Factors affecting high secure forensic mental health nursing workforce sustainability: Perspectives from frontline nurses and stakeholders

focus on the performance of custodial activities that emerged from their review of nursing records. HSH nurses perform extensive custodial activities, for example personal and room searches, restricting access to certain areas of the ward, monitoring communications (Jacob, 2012; Mason et al., 2009; Timmons, 2010). Custodial tasks serve the purpose of Abstract Introduction: There has been no previous study of stakeholders' views on recruitment and retention concerns in high secure forensic settings. Aim: To identify factors affecting recruitment and retention in high secure hospitals, from the perspectives of stakeholders with experience in forensic mental health nursing. Method: Framework analysis of data from fifteen interviews and three focus groups with frontline nurses, nurse leaders, recruitment leads and union representatives from three high secure hospitals in England. Results: Six themes emerged from the data: (a) the unique nature of high secure nursing; (b) the impact of short staffing; (c) wider factors affecting the high secure nursing workforce; (d) the location; (e) staff being on different terms and conditions of work; (f) recruitment strategies. Discussion: Multiple factors are likely to simultaneously affect high secure hospital recruitment and retention. Findings on the unique nature of high secure work reflect previous qualitative research. The themes of location, working terms and condition and recruitment strategies have not been previously identified in forensic nursing research. Implications for practice: Employers should ensure that employment terms and conditions are equitable and consistent. Furthermore, hospital managers should address the effect that movement of staff between wards may have on morale and therapeutic relationships.


Accessible summary
What is known on the subject?
• There are insufficient nurses to meet current demand for mental health care. This is an international concern. Within England, the impact of staff shortages on the quality of patient care in forensic high secure settings has been highlighted by the national regulatory body for hospitals.
• Forensic hospital nursing is a distinct specialism within mental health. Forensic nurses must negotiate the therapeutic, ethical and practical challenges of caring for high-risk patients in a locked environment.

What this paper adds to existing knowledge?
• There has been no previous study to ask frontline high secure forensic nurses, union representatives, senior nurses and workforce leads about what factors may be affecting recruitment and retention in their setting.
• As well as the specialized and challenging nature of the work, participants identified that workforce sustainability was affected by unequal working terms and conditions, the hospital locations and wider national factors, such as changes to how nurse training was funded.
• They also identified that some strategies that were employed to address workforce shortages, such as day-to-day movement of staff within the hospital and incentive packages for new recruits could be demotivating for established staff.

What are the implications for practice?
• Forensic high secure nursing workforce strategies should include training, development and career pathways that are specific to the specialism and extend beyond preceptorship for newly qualified staff.
• There should be clear and equitable employment terms and conditions with remuneration packages that are consistent within and between organizations.

| INTRODUC TI ON
Secure or "forensic" inpatient mental health services provide care and treatment to patients who are detained in hospital under mental health law due to the risk that they present to the public, predominately following a criminal offence. Internationally, provision of forensic mental health care is heterogenous, with Sampson et al. (2016) finding that all 18 countries in the European Union offer some form of long-stay health care for mentally disordered offenders, but with varying treatment philosophies and forms of service provision.
In England, secure hospital provision is categorized as either low, medium or high security. The three high secure hospitals (HSH) in England provide regional and national services for forensic patients requiring a high secure environment, with a combined capacity of around 800 beds. The English national healthcare regulatory body, the Care Quality Commission, expressed concern regarding the impact of shortages of nursing staff on patient care and patient safety in the HSH in 2018 (Care Quality Commission, 2018). This concern should be contextualized within the national and international shortage of nurses, particularly in mental health (Durcan et al., 2017;Harding, 2019), and the increasing demand for mental health services, both nationally and internationally (Rehm & Shield, 2019).

| BACKG ROU N D
Nursing in secure environments is a specialism within mental health nursing. Forensic mental health nurses must be as proficient as any mental health nurse in skills such as risk management and therapeutic engagement; however, the secure environment places specific demands upon these aspects of the nurses' knowledge and skills (Mason, 2002;Mason et al., 2009). There is a distinct ethical dimension to forensic hospital mental health care. Any mental health nurse working with patients detained in hospital against their will must recognize and negotiate the inherent tension between care and custody and the power imbalances within their relationship with patients (Jacob, 2012). Forensic nurses are constantly navigating between the caring and the custodial, as found by Martin and Street's (2003) study of a large Australian forensic hospital. Their analysis contrasted a tendency for nurses to focus on therapeutic relationships during research interviews with the focus on the performance of custodial activities that emerged from their review of nursing records. HSH nurses perform extensive custodial activities, for example personal and room searches, restricting access to certain areas of the ward, monitoring communications (Jacob, 2012;Mason et al., 2009;Timmons, 2010). Custodial tasks serve the purpose of Abstract Introduction: There has been no previous study of stakeholders' views on recruitment and retention concerns in high secure forensic settings.
Aim: To identify factors affecting recruitment and retention in high secure hospitals, from the perspectives of stakeholders with experience in forensic mental health nursing.
Method: Framework analysis of data from fifteen interviews and three focus groups with frontline nurses, nurse leaders, recruitment leads and union representatives from three high secure hospitals in England. Implications for practice: Employers should ensure that employment terms and conditions are equitable and consistent. Furthermore, hospital managers should address the effect that movement of staff between wards may have on morale and therapeutic relationships.

K E Y W O R D S
forensic, nurses, staff perceptions, therapeutic relationships, workforce issues managing risk in the secure ward but the use of therapeutic engagement skills also has a risk management function, as a means of establishing "relational security" (Royal College of Psychiatrists, 2015;Walker et al., 2017). A specific challenge for forensic nurses' use of therapeutic relationship skills is that they are attempting to form therapeutic relationships with individuals who require treatment under conditions of high security on account of their dangerous, violent or criminal propensities (NHS Commissioning Board, 2013, p. 1). Such patients may have caused significant harm to others in the past and may pose a risk of harm to them as nurses in the present. Furthermore, those patients may have experienced significant trauma themselves (Aiyegbusi & Kelly, 2015).
Nursing in the high secure setting has been described as emotional hard labour (Beryl et al., 2018, p. 82), due to the risk of violence, the complex, traumatized and dangerous patient group and the custodial setting. Jacob et al. (2009) characterize forensic nurses as having to cope with feelings of disgust and fear in their encounters with patients who may have committed heinous crimes. Several studies have raised concern about elevated markers of stress and burnout in forensic nurses (Berry & Robertson, 2019;Dickinson & Wright, 2008;Happell et al., 2003;Lee et al., 2015). Nevertheless, some studies have found forensic nurses to have lower prevalence of stress and burnout than fellow mental health nurses in other settings (Chalder & Nolan, 2000;Elliott & Daley, 2013). This may be due to what Beryl et al. (2018) found in their interviews with staff who worked with female HSH patients that some nurses relish the challenge of negotiating the complexities of the setting, particularly valuing the opportunity to work long term and affect positive change incrementally with such a unique patient group. As with analyses of factors affecting workforce sustainability across mental health settings (Durcan et al., 2017), organizational provision of education, support and supervision that reflects the specific and unique features of the HSH setting has been identified as of vital importance (Aiyegbusi & Kelly, 2015;Beryl et al., 2018;Cramer et al., 2020;Mercer & Perkins, 2014). Whilst implications for workforce strategies may be drawn from previous research on the experiences of HSH nurses, there has been no previous study to focus directly on nurse recruitment and retention in the HSH setting from the nurses' point of view. The present study was commissioned by NHS Improvement in 2019 to inform a review of HSH workforce challenges. The study aim was to identify the factors affecting recruitment and retention in HSHs, based on interviews with stakeholders with experience in forensic mental health nursing.

| ME THODS
The study followed a Framework Analysis methodology (Ritchie & Lewis, 2003;Ritchie & Spenser, 1994) to undertake an analysis of data gathered during telephone interviews as well as fieldwork at the three hospital sites. Framework Analysis was an appropriate qualitative methodology, because it was developed specifically to inform government policy (Ritchie & Lewis, 2003), which was the purpose of this study.

| Data collection
Interviews were conducted by two experienced nurse researchers. They had no prior relationship with the participants. The data analysis and administration of the project were undertaken by the interviewers as well as two other researchers. The team were supported by a reference group, comprising nine nurses with expertise in forensic mental health, recruited via the nursing directors of the three sites and the Royal College of Nursing. The reference group helped to shape the study design, the interview and focus group topic guides, to sense-checked the interpretation derived from the data analysis and helped to define the recommendations from the study.
During a six-month data collection period in 2019, there were 15 individual interviews and three focus groups with 16 (n6, n7, n3) participants. Given the potentially contentious topic, it seemed important to gather information in private (interviews) as well as in a discursive context (focus groups). Participants in the interviews were purposively sampled. They were representatives from the nursing leadership team, recruitment leads and staff-side union representatives at each site. They had been identified as key stakeholders with insight into the state of the forensic nursing workforce. Focus group participants were frontline clinical nursing staff recruited via an invitation sent out by each hospital's lead nurse. Two researchers visited two sites. One researcher visited one site. Interviews took place either over the phone or in person on site. Interviews and focus groups took place in private. There were no dropouts; however, focus groups only took place at two sites due to difficulties releasing frontline staff on the fieldwork visit day for the third site. Instead, two individual interviews with frontline staff were undertaken.
Focus groups and individual interviews followed topic guides (Appendix 1) but were semi structured, allowing for discussion to flow according to participants' responses. Interviews and focus groups were between 30 and 70 min in length. Interview participants were offered the opportunity to review copies of their transcripts and offer comments. Field notes were taken to support the interviews. Audio recordings of the interviews were transcribed using a transcription service. The lead researcher then reviewed the transcripts and audio files, anonymizing the transcripts prior to review by the wider team.

| Data analysis
Data analysis was led by one researcher with the wider research team undertaking a collaborative analysis of a sample of three scripts. The five stages of Framework Analysis (Ritchie & Spenser, 1994) were followed. First in the familiarization stage the lead researcher immersed herself in the data, listening to the audio files, reviewing transcripts and field notes and noting recurring themes. Themes were discussed by the research team.
Second, transcripts were reviewed to develop a thematic framework, based on themes and recurrent concerns, a priori (drawn from the research questions) (for example, "the location") and interpretive themes (for example, the role of gender). Third, at the indexing stage, transcripts were reread and annotated with reference to the emerging framework, noting that some passages had multiple index points. The framework was tested out with other members of the research team (as advised by Ritchie & Spenser, 1994). In the fourth stage, a comprehensive task of charting the data took place, using spreadsheets, whereby data from each interviews were arranged according to thematic headings and subheadings. The charts enabled a comparison between cases, with each theme being represented by indicative quotes and summaries. In the fifth mapping and interpretation stage, the objectives of the qualitative analysis were addressed, whereby interpretations were weighed up for salience and a typology of factors affecting recruitment and retention in HSH was developed. The initial complete typology was presented at a meeting with the study reference group, as a means of sense-checking the interpretation and to develop recommendations from the research.

| Ethics
The study design was approved by the lead authors' university ethics committee and the Health Research Authority for the National Health Service of England (IRAS number 265582). AlI participants were given an information sheet about the study in advance. They had an opportunity to discuss the study with the researchers. They signed consent forms. Given the limited pool of possible participants for this study, some participants were concerned about their comments being attributable to them given their specific roles. We assured participants that any reporting of the research would maintain their anonymity.

| Rigour
This study is reported in accordance with COREQ reporting criteria (Tong et al., 2007). The study was conducted with reference to Mays and Pope's (1995) seminal guidance on ensuring rigour in qualitative research. Rigour was ensured through the use of a systematic and reflexive approach to the study design, data collection and analysis, namely the staged Framework Analysis approach. Sampling was purposive, aiming to engage participants who were best placed to assist with addressing the research question. The data collection and analysis were led by experienced researchers, and the process was documented throughout. Interpretations of the data were developed collaboratively by the research team. The plausibility of interpretations was tested through discussion with the project reference group. The final typology of factors affecting recruitment and retention incorporated all cases, whereby there were no deviant cases, rather certain characteristics that featured to a greater or lesser extent in all transcripts. There was some difference in concerns between the three sites, namely the effect that geographical location had on recruitment and retention where one site was more remote than the other two and one was in a much higher cost area to live than the others. Remoteness and cost of living were raised as concerns at all three sites, however.

| FINDING S
Six themes emerged, incorporating sub themes. They are presented in Table 1. They are described below.

| The unique nature of high secure nursing
The first theme encompassed participants comments on what was unique about nurses who worked in the high secure setting and what their motivations were to work there. "This is not a job for everybody" was a common thread, suggesting that high secure nursing is a vocation, and suited nurses who were "resilient, mature and motivated," particularly given the ethically challenging nature of the work. They must be capable of putting in the emotional labour to develop therapeutic relationships in the long term, with limited immediate reward, as typified by these two quotes: You've got to have really, really strong resilience to be able to work in a high secure hospital, to be faced with a high level of abuse from patients, from assaults, and then want to come back the next day.
(Director of nursing, site 3) there isn't that, like I said, that positive feedback that you get from other areas of nursing, where patients genuinely, with sincerity, say to you, 'thank you for looking after me', or you get that feeling that they're actually appreciating it. We work with some very difficult characters, and a lot of them don't want to be in hospital. They don't want to be detained. They don't have the same view of the world as us, in any way, shape or form. They've often come from really dif- (Staff side union representative, site 1) Nurses described being dependent on their fellow nurses not only to remain safe but to remain positive about their work and motivated to return to the setting after incidents happened. As well as having an ongoing sensitivity to risk and the complex dynamics of managing several "risky" patients in one setting, nurses were mindful that many patients were suffering from the effects of trauma, and some patients may be nursed in seclusion and segregation (meaning they were isolated from other patients as a means of managing risk) for long periods of time. The challenge for nurses, then, was to "carry hope" for patients and endeavour to promote quality of life even so: People are in seclusion for so long. Years and years and years people are in seclusion for, but if we can get them up and take them into the garden and make that day a little bit nicer for them, you feel like you can make a little bit of a difference to their life here.
(Focus group member, site 1) There were important caveats made in participants' accounts of high secure nursing. First, they said that there were misconceptions about what they did and how they worked because all that the public heard about HSHs was about a small number of high-profile patients.
They drew distinctions between the different wards and services on their sites, with HSH rehabilitation wards, for example, having a different ethos and approach to risk management than the HSH acute admission wards. Nurses also said that HSHs now were different from in the recent past, with a less "macho" culture. There was some dissent regarding whether the patient population had changed in recent years, with some participants saying patients were more "risky" now due to the increase in medium secure provision, and a consequent heightening of the collective risk profile of the patients that remained in the HSH. Others said that patients were less' "disturbed" than in the past because treatments for mental illness had improved. Some participants said that ward-based nursing had become more difficult over time because therapeutic and occupational activities on offer in the hospitals had decreased and their younger patients were less motivated to engage in craft-related occupational activities than in the past.

| The impact of short staffing
The second theme was the impact that "short staffing" had on nurses' working lives, and consequently on their morale and sense of safety.
All participants were aware of staffing pressures within their hospital and articulated similar perspectives on the impact this had on staff and patients. On a day-to-day basis, staff were moved between TA B L E 1 Factors affecting HSH nursing workforce recruitment and retention

| Wider factors affecting the high secure nursing workforce
The third theme was the impact that wider national nursing workforce pressures had on staffing in HSH services, for example, the recent changes in the funding of preregistration training in England (whereby prospective nurses now had to pay university tuition fees) and public perception of nursing as an unattractive career were common concerns. Participants saw these factors particularly affecting recruitment of certain demographics into the profession, as voiced here: All of the registered nurses coming through are female. So, we're relying very heavily on unregistered being male.
(Service director, site 2) Stakeholder interviewees with responsibility for recruitment described how they were in competition with other services within their own organization for new recruits, with those competing services offering prospective staff less exposure to risk and easier access to place of work, given that the HSH were allocated in the countryside rather than in urban areas. Participants felt that their senior leaders were aware of both the staffing difficulties at their sites and the efforts being made locally to address staff shortages. Staffing in the HSH was seen as being a prominent concern for the executive boards of their employing organizations. However, they considered that forensic mental health nursing as a subspecialism was not high enough on the national nursing agenda.
Changes in the gender and age profile of nurses within HSHs were a key concern. Regarding gender, participants in various roles at all three sites told us that new recruits tended to be young women.
This was reflective of the gender and age profile of people joining the nursing profession. High secure settings require a certain balance of men to women staff, to reflect that most patents are men. The lack of male nursing recruits meant that more men were being employed as healthcare support workers. Male staff were expected to lead on certain tasks, such as searching patients and accompanying them on leave. The changing age and experience profile of ward nursing staff was keenly felt and seen to impact on perceptions of ward safety and team dynamics. We were told that the past few years had been difficult due to a loss of experienced staff who chose to retire at age fifty five. Efforts had been made to encourage such staff to return, for example offering them flexible working patterns. They were also encouraged to return to mentor new starters, thus passing on their expertise. A further commonly held view was that the more recently qualified nursing staff had different expectations of work than previous generations. They were described as "transient," looking for short term (1-2 years') experience in the HSH rather than "a job for life." This was seen as symptomatic of societal changes rather than unique to high secure or forensic nursing.

| The location
The fourth theme was the hospital location. Participants described how the three HSHs were geographically remote. None were within comfortable walking distance of a train station, for example. The cost of housing at one site and availability of attractive local housing for all three sites were seen as discouraging younger, more newly qualified staff with lower incomes who may want to live in more urban areas. A typical comment was: …there's no accommodation out here. You won't get a one bedroom flat for less than six to eight hundred pounds a month. Even room sharing you're looking at five hundred quid a month. And that's a big chunk out of a staff nurse's salary.
(Director of nursing, site 2) Participants told us that in the past there had been "schools of nursing" on site as well as staff housing. This was a means of "feeding in" recruits. Now the sites were not close to their "feeder" higher education institutions, meaning that recently qualified staff may prefer to get jobs in the cities in which they trained. We were told, however, that some staff travelled significant distances to get to work on the site because they were vocationally committed to HSH nursing. We were told that there was a loss of staff to more conveniently located hospitals with better housing prospects (such as local medium secure units) once younger nurses aspired to buy a home or start a family.

| Differential working terms and conditions
Under the fifth theme, participants described salary, rotas and shift patterns as factors affecting workforce retention, in particular the effect of different staff doing the same job role but being on different "terms and conditions." Until the recent past, nurses and other care staff in the HSHs had received a "special hospital lead" additional payment. Some staff of long standing still received it. Newer staff were not eligible for the additional payment, although some received a "retention" payment. This meant that staff on the same job banding doing the same role could have salary differences of several thousand pounds a year and some staff on lower bandings could be earning more than their senior colleagues.
We were told both that the variations in salary were a sources of tension on the wards and that they were something staff were "resigned to." Furthermore, as part of recruitment drives, some new starters had been offered "golden handshakes," relocation subsidies or other financial incentives to stay for a specific amount of time. We were told that this also could be a source of discontent Morale and motivation to stay were also affected by perceived inequities in shift patterns between staff groups. For example, some staff (such as recent retirees who were encouraged to return part time) were offered "flexible working patterns," meaning that other staff had to cover less popular shifts. The hospitals did not use nursing agency or unknown "bank" temporary staff employed by their trust. Rota filling was therefore reliant on staff doing overtime and a pool of "bank" staff allocated to the site. We were told that this was due to the high risk of violence and the need for all staff to have a good working knowledge of the site security policy and procedures. The limitations on available temporary staff were to an extent why staff may be moved day-today between wards to ensure safe enough minimal staffing levels (as described above).

| Recruitment strategies
The final theme was "recruitment," whereby participants from each site critiqued their organizations' multi-faceted nursing workforce recruitment strategies, including rolling recruitment adverts, assessment centres, incentive packages (as described above) and roadshows. Reflecting on the effectiveness of their strategies, human resources and recruitment leads described how important it was to get the "right" nurses, not just any nurse. The personal qualities of "maturity," "confidence" and "resilience" were required. Senior staff at each site considered that they offered "impressive" development packages for newly qualified staff, with lengthy preceptorship training and support programmes. They were aware, though, that newly qualified nurses had many opportunities in the current employment market and that other services within their own organizations were offering similar packages. The newly qualified nurses with whom we spoke agreed that their initial preceptorship packages had been excellent but were unclear about a career trajectory beyond preceptorship. There was consensus that ongoing development and career pathways specific to HSH were important. Short staffing affected HSH being able to fulfil some aspirations about training and development, because day-to-day staffing shortages had to be addressed as the priority.
An important aspect of the recruitment of newly qualified staff was having positive relationships with the local higher education institutions, including having a regular flow of student nurses on clinical placement. One director of nursing said "students are our lifeblood." The recently introduced roles of nurse apprentice and associate nurse were described as being valuable additions to the nursing workforce, but "no substitute for an RMN (Registered Mental Nurse)." One participant with senior responsibility for practice development told us that the associate nurse training programmes were too focused on general adult nursing to prepare trainees to take on some aspects of nursing work in forensic hospitals. In the recent past, secondment into nurse training of healthcare support workers from the HSH had been an effective route to expanding the nursing workforce. A number of focus group members had begun their nursing career this way, graduating from support worker to student to nurse. One site had invested by seconding ten staff to nurse training but the cost of doing this at scale was prohibitive. Family and local connections had previously been common routes into high secure nursing. Whilst some participants still had family members on site, the resource of potential nursing staff in the local population close to one site was described as "exhausted." When asked what might aid recruitment, participants called for a return of nurses being paid to train or some means of mature adults being able to afford to train without getting into debt. They also called for a return to having schools of nursing on the site so they could "grow their own." According to them, not only should nursing be taught on site, but there should be further specialized training in high secure forensic nursing, and coherent career pathways with milestones and educational components.

| D ISCUSS I ON
By proposing a typology of factors affecting HSH nursing workforce recruitment and retention from the perspective of key stakeholders and frontline staff, the study offers valuable insights that should inform national and regional workforce strategy in England and elsewhere, although without comparators, the extent to which the study reflects broader trends within mental health nursing more generally cannot be confirmed. The first identified theme of "the unique nature of high secure nursing" reflected several factors that have emerged in previous studies from the United Kingdom, Canada and Australia on HSH nurses and forensic mental health nurses in general: that HSH work is particularly ethically and emotionally challenging (Mason, 2002;Peternelj-Taylor, 1999), This is particularly due to what Jacob (2012) describes as "the contradictory mandate" of care and custody and possible feelings of fear and abjection towards some patients (Jacob et al., 2009). However, as found in previous studies, forensic nurses who have chosen the high secure setting may relish its ethical and therapeutic complexity (Beryl et al., 2018;Dutta et al., 2016). There can be no doubt that nursing in HSH settings is a rewarding career for nurses who have the resilience and motivation to work long term in a restricted environment with potentially high-risk individuals. Our finding that HSH nurses are committed to their particular ward, team and patients, and do not like to be moved for "risk management" reasons accords with previous interview studies in which forensic nurses were asked to describe their relationships with patients, for example Marshall and Adams' (2018) Canadian study which foregrounded forensic nurses' desire to create a "homely" ward environment given their patients' lengths of stay, and Niebieszczanski et al.'s (2016) study, exploring how the concepts of hope and recovery are articulated by forensic nurses, which found that "use of self" was their primary therapeutic intervention. In Jacob's (2012) analysis of Canadian forensic nurses' navigation of care versus custody, Jacob cites ' Festinger's (1957) theory of cognitive dissonance, whereby a person experiences discomfort when there is a disjunct between their beliefs and their behaviours.
Jacobs' nurses dealt with this cognitive dissonance by reframing their custodial interventions as therapeutic ones (for example seeing the practice of "seclusion" of agitated patients as a therapy not a punishment). Jacobs' analysis is relevant to our study because participants described how the delicate therapeutic balance of staff and patients was upset through, for example, movement of nurses within a hospital to address staffing shortages that might increase risk of incidents. The staff moves consequently affected morale, sense of therapeutic accomplishment and ultimately staff retention.
The message for policy makers here is that where nurse staffing focuses primarily on risk management it foregrounds the custodial role and impedes nurses' sense of being carers, first and foremost. This observation is supported by the findings of Walker et al.'s (2017) recent study of nursing staff in women's medium secure services.
For their participants, relational security was of utmost importance, given the women's trauma histories and how slowly trusting relationships between staff and patients may develop. For their services to be effective, participants called for consistency in therapeutic approaches, meaning use of permanent staff with demonstrable commitment to the client group over use of temporary staff.
Whilst there have been previous studies to consider the role that gender plays on the therapeutic milieu and motivation of staff in high secure forensic settings (Addo, 2006;Aiyegbusi & Kelly, 2015;Mercer, 2013;Mercer & Perkins, 2014, this was the first study to report that nurses in HSH are concerned by the effect on staff retention of the changing gender and age profile of their colleagues, who are getting younger and more female. It was related to nurses' desire to be surrounded by fellow team members whose judgement they trusted, an important feature in limiting burnout and staff distress in the forensic environment (Johnson et al, 2016). In Mercer's work (a discourse analytic study on forensic staff and sexually disordered offenders in an English high secure setting), perceptions of female staff as vulnerable due to patients' histories of sexual violence and "gender role" allocation of tasks was described, with the wards being defined as primarily "male spaces." Our work builds on Mercer's observations, but also suggests that fellow nurses' age (likely a proxy for experience) is also a factor.
Interview participants included recruitment leads for the three hospitals. They related demographic changes to factors affecting nursing recruitment across all settings, namely changes in funding for nurse training that made it less appealing to men or mature students. Furthermore, participants voiced concern that early career forensic nurses' expectations differed from those held by nurses in the past. Early career nurses were seen as more transient, no longer seeking a "job for life." These comments reflect commonly held views about Generation Y and Z's attitudes to work, that they are highly motivated to engage in altruistic pursuits but value work-life balance and are less work-oriented than other generations (Hampton & Welsh, 2019). The evidence on generational differences in approaches to work is mixed (Twenge, 2010), and some studies have found that neophyte nurses are still motivated by traditional nursing values (Hampton & Welsh, 2019;Tuckett, 2015).
Previous studies have also made distinctions between Generation Y and Z nurses' desire for flexibility, yet also for an employer that is committed to them, and offers them good supervision and devel- Our study shows that the national "offer" for forensic mental health nurses should be adapted to address the specialized nature of HSH nursing work and the current discrepancies in terms and conditions between staff.

| Implications for practice
Multiple factors are likely to simultaneously affect HSH recruitment and retention. One approach to addressing recruitment and retention would be for training and career pathways within forensic nursing to be specific to the specialism and extend beyond preceptorship for newly qualified staff, There should be specific training and career pathways for HSH nurses. Also, there should be clear and equitable employment terms and conditions with remuneration packages that are consistent within and between organizations. Furthermore, hospital managers should account for the effect that movement of staff between wards may have on morale, therapeutic relationships and safety culture. Moving staff impacts on their motivation and sense of safety because therapeutic relationships and team dynamics are so important for the management of risk and for the incremental rehabilitation of service users in the high secure setting. This may be a key factor in nursing staff retention.

| Study limitations
The study is limited because it offers a typology of factors based on opinions, some of which require further substantiation (for example notions about generational differences in work motivation); however, this was a qualitative inquiry where researchers were given privileged access to frontline staff and key stakeholders whose work is usually behind closed doors. There was significant consensus between sites and interviewees. Our confidence in the analysis and typology was strengthened through sharing and sense-checking with an expert reference group. The study team's access to frontline staff was limited though and was less than planned. Although we were able to visit all three sites and offer frontline staff the opportunity to take part in focus group interviews, daily pressures on staffing impeded engagement in focus groups on the days we visited.
Data collection took place during Autumn 2019, meaning that findings represent the state of nursing workforce pressures at a point in time. This study did not incorporate views from patients, nor routinely collected data on staffing (for example staff surveys or exit interviews). We were reliant on participants to give their analysis of recruitment and retention, based on their experiences.

| CON CLUS ION
The study was devised to inform nursing workforce policy through analysis of the views of stakeholders and frontline staff.

RELE VAN CE S TATEMENT
There are nursing workforce shortages within mental health settings, including within forensic high secure hospitals.
Stakeholders and frontline nurses from forensic high secure settings in England offer valuable insights into the factors they consider to be affecting nurse recruitment and retention. The findings have relevance to nursing workforce leads and forensic nurse researchers because they should inform workforce sustainability strategies and suggest directions for future research, such as on generational differences in career aspirations between forensic nurses.

ACK N OWLED G EM ENT
Study commissioned by NHS Improvement.

CO N FLI C T S O F I NTE R E S T
During the period of the study, one author (EW) was employed by the funder, NHS Improvement, one author (AMR) was president of the Royal College of Nursing. No other conflicts to declare.

AUTH O R CO NTR I B UTI O N S
JO,AT,IE,EW,AMR made substantial contributions to conception and design,acquisition of data, analysis and interpretation of data; involved in drafting the manuscript or revising it critically for important intellectual content; gave final approval of the version to be published; agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

DATA AVA I L A B I L I T Y S TAT E M E N T
Research data are not shared.