Behavioural and Cognitive Psychotherapy, 2009, 37, 511–526
Recovery After Psychosis (RAP): A Compassion Focused
Programme for Individuals Residing in High Security Settings
Heather Laithwaite, Martin O’Hanlon, Padraig Collins, Patrick Doyle,
Lucy Abraham and Shauneen Porter
The State Hospital, Carstairs, Scotland
University of Glasgow, Scotland
Background: The aim of the study was to evaluate the effectiveness of a recovery group
intervention based on compassionate mind training, for individuals with psychosis. In
particular, the objective was to improve depression, to develop compassion towards self, and to
promote help seeking. Method: A within-subjects design was used. Participants were assessed
at the start of group, mid-group (5 weeks), the end of the programme and at 6 week follow-up.
Three group programmes were run over the course of a year. Nineteen participants commenced
the intervention and 18 completed the programme. Results: Signiﬁcant improvements were
found on the Social Comparison Scale; the Beck Depression Inventory; Other As Shamer
Scale; the Rosenberg Self-Esteem Inventory and the General Psychopathology Scale from the
Positive and Negative Syndrome Scale. Conclusions: The results provide initial indications
of the effectiveness of a group intervention based on the principles of compassionate focused
therapy for this population. The ﬁndings of this study, alongside implications of further research
Keywords: Psychosis, compassionate mind training, recovery, offenders.
In social mentality theory (Gilbert, 1989, 2001, 2005) the interplay in social situations
between emotional, motivational, cognitive, and behavioural processes is conceptualized as
reﬂectioning underlying evolutionary derived systems that shape relationships between the
self and others. Social mentalities are implicated in care-giving, care-eliciting, formation of
interpersonal alliances, social rank and sexual behaviour. They have a critical role in appraising
threat, enhancing safeness, and in regulating the affect associated with these fundamental
evolutionary challenges (MacBeth, Schwannauer and Gumley, 2008). According to whether
the environment is threatening or safe, all organisms must co-ordinate a range of internal
Reprint requests to Heather Laithwaite, Rowanbank Clinic, Balornock, Glasgow G21 3UL, Scotland. E-mail:
© 2009 British Association for Behavioural and Cognitive Psychotherapies
512 H. Laithwaite et al.
processes in order to pursue goals, enact strategies and co-create social roles (Buss, 2003;
Gilbert, 1989, 1992). Whether environments are threatening or safe, humans have (often
rapid) access to an evolved menu or suite of strategic responses (ways of attending, feeling,
behaving and thinking) to aid adaptive responding (Gilbert, 2005).
Social mentality theory refers to the development of the “human warmth syndrome”
whereby human beings develop, through secure attachments with primary care givers, the
ability to have compassion towards themselves and others. A secure attachment facilitates
the development of internal working models of others as “safe, helpful and supportive”.
The internalization of this helps the individual to develop self-soothing and compassionate
behaviours towards themselves and others. This activates the safe(ness) social mentality.
The threat-defence mentality is activated in situations of perceived and actual threat. For
example, social rank may provide a source of threat, whereby dominant individuals will issue
commands and hold power, whilst subordinates will take those commands and be submissive.
Social mentality theory states that the role relationships that exist between people can also
exist within people and arise from internal working models of early relationships. Therefore,
human beings can internalize the voice of a critical other and develop a submissive/subordinate
response to this. This model can help to explain the occurrence of command hallucinations.
It has been demonstrated that people who experience auditory hallucinations often relate
to them as though they were relating to real external others. In particular, the voices are
commonly experienced as malevolent, derogating, shaming and self-critical (Legg and Gilbert,
Developmental theory helps us understand the impact of early attachments on adult
psychopathology and hence the development of safe(ness) or threat focused social mentalities.
Previous research shows that early attachment experiences inﬂuence the ability to develop safe
and secure adult relationships (Bowlby, 1988). Gilbert (2004) refers to two consequences that
result when parents are unable to create (and stimulate) safeness, are threatening or shaming,
and do not convey warmth. First the “under-stimulation” of positive affect and warmth systems;
and second, the child is more likely to be “threat focused”, seeing others as a source of threat.
Subsequently, they are more social rank focused, especially on the power of others to control,
hurt or reject them. Sloman (2000) and Sloman, Gilbert and Hasey (2003) have shown that
those who have not been able to internalize a sense of warmth (able to stimulate positive affect
in the mind of others) and who feel unloved by others, can set out on quests to earn their place,
becoming excessively seeking, competitive and sensitive to rejection (Gilbert, 2004).
People with psychosis who also commit offences often come from backgrounds that reduce
the safe(ness) mentality and result in an activation of the threat focused mentality. Read
et al. (2004) have shown that the very high incidence of childhood trauma (emotional, sexual
and physical abuse or neglect) and a diagnosis of schizophrenia is not attributable to chance.
Experiences of bullying, shame, and other humiliation experiences (Bebbington et al., 2004;
Campbell and Morrison, 2007) trauma and loss (Romme and Escher, 1989) are also associated
with increased risk of developing psychosis. Such traumatic life experiences can lead to
the collapse and disorganization of attachment characterized by impaired mentalization and
theory of mind, fragmentation, dissociation and segmentation of episodic memories; and
use of competing and inconsistent coping responses (Liotti and Gumley, 2008; Read
and Gumley, 2008). Such early experiences may compromise the development of inner warmth.
We know that many people who have psychosis and who have also offended have had such life
experiences (Boswell, 1996; Fonagy et al., 1997) and we understand that this has an impact
Recovery After Psychosis (RAP) 513
on attachment organization and increases propensity for a threat focused social mentality or
“paranoid mind” (Gumley and Schwannauer, 2006).
The potential importance of developing inner warmth came from observations that some high
self-critics could understand the logic of cognitive behavioural therapy, and could generate
alternative thoughts to self-criticism, but rarely felt reassured by such efforts (Lee, 2005).
Similar observations were made when a self-esteem programme was piloted with a group of
patients with psychosis in a high security hospital (Laithwaite and Gumley, 2007). The ﬁndings
of this preliminary study were encouraging and demonstrated an improvement in self-esteem,
and depression. A noticeable change in positive symptomatology was not evident, due to
most participants being remitted of their positive symptoms prior to the group commencing.
Furthermore, participants in the group spoke about their early adverse experiences and how
this contributed to the development of low self-esteem. However, it was clear that many
participants were able to challenge their self-criticism on an “intellectual level” but continued
nevertheless to report feelings of worthlessness and low self-esteem.
The participants in both the above studies (Lee, 2005; Laithwaite and Gumley, 2007) came
from traumatized backgrounds. It is postulated by Gilbert (2004) that individuals with such
experiences are compromised in their ability to generate a model of compassion, and hence
the ability to self-soothe. Further studies have demonstrated that a lack of self-compassion is
associated with increased vulnerability to a number of indicators of psychopathology (Neff,
2003a). We know this is relevant because compassion helps to tap into safeness mode, which
helps to regulate affect. This is signiﬁcant with regards to relapse and recovery after psychosis
as a key aspect in relapse is high levels of emotional distress and affective dysregulation in
the period before, during and following the acute phase of psychosis. For example, ﬁndings
from retrospective and prospective studies have shown that the most commonly reported early
signs of relapse are fearfulness, anxiety, poor sleep, irritability, tension, depression and social
withdrawal (Herz and Melville, 1980; McCandless-Glimcher et al., 1986; Birchwood, Hallett
and Preston, 1989). In terms of recovery, studies by Birchwood, Mason, MacMillan and Healy
(1993) and Rooke, Birchwood and Iqbal (1998) have shown that patients with depression
following an acute psychotic phase were more likely to have experienced more compulsory
admissions and loss of, or drop in, employment status. Gilbert formulates this according to
social rank theory, whereby schizophrenia is a major life event that leads to signiﬁcant loss in
social status and role in society. Those who experience post-psychotic depression may indeed
have greater insight into such losses and fear subsequent relapse for this reason.
Gilbert and colleagues (Gilbert, 1992, 1997, 2000; Gilbert and Irons, 2005) have developed
compassionate mind training (CMT) to help people develop compassion and the ability to
self-soothe, regulate affect and hence provide an antidote to the threat mode. This model
is based on the premise that self-criticism is signiﬁcantly associated with shame-proneness
and that self-criticism is associated with lifetime risk of depression (Murphy et al., 2002).
CMT proposes that some people have not had the opportunity to develop their abilities to
understand sources of their distress, be gentle and self-soothing in the context of set-backs and
disappointments, but are highly (internally and externally) threat focused and sensitive. CMT
seeks to change an internalized dominating-attacking style that elicits a submissive response
to one that elicits a caring and compassionate response.
There is a poverty of published research carried out into people with psychosis in forensic
clinical settings. This is despite the fact that this is a population with complex and long-
term needs. This population has generally experienced past trauma; poor relationships with
514 H. Laithwaite et al.
signiﬁcant others, disrupted attachment histories and has the double stigma of experiencing
severe mental health problems and being offenders (Laithwaite et al., 2007; Boswell, 1996;
Fonagy et al., 1997). Recovery in this population is not just about reduction of symptoms
or distress, but reduction/management of risk of violent offending. It is therefore important
that therapies that have been researched in general mental health settings are adapted and
piloted with this population. A recovery programme that draws on CMT is attractive as it has
a developmental perspective that focuses on the effect of disrupted attachment histories on
the current functioning of the individual and their ability to respond to self-criticism, self-
soothe, and modify distress. Hence a programme that focuses on developing a compassionate
understanding of those vulnerabilities may promote recovery and help those seeking safety
strategies, which in turn may reduce the risk of violent re-offending.
The aim of this group intervention was to evaluate the speciﬁc aims of the Recovery After
Psychosis Programme. The aims of this programme were:
◦To improve depression
◦To improve self-esteem
◦To develop compassion towards self
◦To improve social comparison and to reduce external shame
A within-subjects design was used. Participants were assessed at the start of group, mid-group
(5 weeks), the end of the programme, and at 6-week follow-up.
The State Hospital is the maximum-security hospital for Scotland and Northern Ireland and
provides treatment and care in conditions of special security for individuals with mental
disorder who, because of their dangerous, violent or criminal propensities, cannot be cared for
in any other setting (The State Hospitals Board for Scotland, 2005). There are 11 wards covering
admissions, rehabilitation and continuing care. Patients in the hospital and participants in the
study are familiar with being assessed on a regular basis by health professionals who are
vigilant to issues of risk and mental health.
Participants were considered eligible for the group if they had a primary diagnosis of
schizophrenia, schizo-affective disorder or bi-polar affective disorder (those with bi-polar
affective disorder had a history of psychotic features). Potentially eligible participants were
excluded from the study if they had an organic illness, severe intellectual disability, and were
not able to provide informed consent. Participants were also excluded if they were involved
Recovery After Psychosis (RAP) 515
in other research. All participants in this study had a primary diagnosis of schizophrenia,or
Ethical approval was given by the Local Research Ethics Committee (LREC number
06/s1103/76). Participants were recruited from a high security inpatient NHS setting. Letters
were sent to Responsible Medical Ofﬁcers and clinical psychologists in the hospital in order
to identify potential participants. Prior to seeking informed consent from potentially eligible
patients, the respective patient’s Responsible Medical Ofﬁcers were asked to provide consent
for their patient to be approached. Following consent, patients were approached by a chartered
clinical psychologist (HL) and following a full description of the study, patients were invited
Assessments were administered to participants at the start, at 5 weeks (mid group), and
at the end of the programme, with a 6-week follow-up. All the clinical outcome measures
were standardized measures, either self-report questionnaires or structured interviews with
acceptable psychometric properties.
All psychometric assessments were carried out by the assistant psychologists who had both
received in-house training in the delivery of such assessments. Both were trained to use the
Positive and Negative Syndrome Scale (PANSS) using video assessment (with reliability at
Social Comparison Scale (SCS): this scale was developed by Allan and Gilbert (1995), and
is an 11-item scale that taps global comparisons to others in the domains of attractiveness,
rank and group ﬁt (feeling similar or different to others). A lower total score reﬂects relative
inferiority compared with others, whereas a higher total score indicates relative superiority.
External Shame (the Other as Shamer Scale – OAS): this scale was developed by Goss,
Gilbert and Allan (1994) and Allan, Gilbert and Goss (1994) to measure external shame (how
an individual thinks others see him/her). The scale consists of 18 items asking respondents
to indicate the frequency of their feelings and experiences to items such as, “I feel insecure
about others’ opinion of me” and “other people see me as small and insigniﬁcant” on a 5-point
Likert scale (never, seldom, sometimes, frequently, almost always). A total score is given by
adding up the items; a higher score indicates greater experience of external shame.
Self Compassion Scale (SeCS; Neff, 2003b).This scale is a self-report measure that
explores self-compassion in individuals. It is a 26-item scale that measures self-compassion
(13 items) and coldness towards the self (13 items). There are six subscales – three measure
self-compassion: common humanity, self-kindness and mindfulness. There are also three
subscales to measure coldness towards the self: self-judgment, over identiﬁcation, and
516 H. Laithwaite et al.
isolation. Responses are given on a 5-point Likert scale ranging from 1 =“almost never”
and 5 =“almost always”. Subscale scores are computed by calculating the mean of subscale
item responses. To compute a total self-compassion score, reverse score the negative subscale
items – self-judgment, isolation, and over-identiﬁcation – then compute a total mean. The
higher the total score, the greater the self-compassion (NB: this is recommended scoring by
Neff, personal communication, but not scoring of original 2003b paper).
The Beck Depression Inventory II (Beck, Steer and Brown, 1996) was used as a self-report
measure of mood (score range 0–63). Higher scores reﬂect increase in self-reported low mood.
The Rosenberg Self-Esteem measure (RSE; Rosenberg, 1965; Rosenberg, Schooler,
Schoenbach and Rosenberg, 1995) is a 10-item self-report measure of self-esteem. Higher
scores (range 0–30) are indicative of higher self-esteem.
The Self-Image Proﬁle for Adults (SIP-AD; Butler and Gasson, 2004) consists of 30 self-
descriptions and is a self-report questionnaire. Participants are invited to rate themselves as
they are and how they would like to be (ideal) along each self-description. A self-image score
(SI) represents how the individual feels about him/herself. A high self-image score suggests the
person has a positive view of him/herself. Self-esteem (SE) reﬂects an individual’s evaluation of
him/herself. On the SIP-AD this is operationalized as the discrepancy between how the person
sees him/herself and how they wish to be (ideal). A high score reﬂects a wide discrepancy and
therefore lower scores are interpreted as reﬂecting high self-esteem.
The Positive and Negative Syndrome Scale (PANSS; Kay, Fiszbein and Opler, 1987)
measures 32 symptoms on a 7-point Likert scale, deriving three composite subscales: Positive,
Negative, or General Psychopathology. Higher raw scores indicate higher symptomatology.
The Recovery After Psychosis programme was delivered by a team comprising two chartered
clinical psychologists (HL and PC), an advanced practitioner (MO’H), a trainee clinical
psychologist (LA) and two assistant psychologists (SP and PD). The group was delivered by
three therapists (due to security reasons). The ﬁrst group was facilitated by HL, MO’H and
SP and the last two groups were facilitated by HL, MO’H, PC and LA. SP and PD provided
between group session individual support. The programme was developed by HL and AG and
based on Compassionate Mind Training (Gilbert, 2001). AG provided the group facilitators
with clinical supervision. The programme ran for 10 weeks (20 sessions). This involved two
sessions a week. The programme was divided into the following three modules:
Module one: understanding psychosis and recovery. The aim of this module was to help
patients conceptualize the holistic nature of psychosis and the impact of this on various aspects
of their lives. Patients were encouraged to think about psychosis in relation to their emotions,
their cognitions, their behaviour, relationships and environment (see Figure 1). This model
was then used to understand recovery. Therefore patients were encouraged to think beyond
recovery as symptom reduction, but also to view recovery in terms of their emotions, feelings,
relationships with others and their environment. To help patients with this, the metaphor of
the “pebble in water” was used, so that they could understand how recovery or progress in one
area of their life can have an impact on another area. Another group exercise involved using
Recovery After Psychosis (RAP) 517
Figure 1. “Pebble in the water” conceptualization of recovery and psychosis
the metaphor of “recovery as a journey”, which helped create a visual experience of the many
difﬁculties that they may face in the future, and the “tools” they need to take with them on
their journey to help with this.
Module two: Understanding compassion and developing the ideal friend. In this module the
group explored the concept of compassion and the many features of this (strength, forgiveness,
acceptance, trust, non-judgemental). The strengths and weaknesses of these characteristics
were discussed in depth. This exercise progressed to the creation of the “ideal friend”. The
intention of creating this ideal friend is for patients to be able to refer to “someone” who is
compassionate and, over time, it is anticipated that they will internalize the characteristics of
this ideal friend, to develop their own compassionate responses towards themselves and others.
Guided discovery techniques were used to illicit an image of this ideal friend, and patients
were encouraged to focus on characteristics such as voice tone, facial expressions, and body
posture. Throughout the remainder of sessions, the programme referred to the ideal friend, and
used exercises to help develop compassionate responding. Participants were asked to keep a
diary of any negative emotions and self-critical thoughts they experienced during the week,
and how they responded to this using their “ideal friend”.
Module three: Developing plans for Recovery after Psychosis. This part of the programme
involved the development of a Recovery After Psychosis plan (focusing on triggers, early
warning signs, use of safety behaviours, action plan and agreed coping strategies). This
information was used to create a compassionate letter, which involved participants writing a
518 H. Laithwaite et al.
letter to themselves (as written by their ideal friend). This letter contained encouragement and
support in relation to how to respond to set-backs and how to seek help in the future.
Three groups were run in the hospital. There were 19 (all male) participants in total and
18 participants completed the programme. The mean age of the participants was 36.9 (SD
9.09). The mean duration in hospital was 8 years. Five participants had received a diagnosis of
schizophrenia; 10 paranoid schizophrenia and 3 bi-polar affective disorder (these 3 participants
had experienced auditory hallucinations when elated, although at the time of the group, these
had remitted). Eight of the participants also had a co-morbid personality disorder, namely
anti-social personality disorder. One participant was considered to be in the “borderline”
intellectual disability range.
Analyses were carried out using SPSS for windows (version 14). Descriptive statistics were
conducted and further analyses were carried out using Friedman’s ANOVA. Signiﬁcant overall
effects were followed up with Wilcoxon signed ranks (two-tailed). Effect sizes based on
Wilcoxon signed ranks are provided for all outcome measures for the purposes of transparency.
It should be noted that pmeasures were not adjusted for multiple comparisons. This was a
pilot study and thus we did not want to potentially miss signiﬁcant outcomes by restricting
Primary outcomes measures
Overall signiﬁcant changes were found on the Social Comparison Scale, Other As Shamer
Scale and the Beck Depression Inventory II, the Rosenberg Self-Esteem measure and the Self-
Image proﬁle for Adults. Further analyses using Wilcoxon signed ranks test found signiﬁcant
changes on the Social Comparison Scale between the start and end of the group (Z=1.96,
n-ties =11, p<.05, r=0.3) and this change was maintained at follow-up (Z=2.148,
n-ties =10, p<.05, r=0.36). A small change was found on the Other as Shamer scale between
the start of the group and 6-week follow-up (Z=.801, n-ties =11, p>.5, r=0.15). Signiﬁcant
changes on the Beck Depression Scale were found at the end of treatment (Z=2.332,
n-ties =15, p<.05, r=0.38) and at 6-week follow-up (Z=−2.825, n-ties =16, p<.01,
r=0.47). An overall signiﬁcant change was found on the Rosenberg self-esteem questionnaire.
Further analyses using Wilcoxon signed ranks test demonstrated a signiﬁcant change at 6-week
follow-up (Z=−2.80, n-ties =15, p<.01, r=0.47) from baseline. Signiﬁcant changes were
not found on the Self-compassion scale, the Robson self-concept questionnaire or the Self-
image proﬁle for adults.
Recovery After Psychosis (RAP) 519
Tabl e 1. Primary outcome measures: change in assessment measures over course of treatment (Median, IQR and Friedman’s analysis)
Pre-treatment Mid group Post-treatment 6 week follow-up Effect sizes (r)
Measure median and IQR median and IQR median and IQR median and IQR X21(df)p(t1-t3)∗∗ (t1-t4)
Scale 3.30 (3.1–3.7) 3.57 (3.3–3.9) 3.48 (3.2–4.2) 3.63 (3.1–4.1) 4.87(3) .18 0.22 0.28
Scale 36.00 (29–39) 35.00 (33–40.5) 38.00 (32.5–43.5) 35.00 (33.5–43) 8.54(3) .036∗0.30 0.36
Inventory (11) 9.00 (4.5–15.5) 6.00 (3.0–16) 4.00 (3.0–8.0) 4.00 (1.5–10) 10.05 (3) .018∗0.38 0.47
Other as Shamer Scale 33.00 (23–41.5) 36.50 (25.5–48) 32.50 (22.5–36.3) 31.50 (18.8–46.7) 8.35 (3) .04∗0.04 0.15
Questionnaire 19.00 (18–22) 19.00 (18–22) 20.00 (18.5–23) 22.00 (19–26) 12.5 (3) .006∗0.14 0.47
Questionnaire 126.50 (120–142) 128.50 (120–144.25) 127.50 (115–140.6) 127.50 (112.6–149.7) 1.85 (3) .603 0.01 0.24
SIP-AD-SI 132 (102–150) 129 (109.5–144) 131 (114–149.5) 126 (111–142) 5.09 (3) .165 0.14 0.06
SIP-AD-SE 24 (16.5–37) 25 (17.5–45.5) 20 (12.5–38.5) 22 (14–41) 2.03 (3) .566 0.02 0.07
∗signiﬁcant results ∗∗t1-t3 (pre-treatment to end of treatment) t1-t4 (pre-treatment to 6 week follow-up) Effect sizes calculated on Wilcoxon signed
520 H. Laithwaite et al.
Tabl e 2. Secondary outcome measures: PANSS
Pre-treatment Mid group Post-treatment 6 week follow-up Effect sizes (r)
Measure median and IQR median and IQR median and IQR median and IQR X21 (df)p(t1-t3)∗∗ (t1-t4)
PANSS Positive 9.00 (8–10) No mid group 9.00 (8–10) 8.00 (7–10) 2.79 (2) .248 0.1 0.24
PANSS Negative 10.00 (9–13) No mid group 10.00 (9–16.5) 9.00 (8–12) 5.79 (2) .055 0.02 0.3
Psychopathology 24.00 (20.5–26) No mid group 21.00 (18.5–23.5) 19.00 (16.5–21) 7.61(2) .022∗0.38 0.41
PANSS Depression 9.00 (2.99) No mid group 7.31, 7.00 (2.35) 6.31, 6.00 (2.25) 5.76 (2) .056 0.26 0.31
∗signiﬁcant results ∗∗ t1-t3 (pre-treatment to end of treatment) t1-t4 (pre-treatment to 6 week follow-up) Effect sizes calculated on Wilcoxon
Recovery After Psychosis (RAP) 521
Signiﬁcant changes were found on the PANSS general psychopathology score at the end
of the group (Z=2.23, n-ties =14, p <.05, r=0.38) and this was maintained at follow-up
(Z=2.75, n-ties =12, p<.01, r=0.41). Signiﬁcant changes were not found on the PANSS
positive, negative or depression scales.
This was a pilot, pre-trial study. This was the ﬁrst time that a compassion focused group
intervention has been run at the State Hospital and, to our knowledge, the ﬁrst time that it
has been run with a forensic clinical population. The primary objective of this study was to
evaluate whether the programme would improve depression, improve self-esteem, develop
self-compassion and social comparison and lower the experience of shame compared with
others, and hence improve how an individual perceives how others see him/her.
The ﬁndings of this study demonstrated a large magnitude of change for levels of depression
and self-esteem as measured by the Beck Depression Inventory II, and Rosenberg Self-Esteem
Inventory. A moderate magnitude of change was found for the social comparison scale and
general psychopathology, with a small magnitude of change for shame, as measured by
the Other as Shamer Scale. These changes were maintained at 6-week follow-up. Gilbert
(2005) has shown that self-critical thinking biases are inﬂuential in the development and
maintenance of psychopathology; therefore a programme such as this recovery programme,
which focuses on developing compassionate responses to shame, self-critical and self-attacking
thoughts, will likely lead to a reduction in depression, shame and an increase in self-esteem.
Much of the research on psychopathology has focused on depression; however, we know
that self-critical thinking, shame and low self-esteem also play a role in the development
and maintenance of psychotic experiences (Bentall, Kinderman and Kaney, 1994; Garety,
Kuipers, Fowler, Freeman and Bebbington, 2001; Smith et al., 2006). We observed changes
on the general psychopathology scale that may be associated with a reduction in shame and
self-critical thinking. However, in a larger scale study, investigating the mediating effects of
changes in compassion, shame and self-critical thinking on general psychopathology might
be interesting. Furthermore, anger is a common response to rejection from others, shame
and feeling inferior (Gilbert and Miles, 2000; Baumeister, Smart and Boden, 1997), and
therefore an intervention that focuses on reducing shame, and improving comparison with
others, may have an impact on reducing anger and possibly risk of violent offending. This
again could be explored in a larger scale trial of a compassion focused group on shame,
anger and risk reduction. There is limited published research carried out on interventions for
psychosis with a mentally disordered population. However, although this study drew from
patients in a high security setting, the results sit favourably with a case series study of three
patients with psychosis, anger problems and substance misuse in a low security environment
(Haddock, Lowens, Brosnan, Barrowclough and Novaco, 2004) and with a self-esteem group
intervention carried out in high security (Laithwaite and Gumley, 2007).
A signiﬁcant change was found on the Rosenberg self-esteem questionnaire but not on the
other measures of self-esteem. In the self-esteem group evaluation (Laithwaite and Gumley,
2007) self-esteem was found to be strongly correlated with scores on the BDI II. That is,
lower self-esteem was associated with more severe depressed mood. Therefore it was unclear
522 H. Laithwaite et al.
whether changes in self-esteem were related to changes in depressed mood or vice versa.
Although correlations between scores on the BDI II and the Rosenberg self-esteem measure
were not carried out in this study, it is possible that a similar relationship was present. Indeed,
Rosenberg and colleagues have found that the negative correlation between the two variables
“seems to be due somewhat more to the effect of depression on self-esteem than to the effect
of self-esteem on depression” (Rosenberg et al., 1995, p. 145). Furthermore, the ﬁndings
from Birchwood and Iqbal (1998) draw attention to the fact that depression in psychosis
is particularly common, with prevalence estimates ranging from 22% to75%, depending on
Signiﬁcant changes were not found on the self-compassion scale. However, the median
score on this measure is comparable with norms developed on a general student population
(Neff, 2003a, b). It may be that the self-report of compassion is different for individuals
who have lacked the experience of compassion from others during critical periods of their
development. This would be consistent with the proposals of social mentality theory. There
were several challenges to delivering this programme. The concept of compassion is one that
is not usually discussed in forensic clinical settings where notions of symptom reduction and
risk management prevail. Participants were able to describe the characteristics of compassion
but struggled to relate these characteristics to themselves. For example, acceptance and
forgiveness generated much discussion in the group, with many participants reportedly feeling
uncomfortable about self-forgiveness as it may be interpreted as lack of remorse or empathy
for their victims. The programme focused on developing acceptance for past behaviours but
taking responsibility for future possible outcomes. This seemed to empower many of the
group participants as there was some hope of moving on from the stigma and shame of
the past to being positive about the future. This change in looking at future possibilities
also helped participants respond to self-attacking thoughts that seemed to be mainly past
orientated. There is a movement to promote forgiveness in violent offenders and to promote
the potential to develop a “good life” (Ward and Marshall, 2004), with this being seen as a more
positive approach to offender rehabilitation as it helps to engage individuals in therapy, and
subsequently may reduce risk of future violent offences (Day, Gerace, Wilson and Howells,
Many of the participants initially found it challenging to generate a compassionate image.
This was not just simply that participants in the group found it difﬁcult to access early
memories, as some could clearly describe memories of inconsistent care-giving – it was
that they could not relate to personal experiences of compassion, and therefore found it
challenging to generate an internal working model of a compassion. The research on attachment
theory may help to explain this. When early attachment experiences are compromised,
this may result in insecure adult attachment states of mind. We know from research that
individuals with psychosis and with violent offending histories often have experienced
disrupted attachment histories (Boswell, 1996; Read and Gumley, 2008). For example, limited
early experiences of care giving conducive to secure attachment and limited experience of
mirroring, where needs of the infant are reﬂected on by their care-giver (Fonagy, Gergely,
Jurist and Target, 2002). Such early attachment experiences have an effect on the development
of mentalization and subsequent regulation of affect (Liotti and Gumley, 2008). Therefore
individuals’ ability to reﬂect on their own emotional mental states and memories may be
compromised (Bowlby, 1988., Fonagy et al., 2002). Such early attachment histories might
also have been associated with avoidant/dismissive coping styles. The compassion focused
Recovery After Psychosis (RAP) 523
therapy encouraged participants to reﬂect upon episodic memories that may have resulted in
some participants feeling anxious or distressed and using avoidant coping styles so as not to
think about an image. Furthermore, individuals operating in a threat focused social mentality
may have experienced a degree of fear when generating a compassionate image (Gilbert,
2003). To overcome some of these challenges, group facilitators offered support and helped
the group to generate a group compassionate image, and also suggested that they could think
of a place or non-human object that generated feelings of warmth and safety.
There are several limitations to this study. In particular, the study was conducted with a small
sample of participants without any matched control group. We therefore cannot be fully conﬁd-
ent that the changes observed over time are fully attributable to the effects of the intervention.
Future research could incorporate a larger sample size, and randomization to an appropriate
control condition, which would improve the reliability and generalizability of ﬁndings. In
addition, many of the measures used in the study do not have published norms and have not
been validated with a forensic clinical population. However, comparisons can be drawn with
previous studies that have used these measures. We know that patients in the forensic clinical
population score higher on external shame and lower on social comparison compared with a
student population (Goss et al., 1994; Gilbert, Cheung, Grandﬁeld, Campey and Irons, 2003).
Gilbert and Proctor (2006) developed a group intervention for six patients with major/severe
long term and complex difﬁculties. At the start of this group, the mean score for participants
was much higher on external shame than the forensic clinical population. However, at the end of
the intervention, the scores on external shame and social comparison were comparable with the
forensic clinical population. It is also important to recognize that Bonferonni corrections were
not used in the analysis. One limitation of the study is the accepted pvalue was not corrected
for the number of multiple comparisons and small sample size. However, we considered that
given the pilot nature of the study that the increased risk of type I errors was acceptable. This
was because we wished to estimate which outcomes were more important to measure in a larger
randomized study. Facilitators involved in the delivery of the group were also involved in the
completion of psychometric assessments. To reduce bias, future evaluation of the programme
would be improved by using raters independent of the treatment programme.
In conclusion, this preliminary study evaluated a compassion focused group intervention
for patients with psychosis residing in a high security setting. The ﬁndings demonstrate an
improvement in depression, self-esteem, and rating of self compared with others, and a reduc-
tion in shame, and general psychopathology. Further replication of this study could involve a
waiting list control group, a larger sample size and independent rating of change in outcome.
Further research could also involve extending this protocol to non-forensic populations.
Dr Karen Allan, Lead Consultant Clinical Psychologist at the State Hospital and Ms Patricia
Cawthorne, Clinical Nurse Specialist and Psychological Therapies Operational Manager.
Allan, S. and Gilbert, P. (1995). A social comparison scale: psychometric properties and relationships
to psychopathology. Personality and Individual Differences, 1, 293–299
524 H. Laithwaite et al.
Allan, S., Gilbert, P. and Goss, K. (1994). An exploration of shame measures: II. Psychopathology.
Personality and Individual Differences, 17, 719–722.
Baumeister, R. F., Smart, L. and Boden, J. M. (1997). Relation to threatened egotism to violence and
aggression: the dark side of high self-esteem. Psychological Review, 103, 5–33.
Bebbington, P. E., Bhugra, D., Brugha, T., Singleston, N., Farrell, M., Jenkins, G., Lewis, G and
Meltzer, H (2004). Psychosis, victimization and childhood disadvantage: evidence from the second
British National Survey of psychiatric morbidity. British Journal of Psychiatry, 185, 220–226.
Beck, A. T., Steer, R. A. and Brown, G. K. (1996). Manual for the Beck Depression Inventory –II.San
Antonio TX, The Psychological Corporation.
Bentall, R. P., Kinderman, P. and Kaney, S. (1994). The self, attributional processes and abnormal
beliefs: towards a model of persecutory delusions. Behaviour Research and Therapy, 32, 331–341.
Birchwood, M. J., Hallett, S. E. and Preston, M. C. (1989). Schizophrenia: an integrated approach to
research and treatment. New York: New York University Press.
Birchwood, M., Mason, R., MacMillan, F. and Healy, J. (1993). Depression, demoralisation and
control over psychotic illness: a comparison of depressed and non-depressed patients with a chronic
psychosis. Psychological Medicine, 23, 387–395.
Birchwood, M. and Iqbal, Z. (1998). Depression and suicidal thinking in psychosis: a cognitive
approach. In T. Wykes, N. Tarrier and S. Lewis (Eds.), Outcome and Innovation in Psychological
Treatment of Schizophrenia (pp. 81–100). Hoboken, NJ: John Wiley and Sons Inc.
Boswell, G. (1996). Young and Dangerous: the backgrounds and careers of section 53 offenders.
Bowlby, J. (1988). A Secure Base: clinical applications of attachment theory. London: Routledge.
Buss, D. M. (2003). Evolutionary Psychology: the new sciences of mind (2nd edn). Boston: Allyn and
Butler, R. J. and Gasson, S. L (2004). The Self-Image Proﬁle for Adults (SIP-Adult). San Antonio, TX:
Campbell, M. L. C. and Morrison, A. P. (2007). The relationship between bullying, psychotic-
like experiences and appraisals in 14–16 year olds. Behaviour Research and Therapy, 45, 1579–
Day, A., Gerace, A., Wilson, C. and Howells, K. (2008). Promoting forgiveness in violent offenders:
a more positive approach to offender rehabilitation? Aggression and Violent Behaviour, 13, 195–200.
Fonagy, P., Target, M., Steele, M., Steele, H., Leigh, T., Levinson, A. and Kennedy, R. (1997).
Morality, disruptive behaviour, borderline personality disorder, crime, and their relationships to
security of attachment. In L. Atkinson and K. J. Zucker (Eds.), Attachment and Psychopathology.
New York: Guilford Press.
Fonagy, P., Gergely, G., Jurist, E. L. and Target, M. (2002). Affect Regulation, Mentalisation, and the
Development of the Self. London: Karnac.
Garety, P. A., Kuipers, E., Fowler, D., Freeman, D. and Bebbington, P. E. (2001). A cognitive model
of the positive symptoms of psychosis. Psychological Medicine, 31, 189–195.
Gilbert, P. (1989). Human Nature and Suffering. Hove: Lawrence Erlbaum.
Gilbert, P. (1992). Depression: the evolution of powerlessness. Hove: Lawrence Erlbaum.
Gilbert, P. (2000). Social mentalities: internal “social” conﬂicts and the role of inner warmth and
compassion in cognitive therapy. In P. Gilbert and K. G. Bailey (Eds.), Genes on the Couch;
explorations in evolutionary psychotherapy (pp. 118–150). Hove: Psychology Press.
Gilbert, P. (2001). Evolutionary approaches to psychopathology: the role of natural defences. Australian
and New Zealand Journal of Psychiatry, 35, 17–27.
Gilbert, P. (2003). Evolution, social roles and the differences in shame and guilt. Social Research, 70,
Gilbert, P. (2004). Depression: a biopsychosocial, integrative and evolutionary approach. In M. Power
(Ed.), Mood Disorders: a handbook of science and practice (pp. 99–142). Chichester: Wiley.
Recovery After Psychosis (RAP) 525
Gilbert, P. (2005). Compassion and cruelty: a biopsychosocial approach. In P. Gilbert (Ed.),
Compassion: conceptualisations, research and use in psychotherapy (pp. 9–74). Brighton: Brunner-
Gilbert, P., Cheung, M. S-P., Grandﬁeld, T., Campey, F. and Irons, C. (2003). Recall of threat and
submissiveness in childhood: development of a new scale and its relationship with depression, social
comparison and shame. Clinical Psychology and Psychotherapy, 10, 108–115.
Gilbert, P. and Irons, C. (2005). Focused therapies and compassionate mind training for shame and self-
attacking. In P. Gilbert (Ed.), Compassion: conceptualisations, research and use in psychotherapy
(pp. 263–326). London: Routledge.
Gilbert, P. and Miles, J. N. V. (2000). Sensitivity to social put down: its relationship to perceptions of
social rank, shame, social anxiety, depression, anger and self-other blame. Personality and Individual
Differences, 29, 757–774.
Gilbert, P. and Proctor, S. (2006). Compassionate mind training for people with high shame and self-
criticism: overview and pilot study of a group approach. Clinical Psychology and Psychotherapy, 13,
Goss, K., Gilbert, P. and Allan, S. (1994). An exploration of shame measures I. the “other as shamer
scale”. Personality and Individual Differences, 17, 713–717.
Gumley, A. and Schwannauer, M,(2006). Staying Well After Psychosis: a cognitive interpersonal
approach to recovery and relapse prevention. New York: Wiley.
Haddock, G., Lowens, I., Brosnan, N., Barrowclough, C. and Novaco, R. W. (2004). Cognitive-
behavioural therapy for in-patients with psychosis and anger problems within a low secure
environment. Behavioural and Cognitive Psychotherapy, 32, 77–98.
Herz, M. I. and Melville, C. (1980). Relapse in schizophrenia. American Journal of Psychiatry, 137,
Kay, S., Fiszbein, A. and Opler, I. (1987). The positive and negative syndrome scale (PANSS) for
schizophrenia. Schizophrenia Bulletin, 13, 261–275.
Laithwaite, H. and Gumley, A. (2007). Sense of self, adaptation and recovery in patients with psychosis
in a forensic NHS setting. Clinical Psychology and Psychotherapy, 14, 302–316.
Laithwaite, H., Gumley, A., Benn, A., Scott, E., Downey, K., Black, K. and McEwen, S. (2007).
Self-esteem and psychosis: a pilot study investigating the effectiveness of a self-esteem programme
on the self-esteem and positive symptomatology of mentally disordered offenders. Behavioural and
Cognitive Psychotherapy, 35, 569–577.
Liotti, G. and Gumley, A. (2008). An attachment perspective on schizophrenia: the role of disorganised
attachment, dissociation and mentalisation. Psychosis, Trauma and Dissociation, 14, 117–133.
Lee, D. (2005). The perfect nurturer: a model to develop a compassionate mind within the context
of cognitive therapy. In P. Gilbert (Ed.), Compassion: conceptualisations, research and use in
psychotherapy. London: Bruner-Routledge.
Legg, L. and Gilbert, P. (2006). A pilot study of gender of voice and gender of voice hearer in psychotic
voice hearers. Psychology and Psychotherapy: Theory, Research and Practice, 79, 517–527.
MacBeth, A., Schwannauer, M. and Gumley, A. (2008). The association between attachment style,
social mentalities, and paranoid ideation: an analogue study. Psychology and Psychotherapy: Theory,
Research and Practice, 81, 79–93.
McCandless-Glimcher, L., McKnight, S., Hamera, E., Smith, B. L., Peterson, K. A. and Plumlee,
A. A. (1986). Use of symptoms by schizophrenics to monitor and regulate their illness. Hospital and
Community Psychiatry. 37, 929–933.
Murphy, J. M., Nierenberg, A. A., Monson, R. R., Laird, N. M., Sobol, A. M. and Leighton,
A. H. (2002). Self-disparagement as a feature and forerunner of depression: ﬁndings from the Stirling
County study. Comprehensive Psychiatry, 43, 13–21.
Neff, K. D. (2003a). Self-compassion: an alternative conceptualisation of a healthy attitude towards
oneself. Self and Identity, 2, 85–102
526 H. Laithwaite et al.
Neff, K. D. (2003b). The development and validation of a scale to measure self-compassion. Self and
Identity, 2, 223–250.
Read, J., Goodman, L., Morrison, A., Ross, C., and Aderhold, V. (2004). Childhood trauma, loss and
stress. In J. Read, L. Mosher and R. Bentall (Eds.), Models of Madness: psychological, social and
biological approaches to schizophrenia (pp. 223–252). Hove, Brunner-Routledge.
Read, J. and Gumley, A. (2008). Can attachment theory help explain the relationship between
childhood adversity and psychosis? Attachment: New Directions in Psychotherapy and Relational
Psychoanalysis, 2, 1–35
Romme, M. A. J. and Escher, A. D. M. A. C. (1989). Hearing voices. Schizophrenia Bulletin, 15,
Rooke, O. and Birchwood, M. (1998). Loss, humiliation and entrapment as appraisals of schizophrenic
illness: a prospective study of depressed and non-depressed patients. British Journal of Clinical
Psychology, 37, 259–268.
Rosenberg, M. (1965). Society and Adolescent Self-Image. Princeton, NJ: Princeton University Press.
Rosenberg, M., Schooler, C., Schoenbach, C. and Rosenberg, F. (1995). Global self-esteem and
speciﬁc self-esteem: different concepts, different outcomes. American Sociological Review, 60, 141–
Sloman, L. (2000). How involuntary defeat is related to depression. In L. Sloman and P. Gilbert (Eds.),
Subordination and Defeat: an evolutionary approach to mood disorders and their therapy (pp. 47–66).
Mahwah, NJ: Lawrence Erlbaum.
Sloman, L., Gilbert, P. and Hasey, G. (2003). Evolved mechanisms in depression: the role and
interaction of attachment and social rank in depression. Journal of Affective Disorders, 74, 107–
Smith, B., Fowler, D. G., Freeman, D., Bebbington, P., Bashforth, H., Garety, P., Dunn, G. and
Kuipers, E. (2006). Emotion and psychosis: links between depression, self-esteem, negativeschematic
beliefs and delusions and hallucinations. Schizophrenia Research, 86, 181–188.
The State Hospitals Board for Scotland (2005). The State Hospital Annual Review. Carstairs: Author.
Ward, T. and Marshall, W. L. (2004). Good lives, aetiology and the rehabilitation of sex offenders: a
bridging theory. Journal of Sexual Aggression, 10, 153−169.