Last updated on March 2019

Manualized Group Cognitive-behavioral Therapy for Social Anxiety in First Episode Psychosis

Brief description of study

Social anxiety represents one of the most prevalent comorbid conditions in schizophrenia and related psychosis. Schizophrenia patients with comorbid social anxiety often exhibit impaired social functioning, an increased risk for relapse, and higher rates of suicide. Social anxiety is a treatable condition but has, in the context of psychosis, received only scant attention thus far. There is strong evidence that cognitive-behavioral therapy (CBT) for the treatment of social anxiety is very effective, whether it is delivered individually or in a group setting, and studies have shown that a group setting is more effective than individual therapy. Providing a CBT intervention for social anxiety represents an effective way to empower people with this illness.

The investigators have conducted a preliminary study using an uncontrolled design to assess feasibility and initial benefits of a new manualized group CBT intervention for social anxiety specifically adapted for people with psychosis. The investigators observed a significant reduction in social anxiety symptoms across three groups of first episode psychosis (FEP) participants (n=29) following completion of this 13-week intervention, and observed large effect sizes confirming a significant positive influence of this intervention. The investigators now propose to conduct a randomized controlled trial to fully assess the efficacy of this intervention. The main objective of this research proposal is to contrast the impact of a CBT intervention for the treatment of social anxiety in first episode psychosis with another control condition involving computer assisted cognitive remediation therapy (CACRT). Both interventions will be offered in a group setting, and will therefore have the exact same parameters. A secondary objective of this study is to examine the impact of reduced social anxiety on measures of clinical and functional outcome.

For this trial, 120 patients with recent onset psychotic disorder (defined as within 5 years from their first episode of psychosis) and with social anxiety will be clinically assessed. These participants will be recruited from five different first episode psychosis programs in the Montreal area and referred by their treatment team. They will then be randomly assigned to either the CBT or CACRT conditions. Both interventions will involve 13 weekly group sessions. At the end of group interventions and at two follow-ups (3-month & 6-month), the presence and severity of social anxiety symptoms will be assessed. It is hypothesized that compared to the CACRT group, individuals receiving the CBT intervention will show a reduction in symptoms associated with social anxiety (as determined with multiple self-report and clinician rated measures). This effect will be maintained at follow-ups. In addition, the investigators also hypothesize that the CBT group will show better clinical outcome, defined as the length of symptomatic remission at follow-ups. For functional outcome, they will show significant improvement on a self-report measure a clinician-rated measure of recovery. This study will be one of the first to specifically target social anxiety in people with psychosis using a psychosocial intervention. As such, it will tackle an important problem that is interfering with recovery and with the actualization of functional roles.

Detailed Study Description


Social anxiety is highly prevalent in people with schizophrenia and related psychotic disorders, and represents a major obstacle to a positive functional outcome. The investigators have recently developed a psychosocial intervention for social anxiety that is specifically tailored to people with psychosis. An adapted Cognitive-Behavioural Therapy (CBT) intervention for psychosis must target the stigma attached to the illness, the possible presence of poor social skills, delusional and persecutory ideas, potential limited reading abilities, and associated cognitive deficits that are the hallmark of schizophrenia and related psychoses. Our preliminary data from an uncontrolled evaluation of our new intervention in first episode psychosis revealed a significant positive impact of this 13-week group intervention for manifestations of social anxiety. Following these promising results, the investigators now propose to conduct a multi-center randomized controlled trial to test the efficacy of this intervention relative to an active control condition consisting of a cognitive remediation group. Their primary goal is to show the efficacy of this intervention on symptoms of social anxiety as determined by self-report measures and clinician-rated measures. Their secondary goal is to examine the impact of such an intervention on clinical and functional outcomes. The highly impairing nature of social anxiety in psychosis is well established and the proposed trial will ultimately show that treating this aspect of the condition is possible and will result in improved functioning.

From 2010 to 2012, the investigators have offered their CBT intervention for social anxiety to several groups of patients at the Douglas Institute and now, more recently, at the McGill University Health Centre (MUHC). The investigators have employed an uncontrolled design recommended to evaluate the feasibility and potential benefits of a pilot psychosocial intervention. In their feasibility study, they offered this intervention to five groups of clients who were part of the Prevention and Early Intervention Program for Psychosis (PEPP-Montreal) and to two groups of clients with longer standing illness, from the Psychosis Program at the Douglas Institute and from the MUHC, respectively. All participants were referred by their treating team (e.g. psychiatrist, nurse, case manager, etc.) and were screened by our research team to evaluate the presence of significant social anxiety. They assessed the effects of this intervention using four scales: Social Interaction Anxiety Scale (SIAS), the Social Phobia Inventory (SPIN) and the Brief Social Phobia Scale (BSPS), Structured Clinical Diagnosis (SCID-I) for Diagnostic and Statistical Manual (DSM-IV).

Focusing on their work in FEP at our program PEPP-Montreal, 25 out of 29 participants who were recruited for the feasibility study met diagnostic criteria for social phobia, whereas 4 represented 'borderline' cases that were clearly reporting significant difficulties with social interactions and active avoidance of social situations. From the 29 clients who started the group intervention, 3 did not complete it (completion defined as attending more than 50% of the sessions). The overall attendance rate of completers of the weekly 13-week program was excellent, with an average of 11 sessions attended per participant (ranging from 9-13), confirming that the intervention was well tolerated by the participants. For the 26 patients who completed the intervention, a significant reduction in social anxiety was observed on all three measures (SPIN, SIAS, BSPS) at the end of the 13-week treatment. We observed effect sizes of 1.04 on the SIAS, 0.93 on the SPIN, and 0.95 on the BSPS, after correcting for within-subject dependence among means for the total sample of participants who completed the intervention. These effect sizes are quite large and represent a strong justification to pursue a larger-scale investigation of this novel intervention. Critically, from the 24 participants with a SCID-I based diagnosis of social phobia prior to the treatment, only two still met diagnostic criteria at the end of the study. Moreover, the investigators administered the Internalized Stigma of Mental Illness (ISMI) scale and observed a significant reduction in stigma following this intervention (p<.01), suggesting that our intervention does indeed target stigma in an efficient way.

Having established the potential benefits of our new psychosocial intervention for social anxiety, the investigators devised a control condition to reproduce some of the aspects of our group intervention without the active ingredients of group CBT. An active control condition that could control for spontaneous changes in symptoms of social anxiety, the effect of being in a trial, and therapeutic commitment was chosen. In the selection of an active control condition, it is essential that this intervention does not target social anxiety while still controlling for the non-specific effects described above. As discussed by Everitt, a control intervention that does not target the symptoms of interest should potentially lead to a greater effect size between the experimental treatment and the control condition, and as such requires a smaller sample. Several researchers have attested to the appropriateness of cognitive remediation therapy (CRT) as a control condition for CBT intervention: two recent large multi-center trials involving a group CBT intervention to improve patients' insight into their illness or to improve negative symptoms have used CRT as their active control condition. CRT targetting cognitive performance can be offered in a group format and thus represents a very promising control condition for the proposed study.

Several computer-assisted cognitive remediation therapies (CACRT) have been developed in the last few years. Many of these programs developed specifically for psychiatric patients follow a structured protocol and are given in small groups under supervision of a trained therapist. The goals of these programs are to improve performance on cognitive domains that have been determined to hinder functioning. The intervention aims to provide a positive experience for patients through the promotion of positive attitudes towards learning, and patients are encouraged to develop independent learning skills as well as a sense of competence and confidence. The intervention is also designed to increase awareness of the strengths and weaknesses of the patients' own learning styles. This particular method is advantageous because its goals match other psychiatric rehabilitation models and because it has theoretical bases in neuropsychology, learning, educational psychology, rehabilitation psychology, self-determination, and client-centered therapy. Furthermore, CRT has been shown to be effective at improving patients' cognitive abilities but has less impact on social functioning. Since fall 2011, the investigators have developed a group CRT program in collaboration with the Psychosis Program of the Douglas Institute and have presented their results at a recent scientific conference. The investigators propose to use Dr. Christopher Bowie's protocol in the current study. This psychosocial intervention represents an excellent control condition for their randomized controlled trial: it is conducted in a group format comparable to their proposed group CBT intervention, which controls for potential non-specific effects of simply participating in a group activity. Additionally, although this is a group activity, social interactions are limited as each participant spends most of the time working individually on a computer.

Proposed study

General aim: The goal of this research proposal is to evaluate the efficacy of a group CBT intervention (experimental condition) for social anxiety in first-episode psychosis in comparison to group-based cognitive remediation (control condition) on reducing symptoms of social anxiety.

Primary objective: To compare the efficacy of a 13-week group CBT intervention adapted for people with psychosis for reducing symptoms of social anxiety to an active control condition comprised of a 13-week group cognitive remediation therapy.

Primary hypothesis: Compared to the active control group, individuals receiving the brief manualized CBT intervention will show a reduction in symptoms associated with social anxiety, as determined with the Social Interaction Anxiety Scale (SIAS), the Social Phobia Inventory (SPIN) and the Brief Social Phobia Scale (BSPS) at the end of 13 weeks of intervention and at follow-up assessments (3- and 6-months). In addition, fewer participants receiving CBT will meet DSM-IV diagnostic criteria for social phobia at the end of the study.

Secondary objective: To measure the effect of a group CBT intervention for social anxiety in first episode psychosis on clinical outcome and functional outcome.

Secondary hypothesis: Following the treatment trial the CBT group will show better clinical outcome relative to the control group, defined as the length of symptomatic remission at follow-ups. For functional outcome, they will show significant improvement on a self-report measure of recovery (the Recovery Assessment Scale) and on a clinician-rated measure, the Social and Occupational Functioning Scale (SOFAS) at 3- and 6-month follow-ups.


Study design: A randomized controlled trial will be conducted comparing the experimental group, which will be receiving group CBT for social anxiety plus standard care, to the active control group who will be receiving CACRT and standard care. During the randomization process and repartition of participants, the two groups will be matched on the severity of social anxiety, level of functioning and gender (as described below) in order to avoid any initial group inequality. Participants will be assessed before and after treatment. They will also be asked to return for completion of two follow-up assessments (at 3- and 6-months) in order to measure the effect of treatment over time and verify whether the hypothesized therapeutic gains are maintained. As part of these follow-up evaluations, participants will complete various questionnaires and scales.

Participants: This study will take place in Montreal and will involve five first-episode psychosis clinics affiliated with McGill University or Universit de Montral (UdeM). These include PEPP-Montral, the Douglas Institute (Director Dr Ashok Malla, co-investigator), the First Episode Psychosis Program from the Jewish General Hospital (Director Dr. G. Eric Jarvis, co-investigator), and the First Episode Psychosis Program from the MUHC (Director Dr. Howard Margolese, co-investigator) from the McGill network, and The Clinique Jeunes Adultes Psychotiques from the Centre hospitalier de l'Universit de Montral (CHUM) (Director Dr. Amal Abdel-Baki, co-investigator) and the Programme des Premiers pisodes Psychotiques from Institut Universitaire de Sant Mentale de Montral (Director Dr. Luc Nicole, co-investigator) from UdeM. These related first episode psychosis programs all provide a complete range of services for adults suffering from psychosis, including evaluation, prevention, treatment, hospitalization, outpatient and family services, rehabilitation, modified assertive community case management, and variable community follow-ups. These programs provide follow-up for up to five years. Collectively, these five first-episode programs service the entire population of Montreal (2.2 million) and have more than 1000 active patients in addition to accepting about 300 new clients per year.

Screening of potential participants: Recruitment of participants will be facilitated by the co-investigators, who are clinical heads at the different services involved in this study, and by the culture of integration of clinical care and research that characterizes these different programs.

Clinical assessment: After the screening is performed and informed consent is obtained, the participant will be administered a Structured Clinical Diagnosis (SCID) for DSM-IV to confirm the diagnosis of a psychotic disorder in collaboration with a research psychiatrist. If the diagnosis is confirmed and eligibility for the study is established, the participant will undergo a thorough clinical assessment at intake, 3-month and 6-month follow-ups, and at the end of therapy. Symptom evaluation will include the Scale for Assessment of Positive Symptoms (SAPS) and the Scale for Assessment of Negative Symptoms (SANS), as well as the Calgary Depression Scale. Social anxiety will be assessed via three scales: the SIAS is a 20-item scale which measures anxiety in interpersonal encounters; the SPIN is a 17-item scale assessing multiple dimensions of social anxiety including fear, avoidance and physiological discomfort; the Brief Social Phobia Scale (BSPS) is an 11-item clinician-rated assessment scale developed by Davidson which specifically measures fear, avoidance and autonomic physiological responses that are usually associated with most common social situations. In addition, the social phobia module of the SCID-I will be used to confirm a diagnosis of social phobia. We have extensive experience in administering these instruments and our raters have achieved a high level of inter-rater reliability (ICC 0.79-.92). Finally, the Recovery Assessment Scale (RAS) and the Social and Occupational Functioning Scale (SOFAS) will be administered. The RAS is a 41-item self-report measure examining multiple dimensions of functional recovery as perceived by the patient him/herself. SOFAS is used to evaluate the individual's level of social and occupational functioning independent of the overall severity of psychiatric symptoms. The ratings are done first to allocate a category (1-10) based on the description of that category and how it applies to the individual, followed by a numerical score within that category.

Neuropsychological assessment: Initial assessment of participants will include the Wechsler Abbreviated Scale of Intelligence (WASI), a pen and paper intelligence quotient (IQ) test consisting of 4 sub-scales (Vocabulary, Similarities, Block Design, Matrix Reasoning). It will only be administered once as part of initial assessment before treatment. Participants will also be evaluated using CogState, a computerized cognitive battery of 11 tasks that evaluate 7 cognitive domains (reasoning/problem solving, speed of processing, attention/vigilance, visual learning and memory, verbal learning and memory, working memory and social cognition). Both will be administered by a trained evaluator under the supervision of a neuropsychologist. The CogState is a repeatable neuropsychological battery that was specifically designed for schizophrenia and related psychoses. This assessment will be used to: confirm that participants present an IQ of at least 70; explore any impact of CACRT on neurocognitive performance; explore whether cognitive performance is predictive of treatment response in CBT, as some suggest.

Procedure for participant recruitment and group randomization: Potential participants will be identified by clinical staff, in collaboration with the project coordinator, within the participating early psychosis programs. In all cases, only clients who will have indicated to their treating team a wish to learn more about the research study will be referred to the investigators. Once formal consent is obtained and an initial screening procedure completed, the research coordinator will assess the participant's overall functioning and eligibility for the current study by conducting the Global Assessment Scale (GAF), which has been shown in numerous studies to be a reliable and valid clinical measure of functioning. This measure will allow researchers to match participants not only on cognitive abilities (i.e. intermediate reading level) and language (i.e. informal participant-report of language proficiency) but also on their functionality. The score used for stratification will be a score of 50 or more, which will identify "high functional" versus "low functional" participants. In addition, participants will be asked to fill out the SIAS, SPIN and BSPS, and those with a score above the cut-off values described above will be approached for this study. A brief description of both psychosocial interventions (CBT and CACRT) will be offered by the referral source and the case will be referred to the research coordinator. During this initial session, the research staff will conduct a semi-structured interview, the Social Phobia / Anxiety sub-scale of the SCID-I. Random assignment will be carried out using a computer-run randomization protocol, which improves upon chance allocation to a given group by adjusting assignment probabilities based upon the frequency and characteristics of current group membership on key intake characteristics that could influence outcomes. A biostatistician (co-applicant Norbert Schmitz) not connected with treatment in any capacity and located away from the treatment site will carry out the randomization. Recruitment of participants and symptom assessments will be conducted, respectively, by a research coordinator and a research assistant, and the latter will be blind to treatment. After random assignment is complete, an appointment for the therapy session will be scheduled for those assigned to receive CBT and to those assigned to receive CACRT.

We expect to take four years to complete the recruitment for this study.

Treatment fidelity and quality control measures: To ensure that both interventions are being carried out in accordance with the detailed guidelines provided in the respective manuals, we will put in place quality control measures. For group CBT, we will use the approach developed by one of the co-investigator (T.L.) that is specific to group intervention in early psychosis. Briefly, all sessions will be audiotaped and the group adaptation of the Cognitive- Behavioral Therapy Scale will then be completed by two experienced CBT therapists. The audiotapes of the first 5 sessions of every group will be reviewed by two experienced CBT therapists and they will offer supervision to the therapists. Afterward, approximately 1 in every 3 sessions will be examined, or more if requested. For the active control condition involving cognitive remediation, Dr. Bowie (co-investigator), who has extensive experience in this field, will be responsible for quality control. Dr. Bowie has already conducted several randomized controlled trials using cognitive remediation interventions. Therapists will be trained by a registered clinical psychologist who is a senior clinician with over 12 years of experience with cognitive remediation.

Medication: Medications will be monitored monthly using information from multiple sources (case managers, psychiatrists, and patients) to ensure that results are not due to changes in medication.

Blinding the evaluations: While it is well acknowledged that conducting a classical double blind evaluation in a randomized controlled trial (RCT) of a psychosocial intervention is not feasible, we believe it is important to conduct blind evaluations of the study outcomes. Therefore, all evaluations will be conducted by a trained bilingual rater who will have no involvement in the conduct of the study, the delivery of the interventions or any involvement with any trial-related meetings.

Clinical trial management: The steering committee will meet on a regular basis, about once every three months.

Clinical Study Identifier: NCT02294409

Find a site near you

Start Over