A research paper

ContentServer5.pdf

Home >Social Science homework help >A research paper

Evaluation of anger management groups in a high-security hospital

CLAIRE WILSON1, STACY GANDOLFI2, ALISON DUDLEY1, BRIAN THOMAS1, JAMES TAPP1 AND ESTELLE MOORE1, 1Centralised Groupwork Service, Newbury Therapy Unit, Broadmoor Hospital, Berkshire, UK; 2London Probation Trust, Sexual Offending Treatment Unit, Ilford, Essex, UK

ABSTRACT Background Anger management programmes for offenders typically aim to improve the management of emotion associated with aggressive and antisocial behaviour. Such programmes have been quite extensively evaluated in prison and probation settings, but there is less published research in forensic mental health settings. Aim This study aimed to evaluate anger management groups in a high-security hospital. Method Eighty-six patients were referred for a 20-session anger management intervention. Outcomes were self-reported experiences of anger and changes in institutionally docu- mented incidents of aggression. Incident rates were retrospectively reviewed for all group graduates, where data were available, including a comparison group of graduates who acted as their own controls. Results Group graduates reported sustained reductions in feelings of anger and positive changes in their use of aggression in reaction to provocation. Some reduction in incidents of physical aggression was noted when group completers were compared with non-completers. Incidents of verbal aggression were observed to increase for graduates. There was also a trend towards improvement when treated men were compared with men on the waiting list. Conclusions Our findings contribute to the growing evidence for anger management groups for aggressive men, although the low-base rate of incidents, typical of such a containing and therapeutic hospital setting, rendered the analysis of behavioural outcomes difficult. Implications for practice and research Anger management sessions for male forensic psychiatric patients can be effective in reducing incidents of physical aggression in response to provocation. Evaluation of treatments for anger is particularly difficult in secure and protective settings, where the aim is to keep incidents of actual physical aggression to a minimum. Further research of this kind is needed to test the value of self-reported reduction in angry feelings as an indicator of clinically useful progress. Copyright © 2013 John Wiley & Sons, Ltd.

Criminal Behaviour and Mental Health 23: 356–371 (2013) Published online 23 July 2013 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/cbm.1873

Introduction

The impact of effective anger regulation may be positive, but much depends upon the context in which the anger occurs (Novaco, 2007). It has been argued that expression of anger may improve assertiveness and self-control, and reduce somatic tension (Novaco, 1975; Novaco and Chemtob, 1998). In contrast, nega- tive functions of anger include interpersonal violence (Hollenhorst, 1998; Novaco et al., 2000), substance abuse (DeMoja and Spielberger, 1997), physical health problems (Helmers et al., 1994) and self-harm/suicide (Hillbrand et al., 1994). Anger is also often identified as a salient feature of different types of offending behaviours and therefore a key target for risk reduction (e.g. Groth and Birnbaum, 1979; Bradford, 1982; Rice et al., 1989; Blackburn, 1993).

Given the potential consequences of anger, causal and maintenance factors for it have been investigated to inform interventions for people who have anger dysregulation (Novaco et al., 2000). Accordingly, anger management programmes have become increasingly available within clinical settings. On the basis of research evidence, present day interventions typically follow a cognitive behavioural therapy approach. A wide range of techniques (e.g. cognitive reframing, relaxation techniques and skills training) are taught to help improve anger regulation through measured exposure to provocation and enhanced cognitive skills for processing thoughts of aggression, and provide coping skills for self-regulation and distress tolerance (Wright et al., 2009). Meta-analyses of outcome studies evaluating the effectiveness of such interventions in clinical settings have typically demonstrated a positive effect when treatment groups are compared with control groups (e.g. Beck and Fernandez, 1998; DiGiuseppe and Tafrate, 2003; Del Vecchio and O’Leary, 2004).

For programmes conducted specifically with offender populations, where anger dysregulation is often a core feature of the level of risk that warrants detention and intervention, findings have not been as readily comparable (Jones and Hollin, 2004). Evidence is less compelling from forensic evaluations, with mixed findings of both positive outcomes (Hughes, 1993; Marquis et al., 1996; Dowden et al., 1999; Ireland, 2004) and no reported changes in anger outcomes (Watt and Howells, 1999). A recent review of effective interventions concluded that emotional self-management, interpersonal skills and social problem solving approaches lead to positive changes in anger, but improvements in the evidence base are needed, in both quantity and quality (McGuire, 2008). This conclusion extends to more specialised services that provide care for individuals who present with complex mental health needs in addition to challenging aggressive behaviour. Offender patients in secure forensic inpatient services tend to have high criminogenic needs and persistent violent conduct (e.g. Monahan and Steadman, 1994). There is, however, much less evidence of effectiveness of anger management with this population than in custodial or general clinical settings (Leitner et al., 2006).

Evaluation of groupwork on the management of anger 357

Mental health needs of security hospital patients include personality and organic brain disorders (Blackburn, 2004), and these may impact both on the experience of anger and expression of aggression and capacity to make use of treatment. Originally perceived as an activator to aggression, illness pathology has been associated with the onset of violent behaviour (Taylor et al., 1993), although it has been argued that the relationship between mental illness and violence is more associative than causal (Arboleda-Florez et al., 1996). In principal, interpretations are viewed tentatively, and although this lack of clarity may impact upon methods of treatment, practitioners are still faced with the task of providing interventions to target the problem.

Studies of outcome of anger management programmes within psychiatric settings have provided an initial evidence base for the use of such programmes with male-offender patients (e.g. Stermac, 1986; Jones and Hollin, 2004; McMurran et al., 2001; Hilton and Frankel, 2003). Overall, positive changes in aggression have been reported after intervention, but limitations to the studies are apparent. Use of various operational definitions of the term anger, mixed with overlapping constructs such as hostility and aggression (Spielberger et al., 1995), contributes to the difficulties in comparing studies. Inconsistencies are also evident in the evaluation of anger change with varied use of a range of standardised and un-standardised outcome assessments. Observational methods to assess changes in the frequency and severity of aggressive behaviour are also typically absent (Watt and Howells, 1999).

Our research into anger management groups in a high-security hospital in England covers a 10-year period. In response to some of the methodological issues raised by existing outcome studies, behavioural data in the form of recorded incidents of aggression were investigated, and we included, as a comparison group, men who were waiting to start anger management groups. Individual indices of reliable and clinically important change were conducted to demonstrate variability in response to intervention, which are not accounted for through aggregate outcome evaluation methods (Jacobson and Truax, 1991). As the problems of attrition are considered a significant obstacle to the delivery of programmes (Siddle et al., 2003) and are associated with poorer outcomes (McMurran and Theodosi, 2007), outcomes for participants who dropped out of treatment were also investigated.

Method

Participants

Participants were 86 male-offender patients detained in a high-security psychiatric hospital under the Mental Health Act 1983 (HMSO, 1983) who had been referred for anger management groups over a period of 10 years.

358 Wilson et al.

Procedure

Participants were referred by their clinical teams for anger management. Suitability criteria for the group were evidence of anger management difficulties and violent behaviour, as indicated by a functional analysis of index offences and/or other offending behaviour. Functional analysis is a routine part of the psychological assessment of patients’ needs at this hospital. Participants not considered suitable for the intervention were those whose violent behaviour was not functionally assessed as related to anger, for example offending directly related to psychotic experience. Individual interviews were also conducted with referred patients to determine motivation and suitability for groupwork taking account of mental health (e.g. active symptoms) and risk (e.g. physical aggression). If an individual had an unstable mental state, their anger management referral was either closed or placed on hold and was reviewed periodically until such a time that they were able to attend.

The anger management group is based on the cognitive behavioural model (Novaco, 1975; Meichenbaum, 1997), adapted by Quayle and Shell (1999) for use in high security hospitals. It consists of four core modules: (i) Motivation: introducing key skills; (ii) Introducing key concepts: understanding anger; (iii) Coping with anger; and (iv) Final thoughts: putting it all together. The principal aim of the intervention is to identify personal cognitive, emotional and physical components of anger, the uses of anger and how to control or redirect it in an assertive rather than aggressive manner. The programme aims to help each group member achieve change through feedback within the group and using a range of techniques and materials to promote discussion. It includes exploration of relaxation through music and other exercises, perspective taking in role plays and relapse prevention. Individ- ual 1:1 sessions may also be offered, if required, to support group members outside the group setting.

Facilitators of the group included three to five staff working in multi- disciplinary teams drawn from nurse therapy, psychology and arts therapy disci- plines. Not all facilitators were present each week, but all worked as a team who met for supervision, allowing the group to run with minimal cancellations and maximum facilitator continuity. All staff had experience of group therapy and of working with people with anger difficulties. The contribution of the arts therapies included aiding group cohesion and creativity through music and role-playing of skills (e.g. drama-informed exercises). Sessions were once weekly, for 2 hours, with a total of 20 planned sessions (range 17–21). Delivery was varied only according to the required pace for group participants; all modules were covered in full. Group facilitator supervision was provided throughout on a fortnightly basis by an experienced colleague external to the facilitation team. This promoted intervention integrity based on the core modules and responsivity to the group members, and aided the technical aspects of the delivery of specific manualised techniques.

Evaluation of groupwork on the management of anger 359

Self-report measures of anger assessment were administered pre- and post- the group programme. Behavioural outcome data were obtained retrospectively from electronic incident report forms routinely completed by staff members (independent of the groupwork service and evaluators), recording any incidents of aggression, including self-injurious behaviour, at 3-month intervals for up to 9 months both prior to the groups and following their completion. Additional incident data were collated for participants who were referred for the group and allocated to a waiting list. Referrals to the group came at different times over the course of the years. They were allocated to the waiting list as soon as their referral was received, and collection of their individual incident data started at that point. These participants thus acted as a waiting list comparison group (for number of incidents) prior to engaging in the group. Because of organisational changes in incident reporting systems, incident data were available for only 48 of the total of 70 participants who completed the programme (i.e. there were no incident data for the 22 patients completing the first three programmes) and the 16 patients who dropped out, refused or left the group for other reasons.

Additional case file data on demographic, clinical and forensic characteristics of participants were collated, entered onto an electronic database and analysed using the Statistical Package for the Social Sciences, version 15 (SPSS v.15 IBM United Kingdom Limited, PO Box 41, North Harbour Portsmouth Hampshire, PO6 3AU).

The evaluation was registered, reviewed and approved by the West London Mental Health Trust Clinical Effectiveness and Audit department.

Outcome assessments

RAMAS Anger Assessment Profile (O’Rourke and Hammond, 2000) The RAMAS Anger Assessment Profile (RAAP) self-report questionnaire consists of 112 items, each measured across a five-point Likert scale that builds a profile of an individual’s anger in relation to their cognitive and physiological aspects and actions. It assesses 10 facets of anger: attitude, expression, provocation, somatic tension, duration, sensitivity, victimisation, assault, consequences and anger control. Internal consistency of the RAAP facets from a general population sample ranges from 0.75 to 0.91.

State-Trait Anger Expression Inventory-2 (Spielberger, 1999) The State-Trait Anger Expression Inventory-2 (STAXI-2) self-report measure consists of 57 items, each measured across a four-point Likert scale, concerned with the experience (state and trait anger), expression and control of anger. The measure has six scales: state anger, trait anger, anger expression-out, anger expression-in, anger control-out and anger control-in; five subscales; and an anger expression index, yielding an overall measure of expression and control of anger. Internal consistency of the

360 Wilson et al.

STAXI-2 scales for the normative adult and psychiatric patient samples ranged from 0.75 to 0.90.

Institutional incidents Institutional incidents are recorded on a centralised electronic incident report form. Incidents are recorded immediately after the event, and the report is countersigned to promote reliability, which has been demonstrated (k = 0.925) in previous research at the study site (Vernham et al., 2010). Incident categories are coded and defined into verbal aggression (the use of inappropriate words or behaviour causing distress and/or constituting harassment towards a person), physical aggression (the intentional application of force against a person without lawful justification, resulting in physical injury or personal discomfort) and self-harm (bodily self-injury that was not a suicide attempt).

Analyses

Descriptive analyses were conducted to report the characteristics of referred partic- ipants and to compare group members who completed the programme (completers) with those who did not (non-completers) because they refused to attend the group, elected to drop out or left the group because of poor attendance or deterioration in mental health. Differences between the completers and non-completers were studied across age, length of admission, baseline anger assessment scores and recorded aggressive incidents prior to the group to determine any attrition bias.

A comparison of means was conducted on each of the subscales of the STAXI-2 and RAAP to examine differences between pre-scores and post-scores, for those who completed the programme. To examine individual change for completers across the measures, indices of clinical and reliable change were calculated (for formulae, see Jacobson and Truax, 1991). Clinical change demonstrates a shift to a more ‘functional’ level of anger set by normative ranges of experience. Reliable change is reported to ensure this change occurs beyond chance variation or measurement error.

Mean rates of verbal, physical and self-harm incidents for group completers were compared at 3-month intervals, for up to 9 months both prior to and following completion of the group, to investigate changes in reported trends of incidents of aggression for those who completed. Incident rates were also compared between completers and non-completers, to investigate potential iatrogenic outcomes from intervention attrition. In addition, mean incident rates were compared between group completers and the waiting list comparison group to investigate potential benefits from anger management training compared with ‘treatment as usual’.

All data were screened for outliers and to determine the distribution of anger scores and incident rates necessary for parametric analyses. Where data were not normally distributed, log transformations of scores and rates were conducted to attempt to correct this; where this was not achieved, non-parametric analyses

Evaluation of groupwork on the management of anger 361

were conducted. Effect size estimates for reported pre-group to post-group changes are presented as the standardised mean difference, calculated using G Power 3 (Faul et al., 2009) and Cohen’s d conventions (0.2, small effect; 0.5, medium effect; 0.8, large effect) (Cohen, 1977). Where data were missing on outcome assessments, participant totals are reported.

Results

Ten group cycles were completed over a period of approximately 10years. Of the 86 participants referred, 4 eligible patients refused the programme; of the 82 patients who commenced, 9 chose to stop attending after one or two groups, some saying that they thought the group inappropriate for their needs and others that their medication made it difficult to concentrate and engage. Three other men failed to complete – one because he was discharged from the hospital, one due to a deterioration in his mental health, and one because his attendance was erratic; in both the later cases, discontinuation was negotiated. Thus, 70 men completed the programme with a group session attendance rate average of 85%.

There were no significant differences between completers and non-completers in terms of average length of admission or age (Table 1), whether or not refusers were included in the non-completion group. No significant differences were found on pre-group scores of the RAAP and STAXI-2 between completers and non- completers, reducing the risk of attrition bias, but only 6 of the 16 non-completers consented to and finished the pre-group assessments. Non-completers had a significantly higher rate of physically aggressive incidents compared with com- pleters at 4–6 months pre-group (Z = 2.438, p = 0.015), again whether or not refusers were included in this analysis. There was also significantly higher rates of pre-group self-harm for those who dropped out, at 4–6 months (Z = 2.191, p = 0.028) and 7–9 months (Z = 2.307, p = 0.021).

A comparison of means indicated significant changes across self-report assess- ments (Table 2). On the RAAP, completers reported significant improvement in their capacity to solve problems when experiencing anger (anger control), which remained significant after correcting for multiple comparisons. They also showed a reduction in the duration of feelings of anger (duration) and a significant decrease in acts of aggression (assault). On the STAXI-2, completers reported a significantly greater reduction in angry feelings experienced over time (trait anger) and a reduc- tion in disposition to be provoked into reacting aggressively (temperament). A reported increase in the inward control of angry feelings by ‘cooling off’ (aggression control-in) was also found. On average, this represented change from a level of function that is characteristic of psychiatric patients to one typical of a non-clinical population (Spielberger, 1999).

Individual clinical and reliable change on the RAAP (Table 2) was most marked on scales measuring intensity (somatic tension) and extent (duration) of

362 Wilson et al.

T ab le

1: P ar ti ci pa n t de sc ri pt iv e in fo rm

at io n in cl ud in g by

co m pl et er

st at us

D es cr ip ti ve

C at eg or y

C om

pl et er s

N = 70

(% )

N on

-c om

pl et er sa

N = 16

(% )

T ot al

re fe rr ed

N = 86

(% )

M ea n ag e at

st ar t of

gr ou p

Y ea rs [S D ]

36 .7

[1 0. 3]

36 .1

[6 .6 ]

36 .6

[9 .7 ]

M ea n le n gt h of

ad m is si on

at st ar t of

gr ou p

Y ea rs [S D ]

4. 0 [3 .3 ]

4. 6 [3 .9 ]

4. 1 [3 .4 ]

E th n ic it y

C au ca si an

42 (6 0. 0)

12 (7 5. 0)

54 (6 2. 8)

A fr o- C ar ib be an

15 (2 1. 4)

2 (1 2. 5)

17 (1 9. 8)

O th er /M

ix ed

H er it ag e

7 (1 0. 0)

1 (6 .3 )

8 (9 .3 )

N ot

gi ve n

6 (8 .6 )

1 (6 .3 )

7 (8 .1 )

IC D

di ag n os is

S ch iz op h re n ia , sc h iz ot yp al

an d

de lu si on

al di so rd er s

29 (4 1. 4)

12 (7 5. 0)

41 (4 7. 7)

M oo d (A

ff ec ti ve ) D is or de rs

2 (2 .9 )

0 2 (2 .3 )

D is or de rs of

ad ul t pe rs on

al it y an d

be h av io ur

25 (3 5. 7)

2 (1 2. 5)

27 (3 1. 4)

D is or de rs of

ps yc h ol og ic al

de ve lo pm

en t

2 (2 .9 )

1 (6 .3 )

3 (3 .5 )

N ot

R ep or te d

12 (1 7. 1)

1 (7 .7 )

13 (1 5. 1)

In de x of fe n ce

H om

ic id e

28 (4 0. 0)

4 (2 5. 0)

32 (3 7. 2)

O th er

vi ol en ce

18 (2 5. 7)

10 (6 2. 5)

28 (3 2. 6)

S ex ua l of fe n ce s

14 (2 0. 0)

1 (6 .3 )

15 (1 7. 4)

A rs on

4 (5 .7 )

0 4 (4 .7 )

O th er s (e .g . ro bb er y, ki dn

ap pi n g)

6 (8 .6 )

0 6 (7 .0 )

N on

e 0

1 (6 .3 )

1 (1 .2 )

a I n cl ud es

in di vi du al

w h o re fu se d to

at te n d th e gr ou

p. S D

de n ot es

st an da rd

de vi at io n .

Evaluation of groupwork on the management of anger 363

T ab le

2: G ro up

an d in di vi du al

pr e- gr ou p to

po st -g ro up

ch an ge

in se lf -r ep or t as se ss m en t

S ca le

P re -g ro up

(S D )a

P os t- gr ou p

(S D )

T es t

st at is ti cb

p- va lu ec

E ff ec t si ze

es ti m at es

(d )

C li n ic al

ch an ge

N /( % )

R el ia bl e

ch an ge

N /( % )

C li n ic al

ch an ge /

R el ia bl e ch an ge

R A A P

A tt it ud e (N

= 55 )

33 .7 2 (5 .9 5)

33 .8 1 (4 .9 2)

�0 .1 09

0. 91 4

0. 01

13 (2 3. 6)

7 (1 2. 7)

0 E xp re ss io n (N

= 54 )

26 .4 6 (8 .6 4)

25 .8 3 (7 .3 9)

0. 49 4

0. 62 3

0. 07

7 (1 2. 7)

5 (9 .1 )

P ro vo ca ti on

(N = 52 )

34 .4 2 (1 1. 05 )

33 .6 5 (9 .3 9)

0. 61 1

0. 54 4

0. 07

8 (1 5. 3)

11 (2 1. 1)

5 (9 .6 )

S om

at ic

te n si on

(N = 54 )

25 .8 5 (7 .4 6)

27 .3 1 (6 .5 6)

�1 .4 93

0. 14 1

0. 20

13 (2 4. 0)

12 (2 2. 2)

9 (1 6. 6)

D ur at io n (N

= 54 )

25 .9 8 (9 .1 3)

22 .7 5 (8 .0 7)

2. 46 4*

0 .0 1 7

0 .3 3

15 (2 7. 7)

16 (2 9. 6)

13 (2 4. 0)

S en si ti vi ty

(N = 51 )

30 .7 3 (9 .4 1)

30 .6 4 (7 .3 9)

�0 .1 32

0. 89 5

0. 01

8 (1 5. 6)

9 (1 7. 6)

6 (1 1. 7)

V ic ti m is at io n

(N = 53 )

33 .3 9 (8 .1 3)

35 .1 8 (8 .1 1)

�1 .6 44

0. 10 6

0. 22

7 (1 3. 0)

3 (5 .6 )

1 (1 .9 )

A ss au lt (N

= 53 )

34 .7 9 (8 .6 7)

32 .7 4 (7 .0 4)

�2 .8 23

** 0 .0 0 5

0 .3 3

12 (2 2. 2)

2 (3 .7 )

0 C on

se qu en ce s

(N = 52 )

31 .3 0 (9 .5 7)

29 .9 6 (9 .4 1)

1. 28 3

0. 20 5

0. 14

7 (1 3. 4)

6 (1 1. 5)

4 (7 .6 )

A n ge r co n tr ol

(N = 53 )

30 .3 9 (7 .6 0)

33 .9 4 (8 .2 9)

�3 .7 86 **

0 .0 0 0

0 .5 3

13 (2 4. 5)

10 (1 8. 8)

7 (1 3. 2)

S T A X I- 2

S ta te

an ge r (N

= 52 )

17 .9 4 (6 .7 6)

17 .7 0 (7 .8 5)

�1 .2 56

0. 20 9

0. 03

7 (1 3. 4)

5 (9 .6 )

A n ge r fe el in gs

(N = 52 )

6. 21

(2 .2 2)

6. 06

(2 .4 5)

�1 .0 16

0. 31 0

0. 06

8 (1 5. 3)

2 (3 .8 )

1 (1 .9 )

V er ba l an ge r

(N = 53 )

6. 15

(2 .7 5)

6. 00

(3 .1 4)

�0 .7 61

0. 44 7

0. 05

9 (1 6. 9)

3 (5 .6 )

2 (3 .7 )

P h ys ic al

an ge r

(N = 52 )

5. 60

(2 .4 2)

5. 63

(2 .6 2)

�0 .0 85

0. 93 2

0. 01

2 (3 .8 )

1 (1 .9 )

T ra it an ge r (N

= 52 )

19 .3 7 (6 .4 3)

17 .0 0 (5 .6 2)

�2 .2 21 *

0 .0 2 6

0 .3 6

11 (2 1. 1)

10 (1 9. 2)

6 (1 1. 5)

(C on tin ue s)

364 Wilson et al.

T em

pe ra m en t

(N = 53 )

7. 47

(3 .1 1)

6. 24

(2 .7 2)

�2 .1 12

* 0 .0 3 5

0 .3 9

14 (2 6. 4)

6 (1 1. 3)

5 (9 .4 )

R ea ct io n (N

= 52 )

8. 03

(2 .8 0)

7. 27

(2 .4 4)

�1 .7 93

0. 07 3

0. 28

10 (1 9. 2)

4 (7 .6 )

3 (5 .7 )

E xp re ss io n -o ut

(N = 53 )

15 .9 6 (5 .2 5)

15 .1 0 (4 .1 0)

�0 .0 45

0. 96 4

0. 17

8 (1 5. 0)

4 (7 .5 )

3 (5 .6 )

E xp re ss io n -i n

(N = 51 )

18 .0 8 (5 .5 9)

16 .6 0 (5 .2 8)

1. 40 0

0. 16 8

0. 27

14 (2 7. 4)

13 (2 5. 4)

11 (2 1. 5)

A n ge r co n tr ol -o ut

(N = 52 )

22 .7 8 (5 .1 8)

22 .2 9 (5 .1 4)

0. 59 8

0. 55 3

0. 09

8 (1 5. 4)

7 (1 3. 5)

A n ge r co n tr ol -i n

(N = 50 )

21 .4 4 (6 .0 7)

23 .4 2 (5 .3 7)

�2 .6 7 8 *

0 .0 1 0

0 .3 3

10 (2 0. 0)

7 (1 4. 0)

7 (8 .0 )

A n ge r ex pr es si on

in de x (N

= 49 )

37 .1 2 (1 4. 08 )

34 .2 5 (1 3. 54 )

1. 28 1

0. 20 6

0. 20

12 (2 4. 5)

9 (1 8. 4)

*p < 0. 05 .* *p

< 0. 01 .

a S D

de n ot es

st an da rd

de vi at io n .

b It al ic is ed

te st st at is ti cs

de n ot es

W il co xo n Z st at is ti c.

c C ri te ri on

fo r si gn ifi ca n ce

le ve l us in g th e B on

fe rr on

i co rr ec ti on

m et h od

se t at

p < 0. 00 2.

S T A X I- 2,

S ta te -T ra it A n ge r E xp re ss io n In ve n to ry -2 ; R A A P , R A M A S A n ge r A ss es sm

en t P ro fi le .

T ab le

2: C on tin ue d

S ca le

P re -g ro up

(S D )a

P os t- gr ou

p (S D )

T es t

st at is ti cb

p- va lu ec

E ff ec t si ze

es ti m at es

(d )

C li n ic al

ch an ge

N /( % )

R el ia bl e

ch an ge

N /( % )

C li n ic al

ch an ge /

R el ia bl e ch an ge

Evaluation of groupwork on the management of anger 365

feelings of anger, with improvements in skills to manage it (anger control). On the STAXI-2, significant change is reported on the inward expression of aggression (ex- pression-in) and the experience of anger over time (trait anger). In contrast, the least reported change was on scales assessing attitudes and perceived consequences of aggressive behaviour as well as more reactive and state-based experiences.

There were generally low frequencies of verbal and physical aggression and self- harm (Table 3), and, for completers, there were no significant changes in post- treatment incident rates (measured at 3-month intervals) compared with reported trends prior to the anger management programme. Incidents of verbal aggression increased marginally at each interval following the group, when compared with pre-group incidents, but particularly in the first 3 months. In contrast, rates of physical aggression were reduced a little at 1–3 months (d = 0.12) and more at 7–9 months (d = 0.25) after the group programme compared with before it. A positive trend was also reported for self-harm incidents at 1–3 months post-group, but not after that.

Participants who completed the group had significantly fewer incidents of physical aggression at 7–9 months follow-up (Z = �2.140, p = 0.032, d = 0.7) and self-harm incidents at 4–6 months follow-up (Z = 2.550, p = 0.011) when compared with the post-group incident trends for non-completers; yet, there had been no such differences prior to the group commencing. This significant difference held when comparing completers with those who dropped out of the group (excluding refusers), although this trend was also observable prior to the group commencing. Incidents of verbal aggression were more frequent among participants who dropped out than among completers at 7–9 months (Z = 2.240, p = 0.025, d = 0.33), as were physically aggressive incidents at 1–3 months (Z = 2.338, p = 0.019, d = 0.34) post-group. These were not noted prior to the intervention.

No statistically significant differences for reported incidents rates were found when comparing completers with the waiting list comparison group (Table 3). Completers had fewer incidents of physical aggression at 7–9 months than the waiting list comparison group (d = 0.20). No other comparative improvements were observed.

Discussion

Offender patients in a high-security hospital who complete anger treatment groups report reductions in prolonged feelings of anger and better coping strategies for anger control. This improvement may have been linked to an increased awareness of predisposing and maintaining factors, such as increased physiological arousal. On the basis of these findings, it might be argued that the programme was effective in providing participants with the necessary skills to both identify and manage feelings of anger. At the more immediate stage of experience (state aggression) and poten- tially more critical, changes in anger expression did not seem to be indicated, and

366 Wilson et al.

T ab le

3: A ve ra ge

in ci de n ts fo r co m pl et er s an d w ai ti n g li st co m pa ri so n sa m pl e ac ro ss 3- m on

th in te rv al s pr io r to

th e gr ou p co m m en ci n g an d fo ll ow

-u p at

3- m on

th in te rv al s fo ll ow

in g co m pl et io n of

th e gr ou p

In ci de n ts

1– 3 m on

th s

4– 6 m on

th s

7– 9 m on

th s

P re

(S D )

P os t (S D )

P re

(S D )

P os t (S D )

P re

(S D )

P os t (S D )

C om

pl et er s (n

= 48 )

V er ba l ag gr es si on

0. 37

0. 70

0. 29

0. 35

0. 31

0. 35

(0 .9 5)

(2 .5 0)

(0 .8 9)

(1 .3 4)

(0 .7 7)

(2 .1 6)

P h ys ic al

ag gr es si on

0. 16

0. 06

0. 10

0. 39

0. 04

(0 .7 8)

(0 .2 4)

(0 .5 9)

(0 .3 7)

(1 .3 9)

(0 .2 0)

S el f- h ar m

0. 41

0. 00

0. 02

(0 .2 0)

– –

– (0 .1 4)

(0 .1 4)

W ai ti n g li st gr ou p (n

= 64 )

V er ba l ag gr es si on

0. 49

0. 57

0. 63

0. 61

0. 44

0. 74

(1 .2 0)

(1 .6 8)

(1 .9 2)

(1 .8 7)

(1 .9 9)

(2 .1 0)

P h ys ic al

ag gr es si on

0. 33

0. 22

0. 20

0. 25

0. 12

0. 11

(1 .2 8)

(0 .7 5)

(0 .6 7)

(0 .8 0)

(0 .4 2)

(0 .4 4)

S el f- h ar m

0. 03

0. 07

0. 15

0. 03

0. 11

0. 03

(0 .2 5)

(0 .4 1)

(1 .0 3)

(0 .1 7)

(0 .4 7)

(0 .1 7)

S D

de n ot es

st an da rd

de vi at io n .

Evaluation of groupwork on the management of anger 367

no change in attitude toward the function of anger or the consequences of the anger was observed. It may therefore be beneficial to investigate further factors that influence the expression of immediate feelings of anger compared with those that inhibit aggressive behaviour, which can then be diffused over time. A common but not always prevalent characteristic of offenders/offender patients is the commis- sion of impulsive aggressive acts (Barratt et al., 1997). As we did not include a measure of impulsivity in this study, it is difficult to say whether this may account for the lack of impact on this type of aggressive behaviour.

Consistent with group level comparisons, a higher proportion of individuals demonstrated both clinical and reliable change on scales measuring long-term feelings of anger and methods of expression. The heterogeneity among forensic populations, such as that studied here, is likely to relate to the experience and expression of anger. Consequently, although change is measured in a dedicated ‘direction’, understanding it within the context of each individual will most reliably indicate outcome. An apparent increase in self-reported anger, for example, may represent the improvement of insight and a willingness to report angry feelings, rather than an actual observed increase in anger.

Improvement in rates of incidents of physical aggression was identified for those who completed groups, but effect sizes were small and changes did not reach statis- tical significance. An increase in incidents of verbal aggression for completers immediately following the groups may be explained as an extinction side-effect (Lerman et al., 1999), anger becoming expressed verbally rather than physically. Of note, non-completers were reported to have higher incident rates than completers after the group, most pronounced for those who dropped out of the groups. This is consistent with previous findings (McMurran and Theodosi, 2007).

The use of self-report assessment within a forensic context cannot be ignored when interpreting findings as evidence of impression management on anger assess- ments has been demonstrated (e.g. McEwan et al., 2009). The risk of a type I error occurring with multiple testing is acknowledged, as only one of the reported changes met the corrected criterion for significance testing. The consistency in findings across the RAAP and STAXI-2, however, weighs against this conservative explanation.

As a behavioural indicator of change from reported incidents might circumvent the issue of more subjective measures, use of incident data was limited by the low- base rate of incidents and by only a small proportion of participants accounting for the majority of incidents. This is a recognised phenomenon in institutional aggres- sion studies (e.g. Gudjonsson et al., 1999). The low incident rate possibly reflects the motivation and general stability of participants, where engagement with a group is less likely to be associated with frequent incidents of aggression outside the group itself. Those with a higher number of incidents are less likely to be referred until their difficulties have stabilised. Furthermore, not all types of incidents (e.g. damage to property) were incorporated into the evaluation. Trends should therefore be interpreted cautiously. Equally, use of incident reports can only provide an outside

368 Wilson et al.

view of an individual’s challenging behaviour and not capture instances where someone has appropriately asserted himself or herself and thereby avoided interper- sonal confrontation. The absence of incident data for 22 group completers also needs to be acknowledged as a potential bias; however, comparing those with and without incident data on pre-group self-reported assessments no significant differences were found. As with self-reported outcomes, the problem of multiple significance testing is acknowledged.

It is unlikely that in a generally therapeutic setting, any group treatment will show substantial differences from treatment as usual in terms of its impact on aggres- sive incidents, as the whole milieu is designed to reduce these during residence; however, future outcome research might aim to incorporate additional measures of change to evaluate the impact of anger management programmes more effec- tively. It is rare, for example, in such studies, to measure more directly positive outcomes rather than an absence of negative ones.

Acknowledgements

We would like to thank Sara Spence and Andy Burton at Broadmoor Hospital for the support in collating institutional incident data from the centralised elec- tronic Incident Report (IR1) form.

References

Arboleda-Florez J, Holley H, Crisanti A (1996) Final Report to Health Canada: Mental Illness and Violence: Proof or Stereotype. Calgary: World Health Organization Collaborating Centre for Research and Training in Mental Health.

Barratt ES, Stanford MD, Kent TA, Felthous AR (1997) Neuropsychological and cognitive psychophysiological substrates of impulsive aggression. Biological Psychiatry 47: 1045–1061.

Beck R, Fernandez E (1998) Cognitive-behavioral therapy in the treatment of anger: A meta- analysis. Cognitive Therapy and Research 22: 63–74.

Blackburn R (1993) The psychology of criminal conduct: Theory, research and practice. West Sussex: John Wiley & Sons Ltd.

Blackburn R (2004) What works with mentally disordered offenders. Psychology, Crime and Law 10: 297–308.

Bradford JMW (1982) Arson: A clinical study. Canadian Journal of Psychiatry 27: 188–193. Cohen J (1977) Statistical power analysis for the behavioral sciences (revised ed.). New York:

Academic Press. DeMoja CA, Spielberger CD (1997) Anger and drug addiction. Psychological Reports 81: 152–154. Del Vecchio T, O’Leary KD (2004) The effectiveness of anger treatments for specific anger problems:

A meta-analytic review. Clinical Psychology Review 24: 15–34. DiGiuseppe R, Tafrate RC (2003) Anger treatments for adults: A meta-analytic review. Clinical

Psychology: Science and Practice 10: 70–84. Dowden C, Blanchette K, Serin R (1999) Anger Management Programming for Federal Male

Offenders: An Effective Intervention. Research Branch, Correctional Services of Canada, Ontario.

Evaluation of groupwork on the management of anger 369

Faul F, Erdfelder E, Buchner A, Lang AG (2009) Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behavior Research Methods 41: 1149–1160.

Groth N, Birnbaum J (1979) Men Who Rape. New York: Plenum. Gudjonsson GH, Rabe-Hesketh S, Wilson C (1999) Violent incidents on a medium secure unit

over a 17-year period. Journal of Forensic Psychiatry 10: 249–263. Helmers KF, Posluszny DM, Krantz DS (1994) Associations of hostility and coronary heart disease:

A review of studies. In Siegmen A, Smith T (eds) Anger, Hostility, and the Heart. Hillsdale, NJ: Erlbaum pp. 67–96.

Hillbrand M, Krystal JH, Sharpe KS, Foster HG (1994) Clinical predictors of self-mutilation in hospitalized forensic patients. The Journal of Nervous and Mental Diseases 182: 9–13.

Hilton N, Frankel A (2003) Therapeutic value of anger management programmes in a forensic setting. British Journal of Forensic Practice 5: 8–15.

HMSO (1983) The Mental Health Act 1983, c12. London: HMSO. Hollenhorst PS (1998) What do we know about anger management programs in corrections?

Federal Probation 62: 52–65. Hughes GV (1993) Anger management program outcomes. Research on Offender Programming Issues

5: 5–9. Ireland JL (2004) Anger management therapy with young male offenders: An evaluation of treatment

outcome. Aggressive Behavior 30: 174–185. Jacobson NS, Truax P (1991) Clinical Significance: A statistical approach to defining meaningful

change in psychotherapy research. Journal of Consulting and Clinical Psychology 59: 12–19. Jones D, Hollin CR (2004) Managing problematic anger: The development of a treatment program

for personality disordered offenders in high security. International Journal of Forensic Mental Health 3: 197–210.

Leitner M, Jones S, Barr W, Whittington R, McGuire J (2006) Systematic review of prevention and intervention strategies for populations at high risk of engaging in violent behaviour. Final project report to the NHS National Forensic Mental Health R&D Programme.

Lerman DC, Iwata BA, Wallace MD (1999) Side effects of extinction: Prevalence of bursting and aggression during the treatment of self-injurious behavior. Journal of Applied Behavior Analysis 32: 1–8.

Marquis HA, Bourgon GA, Armstrong B, Pfaff J (1996) Reducing recidivism through institutional treatment programs. Forum on Corrections Research 8: 3–5.

McEwan TE, Davis MR, MacKenzie R, Mullen PE (2009) The effects of social desirability bias on STAXI-2 profiles in a clinical forensic sample. British Journal of Clinical Psychology 48: 431–436.

McGuire J (2008) A review of effective interventions for reducing aggression and violence. Philosophical Transactions of the Royal Society B 363: 2577–2597.

McMurran M, Charlesworth P, Duggan C, McCarthy L (2001) Controlling angry aggression: a pilot group intervention with personality disordered offenders. Behavioural and Cognitive Psychother- apy 29: 473–783.

McMurran M, Theodosi E (2007) Is treatment non-completion associated with increased reconviction over no treatment? Psychology, Crime and Law 13: 333–343.

Meichenbaum D (1997) Cognitive Behavior Modification. New York: Plenum Press. Monahan J, Steadman H (eds) (1994) Violence and Mental Disorder: Developments in Risk

Assessment. Chicago: University of Chicago Press. Novaco RW (1975) Anger Control: The Development and Evaluation of an Experimental Treatment.

Lexington MA: D. C. Heath. Novaco RW, Chemtob CM (1998) Anger and trauma: Conceptualization, assessment, and

treatment. In Follette VM, Ruzek JI, Abueg F (eds) Cognitive-Behavioral Therapies for Trauma. New York: Guilford pp. 162–190.

370 Wilson et al.

Novaco RW, Ramm M, Black L (2000) Anger treatment with offenders. In Hollin CR (ed.) Handbook of Offender Assessment and Treatment. West Sussex: John Wiley & Sons Ltd. pp. 281–296.

Novaco RW (2007) Anger dysregulation. In Cavell T, Malcolm K (eds) Anger, Aggression, and Interventions for Interpersonal Violence. Mahwah, NJ: Erlbaum pp. 3–54.

O’Rourke M, Hammond S (2000) The RAAP Anger Assessment Profile Psychometric Norms and Professional Manual. London. ISBN 1-903165-02-4

Quayle M, Shell Y (1999) Anger Management Programme-Group Sessions and Handouts. Psychology Department: Broadmoor Hospital, Berkshire.

Rice ME, Harris GT, Varney GW, Quinsey VL (1989) Violence in Institutions: Understanding, Prevention and Control. Toronto: Hans Huber.

Siddle R, Jones F, Awenat F (2003) Group cognitive behaviour therapy for anger: a pilot study. Behavioural and Cognitive Psychotherapy 31: 69–83.

Spielberger CD (1999) State-Trait Anger Expression Inventory-2 (STAXI-2). Odessa, FL: Psychological Assessment Resources.

Spielberger CD, Reheiser EC, Sydeman SJ (1995) Measuring the experience, expression and control of anger. In Kassinove H (ed.) Anger Disorders: Definitions, Diagnosis, and Treatment. Washington, DC: Taylor & Francis pp. 49–67.

Stermac LE (1986) Anger control treatment for forensic patients. Journal of Interpersonal Violence 1: 446–457.

Taylor PJ, Mullen P, Wessely S (1993) Psychosis, violence and crime. In Gunn J, Taylor P (eds) Forensic Psychiatry: Clinical, Legal and Ethical Issues. Oxford, UK: Butterworth-Heinemann pp. 329–371.

Vernham Z, Tapp J, Moore E (2010) Observer ratings of interpersonal behaviour as predoictors of aggression and self-harm in a high security sample of male mentally disordered offenders. Unpublished MSc Dissertation. University of Surrey.

Watt BD, Howells K (1999) Skills training for aggression control: Evaluation of an anger management programme for violent offenders. Legal and Criminological Psychology 4: 285–300.

Wright S, Day A, Howells K (2009) Mindfulness and the treatment of anger problems. Aggression and Violent Behavior 14: 396–401.

Address correspondence to: Claire Wilson, Centralised Groupwork Service, Newbury Therapy Unit, Broadmoor Hospital, Crowthorne, Berkshire, RG45 7EG, UK. Email: [email protected]

Evaluation of groupwork on the management of anger 371

Copyright of Criminal Behaviour & Mental Health is the property of John Wiley & Sons, Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.