Psychological Interventions for Psychosis: A Meta-Analysis of Comparative Outcome Studies
Abstract
Objective
Meta-analyses have demonstrated the efficacy of various interventions for psychosis, and a small number of studies have compared such interventions. The aim of this study was to provide further insight into the relative efficacy of psychological interventions for psychosis.
Method
Forty-eight outcome trials comparing psychological interventions for psychosis were identified. The comparisons included 3,295 participants. Categorization of interventions resulted in six interventions being compared against other interventions pooled. Hedges’ g was calculated for all comparisons. Risk of bias was assessed using four items of the Cochrane risk of bias tool, and sensitivity analyses were conducted. Researcher allegiance was assessed, and sensitivity analyses were conducted for robust significant findings.
Results
Cognitive-behavioral therapy (CBT) was significantly more efficacious than other interventions pooled in reducing positive symptoms (g=0.16). This finding was robust in all sensitivity analyses for risk of bias but lost significance in sensitivity analyses for researcher allegiance, which suffered from low power. Social skills training was significantly more efficacious in reducing negative symptoms (g=0.27). This finding was robust in sensitivity analyses for risk of bias and researcher allegiance. Significant findings for CBT, social skills training, and cognitive remediation for overall symptoms were not robust after sensitivity analyses. CBT was significantly more efficacious when compared directly with befriending for overall symptoms (g=0.42) and supportive counseling for positive symptoms (g=0.23).
Conclusions
There are small but reliable differences in efficacy between psychological interventions for psychosis, and they occur in a pattern consistent with the specific factors of particular interventions.
It has been suggested that all psychotherapies are roughly equivalent in efficacy (1–6), although some meta-analyses have suggested differences in relative efficacy between treatments (7). Previous meta-analyses have demonstrated the absolute efficacy of some psychological interventions for psychosis (8–16), while others have been suggested to be unreliable (17). Comparatively little is understood about the relative efficacy of psychological treatments for psychosis. The most extensive meta-analytic evidence was provided by the U.K. National Institute for Clinical Excellence (NICE) (18). However, risk of bias was not assessed, and many comparisons of psychological interventions against other active treatments were underpowered, including subgroup comparisons for positive and negative symptoms (19).
Other comparative meta-analyses have not consistently demonstrated superiority of the intervention of interest. Jones et al. (20) compared cognitive-behavioral therapy (CBT) against other interventions pooled and concluded that CBT was not reliably more efficacious. A limitation of that analysis was that the CBT group also included compliance studies (21). Lynch et al. (22) compared CBT to active control conditions and found a statistically significant benefit (Hedges’ g=0.2) of CBT compared with active controls pooled for positive symptoms. However, the authors concluded that CBT was no better than nonspecific comparison treatments and that the significant effect size could be explained by lack of blinding. There were some methodological criticisms of the Lynch et al. study (23–25), and there remains controversy over which psychological interventions are most efficacious for psychosis.
No meta-analysis since NICE (18) has compiled all randomized controlled trials in which two psychological interventions for psychosis are compared and pooled these as comparison conditions (2). Given the limitations of the NICE meta-analyses and of many new studies that have been published since, a further comparative meta-analysis is warranted. Whereas previous meta-analyses tended to examine only CBT compared with active treatments, in this analysis we considered all intervention types for which sufficient numbers of studies have been conducted. Our aim is to improve our understanding of which therapy is most efficacious, and for which particular symptoms.
Method
Search Strategy
A systematic literature search conducted in May 2013 identified 5,910 articles for potential inclusion. Four databases were included in the search: PubMed (1,539 abstracts), Embase (1,016 abstracts), PsycInfo (2,128 abstracts), and the Cochrane Central Register of Controlled Trials (1,227 abstracts). Abstracts were identified by entering terms indicative of common psychological interventions for psychosis combined with search terms intended to identify all relevant psychotic disorders. MeSH terms, exploded terms, and text words were employed. Reference lists of published meta-analyses were also examined.
Study Selection
We included randomized trials that included a comparison of at least two psychological interventions intended to be therapeutic and to improve psychiatric symptoms in psychosis; included outcome measures intended to assess psychotic or psychiatric symptoms; and included primarily participants with diagnoses of psychotic disorders. Trials that included patients with mood disorders with psychotic features were included only when such patients were in a minority within the sample.
Trials were excluded if the comparison condition could not be deemed an active psychological intervention (e.g., attention controls, treatment as usual, waiting list); if participants were prodromal or ultra-high risk; or if the interventions were primarily aimed at medication adherence or compliance. Only articles in English or German were considered. Interventions were defined as described in Table 1. Two authors (D.T. and M.v.d.G.) categorized interventions into relevant comparisons, and disagreements were resolved by discussion.
Treatment | Definition | Nst | Np |
---|---|---|---|
Befriending | Participants are assigned social support to match therapy hours provided in other conditions. Typically this consists of friendly discussion or social activities, not directly related to symptoms, with a supportive and empathic individual. Discussion instead focuses primarily on neutral topics, such as current affairs or hobbies, and structured group activities may also be provided. Befriending has been suggested as an efficacious intervention in reducing symptoms of psychosis (26, 27). | 11 | 400 |
Cognitive-behavioral therapy (CBT) | CBT aims to promote awareness of the links between thoughts, behaviors, and feelings to help implement changes in symptoms and functioning. Therapists focus on the modification of dysfunctional thoughts and self-defeating behaviors that perpetuate symptoms or suffering. CBT specifically targeting psychosis has been developed primarily since the 1990s and was originally focused on coping with symptoms (28, 29), whereas more recent approaches have focused on challenging maladaptive cognitions through cognitive restructuring and a formulation-based approach (30–32). We identified these as two main subtypes of CBT for the purposes of this meta-analysis, referring to them as “coping enhancement” and “generic CBT.” | 22 | 706 |
Cognitive remediation | Cognitive deficits have been widely implicated as influential in the development and course of psychosis and have therefore been suggested as worthy treatment targets (33). Cognitive remediation refers to those interventions that target basic cognitive processes, such as working memory, attention, and executive function. This intervention is intended to improve these basic cognitive functions and may also be intended to improve various other aspects of functioning. Computer-based tasks are often the chosen method of implementing cognitive remediation. | 11 | 475 |
Psychoeducation | Provision of relevant information to participants about their diagnosis with the aim of improving their understanding of and coping with their diagnosis. Various psychoeducation methods have been developed for psychosis that go further than provision of basic information and therefore may involve development of coping strategies and role playing. A group format is often utilized, and there is often considerable diversity in what may be labeled “psychoeducation.” This modality is often used as a comparator intervention for more standardized forms of intervention. | 8 | 249 |
Social skills training | Behavioral intervention based on behavioral and social learning traditions in which participants’ social functioning is targeted in order to improve their ability to perform in social situations, manage daily life tasks, and reduce social distress. Importance is typically placed on verbal and nonverbal communication alongside learning appropriate perception and responses to social cues. The intervention may also include training in independent living skills. | 16 | 541 |
Supportive counseling | Nondirective talking therapy that may be based on the work of Carl Rogers (34) or may simply be described in studies as a nondirective intervention in which participants have an open forum to discuss their difficulties, without being actively led or challenged by the therapist. Supportive counseling was therefore defined as an intervention in which the common factors of psychotherapy were present without the specific techniques applied in other, more directive therapies, such as CBT. The opportunity to discuss problems with an empathic therapist in a healing setting may provide relief for the participant without any focus on acquiring new skills or challenging cognitive distortions. Supportive counseling is often used as a means of comparing other interventions against only the common factors of psychological interventions (1). | 17 | 529 |
Quality Assessment
To assess the methodological quality of the studies included, we used the first four criteria of the Cochrane Collaboration risk of bias tool—sequence generation, allocation concealment, blinding of assessors, and incomplete outcome data (there is no clear indication that the fifth [selective outcome reporting] and sixth [other sources of bias] items influence validity [19]). Because it is impossible for these studies to employ a double-blind design, the third item (blinding of assessors) was adapted to include only outcome assessors in masking procedures. Two authors (D.T. and E.K.) assessed the risk of bias, and disagreements were resolved by discussion.
Data Extraction and Selection of Outcome Measures
Data were extracted by one of the authors (D.T.) and checked for consistency by another (E.K.). A spreadsheet piloted in a previous meta-analysis was used for data collection. Attempts were made to contact authors in cases of missing or unusable data, and calculations of missing values were carried out in accordance with the Cochrane Handbook (35).
Table 2 summarizes the study characteristics. Statistical data were extracted for outcome measures relevant to psychotic or psychiatric symptoms. In studies where multiple relevant outcome measures were used, data from all outcome measures were collected and combined as a mean effect size. Dichotomous outcome data were also considered in cases where symptom measures had been converted to dichotomous outcomes, such as clinical exacerbations.
Study Authors (Reference Number) | Sample Characteristics | Relevant Comparisons | Symptom Outcome Measures | Format | Bias Risk (0–4) | Duration (weeks) | Follow-Up | Allegiance |
---|---|---|---|---|---|---|---|---|
Barretto et al. (36) | DSM-IV schizophrenia; 6 months clozapine treatment-resistant; outpatients | CBT (N=12) vs. BF (N=10) | CGI, BPRS, PANSS | Individual | 2 | 21 | 6 months | CBT |
Bechdolf et al. (37, 38) | ICD-10 schizophrenia or related disorder; inpatients | CBT (N=40) vs. PE (N=48) | PANSS | Group | 0 | 8 | 6 and 24 months | None |
Bowie et al. (39) | Schizophrenia or schizoaffective disorder; outpatients | SST (N=38) vs. CR (N=38) | PANSS | Group | 1 | 12 | 24 weeks | None |
Cather et al. (40) | Schizophrenia or schizoaffective disorder; outpatients | CBT (N=15) vs. PE (N=13) | PANSS, PSYRATS | Individual | 1 | 16 | N/A | CBT |
Crawford et al. (41) | Schizophrenia; outpatients | BF (N=140) vs. AT (N=140) | PANSS | Group | 0 | 12 | 24 months | AT |
Dobson et al. (42) | DSM-III schizophrenia; outpatients (severe patients excluded) | SST (N=15) vs. BF (N=13) | PANSS | Group | 3 | 11 | 3 months | None |
Drury et al. (43, 44) | Current functional psychosis, excluding bipolar, hypomania, organic syndrome, confusional states, and drug or alcohol disorders | CBT (N=20) vs. BF (N=20) | PAS | Both | 3 | 12 | 5 years | CBT |
Durham et al. (45) | Schizophrenia, schizoaffective disorder, or delusional disorder with positive symptoms; outpatient and inpatient | CBT (N=22) vs. SC (N=23) | PANSS, PSYRATS, GAS | Individual | 0 | 39 | 3 months | CBT |
Eack et al. (46) | DSM-IV schizophrenia or schizoaffective disorder, in early stages of illness; outpatients | CR (N=31) vs. PE (N=27) | Composite symptoms | Group | 2 | 104 | 12 months | CR |
Falloon et al. (47, 48) | DSM-III schizophrenia from families high in expressed emotion; inpatients | SC (N=18) vs. FI (N=18) | Clinical exacerbation; remission; target symptom ratings | Individual | 3 | 39 | 24 months | FI |
Farreny et al. (49) | DSM-IV-TR schizophrenia or schizoaffective disorder; illness duration >2 years; outpatients | CR (N=34) vs. BF (N=28) | PANSS | Group | 2 | 16 | 40 weeks | CR |
Fries et al. (50) | ICD-10 schizophrenia and schizoaffective disorder; at least two hospitalizations; in at least partial remission at baseline | PE (N=23) vs. SC (N=17) | BPRS, SANS | Group | 4 | 25 | 12 months | None |
Garety et al. (51) | Recently relapsed nonaffective psychosis (ICD-10 F2 or DSM-IV), with positive symptoms; carers included in study | CBT (N=27) vs. FI (N=28) | PANSS, PSYRATS, BDI, BAI | Individual | 0 | 52 | 24 months | None |
Haddock et al. (52) | DSM-IV schizophrenia or schizoaffective disorder (< 5 years); current acute ward admission for positive symptoms | CBT (N=9) vs. SC (N=10) | BPRS | Individual | 1 | 5 | N/A | CBT |
Haddock et al. (53) | DSM-IV schizophrenia or schizoaffective disorder; with history of violence; current antipsychotic medication and positive symptoms | CBT (N=38) vs. BF (N=39) | PANSS, PSYRATS | Individual | 0 | 26 | 12 months | CBT |
Hayes et al. (54) | DSM-III-R schizophrenia; noncurrent positive symptoms; from a range of services | SST (N=23) vs. SC (N=22) | BPRS, SANS | Group | 4 | 18 | 6 months | SST |
Hogarty et al. (55, 56) | RDC schizophrenia or schizoaffective disorder from families high in expressed emotion; inpatients | SST (N=23) vs. FI (N=23) | Symptom relapse | Individual | 4 | 104 | N/A | None |
Hogarty et al. (57, 58) | DSM-III-R or DSM-IV schizophrenia or schizoaffective disorder; outpatients | CR (N=67) vs. PE (N=54) | Composite symptoms | Group | 3 | 52 | 24 months | CR |
Horan et al. (59) | DSM-IV schizophrenia or schizoaffective disorder; clinically stable outpatients | SST (N=17) vs. PE (N=17) | BPRS | Group | 2 | 6 | N/A | SST |
Horan et al. (60) | DSM-IV schizophrenia, schizoaffective disorder, delusional disorder, or psychosis not otherwise specified (not secondary to substance disorder); clinically stable outpatients | SST (N=19) vs. CR (N=24) | BPRS | Group | 2 | 12 | N/A | SST |
Jackson et al. (61) | First-episode psychosis, including schizophrenia, schizophreniform, schizoaffective, bipolar, delusional disorder, and psychosis not otherwise specified; inpatient and outpatient | CBT (N=31) vs. BF (N=31) | BPRS, SANS | Individual | 2 | 12 | 12 months | CBT |
Keefe et al. (62) | Chronic DSM-IV schizophrenia, moderate severity | CR (N=27) vs. BF (N=26) | PANSS | Group | 1 | 12 | N/A | CR |
Klingberg et al. (63, 64) | DSM-IV schizophrenia, with at least one negative symptom; positive symptoms excluded; outpatients | CBT (N=99) vs. CR (N=99) | PANSS, SANS, CDSS, CGI, SCL-90 | Individual | 0 | 52 | N/A | CBT |
Lecomte et al. (65, 66) | Early psychosis (< 2 years), with current psychotic symptoms; stabilized outpatients | CBT (N=48) vs. SST (N=54) | BPRS | Group | 0 | 13 | 6 and 12 months | None |
Lewis et al. (30) | DSM-IV schizophrenia, schizophreniform, schizoaffective, or delusional disorder; first or second admission; inpatients and outpatients | CBT (N=101) vs. SC (N=106) | PANSS, PSYRATS | Individual | 0 | 5 | 18 months | CBT |
Liberman et al. (67) | Persistent and unremitting schizophrenia; outpatients | SST (N=42) vs. OT (N=42) | BSI, GAS, BPRS | Both | 3 | 26 | 24 months | None |
Lukoff et al. (68) | DSM-III schizophrenia; inpatients | SST (N=14) vs. PE (N=14) | PAS | Group | 2 | 10 | N/A | None |
Marder et al. (69) | DSM-III schizophrenia; at least two acute episodes or 2 years of psychotic symptoms; male outpatients | SST (N=13) vs. SC (N=14) | BPRS exacerbations | Group | 3 | 104 | N/A | None |
Moritz et al. (70) | Broad psychotic inpatients who met criteria for schizophreniform disorder | CBT (N=24) vs. CR (N=24) | PANSS, PSYRATS | Both | 0 | 4 | N/A | CBT |
Ng and Cheung (71) | DSM-IV schizophrenia; inpatients | SST (N=18) vs. SC (N=18) | BPRS, SANS | Group | 0 | 8 | 6 months | SST |
O’Connor et al. (72) | DSM-IV delusional disorder; stabilized on medication | CBT (N=12) vs. SC (N=12) | MADS, BAI, BDI | Individual | 3 | 24 | N/A | CBT |
Ojeda et al. (73) | DSM-IV schizophrenia; treatment-resistant; inpatients | CR (N=47) vs. OT (N=46) | PANSS | Individual | 2 | 13 | N/A | CR |
Patterson et al. (74) | DSM-IV schizophrenia or schizophreniform disorder; older chronic Latino inpatients | SST (N=21) vs. SC (N=8) | PANSS | Group | 3 | 26 | 12 months | SST |
Patterson et al. (75) | DSM-IV schizophrenia or schizophreniform disorder; older chronic inpatients | SST (N=124) vs. SC (N=116) | PANSS, HAM-D | Group | 2 | 26 | N/A | SST |
Penadés et al. (76, 77) | DSM-IV schizophrenia, chronic, with a prevalence of negative symptoms and cognitive impairment | CBT (N=20) vs. CR (N=20) | PANSS | Individual | 0 | 17 | 6 months | CR |
Penn et al. (78) | Schizophrenia or schizoaffective disorder and current auditory hallucinations; outpatients | CBT (N=32) vs. SC (N=33) | PANSS, BAVQ, PSYRATS | Group | 0 | 12 | 3 and 12 months | CBT |
Pinto et al. (79) | DSM-IV schizophrenia; treatment-refractory; outpatients | CBT (N=19) vs. SC (N=18) | BPRS, SAPS, SANS | Individual | 3 | 26 | N/A | CBT |
Rodewald et al. (80) | DSM schizophrenia or schizoaffective disorder; inpatients | CR (N=44) vs. PST (N=45) | PANSS | Group | 3 | 3 | N/A | PST |
Röhricht and Priebe (81) | DSM-IV schizophrenia; at least two episodes; outpatients | SC (N=21) vs. BP (N=24) | PANSS | Group | 0 | 10 | 4 months | BP |
Sensky et al. (26) and Turkington et al. (82) | DSM-IV or ICD-10 schizophrenia; treatment-resistant; outpatients | CBT (N=46) vs. BF (N=44) | CPRS, SANS, MADRS | Individual | 0 | 39 | 9 months, 5 years | CBT |
Shawyer et al. (83) | DSM-IV schizophrenia or related condition; with command hallucinations in previous 6 months; outpatients | CBT (N=21) vs. BF (N=22) | PANSS, PSYRATS, CH | Individual | 0 | 15 | 6 months | CBT |
Tarrier et al. (28) | DSM-III-R schizophrenia; treatment resistant | CBT (N=15) vs. PST (N=12) | BPRS, PSE | Individual | 3 | 6 | 6 months | CBT |
Tarrier et al. (29, 84, 85, 86) | Schizophrenia according to PSE; acute-ward inpatients | CBT (N=19) vs. SC (N=19) | BPRS, SANS | Individual | 0 | 13 | 12 months | CBT |
Tas et al. (87) | DSM-IV schizophrenia; clinically stable outpatients | SST (N=22) vs. BF (N=27) | PANSS | Group | 0 | 16 | N/A | SST |
Valmaggia et al. (31) | DSM-IV schizophrenia; with residual delusions or auditory hallucinations; medication resistant | CBT (N=36) vs. SC (N=26) | PANSS, PSYRATS | Individual | 0 | 22 | 6 months | CBT |
Wykes et al. (88, 89) | DSM-IV schizophrenia; with >2 years contact with services; outpatients and inpatients | CR (N=20) vs. OT (N=16) | BPRS | Individual | 0 | 13 | 6 months | CR |
Xiang et al. (90) | DSM-IV schizophrenia; clinically stable outpatients | SST (N=48) vs. SC (N=48) | PANSS | Group | 1 | 9 | 6 months | SST |
Xiang et al. (91) | DSM-IV schizophrenia; clinically stable inpatients and outpatients | SST (N=50) vs. PE (N=53) | PANSS | Group | 2 | 4 | 6 and 12 months | SST |
Meta-Analyses
Psychological interventions for psychosis qualified for inclusion in a separate meta-analysis when there were at least five eligible randomized controlled trials comparing that intervention to another psychological intervention. The comparison group for each separate meta-analysis therefore became the pooled set of comparison interventions from these studies (e.g., CBT compared with other interventions pooled). This resulted in meta-analyses for six intervention types. Separate sub-meta-analyses for positive, negative, or general symptoms were undertaken when there were sufficient studies (≥5) assessing these outcomes.
The Comprehensive Meta-Analysis software package, version 2.2.021, was used for all analyses and calculations. For each individual meta-analysis, aggregated effect sizes indicating the pooled difference between the two groups were calculated at end of treatment using Hedges' g. Hedges' g provides a better effect estimate for small sample sizes than similar measures applied to continuous outcome variables, such as Cohen’s d (92). Alpha was set at 0.05, and 95% confidence intervals were computed.
Heterogeneity
A chi-square test provided a Q statistic to determine the presence of heterogeneity alongside an I2 statistic as a description of the percentage of the variance in each meta-analysis that could be explained by heterogeneity between the studies rather than by chance. A value of 0% indicates no heterogeneity, 25% low heterogeneity, 50% moderate heterogeneity, and 75% high heterogeneity (93).
Additional Analyses
Publication bias was assessed for primary outcomes in each of the six meta-analyses by examining funnel plots produced by the Comprehensive Meta-Analysis program (94) as well as by using the trim-and-fill procedure to estimate the effect size after accounting for publication bias (95). Egger’s test of the intercept was conducted to quantify the bias shown by the funnel plots and to determine whether it was significant.
Direct comparisons were made between psychological interventions when there were at least five studies available comparing two specific treatments. Subgroup analyses were conducted for the intervention with the highest number of eligible studies, namely, CBT. This included splitting CBT into two relevant subtypes to determine whether they had similar efficacy. Differential effects of group or individual format were investigated by entering intervention format as a moderator variable.
Researcher allegiance was examined for all studies, using a tool adapted from a previous meta-analysis (96) (see the data supplement that accompanies the online edition of this article). Two researchers independently rated studies and discussed agreement. Subgroup analyses for researcher allegiance were conducted on robust significant findings that survived the sensitivity analyses for high risk of bias, although this was not possible for all such findings because of the limited number of studies available.
Power Calculation
It was expected that a limited number of studies would be available for certain comparisons. Based on the recommendations of Borenstein (97), power calculations were conducted to determine how many studies were required for sufficient statistical power to identify relevant effects. Previous meta-analyses identified small effect sizes (g=0.1 or g=0.2) in favor of specific interventions. Conservatively assuming a high level of between-study variance (tau-squared), a statistical power of 0.80, and an alpha of 0.05, we estimated that 22 studies with a mean of 30 participants in each intervention arm would be required to detect an effect size of g=0.2. To detect an effect size of g=0.1, we estimated that 88 studies would be required.
Results
After removal of duplicates, 4,768 titles and abstracts were examined, of which 489 articles were retrieved for possible inclusion. Figure 1 describes the selection process. In the 48 included studies, a total of 3,295 participants were included in relevant comparisons of psychological interventions. Six common psychological intervention modalities were identified.
Twenty-four studies used the group format, 21 used the individual format, and three used a combination of individual and group sessions. CBT had the highest proportion of studies using only the individual format (77%), followed by supportive counseling (47%), befriending (45%), cognitive remediation (36%), psychoeducation (12.5%), and social skills training (6%). The time from baseline to posttreatment assessment ranged from 3 weeks to 104 weeks. Risk of bias varied among studies (0–4) and among intervention types. CBT had the highest proportion of studies assessed as having no bias risk (59%), followed by befriending (45.5%), supportive counseling (41%), cognitive remediation (36%), social skills training (12.5%), and psychoeducation (12.5%).
Differences Between Psychological Interventions and Other Interventions Pooled
The results of the six meta-analyses comparing psychological interventions with other interventions pooled are presented in Table 3. Separate meta-analyses were conducted for psychosis symptom groupings. Within each symptom grouping, sensitivity analyses were conducted for varying levels of bias risk. Sensitivity analyses were conducted only when at least four studies were available for that comparison.
Comparison | N | Hedges’ g | 95% CI | Z-Score | Q | I2 (%) |
---|---|---|---|---|---|---|
Befriending vs. all other therapies | ||||||
All symptoms | ||||||
All eligible studies | 11 | –0.37* | –0.60, –0.13 | –3.08 | 21.24* | 52.93 |
Excluding high risk of bias (≥3) | 9 | –0.28* | –0.51, –0.05 | –2.39 | 14.84 | 46.08 |
Excluding low risk of bias (≥2) | 6 | –0.22 | –0.50, 0.06 | –1.56 | 10.78 | 53.61 |
Excluding any risk of bias (≥1) | 5 | –0.20 | –0.52, 0.11 | –1.27 | 10.04* | 60.17 |
Positive symptoms | ||||||
All eligible studies/excluding high risk (≥3) | 6 | –0.14 | –0.41, 0.13 | –0.10 | 8.81 | 43.23 |
Excluding any risk of bias (≥1) | 4 | –0.17 | –0.56, 0.22 | –0.86 | 8.50* | 64.72 |
Negative symptoms | ||||||
All eligible studies | 9 | –0.22 | –0.41, 0.04 | –1.69 | 18.12* | 55.85 |
Excluding high risk of bias (≥3) | 8 | –0.18 | –0.45, 0.80 | –1.37 | 15.93* | 56.67 |
Excluding low (≥2) and any risk of bias (≥1) | 5 | –0.10 | –0.44, 0.24 | –0.56 | 11.94* | 66.49 |
General symptoms (PANSS) | 5 | –0.24 | –0.61, 0.13 | –1.26 | 10.42* | 61.61 |
Cognitive behavioral therapy vs. all other therapies | ||||||
All symptoms | ||||||
All eligible studies | 22 | 0.16* | 0.04, 0.28 | 2.64 | 23.91 | 12.18 |
Excluding high risk of bias (≥3) | 18 | 0.12* | 0.00, 0.23 | 2.01 | 14.98 | 0.00 |
Excluding low risk of bias (≥2) | 15 | 0.10 | –0.03, 0.22 | 1.53 | 11.30 | 0.00 |
Excluding any risk of bias (≥1) | 13 | 0.11 | –0.02, 0.24 | 1.72 | 9.16 | 0.00 |
Positive symptoms | ||||||
All eligible studies | 17 | 0.16* | 0.04, 0.28 | 2.67 | 11.17 | 0.00 |
Excluding high risk of bias (≥3) | 15 | 0.14* | 0.02, 0.27 | 2.32 | 9.42 | 0.00 |
Excluding low risk of bias (≥2) | 12 | 0.15* | 0.02, 0.28 | 2.18 | 9.19 | 0.00 |
Excluding any risk of bias (≥1) | 11 | 0.14* | 0.00, 0.27 | 1.97 | 7.44 | 0.00 |
Negative symptoms | ||||||
All eligible studies | 15 | 0.04 | –0.09, 0.16 | 0.55 | 13.94 | 0.00 |
Excluding high risk of bias (≥3) | 14 | 0.02 | –0.10, 0.15 | 0.36 | 13.04 | 0.34 |
Excluding low risk of bias (≥2) | 11 | –0.00 | –0.15, 0.14 | –0.06 | 8.13 | 0.00 |
Excluding any risk of bias (≥1) | 10 | –0.01 | –0.15, 0.14 | –0.06 | 8.14 | 0.00 |
General symptoms (PANSS) | ||||||
All eligible studies/low risk of bias (≥2) | 8 | 0.10 | –0.13, 0.32 | 0.86 | 12.10 | 42.16 |
Excluding any risk of bias (≥1) | 7 | 0.05 | –0.14, 0.24 | 0.54 | 7.60 | 21.06 |
Cognitive remediation vs. all other therapies | ||||||
All symptoms | ||||||
All eligible studies | 11 | 0.13 | –0.05, 0.31 | 1.46 | 14.63 | 31.69 |
Excluding high risk of bias (≥3) | 10 | 0.20* | 0.01, 0.39 | 2.06 | 11.34 | 20.65 |
Excluding low risk of bias (≥2) | 6 | 0.14 | –0.05, 0.33 | 1.41 | 3.21 | 0.00 |
Excluding any risk of bias (≥1) | 4 | 0.12 | –0.11, 0.34 | 1.02 | 2.49 | 0.00 |
Positive symptoms | ||||||
All eligible studies | 6 | 0.16 | –0.17, 0.49 | 0.97 | 14.11* | 64.56 |
Excluding low risk of bias (≥2) | 4 | 0.29 | –0.06, 0.64 | 1.63 | 6.61 | 54.59 |
Negative symptoms | ||||||
All eligible studies | 6 | –0.14 | –0.39, 0.06 | –1.12 | 8.47 | 40.99 |
Excluding high (≥3) and low (≥2) risk of bias | 4 | –0.08 | –0.38, 0.22 | –0.50 | 5.23 | 42.59 |
Psychoeducation vs. all other therapies | ||||||
All symptoms | ||||||
All eligible studies | 8 | 0.10 | –0.27, 0.11 | –0.80 | 8.02 | 12.66 |
Excluding high risk of bias (≥3) | 6 | –0.13 | –0.41, 0.14 | 0.94 | 7.43 | 32.67 |
Positive symptoms | ||||||
All eligible studies/excluding high risk (≥3) | 4 | 0.19 | –0.06, 0.44 | 1.50 | 1.70 | 0.00 |
Negative symptoms | ||||||
All eligible studies | 5 | 0.02 | –0.22, 0.25 | 0.13 | 3.06 | 0.00 |
Excluding high risk of bias (≥3) | 4 | 0.03 | –0.22, 0.28 | 0.23 | 2.97 | 0.00 |
Social skills training vs. all other therapies | ||||||
All symptoms | ||||||
All eligible studies | 16 | 0.06 | –0.17, 0.28 | 0.49 | 45.33* | 66.91 |
Excluding high risk of bias (≥3) | 10 | 0.19* | 0.02, 0.36 | 2.15 | 8.72 | 0.00 |
Excluding low risk of bias (≥2) | 4 | 0.34 | –0.02, 0.70 | 1.87 | 5.47 | 45.13 |
Positive symptoms | ||||||
All eligible studies | 7 | 0.09 | –0.23, 0.41 | 0.56 | 16.44* | 63.51 |
Excluding high risk of bias (≥3) | 6 | 0.09 | –0.26, 0.45 | 0.50 | 16.41* | 69.53 |
Negative symptoms | ||||||
All eligible studies | 9 | 0.27* | 0.01, 0.53 | 2.01 | 17.33* | 53.83 |
Excluding high risk of bias (≥3) | 7 | 0.32* | 0.07, 0.56 | 2.55 | 10.25 | 41.47 |
Excluding low risk of bias (≥2) | 4 | 0.56* | 0.31, 0.82 | 4.29 | 1.99 | 0.00 |
Supportive counseling vs. all other therapies | ||||||
All symptoms | ||||||
All eligible studies | 17 | 0.00 | –0.21, 0.22 | 0.04 | 40.31* | 60.31 |
Excluding high risk of bias (≥3) | 10 | 0.01 | –0.30, 0.32 | 0.06 | 32.97 | 72.70 |
Excluding low risk of bias (≥2) | 9 | –0.12 | –0.30, 0.05 | –1.37 | 6.18 | 0.00 |
Excluding any risk of bias (≥1) | 7 | –0.08 | –0.28, 0.11 | –0.83 | 1.74 | 0.00 |
Positive symptoms | ||||||
All eligible studies | 8 | –0.14 | –0.36, 0.09 | –1.12 | 10.28 | 31.90 |
Excluding high (≥3) and low (≥2) risk of bias | 6 | –0.05 | –0.25, 0.15 | –0.51 | 5.33 | 6.27 |
Excluding any risk of bias (≥1) | 5 | –0.02 | –0.27, 0.23 | –0.17 | 5.00 | 19.98 |
Negative symptoms | ||||||
All eligible studies | 9 | –0.12 | –0.41, 0.17 | –0.83 | 18.55* | 56.87 |
Excluding high (≥3) and low (≥2) risk of bias | 6 | –0.21 | –0.57, 0.15 | –1.13 | 13.34* | 62.52 |
Excluding any risk of bias (≥1) | 5 | –0.09 | –0.45, 0.27 | –0.50 | 7.74 | 48.30 |
Befriending was less efficacious for all symptom outcome measures pooled compared with other therapies pooled (g=−0.366, p<0.05). This effect was robust when studies with a high risk of bias were excluded (g=−0.279, p<0.05) but lost significance when studies with a low risk and no risk of bias were excluded. Removing the studies with a low risk and no risk of bias also limited the power of this comparison. Seven comparisons of befriending and other interventions pooled showed moderate heterogeneity, and two comparisons showed low heterogeneity.
CBT was more efficacious compared with other interventions pooled for all symptom outcome measures pooled (g=0.161, p<0.05). This effect was robust when studies with a high risk of bias were excluded (g=0.118, p<0.05) but lost significance when studies with a low risk and no risk of bias were excluded. For positive symptom outcome measures, CBT was more efficacious (g=0.162, p<0.05). This effect was robust in all three sensitivity analyses when we sequentially removed studies with a high risk (g=0.144, p<0.05), a low risk (g=0.149, p<0.05), and no risk of bias (g=0.137, p<0.05). All comparisons of CBT with other interventions pooled showed no heterogeneity or low heterogeneity.
Cognitive remediation was more efficacious than other interventions pooled for all symptoms in the sensitivity analysis excluding high risk of bias (g=0.202, p<0.05) but was not shown as significantly more efficacious in any other comparisons. Heterogeneity varied among comparisons for cognitive remediation, with two comparisons showing moderate heterogeneity.
Direct Comparisons of Psychological Interventions
The results of direct comparisons between interventions are presented in Table 4. Limited comparisons were possible since few studies were available. CBT was more efficacious than befriending for all symptom measures pooled (g=0.419, p<0.05). CBT was also more efficacious than supportive counseling for positive symptoms (g=0.226, p<0.05).
Comparison | N | Hedges’ g | 95% CI | Z-Score | Q | I2 (%) | p |
---|---|---|---|---|---|---|---|
Direct comparisons of two interventions | |||||||
CBT vs. befriending | |||||||
All symptoms (R) | 6 | 0.42* | 0.15, 0.69 | 3.02 | 7.00 | 28.61 | |
CBT vs. supportive counseling | |||||||
All symptoms (F/R) | 8 | 0.10 | –0.10, 0.30 | 0.99 | 6.09 | 0.00 | |
Positive symptoms (F/R) | 6 | 0.23* | 0.01, 0.44 | 2.07 | 5.00 | 0.04 | |
Social skills training vs. supportive counseling | |||||||
All symptoms (R) | 6 | –0.07 | –0.54, 0.40 | –0.29 | 26.27 | 80.96 | |
CBT subtypes vs. other interventions pooled | |||||||
Coping enhancement subtype | |||||||
All symptoms (F/R) | 6 | –0.01 | –0.19, 0.18 | –0.08 | 1.83 | 0.00 | |
Negative symptoms (F/R) | 5 | –0.04 | –0.23, 0.15 | –0.41 | 2.45 | 0.00 | |
Generic subtype | |||||||
All symptoms (R) | 16 | 0.22* | 0.08, 0.37 | 2.97 | 16.96 | 11.58 | |
Positive symptoms (F/R) | 13 | 0.17* | 0.03, 0.32 | 2.28 | 10.66 | 0.00 | |
Negative symptoms (R) | 10 | 0.01 | –0.08, 0.28 | 1.07 | 10.44 | 13.76 | |
Subgroup analyses of CBT: group vs. individual formatb | |||||||
All symptoms | |||||||
Individual (R) | 18 | 0.18* | 0.05, 0.32 | 2.66 | 19.93 | 14.9 | |
Group (R) | 3 | 0.00 | –0.26, 0.27 | 0.03 | 1.08 | 0.00 | |
Overall (R) | 21 | 0.13 | –0.02, 0.29 | 1.64 | 22.45 | 10.93 | 0.24 |
Positive symptoms | |||||||
Individual (F/R) | 13 | 0.16* | 0.02, 0.30 | 2.17 | 9.04 | 0.00 | |
Group (F/R) | 3 | 0.12 | –0.13, 0.36 | 0.93 | 0.39 | 0.00 | |
Overall (F/R) | 16 | 0.15* | 0.01, 0.34 | 2.04 | 9.50 | 0.00 | 0.80 |
Negative symptoms | |||||||
Individual (F/R) | 12 | 0.09 | –0.06, 0.23 | 1.15 | 12.05 | 8.68 | |
Group (F/R) | 3 | –0.11 | –0.35, 0.14 | –0.85 | 0.16 | 0.00 | |
Overall (F/R) | 15 | 0.02 | –0.17, 0.20 | 0.17 | 13.94 | 0.00 | 0.19 |
Subgroup analyses of researcher allegiance for comparisons with robust significant effects | |||||||
CBT vs. all other therapiesc | |||||||
Positive symptoms (F/M) | |||||||
Excluding high risk of bias (≥3) | |||||||
No allegiance | 3 | 0.10 | –0.15, 0.35 | 0.80 | 0.24 | 0.00 | 0.42 |
Allegiance for CBT | 11 | 0.17 | 0.01, 0.32 | 2.40 | 5.35 | 0.00 | |
Excluding low risk of bias (≥2) | |||||||
No allegiance | 2 | 0.08 | –0.25, 0.40 | 0.50 | 0.21 | 0.00 | 0.60 |
Allegiance for CBT | 9 | 0.18 | 0.03, 0.33 | 2.33 | 5.07 | 0.00 | |
Excluding any risk of bias (≥1) | |||||||
No allegiance | 2 | 0.08 | –0.25, 0.40 | 0.50 | 0.21 | 0.00 | 0.58 |
Allegiance for CBT | 8 | 0.19 | 0.03, 0.34 | 2.36 | 4.96 | 0.00 | |
Social skills training vs. all other therapies | |||||||
Negative symptoms (M) | |||||||
All eligible studies | |||||||
No allegiance | 3 | 0.37 | 0.04, 0.7 | 2.20 | 2.30 | 13.27 | 0.55 |
Allegiance for social skills training | 6 | 0.21 | –0.21, 0.62 | 0.98 | 15.50 | 67.7 | |
Excluding high risk of bias (≥3) | |||||||
No allegiance | 2 | 0.30 | –0.10, 0.71 | 1.48 | 1.51 | 33.82 | 0.83 |
Allegiance for social skills training | 6 | 0.36 | 0.04, 0.69 | 2.19 | 8.97 | 44.25 |
Meta-Analyses for CBT Subtypes
To examine whether there were differences between CBT subtypes (coping enhancement and generic CBT), subgroup analyses were conducted. Results included in Table 4 suggest that generic CBT was more efficacious for all symptom measures pooled and for positive symptoms. The between-group comparisons for group versus individual format were not significant, but this comparison was hampered by low power. No subgroup analyses showed significant heterogeneity.
Researcher allegiance.
Sensitivity analyses for researcher allegiance were conducted for the robust effects of CBT on positive symptoms and social skills training on negative symptoms. The effect of CBT on positive symptoms became nonsignificant in both sensitivity analyses, although only three studies could be included in the no-allegiance group, resulting in low power. The effect of social skills training on negative symptoms remained significant in the sensitivity analyses, although comparison was not possible for the stricter risk of bias categories because of the limited number of studies available.
Publication bias.
Funnel plots and the trim-and-fill procedure suggested the presence of publication bias in some comparisons of the cognitive remediation and social skills training meta-analyses. The funnel plot for all symptoms pooled in the cognitive remediation meta-analysis suggested that three studies with negative findings remained unpublished. Using the trim-and-fill procedure to investigate the significant effect shown for overall symptoms without studies with a high risk of bias (g=0.20), two studies were trimmed, meaning the effect size was reduced to g=0.10 (95% CI=−0.12, 0.32). For the social skills training overall symptoms meta-analyses, the funnel plot suggested that seven studies had not been published when all studies were included. However, when the funnel plot and trim-and-fill procedure were examined for the only significant finding within this meta-analysis, there was no suggestion of publication bias. Similarly, there was no suggestion of publication bias for the significant effects of social skills training found for negative symptoms.
Discussion
This series of meta-analyses comparing psychological interventions for psychosis found significant differences in their relative efficacy for the reduction of psychotic symptoms, as summarized in Figure 2. While some of these differences lost significance when sensitivity analyses were conducted for risk of bias, others were more robust.
CBT also showed superiority when compared directly to befriending for all symptoms and when compared with supportive counseling for positive symptoms. The generic CBT subtype appeared more efficacious in reducing overall symptoms and positive symptoms.
With respect to the much discussed thesis that all psychotherapies produce similar outcomes (1), our results provide evidence that could both support and contradict this proposition.
We are aware that CBT is uniform in its assumption that negative emotions and behavioral problems are the result of the appraisal and interpretation of antecedent events. By changing appraisal and interpretation of events and stimuli, the emotions and the behavior will change. However, there are variants within CBT that differ in degree of emphasis on cognitions or on behavioral experiments. We have the impression that variations in CBT are not typically reflected in results, although the meta-analysis by Wykes et al. (11) found a trend for larger effect sizes in more behavioral CBT. Comparisons could be made to antipsychotic medication, where almost all agents target dopamine D2 receptors. Although the compounds are slightly different from each other, they have about the same efficacy (100). A recent development is that CBT using the same general technique is becoming more focused. For example, CBT protocols are being developed to address command hallucinations (101) and negative symptoms (102, 103). Preliminary results show larger effect sizes for more focused applications compared with generic CBT for psychosis.
Our study had several limitations that affect the extent to which robust conclusions can be drawn from the results. The majority of comparisons had low statistical power (<0.80). Without satisfactory power, there is a high risk of type II errors. A limitation of any meta-analysis categorizing randomized controlled trials into groups by intervention type is that such decisions involve a degree of subjectivity. We attempted to address this issue by having two researchers agree on categorization. There was some controversy regarding study selection following the Lynch et al. CBT meta-analysis (22–24, 99). The risk of bias procedure applied in our meta-analysis addresses the issues raised about inclusion since all but one of the studies excluded in the Lynch et al. meta-analysis were excluded in the most stringent sensitivity analysis. For the aims of this meta-analysis, there did not appear to be any reason to exclude this study (75).
Another limitation concerns our focus on positive, negative, and general symptoms. While CBT, supportive counseling, and befriending target symptom reduction, psychoeducation, social skills training, and cognitive remediation only indirectly target symptoms. Psychoeducation is often intended to improve medication adherence, with secondary symptom improvement, and although the effects on symptoms were not significantly different from all other interventions, this does not mean that psychoeducation was not able to improve adherence. Similarly, cognitive remediation targets the improvement of cognitive functioning, and the absence of an effect on symptoms does not mean that there was no improvement in cognitive functioning. Those effects are beyond the scope of the meta-analyses we present and are not reported. It was also beyond the scope of this study to consider the possibility of patients with better prognosis being channeled into a particular treatment, interaction with pharmacotherapies, and diagnostic heterogeneity among samples, since information on these domains was not reliably available across studies.
There was considerable variety in the quality of studies as assessed by the risk of bias procedure, and there were marked differences in quality between specific intervention types. CBT had the highest proportion of studies assessed as having no risk of bias, and social skills training had the lowest. It is important that future studies on the relative efficacy of social skills training address these issues. Research should continue to compare psychological interventions for psychosis in order to improve statistical power in meta-analyses. It is essential, too, that comparative randomized controlled trials minimize bias risk and that the issue of researcher allegiance be addressed. Meta-analytic studies must also answer related questions about psychosis interventions, such as predictors of treatment outcome and dropout. This includes individual participant data meta-analyses, in which the authors of this study are currently involved. Future research may also focus on dismantling studies, which provide insight into the influence of common and specific factors. Future development of treatment plans may take into account the effects of specific factors on the specific symptom areas and integrate these to optimize both positive and negative symptom reduction.
In conclusion, although the differences observed between interventions for psychosis were small in this meta-analysis, the relatively robust nature of the differences and the pattern by which differences occur have implications for the continued clinical implementation, design, and improvement of psychosocial therapies for psychosis.
1 : The Great Psychotherapy Debate: Models, Methods, and Findings. Mahwah, NJ, Routledge, 2001Google Scholar
2 : Psychotherapy for depression in adults: a meta-analysis of comparative outcome studies. J Consult Clin Psychol 2008; 76:909–922Crossref, Medline, Google Scholar
3 : Comparative efficacy of seven psychotherapeutic interventions for patients with depression: a network meta-analysis. PLoS Med 2013; 10:e1001454Crossref, Medline, Google Scholar
4 : Cognitive-behavioral therapy versus other therapies: redux. Clin Psychol Rev 2013; 33:395–405Crossref, Medline, Google Scholar
5 : Direct comparisons of treatment modalities for youth disorders: a meta-analysis. Psychother Res 2008; 18:5–14Crossref, Medline, Google Scholar
6 : The relative efficacy of bona fide psychotherapies for treating post-traumatic stress disorder: a meta-analysis of direct comparisons. Clin Psychol Rev 2008; 28:746–758Crossref, Medline, Google Scholar
7 : Specificity of treatment effects: cognitive therapy and relaxation for generalized anxiety and panic disorders. J Consult Clin Psychol 2007; 75:513–522Crossref, Medline, Google Scholar
8 : Psychological treatments in schizophrenia, I: meta-analysis of family intervention and cognitive behaviour therapy. Psychol Med 2002; 32:763–782Crossref, Medline, Google Scholar
9 : Efficacy of psychological therapy in schizophrenia: conclusions from meta-analyses. Schizophr Bull 2006; 32(suppl 1):S64–S80Crossref, Medline, Google Scholar
10 : The effect of cognitive behavioral treatment on the positive symptoms of schizophrenia spectrum disorders: a meta-analysis. Schizophr Res 2005; 77:1–9Crossref, Medline, Google Scholar
11 : Cognitive behavior therapy for schizophrenia: effect sizes, clinical models, and methodological rigor. Schizophr Bull 2008; 34:523–537Crossref, Medline, Google Scholar
12 : A meta-analysis of controlled research on social skills training for schizophrenia. J Consult Clin Psychol 2008; 76:491–504Crossref, Medline, Google Scholar
13 : A meta-analysis of cognitive remediation for schizophrenia: methodology and effect sizes. Am J Psychiatry 2011; 168:472–485Link, Google Scholar
14 : Family intervention for schizophrenia. Cochrane Database Syst Rev 2010; 12:CD000088Medline, Google Scholar
15 : The effect of family interventions on relapse and rehospitalization in schizophrenia: a meta-analysis. Schizophr Bull 2001; 27:73–92Crossref, Medline, Google Scholar
16 : Psychoeducation for schizophrenia. Cochrane Database Syst Rev 2011; 6:CD002831Medline, Google Scholar
17 : Psychological treatments in schizophrenia, II: meta-analyses of randomized controlled trials of social skills training and cognitive remediation. Psychol Med 2002; 32:783–791Crossref, Medline, Google Scholar
18
19 (eds): Assessing risk of bias in included studies, in Cochrane Handbook for Systematic Reviews of Interventions. Edited by Higgins JPTGreen S. Chichester, UK, Wiley-Blackwell, 2008, chap 8Crossref, Google Scholar
20 : Cognitive behaviour therapy versus other psychosocial treatments for schizophrenia. Cochrane Database Syst Rev 2012; 4:CD008712Medline, Google Scholar
21 : Compliance therapy in psychotic patients: randomised controlled trial. BMJ 1996; 312:345–349Crossref, Medline, Google Scholar
22 : Cognitive behavioural therapy for major psychiatric disorder: does it really work? A meta-analytical review of well-controlled trials. Psychol Med 2010; 40:9–24Crossref, Medline, Google Scholar
23 : Over-simplification and exclusion of non-conforming studies can demonstrate absence of effect: a lynching party? Psychol Med 2010; 40:25–27Crossref, Medline, Google Scholar
24 : Letter to the editor: a comment on Lynch et al (2009). Psychol Med 2010; 40:877–880Crossref, Google Scholar
25 ;
26 : A randomized controlled trial of cognitive-behavioral therapy for persistent symptoms in schizophrenia resistant to medication. Arch Gen Psychiatry 2000; 57:165–172Crossref, Medline, Google Scholar
27 : Befriending versus CBT for schizophrenia: a convergent and divergent fidelity check. Behav Cogn Psychother 2006; 34:25–30Crossref, Google Scholar
28 : A trial of two cognitive-behavioural methods of treating drug-resistant residual psychotic symptoms in schizophrenic patients, I: outcome. Br J Psychiatry 1993; 162:524–532Crossref, Medline, Google Scholar
29 : Randomised controlled trial of intensive cognitive behaviour therapy for patients with chronic schizophrenia. BMJ 1998; 317:303–307Crossref, Medline, Google Scholar
30 : Randomised controlled trial of cognitive-behavioural therapy in early schizophrenia: acute-phase outcomes. Br J Psychiatry Suppl 2002; 43:s91–s97Crossref, Medline, Google Scholar
31 : Cognitive-behavioural therapy for refractory psychotic symptoms of schizophrenia resistant to atypical antipsychotic medication: randomised controlled trial. Br J Psychiatry 2005; 186:324–330Crossref, Medline, Google Scholar
32 : Cognitive Therapy of Schizophrenia (Guides to Individualized Evidence-Based Treatment). New York, Guilford Press, 2008Google Scholar
33 : Cognitive deficits as treatment targets in schizophrenia. Schizophr Res 2004; 72:21–28Crossref, Medline, Google Scholar
34 : Client-Centered Therapy: Its Current Practice, Implications, and Theory. Boston, Houghton Mifflin, 1951Google Scholar
35 (eds): Selecting studies and collecting data, in Cochrane Handbook for Systematic Reviews of Interventions. Edited by Higgins JPTGreen S. Chichester, UK, Wiley-Blackwell, 2008, chap 7Crossref, Google Scholar
36 : A preliminary controlled trial of cognitive behavioral therapy in clozapine-resistant schizophrenia. J Nerv Ment Dis 2009; 197:865–868Crossref, Medline, Google Scholar
37 : A randomized comparison of group cognitive-behavioural therapy and group psychoeducation in patients with schizophrenia. Acta Psychiatr Scand 2004; 110:21–28Crossref, Medline, Google Scholar
38 : A randomized comparison of group cognitive-behavioural therapy and group psychoeducation in acute patients with schizophrenia: outcome at 24 months. Acta Psychiatr Scand 2005; 112:173–179Crossref, Medline, Google Scholar
39 : Combined cognitive remediation and functional skills training for schizophrenia: effects on cognition, functional competence, and real-world behavior. Am J Psychiatry 2012; 169:710–718Link, Google Scholar
40 : A pilot study of functional cognitive behavioral therapy (fCBT) for schizophrenia. Schizophr Res 2005; 74:201–209Crossref, Medline, Google Scholar
41 ;
42 : Effects of social skills training and social milieu treatment on symptoms of schizophrenia. Psychiatr Serv 1995; 46:376–380Link, Google Scholar
43 : Cognitive therapy and recovery from acute psychosis: a controlled trial, I: impact on psychotic symptoms. Br J Psychiatry 1996; 169:593–601Crossref, Medline, Google Scholar
44 : Cognitive therapy and recovery from acute psychosis: a controlled trial, 3: five-year follow-up. Br J Psychiatry 2000; 177:8–14Crossref, Medline, Google Scholar
45 : Tayside-Fife clinical trial of cognitive-behavioural therapy for medication-resistant psychotic symptoms: results to 3-month follow-up. Br J Psychiatry 2003; 182:303–311Crossref, Medline, Google Scholar
46 : Cognitive enhancement therapy for early-course schizophrenia: effects of a two-year randomized controlled trial. Psychiatr Serv 2009; 60:1468–1476Link, Google Scholar
47 : Family management in the prevention of exacerbations of schizophrenia: a controlled study. N Engl J Med 1982; 306:1437–1440Crossref, Medline, Google Scholar
48 : Family management in the prevention of morbidity of schizophrenia: clinical outcome of a two-year longitudinal study. Arch Gen Psychiatry 1985; 42:887–896Crossref, Medline, Google Scholar
49 : REPYFLEC cognitive remediation group training in schizophrenia: looking for an integrative approach. Schizophr Res 2012; 142:137–144Crossref, Medline, Google Scholar
50 : Wirksamkeit und Prozessmerkmale einer psychoedukativen und bewältigungsorientierten Gruppentherapie für schizophren und schizoaffektiv Erkrankte. Verhaltenstherapie 2004; 13:237–243Crossref, Google Scholar
51 : Cognitive-behavioural therapy and family intervention for relapse prevention and symptom reduction in psychosis: randomised controlled trial. Br J Psychiatry 2008; 192:412–423Crossref, Medline, Google Scholar
52 : A pilot study evaluating the effectiveness of individual inpatient cognitive-behavioural therapy in early psychosis. Soc Psychiatry Psychiatr Epidemiol 1999; 34:254–258Crossref, Medline, Google Scholar
53 : Cognitive-behavioural therapy v social activity therapy for people with psychosis and a history of violence: randomised controlled trial. Br J Psychiatry 2009; 194:152–157Crossref, Medline, Google Scholar
54 : Social skills training with chronic schizophrenic patients: effects on negative symptoms and community functioning. Behav Ther 1995; 26:433–449Crossref, Google Scholar
55 : Family psychoeducation, social skills training, and maintenance chemotherapy in the aftercare treatment of schizophrenia, I: one-year effects of a controlled study on relapse and expressed emotion. Arch Gen Psychiatry 1986; 43:633–642Crossref, Medline, Google Scholar
56 ;
57 : Cognitive enhancement therapy for schizophrenia: effects of a 2-year randomized trial on cognition and behavior. Arch Gen Psychiatry 2004; 61:866–876Crossref, Medline, Google Scholar
58 : Durability and mechanism of effects of cognitive enhancement therapy. Psychiatr Serv 2006; 57:1751–1757Link, Google Scholar
59 : Social cognitive skills training in schizophrenia: an initial efficacy study of stabilized outpatients. Schizophr Res 2009; 107:47–54Crossref, Medline, Google Scholar
60 : Efficacy and specificity of social cognitive skills training for outpatients with psychotic disorders. J Psychiatr Res 2011; 45:1113–1122Crossref, Medline, Google Scholar
61 : Acute-phase and 1-year follow-up results of a randomized controlled trial of CBT versus befriending for first-episode psychosis: the ACE project. Psychol Med 2008; 38:725–735Crossref, Medline, Google Scholar
62 : Feasibility and pilot efficacy results from the multisite Cognitive Remediation in the Schizophrenia Trials Network (CRSTN) randomized controlled trial. J Clin Psychiatry 2012; 73:1016–1022Crossref, Medline, Google Scholar
63 : Negative symptoms of schizophrenia as primary target of cognitive behavioral therapy: results of the randomized clinical TONES study. Schizophr Bull 2011; 37(suppl 2):S98–S110Crossref, Medline, Google Scholar
64 : Adverse effects of cognitive behavioral therapy and cognitive remediation in schizophrenia: results of the Treatment of Negative Symptoms Study. J Nerv Ment Dis 2012; 200:569–576Crossref, Medline, Google Scholar
65 : Group cognitive behavior therapy or social skills training for individuals with a recent onset of psychosis? Results of a randomized controlled trial. J Nerv Ment Dis 2008; 196:866–875Crossref, Medline, Google Scholar
66 : Group CBT for early psychosis: are there still benefits one year later? Int J Group Psychother 2012; 62:309–321Crossref, Medline, Google Scholar
67 : Skills training versus psychosocial occupational therapy for persons with persistent schizophrenia. Am J Psychiatry 1998; 155:1087–1091Link, Google Scholar
68 : A holistic program for chronic schizophrenic patients. Schizophr Bull 1986; 12:274–282Crossref, Medline, Google Scholar
69 : Two-year outcome of social skills training and group psychotherapy for outpatients with schizophrenia. Am J Psychiatry 1996; 153:1585–1592Link, Google Scholar
70 : Antipsychotic treatment beyond antipsychotics: metacognitive intervention for schizophrenia patients improves delusional symptoms. Psychol Med 2011; 41:1823–1832Crossref, Medline, Google Scholar
71 : Social skills training in Hong Kong Chinese patients with chronic schizophrenia. Hong Kong J Psychiatry 2006; 16:14–20Google Scholar
72 : Treating delusional disorder: a comparison of cognitive-behavioural therapy and attention placebo control. Can J Psychiatry 2007; 52:182–190Crossref, Medline, Google Scholar
73 : Efficiency of cognitive rehabilitation with REHACOP in chronic treatment resistant Hispanic patients. NeuroRehabilitation 2012; 30:65–74Medline, Google Scholar
74 : Development and pilot testing of a new psychosocial intervention for older Latinos with chronic psychosis. Schizophr Bull 2005; 31:922–930Crossref, Medline, Google Scholar
75 : Functional adaptation skills training (FAST): a randomized trial of a psychosocial intervention for middle-aged and older patients with chronic psychotic disorders. Schizophr Res 2006; 86:291–299Crossref, Medline, Google Scholar
76 : Cognitive remediation therapy for outpatients with chronic schizophrenia: a controlled and randomized study. Schizophr Res 2006; 87:323–331Crossref, Medline, Google Scholar
77 : Executive function needs to be targeted to improve social functioning with cognitive remediation therapy (CRT) in schizophrenia. Psychiatry Res 2010; 177:41–45Crossref, Medline, Google Scholar
78 : A randomized controlled trial of group cognitive-behavioral therapy vs enhanced supportive therapy for auditory hallucinations. Schizophr Res 2009; 109:52–59Crossref, Medline, Google Scholar
79 : Cognitive-behavioral therapy and clozapine for clients with treatment-refractory schizophrenia. Psychiatr Serv 1999; 50:901–904Link, Google Scholar
80 : Planning and problem-solving training for patients with schizophrenia: a randomized controlled trial. BMC Psychiatry 2011; 11:73Crossref, Medline, Google Scholar
81 : Effect of body-oriented psychological therapy on negative symptoms in schizophrenia: a randomized controlled trial. Psychol Med 2006; 36:669–678Crossref, Medline, Google Scholar
82 : A randomized controlled trial of cognitive-behavior therapy for persistent symptoms in schizophrenia: a five-year follow-up. Schizophr Res 2008; 98:1–7Crossref, Medline, Google Scholar
83 : A randomised controlled trial of acceptance-based cognitive behavioural therapy for command hallucinations in psychotic disorders. Behav Res Ther 2012; 50:110–121Crossref, Medline, Google Scholar
84 : Durability of the effects of cognitive-behavioural therapy in the treatment of chronic schizophrenia: 12-month follow-up. Br J Psychiatry 1999; 174:500–504Crossref, Medline, Google Scholar
85 : Two-year follow-up of cognitive-behavioral therapy and supportive counseling in the treatment of persistent symptoms in chronic schizophrenia. J Consult Clin Psychol 2000; 68:917–922Crossref, Medline, Google Scholar
86 : Are some types of psychotic symptoms more responsive to cognitive-behaviour therapy? Behav Cogn Psychother 2001; 29:45–55Crossref, Google Scholar
87 : Impact of family involvement on social cognition training in clinically stable outpatients with schizophrenia: a randomized pilot study. Psychiatry Res 2012; 195:32–38Crossref, Medline, Google Scholar
88 : The effects of neurocognitive remediation on executive processing in patients with schizophrenia. Schizophr Bull 1999; 25:291–307Crossref, Medline, Google Scholar
89 : Are the effects of cognitive remediation therapy (CRT) durable? Results from an exploratory trial in schizophrenia. Schizophr Res 2003; 61:163–174Crossref, Medline, Google Scholar
90 : Training patients with schizophrenia with the community re-entry module: a controlled study. Soc Psychiatry Psychiatr Epidemiol 2006; 41:464–469Crossref, Medline, Google Scholar
91 : Efficacy of the community re-entry module for patients with schizophrenia in Beijing, China: outcome at 2-year follow-up. Br J Psychiatry 2007; 190:49–56Crossref, Medline, Google Scholar
92 (eds): Analysing data and undertaking meta-analyses, in Cochrane Handbook for Systematic Reviews of Interventions. Edited by Higgins JPTGreen S. Chichester, UK, Wiley-Blackwell, 2008, chap 9Crossref, Google Scholar
93 : Measuring inconsistency in meta-analyses. BMJ 2003; 327:557–560Crossref, Medline, Google Scholar
94 (eds): Addressing reporting biases, in Cochrane Handbook for Systematic Reviews of Interventions. Edited by Higgins JPTGreen S. Chichester, UK, Wiley-Blackwell, 2008, chap 10Crossref, Google Scholar
95 : Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics 2000; 56:455–463Crossref, Medline, Google Scholar
96 : The efficacy of non-directive supportive therapy for adult depression: a meta-analysis. Clin Psychol Rev 2012; 32:280–291Crossref, Medline, Google Scholar
97 : Introduction to Meta-Analysis. Chichester, UK, Wiley, 2009Crossref, Google Scholar
98 : The omnipotence of voices: a cognitive approach to auditory hallucinations. Br J Psychiatry 1994; 164:190–201Crossref, Medline, Google Scholar
99 : Social Skills Training for Schizophrenia: A Step-by-Step Guide, 2nd revised ed. New York, Guilford, 2004Google Scholar
100 : A meta-analysis of head-to-head comparisons of second-generation antipsychotics in the treatment of schizophrenia. Am J Psychiatry 2009; 166:152–163Link, Google Scholar
101 : Cognitive therapy for command hallucinations: randomised controlled trial. Br J Psychiatry 2004; 184:312–320Crossref, Medline, Google Scholar
102 : Cognitive behavioral therapy for negative symptoms (CBT-n) in psychotic disorders: a pilot study. J Behav Ther Exp Psychiatry 2013; 44:300–306Crossref, Medline, Google Scholar
103 : Randomized trial to evaluate the efficacy of cognitive therapy for low-functioning patients with schizophrenia. Arch Gen Psychiatry 2012; 69:121–127Crossref, Medline, Google Scholar