workforce, due to increasing demand and even more staff shortages as a result of shielding, sickness and self-isolation guidance (British Medical Association 2020). The additional pressures of the pandemic are likely to result in an exacerbation of the already significant international nurse shortages (Turale & Nantsupawat 2021).
Much of the literature surrounding nurse turnover and retention is for adult nursing (Chan et al. 2013;Halter et al. 2017;Moseley et al. 2008). Whilst there is some research focused on MHN retention, there has not yet been a systematic review consolidating the evidence on factors that affect retention within this workforce. With the concerning consequences of poor retention on patient care and safety, it is imperative to get a better understanding of the specific factors causing MHNs to leave or contemplating leaving; so that meaningful and targeted retention strategies can be put in place to resolve this current crisis. The aim of this review is to identify what factors affect the retention of MHNs, in order to inform retention strategies and identify areas for future research.

Protocol and registration
This systematic review and thematic synthesis adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines (Moher et al. 2009).

Database searches
A systematic search of six databases was conducted from their year of inception to November 2019: CINAHL, PsychINFO, MEDLINE, Web of Science (Core collection), EMBASE and the British Nursing Index (BNI). Final search strategies were developed using the following key terms: ("Mental health nurs*" OR "psychiatric nurs*" OR RMN* OR "mental nurs*" OR RMHN) AND (Retention OR retain* OR remain* OR leav* OR attrition OR "intent* to leave" OR "intent* to stay" OR turnover OR "sustainable workforce" OR loyalty OR resign* OR shortage). No date limits were applied to ensure a sufficient number of results.

Grey literature and other resources
Although the BNI includes dissertations and theses, it mainly covers papers published within the UK. Therefore, to supplement the electronic searches and minimize publication bias, grey literature databases were searched for any relevant unpublished studies including dissertations and theses. Reference lists from relevant studies and reviews were also screened for eligible studies.

Contacting authors
Primary authors were contacted to try and gain the full-text versions of unpublished articles that could not be located elsewhere.

Inclusion criteria
The following inclusion criteria were used to identify eligible studies: Population • Studies focussing on registered nurses who specialize in mental health or work in mental health settings of any age, experience or staff grade.

Setting
• Studies focussed on registered nurses working in any mental health setting.

Outcomes
• Studies reporting any findings on factors that affect the retention of registered nurses working within mental health settings.
• Studies reporting findings on interventions or strategies that aim to affect retention.
• Studies reporting findings on the predictive factors of retention.

Study design
• Any empirical studies including all study designs.
• Studies were not excluded based on quality due to the paucity of research on this topic.
• Studies not written in the English language at full text were excluded.

Study selection
The search results generated from the databases were exported to Mendeley, and duplications were removed. Titles and abstracts were screened simultaneously to identify relevant papers. Relevant papers were then screened at full text to identify eligible studies for review based on the predetermined eligibility criteria.

Data extraction
A simple standardized data extraction form was created based on the information required to answer the review question using Excel. The extracted data were tabulated and assessed for heterogeneity.

Quality assessment
As the included studies comprised a range of study designs and methodologies, the MMAT tool version 2018 and user guide were used to assess the quality of the studies (Hong et al. 2018). The included studies were categorized and rated using the appropriate quality criteria. These ratings were then used to form a percentage score. Each of the quality criteria within each category will account for 20% of an overall score, so if a study met all the criteria they would score 100%. Scores were accompanied by a summary of the quality by comparing and contrasting the studies within and across categories. Mixed methods studies were rated using the quality criteria for mixed methods studies plus the appropriate quality criteria for each individual study component.
thematic synthesis was conducted to identify and generate descriptive themes across all included studies. Findings from quantitative studies were converted into qualitative form by a process often described as 'qualitizing' or 'data transformation' (Dixon-Woods et al. 2005;and Pluye & Hong 2014). This was done by transforming the statistics derived from survey/questionnaire responses and secondary data sets into textual data, in order to create a narrative of the results that can be synthesized alongside the qualitative findings (Heyvaert et al. 2016). The data transformation process can be seen in Appendix A in the supplementary materials. The qualitized data along with the qualitative findings were then imported to Quirkos software for analysis. Thomas and Harden's (2008) stages of thematic synthesis were followed: the process began with line-by-line coding of the findings of each study; descriptive themes were then developed through organizing the codes into themes and associating themes with similar concepts; the final themes and subthemes were accompanied by a narrative analysis and then contextualized within the existing literature in the Discussion section.

Search results and study selection
The search of the six databases retrieved 4428 potentially relevant papers. No additional papers were identified through searching grey literature or checking the reference lists of relevant papers. After duplicates were merged, the citations and abstracts were screened simultaneously for 2024 papers. This left 45 papers eligible for screening at full text. Primary authors for two of the studies were contacted to access the full-text version of their papers for screening which were unable to be located otherwise (Hanohano & Dee 2017;Walrath 2011). Walrath provided the full-text version for screening. Hanohano and Dee's (2017) paper was excluded at full text as it was unavailable. A total of 22 papers were excluded at full text with reasons. This left 23 papers eligible for inclusion; see Figure 1 for an overview of the study selection process.

Description of the included studies
The detailed characteristics for the twenty-three included studies can be seen in Table 1. Of the studies, twenty were published in journals and peer-reviewed nursing magazines, and three were unpublished PhD theses. The studies span almost three decades from 1991 to 2019 including 18 918 MHNs. The studies were conducted in eight different countries, nine in the United States, six in the United Kingdom, three in Jordan, and one in Canada, China, Japan, Ghana and Israel. Although a lot of the same factors affecting retention are identified in studies conducted in different countries, it is important to note that there are differences in the nurse training programmes and healthcare systems in different countries that generated different factors for nurses working in different countries.
The clinical settings included in the studies comprised of NHS, private, state-run and Ministry of Health settings. Over half of the studies were conducted in hospitals including wide a range of inpatient settings, and six studies included a range of community-based settings which can be seen in Table 1. Six studies did not specify the individual clinical settings (Hamaideh 2011;Murrels & Robinson 1999;Robinson et al. 2005;Robinson & Tingle 2003;Sherring & Knight 2009;Walrath 2011).
The studies included a range of study designs. Two studies were evaluating retention strategies. Rollins (2014) evaluated whether an increase in the length and content of an existing training programme undertaken in MHNs' orientation period impacted retention. The new 'healthy work environment' (HWE) programme was extended to eight weeks and included familiarization to local polices and documentation, skill competencies and seclusion and restraint training. Pelletier et al (2019) introduced a nurse residency programme for newly graduated nurses. The programme incorporated training on a vast number of clinical theories and practice, as well as topics that influence retention derived from the literature, and provided a social support system.
The aims of the studies varied with the majority of the studies (n = 18) directly aiming to explore factors that affect retention, although the factors explored differed among the studies. The rest of the studies (n = 5) had different aims with different foci that can be seen in Table 1. These studies were included as they still generated some important findings on factors that affect retention. Due to the differences in study design and aims, the outcome measures also varied significantly. All the qualitative studies used individual interviews and the quantitative studies used a range of validated and unvalidated questionnaires to explore different factors that affect retention.

Quality assessment
The included studies were categorized into 4 categories of study designs: 'qualitative' (n = 4) (Alexander et al. 2015;Karlowicz & Ternus 2009;Musto & Schreiber 2012;Stacey et al. 2011), 'quantitative non-randomized' (n = 2) (Pelletier et al. 2019;Rollins 2014), 'quantitative descriptive' (n = 15) (Alsaraireh et al. 2014;Bamber 1991;Baum & Kagan 2015;Hamaideh 2011Hamaideh , 2014Ito et al. 2001;Jiang et al. 2019;Kagwe et al. 2019;Murrels & Robinson 1999;Robinson & Tingle 2003;Robinson et al. 2005;Sherring & Knight 2009;Walrath 2011;Yanchus et al. 2015Yanchus et al. , 2017 and 'mixed methods' (n = 2) (Agyapong et al. 2015;Gunn 2015). The details of the individual assessments for each individual study can be seen in Appendix B in the supplementary materials. All of the qualitative studies met 100% of the quality criteria indicating they are high-quality studies. One of the quantitative non-randomized studies met 80% of the quality criteria but did not report on possible confounders (Rollins 2014), whereas Pelletier et al's (2019) quantitative non-randomized study only met 60% of the quality criteria. This study had a response rate of 33% due to attrition over time. Five of      the quantitative descriptive studies met 100% of the quality criteria indicating they are high-quality studies (Hamaideh 2011;Ito et al. 2001;Murrels & Robinson 1999;Robinson & Tingle 2003;Robinson et al. 2005). Six quantitative descriptive studies included a sample that was not representative of their target population due to small sample sizes (Kagwe et al. 2019;Sherring & Knight 2009;Walrath 2011) and samples from single clinical settings which were not representative of MHNs working in other clinical settings in different areas (Alsaraireh et al. 2014;Bamber 1991;Jiang et al. 2019). Seven of these studies were at risk of non-response bias judged by low response rates ranging between 35 and 65%, with no documented reasons for non-response or statistical compensation reported (Bamber 1991;Hamaideh 2014;Kagwe et al. 2019;Sherring & Knight 2009;Walrath 2011;Yanchus et al. 2015;Yanchus et al. 2017). Two of these studies did not use measurements that are validated or pre-tested (Bamber 1991;Kagwe et al. 2019). One used a measurement that was not specific to the topic under study (Yanchus et al. 2017). The two mixed methods studies were of low quality. Neither studies presented an adequate rationale for the use of a mixed methods design, or any information on how or when data integration occurred. The quantitative component met 80% of the criteria for one of the studies (Agyapong et al. 2015), in contrast to the qualitative which met 60%, whereas in Gunn's (2015) study the qualitative component met 100% of the quality criteria and only 60% of the quantitative descriptive criteria.

Findings
Job satisfaction was a widely applied code and citied in seven studies (Alsaraireh et al. 2014;Bamber 1991;Baum & Kagan 2015;Ito et al. 2001;Jiang et al. 2019;Yanchus et al. 2015;Yanchus et al. 2017). Most of these studies demonstrated that a decrease in overall job satisfaction scores was associated with an increase in intent to leave their current positions. Two of these studies demonstrated a reverse trajectory that nurses with higher overall job satisfaction scores were more likely to stay in their jobs (Bamber 1991;Jiang et al. 2019). Whilst it is evident that overall job satisfaction has a profound effect on intent to leave or stay, identifying the individual aspects of job satisfaction that affect retention will be more useful for directing future study and retention strategies. All identified factors were grouped into four main themes: individual characteristics, working within mental health services, training and skills and work environment.

THEME 1: INDIVIDUAL CHARACTERIS-TICS
This theme is made up of three subthemes: demographic characteristics, experience and employmentrelated factors, and values.

Demographic characteristics
Age is the most significant of the demographic characteristics, with seven studies identifying an association between age and intent to leave (Alsaraireh et al. 2014;Bamber 1991;Baum & Kagan 2015;Ito et al. 2001;Jiang et al. 2019;Murrells & Robinson 1999;Robinson et al. 2005). Whilst most of the studies found that as age increases intent to leave decreases, Robinson et al (2005) found that younger nurses that are aged between 20 and 29 and have children had less intentions to leave their job.
For gender, two studies found male MHNs had higher intent to leave (Alsaraireh et al. 2014;Jiang et al. 2019). Conversely, one study found no significant difference by gender on intention to leave (Baum & Kagan 2015). Marital status was identified as a factor in two studies with MHNs with single status being more likely to intend to leave their current positions (Alsaraireh et al. 2014;Jiang et al. 2019).
Ethnicity was cited in two studies, one finding no difference between ethnicity and intent to leave (Baum & Kagan 2015) and another reporting a tentative finding that White Irish and British women were more likely to continue nursing up to ten years postqualification (Robinson et al. 2005). Jiang et al (2019) found that MHNs with better self-rated health were less likely to leave their jobs, whilst MHNs with a higher level of education were more likely. One study found that MHNs with type A personalities were significantly more likely to leave (Bamber 1991).
These findings highlight a need to focus on how the retention of younger MHNs can be addressed and subsequently call for more research around retention and younger nurses specifically. Additionally, these findings could suggest that increasing efforts to recruit mature students into mental health nursing training may be beneficial for retention. More studies are required to form more robust conclusions on how the other demographic characteristics included in this review impact on retention.
Experience and employment-related factors Bamber (1991) and Alsaraireh et al (2014) found that MHNs who had left their jobs were less experienced overall than those that stayed in their jobs. This may be explained by a finding in Alexander et al's (2015) qualitative study stating: As new [mental health nurses], participants reported starting out much more ambitious and idealistic; with experience, they had adopted a tempered realism. This suggests that with experience, some MHNs are able to reconsider what success means to them and their service users and feel that they are able to make positive differences to people's lives (Alexander et al. 2015).
The amount of time spent in their current employment was also associated with MHNs intent to leave. Two studies reported MHNs who had spent longer in their current job were more likely to stay in it (Ito et al. 2001;Jiang et al. 2019). Robinson et al. (2005) found MHNs who entered nursing with a degree and spent less time in their current job had greater intent to leave, whereas those entering nursing with qualifications that were not degree sufficient and had spent less time in their current job were more likely to stay. In addition, if MHNs felt they had alternative opportunities, elsewhere they were more likely to intend on leaving their job (Ito et al. 2001).
The majority of new nurses are now likely to be graduates and at the beginning of their careers with limited experience as qualified registered MHNs. It is important to think about how new nurses can be better supported to spend longer in their first posts to increase their experience, so, in turn, they can support the next newly qualified nurses coming through. Future research may want to focus on the factors affecting the newly qualified nurses' intent to stay or leave their first jobs to gain more insight into their experiences, to identify any shortfalls of the preceptorship programmes that support nurses with the transition from student to qualified nurse to improve their experience and job satisfaction.

Values
Individual values were identified in one study exploring how values affect the practice experience of MHNs (Stacey et al. 2011). They found that when MHNs' values conflicted with their organization's values which restricted their ability to practice in line with their own values, they contemplated leaving.

THEME 2: WORKING WITHIN MENTAL HEALTH SERVICES
This theme is made up of three subthemes: perception of the job, factors related to practice and clinical setting.

Perception of the job
The majority of MHNs (84%) in Gunn's (2015) study reported that the exciting and rewarding nature of the job was their reason for remaining in mental health nursing, with one MHN stating: I have worked in many areas of nursing but this is the most rewarding job I have ever had.
An MHN in another study also explained how the non-economic rewards of the job had contributed to their longevity in mental health nursing: . . .the hook for me is seeing success, no matter how small or how large.
Career pride was a theme that emerged through the findings of Alexander et al.'s (2015) study where MHNs described needing to overcome negative stereotypes held by other professionals and recognize their value, in order to develop career pride and remain in the field.
Stigma associated with working in mental health services was identified in three studies as a factor that almost always negatively affects retention of MHNs (Agyapong et al. 2015;Alexander et al. 2015;Gunn 2015). However, an MHN in Gunn's (2015) study reported that they responded to stigma by returning to the profession to care for mental health service users: 'I returned to MH nursing because I am helping these patients get better that other nurses can't because they have a stigma about it'.
(Participant in Gunn 2015, p.60).  (Bamber 1991;Gunn 2015) were associated with leaving or intent to leave. MHNs reported experiencing moral distress after events that involve the safety of adolescent service users, which impacted their perceived ability to keep service users safe and caused some to leave or consider leaving their job (Musto & Schreiber 2012). One study identified that a higher level of emotional exhaustion increased the likelihood of MHNs contemplating leaving the NHS (Sherring & Knight 2009).

Factors related to practice
Being able to provide successful care and treatment, no matter how small the improvement, and remain hopeful for service users was identified in two qualitative studies (Alexander et al. 2015;Karlowicz & Ternus 2009). MHNs in one study describe their positive experiences of caring for service users which is what influences them to remain in mental health nursing: These findings suggest that working within mental health services has its own set of unique factors and factors that are more relevant to working within mental health. The rewarding nature of the job and helping people through their battles with mental health problems come at a cost to the nurses own mental and physical well-being, resulting in fear, emotional exhaustion and moral distress. It is likely that if nurses possess some of the individual factors from Theme 1, for example limited nursing experience and less time spent in their job, they will be more vulnerable to some of the factors present in this theme. This highlights the relationships between the factors across themes and the impact that one factor could have on other factors increasing the influence on MHNs' intention to leave or stay.

Clinical setting
Interestingly, two studies reported differences in MHNs' intention to leave between different clinical settings. One study found that working on closed wards was associated with higher intent to leave mental health nursing than working on open wards in the hospital under study (Baum & Kagan 2015). Another study found that nurses working in male acute units in the hospital under study had the highest turnover intentions and MHNs working in long-term female units had the lowest turnover intentions (Alsaraireh et al. 2014). Unfortunately, a number of studies did not specify the individual clinical settings in which the MHNs were working, including the four studies conducted within the NHS. This means we cannot be sure which factors identified in these studies influenced retention in which settings. Due to the differences in intent to leave between clinical settings in these two studies, it would seem that factors affecting retention differ not only by organization, but also by the specific clinical settings within organizations. So, using the broad list of factors identified in this review to inform retention strategies or recommendations for change on Trust level may lead to unhelpful and unnecessary efforts that reap few benefits, by missing what are the most important factors in which settings. With the vast range of care environments within the NHS mental health Trusts, it will be important for future studies to identify whether MHNs' intent to leave differs across clinical settings within the NHS and then explore these differences by identifying the specific factors influencing their intentions, so more tailored and meaningful retention strategies can be developed.

THEME 3: TRAINING AND SKILLS
This theme consists of two subthemes: nurse training and developing and maintaining skills.

Nurse training
Course preparation was a significant factor associated with plans to remain working within the NHS (Murrells & Robinson 1999). Two studies found that nurses prepared with a bachelor's degree or qualifications sufficient for degree entry were less likely to remain nursing than those with other types such as associate degrees, access courses or direct entry tests (Alsaraireh et al. 2014;Robinson et al. 2005). Beyond course preparation, one study identified that the length and quality of nurse preceptorships were factors that can cause newly qualified MHNs to consider leaving (Robinson et al. 2005).

Developing and maintaining skills
Three studies demonstrated an association between MHNs' intention to leave and developing and maintaining their skills (Alexander et al. 2015;Kagwe et al. 2019;Robinson et al. 2005). MHNs described being part of a team that encourages and enables them to enhance their skills was a reason they remained working within the profession (Alexander et al. 2015;Gunn 2015). Additionally, being able to use their skills to practice autonomously was a factor that indirectly predicted MHN's turnover intentions through job satisfaction (Yanchus et al. 2015).
CPD and in-service training are often used by MHNs to develop and strengthen their skill sets learned during nurse training. In two studies, a lack of CPD opportunities was a contributing factor to MHN's intentions, or actual decisions, to leave their first job (Robinson et al. 2005;Robinson & Tingle's 2003). MHNs in Karlowicz and Ternus' (2009) study reported that they would like their organization to promote more CPD to improve retention. In Gunn's (2015) study, CPD itself was the reason reported for why one participant left mental health nursing, as they went back to studying to further their education. Two studies identified a lack of in-service training and education negatively affects retention (Bamber 1991;Karlowicz & Ternus 2009). Whilst most of these studies found that an inability to enhance their skill set negatively affected retention, fear of losing nursing skills was also a reason cited for nurses to consider leaving their job working in a mental health hospital (Kagwe et al. 2019).
Whilst CPD and in-service training are important to all nurses for developing their skills, the majority of newly qualified nurses' skill development comes through working in teams with more experienced nurses who promote and encourage skill development. Although earlier findings in Theme 1 suggests that particular attention needs to be paid to the newly qualified nurses, these findings suggest that if we do not also focus on retaining experienced nurses, we may not be able to combat some of the factors affecting retention of newly qualified nurses.

THEME 4: WORK ENVIRONMENT
One study found that MHNs who had left their job were significantly more dissatisfied with their work environment than those still in their jobs (Bamber 1991). Work environment is the largest theme broken down into six subthemes: working relationships, leadership, organizational culture, salary, work schedule and resources.

Working relationships
An association between working relationships and intent to leave or stay was reported in seven studies (Alexander et al. 2015;Gunn 2015;Jiang et al. 2019;Kagwe et al. 2019;Karlowicz & Ternus 2009;Robinson et al. 2005;Robinson & Tingle 2003). Kagwe et al (2019) found that all aspects of working relationships were significantly associated with the MHN's intent to leave their jobs within the next year. Karlowicz and Ternus (2009) reported mixed findings related to team dynamics with currently employed MHNs expressing positive experiences of feeling part of the team, and MHNs that had left their job reporting negative experiences and feeling outcasted. A qualitative study was able to provide a deeper understanding of what aspects of positive working relationships influence MHNs' decisions to remain in the profession (Alexander et al. 2015).
Participants appreciated the sense of belonging and cohesion that the team fostered, and connected that to why they have remained in the specialty (Alexander et al. 2015, p.451).
Another study found that MHNs satisfied with the Trust and coordination between physicians and nurses were more likely to stay in their jobs (Jiang et al. 2019). In two studies, civility, defined as 'courteous and considerate workplace behaviors within the workgroup' (Yanchus et al. 2015, p.222), was shown to indirectly impact on turnover intention through job satisfaction (Yanchus et al. 2015(Yanchus et al. , 2017.

Leadership
Aspects of leadership were identified in seven included studies as factors that affect retention. Receiving support from line managers (Robinson et al. 2005), effective communication (Walrath 2011), high-quality supervision (Karlowicz & Ternus 2009) and leadership opportunities (Alexander et al. 2015) all aided retention.
A lack of supervisory support consisting of guidance, feedback and interpersonal support was associated with intent to leave or turnover in two studies (Ito et al. 2001;Yanchus et al. 2017). Similarly, inadequate support, especially in times of high acuity, was significantly associated with intent to leave (Kagwe et al. 2019). As well as line management, MHNs should have the opportunity to engage in clinical supervision for support and skill development. An unsatisfactory amount of clinical supervision was identified as a factor sufficient to cause newly qualified MHNs to contemplate leaving their job (Robinson et al. 2005).

Organizational culture
MHNs in Karlowicz and Ternus' (2009) study reported that changes within their organizations' attitude and culture should be addressed to improve retention.
Organization's procedural justice concerning performance appraisal was identified as a factor that can predict MHN's turnover intention through job satisfaction in one study (Yanchus et al. 2015). Likewise, being given an opportunity for feedback prior to any disciplinary action was significantly associated with lower intent to leave (Kagwe et al. 2019). MHNs who had left their jobs in Bamber (1991) study reported being significantly more unsatisfied with the quality of decisions made by management.
The role of MHNs within an organization was identified as a factor that can affect retention in two studies. A role that allows MHNs to utilize their diverse set of skills (including dispensing medication) and spend a satisfactory amount of time with service users can increase job satisfaction and improve retention within the profession (Gunn 2015;Karlowicz & Ternus 2009).
Retention strategies were put in place and evaluated in two organizations committed to retaining their MHNs across two studies. Both studies found positive results with a new (HWE) orientation training programme significantly increasing retention rates for one organization (Rollins 2014); a new Nurse Residency Program for new graduates was associated with less nurse turnover for another (Pelletier et al. 2019).

Salary
Salary was one of the most widely applied codes across seven studies (Agyapong et al. 2015;Alsaraireh et al. 2014;Gunn 2015;Jiang et al. 2019;Kagwe et al. 2019;Karlowicz & Ternus 2009;Robinson et al. 2005). Jiang et al. (2019) found that MHNs with a higher monthly income on average are more likely to remain in their jobs. Despite job security with a permanent contract being a factor for staying in their first job, salary that does not reflect their level of responsibility was also a reason sufficient for MHNs to leave or contemplate leaving (Robinson et al. 2005). MHNs in Karlowicz and Ternus' (2009) study suggested pay scales and including more incentives should be addressed within their organization to improve the retention.

Work schedule
Work schedule was explored in four studies. Shift times had no significant impact on MHN's intent to turnover (Alsaraireh et al. 2014). MHNs working over forty hours per week (Jiang et al. 2019) or part-time hours (Baum & Kagan 2015) were more likely to contemplate leaving. More than a third of MHNs in Robinson et al's (2005) were dissatisfied with the combination of work hours and time spent with partners, spouses or on responsibilities for their children, and these factors were sufficient for MHNs to contemplate leaving their first mental health nursing job. MHNs in Karlowicz and Ternus' (2009) study reported that they would like scheduling and work hours to be addressed within their organization in order to improve retention.

Resources
Newly qualified MHNs reported that unsatisfactory staffing levels, including amount of qualified to nonqualified staff, were a sufficient reason to consider leaving their first job (Robinson et al. 2005). Similarly, the ratio of MHNs to service users was not a reason for MHNs to remain in mental health nursing in Gunn's (2015) study.
Workload, and having adequate time to complete their work, was significantly associated with MHNs' likelihood of searching for an alternative job and intent to leave in Kagwe et al's (2019). MHNs' satisfaction with the amount of time spent on paperwork was positively related to intent to remain in the profession in Robinson et al. (2005).
It is clear from these findings that MHNs want to work in an organizational culture that values their role and demonstrates that through respect, procedural fairness and meaningful salary. These findings suggest that the organizational culture must also foster strong leadership, promote positive team dynamics and be sensitive to and attempt to address MHNs' concerns about the work environment such as work schedule and resources, in order to increase nurses' intent to stay in their jobs, their organization and the profession as a whole.

DISCUSSION
This review was concerned with the individual factors that affect the retention of MHNs, and whilst overall job satisfaction is a significant factor, this concept was broken down into individual factors categorized within four interrelated themes: individual characteristics, working within mental health services, training and skills and work environment. Each of the individual factors has the potential to impact on other factors; for example, overwhelming workloads and inadequate staffing will likely expedite burnout and increase moral distress which will intensify MHNs' intention to leave. Individual factors within themes may be enough for MHNs to contemplate leaving, but due to potential intersection between thematic factors, it is likely that no factor occurs in isolation and a combination of the factors across these four themes is what contributes to overall poor job satisfaction and causes MHNs to actually leave their jobs, the NHS and the profession. The most effective retention strategies will need to address multiple yet specific factors to improve job satisfaction. MHN workforce retention will not improve on its own or without the necessary funding to make positive changes and develop retention strategies. Therefore, policymakers and governments around the world must abandon future cuts make real investments to support the nursing profession to change the trajectory of this international crisis (Turale & Nantsupawat 2021).
Many individual and demographic characteristics are associated with MHNs' intent to leave. The most significant individual characteristic identified in this review that impacts retention is age. This finding is corroborated by the review conducted by Health Education England (2014) which found younger nurses are more likely to leave the profession. Whilst the retention of younger nurses needs to be addressed, perhaps trying to increase the uptake of mature students onto training programmes would harbour more fruitful benefits for retention. Lack of financial support is a barrier for mature students which has had a significant effect on mental health fil rates (National Health Service 2019). The findings from this review confirm the importance of NHS long-term plan's promise to explore 'earn and learn' premiums to support mature students (National Health Service 2019).
Newly qualified nurses entering their first job were found to be more at risk of leaving, which also mirrors the review by Health Education England (2014) that found turnover rates are highest in nurses' first and second year of practice. Newly qualified nurses are more likely to be younger and possess other risk factors such as less nursing experience and time spent in their current job increasing the risk of leaving for this group. Future research and retention strategies should focus on how they can better support younger and newly qualified nurses through their first two years of practice. More personalized robust preceptorships tailored to the individual nurses should be given urgent consideration.
Despite review findings suggesting older nurses are more likely to stay, one in three nurses are likely to retire over the next ten years (Royal College of Nursing 2020). Due to the global ageing workforce concern, there is a wealth of research dedicated to the retention of older nurses and preventing early retirement (Duffield et al. 2014;Markowski et al. 2020;Ryan et al. 2018;Uthaman et al. 2016). To approach this, King's College London on behalf of NHS Improvement (2019b) has set out some age-specific recommendations for the retention of older nurses and their experience which include the following: recognizing and capitalizing on their talents and ability to impart knowledge through mentoring opportunities; and promoting flexible working (desirable shift patterns and 'wind down opportunities').
The second theme was more unique to the mental health nursing speciality. Caring for some of society's most vulnerable people with complex conditions is seen as exciting and rewarding, which is a significant factor in attracting and retaining MHNs. However, continuously caring for people with mental health problems can lead to emotional exhaustion, moral distress and burnout which negatively affect MHNs' well-being and retention. Stigma, being assaulted and perceived risk of being assaulted at work are significant factors that cause MHNs to contemplate leaving their jobs and the profession altogether. Interestingly, different clinical settings were shown to affect retention suggesting factors that affect retention differ in their existence and importance across clinical settings. Whilst there is some literature pertaining to differences in vacancy rates across different Trusts and regions, there is an absence of literature detailing retention rates for individual clinical settings within Trusts (Department of Health Social Care 2020). There is scope for future studies to explore the differences in retention rates among clinical settings within the NHS and explore what factors affecting retention exist in which clinical settings to better tailor retention strategies to provide more meaningful outcomes.
The complexities of mental health care and the vast amount of specialist services within mental health require significant training, and experience, to be able to practice competently and autonomously. Nurse training including course preparation, adequate preceptorships, sufficient in-service training and CPD opportunities is a significant factor that affects retention. Similarly, CPD was cited as a top reason for NHS staff leaving in the NHS long-term plan (National Health Service 2019). Effective training will help MHNs to hone and develop their skills and improve their confidence which will positively affect retention. It is necessary to explore in-service training, preceptorships and CPD opportunities across clinical settings to address any shortfalls. Subsequently, ascertaining what inservice training or CPD opportunities would be beneficial to MHNs in which clinical settings will better inform retention strategies.
Positive change could also be catalysed through providing preceptorship packages that are specific to the needs of the individual preceptee rather than a standardized approach. Focussing on the development of more effective preceptorships that encourage and promote CPD opportunities could be an important tool for mitigating the risk of MHNs turning over in their first and second year of practice. Organizations need to ensure they have enough adequately experienced preceptors to meet the standard for preceptorship (Department of Health 2010), and guarantee preceptors spend enough time with their preceptee to build positive relationships and enhance person-centred skill development opportunities, which will contribute to an increase in confidence (Matua et al. 2014). Thereafter, regular in-service training and CPD opportunities specific to their clinical setting to enable skill development would improve MHNs' confidence and ability to work autonomously, which would enhance job satisfaction and improve retention.
Work environment constituted the largest of themes identifying multiple factors on all levels within organizations. Poor organizational cultures, limited MHN roles within the care and treatment of service users, poor leadership, low salaries in relation to responsibility, demanding work schedules, inadequate staffing, large workloads and lack of clinical supervision were factors identified that negatively impact retention. Nurses who believe they have better job opportunities elsewhere are more likely to consider leaving (Ito et al. 2001). This finding is in keeping with Royal College of Nursing's employment survey (Royal College of Nursing 2019) in which 55% of nurses stated they feel confident they could find a similar job with better work environments and salaries elsewhere. Work environments will vary between clinical settings, and it will be important to identify which factors are relevant in different clinical settings to understand what needs addressing to improve MHNs' work environments to improve job satisfaction and retention.
The material related to work environment alludes to a broader discussion about the nature of nursing in contemporary times. We have indicated here that poor leadership, sub-optimal supervisory structures and routines and inadequate staffing (often masked by the use of agency and bank nurses), is indicative of a system that is fragmented and open to continual change. Randall and McKeown (2014) talks of nursing being in a state of 'liquid modernity' (Bauman 2000), where care is compromised as a result of fear, uncertainty, difficulties in establishing relational continuity and the once established leadership structures now being dissipated. This review certainly indicates a workforce that is alienated and marginalized. Senek et al (2020) highlight data from the adult nursing workforce where it is also clear that poor relationships with managers, an inability to solve everyday problems, transient working relationships and perceived poor care, have resulted in alienation and intention to leave the profession.
Two retention strategies, the HWE orientation training and Nurse Residency Programme included in this review, were associated with improved retention and turnover rates of MHNs. Halter et al's (2017) review of systematic reviews on interventions to improve adult nurse turnover also illustrated the importance of some retention strategies. The National Retention Programme is creating tailored strategies on an NHS Trust level using information from exit interviews (NHS Improvement 2019a). However, MHNs turning over within the same Trust are often not considered leaving and their valuable information will not be sought or contributed to the scheme. Future studies focussed on MHNs that are still in post will elicit information that could be used to the implement immediate changes or strategies that can prevent more MHNs from leaving.

LIMITATIONS
The majority of the studies used cross-sectional designs of single hospitals at one point in time, some with small sample sizes unrepresentative of their target populations with poor response rates, limiting the generalizability of the results to other clinical settings and populations. Although a strength of this review is its inclusion of a sufficient number of studies conducted all over the world providing a global picture, different countries have different nurse training programmes and healthcare systems. This limits the generalizability of some of the findings in this review to MHNs working in the UK and more specifically the NHS.
Another limitation is that the factors identified in this review are from studies that span almost three decades. Whilst some factors from the earlier studies will still be relevant to the current workforce healthcare is an ever-changing environment in all countries which will inevitably lead to changes in the factors affecting retention of healthcare workers. Mental health nursing will have changed significantly over the past three decades in most countries. Some of the more recent changes for MHNs in the UK reported by Gournay (2005) include the following: a shift in focus within community services (community teams are caring for and treating people with more severe and enduring mental health problems than before), this has led to a shift in the patient population in inpatient services; more people in inpatient services are now more likely to be detained under the Mental Health Act and have dual diagnoses giving rise to violence and aggression seen in these services; education and training have changed with more recent nurses likely to be graduates and more focus placed on psychosocial interventions to help equip nurses adapt to their new roles. Therefore, some of the factors in this review may not be as important for today's workforce but some may be more significant than before, for example factors relating to inservice training, CPD opportunities and risk of assault. This also highlights the need for more studies to explore the factors that affect the retention of the current workforce to develop relevant and meaningful strategies for today's workforce.

CONCLUSION
Most themes in this review (individual characteristics, training and skills and work environment) identify similar factors that affect retention of the adult nurse workforce (Chan et al. 2013;Moseley et al. 2008). However, it is clear from this review that MHNs encounter some factors unique to working in mental health services, which suggests that retention strategies should be specific to each nursing speciality. Beyond nursing speciality, it has become apparent that the factors identified within this review are likely to vary between clinical settings in mental health due to the differences in work environments and services they provide. So, it is important that future studies now set out to explore what factors exist in which clinical settings to create more tailored retention strategies, which can then be applied on a clinical setting level across all mental health NHS Trusts, to generate better outcomes.
The majority of the studies in this review used quantitative measures that only capture the factors included in the outcome measures used, and thus, the factors identified in this review are not an exhaustive list. There is scope for studies using qualitative methods to further explore the factors affecting MHN retention to identify any important factors that may have been missed. Missing from the current literature is the voice of senior leaders who inform policy around retention. It is important to gain the perspectives of senior leaders to identify whether there are any factors affecting retention of MHNs at an organizational level that are missed by MHNs; and to explore any previous, current or planned retention strategies in order to provide tangible recommendations for change and inform effective retention strategies.

RELEVANCE FOR CLINICAL PRACTICE
This is the first systematic review to collate all the existing evidence relating to factors that affect mental health nurse retention. The findings from this review identify and discuss some of the individual factors that affect retention and identify areas for positive change and future study. These findings can be used to inform more tailored and meaningful retention strategies for mental health NHS Trusts, with an aim to improve job satisfaction of MHNs to combat this current nursing crisis. Improving MHN retention is paramount to the safety and quality of care of our mental health service users.