ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE
A cross-sectional survey of factors related to inpatient assault of staff in a forensic psychiatric hospital Erin L. Kelly, Andrew M. Subica, Anthony Fulginiti, John S. Brekke & Raymond W. Novaco Accepted for publication 17 November 2014
Correspondence to E.L. Kelly: e-mail: [email protected] Erin L. Kelly PhD Post Doctoral Scholar Health Services Research Center, University of California, Los Angeles, California, USA Andrew M. Subica PhD Post Doctoral Scholar Psychology Applied Research Center, Loyola Marymount University, Los Angeles, California, USA Anthony Fulginiti MS Graduate Student School of Social Work, University of Southern California, Los Angeles, California, USA John S. Brekke PhD Professor of Psychology School of Social Work, University of Southern California, Los Angeles, California, USA Raymond W. Novaco PhD Professor of Psychology School of Social Ecology, University of California, Irvine, California, USA
KELLY E.L., SUBICA A.M., FULGINITI A., BREKKE J.S. & NOVACO R.W.
A cross-sectional survey of factors related to inpatient assault of staff in a forensic psychiatric hospital. Journal of Advanced Nursing 71(5), 1110–1122. doi: 10.1111/jan.12609
(2015)
Abstract Aim. To understand staff factors associated with patient aggression towards the staff of an inpatient forensic psychiatric hospital. Background. Violence by patients is a serious concern in psychiatric hospitals and staff are the most frequent targets of physical and verbal assault. Assault and its consequences can severely disrupt the hospital environment and impair the functioning of staff members and patients. This study examined the interplay of staff dispositional and interpersonal factors associated with patient violence. Design. This cross-sectional study surveyed the staff of a large public forensic hospital. Methods. A sample of 348 psychiatric staff participated in an online survey about their workplace experiences, psychosocial characteristics and well-being. Data were collected from November – December 2011. Findings. Nearly all staff reported verbal conflict with patients (99%) and 70% reported being assaulted during the previous 12 months. Verbal conflict with other staff (92%) was also high. Multiple regression analyses indicated that in addition to static risk factors (i.e. staff position, years of experience and gender), the risk of assault was associated with the frequency of conflicts with staff and patients, which in turn was moderated by personal stress reactivity. Conclusion. Physical violence by patients was a pervasive threat for a high proportion of staff. Frequent conflict interactions with volatile patients contributed the most risk, but reactivity to conflict was a dynamic risk factor. The strain associated with assault risk and stress reactivity could be prospectively mitigated by resilience enhancement programming for staff. Keywords: assault, forensic mental health, psychiatric nursing, workplace violence
1110
© 2014 John Wiley & Sons Ltd
JAN: ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE
Why is this research or review needed? ● Workplace violence is a serious issue in psychiatric settings, particularly for nurses and psychiatric technicians. ● Research predicting violence has often been atheoretical and focused on static risk factors instead of examining interactive and dispositional risk factors that can be targeted for change through training, education or other interventions in the workplace.
What are the key findings? ● Study findings indicated that the strongest factors accounting for physical assault exposure were patient–staff conflict, intra-staff conflict and stress reactivity to social conflict ● Using a demands-stressed abilities framework, stress reactivity to social conflict moderated the relationships between conflicts and assault, but the nature of those relationships differs between intra-staff conflict and patientstaff.
How should the findings be used to influence policy/ practice/research/education? ● Assault may be an inescapable workplace hazard in psychiatric care but progress can still be made to reduce assaults and their adverse impact on psychiatric staff. ● Examinations of the dynamics leading to staff assault experiences could be guided by a demands-stressed abilities framework. ● This framework suggests that reducing workplace violence
Patient-on-staff violence in a psychiatric hospital
Many investigators have sought to identify risk and protective factors for violence exposure. A preponderance of studies focused on identifying static characteristics of staff or patients associated with assault, such as gender, age, type of diagnosis or position (Whittington & Wykes 1994, Cunningham et al. 2003, Flannery et al. 2007). Some studies have focused on patient psychopathological symptoms, such as impulsivity, delusions, command hallucinations or anger. Fewer studies have explored the interactional aspects of violent behaviours (Nolan et al. 2003, Whittington & Richter 2005) or the context of the ward environment (Alexander & Bowers 2004). There is substantial evidence that static patient and staff characteristics contribute to patient-on-staff assaults, but a recent meta-analysis (Dack et al. 2013) cautioned that given the small effect sizes and the heterogeneity of the associations between static patient factors and violence, these characteristics may have limited utility in terms of predicting or preventing violence. Much of psychiatric care is about relationships. The quality of staff interactions with patients and among themselves has strong implications for patient well-being and propensity for violence (Cornaggia et al. 2011). Researchers who examined dynamic risk factors found that violence is the result of a complex interplay of variables that influence how staff members and patients interact and then later react to those interactions (Whittington & Richter 2005, Hamrin et al. 2009). Therefore, the current study examined social interaction factors related to patient–staff assaults that could be potential targets for intervention.
demands a more holistic examination of how individual reactions to social conflict interplay with relationships among staff, and between staff and patients.
Introduction Workplace violence is a grave concern throughout the labour market (Privitera et al. 2005), with approximately 572,000 non-fatal violent crimes occurring in 2009 (Harrell 2011). In inpatient psychiatric settings, the rate of violence is particularly high (Carmel & Hunter 1989, Soares et al. 2000, Harrell 2011). Although rates have varied across studies (Gordon et al. 1996, Owen et al. 1998), career rates of assault among psychiatric staff have been as high as 75% in some multinational samples (Poster 1996). Adding to the concern, the scale of patient violence towards mental health professionals is likely underestimated (e.g. Hesketh et al. 2003) and increasing (Soares et al. 2000, Privitera et al. 2005). Nearly 80% of violent or aggressive patient episodes target staff members (e.g. Owen et al. 1998).
© 2014 John Wiley & Sons Ltd
Background The consequences of workplace violence are manifold, extending from the individual to the organization (Fry et al. 2002). Negative outcomes for individual staff members can be prolonged (Lanza et al. 2006) and include traumatic symptomology, negative emotions such as fear, helplessness, anger, despair and resentment (Whittington & Wykes 1992, Erdos & Hughes 2001, Needham et al. 2005, Inoue et al. 2006, Richter & Berger 2006, Bimenyimana et al. 2009). On an organizational level, violence can have deleterious effects on the financial and functional health of an organization in terms of medical and litigation costs, difficulties with staffing recruitment and retention, absenteeism, low morale and impaired treatment delivery (Carmel & Hunter 1989, Hunter & Carmel 1992, Lipscomb & Love 1992, Alexander et al. 1998, Erdos & Hughes 2001, Jackson et al. 2001, Bimenyimana et al. 2009). To avoid these individual and organizational costs, violence prediction
1111
E.L. Kelly et al.
Staff interactions with other staff and patients Studies investigating staff characteristics to explain patienton-staff assault have been relatively limited in scope (Hamrin et al. 2009), primarily focusing on sociodemographic variables (e.g. gender, age), work role factors (e.g. staff positions, years of experience) or patient care tasks most associated with violence (e.g. medication administration, daily living tasks and limit setting [Lanza et al. 1991, Whittington & Wykes 1996, Cunningham et al. 2003, Hamrin et al. 2009]). Missing from such analyses are staff social relationships with other staff and patients. Poor relationships between staff members at work, due to either a lack of social support or interpersonal conflict, have been noted in numerous studies as serious sources of stress for psychiatric employees (Eisenberger et al. 1990, Alexander et al. 1998, Munro et al. 1998, Rathod et al. 2000, Mor Barak et al. 2001, Gilbody et al. 2006, Halbesleben 2006), but staff conflict has been unconnected with violence risk. Scant attention has been paid to how psychiatric staff’s interactions with other staff might create an environment in which assault is more likely to occur. High-conflict social interactions between staff can result from tension related to patient care, poor communication, staff attitudes and managerial problems. Staff members who are in conflict with one another, from the interpersonal level to the managerial level, may contribute to patient aggression by failing to create a therapeutic environment (Hesketh et al. 2003, Lanza et al. 2006) or a harmonious staffing environment which has been shown to lower rates of assault (Cornaggia et al. 2011). 1112
Person–environment fit A useful way to approach the study of workplace violence is through Person-Environment Fit theory (French et al. 1974, 1982, Caplan et al. 1980, Caplan 1987). This theory posits that the mismatch between the demands of a job (e.g. social conflicts, either real or perceived) and the ability of individuals to meet those requirements (i.e. their perceived capacity to manage conflict) will influence the occurrence of negative outcomes such as assault (Edwards & Cooper 1990, Figure 1). Individual abilities can be difficult to measure by self-report because respondents may be disinclined to report deficits in abilities or lack a perspective to self-evaluate their abilities. Consequently, several commonly used measures of workplace person–environment fit have operationalized abilities in terms of the amount of stress that staff feel when presented with demands (Osipow 1998, Spielberger & Vagg 1999). Therefore, this study used the term ‘stressed abilities’ as the operational variable. Based on a demands–stressed abilities framework, a staff member’s ability to cope with social conflict with others may moderate the relationship between the frequency of conflict and patient-on-staff assault risk. For example, staff who are highly sensitive to conflict may be avoidant of it and thereby not acquire conflict management skills or may be viewed as vulnerable by patients and targeted for victimization. It is also possible that being highly reactive to conflict may motivate the development of conflict de-escalation skills. Staff who are less reactive to social conflict may be at higher risk of assault by missing the social cues leading to conflict; conversely, being less reactive to conflict could dispose them to more calmly intervene in conflict situations.
3 Low abilities 2·5 Adverse outcome
efforts have long sought to identify who is likely to be violent and under what conditions (Lyon et al. 2001). However, as violence is relatively rare in most workplace settings, predicting it is challenging. To predict psychiatric inpatient aggression, one framework conceptualizes risk factors on a continuum from static to dynamic (Conroy & Murrie 2007). Static risk factors (e.g. age, gender, ethnicity, violence history) cannot be modified, whereas dynamic risk factors (e.g. mood, mental state, interpersonal functioning) are modifiable and thus represent violence prevention targets (Conroy & Murrie 2007). A review of several dozen aggression management programs found that most focused on identifying potential violence situations and improving high-conflict staff–patient interactions, such as during the de-escalation of aggressive patients or during the use of containment procedures (Farrell & Cubit 2005). Research studies have commonly focused on a single domain of patient or staff factors and infrequently examined them in combination.
High abilities
2 1·5 1 0·5 0 Low demands
High demands
Figure 1 Theoretical relationship of demands and abilities with adverse outcomes. © 2014 John Wiley & Sons Ltd
JAN: ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE
As previously noted, interpersonal conflict represents a major source of stress for staff, therefore it is important to understand whether stress sensitivity and responsivity to conflict affects the risk of assault. Examining these relationships between job demands and stressed abilities could identify staff sub-populations that are particularly challenged by high-risk social interactions that compromise their workplace safety. Using person–environment fit theory, we examined the relationships of conflict, reactivity to conflict and risk of assault. Specifically, we assessed the combination of static risk factors (gender, years of experience and work role) and dynamic staff risk factors (presence of conflict with others and individual stress reactivity to social conflict) in association with assault.
The study Aims The study had three aims: to (1) establish the prevalence of physical assault (patient-on-staff) in a large public forensic hospital; (2) examine the additive relationships between static and dynamic staff risk factors for patient-on-staff assault and (3) explore whether social conflict stress reactivity moderated the relationships between social conflict frequency and patient-on-staff assault.
Design A cross-sectional online survey methodology was used.
Setting The setting was a large mixed-gender hospital in California with a capacity of 1287 beds, but with an average treatment population of more than 1500 patients. Civil commitments are permitted, but the vast majority of patients are judicial commitments. Most individuals treated at the hospital are diagnosed with severe mental illness and co-occurring substance use disorders.
Participants An online survey was sent through the hospital’s internal email system to the hospital’s 1794 total clinical staff members. The executive director and medical director sent a hospital-wide memorandum endorsing the survey and encouraging participation. The survey link was opened by 488 employees (27%). Of those 488, 14 declined to © 2014 John Wiley & Sons Ltd
Patient-on-staff violence in a psychiatric hospital
participate, 19 agreed but did not enter any data and 107 did not answer questions about their assault experiences. The final sample consisted of 348 clinical staff (71% of those who began the survey, 19% of clinical staff). A prior study by Leon et al. (2008), using a similar method and participant population, had a 17% participation rate.
Data collection Data were collected from November–December 2011. Gender was coded as male = 0, female = 1. Race and ethnicity was dummy coded for use in descriptive analyses (1 = Caucasian, 2 = African American, 3 = Hispanic, 4 = Asian and 5 = other). Due to the small number of employees in the other category, this group was not included in analyses. Participants identified their current position (psychiatric technician (which The American Association of Psychiatric Technicians defines ‘psychiatric technicians’ as entry-level employees who give direct care for individuals with mental illnesses or developmental disabilities. In other countries or contexts the titles of a position with similar responsibilities may include psychiatric, mental health or behavioural health in conjunction with aide, worker, technician, assistant or associate), senior psychiatric technician, Registered Nurse, unit supervisor, rehabilitation therapist, psychologist, social worker, psychiatrist or other) and the number of years of work experience in the field of mental health. Staff rated how often they had been physically assaulted during the previous year using an 8-item measure with a 4point Likert scale (0 = never, 1 = one or two times, 2 = three to five times, 3 = 6 or more times) developed for this study. Items varied in their severity: (1) spit at; (2) touched/grabbed aggressively; (3) kissed, fondled or patient sexually exposed themselves; (4) pushed or knocked down; (5) kicked, punched, bit or slapped; (6) hit on the head; (7) had object thrown at them and (8) other). An averaged scaled frequency of physical assault across the eight items was calculated. Participants were also asked whether they had been absent from work due to an injury caused by a patient during the previous 6 months (0 = no, 1 = yes). Patient–staff conflicts were measured with an 8-item scale modified from a subset of items in Psychiatric Nurses Stress Inventory (Sullivan 1993), on which participants reported the frequency of managing patients in high-conflict or combative situations with 5-point ratings (1 = never to 5 = very often). The eight items were classified in three categories: (1) patient–staff conflict, involving the use of containment procedures or restraints (physically restraining a patient, chemically restraining a patient, placing a patient in seclusion); (2) verbal antagonism or challenging behaviours from 1113
E.L. Kelly et al.
patients (patients swearing or yelling at you, patients who tried to intimidate you, difficult or demanding patients) and (3) monitoring potentially dangerous patients (dealing with potentially suicidal patients, continuous observation of patients on a one-to-one basis). A total average scaled score was computed for analyses. Subdomains of patient containment, dealing with antagonistic or challenging patients and monitoring potentially dangerous patients were used for descriptive purposes. The internal reliability of the total scale score was excellent (a = 089). Staff conflict with other employees was assessed using a 3-item subscale of the Psychiatric Nurses Stress Inventory (Sullivan 1993). Respondents rated the occurrence of social conflict with other staff (difficulties in working with particular nurses on the ward, disagreements concerning the treatment of a patient, conflicts with management) during the previous 6 months (1 = never to 5 = very often). Higher scores indicated more conflict with co-workers and other staff. Staff stress reactivity to social conflict (feeling affected, upset or irritated) was assessed by the 5-item subscale of that name from the Perceived Stress Reactivity Scale (Schulz et al. 2005). The scale items have two components: a description of a potentially stressful situation and three potential stress responses, i.e. (a) When I argue with other people, (1) I usually calm down quickly, (2) I usually stay upset for some time or (3) it usually takes me a long time until I calm down; (b) When I argue with others, (1) I usually get very upset, (2) I usually get a little bit upset, or (3) I usually don’t get upset; (c) When I have conflicts with others that may not be immediately resolved, (1) I generally shrug it off, (2) it usually affects me a little, or (3) it usually affects me a lot; (d) When I am wrongly criticized by others, (1) I am normally annoyed for a long time, (2) I am annoyed for just a short time, or (3) in general, I am hardly annoyed at all; and (e) When others say something incorrect about me, (1) I usually get quite upset, (2) I normally get a little bit upset, or (3) in general, I shrug it off. This measure of reactivity to social conflict has had acceptable to good reliability among USA samples and in the current study (a = 073). This scale has also shown good test–retest reliability across 26 days (r = 072; Schlotz et al. 2011). Higher scores indicated more reactivity to social conflict.
Procedure Hospital executive staff members were asked to review and make suggestions about the survey prior to distribution. Following their approval and endorsement of the final survey by email to staff, the first author distributed the online 1114
survey to all clinical staff members. The survey took approximately 1 hour to complete. On clicking the survey link, potential participants were asked to affirm informed consent and then take the survey. All participants were entered into a lottery and 10 participants received $100 for their participation. The hospital received a report of the initial findings.
Ethical consideration This study was approved by the institutional review board of [the University of California, Irvine]. Participation in the study was voluntary and participants received a study information sheet before beginning the survey.
Data analysis All data were inspected for outliers and normality of distributions by examining skew, kurtosis and Q-Q plots. Need for data transformation was not indicated. Descriptive statistics, Pearson’s correlations, one-way analyses of variance and hierarchical linear regression were conducted using SPSS version 18.0 (SPSS Inc. 2009). For the regression model, tests of variance inflation factor and tolerance did not indicate multicollinearity was an issue. To assess differences among staff in time spent with patients, participants were categorized into groups by position according to time spent on the unit. Ward staff (psychiatric technicians, senior psychiatric technicians, unit supervisors, Registered Nurses; n = 231) were indexed as having had the highest levels of patient contact, followed by clinical care staff (rehabilitation therapists, psychologists, social workers, psychiatrists; n = 91) and supervisory staff (administrative supervisors, clinical supervisors; n = 12). This three-level grouping of staff position was used in ANOVA analyses and dummy coded for regression analyses. A four-step hierarchical linear regression was used to assess covariates of patient-on-staff physical assault. Preestablished static staff covariates were entered in the first block. In the second block, the frequencies of patient–staff conflict and staff–staff conflict were entered. In the third block, the dispositional factor (social conflict stress reactivity) was entered. In the final block, two interaction terms were entered with social conflict stress reactivity as a moderator of (1) frequency of staff conflict and (2) frequency of patient care interactions.
Validity and reliability Quantitative survey instruments were derived from well-established surveys with good psychometric validity, © 2014 John Wiley & Sons Ltd
JAN: ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE
Patient-on-staff violence in a psychiatric hospital
Risk factors for patient assault Sociodemographic risk factors. Men experienced higher scaled frequencies of assault than women (scaled mean = 046, SD 052 vs. scaled mean = 033, SD 041, respectively; t[345] = 242, P = 002). With respect to staff position, Ward staff members reported more assaults (scaled mean = 048, SD 050, P < 0001) than Clinical care (scaled M = 014, SD 018) and Supervisory employees (scaled mean = 017, SD 024, P = 0002; F[2, 331] = 2118, P < 0001; per Games-Howell post hoc tests). No significant differences emerged in the rates of assault between clinical care and supervisory staff (P = 094). Years of experience in the mental health field was not significantly related to the frequency of assaults (r = 006, P = 031).
including the Psychiatric Nurses Stress Inventory (Sullivan 1993) and the Perceived Stress Reactivity Scale (Schulz et al. 2005). Reliability of the scales used in this study was assessed using Cronbach’s alpha and all scales demonstrated acceptable to excellent reliability (a = 073–089).
Results Sample characteristics The sample was diverse in terms of gender and race and ethnicity. The majority of respondents were women (69%, n = 239). The sample was racially and ethnically diverse; 37% identified themselves as Caucasian, 21% as African American, 21% as Hispanic, 11% as Asian American and 10% as mixed race or other. On average, staff had worked in mental health for 144 years (SD = 91). In comparison to the hospital’s demographic data (gender, position, race/ethnicity) the sample was slightly over-representative of females and Caucasians.
Social conflict risk factors. Ninety-nine percent of participants reported experiencing conflict with patients, with 179% experiencing conflict often or very often. With regard to antagonism or challenging behaviours from patients during the previous year, only 5% of staff did not report experiencing any of these behaviours and 274% of respondents reported experiencing them often or very often. Ward staff reported significantly more conflict with patients (mean = 335, SD 089, P < 0001) than Clinical (mean = 244, SD 070; F[2, 329] = 3859, P < 0001; per Games-Howell post hoc tests). Supervisory staff did not differ from either Ward or Clinical staff in terms of conflict (mean = 267, SD 004). On average, participants reported occasional use of patient containment procedures such as seclusion, restraints or unscheduled medications (scaled mean = 231, SD 110). The majority of staff (80%) had used these procedures at least once during the prior 6 months, but only 9% reported using them often or very often. Forty-seven percent of
Patient assaults Seventy percent of staff reported experiencing at least one physical assault during the previous year. Figure 2 presents the percentages across categories of assault type. In terms of assault severity, 42% of participants reported at least one incident of non-sexual serious assault (e.g. being kicked, punched, slapped, hit in the head, pushed or knocked down) and 64% reported at least one mild assault (e.g. being grabbed, touched or spat on). The average scaled frequencies across assault types suggested that staff were not often assaulted in multiple forms (scaled M = 037, first interquartile = 000, third interquartile = 057). Twelve percent of respondents reported taking off at least 1 day of work during the previous 6 months due to a patient assault. 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% r
gg
t
n
io
s es
xu
he
c ou
ed
r
G
d/
ch
he
un
s Pu
e
n
i it
he
n
H
P d/
At
ts
ul
a ss
ow
r
th
e
ck
Ki
ad
a Sl
t/
Bi
/ ed
o Kn
ab
d
pp
ov
ck
/T
d be
a al
Se
er
d
tA
a Sp
t ec
th
la
ca
i ys
bj
O
r he
ph
t
O
Figure 2 Percentage of staff reporting each type of assault during the previous year. © 2014 John Wiley & Sons Ltd
1115
E.L. Kelly et al.
Table 2. In the first block of static risk factors, men and Ward staff were at greater risk of physical assault than women and Clinical or Supervisory staff. Experience in the field was not significantly associated with assault. In the second step, patient–staff conflict was positively related to assault risk (semi partial R2 = 181%) but intra-staff conflict was not. In the third block of dispositional risk factors, being less reactive to social conflict was significantly associated with more frequent assaults. In the final step, social conflict stress reactivity significantly moderated the relationship between patient-on-staff assault and intra-staff conflict (Figure 3). Specifically, staff members who reported low social conflict stress reactivity and experienced greater intra-staff conflict were at highest risk of assault. Individuals who were highly reactive to social conflict and experienced more intra-staff conflict were at low risk of assault. Reactivity to conflict was unrelated to risk of assault among participants with low levels of intra-staff conflict. A similar pattern emerged when social conflict stress reactivity was tested as a moderator of patient–staff conflict (Figure 4). Staff who had few conflictual interactions with patients were at low risk of assault, regardless of their social conflict stress reactivity. However, respondents who reported high levels of patient–staff conflict were at higher risk of assault and were most likely to be assaulted if they had low social conflict stress reactivity. Finally, staff with high levels of social conflict stress reactivity were at a slightly lower risk for assault than those with low reactivity.
respondents reported maintaining continuous observation of a patient often or very often. On average, conflict among staff members occurred occasionally (scaled mean = 252, SD 100) although 92% of participants reported some conflict with co-workers. For conflict that happened often or very often, the sources of conflict were roughly equivalent: 225% with fellow staff, 183% about treatment of patients and 208% with management. Ward staff reported significantly more intra-staff conflict (M = 261, SD = 105) than Clinical staff (M = 230, SD = 085, P = 002), yet neither group differed from supervisory staff (M = 256, SD = 067; F[2, 329] = 330, P = 004; per Games-Howell post hoc tests). Dispositional risk factors Reactivity to social conflict was normally distributed across staff members (mean = 194, SD 043). A third of participants reported moderate difficulty with letting go of conflict but only 121% reported having extreme difficulty. Across position groups, there were no significant differences in reactivity to conflict (F[2, 318] = 075, P = 047). Correlations Correlations between main study variables are presented in Table 1. The frequency of patient assaults was substantially correlated with the frequency of patient–staff conflict (r = 054) and had a lesser association with intra-staff conflict (r = 021). Intra-staff conflict and patient–staff conflict were also moderately related (r = 038). Interestingly, reactivity to social conflict was inversely related to the frequency of assault (r = 016) and had a slight association with intra-staff conflict frequency (r = 019), but was statistically unrelated to the frequency of patient–staff conflict (r = 004). This indicated that staff may not react to conflict with patients in the same manner as they would to conflict with other staff.
Discussion Workplace violence may be an unavoidable occupational hazard for psychiatric hospital staff (Anderson & West 2011). The present study sought to enhance understanding of the factors contributing to violence through a demands– stressed abilities framework. It is the first study to test stress reactivity to social conflict as a variable relevant to
Conflict, reactivity and assault Hierarchical linear regression results about the scaled frequency of assault and risk variables are presented in
Table 1 Correlations between main study variables. Variable 1. 2. 3. 4. 5. 6.
Gender Work experience in mental health Patient–staff conflict Staff conflict Reactivity to social conflict Patient assault frequency
1
2
009 011 002 012 013
3
011 009 000 006
4
038** 004 054**
019** 021**
5
016*
*P < 001; **P < 0001. Gender is coded as 0 = male, 1 = female. 1116
© 2014 John Wiley & Sons Ltd
0031 0119 0024 0100
A cross-sectional survey of factors related to inpatient assault of staff in a forensic psychiatric hospital Erin L. Kelly, Andrew M. Subica, Anthony Fulginiti, John S. Brekke & Raymond W. Novaco Accepted for publication 17 November 2014
Correspondence to E.L. Kelly: e-mail: [email protected] Erin L. Kelly PhD Post Doctoral Scholar Health Services Research Center, University of California, Los Angeles, California, USA Andrew M. Subica PhD Post Doctoral Scholar Psychology Applied Research Center, Loyola Marymount University, Los Angeles, California, USA Anthony Fulginiti MS Graduate Student School of Social Work, University of Southern California, Los Angeles, California, USA John S. Brekke PhD Professor of Psychology School of Social Work, University of Southern California, Los Angeles, California, USA Raymond W. Novaco PhD Professor of Psychology School of Social Ecology, University of California, Irvine, California, USA
KELLY E.L., SUBICA A.M., FULGINITI A., BREKKE J.S. & NOVACO R.W.
A cross-sectional survey of factors related to inpatient assault of staff in a forensic psychiatric hospital. Journal of Advanced Nursing 71(5), 1110–1122. doi: 10.1111/jan.12609
(2015)
Abstract Aim. To understand staff factors associated with patient aggression towards the staff of an inpatient forensic psychiatric hospital. Background. Violence by patients is a serious concern in psychiatric hospitals and staff are the most frequent targets of physical and verbal assault. Assault and its consequences can severely disrupt the hospital environment and impair the functioning of staff members and patients. This study examined the interplay of staff dispositional and interpersonal factors associated with patient violence. Design. This cross-sectional study surveyed the staff of a large public forensic hospital. Methods. A sample of 348 psychiatric staff participated in an online survey about their workplace experiences, psychosocial characteristics and well-being. Data were collected from November – December 2011. Findings. Nearly all staff reported verbal conflict with patients (99%) and 70% reported being assaulted during the previous 12 months. Verbal conflict with other staff (92%) was also high. Multiple regression analyses indicated that in addition to static risk factors (i.e. staff position, years of experience and gender), the risk of assault was associated with the frequency of conflicts with staff and patients, which in turn was moderated by personal stress reactivity. Conclusion. Physical violence by patients was a pervasive threat for a high proportion of staff. Frequent conflict interactions with volatile patients contributed the most risk, but reactivity to conflict was a dynamic risk factor. The strain associated with assault risk and stress reactivity could be prospectively mitigated by resilience enhancement programming for staff. Keywords: assault, forensic mental health, psychiatric nursing, workplace violence
1110
© 2014 John Wiley & Sons Ltd
JAN: ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE
Why is this research or review needed? ● Workplace violence is a serious issue in psychiatric settings, particularly for nurses and psychiatric technicians. ● Research predicting violence has often been atheoretical and focused on static risk factors instead of examining interactive and dispositional risk factors that can be targeted for change through training, education or other interventions in the workplace.
What are the key findings? ● Study findings indicated that the strongest factors accounting for physical assault exposure were patient–staff conflict, intra-staff conflict and stress reactivity to social conflict ● Using a demands-stressed abilities framework, stress reactivity to social conflict moderated the relationships between conflicts and assault, but the nature of those relationships differs between intra-staff conflict and patientstaff.
How should the findings be used to influence policy/ practice/research/education? ● Assault may be an inescapable workplace hazard in psychiatric care but progress can still be made to reduce assaults and their adverse impact on psychiatric staff. ● Examinations of the dynamics leading to staff assault experiences could be guided by a demands-stressed abilities framework. ● This framework suggests that reducing workplace violence
Patient-on-staff violence in a psychiatric hospital
Many investigators have sought to identify risk and protective factors for violence exposure. A preponderance of studies focused on identifying static characteristics of staff or patients associated with assault, such as gender, age, type of diagnosis or position (Whittington & Wykes 1994, Cunningham et al. 2003, Flannery et al. 2007). Some studies have focused on patient psychopathological symptoms, such as impulsivity, delusions, command hallucinations or anger. Fewer studies have explored the interactional aspects of violent behaviours (Nolan et al. 2003, Whittington & Richter 2005) or the context of the ward environment (Alexander & Bowers 2004). There is substantial evidence that static patient and staff characteristics contribute to patient-on-staff assaults, but a recent meta-analysis (Dack et al. 2013) cautioned that given the small effect sizes and the heterogeneity of the associations between static patient factors and violence, these characteristics may have limited utility in terms of predicting or preventing violence. Much of psychiatric care is about relationships. The quality of staff interactions with patients and among themselves has strong implications for patient well-being and propensity for violence (Cornaggia et al. 2011). Researchers who examined dynamic risk factors found that violence is the result of a complex interplay of variables that influence how staff members and patients interact and then later react to those interactions (Whittington & Richter 2005, Hamrin et al. 2009). Therefore, the current study examined social interaction factors related to patient–staff assaults that could be potential targets for intervention.
demands a more holistic examination of how individual reactions to social conflict interplay with relationships among staff, and between staff and patients.
Introduction Workplace violence is a grave concern throughout the labour market (Privitera et al. 2005), with approximately 572,000 non-fatal violent crimes occurring in 2009 (Harrell 2011). In inpatient psychiatric settings, the rate of violence is particularly high (Carmel & Hunter 1989, Soares et al. 2000, Harrell 2011). Although rates have varied across studies (Gordon et al. 1996, Owen et al. 1998), career rates of assault among psychiatric staff have been as high as 75% in some multinational samples (Poster 1996). Adding to the concern, the scale of patient violence towards mental health professionals is likely underestimated (e.g. Hesketh et al. 2003) and increasing (Soares et al. 2000, Privitera et al. 2005). Nearly 80% of violent or aggressive patient episodes target staff members (e.g. Owen et al. 1998).
© 2014 John Wiley & Sons Ltd
Background The consequences of workplace violence are manifold, extending from the individual to the organization (Fry et al. 2002). Negative outcomes for individual staff members can be prolonged (Lanza et al. 2006) and include traumatic symptomology, negative emotions such as fear, helplessness, anger, despair and resentment (Whittington & Wykes 1992, Erdos & Hughes 2001, Needham et al. 2005, Inoue et al. 2006, Richter & Berger 2006, Bimenyimana et al. 2009). On an organizational level, violence can have deleterious effects on the financial and functional health of an organization in terms of medical and litigation costs, difficulties with staffing recruitment and retention, absenteeism, low morale and impaired treatment delivery (Carmel & Hunter 1989, Hunter & Carmel 1992, Lipscomb & Love 1992, Alexander et al. 1998, Erdos & Hughes 2001, Jackson et al. 2001, Bimenyimana et al. 2009). To avoid these individual and organizational costs, violence prediction
1111
E.L. Kelly et al.
Staff interactions with other staff and patients Studies investigating staff characteristics to explain patienton-staff assault have been relatively limited in scope (Hamrin et al. 2009), primarily focusing on sociodemographic variables (e.g. gender, age), work role factors (e.g. staff positions, years of experience) or patient care tasks most associated with violence (e.g. medication administration, daily living tasks and limit setting [Lanza et al. 1991, Whittington & Wykes 1996, Cunningham et al. 2003, Hamrin et al. 2009]). Missing from such analyses are staff social relationships with other staff and patients. Poor relationships between staff members at work, due to either a lack of social support or interpersonal conflict, have been noted in numerous studies as serious sources of stress for psychiatric employees (Eisenberger et al. 1990, Alexander et al. 1998, Munro et al. 1998, Rathod et al. 2000, Mor Barak et al. 2001, Gilbody et al. 2006, Halbesleben 2006), but staff conflict has been unconnected with violence risk. Scant attention has been paid to how psychiatric staff’s interactions with other staff might create an environment in which assault is more likely to occur. High-conflict social interactions between staff can result from tension related to patient care, poor communication, staff attitudes and managerial problems. Staff members who are in conflict with one another, from the interpersonal level to the managerial level, may contribute to patient aggression by failing to create a therapeutic environment (Hesketh et al. 2003, Lanza et al. 2006) or a harmonious staffing environment which has been shown to lower rates of assault (Cornaggia et al. 2011). 1112
Person–environment fit A useful way to approach the study of workplace violence is through Person-Environment Fit theory (French et al. 1974, 1982, Caplan et al. 1980, Caplan 1987). This theory posits that the mismatch between the demands of a job (e.g. social conflicts, either real or perceived) and the ability of individuals to meet those requirements (i.e. their perceived capacity to manage conflict) will influence the occurrence of negative outcomes such as assault (Edwards & Cooper 1990, Figure 1). Individual abilities can be difficult to measure by self-report because respondents may be disinclined to report deficits in abilities or lack a perspective to self-evaluate their abilities. Consequently, several commonly used measures of workplace person–environment fit have operationalized abilities in terms of the amount of stress that staff feel when presented with demands (Osipow 1998, Spielberger & Vagg 1999). Therefore, this study used the term ‘stressed abilities’ as the operational variable. Based on a demands–stressed abilities framework, a staff member’s ability to cope with social conflict with others may moderate the relationship between the frequency of conflict and patient-on-staff assault risk. For example, staff who are highly sensitive to conflict may be avoidant of it and thereby not acquire conflict management skills or may be viewed as vulnerable by patients and targeted for victimization. It is also possible that being highly reactive to conflict may motivate the development of conflict de-escalation skills. Staff who are less reactive to social conflict may be at higher risk of assault by missing the social cues leading to conflict; conversely, being less reactive to conflict could dispose them to more calmly intervene in conflict situations.
3 Low abilities 2·5 Adverse outcome
efforts have long sought to identify who is likely to be violent and under what conditions (Lyon et al. 2001). However, as violence is relatively rare in most workplace settings, predicting it is challenging. To predict psychiatric inpatient aggression, one framework conceptualizes risk factors on a continuum from static to dynamic (Conroy & Murrie 2007). Static risk factors (e.g. age, gender, ethnicity, violence history) cannot be modified, whereas dynamic risk factors (e.g. mood, mental state, interpersonal functioning) are modifiable and thus represent violence prevention targets (Conroy & Murrie 2007). A review of several dozen aggression management programs found that most focused on identifying potential violence situations and improving high-conflict staff–patient interactions, such as during the de-escalation of aggressive patients or during the use of containment procedures (Farrell & Cubit 2005). Research studies have commonly focused on a single domain of patient or staff factors and infrequently examined them in combination.
High abilities
2 1·5 1 0·5 0 Low demands
High demands
Figure 1 Theoretical relationship of demands and abilities with adverse outcomes. © 2014 John Wiley & Sons Ltd
JAN: ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE
As previously noted, interpersonal conflict represents a major source of stress for staff, therefore it is important to understand whether stress sensitivity and responsivity to conflict affects the risk of assault. Examining these relationships between job demands and stressed abilities could identify staff sub-populations that are particularly challenged by high-risk social interactions that compromise their workplace safety. Using person–environment fit theory, we examined the relationships of conflict, reactivity to conflict and risk of assault. Specifically, we assessed the combination of static risk factors (gender, years of experience and work role) and dynamic staff risk factors (presence of conflict with others and individual stress reactivity to social conflict) in association with assault.
The study Aims The study had three aims: to (1) establish the prevalence of physical assault (patient-on-staff) in a large public forensic hospital; (2) examine the additive relationships between static and dynamic staff risk factors for patient-on-staff assault and (3) explore whether social conflict stress reactivity moderated the relationships between social conflict frequency and patient-on-staff assault.
Design A cross-sectional online survey methodology was used.
Setting The setting was a large mixed-gender hospital in California with a capacity of 1287 beds, but with an average treatment population of more than 1500 patients. Civil commitments are permitted, but the vast majority of patients are judicial commitments. Most individuals treated at the hospital are diagnosed with severe mental illness and co-occurring substance use disorders.
Participants An online survey was sent through the hospital’s internal email system to the hospital’s 1794 total clinical staff members. The executive director and medical director sent a hospital-wide memorandum endorsing the survey and encouraging participation. The survey link was opened by 488 employees (27%). Of those 488, 14 declined to © 2014 John Wiley & Sons Ltd
Patient-on-staff violence in a psychiatric hospital
participate, 19 agreed but did not enter any data and 107 did not answer questions about their assault experiences. The final sample consisted of 348 clinical staff (71% of those who began the survey, 19% of clinical staff). A prior study by Leon et al. (2008), using a similar method and participant population, had a 17% participation rate.
Data collection Data were collected from November–December 2011. Gender was coded as male = 0, female = 1. Race and ethnicity was dummy coded for use in descriptive analyses (1 = Caucasian, 2 = African American, 3 = Hispanic, 4 = Asian and 5 = other). Due to the small number of employees in the other category, this group was not included in analyses. Participants identified their current position (psychiatric technician (which The American Association of Psychiatric Technicians defines ‘psychiatric technicians’ as entry-level employees who give direct care for individuals with mental illnesses or developmental disabilities. In other countries or contexts the titles of a position with similar responsibilities may include psychiatric, mental health or behavioural health in conjunction with aide, worker, technician, assistant or associate), senior psychiatric technician, Registered Nurse, unit supervisor, rehabilitation therapist, psychologist, social worker, psychiatrist or other) and the number of years of work experience in the field of mental health. Staff rated how often they had been physically assaulted during the previous year using an 8-item measure with a 4point Likert scale (0 = never, 1 = one or two times, 2 = three to five times, 3 = 6 or more times) developed for this study. Items varied in their severity: (1) spit at; (2) touched/grabbed aggressively; (3) kissed, fondled or patient sexually exposed themselves; (4) pushed or knocked down; (5) kicked, punched, bit or slapped; (6) hit on the head; (7) had object thrown at them and (8) other). An averaged scaled frequency of physical assault across the eight items was calculated. Participants were also asked whether they had been absent from work due to an injury caused by a patient during the previous 6 months (0 = no, 1 = yes). Patient–staff conflicts were measured with an 8-item scale modified from a subset of items in Psychiatric Nurses Stress Inventory (Sullivan 1993), on which participants reported the frequency of managing patients in high-conflict or combative situations with 5-point ratings (1 = never to 5 = very often). The eight items were classified in three categories: (1) patient–staff conflict, involving the use of containment procedures or restraints (physically restraining a patient, chemically restraining a patient, placing a patient in seclusion); (2) verbal antagonism or challenging behaviours from 1113
E.L. Kelly et al.
patients (patients swearing or yelling at you, patients who tried to intimidate you, difficult or demanding patients) and (3) monitoring potentially dangerous patients (dealing with potentially suicidal patients, continuous observation of patients on a one-to-one basis). A total average scaled score was computed for analyses. Subdomains of patient containment, dealing with antagonistic or challenging patients and monitoring potentially dangerous patients were used for descriptive purposes. The internal reliability of the total scale score was excellent (a = 089). Staff conflict with other employees was assessed using a 3-item subscale of the Psychiatric Nurses Stress Inventory (Sullivan 1993). Respondents rated the occurrence of social conflict with other staff (difficulties in working with particular nurses on the ward, disagreements concerning the treatment of a patient, conflicts with management) during the previous 6 months (1 = never to 5 = very often). Higher scores indicated more conflict with co-workers and other staff. Staff stress reactivity to social conflict (feeling affected, upset or irritated) was assessed by the 5-item subscale of that name from the Perceived Stress Reactivity Scale (Schulz et al. 2005). The scale items have two components: a description of a potentially stressful situation and three potential stress responses, i.e. (a) When I argue with other people, (1) I usually calm down quickly, (2) I usually stay upset for some time or (3) it usually takes me a long time until I calm down; (b) When I argue with others, (1) I usually get very upset, (2) I usually get a little bit upset, or (3) I usually don’t get upset; (c) When I have conflicts with others that may not be immediately resolved, (1) I generally shrug it off, (2) it usually affects me a little, or (3) it usually affects me a lot; (d) When I am wrongly criticized by others, (1) I am normally annoyed for a long time, (2) I am annoyed for just a short time, or (3) in general, I am hardly annoyed at all; and (e) When others say something incorrect about me, (1) I usually get quite upset, (2) I normally get a little bit upset, or (3) in general, I shrug it off. This measure of reactivity to social conflict has had acceptable to good reliability among USA samples and in the current study (a = 073). This scale has also shown good test–retest reliability across 26 days (r = 072; Schlotz et al. 2011). Higher scores indicated more reactivity to social conflict.
Procedure Hospital executive staff members were asked to review and make suggestions about the survey prior to distribution. Following their approval and endorsement of the final survey by email to staff, the first author distributed the online 1114
survey to all clinical staff members. The survey took approximately 1 hour to complete. On clicking the survey link, potential participants were asked to affirm informed consent and then take the survey. All participants were entered into a lottery and 10 participants received $100 for their participation. The hospital received a report of the initial findings.
Ethical consideration This study was approved by the institutional review board of [the University of California, Irvine]. Participation in the study was voluntary and participants received a study information sheet before beginning the survey.
Data analysis All data were inspected for outliers and normality of distributions by examining skew, kurtosis and Q-Q plots. Need for data transformation was not indicated. Descriptive statistics, Pearson’s correlations, one-way analyses of variance and hierarchical linear regression were conducted using SPSS version 18.0 (SPSS Inc. 2009). For the regression model, tests of variance inflation factor and tolerance did not indicate multicollinearity was an issue. To assess differences among staff in time spent with patients, participants were categorized into groups by position according to time spent on the unit. Ward staff (psychiatric technicians, senior psychiatric technicians, unit supervisors, Registered Nurses; n = 231) were indexed as having had the highest levels of patient contact, followed by clinical care staff (rehabilitation therapists, psychologists, social workers, psychiatrists; n = 91) and supervisory staff (administrative supervisors, clinical supervisors; n = 12). This three-level grouping of staff position was used in ANOVA analyses and dummy coded for regression analyses. A four-step hierarchical linear regression was used to assess covariates of patient-on-staff physical assault. Preestablished static staff covariates were entered in the first block. In the second block, the frequencies of patient–staff conflict and staff–staff conflict were entered. In the third block, the dispositional factor (social conflict stress reactivity) was entered. In the final block, two interaction terms were entered with social conflict stress reactivity as a moderator of (1) frequency of staff conflict and (2) frequency of patient care interactions.
Validity and reliability Quantitative survey instruments were derived from well-established surveys with good psychometric validity, © 2014 John Wiley & Sons Ltd
JAN: ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE
Patient-on-staff violence in a psychiatric hospital
Risk factors for patient assault Sociodemographic risk factors. Men experienced higher scaled frequencies of assault than women (scaled mean = 046, SD 052 vs. scaled mean = 033, SD 041, respectively; t[345] = 242, P = 002). With respect to staff position, Ward staff members reported more assaults (scaled mean = 048, SD 050, P < 0001) than Clinical care (scaled M = 014, SD 018) and Supervisory employees (scaled mean = 017, SD 024, P = 0002; F[2, 331] = 2118, P < 0001; per Games-Howell post hoc tests). No significant differences emerged in the rates of assault between clinical care and supervisory staff (P = 094). Years of experience in the mental health field was not significantly related to the frequency of assaults (r = 006, P = 031).
including the Psychiatric Nurses Stress Inventory (Sullivan 1993) and the Perceived Stress Reactivity Scale (Schulz et al. 2005). Reliability of the scales used in this study was assessed using Cronbach’s alpha and all scales demonstrated acceptable to excellent reliability (a = 073–089).
Results Sample characteristics The sample was diverse in terms of gender and race and ethnicity. The majority of respondents were women (69%, n = 239). The sample was racially and ethnically diverse; 37% identified themselves as Caucasian, 21% as African American, 21% as Hispanic, 11% as Asian American and 10% as mixed race or other. On average, staff had worked in mental health for 144 years (SD = 91). In comparison to the hospital’s demographic data (gender, position, race/ethnicity) the sample was slightly over-representative of females and Caucasians.
Social conflict risk factors. Ninety-nine percent of participants reported experiencing conflict with patients, with 179% experiencing conflict often or very often. With regard to antagonism or challenging behaviours from patients during the previous year, only 5% of staff did not report experiencing any of these behaviours and 274% of respondents reported experiencing them often or very often. Ward staff reported significantly more conflict with patients (mean = 335, SD 089, P < 0001) than Clinical (mean = 244, SD 070; F[2, 329] = 3859, P < 0001; per Games-Howell post hoc tests). Supervisory staff did not differ from either Ward or Clinical staff in terms of conflict (mean = 267, SD 004). On average, participants reported occasional use of patient containment procedures such as seclusion, restraints or unscheduled medications (scaled mean = 231, SD 110). The majority of staff (80%) had used these procedures at least once during the prior 6 months, but only 9% reported using them often or very often. Forty-seven percent of
Patient assaults Seventy percent of staff reported experiencing at least one physical assault during the previous year. Figure 2 presents the percentages across categories of assault type. In terms of assault severity, 42% of participants reported at least one incident of non-sexual serious assault (e.g. being kicked, punched, slapped, hit in the head, pushed or knocked down) and 64% reported at least one mild assault (e.g. being grabbed, touched or spat on). The average scaled frequencies across assault types suggested that staff were not often assaulted in multiple forms (scaled M = 037, first interquartile = 000, third interquartile = 057). Twelve percent of respondents reported taking off at least 1 day of work during the previous 6 months due to a patient assault. 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% r
gg
t
n
io
s es
xu
he
c ou
ed
r
G
d/
ch
he
un
s Pu
e
n
i it
he
n
H
P d/
At
ts
ul
a ss
ow
r
th
e
ck
Ki
ad
a Sl
t/
Bi
/ ed
o Kn
ab
d
pp
ov
ck
/T
d be
a al
Se
er
d
tA
a Sp
t ec
th
la
ca
i ys
bj
O
r he
ph
t
O
Figure 2 Percentage of staff reporting each type of assault during the previous year. © 2014 John Wiley & Sons Ltd
1115
E.L. Kelly et al.
Table 2. In the first block of static risk factors, men and Ward staff were at greater risk of physical assault than women and Clinical or Supervisory staff. Experience in the field was not significantly associated with assault. In the second step, patient–staff conflict was positively related to assault risk (semi partial R2 = 181%) but intra-staff conflict was not. In the third block of dispositional risk factors, being less reactive to social conflict was significantly associated with more frequent assaults. In the final step, social conflict stress reactivity significantly moderated the relationship between patient-on-staff assault and intra-staff conflict (Figure 3). Specifically, staff members who reported low social conflict stress reactivity and experienced greater intra-staff conflict were at highest risk of assault. Individuals who were highly reactive to social conflict and experienced more intra-staff conflict were at low risk of assault. Reactivity to conflict was unrelated to risk of assault among participants with low levels of intra-staff conflict. A similar pattern emerged when social conflict stress reactivity was tested as a moderator of patient–staff conflict (Figure 4). Staff who had few conflictual interactions with patients were at low risk of assault, regardless of their social conflict stress reactivity. However, respondents who reported high levels of patient–staff conflict were at higher risk of assault and were most likely to be assaulted if they had low social conflict stress reactivity. Finally, staff with high levels of social conflict stress reactivity were at a slightly lower risk for assault than those with low reactivity.
respondents reported maintaining continuous observation of a patient often or very often. On average, conflict among staff members occurred occasionally (scaled mean = 252, SD 100) although 92% of participants reported some conflict with co-workers. For conflict that happened often or very often, the sources of conflict were roughly equivalent: 225% with fellow staff, 183% about treatment of patients and 208% with management. Ward staff reported significantly more intra-staff conflict (M = 261, SD = 105) than Clinical staff (M = 230, SD = 085, P = 002), yet neither group differed from supervisory staff (M = 256, SD = 067; F[2, 329] = 330, P = 004; per Games-Howell post hoc tests). Dispositional risk factors Reactivity to social conflict was normally distributed across staff members (mean = 194, SD 043). A third of participants reported moderate difficulty with letting go of conflict but only 121% reported having extreme difficulty. Across position groups, there were no significant differences in reactivity to conflict (F[2, 318] = 075, P = 047). Correlations Correlations between main study variables are presented in Table 1. The frequency of patient assaults was substantially correlated with the frequency of patient–staff conflict (r = 054) and had a lesser association with intra-staff conflict (r = 021). Intra-staff conflict and patient–staff conflict were also moderately related (r = 038). Interestingly, reactivity to social conflict was inversely related to the frequency of assault (r = 016) and had a slight association with intra-staff conflict frequency (r = 019), but was statistically unrelated to the frequency of patient–staff conflict (r = 004). This indicated that staff may not react to conflict with patients in the same manner as they would to conflict with other staff.
Discussion Workplace violence may be an unavoidable occupational hazard for psychiatric hospital staff (Anderson & West 2011). The present study sought to enhance understanding of the factors contributing to violence through a demands– stressed abilities framework. It is the first study to test stress reactivity to social conflict as a variable relevant to
Conflict, reactivity and assault Hierarchical linear regression results about the scaled frequency of assault and risk variables are presented in
Table 1 Correlations between main study variables. Variable 1. 2. 3. 4. 5. 6.
Gender Work experience in mental health Patient–staff conflict Staff conflict Reactivity to social conflict Patient assault frequency
1
2
009 011 002 012 013
3
011 009 000 006
4
038** 004 054**
019** 021**
5
016*
*P < 001; **P < 0001. Gender is coded as 0 = male, 1 = female. 1116
© 2014 John Wiley & Sons Ltd
0031 0119 0024 0100
