Staff issues and perspectives
(Studies that were specifically identified for this review)

SourceStudy designAims of studyOutcome measuresResults
Baxter et al. (1992)

Country: Australia

Evidence level: 2-
Survey.

Setting: suburban psychiatric hospital.

Population: all nurses (425).
To elicit nurses' experiences and attitudes towards physical assault.
  • Assault rate (measured in terms of status - grade and specialism - length of employment, location – ward- gender and basic training in assault avoidance) - assessed by biographical and demographic form.
  • Attitude towards assault - assessed by specially designed attitudes to assault questionnaire.
  • Sex role beliefs - assessed by the Bem Sex role inventory.
  • 61% response rate.
  • Assault rate positively correlated with:
    -

    length of employment (p<0.001)

    -

    status (p<0.001 - attributable to much higher rate of assault amongst student nurses).

  • Attitudes towards assault
    -

    most nurses thought that gender was irrelevant to assault rate. However, males were significantly more likely than females to consider male nurses better at handling acutely assaultive service users (p<0.001).

  • Nurses concerned about:
    -

    high levels of violence

    -

    management of assaultive service users

    -

    staff gender related violence issues

    -

    assumed ability to predict violence

    -

    support/protection provided by hospital.

Reviewer's comments
  • Authors note that results amalgamate neutral response into either + or - outcomes.
  • Authors note that results may not be generalisable.
  • Questionnaire not tested for reliability.
  • Results may be country/setting specific.
  • This study has a number of quality related issues.
Carlsson et al. (2000)

Country: Sweden

Evidence level: 2+
Phenomenological study (transcribed accounts followed up by interviews).

Setting: psychiatric clinic.

Population: two nurses and three nurse assistants.
To assess the use of tacit knowledge in bringing about a positive outcome in encounters with aggressive and violent service users.Themes emerging from transcripts relating to the study aims.Seven themes emerged: respecting one's fear, respecting the client, touch, dialogue, situated knowledge, stability, mutual regard and pliability.
Reviewer's comments
  • Authors argue that inter-subjective knowledge is a crucial aspect of caring practice and warrants further attention.
  • This study is of reasonable quality, however further interviews are required to build up a picture of how these themes compare to those put forward by other individuals in similar and different in-patient settings.
Crichton et al. (1998)
Country: Canada and UK

Evidence level: 2-
Cohort study with historic control.

Setting: psychiatric in-patient settings (various).

Population: 132 Canadian nurses, controls = 32 British nurses (all grades) (used sample from earlier pilot study - Crichton JHM, 1997 as control).
To compare the attitudes of Canadian and British nursing staff towards the management of service user misdemeanours.
  • Responses to video vignettes (measured by semi-structured questionnaires).
  • Demographic variables.
  • No statistical differences were found for demographic variables within the Canadian sample or between the Canadian sample and the UK sample.
  • The following results were significantly different. More Canadian nurses opted for PRN medication if the service user would accept it (p<0.05). More UK nurses would talk to the service user about what had happened (p<0.01). More Canadian nurses would recommend a short time in seclusion (p<0.01). More UK nurses would use de-escalation techniques (P<0.05). More Canadian nurses would use PRN medication against the service user's wishes (p<0.05). More Canadian nurses would request police involvement (p<0.05).
  • The following results related to the Canadian study only.
    -

    Where nurses attributed an incident to service user choice, the following responses were deemed most helpful: police involvement (p<0.001), telling the service user their action was wrong and unacceptable (p<0.05), the use of sanctions (p<0.05).

    -

    Those who rated some degree of personal threat rated the following as helpful: seclusion (p<0.05), sanctions (p<0.01) and police involvement (p<0.05).

    -

    Those who rated any degree of threat to the ward rated mechanical restraints as helpful (p<0.05). The authors note that the Canadian nurses tended to opt for more restrictive measures to control incidents than UK nurses.

Reviewer's comments
  • The authors note that the response to vignettes may not compare fully to real life responses.
  • The authors note that minor differences in terminology usage may have been unavoidable and so may have skewed the results slightly.
  • This study has quality related concerns, since the control group is a selected group of nurses from a larger study. The manner in which the nurses were selected from the previous study of UK nurses (n=192) is not specified. This could leave to a serious bias that jeopardises any comparison of attitudes by UK and Canadian nurses.
Crichton (1997)

Country: UK

Evidence level: 2+
Single sample cohort study.

Setting: psychiatric in-patient (various).

Population: 192 nurses (all grades).
To assess the attitudes of nurses towards the management of a service user misdemeanour.
  • Level of threat (personal/general).
  • Whether staff reported different strategies for different diagnoses.
(All assessed by semi-structured questionnaire responses to video vignettes).
  • When the service user was diagnosed with schizophrenia, the following were rated as more helpful by staff: voluntary medication (p<0.001), and non-voluntary medication (p<0.001).
  • When the service user was diagnosed with personality disorder, the following were rated as more helpful: sanction (p=0.003), telling the service user their action was wrong and unacceptable (p=0.036), encouraging the service user to apologise (p=0.011), police investigation (p=0.034).
  • When the service user was female, had personality disorder and was involved in a violent incident, sanction was rated as most helpful (p=0.034).
  • Seclusion was rated more helpful if the service user was male (p=0.047).
  • When the service user had a history of violence, seclusion (p=0.027) and sanctions (p=0.047) were rated as more helpful.
  • There were no significant differences according to the race of the service user
  • Those who felt personal threat, found the following more helpful: seclusion (p<0.001), voluntary medication (p<0.001), compulsory medication (p=0.018), sanctions (p=0.018), telling off (p=0.002), police involvement (p<0.001). Those who rated any degree of threat, found the following more helpful: seclusion (p=0.02), sanction (p<0.001), telling off (p<0.001) and encouragement of an apology (P<0.001). However, the choice of sanction depended on the nature of the incident. Where the incident involved punching rather than the use a firearm, the following were rated as more helpful: seclusion (p=0.001), non-voluntary medication (p=0.048) and encouraging an apology (p=0.023).
  • Patients with personality disorder (p<0.001) and those with a history of violence (p=0.008) were said to have acted out of choice rather than lack of self-control. Only those service users with a history of violence and schizophrenia were said to have acted out of lack of self-control (p=0.032).
  • When untrained staff were compared with trained staff, every response - except de-escalation techniques and police involvement - were rated as significantly more helpful.
Reviewer's comments
  • Author notes that vignettes may not fully correspond to real life.
  • This study was of reasonable quality but the application of Chi squared analysis to content analysis appears questionable.
Cutcliffe (1999)

Country: UK

Evidence level: 2+
Hermeneutic study.

Setting: unit with high incidence of violence in psychiatric hospital.

Population: six qualified nurses (all from unit).
To examine the experiences of nurses who experience violence from individuals suffering from long-term mental health problems.Themes emerging from participant interviews.
  • Three overlapping key themes, describing the lived experience of violence, emerged: personal construct of violence; feeling equipped; and feeling supported.
  • The importance of training and debriefing were stressed.
Reviewer's comments
  • Author argues that there is an important relationship between a nurse's ability to deal with an incident therapeutically and feeling supported. This should be considered in strategic plans.
  • Study notes how different perceptions of an individual's behaviour as violent or not violent (based on their therapeutic relationship with the client) influence nurse decisions to report an incident as violent.
  • Author notes that the findings of the study are not generalisable.
  • This study is of reasonable quality, however further interviews are required to build up a picture of how these theme compare to those put forward by other individuals in similar and different in-patient settings.
Delaney et al. (2001)

Country: Australia

Evidence level: 2-
Multi-method approach using survey and focus groups and analysis of incident forms:
  • survey - n=95 registered nurses
  • focus groups- nursing staff (number of focus groups or staff involved not stated)
  • incident forms n=60.
Setting: four acute in-patient facilities.

Population: psychiatric nurses.
To study the clinical management of service users identified as potentially aggressive in psychiatric in-patient settings.Nursing management strategies.
  • 59/95 (62%) surveys returned. - 88% respondents indicated that they had been assaulted. 53% had participated in training in the past 12 months.
  • Six risk factors for violence emerged from the focus groups - history, status and arrival mode; ongoing informal nursing assessment; individualised care; peer support and administrative responsiveness; aggression and nursing stress and policies/manuals.
  • 75% of service users instigating aggressive behaviour had been identified as at risk on admission. Progress noted documented this risk for 73%.
Reviewer's comments
  • There are several quality related problems with this study: the number of focus groups and the number of participants overall or in each is not stated. The method of analysis used for the various studies is not stated. It is not clear whether the studies were undertaken contemporaneously.
Duxbury (1999)

Country: UK

Evidence level: 2-
Qualitative.

Setting: acute inpatient mental health settings and general settings (medical and surgical).

Population: 34 qualified mental health nurses and 32 qualified general nurses.
To compare nurse experiences of violence in the two settings.

To see if similarities/ differences exists between experiences in these settings that can be used as learning points.
  • Critical incident technique. (Critical incident involves providing participants with a blank sheet of paper with a written prompt -i.e. a sentence asking them to expand on an assault incident).
  • Reliability/validity of approach tested by an independent researcher
  • Indicated findings:
    -

    nurses attributed violence to internal factors

    -

    emphasis on biomedical approach

    -

    verbal and physical abuse most common

    -

    Lack of reference to de-escalation

    -

    MH nurses controlled events themselves, general nurses sought external support.

Reviewer's comments
  • Author states sampling and size appropriate to study design.
  • Author comments that lack of prompting for reference to de-escalation may reflect study design weakness.
  • The methodological approach adopted by this study makes it a weak study in terms of quality. These concerns are not, however, overwhelming.
Duxbury (2000)

Country: UK

Evidence level: 2-
Pluralistic evaluation design (triangulation methodology).

Setting: three acute wards (one PICU, one high dependency, one open acute ward).

Population: service users (n=80) nurses (n=72) medical staff (n=10) (qualified and students).
To record the nature and management of aggressive and violent incidents.

To survey staff and service users' views about causes of violence and its management.

To compare views of staff with views of service users.

To determine the impact of internal, external and situational models on views and practice.
Aim 1 - adapted incident forms - modified staff observation aggression scale (MSOAS).

Aim 2, 3 and 4 - the management of aggression and violence attitude scale (MAVAS).

Semi-structured group interviews (used to corroborate finding on attitude scale - not incorporated into results).
  • 221 violent incidents recorded.
  • Statistic analysis of MSOAS and MAVAS were conducted using descriptive and inferential statistics.
  • 70% of incidents involved verbal threats or abuse only. Physical violence accounted for 13.5% of incidents.
  • Staff most commonly reported problematic interaction and restrictive environments as causes of violence/aggression. However, staff did not view their own interactions with service users as problematic when surveyed. 26% of all incidents were not attributed to a specific cause.
  • Strong correlation between type of aggression reported and intervention used (p<0.000). Most common strategies used in isolation were de-escalation (22%) and medication (25%). However, range of techniques, involving restraint, medication and seclusion, was reported for 47% of all incidents.
  • Service users and staff had a number of opposing views about the causation and management of violence/aggression.
  • Service users believed that external and situational factors (such as staff interaction and restrictive regimes) were largely to blame p<0.001). Reactive aggressive incidents resulting in assault =42.9%.
  • Staff emphasised internal factors (i.e. service user illness).
  • Staff want practices such as seclusion to continue; service users, in many instances, do not (p<0.000).
  • Service users are not aware of the use of de-escalation techniques (p<0.000).
Reviewer's comments
  • Author notes that neither staff nor service users were happy with the current approach (which used standard interventions such as seclusion and restraint that the author argues can be attributed to a philosophy based on control). Both parties felt that violent incidents could be managed better.
  • There are methodological weaknesses with this study. In particularly, only selected results are taken from staff and service user interviews. Most results are taken from the incident forms. Whilst these quality concerns are not overwhelming, they threaten the findings, since it is not clear whether what was not reported challenged the findings of the incident reports.
Lowe et al. (2003)

Country: UK

Evidence level: 2+
Questionnaire survey.

Setting: acute psychiatric admissions unit.

Population: 100 psychiatric nurses (multiple grades).
To examine nurse perceptions about conflict situations, (for example, importance of and details of each event); and whether there are differences between individual nurses and grades of nurses (including nursing assistants).10 scenarios were rated using 10 response statements. Each statement represented a specific type of intervention, based on previously identified theme. A 20-point scale was used to register agreement between nurses and nursing assistants.
  • Response rate= 72%.
  • Three main themes emerged - support/control, communication, face-saving/personal control. Support/control was considerably more important than the other two.
  • Nurses with higher grades tended to use less limit setting interventions.
  • Authors conclude that training and education may underlie this difference.
  • Authors posit that since limit setting has been associated with triggering conflict situations, nurse training may reduce in-patient aggression.
Reviewer's comments
  • The authors noted that one weakness with the study design is that clinical experience was not measured. Grade was used as a proxy.
  • Authors also note that higher-grade nurses tend to have less contact with service users, but do not consider this a design weakness.
  • Authors recognised that further research is needed to assess nurses' attitudes and limit setting.
  • The study was of a reasonable quality.
Morrison (1993)

Country: US

Evidence level: 2-
Controlled cohort study.

Setting: two inpatient psychiatric settings (short-term admission and evaluation units).

Population: 69 nurses - 34 psychiatric, 35 doctoral student nurses (as control).
To assess nurses perceptions of serious violent behaviour and whether perceptions vary between nurses with or without extensive psychiatric experience.Author notes that the design meant that no outcome measures were used.
Demographic details were collected
The study consisted of rating pairs of items in terms of seriousness using the violence scale.
  • There were no significant differences in terms of demographic details between nurses and those enrolled on doctoral programme
  • There was a high degree of agreement between psychiatric nurses and doctoral students over which types of violence were most serious. Only small difference were noted.
  • However, psychiatric nurses themselves displayed disagreement about the seriousness of types of violence.
  • In rating types of violence as serious, differences were noted with students rating violence to others and violence to self as equally serious, and verbal violence to others as standard. Psychiatric nurses rated violence to property as standard and rated violence to others and verbal violence to others as similar. Again there were discrepancies within the groups of nurses.
Reviewer's comments
  • Author notes that the study design made it impossible to determine the cause of any differences.
  • Author notes that the results may not be generalisable to other psychiatric professional groups.
  • Author argues that the study has implications in relation to how nurses predict violence. It is not clear how the findings of this study can be transferred to the clinical environment.
  • The decision to compare psychiatric nurses with doctoral students is not explained and the decision seems odd. It is possible that the aim is to test the effect of education and lack of psychiatric experience on ratings of seriousness for violent events, but the manner in which these two variables interact and confound one another is not explored. The sample size seems too small for patterns to emerge.
Morrison (1998)

Country: US

Evidence level: 2-
Single sample cohort (causal modelling).

Setting: three hospitals (state, federal, private).

Population: 162 Staff.
To investigate whether there is a correlation between relationship, personal growth and factors which maintain the current hospital approach/system (system maintenance) - for example, authoritarianism and social restrictiveness.The following antecedent variables were assessed:
  • Relationship (extent to which employees were concerned about their jobs, supportive and supported)
  • Growth (extent to which employees encouraged to be self-sufficient and efficient, also time constraints)
  • System maintenance (extent to which employees know what to expect and to which rules were communicated. Way that rules were used to keep employees under control)
The following beliefs were assessed:
  • authoritarianism (people with mental health problems are inferior and require coercive handling)
  • social restrictiveness (mentally unwell are a threat to society and functioning must therefore be restricted during hospitalisation).
Measured by:
  • the work environment scale (WES) - used to assess relationship, growth and system maintenance
  • opinions about mental illness scale (OMI) - used to assess authoritarianism
  • violence scale (VS) - used to assess violent incidents
  • social desirability scale (SDS) - used to assess social restrictiveness
  • demographic data collection instrument.
  • Most staff rated themselves as able to form good relationships with service users (56.8%). Most rated themselves better then other staff in terms of performance (65%).
  • Regression analysis was used to determine the relationship between authoritarianism and social restriction and aggression (the latter is the dependent variable).
  • Author notes that although authoritarian ideology is not predictive of aggression in either model, aspects of control are predictive in both models. Satisfaction with the hospital emerged as the most predictive of all the model variables
Reviewer's comments
  • Several factors were not included within the analysis - type of hospital, or level and education of participating nursing staff- which is a weakness of the study, since they may not be adjusted for.
  • The author notes a number of weaknesses with the study design, such as an unquantified relationship between attitudes and behaviour; only marginal consistency in three of the scales; and a lack of clarity as to whether the OMI is the appropriate tool to assess control in nurses. This last query seems to result from the fact that the model is not entirely upheld by the study.
  • The study is theoretical.
  • It is not clear that the model adds anything to the study beyond that which is achieved by assessing the outcome measures and then correlating results. This approach makes the results confusing, and although the quality issues are not overwhelming, it is not clear how much this study adds to our understanding of the causes of violent incidents.
Nolan et al. (2001)

Country: Sweden and UK

Evidence level: 2-
Questionnaire survey.

Setting: eight districts in Sweden, three mental health trusts in UK.

Population: 720/1,090 Swedish nurses, 296/661 UK nurses.
To compare the extent and nature of violence encountered by Swedish and UK mental health nurses.
  • Number of times exposed to violence.
  • Type of incident.
  • Severity of incident.
  • Self-esteem.
  • Satisfaction with working environment.
  • UK response rate (45%), Swedish response rate (68%).
  • Over preceding 12 months, 71% of UK nurses compared to 59% of Swedish nurses had been exposed to violence. More UK nurses had been exposed very often (P<0.01). UK nurses reported a higher level of violence involving service user's relatives (p<0.05).
  • Significantly more UK nurses had been victims of aggressive behaviour, i.e. spitting (p<0.01) and had experienced violence involving a weapon (p<0.05).
  • More UK nurses described receiving minor injuries.
  • More support was available to Swedish nurses following an incident (p<0.01).
  • UK nurses were more likely to report low self-esteem (p<0.05).
  • UK nurses who had experienced violence in the preceding 12 months were more likely to find their job always psychologically taxing (P<0.05).
  • Swedish nurses were more likely to say that they were satisfied with their work duties (p<0.05).
  • A negative correlation was found between high self-esteem and positive attitudes to work (p<0.01).
  • A positive correlation was found between self-esteem and influence over work duties (p<0.17).
  • A positive correlation was found between self-esteem and feedback from line managers (p<0.05).
Reviewer's comments
  • 30% of the UK nurses worked in the community; all Swedish nurses worked in in-patient departments.
  • Authors note that the study raises questions about the role of nurses in the UK, public perception of their role and need to address training issues.
  • This study was of reasonable quality.
Omérov et al. (2001)

Country: Sweden

Evidence level: 2-
Cross sectional study (over three years -18 months of study period in total).

Setting: two 16-bedded psychosis wards.

Population: nursing staff (137 interviews).
To ascertain number and type of violent incidents over 18 months and to assess staff reactions to assault.
  • Staff observation aggression scale (SOAS) used to assess number and type of incident.
  • Semi-structured interviews using questionnaire to assess staff responses to violence.
  • Demographic details of service users and staff (method of collection not specified).
  • 137 incidents. There was a gradual decrease in violent incidents during the study.
  • Violent female service users tended to be older than their male counterparts (p<0.001).
  • 30% of violent incidents were regarded as unprovoked. 47% of incidents were associated with limit setting, or instruction giving re medication.
  • 81% of the violence was directed towards staff
  • 53% of the violence was committed by men; 47% by women.
  • Most violent acts were directed against women (93% of female attacks were against women, and 67% of male attacks - p<0.001).
  • 17% of violent incidents were between service users.
  • Most staff felt insulted by the assault (43%); one-third felt angry. More men felt frightened (p<0.05); more women felt surprised (p<0.01). Most staff felt very uncomfortable after the assault, brought the incident home with them, found it hard to relax, nightmares were frequent, and returning to work could be difficult.
  • All except one would have liked training in self-defence and update courses.
  • Majority wanted training in psychopathology.
  • Wanted debriefing (saw the interviews in this light).
Reviewer's comments
  • Authors note that one change was the introduction of zuclopethixol acetate to control psychotic service users in the course of the three years evaluated, which could act as a to confounder.
  • The way that the questionnaires were used changed over the course of the study.
  • Authors argue that violence between service users is likely to be more prevalent than the study indicates, due to under-reporting.
  • This study was of reasonable quality.
Ray & Subich (1998)

Country: US

Evidence level: 2-
Survey.

Setting: psychiatric hospital.

Population: 150 staff members.
To assess whether there is a relationship between staff attitudes and service user assault.
  • External locus of control -assessed by internal-external locus of control scale (I-E).
  • Anxiety - assessed by state-trait anxiety inventory (STAI).
  • Authoritarianism - assessed by right wing authoritarianism scale (RWA).
  • Number of years working in hospital.
  • Number of injuries over previous year and throughout employment at the hospital.
  • Number of assaults over previous year and throughout employment at the hospital.
  • Demographic information.
  • Response rate 78/150.
  • Results were analysed using MANCOVA and regression analysis.
  • A significant relationship between: STAI score and injury during the last year was noted (p=0.01), external locus of control and ever being injured (p=0.05), STAI score and ever being injured (p=0.02), external locus of control and ever being injured (p=0.02), STAI scores, injury in the last year and level of trait anxiety (p+0.01).
  • Prediction of the mean number of assaults was significantly related to trait anxiety, locus of control and authoritarianism.
  • A lower RWA score was associated with more annual assaults.
Reviewer's comments
  • The authors conclude that the results indicate that external locus of control, anxiety and authoritarianism are connected with service user assault and injury.
  • The authors noted the following limitations: relatively small sample size, that the operational definition of assault and injury may have been too vague for the expected MANCOVA differences to emerge, and that accurate results in the regression analysis may have been affected by lack of recall of assault by the participants.
  • This study was of reasonable quality.
Spokes et al. (2002)
Country: UK

Evidence level: 2+
Single sample cohort (semi-structured interviews).

Setting: 13 psychiatric inpatient units across five sites (10 adult acute admission wards, two PICUs and one low secure unit).

Population: 350 nursing staff (qualified and unqualified).
To obtain the views of a sample of mental health nurses about staff-related factors that they perceive contribute to or reduce in-patient violence (physical and psychological).
  • Standard content analysis (QSR NUD*IST software package (V 4.0, QSR, 1997).
  • SPSS was used to assess demographic data.
  • 108/350 staff took part (63% qualified, 40% unqualified care assistants). (108/350 = 29%).
  • 27 characteristics were identified as having a positive influence on the potential outcome of an incident
  • 19 characteristics were identified as having a negative influence on the potential outcome of an incident.
  • Three main themes were identified - clinical skills, personal characteristics, interpersonal skills - which nurses believe have the potential to both positively or negatively affect in-patient violence.
Reviewer's comments.
  • The authors noted that there is a possibility that nurses' responses were not entirely honest. However, they argue that the level of responsibility that nurses assumed for the occurrence of violent incidents suggests most nurses were honest.
  • Authors also note that higher-grade nurses tend to have less contact with service users, but do not consider this a design weakness.
  • Authors suggest that training programmes should be developed that focus not only on clinical skills, but also on interpersonal skills and characteristics.
  • This study was of reasonable quality
Wykes & Whittington (1998)

Country: UK

Evidence level: 2-
Case control.

Setting: six acute psychiatric wards in a major teaching hospital.

Population: 39 psychiatric nurses. 34 concurrently selected psychiatric nurse acted as control matched by grade and sex and age.26 nurses from a PICU were used to establish baseline scores.
To investigate the effects of workplace violence on psychiatric nurses.
  • Trauma related stress -assessed using impact of events scale (IES) and PTSD symptom scale (PTSDSS).
  • General distress - assessed by state-trait anxiety inventory (STAI), state-trait anger expression inventory (STAXI), the general health questionnaire (GHQ-28) and the Beck depression inventory (BDI).
  • Physical injury (coded present or absent).
  • Concurrent stressors -assessed by daily hassles questionnaire (DHQ) and 118-item schedule.Assault = physical contact.
  • Baseline was established by selecting an initial group of nurses (n=26) from a PICU who had not been assaulted by a service user at work in the preceding months (time 0).
  • Assaulted group (n=39) were assessed twice (within 10 days of assault (time 1) and one month after (time 2). (Some were also assessed at baseline which enabled pre and post changes to be measured n=10).
  • A control group (n=34) were recruited concurrently with assaulted group (not assaulted in previous six months) and matched as far as possible to assault group.
  • PTSD was noted (PTSDSS). Two nurses had PTSD at time 1; two different nurses had PRSD at time 2.
  • The mean scores on all measures for general distress were below those that would indicate severe psychological effects in assaulted nurses.
  • Where physical injury was sustained, PTSDSS and IES were significantly higher at time 1 (p<0.05). Intrusion and avoidance symptoms were also significantly higher (p<0.05) at time 1 (symptoms of trauma related stress and general distress).
  • There was no significant relationship between concurrent stressors and either trauma related stress or general distress.
  • Victims who experienced another assault between time 1 and time 2 were significantly more likely to experience very high or very low symptoms of stress at time 2 (p<0.005).
  • For PTSDSSs and GHQ, arousal measures at time 1 were significantly related to increased avoidance at time 2 (PTSDSS p<0.001, GHQ p<0.003).
  • Significant relationships were found between avoidance at time 1 and stress symptoms at time 2 on PTSDSS (p<0.03) and IES (p<0.004).
  • Six of the 17 participants who reported avoidance (scored at lease 1 avoidance item on PTSDSS) at time 1 had either the sub-syndrome or PTSD at time 2 (p=0.008).
  • Authors note that no staff member took time off work. They argue that the high-risk environment in which staff work means that they are less likely to make attributions of self-blame.
Reviewer's comments
  • There are several quality issues raised by this study. The most serious is the inclusion of results from the baseline group where these nurses were later assaulted. This group is an uncontrolled cohort, which is not related to the main study participants. The results are reported in a confusing way, making it difficult to disentangle which group is being referred to. The sample size is also small.

From: Appendix 5, Evidence tables – included studies

Cover of Violence
Violence: The Short-Term Management of Disturbed/Violent Behaviour in In-Patient Psychiatric Settings and Emergency Departments.
NICE Clinical Guidelines, No. 25.
National Collaborating Centre for Nursing and Supportive Care (UK).
Copyright © 2006, Royal College of Nursing.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means electronic,mechanical, photocopying, recording or otherwise,without prior permission of the Publishers or a licence permitting restricted copying issued by the Copyright Licensing Agency, 90 Tottenham Court Road, London W1T 4LP. This publication may not be lent, resold, hired out or otherwise disposed of by ways of trade in any form of binding or cover other than that in which it is published,without the prior consent of the Publishers.

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.