Nurses' experiences of racism in mental health settings through patient and family interactions: A systematic review

Nursing staff engage readily with patients and associates in mental health/forensic inpatient settings. These settings are known to have instances of workplace violence directed towards staff and such violence includes racism. Racism is a form of workplace violence that must be better understood and supported within this complex setting. Completing a systematic review to coalesce preexisting research and suggested interventions can be beneficial to supporting nurses. Systematic review following PRISMA guidelines. CINAHL, PsycInfo, Medline, British Nursing Database and Web of Science databases were searched. Reviewers screened the papers for inclusion (29 articles out of 7146 were selected for inclusion) and completed the quality appraisal using the Mixed Methods Appraisal Tool. Subsequently, data extraction was completed, and findings were summarised through narrative synthesis. The way racism was conceptualised impacted how data was collected, reported and interpreted; racism was silenced or exposed depending on how studies were undertaken. If exposed, evidence indicates racism is a problem but is not always acknowledged or acted upon. Some evidence determined racism led to negative work-related outcomes. The literature provided limited examples of interventions. These included changing education/orientation for staff, openly discussing racist events and better planning for patients among colleagues and management. Increasing diversity within the workforce requires more research exploring and addressing issues related to racism towards nurses. Narratives of racism being normalised and embedded in mental health/forensic settings need to be challenged.

defined as 'threatening conduct based on race, colour, language, national origin, religion, association with a minority, birth or other status' (International Labour Office, 2002: 4).The focus of existing research on violence in mental health settings has however tended to focus on specific types of workplace violencepredominantly physical violence-and not on other aspects such as verbal aggression, including racist attitudes and behaviours (Stewart & Bowers, 2013).Moreover, the extent to which racism features in all aspects of workplace violence has not been a prominent research focus.
Racism creates inequitable opportunities and outcomes for people based on race and can come in the form of harassment, humiliation, violence or intimidating behaviour manifested through institutions and individuals (Australian Human Rights Commission, n.d.).Racism from patients and their relatives can seriously affect nurses' mental and physical well-being and can be experienced through interpersonal and structural means (Centre for Disease Control and Prevention, 2023;Ontario Human Rights Commission, n.d.).Individuals in inpatient mental health units, which for this review include forensic/secure settings, are often required to stay on these units against their wishes, under some form of legal framework (Bettridge & Barbaree, 2012).The tensions and conflict associated with restricting individuals' freedoms and choices can heighten nurses' experiences of racism, particularly when considered alongside the complexities involved in caring, daily, for individuals with acute mental illnesses and disorders (Fletcher et al., 2021;Odes et al., 2020).However, when racism is a part of professional discussions within nursing, the focus tends to be more on the cultural considerations of the patient, or discrimination through a systematic lens rather than on the experiences of nurses in racially complex engagements (Iheduru-Anderson et al., 2021).
This systematic review focuses on nurses' experiences of racism while working in mental health settings, including forensics.Aspects such as minority groups, geographical locations, varying terminology and cultural dynamics will be explored in terms of how these nuances can alter how racism is encountered by nurses.Lastly, nurses are leaving the field, shortages are apparent and burnout is being reported by nursing staff affecting the care patients receive and at a costly expense to the public (Renwick et al., 2019).All forms of racism are not holistically being studied.

A I M A N D OB J ECT I V E S
The overall aim of this systematic literature review was to determine what is currently known about the experiences of racism among mental health nurses working in inpatient mental health settings.We sought to discover how racism is defined and codified in the literature, establish the methodological approaches that have been used to study racism, understand the factors that can influence how racism is experienced by nurses, identify interventions implemented to address racism, and finally consider the impact of nurses' experiences both of racism and of existing support mechanisms on workforce issues.Furthermore, the review will make recommendations for future research, policy, interventions and applications in practice with a focus on supporting the retention of mental health nurses in inpatient settings.

M ET HOD S Design
A systematic review with a narrative synthesis of findings was undertaken.This approach provides a rigorous process to identify and critically appraise relevant literature on the experiences of racism among mental health nurses while identifying any interventions developed to address this issue (Popay et al., 2006).The decision for this approach, without a meta-synthesis or metaanalysis, was informed by the scoping searches that returned sources that were produced within different paradigms and methodologies.The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA-2015) guideline was used in the preparation of this review's protocol (Shamseer et al., 2015) and the PRISMA checklist for reporting was followed as relevant (Page et al., 2021).The protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO registration number: CRD42022308902) (Henshall et al., 2022).
A grey literature review was also conducted, an important step in completing a comprehensive review (Data S1 for grey literature search strategy).The search was conducted using combinations of search terms shared with our search strategy for the systematic review.The databases Google Advanced, EBSCO and Proquest were searched.However, results from the grey literature search did not meet our inclusion criteria for the systematic review.Instead, where relevant, papers from the grey literature search informed our discussion.

Search strategy
The databases searched were CINAHL, PsycINFO, Medline, British Nursing Database and Web of Science with no restrictions on language, date of publication and country of publication.The Medical Subject Headings were used with search terms containing variations of 'racism' and 'mental health OR forensic nurses' (Data S2 for full search strategy).The database search was supplemented through the screening of the reference lists of all included papers.

Study selection
Search results were compiled, screened and extracted using Covidence (Veritas Health Innovation, Australia), a web-based collaboration software for systematic and other literature reviews.After the initial search, all duplicates were removed.All titles and abstracts were screened by five researchers [AW, SK, AK, RS and OK].Any conflicts were reviewed by a group meeting with a minimum of three reviewers present, and unresolved conflicts were then discussed at the full team meetings.The same strategy was applied to the full-text review process.
Primary research of any design published in peerreviewed journals was included.Articles were included if they reported on experiences of nursing staff within a mental health inpatient setting with regards to racism directed from patients/families/visitors.Articles reporting on nursing staff (including nurses, student nurses and nurse auxiliary staff) based on inpatient mental health wards were included.If articles did not specify the clinician type or setting the authors were contacted for clarification.Any full texts of non-English articles that were included following title and abstract screening were screened by native-speaking healthcare researchers external to the team if the team members did not speak the language.The full text was then translated into English if deemed eligible for inclusion through a professional translation service and was then re-screened by the internal research team for possible inclusion in the review.
Review articles, commentaries, editorials or secondary data analysis were excluded.Articles which did not specify clinician type or setting were excluded.Studies which specified multiple clinician types but did not break down findings by clinician type, or if nursing staff did not comprise a substantial proportion (50% or more) were excluded.Studies that addressed workplace harassment but did not identify racism as a sub-type of harassment were not included unless the study identified a relationship between staff ethnicity and workplace harassment.

Data extraction
Two researchers independently completed data extraction, with a third researcher joining the discussions when there was a lack of consensus.As part of data extraction, using a structured bespoke template, researchers identified the year of publication, study period, language of publication, study setting, patient and/or staff characteristics, definition and incidence of racism, methodology and a summary of findings (Data S3 for the data extraction table).

Quality appraisal
Quality assessments of all included studies were completed independently by two researchers using the Mixed Methods Appraisal Tool (MMAT; Hong et al., 2018), with the third researcher providing consensus.The methodology was identified for each included text and then the relevant MMAT questions were answered.The MMAT required identifying the study design and then assessing the methodological quality criteria.The domains reviewed were dependent on the study design (MMAT; Hong et al., 2018).No studies were excluded as a result of the quality assessment (Tables 1-3).

Narrative synthesis
The initial searches and scoping of the literature informed our decision to proceed with the narrative synthesis of data from qualitative, quantitative and mixed-methods studies (Popay et al., 2006).The narrative synthesis, a process of 'telling a story' of the findings from the individual studies, was appropriate to synthesise the evidence coming from heterogeneous studies on experiences and responses to racism.It also transpired during the review that there was a need to explore and report on the ways researchers conceptualised and studied racism.This is more aligned with the meta-narrative review approach (Wong et al., 2013).

Review authors' reflexivity
The role of reflexivity is to explore the perspectives of each researcher and consider the influence their perspectives might have on their research.Whilst reflexive practice is valued in primary research and guidance is available, little has been written about reflexivity in systematic reviews (Rees et al., 2017).Reflexivity in a literature review process is concerned with the ongoing, mutually shaping interaction between the review team and the review (Glenton et al., 2020).
Using self-reflexivity statements and team meetings, we were able to capture the perspectives that were brought to the review process.We applied both 'prospective' (the influence that the reviewers have on the review) and 'retrospective' (the influence that the review has on reviewers-how the review process and findings influence their prior positions) reflexivity (Glenton et al., 2020).The review team members' prior positions and presumptions were discussed.Reflecting on our ethnic backgrounds (our team had diverse ethnic backgrounds), our professional roles (registered nurses, including mental health nurses, and healthcare researchers) and the respective socio-geographical contexts we worked and lived in (teams in Canada and England) helped us to understand what our individual experiences and understandings were regarding racism in nursing.We also learnt to appreciate our distinctive backgrounds and experiences and to recognise the value our differing perspectives brought to the review process.In the self-reflexivity statements, team members reflected on their experiences of being exposed to racism, their understanding of the scale of the problem and their expectations of how racism is depicted in the literature.All team members expected the topic to be more wellresearched-reflecting what we have observed in clinical academic practice; we expected to find rich data on nurses' experiences of racism and numerous interventions to tackle racism in mental health inpatient settings.We sought validation of our personal experiences, we wanted to explore how others conceptualise and respond to racism, and we looked for answers on how to tackle racism.
We continued to reflect and discuss our individual experiences of working on the review throughout the process, the main review team met weekly to discuss progress and conflicts of opinion.Chart 1 depicts a variety of reflexive practices we engaged in at different stages of the review.

R E SU LT S
The initial search of the databases generated 7146 records (Figure 1).Out of 7146, 2452 duplicated records were removed.Of the remaining 4694 records, 4224 irrelevant records were excluded after screening the title and abstract.The remaining 470 records were then screened by examining full texts against the eligibility criteria resulting in the exclusion of 441 studies.Reasons for exclusion of these 441 studies were (i) unrelated topic (348); (ii) non-primary research (42), (iii) not including inpatient mental health units (33), (iv) unable to translate (6), (v) duplicate ( 5), (vi) not research articles (4), (vii) not focused on nurses (2), (viii) text not found (1).This resulted in 29 papers remaining for inclusion in the review.The studies included in the review were diverse in terms of geographical locations, target populations and health systems.The studies represent an appropriate coverage of racism experienced by nurses from diverse ethnic backgrounds.

Characteristics of included studies
The 29 articles included in this review were (Ferns & Meerabeau, 2008) published between 1991 and 2019.These papers covered data from 11 countries; the majority reported on data from the United Kingdom (10 papers) and the United States (8 papers).Most papers focused on nurses and auxiliary nurses (solely or among other professions) while two studies reported on nursing students (Ferns & Meerabeau, 2008;Zwane & Poggenpoel, 2000).In terms of reporting participants' demographic characteristics describing their identity, 18 studies did not provide any details while 11 acknowledged participants' ethnicity (9 studies), religion (Isaiah  (Schablon et al., 2018).When reporting ethnicity, the level of detail varied, for example, one study made a general observation about a predominantly White or Black institution the participants were coming from (Zwane & Poggenpoel, 2000).Only two studies reported patients' ethnicity (Gillig et al., 1998;Pulsford et al., 2013).Twenty-one studies were quantitative, six were qualitative and two applied a mixed-methods approach.Table 4 presents details of the included studies and Table 5 provides the characteristics of participants they reported on.Table 6 presents each study's aims and findings.

The prominence and expression of racism
Racism was not the focus in the majority of the included studies, with only a few identifying racism as a key topic (Deacon, 2011;Kavanagh, 1991;Sullivan & Yuan, 1995).Primarily when racism was covered it was included under umbrella terms such as violence, workplace violence and aggression (verbal, psychological, physical) directed against staff working in mental health care settings.One of the studies mentioned racial abuse as an example in the context of 'staff experiences of harm' but then provided no additional analysis of harm related to racialised experiences (Cooper & Inett, 2018).In addition, various terms were used to directly refer to racism, for instance, racial harassment, racial abuse, racial comments, personal racism and discrimination.In general, racism and related terms lacked definitions.
There was also variability in research articles concerning data collection and analysis sections in association with how racism was established by the authors.For instance, in one research study, the author (Deacon, 2011) categorised interpersonal racism as racist comments made by patients that were directed towards nurses, directed towards other patients, stated to nurses that were about other ethnic minority patients and non-person-centred comments stated to nurses.Others defined racial harassment as a type of psychological violence in terms of any threatening behaviour related to race, skin, colour, language, nationality, religion, minority race, birthplace or identity (Chen et al., 2008(Chen et al., , 2009)).The articles in addition described examples of racism in diverse ways, e.g.lack of respect for racialized/immigrant workers, asking for health care professionals belonging to a certain race (Caucasian) (Brophy et al., 2018), discrimination of minorities and underserved populations and racism as part of verbal aggression or general aggression (Jackson & Ashley, 2005).In one study, the authors (Jacobowitz et al., 2015) collected data on race as a demographic and considered it in relation to the traumatic event, resilience to stress, attitude/confidence in managing violent patient situations and compassion fatigue to Post Traumatic Stress Syndrome.Out of 7146, 2452 duplicated records were removed by the Covidence (Veritas Health Innovation, Australia) automatically.Of the remaining 4694 records, they were screened by 5 researchers (AW, SK, AK, RS, OK).In this study, 4224 irrelevant records were excluded after screening the title and abstract.The remaining 470 records were then screened by examining full texts against the eligibility criteria resulting in the exclusion of 441 studies.Reasons for exclusion of these 441 studies were (i) unrelated topic (348); (ii) non-primary research ( 42), (iii) not including inpatient mental health units ( 33), (iv) unable to translate ( 6), (v) duplicate ( 5), (vi) not research articles (4), (vii) not focused on nurses (2), (viii) text not found (1).This resulted in 29 papers remaining for inclusion in the review.
The expression of racism in these research articles varied somewhat.The most common form of racism mentioned in the majority of studies was verbal.Further, there were references to both overt racism as well as covert/subtle racism.The more explicit expressions of racism included verbal comments, threats, discrimination and patients from certain ethnicities demonstrating a superiority complex in their verbal and non-verbal behaviours (Chen et al., 2008(Chen et al., , 2009;;Cooper & Inett, 2018;Flannery et al., 2001;Ferns & Meerabeau, 2008;Stewart & Bowers, 2013;Zwane & Poggenpoel, 2000) and physical violence (Lawoko et al., 2004;Sullivan & Yuan, 1995).On the other hand, the expressions of subtle/covert racism ranged from microaggressions and general disrespect for ethnic minorities (Tonso et al., 2016) to data indicating higher rates of burnout, and violent incidents in staff belonging to certain ethnic groups (Isaiah et al., 2019;Jacobowitz et al., 2015;Lawoko et al., 2004).

The impact of the methodological approach on reporting experiences of racism
The methodological approach chosen to explore instances of racism determined if and to what extent racism was reported.The quantitative studies (22 studies) included data collected in prospective and retrospective incident reports, self-reported questionnaires of experiences and compensation databases.These studies further varied in the way racism was determined, collecting data on racial harassment among the ethnic group(s) or on abuse/violence/trauma and breaking it down by ethnicity.
The included retrospective studies routinely collected incident data on assaults; data was broken down into racial harassment (Dickens et al., 2013;Flannery & Walker, 2003;Stewart & Bowers, 2013) or by nationality (Hamadeh et al., 2003).This type of study may be subject to bias through under-reporting and inconsistency (Dickens et al., 2013).To address some of the challenges of retrospective studies, prospective incident reports were used instead by other studies (Flannery et al., 2001;Flannery et al., 2011; the second part of Ridenour et al., 2015;Whittington & Wykes, 1994).Another approach to exploring the extent of racism was through an analysis of a workers' compensation database with ethnicity coded to study non-fatal workplace assaults and comparing potential assaults on non-White versus White healthcare workers (Sullivan & Yuan, 1995).
In studies reporting the percentage of racial harassment compared to other types of violence, participants had an option of indicating one or multiple answers (types of violence) when reflecting on an incident.When there was no option for respondents to indicate multiple types of violence in a single incident (e.g.co-occurring physical and racial harassment), racial harassment was infrequent ranging from 1% (Dickens et al., 2013); to 4% (Chen et al., 2009;Stewart & Bowers, 2013) to 4.5% (Chen et al., 2008).Some authors acknowledged that they did not estimate the extent to which participants may have experienced multiple forms of violence within a single occurrence (Jackson & Ashley, 2005;Jacobowitz et al., 2015).When participants were permitted to choose multiple forms of violence for a single incident, racism was more likely to be reported from 13% (Tonso et al., 2016) to 40.7% (Samadzadeh & Aghamohammadi, 2018).
Minimal qualitative research was found (Brophy et al., 2018;Cooper & Inett, 2018;Deacon, 2011;Richter, 2014).Many of these studies focused on experiences of aggression and violence with the exploratory nature of these studies providing opportunities to mention experiences of culturally and racially based conflicts and discrimination.However, these instances were not fully acknowledged due to methodological issues, for example conducting focus groups on a sensitive topic (Richter, 2014), the small number of participants representing various racial and ethnic backgrounds (Brophy et al., 2018), poor choice of quotes and limited discussion of racism (Cooper & Inett, 2018).Only one datadriven ethnographic research study investigated racism as its main topic, including racist insults by patients to nurses.This surfaced as an uncomfortable nursing issue when researching nursing practice (Deacon, 2011).Finally, with regard to mixed methods studies, one investigated racism alongside sexism, exploring rich and nuanced experiences of racism through participant observation, in-depth interviews and attitudinal surveys (Kavanagh, 1991).Two other mixed-method studies included open-ended questions that provided the opportunity for narratives on racism to emerge among participants (Ferns & Meerabeau, 2008;Stewart & Bowers, 2013).

Experiences and response to racism
Of the 29 studies included in the review, only one focused solely on racism (Deacon, 2011) with another study doing so alongside sexism (Kavanagh, 1991).Other studies researched racism to a lesser or greater degree within the context of discussing (verbal) violence/aggression/ assault/harm and applying a variety of conceptual and methodological approaches.
The literature covered experiences of racism across the world in a variety of culturally diverse groups including migrant nurses and nurses from ethnic minority backgrounds.It was inconclusive in reporting the extent and impact of racial harassment on nursing staff.There were examples of nurses being exposed to racist verbal abuse (e.g.'unpleasant and offensive language' [Deacon, 2011 p. 454]) or patients discriminating between members of staff depending on their ethnicity.These examples included: keeping distance (e.g.due to suspected contagious illnesses), not cooperating during conversations or refusing treatment (Stewart & Bowers, 2013;Wilson & Kneisl, 1992).Exposure to, and involvement in, racially motivated incidents were routinely experienced (Deacon, 2011;Ferns & Meerabeau, 2008;Kavanagh, 1991), starting as early as nursing school (Ferns & Meerabeau, 2008;Wilson & Kneisl, 1992;Zwane & Poggenpoel, 2000).
Prevalence studies provided a mixed picture.In studies where incidents of aggression were recorded over a year, racial incidents made up 0.9% (Dickens et al., 2013), 4% (Stewart & Bowers, 2013), 4.5% (combined for patients and co-workers as perpetrators) (Chen et al., 2008), 13% (Tonso et al., 2016) and 25.6% (Schablon et al., 2018) of all incidents.Data on the prevalence of aggression directed at nurses, broken down to ethnicity/race, provided further mixed evidence on racially motivated incidents.The data is contradictory with some studies stating that the lowest percentage of workers to report physical aggression were from ethnic minorities (Ridenour et al., 2015), whilst others reported this was the highest percentage (Hamadeh et al., 2003;Isaiah et al., 2019).In one study participants felt that racism was a major cause of the experience of physical aggression in the psychiatric unit (45% of patients and 21% of staff) (Gillig et al., 1998), while in another it was concluded (predominantly White sample) neither staff nor patients highlighted ethnicity or culture as being significant factors in aggression management (Pulsford et al., 2013).
In response to racial attacks, nurses reported experiencing a range of emotions including feeling upset, fear, anger, disappointment, helplessness and shock (Cooper & Inett, 2018;Richter, 2014), with these feelings translating into poorer relationships with patients (Cooper & Inett, 2018).There was an acknowledgement that constant verbal abuse may have a profound psychological effect on staff (Bowers et al., 2009;Tonso et al., 2016).However, again, the evidence was mixed with some concluding that when confronted with physical violence, better mental health was associated with nurses from an ethnic minority background (Renwick et al., 2019) and lower burnout (associated with verbal abuse) was present in certain ethnic minorities (Bowers et al., 2009).
On a micro level, nurses were reported to deal with this difficult issue by using strategies normally used for the management of anti-social behaviour on the ward.These strategies included conceptualising patients' racism as a consequence of their mental ill-health (also in Kavanagh, 1991;Stewart & Bowers, 2013;Richter, 2014) and using nursing methods of direct engagement, trouble avoidance and the minimisation of strangeness (Deacon, 2011).Avoidance of confronting issues related to racism was enacted through the heavy reliance on humour; staff members responded and consoled themselves and their co-workers by making light of the tense circumstances in which they worked (Kavanagh, 1991).Other nurses described growing 'thick skins' (Richter, 2014, p. 497).As coping strategies, nurses utilised personal support from family and friends, having time away from the ward, informal support from colleagues and formal support from senior management colleagues (Cooper & Inett, 2018).Some participants reported a general sense of acceptance of harmful incidents and the distress they had caused (Cooper & Inett, 2018;Kavanagh, 1991).The normalisation of racism led to nurses experiencing a lack of support and feeling let down by management.There was a reported tendency of staff to avoid critical examination of their own and co-workers' ethnicities as those characteristics influenced their life experiences, occupational roles and statuses and hierarchical relationships (Kavanagh, 1991).Ethnic issues were minimised, or discussions altogether avoided, if brought up-ignored or overlooked.There was a reported illusion of sameness, minimising and discouraging acknowledgement of real and significant experiential differences (Kavanagh, 1991).

Interventions to prevent and address racism
Data from the included studies reported on a range of interventions to prevent and address racism, and therefore reduce racially motivated incidents, identifying certain conditions to optimise their effectiveness.One study reported the need to acknowledge that workplace violence is preventable and not part of the job (Chen et al., 2009), whilst others identified the need for workplaces to transparently confront racism.Some studies indicated that until racism is openly discussed changes cannot take place (Deacon, 2011;Ferns & Meerabeau, 2008;Schablon et al., 2018).Recommendations included talking about racism in education and research with references to the realities of clinical practice (Deacon, 2011;Ferns & Meerabeau, 2008).
One paper stated that the issue of racism had to be approached systematically (Schablon et al., 2018), whilst another commented that the incidents of racism be overtly confronted and managed by healthcare teams (Deacon, 2011).A third reported that better evaluation of the environment that creates the risk of violence for staff members is needed to guide the formulation of meaningful interventions (Jackson & Ashley, 2005).Identified risk management strategies included a system of nonviolent self-defence for employees, alternative approaches to restraint and seclusion, an awareness of the early warning signs of impending loss of control, early interventions with pro re nata (PRN, take as needed) medications, tighter links to emergency backup services and a post-incident crisis intervention approach (Flannery et al., 2001).
Part of the proposed solution was linked to training.Some studies observed that instead of 'race equality training', there was a need for 'work-based practice development' focused on developing the nursing methods of managing personal racism by carefully guided exposure to these methods (Deacon, 2011).Examples of training in the literature are pre-incident preparedness training, used for high-risk groups including non-violent crisis intervention (Chen et al., 2008;Chen et al., 2009), training to de-escalate aggressive situations (Richter, 2014); training on the reasons for inpatient aggression and how to react to threatening situations (Isaiah et al., 2019).One study researching migrant nurses suggested an orientation programme acquainting the newly employed nurses with the culture, values and religious beliefs of the host country as part of their pre-employment training (Hamadeh et al., 2003).
Post-incident crisis intervention programs such as the Assaulted Staff Action Program (ASAP) were shown to provide needed support to staff victims of patient assault and to result in sharp decrements in assault facility-wide (Flannery et al., 2011;Flannery & Walker, 2003).Other post-incident interventions suggested were staff inclusion in decision-making processes, staff reflective practice, patient-focused meetings, supervision, immediate practical, emotional and social support, follow-up support, forming a watching waiting approach, restorative practice and debriefing (Cooper & Inett, 2018).One suggestion was the implementation of 'an outside consultant' based in hospitals such as the Black Task Force, to provide staff members an avenue to raise their concerns and for action to be taken on their behalf (Kavanagh, 1991, p. 264).
Some risk management strategies were focused on patients and patient management.One suggestion was to mark violent incidents in patients' charts for consideration in future treatment plans (Chen et al., 2009).These incidents would need to be discussed at team meetings and patient care rounds (Chen et al., 2009).Furthermore, patients' conditions and behaviours are to instead be managed by stress management, impulse control and/ or behavioural contracts with escalation to restriction, seclusion and immediate medication (Chen et al., 2009).Some studies went further, suggesting policies that led to greater patient self-control (Bowers et al., 2009) and teaching patients better social skills, including respecting the boundaries of others, providing better training in daily living skills by community placement and practising verbal conflict-resolution skills (Flannery et al., 2001).
Lastly, studies highlighted the importance of recording aggression, including racist incidents, using standardised tools with demonstrable validity and reliability (Dickens et al., 2013).One study stated it was crucial to establish a formal reliable reporting system (Chen  et al., 2009) and to develop a supportive work environment to encourage reporting of violent incidents in psychiatric hospitals (Chen et al., 2009).This was especially important given findings that indicated up to 60% of staff who experienced racial harassment took no action (Chen et al., 2009).

DI SC US SION
This systematic review provides valuable insights into how racism has been codified in previous research.The literature review findings provide little insight into the experiences of racism targeted at nurses and are inconclusive in describing its extent, impact and strategies to prevent and respond to racism.The evidence that is presented is weak, lacks comprehensive data exploring racism and is not systematically researched.Findings related to experiences of racism and interventions are inconsistent, with some finding a difference in experienced aggression depending on one's ethnicity/culture/ race/nationality/religion (e.g.Hamadeh et al., 2003;Isaiah et al., 2019;Renwick et al., 2019;Samadzadeh & Aghamohammadi, 2018;Sullivan & Yuan, 1995) and others concluding no difference (e.g.Chen et al., 2008;Ridenour et al., 2015;Whittington & Wykes, 1994).Overall, the insights from the available evidence, especially qualitative research, are concerning and provide a clear indication that more research is needed.

Methodological choices and their impact on reporting racism
The way studies are designed and conducted can influence the extent to which data relating to racism features in the analysis.Such choices can result in either the exposure or the silencing, of racism.The methodological approaches used in the studies to date do not give prominence to racism, leading to the underreporting of its prevalence.The quality of the reporting of racism is dependent on the choice of data collection tool(s) used and the approaches to analysis undertaken.Furthermore, depending on the stance of the researcher undertaking the study, findings related to racism may be minimised or not acted upon.
The review of the methodological approaches used in the included studies to explore racism emphasises the importance of transparency in data collection.Tools that were not expounded on (questionnaires, surveys and scripted interviews) made it difficult to conclude whether a tool was enabling or inhibiting reports of racism.For example, specifically reporting racial harassment versus reporting it using umbrella categories like verbal abuse, led to challenges in grasping the scale of the problem.Another methodological issue was the underrepresentation of participants and findings with people who identified to be a part of minority groups (e.g.Brophy et al., 2018).A paper declaring to be examining characteristics of staff as determinants that lead to an assault recognised the diversity/ethnicity of the staff (England, UK) as a potential determinant.However, its conclusion focussed only on White staff and whether they were assaulted by White or non-White aggressors (Whittington & Wykes, 1994).Other examples are two papers by Chen et al. (2008 and2009) that used the same data set collected in Taiwan.The data set reported on the number of racial harassment incidents during 1 year, declaring a 4.5% prevalence rate, however, it did not include the ethnicities of participants or patients in the sample leading to incomplete findings regarding racism.Furthermore, there is an overall lack of qualitative studies which provide opportunities for nurses to share experiences of racism in their own words.We argue that this is an important area for further research if the nature and impact of racism against mental health nurses is to be better understood.
There was a lack of homogeneity in the tools used across the studies.Some studies produced novel tools to complete their studies, but a lack of transparency prevented evaluations of such tools and made a comparison between studies much harder.In addition, other data, such as basic demographic data on staff ethnicities, were not collected hindering the formulation of a more fulsome conclusion.Many studies, examined during the screening process, were excluded because nuances such as racism were not accounted for in the tools used; this resulted in missed opportunities to gain a baseline understanding of the racism nurses may have experienced.In addition, standardised initiatives for organisations such as the collection of incident reviews or compensation claims did not collect information on racism and/or ethnicity such as with the Assault Staff Action Program, further making the issue harder to appreciate (Flannery et al., 2001;Sullivan & Yuan, 1995).
A better understanding is needed of how different ways of collecting data can impact study findings, such as how group interviews versus individual interviews on violence in mental health nursing may impact willingness to share, or how questions are worded (Ayres et al., 2023;Isobel, 2021).For qualitative studies focused on experiences of racism, a trauma-informed qualitative research approach is recommended for researchers (Alessi & Kahn, 2022).According to Alessi and Kahn (2022), a trauma-informed structure would require researchers to have better insight into the surrounding information of the traumatic event (i.e.historical and political context) and to ensure a safe and trusting space for participants whether participating in an interview or completing a survey.In addition, avoiding (re)traumatization of participants through thoughtful interviews/wording and ensuring regular self-reflection and self-care of researchers (Alessi & Kahn, 2022).Similar concepts could also be applied to quantitatively designed studies.In conducting this systematic review, the research team ensured a reflexive stance throughout to help ground the team and prepare them for the interpretation of findings, an aspect lacking in the studies incorporated in the review.
In the included studies, one provided questionnaires about workplace violence to nurses with a ballpoint pen being awarded for timely completion, another provided cross-sectional surveys about violence to be completed by nurses on the unit during their shift, and another study collected data using a 14-page long questionnaire containing 135 questions focused around violence towards psychiatric staff (Chen et al., 2008;Isaiah et al., 2019;Lawoko et al., 2004).These examples are indicators of researchers collecting imprecise data through insincere compensation for participation, in environments that participants may consider unsafe, and with daunting questionnaires that can all hinder candid involvement.A topic like racism can be exceptionally challenging and traumatic to engage with, and without the application of trauma-informed research principles the collected results could be skewed, participants retraumatized and chances of harm increased for unprepared researchers (Alessi & Kahn, 2022).As previously indicated, nurses have a tendency towards minimising reporting of racist incidents and researchers must, in turn, minimise the chances of further contributing to this (Kavanagh, 1991).

Problems researching racism
Challenges in researching racism are partly due to the various ways racism is expressed and understood.The same instance of racism may be perceived differently by different people.This review included studies undertaken across a diverse range of historical, political and social contexts which when combined with researchers' subjectivity can lead to a lack of acknowledgement of the existence of racism.An example of this is the Bahrain study (Hamadeh et al., 2003), which contained data indicating a five times higher chance of non-Bahrain staff being assaulted; this was interpreted as the fault of the clinician for not understanding local culture and lifestyles.A different researcher with a different background may have a completely different interpretation and methodological approach.Disagreement on how to structure and design studies is not uncommon between researchers, but the additional layer of how researchers are related to racial dynamics can greatly alter how information is understood.Furthermore, racism in healthcare is not as readily studied from the viewpoint of clinicians, perhaps because racism is deemed too complex, leading to an avoidance that is not specific to mental health nursing.However, the additional challenges within this field such as the diagnoses of the patients, and containment on units against patients' wishes can pre-empt strenuous relationships with caregivers and healthcare staff, making the topic more likely to be a taboo one.Our analysis of how racism was researched and the lack of racismoriented studies, through the synthesis of evidence, leads to a conclusion that research of racism directed towards mental health nursing poses a daunting task for researchers and organisational leadership strategies.

Accountability to patients
Nurses, like many healthcare professionals, commonly have standards and codes of conduct that they are accountable to in terms of how patients are cared for and to ensure patient-centred care.Codes of conduct such as the one from the College of Nurses of Ontario focus on ensuring patient dignity, individual autonomy and the provision of competent care, and that trust is maintained throughout (College of Nurses of Ontario (CNO), 2023).Accountability to patients is displayed through being able to fulfil the therapeutic needs/decisions of patients, but in mental health nursing there is the additional challenge of navigating whether patients can provide sound decision-making and whether they can be held responsible for their decisions (Manuel & Crowe, 2014).Racism can be a manifestation of a person's mental health diagnosis, such as in the case of a psychotic disorder through delusional thoughts (Poussaint, 2002).Further, a patient that is racist towards a nurse regardless of their diagnosis, makes the application of nursing standards challenging.An example is shown in the research completed by Deacon (2011), where providing care to a patient who utilises racial insults requires exceptional professionalism to continue that care provision, and additionally places the nurse in a vulnerable situation in a mental health setting.An organisation and culture that does not support nurses during these times can lead to negative consequences for both the nurse and the patient.With regards to the patient, there is literature on nurses labelling patients as difficult; this can lead to avoidance and reduction in care by nurses (Khalil, 2009).Being able to support nurses and have a framework/guidance to inform care practices can lead to better outcomes for both staff and patients, even when patients display racist behaviours towards staff.The mental health of a patient does not excuse organisational inaction.

Implications for the workforce (human resources)
The well-being of staff and retention strategies have been challenging in nursing.Neglecting to address racism towards nurses by patients risks contributing to these challenges and reflects poorly on organisations.Furthermore, there is some evidence to indicate that nurses who identified as an ethnic group differing from the majority were more likely to leave their mental health nursing job (Robinson et al., 2006).Organisations that can implement a culture of leadership that addresses violence such as racism in a timely fashion may improve employees' morale by demonstrating their commitment to current and future staff (Totman et al., 2011).

Recommendations
Based on our findings, we strongly recommend additional studies to provide insights into the racism nursing staff experience as there are currently few such studies.For future research, we suggest using traumainformed research methods, the creation of more qualitative studies that explore the experiences of nursing staff in-depth, and more transparency/consistency in the data collection tools being used by researchers.Furthermore, ensuring standardised data collection can better help organisations and researchers understand the prevalence of racism.This can be demonstrated through data collection that includes aspects such as baseline demographic information (ethnicity, of staff and patients) in post-incident crisis intervention programs, or general incident reporting, and ensuring options in place for staff to report racially traumatic events readily for more representative data.In association, by ensuring there is an open and active dialogue regarding racism from patients with receptive colleagues and management, staff can be better supported.Doing so will further encourage staff to utilise tools such as incident reviews and be more willing to be involved and transparent for research/organisational initiatives.Nurses should not need to rely on resilience, avoidance or humour to cope (Deacon, 2011;Kavanagh, 1991;Richter, 2014).Improved training and education for staff is another option to help prepare them, but organisations must communicate that this is not the only intervention available as it puts the onus (or responsibility) on clinicians to prevent and manage racist events.Finally, further evaluation of interventions is recommended through future research that can help to validate the application of these interventions in clinical practice, as well as fine-tuning their contextspecific implementation.

Strengths and limitations
The paper has detailed a comprehensive review of international evidence on racism in nurse-patient relationships, focusing on a rarely researched issue of violence by patients towards nurses in mental health settings.It is inclusive of studies from across the world, published in different languages, and applying a variety of approaches to research the instances of racism.The review goes beyond the critical appraisal of the collected evidence understood as its methodological soundness; while individual studies were mostly scored high for their rigour in applying a particular methodology and methods, the overall evidence was judged as of questionable quality.This insight was achieved due to the study team applying the principles of reflexivity and critical theory to ask fundamental questions about the development of knowledge about racism.Several studies were not included in the review due to missing data regarding eligibility.This may have led to a number of studies being excluded which may have provided insightful data on this topic.

CONC LUSION
This review furthers understanding of how and to what extent research has addressed racism in mental health nursing when it is a nurse who is a victim of a patient's actions.The review found that racism was explicitly acknowledged by a small proportion of the literature; the majority facilitated/perpetuated the silence instead of an open discussion of the problem of racism.With the increasing mobilisation and diversity of the nursing workforce, there is a need for responsible research based on the best evidence to inform nursing education and practice internationally.

R E L EVA NC E FOR C L I N ICA L PR AC T IC E
Racism experienced by nursing staff in a mental health setting is an important issue that can impact the nursepatient relationship and the well-being and job satisfaction of nurses.There are some recommendations on interventions that hospital leadership and nurses can implement to manage racism directed towards nurses.These include increasing awareness, both from colleagues and management, such as through corporate protocols/policies to support staff, and by ensuring staff are aware of these mechanisms.

AC K NOW L E DGE M E N T S
We would like to thank our colleagues who supported the review by screening papers in languages other than English; Ayla Cetinkaya, Alicia Diaz-Gil, Verena Hinze, Miran Kim, Ayse Kurtulmus, Edoardo Ostinelli, Camilo Yang.Aaron Worsley supported the review as a health and social care librarian at the Oxford Brookes University.

F U N DI NG I N FOR M AT ION
No Funding was provided for this systematic review.

CON F L IC T OF I N T E R E ST STAT E M E N T
None.

C H A R T 1
Reflexive practice throughout the reviewing process.Unrelated topic (n = 348) Not including inpatient mental health unit (n = 33) Text not found (n = 1) Unable to translate (n = 6) Non-primary research (n = 42) Non research article (n = 4) Not Focused on nurses (n = 2) Included Studies included in review (n = 29) Included studies ongoing (n = 0) Studies awaiting classificaƟon (n = 0) Screening Studies from databases/registers (n = 7146)14470349, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/inm.13317by Test, Wiley Online Library on [25/03/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License F I G U R E 1 PRISMA flowchart.Shows the study identification and selection process.The original database search resulted in 7146 records.

T A B L E 1 MMAT qualitative. Author Are there clear research questions? Does the collected data allow us to address the research questions? Is the qualitative approach appropriate to answer the research question? Are the qualitative data collection methods adequate to address the research question? Are the findings adequately derived from the data? Is the interpretation of results sufficiently substantiated by data? Is there coherence between qualitative data sources, collection, analysis and interpretation?
Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/inm.13317by Test, Wiley Online Library on [25/03/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

Is there an adequate rationale for using a mixed methods design to address the research question? Are the different components of the study effectively integrated to answer the research question? Are the outputs of the integration of qualitative and quantitative components adequately interpreted? Are divergences and inconsistencies between quantitative and qualitative results adequately addressed? Do the different components of the study adhere to the quality criteria of each tradition of the methods involved?
T A B L E 3 MMAT mixed methods.14470349, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/inm.13317by Test, Wiley Online Library on [25/03/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

:
Staff Observation Aggression Scale-Revised (SOAS-R);Tonso et al., 2016 the World Health Organization Joint Program on Workplace Violence in the Health Sector questionnaire).However, data collected regarding racism was not always acknowledged when reporting findings (e.g.Jackson & Ashley, 2005: aggregated data on incidents of physical versus psychological violence without breaking data down to racial harassment; Study details.Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/inm.13317by Test, Wiley Online Library on [25/03/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/inm.13317by Test, Wiley Online Library on [25/03/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Study participants.Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/inm.13317by Test, Wiley Online Library on [25/03/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Female-67% Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/inm.13317by Test, Wiley Online Library on [25/03/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License T A B L E 4T A B L E 5 Study aims and findings.Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/inm.13317by Test, Wiley Online Library on [25/03/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/inm.13317by Test, Wiley Online Library on [25/03/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/inm.13317by Test, Wiley Online Library on [25/03/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/inm.13317by Test, Wiley Online Library on [25/03/2024].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License T A B L E 6