Although the relationship has not been studied extensively, assertion may play а significant role in the inpatient setting as aggression towards staff is frequently precipitated by conflict over demands, or after requests have been refused by staff. As part of an ongoing program of research into inpatient aggression by the Victorian Institute of Forensic Mental Health, the study described here was initiated to ascertain the extent to which various demographic, clinical, affective, and behavioural characteristics relate to aggression. (Paxton, 1997, p. 149-167)
The identification of characteristics associated with aggression may contribute to risk assessment research and assist in the prevention and treatment of aggressive behaviour. Following previous research it was predicted that psychotic symptoms, recent substance use, prior aggression, poor anger control, dysfunctional impulsivity, and limited assertion would be associated with an increased likelihood of aggression.
The evaluation and administration of aggression have become critical mechanism of psychiatric inpatient care and the centre of considerable research attention. This study analysed the associations between features of psychiatric inpatients with aggressive behaviour in а forensic psychiatric hospital. Clinical characteristics, patients’ social behaviour, and accounts of aggression and matter used were incurred for every patient admitted to the Thomas Embling Hospital, а safe and sound forensic psychiatric hospital, during 2002. Of these patients, 110 completed an extra evaluation of psychotic indications and а series of psychological tests measuring out anger expression and control, assertiveness, and impulsivity. Results suggested that а fresh history of matter use and well-established history of aggression, recent antisocial behaviour, and indications of psychosis including thought interruption, auditory hallucinations, and conceptual disorganisation contributed to aggression.
All too many research studies have relied upon the number of official reports of violent incidents made by nurses. Unfortunately, many factors determine whether nurses make an official report of an incidents, including local norms, local policies, how busy the ward is, how big is the report form, sensitivity to violence, differing perceptions of aggression, etc. Some useful work can and has been done with official reports (Rixtel, 1997, p. 111–119). However, studies which have sought to compare violent incident reports with alternative measures show repeatedly that many more incidents occur than are reported. This is in comparison with medical and nursing records other violent incident instruments (Palmstierna, 1987, p.657–663) and video cameras. All these methods, plus recent work in the East End of London demonstrate that official reports only capture а pro portion of violent incidents, that incidents which result in minor or no injury often go unreported, and that the willingness of nurses to use them varies from ward to ward. Official reports may be easy to use, as they are produced routinely by the organisation and require no extra effort on the part of the researcher. However, they do not constitute а good method of measuring objectively the rate of violent incidents.
Overt Aggression Scale (OAS)
The OAS (Spokes, 2002, p. 199–209) is the grandfather of all research tools in this area, and has either influenced or been incorporated into many of the scales that will be discussed subsequently. The method used is incident-based recording by nurses, immediately after the event, i.e. the nurse completes an OAS form after each violent incident. The scale includes verbal aggression, physical aggression, and property damage and self-harm. Each is rated four degrees of severity; for example, physical aggression can be ticked off under the following four headings: makes threatening gesture, swings at people, grabs at clothes; strikes, kicks, pushes, pulls hair (without injury to them); attacks others causing mild-moderate physical injury (bruises, sprains, welts); attacks others causing severe physical injury (broken bones, deep lacerations, internal injury). Duration, date and time, patient name, shift, and interventions used are also recorded.
Three inter-rate reliability trials are reported, with infraclass correlation coefficients of 0.7 for aggression and 0.87 for most restrictive intervention. Internal consistency (between degree of aggressive behaviour and degree of staff intervention) was poor; however, this is an argument for further research on nursing interventions, not а challenge to the validity of the scale when these two aspects are scored separately.
Scoring of the OAS does have some lack of clarity, as it appears to be different in the 1986 and 1991 publications. In the later publication the dimensions of the scale are weighted differently, with physical aggression having the highest weight and verbal aggression the lowest. It is unclear whether summing scores across these differing dimensions is to be recommended, as for instance, а score of rive makes а mild insult plus aggression against objects equal to physical aggression against а person. Not everyone would necessarily agree with this equivalency.
This scale has two major problems. First, it groups quite different forms of behaviour together as 'aggression' and, secondly, in the rating of physical aggression it conflates severity with outcome. Both these issues will be discussed in more detail below.
Modified Overt Aggression Scale (MOAS)
The MOAS (Van, 1999) basically consists of the first half of the OAS, with slightly different definitions and the addition of а zero point on each of the four dimensions. The scale is used weekly by а visiting psychologist, who consults all written records and interviews the ward staff. а scale is then completed for each patient, giving details of the severity of the most serious act of aggression committed over the past week.
Inter-rate reliability scoring of the different dimensions is similar to the 1991 OAS, with physical aggression having four times the weight of verbal aggression. Obtaining one score for а week results in the loss of а great deal of information. Using this scale nothing can be known about interventions used by nurses, duration of incidents or, indeed, number of incidents in а 1-week period for any one patient. It does not allow the study of close precursors or triggers (e.g. nurse-patient interactions prior to an incident), or of immediate consequences for nurses and patients involved. However, by using interviews of nurses plus collateral information from records it restricts the influence of nurses' differing propensities to complete individual forms on each incident.
In addition, it does allow the investigation of the relationship between aggressive behaviour and enduring ward characteristics (e.g. size, type, clientele, physical environment, etc.), or between aggressive behaviour and other patient characteristics (e.g. diagnosis). It could also be used as а dependent measure in an intervention study. Unfortunately the scale carries with it the two major deficits of the OAS, grouping diverse behaviours together as exemplars of 'aggression' and conflating outcome with severity.
Staff Observation Aggression Scale (SOAS)
The SOAS (Palmstierna & Wistedt, 1987) is an incident-based method of recording for use by nurses, who complete scales immediately after the event. It includes verbal and physical aggression, plus property damage, but not self-harm. These are not separately rated as in the OAS, but are split across three main dimensions: means, aims and results. а fixed number of precipitants and nursing containment measures used are also recorded.
Reliability of the scale was tested by asking а number of staff to rate described aggressive events. The intra-class correlation for total scores was very less. Some evidence on predictive validity is available, with а correlation between high Brief Psychiatric Rating Scale (Wistedt, 1990, p. 249–252) hostility score and serious violence as measured by SOAS within 8 days. (Pasmans, 1995, p. 46–52.)
The SOAS makes several significant advances on previous scales. By producing separate scores for means, aims and results, the capacity is developed to measure the violent behaviour of the patient separately from its consequences. Unfortunately, the individual scale items for means are rudimentary and include verbal aggression.
Similarly aims are simply construed as the person who is the object of the behaviour. Thus the SOAS also has problems in rating the severity of incidents, and has to resort to the use of result scale based upon а ranked hierarchy of injuries.
Violence Scale (VS)
The VS (Carmel, 1990, p. 558–560.) requires nurses to rate each patient for the frequency of aggressive behaviours since admission, at the time of discharge. The scale includes items relating to violence to self, others and property. Each item is Likert scaled for frequency of aggressive behaviours. Unlike the other scales, high scores represent frequency of aggression, not severity.
Internal consistency of the three sub scales (violence to self, others and property) is high. Inter-rate reliability was tested using the vignette method (as with the SOAS) and was modest. Scores from the subscales can be totalled, but this is not recommended by the author.
Use of this scale results in а similar loss of information, as the MOAS and ROAS in that it does not accurately count the number of incidents, nor their close precursors or consequences. It therefore has the same potential uses and handicaps as these scales. Its one possible advantage over them is that rating is performed only once at discharge, reducing the number of measures which need to be taken.
This conceals, however, the scale's potentially fatal flaw: the low inter-rate reliability, even working from the same case vignettes. It seems likely that this would be even lower when nurses are asked to make а retrospective assessment about а whole admission, even though it is specified that any rate must have cared for the patient on а 'daily basis'. Had the patient been transferred from ward to ward during admission, or if considerable nursing staff turnover occurred during the admission, the reliability and validity of any rating would be considerably reduced. Perhaps this scale is of greater utility in those sectors of the US psychiatric services, where admissions are of uniformly brief duration.
Retrospective Overt Aggression Scale (ROAS)
The ROAS (Wistedt, 1990, p. 125-150) is essentially the OAS made retrospective for 1 week with Likert scale measures of frequency of incidents. Ratings are made by nurses. Scoring uses the later OAS method of weighted scores multiplied by the Likert interval value. All items are then summed.
Twenty-eight ratings performed by two people gave а Pearson are correlation but it was very minor. Cronbach's alpha coefficient was 0.75 that is also very minute, indicating а good degree of internal consistency between the four OAS dimensions. However, perhaps this should be expected, as patients are seldom quietly destructive of property or physically aggressive to others, such acts usually being accompanied by verbal abuse and hostility.
The method used by this tool has similarities to the MOAS and the VS, having the same advantages and disadvantages. Yet again the deficits of the OAS in terms of grouping diverse behaviours and conflating outcome with severity have been brought forward into а new instrument.
Aggressive Incident Record Form (AIRF)
The AIRF (Wistedt, 1990, p. 301-310) is an incident-based recording form for use by nurses immediately after the event. It consists of the OAS, padded on either side by extensive antecedent information and intervention information, and attempts to cover all the items required for clinical and managerial, as well as research purposes. Much of the antecedent information requires the person completing the form to speculate about imminent and long term causes.
The theoretical source of these items is not clearly explained, neither is empirical evidence advanced that they are relevant. It includes а subjective judgement of the injury potential of an incident, and expands upon the OAS version of intervention methods used. No details of the scoring system are provided. Inter-rate reliability data are only available on 16 incidents. It appears to be good, but simple percentages of agreement only are given. Face validity of the scale was judged to be good by а panel of psychologists.
This is а lengthy scale that is highly likely to induce 'form-filling fatigue', unless it can be used to replace all other forms of official record keeping as the authors suggest. The absence of а clear scoring system is а major deficit, and inasmuch as the scale follows the OAS it has the same weaknesses.
Video camera surveillance
Using an entirely different approach, а set of researchers (Bowers L, 2002, p.106–109.) installed video equipment in the four comers of the day room of а 14-bedded psychiatric intensive care unit. Videotapes were then reviewed to identify assaults (violent incidents), defined behaviourally as attacks upon another person 'including hitting, kicking, slapping, biting, choking and throwing objects '. The degree of hostility of incidents was also rated by taking into account accompanying gestures, verbal threats and avoidance by the victim. In the second study incident severity was operationalised in behavioural terms as composed of the forcefulness of blows, the body area targeted, and staff reactions. Infraclass correlation coefficients were high for the identification of incidents, and for the classification between high and low hostility.
This method of approaching violent incidents represents а significant advance in that behavioural definitions are used, the attempt to count all types of aggression together is abandoned, and severity is no longer conflated with outcome. However the use of video cameras introduced new problems. Only incidents in the day room of the ward could be recorded. Staffs were reluctant to be videotaped and insisted that а red light would come on when recording was taking place, leading to potential observer effect. а researcher had to be present when recording was taking place, so only some time segments were sampled. Finally, the ethics of videotaping patient behaviour without their consent is contentious, and these researchers had to obtain special authority locally to conduct their studies.
Issues rose what is а violent incident?
If we wish to make progress with ward-based research in psychiatric nursing, we must develop more sophisticated methods and refine our analytical tools. The broad definition of an aggressive incident that is encompassed by the OAS and its followers is no longer useful. In fact, it is likely to be obscuring real differences in the precipitants and consequences of а wide variety of behaviours currently classed together.
Threatening Aggression
Most patients who harm themselves never pose а threat to others and, vice versa, most patients who pose а threat to others rarely attempt to harm themselves. These behaviours are likely to have different determinants, different causes, and need to be dealt with in different ways. There is little to be gained by grouping them together, and the most likely consequence of doing so is that finer-grained information will be lost. To count and rate self-harm as an equivalent to property damage, etc. is like adding apples and cabbages.
Verbal aggression is ubiquitous, and relatively harmless. It is also difficult to observe, hard to rate for its capacity for harm, differently culturally acceptable and tolerated across social classes, and difficult to define behaviourally. On occasion it can be seen as а positive way of avoiding physical aggression. It requires separate study, in detail, rather than grouping together with other diverse behaviours. Damage to property is not an area of high concern. If one had to evaluate this, it would be in terms of the financial cost. This behaviour may also be seen as а positive way of not hurting other people.
Physical Aggression
Physical aggression to others is the behaviour that really concerns us here, as it carries the capacity for permanent injury, is clearly understood by everyone as violence, and is the highest priority for nurses to prevent. Swearing and furniture damage are of little consequence in comparison to physical injury to а person. This is what out research tools should concentrate on measuring.
Having settled upon the measurement of physical aggression, it becomes apparent that the two main measures required in the first instance are the frequency of incidents and their severity. Other aspects of measurement depend on the focus of the research being undertaken --whether the interest is in interaction, explanation, causation or consequences, etc. Severity will be dealt with separately below. Measures of frequency can be obtained by retrospective judgement (e.g. ROAS) or by incident-based recording (e.g. AIRF) and the choice here again depends upon the precise focus of the research, except to say that some measure of frequency will nearly always be required.
Sexual Violence
Sexual violence is another issue of great current concern. No published scales to date have endeavoured to measure sexual harassment or assault. This can range from verbal harassment, indecent proposition, indecent exposure, to assault or rape. Of course, this is an urgent and important topic for consideration and research, but one that should be separately studied in its own right rather than submerged in the amorphous context of 'aggression'.
Violence from Playfulness or Intimidation
To arrive at а clear assessment of frequency when using an incident-based approach, it is necessary to be able to find ways to deal with counting continuous violence, and ways to separate actual violence from playfulness or intimidation. The standard method for dealing with continuous events, used by both the OAS and the SOAS, is to define arbitrarily а period of time between violence for two incidents to be considered separately. Although slightly artificial, this does give а pragmatic way of overcoming the difficulty, and ensures reliability and consistency. Playfulness and intimidation are more difficult to handle. It is hard to distinguish the playfulness of young men from real physical aggression. The two merge into one another. Is а playful kick just а playful kick, or is it an attempted assault? (Bowers L, 2002, p.80-90) in their video work showed that 62% of incidents were of low hostility or playful.
Distinguishing intimidation from actual attempted assault can be equally as difficult. Patients can show off, pretend to have been slighted, or pretend to take offence. Such behaviour can result in two patients facing off and making aggressive gestures until one backs down. Sometimes these incidents result in actual violence, at other times they pass without an outcome. The success of Brizer et al. (1988) in drawing up consistently applicable criteria to separately identify high hostility physical aggression shows that this can be done and has the capacity to be included in the definitions of and training for the use of а violent incident measure.
The concept of severity
All current scales, save those based on video camera surveillance, poorly conceptualise the severity of а violent incident and confound it with outcome. In terms of the OAS, а severe incident is one that causes severe injury. However, actual outcome reflects nursing efficacy as much as it reflects capacity to cause injury on the part of the patient.
To give an extreme example: suppose а patient who pulls out an gun and sprays bullets around the ward but who is immediately tackled by an intrepid nurse, disarmed and restrained without injury to anyone, scores at the most 8 on the MOAS physical aggression scale; however, а confused elderly man swinging his fists around, who is badly restrained by nurses who miss-time their moves, resulting in one of them being knocked out, gives а score double of that, i.e. 16.
Severity is not intention to cause injury, either. To grasp intention one would require the honest report of the perpetrator plus an assessment of how well they understood the potential for injury of what they did. Although patients have been given post incident interviews to get their perceptions, it is hard to see how this could be built into а tool to regularly measure the severity of all incidents as many patients would not be in а suitable state to answer, or consent to do so.
Severity of а violent incident is best conceptualised as the injury potential regardless of intent. The intent of the perpetrator is not the primary concern of most research -- it is their capacity to cause harm that matters. Conceptualising severity in this way means that it can be objectively and behaviourally defined, so increasing the validity and reliability of any measure. The most severe possible violent incident may thus be defined as one in which the assailant attempts to strike, repeatedly, without warning, with total commitment, speedily, to the most vulnerable areas of the victim's body, with а dangerous weapon.
Validity, reliability and standardisation of incident-based scales
These are systematic behavioural recording tools, not tests of ability or psychological characteristics. Not every type of reliability or validity is therefore relevant to an assessment of their relative merit or utility for research.
Test-retest reliability is relevant, but could only be assessed if incidents were recorded on videotape or with the use of written vignettes of violent incidents. In the real everyday life of the ward the same incident is never repeated twice. Similar problems face the assessment of inter-rate reliability, as few ward incidents are witnessed from the beginning by more than one nurse.
Artificial Testing
Artificial testing of these characteristics using incident vignettes is thus the usual recourse of researchers in this field. Reliability construed as internal consistency makes no sense when applied to an observational tool of this nature, and is not relevant to the assessment of incident based recording.
Face validity is always required so that users of the instrument will find it believable and acceptable and in the case of violent incident-recording face and content validity are identical. Similarly, concurrent and predictive validity may be felt to be of some use in justifying an incident based record to others. Concurrent validity can be tested by using other measures of hostility, official records, or by comparing the incident-based accounts with another tool using the retrospective method.
In any case, with incidents of physical aggression or face validity is so high that the necessity for concurrent validation may be questioned. Tests of predictive validity are dependent in this case upon the known relationship between past and future violence. (Lanza, 1995, p. 129-141)
To the degree that this link can be assumed, then violent incidents of particular patients at one point in time should correlate with violent incidents at а later time. Full construct validity can only be achieved by а wide-ranging set of research studies showing relationships between violent incidents and other variables. As this is the purpose for which the tool is developed, it cannot be provided in the first instance (Ray, 1998, p. 277–289) make heavy weather of the fact that no violent incident measures have been standardised. However, this is to misunderstand the reasons for standardisation and their application to violent incident measures. The purpose of standardisation is to provide normative measures against which to compare individual scores.
It is not possible to determine norms for all of humanity, as in the standardisation of IQ tests. Norms would have to be produced for each type of psychiatric ward, and probably for different countries. It is to discover such differences and the explanations for them that the instrument is constructed in the first place. It is therefore meaningless to demand their provision in advance of the research taking place.
CHAPTER 3: METHODOLOGY
Method
The methods of setting for this project were presented by Thomas Embling Hospital (TEH), the secure inpatient hospital of the Victorian Institute of Forensic Mental Health in Melbourne, Australia. The TEH provides psychiatric assessment and treatment for prisoners with а serious mental illness requiring involuntary hospital treatment, people detained as being unfit to plead (i.e., incompetent to stand trial) or not guilty because of mental impairment, offenders or alleged offenders referred by courts for psychiatric assessment and/or treatment, and people with serious mental illness referred from general mental health services who are а risk to the community and cannot be managed in а general psychiatric ward.
At the commencement of the study, the TEH encompassed an acute care program and а continuing care program. The acute care program comprised 40 beds: two 15-bed wards for acutely ill males and а 10-bed ward for acutely ill women. The continuing care program comprised 40 beds in two wards: а 20-bed extended care ward, and а 20-bed intensive psychosocial rehabilitation ward. On October 9, 2002 а third 20-bed rehabilitation ward was opened, bringing the total number of beds to 100.
Research Study
In а 24–week prospective SOAS study on а 20–bed ward (Nijman, 2002, p. 198–200.) for instance, about 10 percent of the total of 164 reported aggressive incidents had physical consequences (e.g., pain, bruises), and one of these assaults required somatic treatment of а staff member. Since the team of this ward comprised the equivalent of about 20 full-time nurses, at most, about 11 percent of nurses could have experienced such а severe incident in а year’s time. These numbers, however, were obtained on а ward in а Dutch city of about 100,000 inhabitants, and are hard to compare to а ward in London.
Nevertheless, self-report methods of aggression may lead to over-reporting. Alternatively, prospective incident-based aggression registration for research purposes still suffers from underreporting. Selection bias may also have raised the aggression frequencies reported in the current study, since staff members who had experienced more frequent or severe aggression may have been more inclined to participate.
Nevertheless, both the current and earlier results (Hunter, 1992, p. 596–598) stress that the chance psychiatric nurses will be severely injured at а certain point during their careers is real.
Procedure
During the first week of admission the patient’s case coordinator or primary nurse completed а shortened version of the Social Behaviour Schedule (Hunter, 1992, p. 556–580) The SBS covers 21 behaviour areas exhibited by patients with long-term impairments caused by psychiatric illness.
In the current study only those items derived from а classification system of four behavioural syndromes developed by were used. Using the results of а principal components analysis of SBS items, Individual characteristics and aggression 731 identified four behavioural syndromes:
- Social withdrawal
- Thought disturbance
- Antisocial behaviour and
- Depressed behaviour
Additionally, а total SBS score was obtained by summing the four subscales, representing global impairment in social behaviour.
Observations
During the first week of admission the patient’s psychiatric registrar completed а short version of the Brief Psychiatric Rating Scale (Wistedt, 1990, p. 2899–200). The BPRS is а clinician-based rating scale that provides а means to evaluate а range of psychiatric symptom constructs. The short version of the BPRS used in this study included three subscales taken from the original BPRS: conceptual disorganisation, hallucinatory behaviour, and unusual thought content. а total BPRS score was also calculated by summing each of the three subscales.
The BPRS has been used in numerous studies examining psychiatric illness and has been shown to be а reliable and valid measure of psychiatric symptoms (Wistedt). In previous studies of inpatient aggression BPRS conceptual disorganisation, unusual thought content, and hallucinatory behaviour scales differentiated aggressive from non-aggressive patients (Wistedt, 1990, p. 325-330)
Patients admitted during 2002 were approached during their first week of admission to participate in the assessment of affective and behavioural characteristics, and а further assessment of psychotic symptoms. In an attempt to ensure patients were assessed before any aggressive behaviour occurred, those patients residing in the hospital between October 1, 2001 and December 31, 2001 were assessed during this period.
Only those patients residing in the hospital on January 1, 2002, or admitted during 2002, were included in the final analysis. Participation in assessment involved completion of а semi-structured interview and а number of self-report inventories.
The use of several types of legal and illegal substances (alcohol, marijuana, cocaine, heroin, and amphetamines) was recorded by asking patients whether they had used these substances in the preceding year and during their lifetime. The patient’s history of aggression was recorded using the Violence Rating Scale (Pasmans, 1995, p. 46–52.)
The VRS has been used in studies of the violence history of other forensic psychiatric hospital patients and has been used as а screening instrument to describe the violence profile of an offender population.
The VRS is а combination of ratings of violence associated with the index offence and ratings derived from previous convictions.
Individual characteristics and aggression
When compared with the original RAS the utility of the simplified RAS was considered high. Statistical properties have been reported as acceptable with а mean inter-item correlation between the simplified and the original RAS of .79 and а total score correlation of .94 on the two tests.
Diagnosis of Aggression
Diagnosis was identified from discharge summaries that were completed by psychiatric registrars. For patients remaining in the hospital on December 31, 2002, working diagnosis was obtained through interview with the psychiatric registrar responsible for the patient’s care on that date. The most common diagnosis was schizophrenia. Disorders characterised by the presence of psychotic symptoms, including existing methods of treatment.
CHAPTER 4: ANALYSIS & RESULTS
After going through а complete process of observation using different methods and scales, approximately one in six staff members (16 percent) reported being the victim of severe physical violence at work during the last year. This severe form of aggression also turned out to be the strongest predictor of calling in sick according to Spearmans’s rating scale.
Personnel experiencing severe physical violence stayed at home an average of 3.7 days per year, whereas the mean number of days lost from work per nurse was 1.2. Although these results are hard to compare to findings from prospective incident-based aggression studies, 16 percent of staff being victimized in such а severe way seems to be rather high.
Analysis
In this study information from case notes and interviews with patients was used to score the rate of violence. а total score was obtained from these surveys was proved very helpful. Later on, those files were reviewed to record details of substance use and prior violent behaviour. Patients who consented to participation then completed а number of self-report questionnaires: the Novaco Anger Scale (Kay, 1988, p. 539–546.) the simplified version of the Rathus Assertiveness Scale, the Psychotic Symptom Rating Scales (Bowers L, 2002, p.106–109.), and the Functional and Dysfunctional Impulsivity scale.
The NAS is а self-report measure of anger arousal and control that has cognitive, behavioural, and arousal dimensions. In addition to these three domain scales, а total anger score was calculated by summing the three subscales. The NAS was developed and validated for use with mentally disordered as well as normal populations. Its statistical properties are well established.
In studies with psychiatric patients in California state hospitals the NAS was found to have an internal reliability of .95 and а test – retest reliability of .84 (Bowers, p. 115-125) Further, it was significantly related to а number of anger and aggressive behaviour criteria evaluated in concurrent, retrospective, and prospective analyses, which also included comparative measures.
Results
Aggressive incidents occurring during 2002 were recorded using an adapted version of the Overt Aggression Scale (Nijman, 2002, p. 187–195). This scale categorises aggressive behaviour into verbal aggression, physical aggression against objects, and physical aggression against self and physical aggression against other people. Within each category aggressive behaviours are arranged hierarchically according to severity. For the purpose of this study an aggressive episode was defined as the occurrence of any behaviour listed on the adapted OAS. In this version of the OAS, items relating to physical aggression against self were eliminated. The severity of an aggressive behaviour was ranked, with physical aggression towards others the most severe form of aggression and property damage the least severe. All ward staff were trained in how to record incidents of aggression and were instructed to record an incident either after it occurred or when reviewing their patients at the completion of а shift.
Where several forms of aggression occurred during one incident (e.g., verbal and physical aggression), the most severe form of aggression was rated. The name of the aggressive patient, the date aggression occurred, whether the victim of aggression was а patient or а member of staff, and the ward location were also recorded at the time of the aggressive incident or through review of patient notes and interview of ward staff by me, which occurred after aggressive behaviour. I visited each ward three times every week and provided support and encouragement to ward staff to ensure they remained mindful of the project and recorded aggressive behaviours reliably.
In this study an ‘aggressive’ patient was one who recorded at least one of the aggressive behaviours recorded by the OAS during their admission. As there may be some differences in the characteristics associated with patients who are physically aggressive compared with those who may be verbally aggressive or aggressive to property, further analyses were conducted whereby patients were classified as ‘violent’ if they had recorded at least one physically aggressive behaviour or ‘non-violent’ if they did not record any aggressive behaviours or recorded only verbally aggressive behaviours or damage to property. Experiences with Inpatient Aggression as to the psychometric properties of the POPAS, this initial test suggested rather good internal consistency.
CHAPTER 5: DISCUSSION & CONCLUSION
Discussion
From the above research and analysis, it can be found that nurses working solely with involuntarily admitted patients, in particular, seem to have а high likelihood of becoming injured. Non-physical forms of aggression (e.g. verbal threats) were experienced by most staff members, and on а regular basis. а majority of respondents also felt they had been sexually harassed or intimidated at least once during а year’s time. Female and young staff members, especially, appear to be at risk in this respect.
Furthermore, the chance of being confronted with patients’ severe self-mutilating behaviour, or with the loss of а patient due to suicide seems to be substantial when working as а psychiatric nurse. If POPAS reports are to be generalized, the mean reported number of 0.4 suicides per staff member would mean that psychiatric nurses in London are confronted with а completed suicide of one of their patients every two and а half years.
As to variables associated with aggression, the current results did not support the finding that staff members who participated in trainings on how to prevent or manage aggression experience less inpatient violence. However, it may well be that nurses already working with difficult patients (i.e., patients who display much aggression) were more likely to have followed such trainings, which would have made this comparison unfair.
The fact that various outwardly directed aggressive acts (e.g., verbal abuse, verbal threats, destruction of property, interpersonal violence), and also self-harming behaviour, appear to occur in combination in the same patients (Nijman, 2002) may play а role in this high internal consistency. High but meaningless correlations between items as а result of answering tendencies cannot, however, are ruled out.
For instance, it is possible that staff members with strong traits of neuroticism may have had better recall for, or overestimated various types of, incidents that stimulate anxiety, as neuroticism indicates а susceptibility to fear, anxiety, and hyper-vigilance (Van, 1999, p. 55-80) Alternatively, it could be argued that the high internal consistency occurs because some staff members, for whatever reason, are more frequently victims of aggression. The idea that some staff is more prone to being assaulted has been much discussed in the psychiatric literature on violence, and has received some empirical support (Palmstierna, 1987, p.157–163). Intriguingly, brings both these potential explanations of POPAS internal consistency together, showing that staff with high trait anxiety was more likely to be assaulted. (Bowers 2002) has also highlighted the importance of anxiety, demonstrating that although most staff victims report anxiety at the time of the incident or shortly afterwards, some become more cautious and fearful in the longer term. (Lion, 1981, p. 815-825)
Summary
The focus of this paper is violent incident instrumentation, or how violent incidents are defined, counted and measured for the purposes of research. Female and young staff members clearly reported more experiences of sexual harassment during their work, and sexual harassment was found to be significantly associated with calling in sick according to Spearmans’s rating scale. (Brizer, 1988, p.751-752)
Perhaps aggression management training should take such differences between staff groups into account, for instance, by teaching specific skills for dealing with sexually intimidating behaviour. Further, female nurses appeared to be less often confronted with severe violence (i.e., severe physical violence acts and completed suicides), which raises the question whether female nurses might be better at preventing such extreme aggression. Alternatively, however, they may be less often called upon when severe physical violent situations (are expected to) occur.
The validity of the POPAS assessments still needs to be established, although some of the current findings seem to contain face-validity (e.g., encountering more severe physical violence when working with involuntarily admitted patients, more sexual harassment of female staff members). The results of the current study, however, need to be regarded with caution until further validation of POPAS assessments has been conducted. Results from а POPAS survey probably cannot be regarded as assessments of the actual prevalence of discrete aggressive occurrences on psychiatric wards; multiple reports of the same incidents, particularly in cases with high emotional and/or physical impact (i.e., completed suicides) are likely to occur. Therefore, in studying the prevalence, nature, and also severity of aggression, prolonged registration periods with staff observation scales are to be preferred. Detailed assessment of severe assaults, for instance, can be conducted with the Attacks (Bowers, 2002), whereas more global registrations of aggressive incidents ranging from mild to severe may be performed with the SOAS-R (Nijman, 1999, 2002). One advantage of these scales, compared to а general survey instrument like the POPAS, is that they aim to record exact information on the kind of behaviour, and combination of behaviours, displayed by the patient. (Brizer, 1987, p. 769-770)
In practice, combinations of aggressive conducts (e.g., screaming, breaking objects, and attacking persons) during an assault are rather common, whereas, the POPAS records isolated expressions of aggression. To gain more insight into the accuracy of POPAS assessments, а cross-validation study of the POPAS with an aggression observation instrument might be useful. (Wykes, 1994)
In such а study, staff members could be asked to provide their estimates of aggression frequencies on the POPAS, directly after а one-year period of aggression registration with an incident based aggression observation scale has been completed. By including brief self-reports measures of staff members’ levels of neuroticism and anxiety, the influence of such traits on the way aggressiveness of psychiatric patients is perceived could also be obtained. In this way, more knowledge of the accuracy of retrospective staff surveys on levels of aggression on psychiatric wards may be gathered.
References
Bowers, L. & Clark, N. (1998). Incident Data from Mental Health Wards, Tower Hamlets Healthcare Trust. Unpublished.
Brizer D.A., Convit, A. & Krakowski, M. (1987). а rating scale for reporting violence on psychiatric wards. Hospital and Community Psychiatry, 38, p. 769-770.
Brizer, D.A., Crowner, M.L., Convit, A. & Volavka, M.D. (1988). Videotape recording of inpatient assaults: а pilot study. American Journal of Psychiatry, 145, 751-752.
Bech, P. 1994. Measurement by observations of aggressive behavior and activities in clinical situations. 4:290–302.
Bjrkly S. 1995. Prediction of aggression in psychiatric patients: а review of prospective prediction studies. Clinical Psychological Review 15:475–502.
Bjrkly S. 1996 Report form for aggressive episodes: preliminary report. Percept Mot Skills 83:1139–1152.
Bowers L. 1999. а critical appraisal of violent incident measures. J Ment Health 8:339–349.
Bowers L. 2002. Dangerous and Severe Personality Disorder: Response and Role of the Psychiatric Team. London: Routledge.
Bowers L, Nijman H, Palmstierna T, Crowhurst N. 2002. Issues in the measurement of violent incidents and the introduction of а new scale: the Attacks (Attempted and Actual Assault Scale). 106:106–109.
Broers E, de Lange J. 1996. Agressie in de psychiatrie [Aggression in psychiatry]. Utrecht, the Netherlands: NCGV.
Buss AH, Perry, M. 1992. The aggression questionnaire. J Pers Soc Psychol 63:452–459.
Carmel H, Hunter M. 1989. Staff Injuries from inpatient violence. Hosp Commun Psychiatry 40:41–46.
Carmel H, Hunter M. 1990. Compliance with training in managing assaultive behavior and injuries from inpatient violence. Hosp Comm Psychiatry 41: 558–560.
Crowner, M.L., Stepic, F., Peric, G. & Czobor, P. (1994). Typology of patient-patient assaults detected by video cameras. American Journal of Psychiatry, 151, 1669-1672.
Haller RM, Deluty RH. 1988. Assaults on staff by psychiatric in-patients: а critical review. Br J Psychiatry 152:174–179.
Hodgkinson P, Hillis T, Russell D. 1984. Assaults on staff in а psychiatric hospital. Nurs Times 80: 44–46.
Hunter M, Carmel H. 1992. The cost of staff injuries from inpatient violence. Hosp Community Psychiatry 43:586–588
Honigfeld, G., Gillis, R.D. & Klett, C. (1965). The Nurse's Observation Scale for Inpatient Evaluation. Journal of Clinical Psychology, 21, 65-71.
Infantino JA, Musingo S. 1985. Assault and injuries among staff with and without training in aggression control techniques. Hosp Community Psychiatry 36:1312–1314.
Kay SR, Wolkenfield F, Murril L. 1988. Profiles of aggression among psychiatric patients I. Nature and prevalence. J Nerv Ment Dis 176:539–546.
Kay, S.R., Wolkenfeld, F. & Murrill, L.M. (1988). Profiles of aggression among psychiatric patients pt 2: Co-variates and predictors. Journal of Nervous and Mental Disease, 176, 547-557.
Kay, S.R.; Wolkenfeld, F. & Murrill, L.M. (1988). Profiles of aggression among psychiatric patients. Journal of Nervous and Mental Disease, 176, 539-546
Lanza, M.L. & Kayne, H.L. (1995). Patient assault: а comparison of patient and staff perceptions. Issues in Mental Health Nursing, 16, 129-141.
Lion, J. R., Snyder, W. & Merrill, G. L. (1981). Underreporting of assaults on staff in а state hospital. Hospital and Community Psychiatry, 32, 815-825.
Morrison, E.F. (1993a). The measurement of aggression and violence in hospitalized psychiatric patients. International Journal of Nursing Studies, 30, 51-64.
Morrison, E.F. (1993b). а comparison of perceptions of aggression and violence by psychiatric nurses. International Journal of Nursing Studies, 30, 261-268.
McCrae RR, Costa PT. 1987. Validation of the fivefactor model of personality across instruments and observers. J Pers Soc Psychol 52:81–90
Nijman H. 1999. Aggressive behavior of psychiatric inpatients. Measurement, prevalence, and determinants. Maastricht, the Netherlands:University Press Maastricht.
Nijman H, a` Campo J. 2002. Situational determinants of inpatient self-harm. Suicide Life Threat Behav 32:167–175.
Nijman H, Allertz W, Merckelbach H, a´ Campo J, Ravelli D. 1997. Aggressive behavior on an acute psychiatric admissions ward. Eur J Psychiatry 11: 106–114.
Nijman H, Muris P, Merckelbach H, Palmstierna T, Wistedt B, Vos, van Rixtel A, Allertz W. 1999. The Staff Observation Aggression Scale- Revised (SOAS-R). Aggr Behav 25:197–209.
Nijman H, Evers C, Merckelbach H, Palmstierna T. 2002. Assessing aggression severity with the revised Staff Observation Aggression Scale (SOAS-R). J Nerv Ment Dis 190:198–200. Oud N. 2001. The Percerptions of Prevalence Of Aggression Scale (POPAS).
Overall, J.E. & Goreham, D.R. (1962). The Brief Psychiatric Rating Scale. Psychological Reports, 10, 799-812.
Palmstierna, T. & Wistedt, B. (1987). Staff observation aggression scale, SOAS: Presentation and evaluation. Acta Psychiatrica Scandinavica, 76, 657-663.
Palmstierna, T., Lassenius, R. & Wistedt, B. (1989). Evaluation of the Brief Psychopathological Rating Scale in relation to aggressive behaviour by acute involuntarily admitted patients. Acta Psychiatrica Scandinavica, 79, 313-316.
Paxton, R., Anslow, P., Milne, D. & Grant, B. (1997). Evaluation of а new record system for aggressive incidents in mental health services. Journal of Mental Health, 6, 149-167.
Powell, G., Caan, W. & Crowe, M. (1994). What events precede violent incidents in psychiatric hospitals? British Journal of Psychiatry, 165, 107-112.
Palmstierna T, Wistedt B. 1987. Staff observation aggression scale. Presentation and evaluation. 76:657–663.
Pasmans PW. 1995. Rogerianen en het passief-agressieve patroon [Rogerians and the passive-aggressive pattern].
Ray CL, Subich LM. 1998. Staff assaults and injuries in а psychiatric hospital as а function of three attitudinal variables. Issues Ment Health Nurs 19:277–289.
Silver, J.M. & Yudofsky, S.C. (1991). The overt aggression scale. Journal of Neuropsychiatry, 3, S22-S29.
Silver, J.M. & Yudofsky, S.C. (1987). Documentation of aggression in the assessment of the violent patient. Psychiatric Annals, 17, 375-384.
Sorgi, P., Ratey, J, Knoedler, D.U. Markert, R.J. & Reichman, M. (1991). Rating aggression in the clinical setting: а retrospective adaptation of the overt aggression scale: preliminary results. Journal of Neuropsychiatry, 3, S52-S56.
Silver JM, Yudofsky SC. 1991. The Overt Aggression Scale. J Neuropsychiatry 3:S22–S29
Spokes K, Bond K, Lowe T, Jones J, Illingworth P, Brimblecombe N, Wellman N. 2002. Hovis – The Hefordshire / Oxfordshire violent incident study. J Psychiatr Ment Health Nurs 9:199–209.
van de Werf B, Staverman K. 1999. Registratieformulier Agressie Incidenten [Report form aggressive incidents]. Bennebroek, the Netherlands: Psychiatric Hospital Vogelenzang.
Wykes, T., Whittington, R. & Sharrock, R. 1994). The assessment of aggression and potential for violence. In Barnes, T. & Nelson, H. (Eds.), The Assessment of Psychoses: а practical handbook. London: Chapman and Hall. Wistedt B, Rasmussen A, Pedersen, L, Malm U, Tra¨ skman-Bendz T, Wakelin J, Bech P. 1990. The Development of an bserver-sScale for measuring social dysfunction and aggression. Pharmacopsychiatry 23:249–252.
Yudofsky, S.C., Silver, J.M., Jackson, W., Endicott, J. & Williams, D. (1986). The overt aggression scale for the objective rating of verbal and physical aggression. American Journal of Psychiatry, 143, 35-39.
Yudofsky SC, Silver JM, Jackson W, Endicott J, Williams D. 1986. The overt aggression scale for the objective rating of verbal and physical aggression. A. J Psychiatry 143:35–39. Experiences With Inpatient Aggression 227.