Paper
Didier Camus, Mehdi Gholam-Rezaee, Nicolas Kühne, Jean-Philippe Duflon, Jean-Michel Kaision, Valérie Moulin.
CHUV, Psychiatric Department, Vevey, Switzerland
Keywords: Psychiatric hospitals, Violence, Aggression, Assessment, SOAS-R
Introduction and/or background
Since 2008, the psychiatric department of the Lausanne University Hospital (DP-CHUV), Switzerland, has employed the scale “Staff Observation Aggressive Scale- Revised” (SOAS-R) to assess aggressive and violent incidents in their in-patient units. A large group of patients with different psychiatric disorders and age groups (children/adolescents, adults, elderlies) were concerned by this assessment.
Method
The aim of this study is to describe the results of the assessment (nature and prevalence of aggressive behavior towards staff or others targets, patients’ and victims’ characteristics, differences among victims’
from different services,…).
• The study was performed in the DP-CHUV in seven specific acute hospitals located in the Canton de Vaud:
• Two adult psychiatric services (PGE: 7 wards, 95 beds; PRANGINS: 5 wards, 74 beds).
• Two psychogeriatric services (SUPAA: 5 wards, 80 beds; GIMEL: 1 ward, 20 beds).
• The child and adolescent psychiatric service (SUPEA – 1 ward – 10 beds).
• The toxidependency service (PCO – 1 ward – 8 beds).
• A “mixed” (adult and psychogeriatric) service (CPNVD, 4 wards, 42 adult beds + 14 psychogeriatric beds).
The SOAS-R is a standardized scale allowing classification of aggressive behaviors (physical assaults;
verbal assaults and assaults against property). This scale consists of five columns (1) provocation; 2) means used; 3) target of the incident; 4) consequence(s) for the victim; 5) measure(s) used to stop the aggression), pertaining to describe the contexts of the violent or aggressive acts occurring in the wards involved (Nijman, 1999; Nijman, 2002). The severity of the incidents is measured using the SOAS-R-scoring system (0-22 points) and the SOAS-gravity score (0-100 with a VAS) filled in by the victim. The SOAS is reported after each violence incident.
The assessment of violent incidents in DP-CHUV using the SOAS-R shows two specificities compared to other institutions using the same scale:
• Considering that the problem of violence concerned and implied all the people working in our institutions, all personnel working in these hospitals (nursing staff, medical staff, administrative staff, security staff, students, etc…) have been consequently advised to the possibility of using this tool in case of observing any violent act.
• To limit what is called “the reporting industry” (De Niet, 2005), in case of verbal violence, the victims should fill the SOAS only if they feel threatened.
More information on patients were collected (age, gender, type of admission, principal and secondary psychiatric diagnoses according to ICD-10) using the hospitals data bases. Data were analysed using the Statistical Package for Social Sciences (SPSS) and R environment for statistical analysis.
Results
General characteristics
The analysed data set contains 11902 hospital stays with a total of 361045 treatment days. A total of 1239 incidents (1316 declarations) were registered during the study period. Aggressive incidents were recorded from 500 patients (7.5%) of the 6623 patients hospitalised during this time. Among 500 violent patients, 216 of them (43.2%) were responsible for more than one incident, while 45 patients (9%) were responsible
for more than 5 incidents. 40 violent episodes (3.23%) were caused by external persons (family, relatives, friends). The global incidence rate per 100 treatment days for the DP-CHUV was 0.34. This rate varies among different services, from 0.54 (PGE) to 0.13 (SUPAA).
Table 1: General data
PGE 95 834 5660 155550 2790 305 8.0 10.9 0.54 07.2008 - 12.2012
SUPAA 80 137 1846 104880 1442 65 4.0 4.5 0.13 01.2009 - 12.2012
SUPEA 10 59 449 13740 299 27 8.0 9.0 0.43 01.2009 - 12.2012
PCO 10 50 657 10223 409 31 5.0 7.6 0.49 01.2009 - 12.2012
PRANGINS 74 62 1819 40790 1281 29 1.0 2.3 0.15 06.2011 - 12.2012
GIMEL 20 35 210 11823 176 13 3.0 7.4 0.30 05.2011 - 12.2012
CPNVD 56 62 1261 24039 891 30 1.0 3.4 0.26 11.2011 - 12.2012
DP-CHUV 345 1239 11902 361045 6623* 500 //////////// 7.5 0.34 ////////////////// ////////////
* During an hospital stay, a patient may be transferred into another service
Verbal aggressions, combined or not with other types of violence (physical or object aggressions) represented 66.5% (n = 875) of the situations. On the other hand physical aggression was present in 53.2%
(n = 687) of violent events combined or not with other types of violence. Most violent incidents occured during the daytime, more frequently in the afternoon (2 pm – 8 pm; n = 491; 39.5%) than in the morning (7 am – 1 pm; n = 342; 27.6%) or the night (22 pm – 7 am; n = 273; 22.1%). Distribution was uniform throughout the days of the week. A quarter of incidents (n = 326; 27.2%) occurred within the first week but 44.2% of the incidents (n = 530) occurred from the second month of hospitalization. The corridors of unit (n = 342, 28 % of the incidents), the seclusion room (n = 193, 15.6% for the DP-CHUV but 19.6% (n = 188 incidents) for the adult services) and the patient’s room (n= 221, 17.8% for the DP-CHUV but 51.8%
for the psychogeriatric services) were the places where incidents occurred more frequently. Hospital staff (mostly nursing and medical staff) were most frequently the targets of the aggressions (n = 1014; 77.1%).
Nurses were the principal targets of the aggressions. Almost a quarter of the declarations showed that the patients could also be the target (Total: 22.8%). In the majority of the situations, the victims expressed their feeling as being threatened in the situation (67.4%). More than a quarter of aggressive situations (25.3%) had caused physical pain but, fortunately, only 40 victims (3 %) needed a treatment by a physician.
Regarding the type of intervention, the staff controlled the situation more often by talking with the patients (52.3%), by administering medication (35.3%) or by using seclusion room (29.8%).
Patients’ characteristics
Regarding patients characteristics, the results (table2) showed some differences between SOAS and non-SOAS patients.
Table 2: Differences between SOAS and non SOAS patients
Variable Non-SOAS SOAS Test Statistic P-value
Age (mean) 48.53 42.64 t-test 5.96 <0.0001
Gender (Man) [#(%)] 2815 (90.37) 300 (9.63)
Chi-square 63.12 <0.0001
Gender (Woman) [#(%)] 3346 (95.38) 162 (4.62)
Involuntary admission 4621 (91.31) 440 (8.69)
Voluntary admission 6643 (97.11) 198 (2.89)
Average lenght of stay 30.14 56.53 t-test -8.68 <0.0001
Schizophrenia, schizotypal, delusional disorders - Mental
disorders due to psychoactive substance use 70.12 29.88
Chi-square Schizophrenia, schizotypal, delusional disorders - Others 90.97 9.03
Mood (affective) disorders - Mental disorders due to
psychoactive substance use 89.92 10.08
Chi-square 0.0066
Mood (affective) disorders - Others 95.79 4.21
Violent patients (SOAS patients) differed from non-violent subjects (Non-SOAS patients) in being younger, were more often men and more often hospitalised without their consent (involuntary admission).
According to the ICD 10, a higher risk was also found 1) in patients with a principal diagnosis F2 (Schizophrenia, schizotypal, delusional disorders) associed with secondary diagnosis F1 (Mental disorders due to psychoactive substance use), and 2) with a principal diagnosis F3 (Mood (affective) disorders) associed with secondary diagnosis F1.
SOAS scores and gravity scores
The severity of all declarations on the SOAS scale ranged from 1 to 21, with a mean severity to 13.05 and a median to 13.00. SOAS scores vary according to the type of violence. When violence is physical, associated or not with another form of violence, SOAS scores were higher. SOAS scores differed too according to the services (figure 1): PCO (mean = 10.7 ± 0.53; median = 11.0) present the lowest scores. On the contrary, GIMEL (mean = 15.3 ± 0.55; median = 16.0), the SUPAA (mean = 14.9 ± 0.31; median = 15.5) and the SUPEA (mean = 14.5 ± 0.48; median = 15.0) present highest scores.
n= 65
Figure 1: Boxplots of SOAS-R Figure 2: Boxplots of gravity scores according services
Gravity scores range from 1 to 100 points on the VAS. More than half of the declarations (54.8%) present gravity scores > 50 points and 30.5% present gravity scores > 75. According services, results present some differences (figure 2). The mean and the median of the gravity scores may be increased according to certain elements present or not in the declaration (table 3).
Table 3: Means ans median og gravity score versus different elements
Name Mean 0* Mean 1** Median 0* Median 1** Obs 0* Obs 1** P-value
Provocation not understandable 53.1 58.2 52 56.5 1040 274 0.0055
Means used: Hand (hitting, punching, etc...) 53.4 56.2 52 58 955 359 0.0021
Dangerous method: Knife 53.6 76.8 53 86 1282 32 0.0000
Target: Patient self 53.1 66.7 52 71 1209 105 0.0000
Target: Other persons 53.7 64.3 53 68 1258 56 0.0070
Consequences: No 56.5 38.7 56 37 1145 169 0.0000
Consequences:Felt threatened 47.3 57.5 48 58 427 887 0.0000
Consequences:Pain > 10 mn 52.8 65 51 69.5 1160 154 0.0000
Consequences: Visible injury 53.3 60.5 52 65 1157 157 0.0005
Consequences: Need a treatment 53.6 68.2 53 68 1263 51 0.0003
Consequences: Need a treatment by a physician 53.6 71.7 53 70.5 1274 40 0.0000
Measures to stop aggression: Peroral medication 52.8 58.5 52 59 996 318 0.0079
Measures to stop aggression: Parenteral medication 53.4 60.7 52 61 1168 146 0.0100
Measures to stop aggression: Held with force 52.5 58.1 51 58 915 399 0.0000
Measures to stop aggression: Seclusion 51.5 61 51 62 948 366 0.0000
Measures to stop aggression: Physical restrant 53.8 71.8 53 73 1283 31 0.0002
0*: With negative response in the items 1**: With positive response in the items
Finally, we were curious to assess the correlation between SOAS scores and SOAS gravity scores. Figure 3 scattered these scores for all reported incidents. A correlation of 0.26 was calculated which ascertains the existence of a positive linear dependence. In the SUPAA, a psychogeriatric service (figure 4), the correlation between SOAS scores and gravity scores is not reported.
5 10 15 20
SOAS-R Score Gravity score 020406080100
5 10 15 20
SOAS-R Score Gravity score 020406080100
Figure 3:
A scatter plot of gravity score versus SOAS-R score for DP-CHUV with their regression line (ϱ=0.26) and the corresponding 95% CI
Figure 4:
A scatter plot of gravity score versus SOAS-R score for SUPAA with their regression line (ϱ=0.09) and their corresponding 95% CI
20 40 60 80
Age (years) Gravity score 020406080100
Figure 5:
A scatter plot of gravity score versus age(ϱ=−0.18)
Comparing the age for SOAS patients and SOAS gravity scores (figure 5), we can note the negative correlation (ϱ=−0.18) between the age of the patients and the SOAS-gravity scores.
Discussion
This study is the first one realized in the DP-CHUV. The important data collected since five years provides in depth information about the nature, frequency and severity of violence acts occurred in different acute psychiatric services. Some of these results are in agreement with several previous studies developed in the
international literature: In comparison with non-violent patients, violent patient were younger, men. Some psychiatric disorders can increase the risk for a violent act (e.g. F2 and F3 diagnoses, associated with an abuse substance disorder - F1-). One third of the incidents were serious. Nursing staff are the principal target of the incidents. Special attention must be given to develop training and strategies to prevent and to cope with patient’s violence behaviour. After the incidents, special attention on victim’s psychological support is needed. The link between violence and constraint or coercion seems to be important. A patient can be aggressive because he is involuntarily hospitalised, or wants to leave the seclusion room, the hospital or the ward without authorisation. Moreover, putting the patient in a seclusion room is a practice used to prevent and contain violence and to decrease symptomatology. It can be also a place with risks for nursing or medical staff. The total incidence rate per 100 hospitalization days is rather low (0.34 incident) comparing those reported from literature (Abderhalden 2007; Salamin 2010; Nijman 2005). Furthermore, differences between services can be questioned. These results must allow us some assumptions to explain this low incident rate for the DP-CHUV and the differences between services: Under reporting? Better violence prevention management? Specific procedures to report incidents? No uniform violence definition between caregivers? …). Whatever these assumptions, we should not either forget that comparisons between institutions are difficult to do or must be done with precautions. “The expression of ward incident rates has been unclear and disorganized, resulting in incomparability between studies and lack of precision”
(Bowers 2002, cited by Abderhalden, 2007). The results of these study shows that verbal violence can have an impact equally important to victims than physical violence. This point is really interesting. It reminds over the subjective aspect on the violence act impact on victims. It also raises questions about the structure of the SOAS-R scale and the SOAS-R scores structure.
Finally, the results underscore the necessity for our institution to assess violent incidents. The information collected help us do violence risk management and to better support victims. However, questions remain regarding the meanings attributed in using this tool as information and communication support with the patient and multidisciplinary team: How do we use the SOAS-R in clinic? What is the support perceived by staff?
Acknowledgements
We gratefully aknowledge the staff of the participating wards for their collaboration in the collection data.
References
Abderhalden Christoph § al., Frequency and severity of aggressive incidents in acute psychiatric wards in Switzerland.
Abderhalden Clinical Practice and Epidemiology in Mental Health, 2007, 3: 30
Bowers L., The expression and comparison of ward incident rates. Issues Mental Health Nurs 2000, 21 (4): 365-374 Grassi L. § al., Characteristics of violent behaviour in acute psychiatric in-patients: a 5-years Italian study. Acta Psychiatr.
Scand 2001: 104; 273-279
Grassi L. § al., Violence in psychiatric units: a 7-year Italian study of persistently assaultive patients. Soc Psychiatry Psychiatr Epidemiol (2006) 41:698-703
De Niet G. J., G. J. M. Hutschemakers, B.H. H. G. Lendenmeijer, Is the reducing effect of the Staff Observation Aggression Scale owing to a learning effect? An explorative study. Journal of Psychiatric and Mental Health Nursing, 2005, 12, 687-694 Nijman H.L. § al., Fifteen years of research with the Staff Observation Aggression Scale: a review. Acta Psychiatr. Scand
2005: 111; 12-21
Nijman H.L. § al., The Staff Observation Aggression Scale - Revised (SOAS-R). Aggressive Behavior. 1999; 25:197–209 Nijman H.L. § Palmsternia T., Measuring aggression with the Staff observation aggression scale-revised, Acta Psychiatrica
Scand, 2002, 106 (suppl.412)
Salamin Virginie, Schuwey-Hayoz Aline, Giacometti Bickel Graziella, Epidémiologie des comportements agressifs en psychiatrie hospitalière: état des lieux dans le canton de Fribourg. Schweitzer Archiv für neurologie und psychiatrie 2010; 161(1): 23-9