Schizophrenia Research 125 (2011) 54–61

Contents lists available at ScienceDirect

Schizophrenia Research
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s c h r e s

A randomized controlled trial of cognitive behavioral therapy for individuals

at clinical high risk of psychosis
Jean Addington a,b,c,⁎, Irvin Epstein c, Lu Liu a, Paul French d,e,
Katherine M. Boydell b,f,g, Robert B. Zipursky h
a
Department of Psychiatry, University of Calgary, Alberta, Canada
b
Department of Psychiatry, University of Toronto, Toronto, Ontario, Canada
c
Centre for Addiction and Mental Health, Toronto, Ontario, Canada
d
Greater Manchester West Mental Health NHS Foundation Trust, Manchester, UK
e
Department of Psychology, University of Manchester, Manchester, UK
f
Community Health Systems Resource Group, The Hospital for Sick Children, Canada
g
School of Public Health, University of Toronto, Canada
h
Department of Psychiatry and Behavioural Neurosciences, McMaster University, Hamilton, Ontario, Canada

a r t i c l e i n f o a b s t r a c t

Article history: Background: There has been increasing interest in early detection during the prodromal phase of a
Received 31 July 2010 psychotic disorder. To date a few treatment studies have been published with some promising
Received in revised form 10 October 2010 results for both pharmacological treatments, using second generation antipsychotics, and
Accepted 14 October 2010
psychological interventions, mainly cognitive behavioral therapy. The purpose of this study was
Available online 12 November 2010
to determine first if cognitive behavioral therapy (CBT) was more effective in reducing the rates of
conversion compared to a supportive therapy and secondly whether those who received CBT had
Keywords: improved symptoms compared to those who received supportive therapy.
Psychosis
Method: Fifty-one individuals at clinical high risk of developing psychosis were randomized to
Clinical high risk
CBT or a supportive therapy for up to 6 months. The sample was assessed at 6, 12 and 18 months
Cognitive behavior therapy
post baseline on attenuated positive symptoms, negative symptoms, depression, anxiety and social
functioning.
Results: Conversions to psychosis only occurred in the group who received supportive therapy
although the difference was not significant. Both groups improved in attenuated positive
symptoms, depression and anxiety and neither improved in social functioning and negative
symptoms. There were no differences between the two treatment groups. However, the
improvement in attenuated positive symptoms was more rapid for the CBT group.
Conclusions: There are limitations of this trial and potential explanations for the lack of differences.
However, both the results of this study and the possible explanations have significant implications
for early detection and intervention in the pre-psychotic phase and for designing future
treatments.
© 2010 Elsevier B.V. All rights reserved.

1. Introduction psychotic illness. One of the major directions of this work is

In schizophrenia research there is a key focus on identifying
individuals who may be in the pre-psychotic stage of a
⁎ Corresponding author. Centre for Mental Health Research and Education,
University of Calgary, 3280 Hospital Drive NW, Calgary, Alberta, T2N 4Z6,
Canada.
E-mail address: jmadding@ucalgary.ca (J. Addington).
determining the risk of conversion to a full blown psychotic
illness and developing predictors of conversion (Cannon et al.,
2008; Ruhrmann et al., 2010; Yung et al., 2003). Criteria, mainly
based on attenuated positive symptoms, have been established
for identifying these young people at “clinical high risk” (Miller
et al., 2002; Yung et al., 1996). However, individuals who meet
at-risk criteria experience a diverse array of symptoms and
behaviors and the potential benefits of treatment for these

0920-9964/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.schres.2010.10.015
 J. Addington et al. / Schizophrenia Research 125 (2011) 54–61
young people have been recognized for some time. Four
controlled trials addressing interventions in this clinical high-
risk (CHR) population have been published.
In the first study 59 “ultra-high-risk” participants were
randomized to six months of active treatment (risperidone
1–3 mg/day plus a modified CBT) or to a needs-based
intervention (McGorry et al., 2002). By the end of treatment,
significantly fewer individuals in the active treatment group
had progressed to a first-episode of psychosis (9.7% vs 36%).
Despite some limitations, this represented a landmark study.
A second, more rigorous study, was a randomized, double-
blinded trial of 60 help-seeking prodromal patients com-
paring the efficacy of an antipsychotic (olanzapine) vs
placebo in preventing or delaying the onset of psychosis.
Although not statistically significant, at one-year follow-up
16% of olanzapine-treated participants had converted to
psychosis compared with 35% of placebo-treated partici-
pants. Olanzapine was associated with significantly greater
symptomatic improvement in prodromal symptoms than
placebo (McGlashan et al., 2006).
A third trial, the Early Detection and Intervention
Evaluation (EDIE), was a single-blinded, randomized trial of
cognitive behavior therapy (CBT) with individuals at high risk
for psychosis (Morrison et al., 2004). Fifty-eight patients were
randomized to either CBT or to monitoring. CBT significantly
reduced the likelihood of progression to psychosis over
12 months and improved positive symptoms with some
benefits maintained at 3-year follow-up (Morrison et al.,
2007). The fourth study was a twelve week trial comparing
eicosapentoic acid (EPA) with placebo (Amminger et al.,
2010). At 12 months 4.9% (2/41) of individuals in the EPA
group compared to 27.5% (11/40) in the placebo group
developed psychosis. Furthermore, there were significant
group differences in positive and negative symptoms at
12 weeks and 12 months in favor of the treatment group.
These studies provide preliminary evidence for the
appropriateness of psychological and pharmacological inter-
ventions in the treatment of young people at CHR. However,
treating young people in the putative prodromal phase does
cause some concern that these young people may be exposed
to unnecessary and potentially harmful treatments. For
example, there have been some concerns about the use of
antipsychotic medication (Bentall and Morrison, 2002). Thus,
psychological interventions might be expected to be prom-
ising in this pre-psychotic period particularly when the
symptoms are less severe and also less specific. French and
Morrison (2004) present several arguments to support why
CBT may be a beneficial psychological intervention for this
clinical high-risk group. It addresses the range of symptoms
and concerns present in the clinical high-risk period and
teaches potentially effective strategies to protect against the
impact of environmental stressors that may contribute to the
emergence of psychosis.
The primary aim of this study was to test the effectiveness
of CBT compared to a supportive type therapy in preventing
or delaying the onset of conversion to psychosis. The
hypothesis was that those receiving CBT would have a
lower conversion rate than those in the supportive therapy
group. Secondary aims were to examine the effectiveness of
CBT in reducing prodromal symptoms, depression, anxiety
and poor functioning in help-seeking individuals at CHR of
55
psychosis. Hypotheses would be that greater reduction in
symptoms and functioning would be observed in the group
receiving CBT.
2. Methods
2.1. Setting
This study was conducted at the PRIME Clinic at the Centre
for Addiction and Mental Health (CAMH) in Toronto, Canada.
Further details about the Toronto PRIME Clinic are provided
elsewhere (Addington et al., 2008). The study was approved
by the Human Subjects Review Committee of the University
of Toronto at CAMH. Informed consent was obtained from
those who met criteria and were judged fully competent to
give consent. Parental consent was obtained from parents/
guardians of participants who were under age 16.
2.2. Design
This was a single blind RCT of CBT vs supportive therapy
with a 6-month treatment phase and a 12-month follow-up
phase. Fifty-one participants were randomized immediately
after the baseline assessment using concealed stratified
randomization with minimization (Popcock, 1983). Partici-
pants were stratified by sex and severity of the prodromal
symptoms. Treatment began 1–2 weeks after completion of
baseline assessment. Treatment was available for up to 20
sessions over a 6-month period. Comprehensive assessments
were conducted at baseline, 6, 12 and 18 months. Prodromal
symptoms were monitored monthly during the 6-month
treatment phase. Clinical raters and attending psychiatrists
were blind to the treatment group.
2.3. Participants
Recruitment of participants between 14 and 30 years old
was sought from a variety of sources including family
physicians, student counselors, and community mental
health teams and practitioners. Recruitment and ascertain-
ment methods included advertisement on radio, public
transit and local newspaper and have been described
elsewhere (Addington et al., 2008). All CHR participants
were required to meet the Criteria of Prodromal States
(COPS) using the Structured Interview for Prodromal Symp-
toms (SIPS)(Miller et al., 2003). The Presence of Psychotic
Symptoms (POPS)(McGlashan et al., 2003) was used to
determine conversion to psychosis. Participants were exclud-
ed if they met criteria for any current or lifetime axis I
psychotic disorder, prior history of treatment with an
antipsychotic, IQ b70 or past or current history of a clinically
significant central nervous system disorder which may
confound or contribute to prodromal symptoms. A thorough
clinical assessment was conducted by the PRIME psychiatrist
or psychologist (IE, JA) to determine if entry criteria were
met. In designing this study sample size calculations were
based on current reported rates in the literature. We expected
a conversion rate of 40% in the control group with a 50%
reduction in conversion for the active treatment group, i.e. a
reduction of conversion rate from 40% to 20%, a difference
which would be clinically significant. Using a formula based
56 J. Addington et al. / Schizophrenia Research 125 (2011) 54–61
on comparing the proportions of subjects in two groups who
exhibit an outcome (40% to 20%) (Streiner, 1990) sample size
estimates for two-tailed tests with a significance level of 0.05
and a power of 80% were 83 per group.
2.4. Measures
In addition to the SCID and the SIPS which were used as
described above to identify operationally the presence of
prodromal symptoms and measure severity over time, the
following measures were used at 6, 12 and 18 months: the
Calgary Depression Scale (CDSS) (Addington et al., 1993);
Social Anxiety Scale (SAS) and the Social Interaction Anxiety
Scales (SIAS) (Olivares et al., 2001); the Social Functioning
Scale (SFS) (Birchwood et al., 1990). The Working Alliance
Inventory-Short Form (WAI-SF) (Tracey and Kokotovic,
1989), a client and therapist self report measure, based on
12 statements rated on a seven point scale ranging from
“never” to “always” was used. The scale is based on the
conceptualization of therapeutic alliance as being rooted in
three components: therapeutic bond, client and therapist
agreement on goals and client and therapist agreement on
tasks. If possible both client and therapist versions were used
twice, early (session 3–5) and later (session 10–20).
2.5. Interventions
The CBT intervention was a manualized problem-focused
time limited treatment of up to 20 sessions to be completed
within 6 months (French and Morrison, 2004). This inter-
vention was based on the experiences of the Manchester EDIE
Trial in using CBT with a CHR population and specifically
examined strategies for change (French and Morrison, 2004).
These included normalization, generating and evaluating
alternative beliefs, safety behaviors, metacognitive beliefs,
core beliefs, social isolation and relapse prevention. The CBT is
a formulation based approach. Treatment strategies are
selected within the context of a collaboratively derived
formulation and related to the problems that are agreed
upon and prioritized by the client. Additionally the French
and Morrison text represents many of the common themes
and experiences of this client group and provides examples of
how they may be addressed in therapy.
The purpose of the alternative treatment was to match CBT
for nonspecific effects of therapist contact and interest, social
interaction and support. Common factors include client
expectancy, providing a rationale for change, therapist factors
and therapeutic alliance (Tarrier et al., 2004). The supportive
therapy (ST) was an active psychological treatment directly
assisting individuals to cope with current problems. Therapists
followed brief guidelines as to what they could and could not
do. The therapy consisted of finding out how the previous week
had been. Any crises were dealt with, and advice was offered to
help with any immediate problems. No active CBT techniques
were taught or used. Psychoeducational information about
psychosis and managing stress was offered. There was a focus
on listening, reflecting and empathizing, and demonstrating
uncritical acceptance and genuineness. The therapy was non-
confrontational, supportive and accepting.
All therapy sessions were delivered by two master's level
clinicians who were trained in both therapies and had
experience in CBT, supportive therapy and case management.
Approximately 40 h of specific training (including live super-
vision) in CBT and ST occurred over a 3–6 week period prior to
enrollment of participants. Two days of intensive training in
CBT was conducted by Dr French. Drs Morrison and French
were available for consultation in the first few months of the
trial. Weekly meetings were held with the therapists and JA to
discuss treatment fidelity and general patient management and
supervision.
2.6. Treatment fidelity
All sessions were audiotaped. A random selection of 20% (90
tapes) of taped sessions was assessed by two independent
raters who were blind to all client data and to the audiotape
selection procedure. Selected tapes were stratified using
minimization according to the following strata: therapy (CBT
or ST), therapist (A or B), stage of therapy (early, middle or late)
and entry to study (early, middle or late). The tapes were
monitored for fidelity to CBT using the Cognitive Therapy Scale
(CTS) (Dobson et al., 1985; Vallis et al., 1986) which was
designed for rating CBT and has good interrater reliability and
validity. The integrity of ST was determined by a lack of CBT
components on the CTS and by the presence of the central
components outlined for ST. Agreement between raters was
good (intraclass correlation of 0.90 on CTS). The quality of CBT
was assessed as adequate. Means on the cognitive techniques
sub-scale were 19.6 (SD = 8.3, range 15–28) for the CBT group
compared to 0.3 (SD = 1.0, all scores were zero except for 2
scores of 3 and 5) for the ST group. Means for the general
therapy techniques sub-scale were 23 (SD = 3.2, range 16–28)
for the CBT group and 19.7 (SD = 3.5, range 16–29) for the ST
group. The CBT group rated significantly higher than the ST
group both on cognitive techniques (t= 13.4, p b 0.001) and
general techniques (t= 41. p b 0.001). Raters incorrectly clas-
sified 8 of the 90 tapes. Four CBT tapes were rated as ST and four
ST tapes rated as CBT.
2.7. Procedures
Raters were experienced research clinicians who demon-
strated adequate reliability at routine reliability checks. Gold
standard post-training agreement on the distinction between
prodromal and psychotic levels of intensity on the positive
symptom items (i.e., the critical threshold for determining
initial eligibility and subsequent conversion status) was
excellent (kappa = 0.90). The DSM-IV diagnoses were made
using the SCID-I. Interrater reliability was determined at the
start of the study and annually by 100% agreement on the
diagnosis and at least 80% agreement for symptom presence.
Extensive steps were taken to maintain blindness of raters.
Therapists and raters did not communicate details about
individual clients to each other. Office space and data storage
were kept separate and secure.
3. Results
3.1. Participants
Fifty-six consented post-screening, two were psychotic at
baseline, and three did not return after screening and
 J. Addington et al. / Schizophrenia Research 125 (2011) 54–61 57

consenting to participation. Fifty-one individuals (36 males,
15 females) were randomized. Details are presented in the
Consort Diagram in Fig. 1.
All participants met the criteria for the Attenuated Positive
Symptom Syndrome which included the emergence or
worsening over the past year of a non-psychotic disturbance
of thought content, thought process or perceptual abnormal-
ity. We did not have participants who met the other COPS
criteria of Genetic Risk with Deterioration or Brief Intermit-
tent Psychotic Symptoms. The Structured Clinical Interview
for DSM-IV (SCID-I) was used to determine the presence of
any axis I disorders. Forty-one percent (n = 21) had a
comorbid diagnosis of major depressive disorder, 43%
(n = 22) had an anxiety disorder, 10% (n = 5) had alcohol
abuse and 19% (n = 10) had cannabis abuse. Participants'
baseline characteristics are presented in Table 1. Means were
compared using the Student t-test. Where appropriate
categorical variables, proportions, or percentages were
compared between the groups using chi-square or Fisher's
exact tests. There were no significant differences between the
two groups in demographic variables or comorbid diagnoses.
The one exception was that the CBT group had significantly
more students (χ2 = 4.40, p = 0.04).
Thirteen participants dropped out of the trial before the 6-
month assessment. There was no significant difference
between these 13 and the 35 participants who completed at
least one of follow-up assessments in demographics, depres-
sion, ratings on severity of attenuated psychotic symptoms, or
social functioning.
3.2. Outcomes
There were no conversions in those who received CBT and
three in the ST group. All three conversions had a final
diagnosis of schizophrenia. Two of the conversions occurred
approximately 15 weeks after baseline, and the third oc-
curred 10 weeks after baseline. A survival analysis using
Kaplan–Meier plots with log-rank test was employed to
compare the difference in the rate of conversion to psychosis
between the groups. As shown in Fig. 2, Kaplan–Meier
survival estimates showed a higher likelihood of conversion
to psychosis for the ST group compared to the CBT group (log-
rank test p = 0.059). However, in view of the small sample
size and the small or zero events we cannot preclude the
potential difference in the rate of conversion to psychosis
between the two groups.

Referrals(N= 562)

Assessed suitable after phone screen

(N= 302)

Met COPS criteria (N= 112)

Refused any study (N= 37)

Refused but consented to a non-treatment study (N= 19)
Consented (N= 56)

Randomized (N= 51)

Completed baseline Completed baseline

Randomized to CBT Randomized to ST
(N= 27) (N= 24)

Completed 6 month Converted in 1st 6 Converted in 1st Completed 6 month

assessment (N= 19) months (N= 0) 6 months (N= 3) assessment (N= 16)
70% 0% 12.5% 76%

Completed 12 month Completed 12 month

assessment (N= 16) assessment (N= 15)
59% 71%

Completed 18 month Completed 18 month

assessment (N= 15) assessment (N= 13)
56% 62%

Fig. 1. Consort Diagram.

58 J. Addington et al. / Schizophrenia Research 125 (2011) 54–61

Table 1 improvement in SOPS positive symptoms over time

Baseline characteristics. (p b 0.001) but no statistically significant difference in the
Variable CBT (N = 27) ST(N = 24)
change rate of SOPS positive symptoms over time between
the two groups (p = 0.44).
Mean age in years (SD) 20.8 (4.51) 21.1 (3.74)
There were no significant group differences on the CDSS
Years of education SD) 12.7 (3.46) 12.9 (2.41)
Gender, n (%) and no time effects except for a significant improvement for
Male 18 (35.3) 18 (35.3) the ST group between baseline and 6 months (t = 4.94,
Female 9 (17.7) 6 (11.8) p b 0.05). For anxiety there was no significant differences
Currently working, n (%) 13 (25.5) 11 (21.6)
between the groups although both groups improved over
Racial background, n (%)
White 13 (25.5) 16 (31.4) time on the SIAS (F = 0.09, P N 0.05,) and on the SAS there was
Black 2 (3.9) 1 (2.0) a significant time effect for the ST group between baseline and
Asian 7 (13.8) 5 (9.8) 18 months (t = 3.55,p b 0.05). There were no significant
Other 5 (9.8) 2 (3.9) differences between the groups in negative symptoms or
Marital status, n (%)
social functioning nor was there any significant time effect
Single, never married 24 (47.1) 23 (45.1)
Cohabiting with a significant other 3 (5.9) 1 (2.0) within the groups for these measures. Follow-up means are
*Current student, n (%) 22 (43.1) 13 (25.5) presented in Table 2.
Comorbid diagnosis We also examined the change in SOPS positive symptom
Mood disorder, n (%) 13 (25.5) 13 (25.5)
ratings over months 1–5. There were no significant differ-
Anxiety disorder, n (%) 9 (17.7) 9 (17.7)
Alcohol abuse, n (%) 2 (3.9) 3 (5.9)
ences between the groups, but there was a significant time
Cannabis abuse, n (%) 5 (9.8) 5 (9.8) effect within the CBT group between baseline and 1,2,3,4 and
5 months (t ranged from 3.95 to 5.2, p b 0.05 to b0.0001) with
*p b 0.05.
most improvement occurring in the first three months.

3.3. Treatment exposure

In order to accommodate missing data and account for
intra-participant correlation over time, mixed effects models Overall the mean number of sessions was 12 (SD = 6.2,
for longitudinal data were used to compare the mean range 1–26). Thirty-one percent (n = 16) received less than 7
difference in positive and negative symptoms on the SOPS, sessions. Those who dropped out before the 6-month follow-
depression, anxiety, and social functioning over the 6, 12 and up had significantly fewer sessions (5 vs 13.4; t = 7.1,
18 month follow-up period. The analyses included explor- p b 0.0001). The reasons for not continuing are presented in
atory data analysis prior to building a mixed effects model, Table 3. Therapists completed rating forms after each session
mixed effects model development and interpretation and recording presenting problems for the sessions and the
model assessment. Means and standard deviations for both intervention strategies or techniques used in each session.
groups are presented in Table 2. There were no statistically The most common problems addressed for CBT were anxiety,
significant differences found between the groups on any of positive symptoms, relationships and work and school and
these measures at baseline. for ST the latter two problems only. The most often used
There was a rapid decline in the SOPS positive symptoms strategies presented in Table 4 support that ST focused on
for both groups, i.e. a strong negative relationship between engagement and support. However, much of the CBT sessions
visit and total SOPS positive score. There was a significant focused on assessment, and engagement and less time was
Probability of conversion to psychosis
1.00

ST
CBT
0.75
0.50
0.25
0.00

0 3 6 9 12 15 18 21 24 27
Follow-up time (months)
Number at risk*
ST 24 (2) 17 (1) 14 (0) 14 (0) 14 (0) 13 (0) 10 (0) 1 (0) 1 (0) 0
CBT 27 (0) 22 (0) 20 (0) 19 (0) 17 (0) 16 (0) 5 (0) 1 (0) 0 (0) 0

Note: ST = Support Therapy, CBT =Cognitive Behavior Therapy

*number of subjects converted to psychosis in parentheses

Fig. 2. Kaplan–Meier plot.

 J. Addington et al. / Schizophrenia Research 125 (2011) 54–61 59

Table 2
Outcome measures (means and standard deviations) for baseline, 6, 12 and 18 months.

Baseline 6 month 12 months 18 months

Measures CBT ST CBT ST CBT ST CBT ST Range

SOPS positive 10.8 (4.1) 12.3 (5.0) 6.4 (5.2) 7.6 (4.9) 5.2 (5.6) 6.6 (4.7) 4.6 (4.6) 4.5 (4.1) 0–30
SOPS negative 7.4 (4.5) 8.5 (5.9) 5.2 (5.8) 5 (4.8) 4.4 (5.1) 4.1 (5.6) 4.4 (4.3) 4.9 (5.3) 0–36
CDSS 3.5 (3.7) 6.5 (5.6) 2.9 (4) 2.1 (3.3) 3.3 (4.6) 2.5 (3.7) 2.6 (3.5) 1.9 (4.2) 0–27
SAS 48.7 (10.3) 54.6 (13) 41.5 (9) 48.1 (13) 42.5 (8.7) 45.3 (17.4) 43.2 (10.6) 46.8 (12.7) 0–100
SIAS 34.1 (15) 36.3 (16.2) 24 (15.9) 30.9 (15.7) 25.4 (18.2) 28.4 (18.6) 26.6 (15.9) 29.1 (18.6) 0–80
GAF 59.1 (13.2) 58.6 (11.1) 64.2 (14.4) 61.3 (9.9) 62.7 (12.3) 62.6 (10.2) 60.2 (17.9) 63.4 (11) 1–100
SFS 118.5 (15.9) 117.9 (21.9) 122 (22.8) 117.4 (15.7) 126.7 (19.8) 128.3 (16.6) 133.6 (16.3) 124.5 (22.5) 0–225

spent on core CBT strategies such as alternative solutions and
explanations, and metacognitions.
3.4. Alliance in therapy
The three sub-scores from the WAI-SF (therapeutic bond,
client and therapist agreement on goals, client and therapist
agreement on tasks) can range from 7 to 28. Mean sub-scores for
all three were between 21 and 27 suggesting a positive
relationship between therapists and participants. These sub-
scores did not differ between the two treatment groups. The one
exception was that at the second rating the CBT group felt that
the therapists were significantly more goal-directed (t=2.13,
pb 0.05). A comparison of the therapists' ratings with the
participants' ratings revealed no difference except for the
therapeutic bond. The therapists perceived the bond as greater
than the participants did at both times of rating (t=4.12,
pb 0.0001; t=2.29, pb 0.05).
4. Discussion
To the best of our knowledge these are the first published
results of an RCT comparing two different non-pharmacological
treatments for those at CHR, although there are other studies

Table 3
Non-completers and dropouts.

Subject Final assessment completed # Therapy sessions Reasons

Dropped out before 6-month assessment

4 baseline 1 ST Did not feel he needed treatment, reluctant to continue, then moved to
another province
34 baseline 3 ST Only wanted the assessment, was seeing a psychoanalyst 3× per week.
46 baseline 4 CBT Stopped coming and no contact
8 1 month 3 CBT Moved and dropped out
33 2 months 5 CBT Referred because he had to leave long-term youth therapist when he turned
18. Would not disengage from previous therapist, did not like structure/
homework of CBT, did not convert
10 3 months 10 CBT Doing well, never returned
12 3 months 7 CBT Felt better, chose to leave got a full time job, did not convert
18 3 months 5 CBT Mother only brought sporadically, then refused
20 3 months 6 ST Did well, work was an issue, chose not to come back
41 3 months 2 CBT Dropped out and unable to contact
50 3 months 5 ST Did not want to come, felt attending increased his coincidences and at times
paranoia
40 4 months 9 CBT Young girl keen to come but father refused to bring her, so dropped out
26 5 months 16 ST Relocated to Europe, did not convert

Dropped out after 6-month assessment

39 6 months only 3 ST Came because he wanted CBT — particularly strategies to help with his anxiety
19 6 months only 4 CBT Disorganized, lived alone far away enjoyed coming when he showed up
49 6 months only 10 CBT Did well but work and distance made it hard to continue with study
25 6 months only 19 CBT Did very well but work made him decide to leave the study after he attended
the 9-month follow-up for symptoms check

Dropped out after 12-month assessment

51 6 and 12 months 10 CBT Did well, liked therapy, moved away before end of 6 months to go to college
but returned for assessments
42 6 and 12 months 9 ST Had done well, completed University, then moved to UK
22 12 months only 6 ST Had been receiving therapy from a private therapist. Wanted more of a
problem solving approach, went back to his therapist. Did come back for the
12-month assessment

Completers of 18-month assessment with a low number of sessions

31 18 months only 4 CBT Moved too distant to attend regularly, returned for 18 months
21 12 and 18 months 5 CBT Poor attendance, did not like CBT strategies, wanted just to chat
60 J. Addington et al. / Schizophrenia Research 125 (2011) 54–61
Table 4
Percentage of time given to different interventions.
Intervention CBT
Assessment 42%
Goal setting 26%
Engagement 47%
Formulation 27%
Normalization 30%
Education 23%
Alternative solutions 13%
Alternative explanations 17%
Safety behaviors 5%
Metacognitive beliefs 7%
Core beliefs 24%
Relapse prevention 9%
Termination 15%
Crisis intervention 0%
Support 0%
Befriending 0%
underway or recently completed (Phillips et al., 2009). All three
conversions occurred in the supportive therapy group, but this
was not a significant difference. Participants in both treatment
groups made significant improvements in attenuated positive
symptoms, anxiety and depression and neither treatment
impacted negative symptoms nor poor functioning. Reductions
in attenuated positive symptoms, depression, and anxiety were
also observed over time in both the treated and untreated
group in the EDIE trial (French et al., 2007).
Although both groups had improved at six months, there
were no differences in positive symptoms between the
groups. However, an examination of the change in positive
symptoms over the first 5 months demonstrated that the CBT
group had an earlier and thus faster reduction in their
positive symptoms. This is similar to the results of the EDIE
trial (French et al., 2007). Interestingly, in the SoCRATES
study, a large CBT trial of first-episode subjects, one of the
main findings was that compared to the supportive therapy
group and the treatment as the usual group, the CBT group
made a more rapid recovery within the first 70 days of the
trial (Lewis et al., 2002).
There are four main reasons why there is a lack of difference
between the groups. First there may be no difference between
the treatments, i.e., both treatments are equally effective, and
CBT is not a superior treatment. Secondly, both treatments
could be equally ineffective and that the participants may have
recovered without any intervention. Thirdly, the study is
underpowered to detect a difference. Fourthly, the conversion
rate is lower than would have been anticipated. These four
reasons are addressed in detail below.
First there are several explanations that may, contrary to
expectations, account for CBT's lack of superiority. The CBT
group received an inadequate dose of CBT treatment. Table 3
reveals that, for many of the CBT cases, the interventions
focused primarily on engagement and less on the strategies
that are the core of CBT. Furthermore, the number of sessions
was limited which may have accounted for less time spent on
core CBT strategies. It may also be hard, therapeutically, to
engage young people who are uncertain about acknowledg-
ing a potential illness. These are potentially therapist issues in
that the therapists were not doctoral level CBT specialists.
Perhaps these clients are extremely complex in their
presentation and require highly skilled CBT specialists. The
study therapists were skilled and had CBT training, but may
Support
be more typical of therapists that would be available to this
population in routine mental health clinics. Future studies
31%
will have to test the need to have more expert therapists or at
0%
43% least more intensive supervision. A fourth explanation is that
0% the CBT was used to address several presenting problems, and
0% the participants had different needs. This issue was reported
0%
in an excellent, methodologically sound, large CBT for
0%
0% psychosis relapse trial in the UK where there was no
0% difference in CBT vs ST (Garety et al., 2008). The UK study
0% reported general vs specific benefits to patients when there
0% was at times an absence of target symptoms. A similar
0%
recommendation may apply here in that CBT should be
0%
1% directed at targeting and improving attenuated symptoms.
75% Secondly, in the absence of a treatment as usual/standard
18% care control condition definitive conclusions are limited and
the groups may have improved without treatment. There is
no “treatment as usual” for this population, and as they were
help-seeking “no treatment” was not considered a reasonable
option. The ST was an attempt to create a nonspecific but
standardized treatment as usual that would serve as an
appropriate comparison to test the efficacy of CBT.
Thirdly, the study is underpowered, and the sample is
clearly too small to detect a difference although all effects
were in the predicted direction.
Finally, the conversion rates were much lower than
expected, and for a few the final conversion status is
unknown. Approximately 40% of the sample did not complete
the 18-month follow-up. Table 2, however, suggests that
several of these young people left the study when they felt
they had made some improvement; for others the time
involved and travel were an issue. However, we were able to
contact some of the “dropouts” to ascertain their status with
respect to conversion after 1 year and of those who did not
complete one year we are uncertain about the status of only
12 participants. Compared to earlier studies the conversion
rate in this sample is low, a phenomenon that is being
reported in other research centers (Yung et al., 2007). This is
clearly a problem for future treatment studies as sample sizes
will have to be considerably larger than the sample originally
planned for this study.
This area is relatively new, and future research into the
longitudinal development of psychosis is indicated as is an
improved understanding of the potential impact of different
interventions. Although clearly limited by its sample size, this
study raises stimulating questions in terms of future studies
of psychological treatments for this population. How much
therapy would be enough to make a difference? Should
supportive therapy, which is potentially a less costly therapy,
be offered earlier than CBT in a stage model of treatment
(McGorry et al., 2006)? Should CBT be used specifically to
target attenuated positive symptoms, and other therapies
used for different problems e.g. social skills deficits? That is,
simple interventions offering support and problem solving
may be helpful for these CHR young people when they first
seek help. Structured CBT strategies possibly requiring more
experienced therapists should be reserved for targeting specific
problems such as severe attenuated positive symptoms.
In conclusion, help-seeking participants who meet criteria
for being at-risk for psychosis do improve when provided
 J. Addington et al. / Schizophrenia Research 125 (2011) 54–61
with therapy. However, this study did not find that CBT was
more effective than a supportive therapy. The results of this
study underscore critical points that now need to be
considered in future trials including the need to ensure that
the treatments can be clearly distinguished from one another,
that treatment dose and duration are adequate, level of
therapist expertise be defined and that desired outcomes be
clearly operationalized.
Role of funding source
This work was supported by a grant to Jean Addington from the Ontario
Mental Health Research Foundation, Ontario Canada. The funding source had
no involvement in the decision to write the paper or in the decision to submit
the paper for publication.
Contributors
Jean Addington was responsible for the design of the study, writing the
protocol, managing and overseeing all of the treatment and data collection,
and writing the first draft of the paper. Drs Zipursky, Boydell and French
made contributions to the design of the study and writing the protocol. Drs
Zipursky, Epstein and French made contributions to the managing and
overseeing various aspects of the study, Ms Liu was responsible for statistical
analysis and all authors contributed to and approved the final manuscript.
Conflict of Interest
Dr Jean Addington currently receives funds from NIMH and Alberta Heritage
Foundation for Medical Research. Drs Addington, Boydell, French, Zipursky and
Epstein report no financial relationships with commercial interests.
Acknowledgements
We would like to thank the following people for their work on this
project: I. Furimsky, M. Haarmans, D. Kirsopp, E. Mancuso, A. McCleery, S.
McMillan, R. Rabin, H. Saeedi and L. Tran.
References
Addington, D., Addington, J., Maticka-Tyndale, E., 1993. Assessing depression
in schizophrenia: the Calgary Depression Scale. British Journal of
Psychiatry 163, 39–44.
Addington, J., Epstein, I., Reynolds, A., Furimsky, I., Rudy, L., Mancini, B.,
McMillan, S., Kirsopp, D., Zipursky, R.B., 2008. Early detection of psychosis:
finding those at clinical high risk. Early Intervention in Psychiatry 2,
147–153.
Amminger, G.P., Schafer, M.R., Papageorgiou, K., Klier, C.M., Cotton, S.M., Harrigan,
S.M., Mackinnon, A., McGorry, P.D., Berger, G.E., 2010. Long-chain omega-3
fatty acids for indicated prevention of psychotic disorders: a randomized,
placebo-controlled trial. Archives of General Psychiatry 67, 146–154.
Bentall, R.P., Morrison, A.P., 2002. More harm than good: the case against
using anti-psychotic drugs to prevent severe mental illness. Journal of
Mental Health 11, 351–356.
Birchwood, M., Smith, J., Cochrane, R., Wetton, S., Copestake, S., 1990. The Social
Functioning Scale: the development and validation of a new scale
adjustment for use in family intervention programmes with schizophrenic
patients. British Journal of Psychiatry 157, 853–859.
Cannon, T.D., Cadenhead, K., Cornblatt, B., Woods, S.W., Addington, J., Walker, E.,
Seidman, L.J., Perkins, D., Tsuang, M., McGlashan, T., Heinssen, R., 2008.
Prediction of psychosis in youth at high clinical risk: a multisite longitudinal
study in North America. Archives of General Psychiatry 65, 28–37.
Dobson, K., Shaw, B., Vallis, T., 1985. The reliability of competence ratings on
cognitive–behavior therapists. British Journal of Clinical Psychology 24,
295–300.
French, P., Morrison, A.P., 2004. Early Detection and Cognitive Therapy for
People at High Risk of Developing Psychosis. Wiley, London.
French, P., Shryane, N., Bentall, R.P., Lewis, S.W., Morrison, A.P., 2007. Effects of
cognitive therapy on the longitudinal development of psychotic experi-
ences in people at high risk of developing psychosis. British Journal of
Psychiatry 51, s82–s87.
Garety, P.A., Fowler, D.G., Freeman, D., Bebbington, P., Dunn, G., Kuipers, E.,
2008. Cognitive–behavioural therapy and family intervention for relapse
prevention and symptom reduction in psychosis: randomised controlled
trial. British Journal of Psychiatry 192, 412–423.
Lewis, S., Tarrier, N., Haddock, G., Bentall, R., Kinderman, P., Kingdon, D.,
Siddle, R., Drake, R., Everitt, J., Leadley, K., Benn, A., Grazebrook, K., Haley,
61
C.A.S., Davies, L., Palmer, S., Faragher, E.B., Dunn, G., 2002. Randomised
controlled trial of cognitive–behavioral therapy in early schizophrenia:
acute-phase outcomes. British Journal of Psychiatry 181, s91–s97.
McGlashan, T.H., Zipursky, R., Perkins, D.O., Addington, J., Miller, T.J., Woods,
S.W., Hawkins, K.A., Hoffman, R.E., Lindborg, S., Ohen, M., Reier, A., 2003.
The PRIME North America randomized double blind clinical trial of
olanzapine vs placebo in patients at risk for being prodromally
symptomatic for psychosis: I Study rationale and design. Schizophrenia
Research 61, 7–18.
McGlashan, T.H., Zipursky, R.B., Perkins, D., Addington, J., Miller, T., Woods, S.W.,
Hawkins, K.A., Hoffman, R.E., Preda, A., Epstein, I., Addington, D., Lindborg,
S., Trzaskoma, Q., Tohen, M., Breier, A., 2006. Randomized, double-blind
trial of olanzapine versus placebo in patients prodromally symptomatic for
psychosis. American Journal of Psychiatry 163, 790–799.
McGorry, P.D., Yung, A., Phillips, L., Yuen, H., Francey, S., Cosgrave, E.,
Germano, D., Bravin, J., McDonald, T., Blair, A., Adlard, S., Jackson, H.,
2002. Randomized controlled trial of interventions designed to reduce
risk of progression to first-episode psychosis in a clinical sample with
subthreshold symptoms. Archives of General Psychiatry 59, 921–928.
McGorry, P.D., Hickie, I.B., Yung, A.R., Pantelis, C., Jackson, H.J., 2006. Clinical
staging of psychiatric disorders: a heuristic framework for choosing
earlier, safer and more effective interventions. Australian and New
Zealand Journal of Psychiatry 40, 616–622.
Miller, T.J., McGlashan, T.H., Rosen, J.L., Somjee, L., Markovich, P.J., Stein, K.,
Woods, S.W., 2002. Prospective diagnosis of the initial prodrome for
schizophrenia based on the structured interview for prodromal
syndromes: preliminary evidence of interrater reliability and predictive
validity. Am J Psychiatry 159, 863–865.
Miller, T.J., Zipursky, R.B., Perkins, D., Addington, J., Woods, S.W., Hawkins, K.A.,
Hoffman, R., Preda, A., Epstein, I., Addington, D., Lindborg, S., Marquez, E.,
Tohen, M., Breier, A., McGlashan, T.H., 2003. The PRIME North America
randomized double-blind clinical trial of olanzapine versus placebo in
patients at risk of being prodromally symptomatic for psychosis. II. Baseline
characteristics of the “prodromal” sample. Schizophrenia Research 61, 19–30.
Morrison, A.P., French, P., Walford, L., Lewis, S.W., Kilcommons, A., Green, J.,
Parker, S., Bentall, R.P., 2004. Cognitive therapy for the prevention of
psychosis in people at ultra-high risk: randomised controlled trial.
British Journal of Psychiatry 185, 291–297.
Morrison, A.P., French, P., Parker, S., Roberts, M., Stevens, H., Bentall, R.P.,
Lewis, S.W., 2007. Three-year follow-up of a randomized controlled trial
of cognitive therapy for the prevention of psychosis in people at
ultrahigh risk. Schizophrenia Bulletin 33, 682–687.
Olivares, J., Garcia_Lopez, L.J., Hidalgo, M.D., 2001. The Social Phobia Scale
and the Social Interaction Anxiety Scale: factor structure and reliability
in a Spanish speaking population. Journal of Psychoeducational Assess-
ment 19, 69–80.
Phillips, L.J., Nelson, B., Yuen, H.P., Francey, S.M., Simmons, M., Stanford, C.,
Ross, M., Kelly, D., Baker, K., Conus, P., Amminger, P., Trumpler, F., Yun, Y.,
Lim, M., McNab, C., Yung, A.R., McGorry, P.D., 2009. Randomized
controlled trial of interventions for young people at ultra-high risk of
psychosis: study design and baseline characteristics. Australian and New
Zealand Journal of Psychiatry 43, 818–829.
Popcock, S., 1983. Clinical Trials. John Wiley & Sons Ltd, Chicester UK.
Ruhrmann, S., Schultze-Lutter, F., Salokangas, R.K., Heinimaa, M., Linszen, D.,
Dingemans, P., Birchwood, M., Patterson, P., Juckel, G., Heinz, A.,
Morrison, A., Lewis, S., von Reventlow, H.G., Klosterkotter, J., 2010.
Prediction of psychosis in adolescents and young adults at high risk:
results from the prospective European prediction of psychosis study.
Archives of General Psychiatry 67, 241–251.
Streiner, D.L., 1990. Sample size and power in pscyhiatric research. Canadian
Journal of Psychiatry 35, 616–620.
Tarrier, N., Lewis, S., Haddock, G., Bentall, R., Drake, R., Kinderman, P.,
Kingdon, D., Siddle, R., Everitt, J., Leadley, K., Benn, A., Grazebrook, K.,
Haley, C.J., Akhtar, S., Davies, L., Palmer, S., Dunn, G., 2004. Cognitive–
behavioral therapy in first-episode and early schizophrenia. British
Journal of Psychiatry 184, 231–239.
Tracey, T., Kokotovic, A., 1989. Factor Structure of the Working Alliance
Inventory. Psychological Assessment 1, 207–210.
Vallis, T., Shaw, B., Dobson, K., 1986. The Cognitive Therapy Scale:
psychometric properties. Journal of Consulting and Clinical Psychology
54, 381–385.
Yung, A.R., McGorry, P.D., McFarlane, C.A., Jackson, H.J., Patton, G.C., Rakkar,
A., 1996. Monitoring and care of young people at incipient risk of
psychosis. Schizophrenia Bulletin 22, 283–303.
Yung, A.R., Phillips, L.J., Yuen, H., Francey, S., McFarlane, C.A., Hallgreen, M.A.,
McGorry, P.D., 2003. Psychosis prediction: a 12-month follow-up of a
high-risk (“prodromal”) group. Schizophrenia Research 60, 21–32.
Yung, A.R., Yuen, H.P., Berger, G., Francey, S., Hung, T.C., Nelson, B., Phillips, L.,
McGorry, P., 2007. Declining transition rate in ultra high risk (prodromal)
services: dilution or reduction of risk? Schizophrenia Bulletin 33, 673–681.