563845 research-article2014
APY0010.1177/1039856214563845Australasian PsychiatryKasinathan et al.
Australasian
Psychiatry
Forensic psychiatry
Assessing the risk of imminent aggression in mentally ill young offenders
Australasian Psychiatry 2015, Vol 23(1) 44–48 © The Royal Australian and New Zealand College of Psychiatrists 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1039856214563845 apy.sagepub.com
John Kasinathan Consultant Forensic Psychiatrist, Adolescent Unit, The Forensic Hospital, Justice Health, New South Wales Health, Sydney, NSW, and; Conjoint Lecturer, Faculty of Medicine, School of Psychiatry, University of New South Wales, Sydney, NSW, and; Forensic Mental Health Service Australian Capital Territory Health, Canberra, ACT, and; Visiting Fellow, Centre for Mental Health Research, The Australian National University, Canberra, ACT, Australia Christopher Marsland Clinical Nurse Consultant, Adolescent Unit and Justice Health, The Forensic Hospital, New South Wales Health, Sydney, NSW, Australia
Philip Batterham Fellow in Mental Health Research, Centre for Mental Health Research, The Australian National University, Canberra, ACT, Australia
Claire Gaskin Clinical Director Adolescent Mental Health, Justice Health, The Forensic Hospital, New South Wales Health, Sydney, NSW, and; Conjoint Lecturer, Faculty of Medicine, School of Psychiatry University of New South Wales, Sydney, NSW, Australia Jonathon Adams Consultant Forensic Psychiatrist, Justice Health, The Forensic Hospital, New South Wales Health, Sydney, NSW, and; Conjoint Lecturer, School of Psychiatry, Faculty of Medicine, University of New South Wales, Sydney, NSW, Australia
Michael Daffern Principal Consultant Psychologist, Centre for Forensic Behavioural Science University of Swinburne, Melbourne, VIC, and; Victorian Institute of Forensic Mental Health, Melbourne, VIC, Australia
Abstract Objectives: Aggression in adolescents presents a significant problem for psychiatric units. The Dynamic Appraisal of Situational Aggression (DASA) is an empirically validated measure designed to appraise the risk of imminent aggression (within the next 24 hours) in adult patients. Our aim was to examine the predictive validity of the DASA: Youth Version (DASA:YV) with youth-specific items, in young offenders hospitalised with a mental illness. Methods: This prospective validation study involved 4440 DASA:YV ratings of mentally ill adolescents in a secure hospital. At 24 hours post-assessment, the nursing staff documented whether patients had behaved aggressively: physically, verbally or towards property. Predictive accuracy was assessed using the area under the curve (AUC) of the receiver operating characteristic (ROC) curve. Results: The DASA:YV significantly predicted any imminent aggression (AUC = 0.754). Additional youth-specific items conferred a greater predictive yield, as compared to adult-derived items (p = 0.014). Conclusions: It is possible to monitor the risk state of hospitalised mentally ill youth, so that heightened states can be detected early, thus facilitating interventions to reduce the risk of violence. Keywords: aggression, appraisal methods, dynamic appraisal, mental illness, risk, violent behaviour, youth
A
ggression occurs frequently and is problematic in forensic, civil and adolescent psychiatric units.1,2 Aggressive behaviours have a negative impact on the well-being of victims and those whom witness aggression.3 Adolescent psychiatric units are unique environments where youth with mental illness are frequently detained against their will.4 Patients perceived to be at risk of aggression are often subjected to restraint
and seclusion,5,6 which can be intrusive, contentious7,8 and linked to deaths.9 Barriers to reducing restraint and
Corresponding author: John Kasinathan, Adolescent Unit, Forensic Hospital, PO Box 150, Matraville 2036, Sydney, NSW, Australia. Email: [email protected]
44 Downloaded from apy.sagepub.com at UNIV CALIFORNIA SAN DIEGO on June 9, 2015
Kasinathan et al.
seclusion include a lack of staff knowledge about early warning signs of aggression.5 Violence risk assessment fulfils a critical role in violence prevention and the reduction of restrictive practices. Appraisal of patients’ risk state (the intra-individual variability in violence potential)10 may facilitate patient management, guiding treatments and activities that are adversely affected by violence.11 Structured violence risk assessment schemes were developed in recent years, following the repeated demonstration that empirically-derived instruments have stronger predictive accuracy than unaided clinical judgements.12 Structured measures developed to assess short-term violence risk in psychiatric units include the Brøset Violence Checklist (BVC)13 and the Dynamic Appraisal of Situational Aggression (DASA).14 Although the predictive accuracy of DASA was established in local and international studies (area under the curve (AUC) = 0.61 – 0.82),11,14–18 the impact on effectiveness of interventions remains unclear; however, several studies suggest that structured violence risk assessments can contribute to informed management, resulting in reduced violence and restrictive practices.19–21 Studies have tested the predictive validity of violence risk assessments within youth institutions, including the Structured Assessment of Violence Risk in Youth (SAVRY),22 but these were designed to predict violence over the medium to long-term (months to years). Such measures are too cumbersome and lack the capacity to measure change on a daily basis. Developmental factors may weaken the accuracy of adult-derived measures used with youth.23 Chu et al.23 found that the predictive validity of the DASA in detained young Singaporean offenders without mental illness was only modest (AUC = 0.59). This poorer predictive validity, compared with DASAbased adult studies, was hypothesised as being because aggression in generic young offenders was related to different factors, compared to psychiatric patients. We aimed to prospectively determine the predictive validity of the DASA in young offenders hospitalised with a mental illness. We hypothesised that additional youthspecific measures would confer greater predictive validity.
Methods
Aboriginal (n = 15; 43%). Diagnoses, made using the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR),24 revealed that most (n = 33; 94%) had a mental illness, while two (6%) had conduct disorder only: schizophrenia spectrum disorder (n = 29; 83%) was the most common mental illness, 4 (11%) patients had bipolar disorder and 31 (89%) patients had conduct disorder. During the study, two patients (6%) graduated to rehabilitation units within the Forensic Hospital and 27 patients (77%) were discharged to the community or detention. Measures: Dynamic Appraisal of Situational Aggression: Youth Version The DASA: Youth Version (DASA:YV) comprises 11 items incorporating the seven following DASA items:14 negative attitudes and impulsivity from the HCR-2025; irritability and verbal threats from the BVC13; and sensitivity to perceived provocation, easily angered when requests are denied and unwillingness to follow directions from the DASA development study.14 Four youth-specific items were added to form the DASA:YV: anxious or fearful, low empathy/remorse and significant peer rejection from the SAVRY22 (with permission) and outside stressors from the generic young offender validation study.23,26 Procedures First, staff were trained to score the DASA:YV by the first and second authors (JK and CM). Then daily, from 15 March 2011 to 28 November 2013 at 9:00 p.m., the trained nurses completed the DASA:YV for each inpatient, scoring each item for its presence or absence in the 24 hours prior. Patients with an increase in the behaviour were scored as 1. Staff was not instructed to act in any particular way if there were elevated scores, apart from standard clinical practice. Additionally, staff rated their final judgment of imminent aggression (final risk rating) as low, medium or high. To test for predictive validity, at the time of the assessment staff completed a daily recording of the patient’s aggression during the preceding 24 hours. Each day’s DASA:YV and final risk rating was matched with the subsequent day’s recording of aggression.
Setting
The Justice Health Human Research and Ethics Committee approved this study. The study setting was a 6-bed adolescent forensic psychiatric unit providing acute care and rehabilitation to adolescent patients within the Forensic Hospital (a 135-bed high-security psychiatric hospital) in Sydney, Australia. Participants Participants comprised 35 patients (34 male and 1 female patient) from 14 – 21 years of age. Under one-half were
Analytic approach Statistical analyses were undertaken using SPSS version 19.0 (IBM, Chicago, IL, USA) and we conducted AUC comparisons with SAS version 9.3 (SAS Institute, Cary, NC, USA). The predictive validity of the DASA, DASA:YV and final risk rating were assessed using the AUC of the ROC, which plots the true positive rate (sensitivity) against the false positive rate (specificity). It is less dependent on the base rate of violence than other measures of predictive accuracy.27 The AUC ranges from 0
45 Downloaded from apy.sagepub.com at UNIV CALIFORNIA SAN DIEGO on June 9, 2015
Australasian Psychiatry 23(1)
Table 1. AUC for risk predictions for aggression Final risk rating
DASA (7 item)
DASA:YV (11 item)
AUC
95% CI
AUC
95% CI
AUC
95% CI
Any aggression Physical aggression toward others Verbal aggression Physical aggression toward objects
0.687a 0.667a 0.694a 0.683a
0.65 – 0.73 0.58 – 0.75 0.65 – 0.74 0.63 – 0.74
0.738a 0.711a 0.724a 0.744a
0.70 – 0.78 0.63 – 0.79 0.68-0.77 0.69 – 0.79
0.754a 0.720a 0.742a 0.751a
0.72 – 0.79 0.64 – 0.80 0.70 – 0.79 0.70 – 0.80
ap
< 0.001. AUC: Area under the curve; DASA: Dynamic Appraisal of Situational Aggression; DASA:YV: Youth Version of DASA.
(perfect negative prediction) to 0.5 (chance prediction) to 1.0 (perfect positive prediction). An AUC of 0.70 implies there is a 70% chance that a truly violent person would score higher than the cut-off for violence on the instrument; and a truly non-violent person would score below the cut-off value.
Both the DASA:YV and DASA had greater predictive validity than the ‘final risk rating’, a clinical judgement of risk (respectively, the χ2 = 14.57 and p < 0.001, and χ2 = 8.28 and p = 0.004; Figure 1).
Results
Our findings support the hypothesis that structured behaviour ratings by clinicians can assist the appraisal of aggression risk for the next 24 hours, in young offenders hospitalised with a mental illness. Both the DASA and DASA:YV showed strong predictive validity that was significantly greater than the clinician’s judgment of imminent risk, similar to findings by Griffith et al.18 The hypothesis that the DASA:YV would confer greater predictive validity was supported. The predictive validity of the DASA and DASA:YV for different types of aggression was uniform, in contrast with prior DASA studies15 that had more AUC variation for different aggression types. Any score of 1 or higher on the DASA and DASA:YV maximised sensitivity and specificity for both instruments.
DASA:YV ratings A total of 4440 DASA:YV ratings were completed. The mean total was 0.65, the median was 0 (SD 1.6) and they ranged from 0 to 11. Ratings were skewed (3.04). In total, 78% of DASA:YV ratings were scored as 0; 7.2% were scored as 1; 4.0% were scored as 2; 3.1% were scored as 3; 2.5% were scored as 4; 1.5% were scored as 5; 1.2% were scored as 6; 0.8% were scored as 7 and 1.0% were scored as 8 or higher. Aggressive behaviour There were 228 incidents of aggression in the study period: there were 61 episodes of physical aggression to others, 180 episodes of verbal aggression and 127 episodes of aggression toward objects. Predictive validity of the DASA, DASA:YV and clinician risk rating All AUC values for the DASA, DASA:YV and final risk rating were significantly greater than chance (Table 1). Youden indices indicated that a score of 1 or higher on the DASA:YV maximised sensitivity (65.9%) and specificity (81.1%) in predicting any aggression. Likewise, a score of 1 or higher on the DASA maximised sensitivity (60.9%) and specificity (83.1%). A final risk rating of medium or greater had 47.5% sensitivity and 89.3% specificity. AUC comparisons for any form of aggression revealed that the DASA:YV (AUC = 0.754) conveyed greater predictive validity than the DASA (χ2 = 6.04; p = 0.014).
Discussion
The predictive validity for imminent aggression in the present study was comparable to adult studies11,14,15,17 and greater than a generic young offender study.23 Our setting had a higher staff-to-youth ratio (about 1:3), compared to the generic youth study (about 1:30)23; thus, our staff probably dedicated sufficient time to ensure a valid assessment, facilitating stronger predictive validity of the instrument. This study had limitations. Gender imbalance was present. Sample size was limited. Although there were a great number of risk assessments, ongoing validation in larger adolescent populations may provide further evidence for the predictive validity of the DASA:YV and scope for improving the instrument further, by including lower-severity risk factors to provide greater differentiation of the risk status. We felt it was worthwhile for the clinicians to complete the additional four items of the DASA:YV, beyond the 7-item DASA. First, because we found the DASA:YV was significantly more predictive than the DASA. Second,
46 Downloaded from apy.sagepub.com at UNIV CALIFORNIA SAN DIEGO on June 9, 2015
Kasinathan et al.
Figure 1. ROC curves for the DASA:YV, the 7-item DASA and the clinicians’ final risk rating. Diagonal segments are produced by ties. DASA: Dynamic Appraisal of Situational Aggression; ROC: receiver operating characteristic; DASA: YV: Youth Version of DASA. because three of the four youth-specific items were derived from research demonstrating robust links between these factors and youth violence.22 Third, because the DASA:YV is quick and easy to use, as scoring the additional four items occupied only 30 seconds more for each rating, compared to the DASA. The strong predictive validity of the DASA:YV renders it useful in conjunction with clinical judgment for imminent aggression risk in mentally ill youth. The presence of items measured by the DASA:YV should indicate the need for intervention, particularly with higher total scores.26 We found support for the DASA:YV to assist clinicians in day-to-day assessments of imminent aggression risk in mentally ill youth. By monitoring the risk state of mentally ill youth, heightened risk states can be detected early, appropriate interventions can be implemented and the likelihood of violent outcomes may be reduced.
2. Barzman DH, Brackenbury L, Sonnier L, et al. Brief Rating of Aggression by Children and Adolescents (BRACHA): Development of a tool for assessing risk of inpatients’ aggressive behavior. J Am Acad Psychiatry Law 2011; 39: 170–179. 3. Royal College of Psychiatrists (RCP). Healthcare Commission national audit of violence 2006–2007. London: RCP, 2007. 4. Gadon L, Johnstone L and Cooke D. Situational variables and institutional violence: A systematic review of the literature. Clin Psychol Rev 2006; 26: 515–534. 5. Royal Australian and New Zealand College of Psychiatrists (RANZCP). Position Statement 6: Minimising the use of seclusion and restraint in people with mental illness. RANZCP, 2010. 6. Ching H, Daffern M, Martin T, et al. Reducing the use of seclusion in a forensic psychiatric hospital: Assessing the impact on aggression, therapeutic climate and staff confidence. J Forensic Psychiatry Psychol 2010; 21: 737–760. 7. Fisher W. Restraint and seclusion: A review of the literature. Am J Psychiatry 1994; 15: 1584–1591. 8. Sailas E and Fenton M. Seclusion and restraint for people with serious mental illness. Cochrane Datab Systemat Rev 2012; 2: CD001163.
Acknowledgements
9. Department of Health. A new Mental Health Act for Victoria: Summary of proposed reforms. Melbourne: State Government of VIC, 2012.
We thank all the nursing staff of the Adolescent Unit of The Forensic Hospital for their commitment in taking up and using the DASA:YV. We especially thank Nicole Beaumont for her tireless assistance with data entry.
10. Douglas KS and Skeem JL. Violence risk assessment: Getting specific about being dynamic. Psychol Public Policy Law 2005; 11: 347–383.
Disclosure
11. Daffern M and Howells K. The prediction of imminent aggression and self- harm in personality disordered patients of a high security hospital using the HCR-20 clinical scale and the Dynamic Appraisal of Situational Aggression. Int J Forens Mental Health 2007; 6: 137–143.
The authors declare that there are no conflict of interest. The authors alone are responsible for the content and writing of the paper.
12. Quinsey V, Harris G, Rice M, et al. Violent offenders: Appraising and managing risk. 2nd ed. Washington, DC: American Psychological Association, 2006.
References 1. Daffern M, Mayer M and Martin T. A preliminary investigation into patterns of aggression in anAustralian forensic psychiatric hospital. J Forens Psychiatry Psychol 2003; 14: 67–84.
13. Almvik R, Woods P and Rasmussen K. The Brøset Violence Checklist: Sensitivity, specificity, and inter-rater reliability. J Interpers Viol 2000; 15: 1284–1296.
47 Downloaded from apy.sagepub.com at UNIV CALIFORNIA SAN DIEGO on June 9, 2015
Australasian Psychiatry 23(1) 14. Ogloff JRP and Daffern M. The dynamic appraisal of situational aggression: An instrument to assess risk for imminent aggression in psychiatric inpatients. Behav Sci Law 2006; 24: 799–813.
21. Van de Sande R, Nijman HL, Noorthoorn EO, et al. Aggression and seclusion on acute psychiatric wards: Effect of short-term risk assessment. Brit J Psychiatry 2011; 199: 473–478.
15. Barry-Walsh J, Daffern M, Duncan S, et al. The prediction of imminent aggression in patients with mental illness and/or intellectual disability using the Dynamic Appraisal of Situational Aggression instrument. Australas Psychiatry 2009; 17: 493–496.
22. Borum R, Bartel P and Forth AE. Manual for the Structured Assessment of Violence Risk in Youth (SAVRY) Version 1.1. Odessa, FL: Psychological Assessment Resources, 2003.
16. Daffern M, Howells K, Hamilton L, et al. The impact of structured risk assessments followed by management recommendations on aggression in patients with personality disorder. J Forens Psychiatry Psychol 2009; 20: 661–679. 17. Vojt G, Marshall LA and Thomson LDG. The assessment of imminent inpatient aggression: A validation study of the DASA-IV in Scotland. J Forensic Psychiatry Psychol 2010; 21: 789–800. 18. Griffith JJ, Daffern M and Godber T. Examination of the predictive validity of the Dynamic Appraisal of Situational Aggression in two mental health units. Int J Mental Health Nurs 2013; 22: 485–492. 19. Needham I, Abderhalden C, Meer R, et al. The effectiveness of two interventions in the management of patient violence in acute mental inpatient settings: Report on a pilot study. J Psychiatric Mental Health Nurs 2004; 11: 595–601. 20. Abderhalden C, Needham I, Dassen T, et al. Structured risk assessment and violence in acute psychiatric wards: Randomised controlled trial. Brit J Psychiatry 2008; 193: 44–50.
23. Chu CM, Hoo E, Daffern M, et al. Assessing the risk of imminent aggression in institutionalized youth offenders using the dynamic appraisal of situational aggression. J Forens Psychiatry Psychol 2012; 23: 168–183. 24. American Psychiatric Association. Diagnostic and statistical manual of mental disorders, text revision (DSM-IV-TR). 4th ed. Washington, DC: American Psychiatric Press; 2000. 25. Webster CD, Douglas KS, Eaves D, et al. HCR-20: Assessing risk for violence. Version 2 ed. British Columbia, Canada: Mental Health, Law and Policy Institute of Simon Fraser University, 1997. 26. Daffern M, Ogloff JRP, Chu CM, et al. The Dynamic Appraisal of Situational Aggression: Youth Version (DASA:YV). Victoria: Centre for Forensic Behavioural Science at Monash University, 2008. 27. Douglas KS and Webster CD. Predicting violence in mentally and personality disordered individuals. In: Roesch R, Hart SD and Ogloff JRP (eds) Psychology and law: The state of the discipline. New York: Plenum, 1999, pp. 175–239.
48 Downloaded from apy.sagepub.com at UNIV CALIFORNIA SAN DIEGO on June 9, 2015
APY0010.1177/1039856214563845Australasian PsychiatryKasinathan et al.
Australasian
Psychiatry
Forensic psychiatry
Assessing the risk of imminent aggression in mentally ill young offenders
Australasian Psychiatry 2015, Vol 23(1) 44–48 © The Royal Australian and New Zealand College of Psychiatrists 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1039856214563845 apy.sagepub.com
John Kasinathan Consultant Forensic Psychiatrist, Adolescent Unit, The Forensic Hospital, Justice Health, New South Wales Health, Sydney, NSW, and; Conjoint Lecturer, Faculty of Medicine, School of Psychiatry, University of New South Wales, Sydney, NSW, and; Forensic Mental Health Service Australian Capital Territory Health, Canberra, ACT, and; Visiting Fellow, Centre for Mental Health Research, The Australian National University, Canberra, ACT, Australia Christopher Marsland Clinical Nurse Consultant, Adolescent Unit and Justice Health, The Forensic Hospital, New South Wales Health, Sydney, NSW, Australia
Philip Batterham Fellow in Mental Health Research, Centre for Mental Health Research, The Australian National University, Canberra, ACT, Australia
Claire Gaskin Clinical Director Adolescent Mental Health, Justice Health, The Forensic Hospital, New South Wales Health, Sydney, NSW, and; Conjoint Lecturer, Faculty of Medicine, School of Psychiatry University of New South Wales, Sydney, NSW, Australia Jonathon Adams Consultant Forensic Psychiatrist, Justice Health, The Forensic Hospital, New South Wales Health, Sydney, NSW, and; Conjoint Lecturer, School of Psychiatry, Faculty of Medicine, University of New South Wales, Sydney, NSW, Australia
Michael Daffern Principal Consultant Psychologist, Centre for Forensic Behavioural Science University of Swinburne, Melbourne, VIC, and; Victorian Institute of Forensic Mental Health, Melbourne, VIC, Australia
Abstract Objectives: Aggression in adolescents presents a significant problem for psychiatric units. The Dynamic Appraisal of Situational Aggression (DASA) is an empirically validated measure designed to appraise the risk of imminent aggression (within the next 24 hours) in adult patients. Our aim was to examine the predictive validity of the DASA: Youth Version (DASA:YV) with youth-specific items, in young offenders hospitalised with a mental illness. Methods: This prospective validation study involved 4440 DASA:YV ratings of mentally ill adolescents in a secure hospital. At 24 hours post-assessment, the nursing staff documented whether patients had behaved aggressively: physically, verbally or towards property. Predictive accuracy was assessed using the area under the curve (AUC) of the receiver operating characteristic (ROC) curve. Results: The DASA:YV significantly predicted any imminent aggression (AUC = 0.754). Additional youth-specific items conferred a greater predictive yield, as compared to adult-derived items (p = 0.014). Conclusions: It is possible to monitor the risk state of hospitalised mentally ill youth, so that heightened states can be detected early, thus facilitating interventions to reduce the risk of violence. Keywords: aggression, appraisal methods, dynamic appraisal, mental illness, risk, violent behaviour, youth
A
ggression occurs frequently and is problematic in forensic, civil and adolescent psychiatric units.1,2 Aggressive behaviours have a negative impact on the well-being of victims and those whom witness aggression.3 Adolescent psychiatric units are unique environments where youth with mental illness are frequently detained against their will.4 Patients perceived to be at risk of aggression are often subjected to restraint
and seclusion,5,6 which can be intrusive, contentious7,8 and linked to deaths.9 Barriers to reducing restraint and
Corresponding author: John Kasinathan, Adolescent Unit, Forensic Hospital, PO Box 150, Matraville 2036, Sydney, NSW, Australia. Email: [email protected]
44 Downloaded from apy.sagepub.com at UNIV CALIFORNIA SAN DIEGO on June 9, 2015
Kasinathan et al.
seclusion include a lack of staff knowledge about early warning signs of aggression.5 Violence risk assessment fulfils a critical role in violence prevention and the reduction of restrictive practices. Appraisal of patients’ risk state (the intra-individual variability in violence potential)10 may facilitate patient management, guiding treatments and activities that are adversely affected by violence.11 Structured violence risk assessment schemes were developed in recent years, following the repeated demonstration that empirically-derived instruments have stronger predictive accuracy than unaided clinical judgements.12 Structured measures developed to assess short-term violence risk in psychiatric units include the Brøset Violence Checklist (BVC)13 and the Dynamic Appraisal of Situational Aggression (DASA).14 Although the predictive accuracy of DASA was established in local and international studies (area under the curve (AUC) = 0.61 – 0.82),11,14–18 the impact on effectiveness of interventions remains unclear; however, several studies suggest that structured violence risk assessments can contribute to informed management, resulting in reduced violence and restrictive practices.19–21 Studies have tested the predictive validity of violence risk assessments within youth institutions, including the Structured Assessment of Violence Risk in Youth (SAVRY),22 but these were designed to predict violence over the medium to long-term (months to years). Such measures are too cumbersome and lack the capacity to measure change on a daily basis. Developmental factors may weaken the accuracy of adult-derived measures used with youth.23 Chu et al.23 found that the predictive validity of the DASA in detained young Singaporean offenders without mental illness was only modest (AUC = 0.59). This poorer predictive validity, compared with DASAbased adult studies, was hypothesised as being because aggression in generic young offenders was related to different factors, compared to psychiatric patients. We aimed to prospectively determine the predictive validity of the DASA in young offenders hospitalised with a mental illness. We hypothesised that additional youthspecific measures would confer greater predictive validity.
Methods
Aboriginal (n = 15; 43%). Diagnoses, made using the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR),24 revealed that most (n = 33; 94%) had a mental illness, while two (6%) had conduct disorder only: schizophrenia spectrum disorder (n = 29; 83%) was the most common mental illness, 4 (11%) patients had bipolar disorder and 31 (89%) patients had conduct disorder. During the study, two patients (6%) graduated to rehabilitation units within the Forensic Hospital and 27 patients (77%) were discharged to the community or detention. Measures: Dynamic Appraisal of Situational Aggression: Youth Version The DASA: Youth Version (DASA:YV) comprises 11 items incorporating the seven following DASA items:14 negative attitudes and impulsivity from the HCR-2025; irritability and verbal threats from the BVC13; and sensitivity to perceived provocation, easily angered when requests are denied and unwillingness to follow directions from the DASA development study.14 Four youth-specific items were added to form the DASA:YV: anxious or fearful, low empathy/remorse and significant peer rejection from the SAVRY22 (with permission) and outside stressors from the generic young offender validation study.23,26 Procedures First, staff were trained to score the DASA:YV by the first and second authors (JK and CM). Then daily, from 15 March 2011 to 28 November 2013 at 9:00 p.m., the trained nurses completed the DASA:YV for each inpatient, scoring each item for its presence or absence in the 24 hours prior. Patients with an increase in the behaviour were scored as 1. Staff was not instructed to act in any particular way if there were elevated scores, apart from standard clinical practice. Additionally, staff rated their final judgment of imminent aggression (final risk rating) as low, medium or high. To test for predictive validity, at the time of the assessment staff completed a daily recording of the patient’s aggression during the preceding 24 hours. Each day’s DASA:YV and final risk rating was matched with the subsequent day’s recording of aggression.
Setting
The Justice Health Human Research and Ethics Committee approved this study. The study setting was a 6-bed adolescent forensic psychiatric unit providing acute care and rehabilitation to adolescent patients within the Forensic Hospital (a 135-bed high-security psychiatric hospital) in Sydney, Australia. Participants Participants comprised 35 patients (34 male and 1 female patient) from 14 – 21 years of age. Under one-half were
Analytic approach Statistical analyses were undertaken using SPSS version 19.0 (IBM, Chicago, IL, USA) and we conducted AUC comparisons with SAS version 9.3 (SAS Institute, Cary, NC, USA). The predictive validity of the DASA, DASA:YV and final risk rating were assessed using the AUC of the ROC, which plots the true positive rate (sensitivity) against the false positive rate (specificity). It is less dependent on the base rate of violence than other measures of predictive accuracy.27 The AUC ranges from 0
45 Downloaded from apy.sagepub.com at UNIV CALIFORNIA SAN DIEGO on June 9, 2015
Australasian Psychiatry 23(1)
Table 1. AUC for risk predictions for aggression Final risk rating
DASA (7 item)
DASA:YV (11 item)
AUC
95% CI
AUC
95% CI
AUC
95% CI
Any aggression Physical aggression toward others Verbal aggression Physical aggression toward objects
0.687a 0.667a 0.694a 0.683a
0.65 – 0.73 0.58 – 0.75 0.65 – 0.74 0.63 – 0.74
0.738a 0.711a 0.724a 0.744a
0.70 – 0.78 0.63 – 0.79 0.68-0.77 0.69 – 0.79
0.754a 0.720a 0.742a 0.751a
0.72 – 0.79 0.64 – 0.80 0.70 – 0.79 0.70 – 0.80
ap
< 0.001. AUC: Area under the curve; DASA: Dynamic Appraisal of Situational Aggression; DASA:YV: Youth Version of DASA.
(perfect negative prediction) to 0.5 (chance prediction) to 1.0 (perfect positive prediction). An AUC of 0.70 implies there is a 70% chance that a truly violent person would score higher than the cut-off for violence on the instrument; and a truly non-violent person would score below the cut-off value.
Both the DASA:YV and DASA had greater predictive validity than the ‘final risk rating’, a clinical judgement of risk (respectively, the χ2 = 14.57 and p < 0.001, and χ2 = 8.28 and p = 0.004; Figure 1).
Results
Our findings support the hypothesis that structured behaviour ratings by clinicians can assist the appraisal of aggression risk for the next 24 hours, in young offenders hospitalised with a mental illness. Both the DASA and DASA:YV showed strong predictive validity that was significantly greater than the clinician’s judgment of imminent risk, similar to findings by Griffith et al.18 The hypothesis that the DASA:YV would confer greater predictive validity was supported. The predictive validity of the DASA and DASA:YV for different types of aggression was uniform, in contrast with prior DASA studies15 that had more AUC variation for different aggression types. Any score of 1 or higher on the DASA and DASA:YV maximised sensitivity and specificity for both instruments.
DASA:YV ratings A total of 4440 DASA:YV ratings were completed. The mean total was 0.65, the median was 0 (SD 1.6) and they ranged from 0 to 11. Ratings were skewed (3.04). In total, 78% of DASA:YV ratings were scored as 0; 7.2% were scored as 1; 4.0% were scored as 2; 3.1% were scored as 3; 2.5% were scored as 4; 1.5% were scored as 5; 1.2% were scored as 6; 0.8% were scored as 7 and 1.0% were scored as 8 or higher. Aggressive behaviour There were 228 incidents of aggression in the study period: there were 61 episodes of physical aggression to others, 180 episodes of verbal aggression and 127 episodes of aggression toward objects. Predictive validity of the DASA, DASA:YV and clinician risk rating All AUC values for the DASA, DASA:YV and final risk rating were significantly greater than chance (Table 1). Youden indices indicated that a score of 1 or higher on the DASA:YV maximised sensitivity (65.9%) and specificity (81.1%) in predicting any aggression. Likewise, a score of 1 or higher on the DASA maximised sensitivity (60.9%) and specificity (83.1%). A final risk rating of medium or greater had 47.5% sensitivity and 89.3% specificity. AUC comparisons for any form of aggression revealed that the DASA:YV (AUC = 0.754) conveyed greater predictive validity than the DASA (χ2 = 6.04; p = 0.014).
Discussion
The predictive validity for imminent aggression in the present study was comparable to adult studies11,14,15,17 and greater than a generic young offender study.23 Our setting had a higher staff-to-youth ratio (about 1:3), compared to the generic youth study (about 1:30)23; thus, our staff probably dedicated sufficient time to ensure a valid assessment, facilitating stronger predictive validity of the instrument. This study had limitations. Gender imbalance was present. Sample size was limited. Although there were a great number of risk assessments, ongoing validation in larger adolescent populations may provide further evidence for the predictive validity of the DASA:YV and scope for improving the instrument further, by including lower-severity risk factors to provide greater differentiation of the risk status. We felt it was worthwhile for the clinicians to complete the additional four items of the DASA:YV, beyond the 7-item DASA. First, because we found the DASA:YV was significantly more predictive than the DASA. Second,
46 Downloaded from apy.sagepub.com at UNIV CALIFORNIA SAN DIEGO on June 9, 2015
Kasinathan et al.
Figure 1. ROC curves for the DASA:YV, the 7-item DASA and the clinicians’ final risk rating. Diagonal segments are produced by ties. DASA: Dynamic Appraisal of Situational Aggression; ROC: receiver operating characteristic; DASA: YV: Youth Version of DASA. because three of the four youth-specific items were derived from research demonstrating robust links between these factors and youth violence.22 Third, because the DASA:YV is quick and easy to use, as scoring the additional four items occupied only 30 seconds more for each rating, compared to the DASA. The strong predictive validity of the DASA:YV renders it useful in conjunction with clinical judgment for imminent aggression risk in mentally ill youth. The presence of items measured by the DASA:YV should indicate the need for intervention, particularly with higher total scores.26 We found support for the DASA:YV to assist clinicians in day-to-day assessments of imminent aggression risk in mentally ill youth. By monitoring the risk state of mentally ill youth, heightened risk states can be detected early, appropriate interventions can be implemented and the likelihood of violent outcomes may be reduced.
2. Barzman DH, Brackenbury L, Sonnier L, et al. Brief Rating of Aggression by Children and Adolescents (BRACHA): Development of a tool for assessing risk of inpatients’ aggressive behavior. J Am Acad Psychiatry Law 2011; 39: 170–179. 3. Royal College of Psychiatrists (RCP). Healthcare Commission national audit of violence 2006–2007. London: RCP, 2007. 4. Gadon L, Johnstone L and Cooke D. Situational variables and institutional violence: A systematic review of the literature. Clin Psychol Rev 2006; 26: 515–534. 5. Royal Australian and New Zealand College of Psychiatrists (RANZCP). Position Statement 6: Minimising the use of seclusion and restraint in people with mental illness. RANZCP, 2010. 6. Ching H, Daffern M, Martin T, et al. Reducing the use of seclusion in a forensic psychiatric hospital: Assessing the impact on aggression, therapeutic climate and staff confidence. J Forensic Psychiatry Psychol 2010; 21: 737–760. 7. Fisher W. Restraint and seclusion: A review of the literature. Am J Psychiatry 1994; 15: 1584–1591. 8. Sailas E and Fenton M. Seclusion and restraint for people with serious mental illness. Cochrane Datab Systemat Rev 2012; 2: CD001163.
Acknowledgements
9. Department of Health. A new Mental Health Act for Victoria: Summary of proposed reforms. Melbourne: State Government of VIC, 2012.
We thank all the nursing staff of the Adolescent Unit of The Forensic Hospital for their commitment in taking up and using the DASA:YV. We especially thank Nicole Beaumont for her tireless assistance with data entry.
10. Douglas KS and Skeem JL. Violence risk assessment: Getting specific about being dynamic. Psychol Public Policy Law 2005; 11: 347–383.
Disclosure
11. Daffern M and Howells K. The prediction of imminent aggression and self- harm in personality disordered patients of a high security hospital using the HCR-20 clinical scale and the Dynamic Appraisal of Situational Aggression. Int J Forens Mental Health 2007; 6: 137–143.
The authors declare that there are no conflict of interest. The authors alone are responsible for the content and writing of the paper.
12. Quinsey V, Harris G, Rice M, et al. Violent offenders: Appraising and managing risk. 2nd ed. Washington, DC: American Psychological Association, 2006.
References 1. Daffern M, Mayer M and Martin T. A preliminary investigation into patterns of aggression in anAustralian forensic psychiatric hospital. J Forens Psychiatry Psychol 2003; 14: 67–84.
13. Almvik R, Woods P and Rasmussen K. The Brøset Violence Checklist: Sensitivity, specificity, and inter-rater reliability. J Interpers Viol 2000; 15: 1284–1296.
47 Downloaded from apy.sagepub.com at UNIV CALIFORNIA SAN DIEGO on June 9, 2015
Australasian Psychiatry 23(1) 14. Ogloff JRP and Daffern M. The dynamic appraisal of situational aggression: An instrument to assess risk for imminent aggression in psychiatric inpatients. Behav Sci Law 2006; 24: 799–813.
21. Van de Sande R, Nijman HL, Noorthoorn EO, et al. Aggression and seclusion on acute psychiatric wards: Effect of short-term risk assessment. Brit J Psychiatry 2011; 199: 473–478.
15. Barry-Walsh J, Daffern M, Duncan S, et al. The prediction of imminent aggression in patients with mental illness and/or intellectual disability using the Dynamic Appraisal of Situational Aggression instrument. Australas Psychiatry 2009; 17: 493–496.
22. Borum R, Bartel P and Forth AE. Manual for the Structured Assessment of Violence Risk in Youth (SAVRY) Version 1.1. Odessa, FL: Psychological Assessment Resources, 2003.
16. Daffern M, Howells K, Hamilton L, et al. The impact of structured risk assessments followed by management recommendations on aggression in patients with personality disorder. J Forens Psychiatry Psychol 2009; 20: 661–679. 17. Vojt G, Marshall LA and Thomson LDG. The assessment of imminent inpatient aggression: A validation study of the DASA-IV in Scotland. J Forensic Psychiatry Psychol 2010; 21: 789–800. 18. Griffith JJ, Daffern M and Godber T. Examination of the predictive validity of the Dynamic Appraisal of Situational Aggression in two mental health units. Int J Mental Health Nurs 2013; 22: 485–492. 19. Needham I, Abderhalden C, Meer R, et al. The effectiveness of two interventions in the management of patient violence in acute mental inpatient settings: Report on a pilot study. J Psychiatric Mental Health Nurs 2004; 11: 595–601. 20. Abderhalden C, Needham I, Dassen T, et al. Structured risk assessment and violence in acute psychiatric wards: Randomised controlled trial. Brit J Psychiatry 2008; 193: 44–50.
23. Chu CM, Hoo E, Daffern M, et al. Assessing the risk of imminent aggression in institutionalized youth offenders using the dynamic appraisal of situational aggression. J Forens Psychiatry Psychol 2012; 23: 168–183. 24. American Psychiatric Association. Diagnostic and statistical manual of mental disorders, text revision (DSM-IV-TR). 4th ed. Washington, DC: American Psychiatric Press; 2000. 25. Webster CD, Douglas KS, Eaves D, et al. HCR-20: Assessing risk for violence. Version 2 ed. British Columbia, Canada: Mental Health, Law and Policy Institute of Simon Fraser University, 1997. 26. Daffern M, Ogloff JRP, Chu CM, et al. The Dynamic Appraisal of Situational Aggression: Youth Version (DASA:YV). Victoria: Centre for Forensic Behavioural Science at Monash University, 2008. 27. Douglas KS and Webster CD. Predicting violence in mentally and personality disordered individuals. In: Roesch R, Hart SD and Ogloff JRP (eds) Psychology and law: The state of the discipline. New York: Plenum, 1999, pp. 175–239.
48 Downloaded from apy.sagepub.com at UNIV CALIFORNIA SAN DIEGO on June 9, 2015
