Cognitive behaviour therapy for psychosis can be adapted for minority ethnic groups: A randomised controlled trial ^☆

• Shanaya Rathod^a, ,
• Peter Phiri^{a, b, ,} ,
• Scott Harris^c,
• Charlotte Underwood^a,
• Mahesh Thagadur^a,
• Uma Padmanabi^a,
• David Kingdon^b

• ^a Southern Health NHS Foundation Trust, UK
• ^b Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, UK
• ^c Public Health Sciences & Medical Statistics, Faculty of Medicine, University of Southampton, UK

Corresponding author at: Research & Outcomes Dept., Tom Rudd Unit, Moorgreen Hospital Botley Road, West End, Southampton, SO30 3JB, UK. Tel.: + 44 23 8047 5258; fax: + 44 23 8047 3022.

Received 23 August 2012

Revised 19 October 2012

Accepted 3 November 2012

Available online 8 December 2012

• http://dx.doi.org/10.1016/j.schres.2012.11.007, How to Cite or Link Using DOI

---

Keywords

• Psychosis;
• Culturally adapted;
• Cognitive therapy;
• Ethnicity;
• African Caribbean;
• South Asian Muslim;
• Ethnic minority;
• Schizophrenia

---

1. Introduction

Cultural influences are reported in the clinical manifestation, prevalence, treatment access and outcomes for individuals with schizophrenia. Cantor-Graae and Selten (2005) meta-analysis of 18 studies demonstrated a significantly increased risk of schizophrenia in migrant groups from developing countries with variation of risk by both host countries and countries of origin.

Cognitive Behavioural Therapy (CBT) for schizophrenia is an evidence-based adjunct to medication and recommended internationally (Dixon et al., 2009 and National Institute for Clinical Excellence, 2009). However, clients from some black and minority ethnic (BME) groups e.g. the African Caribbean and Black African groups have shown higher dropout rates and poorer outcomes compared with the White group (Rathod et al., 2005). Cultural adaptation and understanding of ethnic, cultural and religious interpretations remains an underdeveloped area (Rathod et al., 2008).

Griner and Smith (2006) have shown an effect size of 0.45 for culturally adapted evidence based interventions in comparison to traditional treatments (Wykes et al., 2008) whilst the meta-analysis by Huey and Polo (2008) was inconclusive. Evidence from small pilot studies suggests that locally adapted CBT with minority populations has been successful (Carter et al., 2003 and Kubany et al., 2003; Hinton et al., 2004 and Hinton et al., 2005; Patel et al., 2007, Rojas et al., 2007 and Rahman et al., 2008). Rathod et al. (2010) following a recent qualitative study reported that in principle, culturally adapted CBT for psychosis would be acceptable to BME clients. The recommendations from this study were incorporated into a CBT manual (Kingdon and Turkington, 2005) using Tseng et al.'s (2005) framework of cultural adaptations of psychological therapies. The authors considered several adaption models (including Bernal et al., 1995, Domenech Rodriguez and Weiling, 2004, Barrera and Castro, 2006, Hays and Iwamasa, 2006, Hwang, 2006 and Leong and Lee, 2006) and agreed that Tseng's framework would allow fidelity to the core principles of CBT with adequate flexibility for adapting the therapy to cultural beliefs, thereby preserving validity to the original treatment. The authors acknowledged that in practice, even if desirable, it may not be feasible to develop a different CBT for every cultural group and subgroup within them as every cultural group and subgroup had their own uniqueness and could not be considered as one.

2. Aims

This study aimed to:

a)

Assess the feasibility of the culturally adapted CBT for psychosis (CaCBTp) with specified BME groups.

b)

Further modify CaCBTp in accordance with emerging findings.

3. Methodology

3.1. Study design

This was a single blind multi-site randomised trial of CaCBTp for ethnic minority participants compared to treatment as usual (TAU) with a 6 month follow-up period after completion of intervention.

3.1.1. Study centres

The study was conducted at two sites in the UK—Hampshire (Southampton/Portsmouth) and London (Central and North West London).

3.2. Participants

3.2.1. Inclusion criteria

Participants were eligible if they were:

1.

Between ages 18 and 65 with a diagnosis of schizophrenia, schizoaffective disorder or delusional disorders using ICD-10 Research Criteria.

2.

From the following groups:

•

Black British, Black Caribbean or African Caribbean (all three terms usually refer to people of Caribbean origin with Caribbean origin parents and heritage, even if they are born in the UK themselves).

•

South Asian Muslim (Pakistani and Bangladeshi—refer to people of Muslim religion who either have their origins in South Asia or their parents and heritage are).

The rationale for choosing the specified ethnic groups has been discussed in a publication of the previous study (Rathod et al., 2010).

3.

Willing to participate in the interview and/or be tape recorded.

4.

Had mental capacity to consent and participate.

5.

Able to speak English or were willing to participate with interpreters.

3.2.2. Exclusion criteria

1.

Severe illness which would affect ability to participate in assessments or therapy e.g. very thought disordered or distressed by symptoms.

2.

Lacked mental capacity or denied consent.

3.

The treating clinical team thought was inappropriate. For e.g. if they were due to receive CBT through their services as standard treatment and being in the trial could mean they may be randomised to TAU arm.

3.2.3. Recruitment of participants

Power calculations used previous pilot studies (e.g. Turkington and Kingdon, 2000). This study aimed to recruit forty participants to ensure that after the expected loss to follow-up, there would be at least 12 participants per group for analysis to be able to make reasonable estimates of the treatment effects (FDA guidance http://www.fda.gov/cder/guidance/5356fnl.pdf). Recruitment was conducted from May 2009 to December 2010 through Community Mental Health Teams (CMHTs), specialist services including Early Intervention in Psychosis (EIP) teams, Assertive Outreach Teams (AOT), Rehabilitation and Inpatient services. Permission to approach suitable potential participants was obtained from consultant psychiatrists and clinical teams. Participant information sheets were given by the consultant and/or care co-ordinators who explained the study in detail.

3.3. Ethical considerations

Ethics approval was obtained from Southampton Committee (B); REC: 08/0504/5—Trial number: ISRCTN95603741. Research and Governance approval was granted at all sites. Confidentiality was protected by anonymisation of data, and only where written informed consent was unambiguously given, participants were enrolled in the trial. Data was collected in an overtly non-coercive manner. The research team included registered practitioners who were CRB (Criminal Records Bureau) approved.

Written informed consent was obtained following assessment of mental capacity to consent specifically to the research study. This was a requirement by the Ethics committee. Assessment of capacity was based on the five core principles of the Mental Capacity Act (2005) Section 3 (Department of Constitutional Affairs, 2007). Capacity to consent was specific to the research study.

The Consultant and/or Care co-ordinator was encouraged to take necessary steps to help patients make decisions. These steps included recognising language and cultural differences, giving right amount of detail, pacing the information, checking for understanding and repeating if necessary and/or involving an advocate. The consultant or care co-ordinator recorded a summary of the capacity assessment in the participant's clinical notes.

3.4. Randomisation

Once written informed consent was obtained, baseline assessments were conducted by assessors blind to allocation followed by randomisation conducted at a remote location by an independent administrator who was contacted by the researcher from the specific site. Block randomisation with randomly permuted block size was used to ensure similar numbers of participants were allocated to each arm of the trial (Altman and Bland, 1999). Randomisation was stratified by centre and ethnic group.

3.5. Intervention and treatment as usual

The intervention group was offered 16 sessions of CaCBTp over a period of 16 to 20 weeks by trained CaCBTp therapists (PP, LW, AS). Each session varied in duration from 40 min to 1.5 h and settings varied based on participant preferences in line with recommendations from the National Institute for Health and Clinical Excellence (NICE, 2009). All therapists were accredited by the British Association for Behavioural and Cognitive Psychotherapies (BABCP). Participants in the TAU group had details of how to access CBT at the end of the trial. Therapy was audiotaped where informed consent was given.

Prior to the trial, therapists attended training led by the authors DK, PP & SR using recommendations from the qualitative study. During the trial, all therapists received ongoing CaCBTp supervision. The principles underlying the cultural adaptation are published (Rathod et al., 2010). Therapists were blind to treatment allocation and only became aware once participants were assigned to them for treatment in the CaCBTp group.

TAU can vary within NHS organisations in the UK. It involves pharmacotherapy, regular monitoring by a multi-disciplinary team including outpatient clinic appointments with a psychiatrist, reviews by a care co-ordinator and input from other professionals like support workers for social activities, occupational therapist or attendance at day centre for meaningful activities (Foster et al., 2010). In some first episode (EIP) services CBT is provided as standard TAU, therefore participants from those teams were excluded from the trial.

3.6. Outcome measures

3.6.1. Primary outcome measure

The primary outcome was symptomatic reduction measured by the Comprehensive Psychopathological Rating Scale (CPRS; Asberg et al., 1978).

Several subscales derived from the CPRS were analysed:

•

The Montgomary–Asberg Depression Rating Scale (MADRAS; Montgomary and Asberg, 1979)

•

The Schizophrenia Change Rating Scale (SCS; Montgomary et al., 1978)

•

Brief Anxiety Rating Scale (Tyrer et al., 1984)

•

Brief Rating Instrument for Assessment of Negative Symptoms Scale (BRAINS; Hansen et al., 2003).

3.6.2. Secondary outcome measures

•

Insight in Psychosis scale (David, 1990)

•

Patient Experience Questionnaire (PEQ; IAPT, 2008).

Fidelity was measured by review of therapy audio tapes.

3.7. Assessments

The assessors were blind to group allocation. All assessors were trained in conducting the CPRS and Insight Scale and reliability established. In CBT for Psychosis, the style of questioning is fundamental to engagement especially when patients experience florid symptoms. Kingdon and Turkington (2005) recommended use of a conversational rather than a staccato questioning style interview and being gentle in keeping a flow of the discussion. This would allow sensitivity and ability to retreat if distress increased. During training, assessors were encouraged to adopt a conversational style in using the interview schedules that was culturally sensitive to concerns expressed by ethnic groups about their involvement in research (Crozier, 2003 and Rathod et al., 2010). Outcome measures were completed at baseline, post-treatment and at six months follow-up.

3.8. Medication

Medication was recorded at baseline for each participant. Any changes during the course of the trial were recorded. Antipsychotic medication was calculated as Chlorpromazine (CPZ) equivalents (mg/day) using the formulae recommended by Atkins et al. (1997).

3.9. Statistical analysis strategy

The aim of the analysis was to estimate the difference between the CaCBTp and TAU groups with regard to each outcome measure at post-treatment and six month follow-up taking into account baseline values (Table 2). Intention to treat (ITT) analysis was performed on the primary outcome measure of overall CPRS scores pre and post-treatment (Altman et al., 2001). Comparison between the two groups used analysis of covariance post-treatment and at six months follow-up. Secondary continuous outcomes were also analysed using analysis of covariance, with the binary outcomes analysed using logistic regression. Statistical models were additionally adjusted for age, gender and medication use, with both unadjusted and adjusted results presented. All continuous outcome variables were checked for normality. If the assumption was not met, data transformations or equivalent non-parametric tests were conducted.

P-values of less than 0.05 were considered statistically significant and estimated treatment differences are presented with 95% confidence intervals. IBM Statistical Package for the Social Sciences version 19 for Windows (SPSS Inc., Chicago; Gray and Kinnear, 2012) was used to analyse the data. The study is reported in accordance with the CONSORT (Consolidated Standards of Reporting Trials; Moher et al., 2001) statement and ICH Guidelines for Good Clinical Practice.

4. Results

4.1. Demographics

Fig. 1 presents the flow of participants in the trial. Two participants randomised to either group were excluded from the study and analysis following randomisation when it was discovered that their ethnicity met the exclusion criteria (both were of Indian Hindu ethnicity) but recorded as Pakistani on clinical records.

Table 1 presents the demographic characteristics of the participants. 35 participants were randomly allocated to either CaCBTp group (n = 16) or TAU group (n = 17). Post-treatment assessments were completed on n = 13 in the CaCBTp group and n = 14 in the TAU group. The mean (SD) age of participants was 33.55 (11.60) years.

Anti-psychotic medication were recorded at baseline and converted to Chropromazine equivalents (CPZ). Mean dosage of CPZ equivalents were: CaCBTp group 407.30 mg (SD = 392.62); TAU group 502.54 mg (SD = 544.59). Medication was higher in the TAU group compared to the CaCBTp group by a mean dosage of 95.24 mg/day (CPZ).

Duration of illness differed by average of 4 years between the treatment groups. TAU had a longer duration than CaCBTp. The analysis was adjusted for medication usage (encompassing differences in medication usage and duration) and standard factors used in CBT studies of age and gender.

4.2. Attrition rates

Attrition rates between baseline and post-treatment were similar in both groups—9% withdrew/missing. Attempts were made to follow-up all randomised participants (White et al., 2011). This led to obtaining end point measures on (n = 2) participants who had discontinued the study but not withdrawn consent to the assessments—the first discontinued therapy after six sessions during the holy month of Ramadan when the participant requested a break, discontinued medication and suffered a relapse of symptoms. The second participant withdrew before commencing therapy in the CaCBTp arm. One participant in the TAU group died prior to follow-up ratings.

4.3. CaCBTp sessions

The mean (SD) number of CaCBTp sessions for the 15 participants in the treatment group was 13.6 (4.9).

4.4. Primary outcome measures

Table 2 presents the results of the outcome measures post-treatment and at 6 months follow-up. The table also presents both change scores (95% CI) on unadjusted and adjusted results. Positive mean differences indicate a positive result for CaCBTp over TAU. Post-treatment, the intervention group showed statistically significant reductions in symptomatology on overall CPRS scores, CaCBTp Mean (SD) = 16.23 (10.77), TAU = 18.60 (14.84); p = 0.047, with a difference in change of 11.31 (95% CI: 0.14 to 22.49). Fig. 2 presents the CPRS mean total scores and subscales of CPRS.

4.5. Secondary outcome measure: Insight Scale

The Insight Scale results are presented in Table 2. Post treatment, mean (SD) total insight scores were: CaCBTp—10.23 (3.49) and TAU—11.07 (3.70). At six months follow-up the mean total insight scores were: CaCBTp Mean (SD) = 10.14 (3.32) and TAU = 11.06 (2.90). Insight domains 2 (acceptance of illness) and 3 (re-labelling of psychotic experiences) when adjusted for baseline scores post-treatment were statistically significant at p < 0.001 and p < 0.001 respectively, a 0.11 unit improvement in CaCBTp when compared to TAU.

4.6. Medication

Analyses of CPZ equivalents at baseline revealed CaCBTp group M (SD) = 407.30 (392.62) and TAU group 502.54 (544.59). Two participants were on both 1st and 2nd generation antipsychotics: CaCBTp (n = 1) and TAU (n = 1). Participants on Clozapine constituted 18% of the sample. At six months follow-up, the mean medication change for the TAU group was M (SD) = 2.29 (1.10) and CaCBTp group 3.13 (0.61) a statistical significant difference between the two groups: β = − 0.83; p < 0.001.

4.7. PEQ

Post-treatment, overall satisfaction with treatment experience was based on PEQ item 10 (PEQ) (n = 14), M (SD) = 4.36 (1.082) (88.2% CI; 3.71 to 4.71). Overall satisfaction was significantly correlated with the number of sessions attended (r = 0.563; p = 0.003). PEQ items 10 (How satisfied are you with the overall experience of using this service: CaCBTp?) and 8 (How involved were you with important decisions about your care/treatment?) were significantly correlated (r = 0.563; p = 0.003). There was a strongly significant correlation between PEQ items 10 and 5 (How satisfied are you with therapist that treated you?) (r = 0.78; p < 0.001). Males had a slightly higher satisfaction than females (M = 4.50 vs. M = 4.17) p = 0.58 although not statistically significant. Overall satisfaction was associated with accessibility, type of therapy, therapist and involvement in decision-making process.

5. Discussion

The main findings of this study were that culturally adapted CBT for psychosis was acceptable and effective. The CaCBTp group engaged well in therapy as indicated by low attrition rates, mean number of sessions attended and the high scores on the PEQ questionnaire. There was a significant reduction of symptomatology on CPRS total and subscale scores post treatment. The gains were not maintained at follow-up except in the MADRAS subscale although a positive trend towards reduction was noted. This may mean that the amount of therapy at 16 sessions was sub-optimal.

Insight score change did not show a statistical significance despite statistically significant changes in symptomatology measured by the CPRS. The insight schedule may be insensitive to such changes: “How do you explain these phenomena [the belief that…hearing that voice/seeing that image, etc.]?” is scored in a way inconsistent with CBT models that stress and vulnerability can cause mental illness including psychosis (Zubin and Spring, 1977 and Rutten et al., 2009). Participants who responded by selecting: ‘part of my illness’ are deemed to have full insight whereas, ‘Reaction to outside event/s’ (e.g. tiredness and stress) is scored as only partial insight. Literature suggests that attribution of symptoms of mental illness to external stressors is a common occurrence amongst individuals from different cultural groups (Al-Krenawi et al., 2001, McCabe and Priebe, 2004 and Cabassa et al., 2008). In future studies, assessors may need to consider rating participants who attribute experiences to reaction to outside events alongside those who attribute them to illness. A sub analysis from the study revealed change in domains two and three (acceptance of illness and re-labelling of psychotic symptoms) as statistically significant when adjusted for baseline scores.

These are encouraging results as this is the first randomised study testing the feasibility of culturally adapted cognitive behaviour therapy for psychosis. The findings are timely given the on-going attempts to reduce the disparities in the provision of psychological therapies across diverse populations. Findings from this trial add to evidence from trials on cultural adaptations in other mental disorders (Rossello and Bernal, 1999, Kohn et al., 2002, Naeem et al., 2010, Grant et al., 2011 and Jimenez et al., 2011).

5.1. Limitations and strengths

This study did not compare standard CBT with CaCBTp. Authors acknowledge that intervention with CBT implies close follow-up and regular appointments than usually expected in TAU.

There was a lack of choice of therapist gender and language matching (Griner and Smith, 2006) but this was not in the study design. The significant mean differences at baseline may have influenced outcome. It was not possible to control this due to the randomisation process. The insight schedule may not have been sensitive to this population group and raises questions about the need for development of culturally sensitive outcome measures as adapting an intervention and then using measures that might have been validated on predominantly western populations may defeat the objective.

5.2. Implications

The findings provide evidence for the use of psychological interventions for minority populations and a benchmark for culturally adapted CBT for psychosis for BME populations. Low attrition rates in the CaCBTp arm and high levels of satisfaction suggest that this intervention was sensitive to the needs of participants from BME groups. The benefits of this project will be realised at different levels:

1.

The current study can inform a fully-powered PHASE III trial.

2.

An effective treatment based on cultural explanations will ensure better quality of care for this group of patients due to better engagement and outcomes. Offering CaCBTp offers choice and an evidence based alternative or adjunct to psychotropic medication (DoH:DRE, 2003; Inside Outside: Sashidharan, 2003).

3.

Implications for policy makers include the training of staff and dissemination of this therapy within resource constraints. Provision of CBT is a cost effective way of managing patients with psychosis in the community as the relapse rate is found to be low (Turkington et al., 2006a and Turkington et al., 2006b).

4.

There are implications for further development of clinical or public health practice due to the role of this treatment in early intervention, relapse prevention and effective community care.

6. Conclusions

The study contributes to the growing evidence for the need for adaptation of evidenced based interventions and efficacy and acceptability of culturally adapted CBT for psychosis. When working with diverse cultures, therapist cognisance and flexibility in the delivery of therapy in the context of culturally derived attitudes, beliefs, values and norms are vital for therapy to be effective and acceptable.

Role of funding source

This trial was part funded by the DRE, Clinical Trailblazers programme and Southern Health NHS Foundation Trust and forms part of author PPs doctoral thesis.

Contributors

Rathod, S., Phiri, P & Kingdon, D. designed the study and wrote the protocol. Rathod, S is the CI and grant holder. Phiri, P managed the literature review and analysis and undertook statistical analysis supervised by Harris, S. Phiri, P., wrote the first draft of the manuscript. All authors (SR, PP, SH, CU, MT, UP & DK) contributed to manuscript revisions and approved the final version.

Conflict of interest

My co-authors and I do not have any interests that might be interpreted as influencing the research and local research ethical standards were followed in the conduct of the trial.

Copyright © 2012 Elsevier B.V. All rights reserved.