Most meta-analytic studies of cognitive–behavioural and otherpsychotherapies for adult depression have found that thesetherapies have moderate to large effects on depression comparedwith control conditions,1–5 effects that are comparable to thoseof antidepressive medication.6 It may be possible, however, thatthese effects are overestimated because of publication bias – thetendency for increased publication rates among studies that showa statistically significant effect of treatment.7,8 Publication bias canbe considered as one of the major drawbacks of meta-analyticstudies and a threat to their validity. Evidence of such bias hasbeen found in many intervention fields, including that ofdepression treatment. In two recent meta-analytic reviews ofstudies examining the effects of antidepressive medication,9,10 itwas found that considerably more positive than negative trialswere reported, which resulted in a considerable overestimationof the effect sizes of drugs. Although many meta-analytic studieshave examined the effects of cognitive–behavioural and otherpsychotherapies for depression,11 publication bias has not beenexamined thoroughly in these studies. Most comprehensivemeta-analyses examining the effects of psychotherapies have notexamined publication bias at all.12,13 One meta-analysis didexamine publication bias,2 but did not test whether the effectsof published and unpublished studies differed significantly fromeach other, nor did it use advanced techniques to formally testwhether significant publication bias was present. That study alsodid not calculate effect sizes that were adjusted for publication biasby imputing missing studies, and assessed the mean effect sizesof psychotherapies after adjustment for publication bias;furthermore, the meta-analysis included only a limited numberof the published studies. However, the study found someindications that publication bias might be present in research oncognitive–behavioural and other psychotherapies for adultdepression. We decided, therefore, to conduct a new meta-analysisof studies examining the effects of psychotherapies for adultdepression and focus on the question whether publication biasmay have resulted in an overestimation of the mean effect size.

Method

Identification and selection of studiesWe used a database of 1036 papers on the psychological treatmentof depression, which includes studies on combined treatments andcomparisons with pharmacotherapies. This database has beendescribed in detail elsewhere,14 and has been used in a series ofearlier meta-analyses (www.evidencebasedpsychotherapies.org).It was developed through a comprehensive literature search (from1966 to January 2009) in which we examined 9011 abstracts inPubMed (1629 abstracts), PsycINFO (n= 2439), Embase (n= 2606)and the Cochrane Central Register of Controlled Trials (n= 2337).These abstracts were identified by combining terms indicative ofpsychological treatment and depression (both MeSH terms andtext words). For this database we also collected the primary studiesfrom 42 meta-analyses of psychological treatment for depression(www.evidencebasedpsychotherapies.org). For the study reportedhere we examined the full texts of these 1036 papers. The abstractsfrom the electronic bibliographic databases were screened by thefirst author (P.C.) and the 1036 retrieved reports were examinedindependently by two reviewers for possible inclusion. When thetwo reviewers disagreed, they discussed the differences with thethird reviewer until agreement was reached.

We included published studies in which the effects of apsychological treatment on adults with a diagnosed depressivedisorder or scoring above a cut-off point on a depression measure-ment instrument were compared with a control condition in arandomised controlled trial. Depressive disorders could be diag-nosed with one of the versions of the DSM, the ICD, ResearchDiagnostic Criteria or another diagnostic system. ‘Psychologicaltreatments’ were defined as interventions in which verbalcommunication between a therapist and a client was the coreelement, or in which a psychological treatment was written downin book format (bibliotherapy) which the client worked throughmore or less independently, but with some kind of personalsupport from a therapist (by telephone, email or otherwise).

173

Efficacy of cognitive–behavioural therapyand other psychological treatments for adultdepression: meta-analytic study of publication biasPim Cuijpers, Filip Smit, Ernst Bohlmeijer, Steven D. Hollon and Gerhard Andersson

BackgroundIt is not clear whether the effects of cognitive–behaviouraltherapy and other psychotherapies have been overestimatedbecause of publication bias.

AimsTo examine indicators of publication bias in randomisedcontrolled trials of psychotherapy for adult depression.

MethodWe examined effect sizes of 117 trials with 175 comparisonsbetween psychotherapy and control conditions. As indicatorsof publication bias we examined funnel plots, calculatedadjusted effect sizes after publication had been taken intoaccount using Duval & Tweedie’s procedure, and tested the

symmetry of the funnel plots using the Begg & Mazumdarrank correlation test and Egger’s test.

ResultsThe mean effect size was 0.67, which was reduced afteradjustment for publication bias to 0.42 (51 imputed studies).Both Begg & Mazumbar’s test and Egger’s test were highlysignificant (P<0.001).

ConclusionsThe effects of psychotherapy for adult depression seem tobe overestimated considerably because of publication bias.

Declaration of interestNone.

The British Journal of Psychiatry (2010)196, 173–178. doi: 10.1192/bjp.bp.109.066001

Review article

In earlier meta-analyses of psychotherapies for adultdepression, few indications were found that different types ofpsychotherapy have differential effects on depression. Fewindications for significant differences are found in multivariatemeta-regression analyses,1 as well as in meta-analytic research inwhich direct comparisons of different types of psychotherapyare examined.15,16 However, we decided to examine the effectsof the full sample of psychotherapies, as well as the sample ofstudies examining cognitive–behavioural therapies. We definedcognitive–behavioural therapy as a psychological treatment inwhich the therapist focuses on the impact a patient’s presentdysfunctional thoughts have on current behaviour and futurefunctioning.16 The therapy is aimed at evaluating, challengingand modifying the patient’s dysfunctional beliefs (cognitiverestructuring). The definitions of other psychotherapies wedistinguished have been given elsewhere.16

We excluded studies on children and adolescents (below 18years of age). Studies in which the psychological interventioncould not be distinguished from other elements of the inter-vention were also excluded (managed care interventions anddisease management programmes), as were studies in which astandardised effect size could not be calculated (mostly because notest was performed in which the difference between experimentaland control groups was examined), studies in which patients werenot randomly assigned to conditions, and studies of in-patients.We also excluded unpublished dissertations, studies aimed atmaintenance treatments and relapse prevention, and studies thatalso included participants with anxiety disorders. Comorbidgeneral medical or psychiatric disorder was not used as anexclusion criterion. No language restriction was applied.

Data extraction

Studies were coded according to patient characteristics, inter-vention characteristics and general characteristics of the study.We coded the following characteristics: type of recruitment(through the community, from clinical samples, other recruitmentmethod); diagnosis of depression (diagnosed mood disorder,other definition of depression – usually a high score on aself-rating instrument); target group (adults in general, olderadults, student populations, women with postpartum depression,general medical patients with depression, other target groups);type of psychotherapy (cognitive–behavioural therapy accordingto the manual by Beck et al,17 other cognitive–behavioural therapyin which cognitive restructuring is the core element, interpersonalpsychotherapy, problem-solving therapy, non-directive supportivetherapy, behavioural activation treatment, other psychotherapy;full definitions of these treatments are given elsewhere);16 treat-ment format (individual, group, guided self-help); type of controlgroup (waiting list, care as usual, pill placebo, other controlgroup); and data analyses (intention to treat, completers only).

All studies were coded by two independent assessors. Whenthey disagreed, they discussed the differences with a third revieweruntil agreement was reached. A study of the quality of theincluded studies has been reported elsewhere.18

Statistical analysis

We first calculated effect sizes (d) for each study by subtracting (atpost-test) the average score of the control group from the averagescore of the experimental group and dividing the result by thepooled standard deviations of the experimental and controlgroup. An effect size of 0.5 thus indicates that the mean of theexperimental group is half a standard deviation larger than themean of the control group. Effect sizes of 0.80 can be assumed

to be large, effect sizes of 0.50 are moderate and effect sizes of0.20 are small.19 In the calculations of effect sizes, onlystandardised instruments were used that explicitly measureddepression (online Table DS1). If more than one depressionmeasure was used, the mean of the effect sizes was calculated, suchthat each study (or contrast group) contributed only one effectsize to the meta-analysis. When means and standard deviationswere not reported we used other statistics (t, P) to calculate effectsizes.

To calculate the pooled mean effect size we used the computerprogram Comprehensive Meta-analysis version 2.2.021 forWindows, developed for support in meta-analysis (www.meta-analysis.com). We conducted all analyses using the random effectsmodel.20 Subgroup analyses were conducted according to theComprehensive Meta-analysis procedures. In the subgroupanalyses we used mixed effects analyses that pooled studies withinsubgroups with the random effects model but tested for significantdifferences between subgroups with the fixed effects model.

In order to assess the heterogeneity of effect sizes we calculatedthe I 2 statistic, which is an indicator of heterogeneity inpercentages. A value of 0% indicates no observed heterogeneity;larger values show increasing heterogeneity, with 25% as low,50% as moderate and 75% as high heterogeneity.21 We alsocalculated the Q statistic, but only report whether this wassignificant or not.

Sensitivity analyses

In our analyses we included studies in which two or morepsychological treatments were compared with a control group.This means that multiple comparisons from these studies wereincluded in the same analysis. These multiple comparisons,however, are not independent of each other, which may haveresulted in an artificial reduction of heterogeneity. This mayinfluence the overall effect size and our estimates of publicationbias. Therefore, we conducted additional meta-analyses in whichwe included only one comparison per study. First, we includedonly the comparison with the largest effect size (i.e. the largestdifference between the psychotherapy and control group),followed by another analysis in which only the comparison withthe smallest effect size was included.

Because many different effect measures were used in thestudies, we also conducted separate analyses in which the effectsizes were based on the two measures that were used in moststudies: the Beck Depression Inventory (BDI) and the HamiltonRating Scale for Depression (HRSD). Furthermore, we conductedanalyses in which possible outliers (defined as studies with aneffect size of 1.5 or larger) were removed, because these mightdistort the overall results.

Indications of publication bias

In order to examine the possibility of publication bias, weconducted several tests.

Funnel plot. Perhaps the most common method that hasbeen proposed to detect the existence of publication bias in ameta-analysis is the funnel plot.22 This plots a measure of studysize (the standard error) on the vertical axis as a function of effectsize on the horizontal axis. Large studies appear at the top of thegraph and tend to cluster near the mean effect size. Smaller studiesappear towards the bottom of the graph. As there is moresampling variation in effect size estimates in the smaller studies,they will be dispersed across a range of values.23 Visual inspectionof a funnel plot can give an indication of publication bias. The

174

Cuijpers et al

Publication bias in psychotherapy research

studies can be expected to be distributed symmetrically about thepooled effect size when publication bias is absent. In the presenceof bias, it can be expected that the lower part of the plot will showa higher concentration of studies on one side of the mean than onthe other. This is caused by the fact that smaller studies (appearingtowards the bottom of the funnel plot) are more likely to be pub-lished if they have larger than average effects, which makes themmore likely to meet the criterion for statistical significance.23

Although funnel plots can be constructed on the basis of thestandard error or on trial size,24 we decided to use the standarderror on the vertical axis because this is the most widely usedmethod, and because it is readily available in the ComprehensiveMeta-analysis software.

Duval & Tweedie’s trim and fill procedure. If a meta-analysishas included all relevant studies, the funnel plot can be expectedto be symmetric and dispersed equally on either side of the meaneffect.23 If, on the other hand, the funnel plot is asymmetric (withmore studies on the right of the mean effect size than on the left)this could indicate publication bias. Duval & Tweedie developed amethod of imputing missing studies, based on the assumptionthat the studies should be equally distributed on both sides ofthe mean effect size.25 This procedure, usually called the Duval& Tweedie trim and fill procedure, yields an estimate of the effectsize after the publication bias has been taken into account(adjusted effect size) and also indicates how many studies wereimputed to correct for publication bias.

Begg & Mazumdar rank correlation test. The Begg & Mazumdarrank correlation test is based on the assumption that studies withlarger sample sizes are published more often and that studies withan equal sample size are published less often when the effect size issmaller.26 Therefore, it can be expected that in the case ofpublication bias there is a negative correlation between thestandardised effect size and the standard errors of these effects.This correlation is tested with Kendall’s tau: a significant valueindicates possible publication bias. Because publication bias isexpected to reduce the mean effect size, the significance test isone-tailed.

Egger’s test of the intercept. Egger’s linear regression method,27

like the Begg & Mazumdar rank correlation test, is intended toquantify the bias captured by the funnel plot. In the Egger testthe standard normal deviation is regressed on precision, definedas the inverse of the standard error.23 The intercept in thisregression corresponds to the slope in a weighted regression ofthe effect size on the standard error. In our results we report theintercept, the 95% confidence interval and the significance (one-tailed). Power for this test is generally higher than power for therank correlation method, but is still low unless there is severe biasor a substantial number of studies.28

Results

The selection and inclusion of studies are summarised in Fig. 1.29

All inclusion criteria were met by 117 studies, in which 175psychological treatment conditions were compared with a controlgroup. These studies included a total of 9537 participants (5481 inthe psychotherapy conditions and 4056 in the control conditions).An overview of selected characteristics of the included studies ispresented in online Table DS1 together with a full list ofreferences.

Indicators of publication bias in the full sample

The overall effect size of the 175 comparisons between psycho-therapy and a control condition was 0.67 (95% CI 0.60–0.75),with high heterogeneity (I 2 = 70.27). Adjustment for publicationbias according to Duval & Tweedie’s trim and fill procedureresulted in a mean effect size of 0.42 (95% CI 0.33–0.51) with51 studies missing. Both Begg & Mazumbar’s test and Egger’s testresulted in highly significant indicators of publication bias(P50.001). The results of these analyses are summarised inTable 1. The funnel plot of the effect sizes of the studies (withthe standard error on the vertical axis and the effect size at thehorizontal axis), first without and then with the imputed studies,clearly shows that smaller studies with lower effect sizes aremissing (Fig. 2) and this is confirmed after the imputation ofmissing studies.

Because the overall results of our analyses may have beeninfluenced by possible outliers, we conducted a new meta-analysisin which all effect sizes from 1.5 or larger were removed. As can beseen in Table 1, the mean effect size dropped somewhat (d= 0.51),but all indicators of publication bias remained highly significant(P50.001), and Duval & Tweedie’s trim and fill procedureestimated the adjusted effect size to be considerably smaller(d= 0.39), with a total of 38 missed studies.

In order to examine the influence of multiple comparisonsfrom one study, we conducted another meta-analysis in whichwe included only one comparison per study (online Table DS2).From the 46 studies with multiple comparisons we included onlythe comparison with the largest effect size (i.e. the largestdifference between the psychotherapy and control groups). Wethen conducted another analysis in which only the comparisonwith the smallest effect size was included. As can be seen in Table1, the resulting mean effect sizes differed somewhat from theoverall analyses, and heterogeneity dropped a little, but allindicators of publication bias remained highly significant(P50.001).

When we limited the analyses to the effect sizes found for theBDI, somewhat larger effect sizes were found, but again allindicators for publication bias remained highly significant(P50.001). The same was true when we limited the analyses tothe effect sizes found for the HRSD.

Publication bias in subgroups of studies

It is possible that the publication bias we found differs forsubtypes of studies. We therefore conducted a series of analysesin which we first selected a subgroup of studies based on a specificcharacteristic, and then examined the indicators of publicationbias within this subgroup (online Table DS2). Significantindicators of publication bias were found for most subgroups ofstudies. No indication of publication bias was found for studiesexamining interpersonal psychotherapy, studies examiningpsychotherapy for women with postpartum depression, andstudies in which patients were not recruited through thecommunity or from clinical samples (usually from generalmedical patient groups or systematic screening).

Publication bias in studies of cognitive–behaviouraltherapy

Because more than half of the comparisons examined cognitive–behavioural therapy, we decided to investigate publication biasin these studies in more detail. The results of these analyses arepresented in Fig. 2 and in online Table DS3. The main analysespointed at a considerable and significant risk of publication bias

175

Cuijpers et al

among studies examining cognitive–behavioural therapy, and thisremained high in all sensitivity analyses. The overall effect size ofthe 89 comparisons between cognitive–behavioural therapy and acontrol condition was 0.69, and after adjustment for publicationbias this was reduced to 0.49 with 26 studies missing. Both Begg& Mazumbar’s test and Egger’s test were highly significant(P50.001).

In the subgroup analyses we merged several subgroups becauseof the small number of studies per group (the target group hadtwo categories: adults and ‘more specific target group’; andplacebo and other control groups were merged into ‘other controlgroup’). We also conducted a subgroup analysis in which weexamined cognitive–behavioural therapy according to the manualby Beck et al,17 and other types of cognitive–behavioural therapy:in the majority of subgroups, all indicators pointed at significantpublication bias (online Table DS3).

Discussion

We used a large sample of controlled studies of psychotherapy foradult depression to examine the possibility of publication bias. Alltests for publication bias gave strong and highly significantindications of publication bias. The overall mean effect size ofpsychotherapy was 0.67, which corresponds with a numberneeded to treat (NNT) of 2.75.30 After adjustment for publicationbias the effect size was reduced to 0.42, which corresponds to anNNT of 4.27. When we examined the subsample of studiesexamining cognitive–behavioural therapy, the results werecomparable. The overall effect size of cognitive–behaviouraltherapy was 0.69, and after adjustment for publication bias thiswas reduced to 0.49, with 26 studies missing. There were somesubgroups of studies for which we found no indication ofsignificant publication bias, including research on interpersonalpsychotherapy and research on psychotherapy for women withpostpartum depression. These results have to be considered withcaution, however, because these subsamples were relatively smalland may have been the result of random error.

Limitations

This meta-analytic review has several limitations. The mostimportant is that our tests for publication bias do not providedirect evidence of such bias. These procedures only test whetherthe funnel plot is symmetrical and whether small negative studiesare missing. This cannot be considered as direct evidence, andthere may in principle be other reasons why these studies aremissing. It may be possible, for example, that early pilot studiesof new treatments result in large effect sizes because the developerof such a new treatment is the best expert, realising larger effectsizes than later studies by other groups who test the effects of thistreatment in routine care. Alternatively, it may be possible thatsuch early pilot projects attract a specific type of patient, who iswilling to undergo a new treatment. If these patients were morereceptive to change or to treatment in general, that would resultin larger effect sizes of smaller pilot studies. However, funnel plotasymmetry cannot be considered to be proof of bias in a meta-analysis,31 and no statistical imputation method can recover thehidden and missing truth. With the highly significant indicatorsof funnel plot asymmetry, it is very likely that meta-analyses ofthese studies overestimate the true effect size of psychotherapyfor adult depression.

We also used a rather narrow definition of publication bias inthis study. A broader definition of publication bias might indicate

176

Literature search9011 references identified:

PubMed 1629PsychINFO 2439

Embase 2606Cochrane Central Registerof Controlled Trials 2337

Retrieved:1036 publications

Analysed:117 controlled studies

of psychotherapy

Excluded: 919Studies with adolescents 67No random assignment 53Duplicate publication 238Not only depression 120No psychotherapy 133

No control condition 188Maintenance trial 35

Other reasons 85

Fig. 1 Selection of studies.

Table 1 Mean effect sizes (d) for psychological treatments of adult depression from published studies: tests for publication bias

in the overall sample

Unadjusted values Adjusted valuesa Egger’s test

Nc d (95% CI) I 2 d (95% CI) Nimpa Taub I c (95% CI)

All comparisons 175 0.67 (0.60–0.75) 70.27*** 0.42 (0.33–0.51) 51 0.37*** 1.58 (1.06–2.11) ***

Outliers excludedd 153 0.51 (0.45–0.57) 43.98*** 0.39 (0.32–0.45) 38 0.26*** 1.01 (0.58–1.44) ***

Multiple comparisons excluded (lowest retained) 117 0.57 (0.49–0.66) 67.98*** 0.39 (0.29–0.49) 28

Multiple comparisons excluded (highest retained) 117 0.70 (0.60–0.80) 74.87*** 0.45 (0.33–0.56) 33 0.37*** 1.63 (0.92–2.34) ***

BDI only 113 0.76 (0.65–0.87) 66.81*** 0.45 (0.33–0.56) 38 0.42*** 2.39 (1.78–2.99) ***

HRSD only 67 0.83 (0.70–0.96) 65.61*** 0.54 (0.40–0.69) 21 0.47*** 1.68 (0.99–2.36) ***

BDI, Beck Depression Inventory; HRSD, Hamilton Rating Scale for Depression; Nc, number of comparisons; Nimp, number of imputed studies.a. Based on Duval & Tweedie’s trim & fill procedure.b. Kendall’s tau based on Begg & Mazumbar rank correlation.c. Intercept based on Egger’s test.d. In these analyses, mean effect sizes of 1.5 and higher were excluded.*P<0.05, **P<0.01, ***P<0.001.

Publication bias in psychotherapy research

not only selective publication of studies but also selectivereporting of outcomes.32,33 Such selective reporting of outcomesmight also cause asymmetry of the funnel plot. A furtherlimitation is that many of the included studies were small andalthough they were all randomised controlled trials they maynot meet standard quality criteria for clinical trials. On the otherhand the number of included studies was high, enabling us tocontrol for several basic characteristics of the populations,interventions and study designs.

The statistical tests we used to assess the asymmetry of thefunnel plot have several weaknesses.20,34 A problem with Duval& Tweedie’s trim and fill procedure is that it depends stronglyon the assumptions of the model for why studies are missing,and the algorithm for detecting asymmetry can be influenced byone or two aberrant studies.34 Begg & Mazumdar’s test and Egger’stest may also yield a different picture depending on the index usedin the analyses,24 and they tend to have low power. Furthermore,they only make sense if there is a reasonable amount of dispersionin the sample sizes and a reasonable number of studies. In ourstudy, however, we included a large number of studies and wefound strong indications of publication bias. In this situation,these weaknesses do not suggest that the main conclusion – thatthere is considerable risk of publication bias – should be doubted.

Implications

Research on psychotherapy for adult depression does not seem tobe any freer from publication bias than research on medicationtreatment, although it is likely that the sources of this bias differbetween the two types of studies. There have long been concernsthat the large pharmaceutical companies that fund much of the

drug research have an economic incentive to make the mostpositive case possible for the medications that they sell and thatthis incentive may influence the investigators whose research theysupport. Although psychological treatments are not supported bylarge pharmaceutical companies with strong economic interests inlarger effect sizes, there appears to be a considerable risk ofpublication bias in studies of psychotherapy as well. Researchersof psychological treatments do have personal interests inpublication of (larger) effects, as these are more likely to lead totenure and lucrative workshop fees. Pharmaceutical companieshave clear financial reasons to inflate research findings, andpsychological investigators have both personal and professionalreasons for doing the same. Results of a psychotherapy studymight not be published because of authors’ failure to submit theresults of negative studies, journal editors preferring largesignificant outcomes over small non-significant effects, andnegative reports by reviewers.9 A recent study examiningpsychotherapy for depression among children and adolescents alsofound strong indications of publication bias, suggesting that thisproblem is not restricted to psychotherapy for adult depressionalone.35 Furthermore, there is a growing concern not only thatresearch findings in the field of depression treatment are biased,but that most current published research findings are false andthat claimed research findings may be simply accurate measuresof prevailing bias.36

We strongly encourage psychotherapy researchers to registertheir studies in trial registries, as this will facilitate laterinvestigations into publication bias. We also believe that theinvestigation of publication bias is a concern not only forpsychotherapy research but for psychology in general. The ‘filedrawer’ effect is probably present in all areas of psychological

177

0.0 –

0.2 –

0.4 –

0.6 –

0.8 –

Stan

dar

de

rro

r

74 73 72 71 0 1 2 3 4 74 73 72 71 0 1 2 3 4

74 73 72 71 0 1 2 3 4 74 73 72 71 0 1 2 3 4

0.0 –

0.2 –

0.4 –

0.6 –

0.8 –

0.0 –

0.2 –

0.4 –

0.6 –

0.8 –

0.0 –

0.2 –

0.4 –

0.6 –

0.8 –

Stan

dar

de

rro

r

Stan

dar

de

rro

r

Effect size

Effect size

Effect size

Effect size

Stan

dar

de

rro

r

(a) (b)

(c)

(d)

Fig. 2 Funnel plots. (a) All psychotherapy studies, without imputed studies; (b) studies of cognitive–behavioural therapy (CBT) only,without imputed studies; (c) all psychotherapy studies, with imputed studies (black circles); (d) CBT studies only, with imputed studies.Imputation according to Duval & Tweedie trim and fill procedure.

Cuijpers et al

research,37 and may have implications for how research findingsare interpreted. Meta-analysts are therefore encouraged to quanti-fy systematically the number of studies likely to have been missedin their review, and to conduct sensitivity analyses as exemplifiedin this study to obtain an impression of the impact of publicationbias on the overall estimate.

Not only are the effects of antidepressant medication overesti-mated because of publication bias, but the same seems to be truefor psychotherapy for adult depression. The two most importantfirst-line treatments of adult depression appear not to be as effec-tive as is often assumed and it may be that the presumably activeingredients of treatment account for a smaller proportion of theoutcomes observed than is widely believed. It may be time tosearch for new and more effective treatments of depression orto isolate the active ingredients of those that already exist, andto examine in more depth which treatments are most effectivefor which patients.

Pim Cuijpers, PhD, Department of Clinical Psychology, Vrije Universiteit (VU), andEMGO Institute, VU University Medical Centre, Amsterdam; Filip Smit, PhD,Department of Clinical Psychology, VU University, EMGO Institute, VU UniversityMedical Center, Amsterdam, and Trimbos Institute, Utrecht; Ernst Bohlmeijer, PhD,Technical University Twente, Deventer, The Netherlands; Steven D. Hollon, PhD,Vanderbilt University, Nashville, Tennessee, USA; Gerhard Andersson, PhD,Department of Behavioural Sciences and Learning, Swedish Institute for DisabilityResearch, Linkoping University, and Department of Clinical Neuroscience, PsychiatrySection, Karolinska Institute, Stockholm, Sweden

Correspondence: Professor Pim Cuijpers, Department of Clinical Psychology,VU University Amsterdam, Van der Boechorststraat 1, 1081 BT Amsterdam,The Netherlands. Email: p.cuijpers@psy.vu.nl

First received 10 Mar 2009, final revision 20 Sep 2009, accepted 28 Oct 2009

References

1 Cuijpers P, van Straten A, Warmerdam L, Smits N. Characteristics of effectivepsychological treatments of depression: a meta-regression analysis.Psychother Res 2000; 18: 225–36.

2 Churchill R, Hunot V, Corney R, Knapp M, McGuire H, Tylee A, et al.A systematic review of controlled trials of the effectiveness and cost-effectiveness of brief psychological treatments for depression. HealthTechnol Assess 2001; 5: 35.

3 Gaffan EA, Tsaousis I, Kemp-Wheeler SM. Researcher allegiance and meta-analysis: the case of cognitive therapy for depression. J Consult Clin Psychol1995; 63: 966–80.

4 Gloaguen V, Cottrauxa J, Cucherata M, Blackburn IM. A meta-analysis of theeffects of cognitive therapy in depressed patients. J Affect Disord 1998; 49:59–72.

5 McDermut W, Miller IW, Brown RA. The efficacy of group psychotherapyfor depression: a meta-analysis and review of the empirical research.Clin Psychol Sci Pract 2001; 8: 98–116.

6 Cuijpers P, van Straten A, van Oppen P, Andersson G. Are psychologicaland pharmacological interventions equally effective in the treatment ofadult depressive disorders? A meta-analysis of comparative studies.J Clin Psychiatry 2008; 69: 1675–85.

7 Rothstein HR, Sutton AJ, Borenstein M, eds. Publication bias. In PublicationBias in Meta-analysis: Prevention, Assessment and Adjustment: 277–302.Wiley & Sons, 2005.

8 Sterne JAC, Egger M. Regression methods to detect publication and otherbias in meta-analysis. In Publication Bias in Meta-analysis: Prevention,Assessment and Adjustment (eds HR Rothstein, AJ Sutton, M Borenstein):88–111. Wiley, 2005.

9 Turner EH, Matthews AA, Linardatos E, Tell RA, Rosenthal R. Selectivepublication of antidepressant trials and its influence on apparent efficacy.N Engl J Med 2008; 358: 252–60.

10 Kirsch I, Deacon BJ, Huedo-Medina TB, Scoboria A, Moore TJ, Johnson BT.Initial severity and antidepressant benefits: a meta-analysis of datasubmitted to the food and drug administration. PLoS Med 2008; 5: e45.

11 Cuijpers P, Dekker J. Psychologische behandeling van depressie: eensystematisch overzicht van meta-analyses [Psychological treatment of

depression: a systematic overview of meta-analyses]. Ned Tijdschr Geneesk2005: 149: 1892–7.

12 Westen D, Morrison K. A multidimensional meta-analysis of treatments fordepression, panic, and generalized anxiety disorder: an empiricalexamination of the status of empirically supported therapies. J ConsultClin Psychol 2001; 69: 875–99.

13 Robinson LA, Berman JS, Neimeyer RA. Psychotherapy for the treatmentof depression: a comprehensive review of controlled outcome research.Psychol Bull 1990; 108: 30–49.

14 Cuijpers P, van Straten A, Warmerdam L, Andersson G. Psychologicaltreatment of depression: a meta-analytic database of randomized studies.BMC Psychiatry 2008; 8: 36.

15 Wampold BE, Mondin GW, Moody M, Stich F, Benson K, Hyun-nie A.A meta-analysis of outcome studies comparing bona fide psychotherapies:empirically, ‘All must have prizes’. Psychol Bull 1997; 122: 203–15.

16 Cuijpers P, van Straten A, Andersson G, van Oppen P. Psychotherapy fordepression in adults: a meta-analysis of comparative outcome studies.J Consult Clin Psychol 2008; 76: 909–22.

17 Beck AT, Rush AJ, Shaw BF, Emery G. Cognitive Therapy of Depression.Guilford, 1979.

18 Cuijpers P, van Straten A, Bohlmeijer E, Hollon SD, Andersson G. The effectsof psychotherapy for adult depression are overestimated: a meta-analysis ofstudy quality and effect size. Psychol Med 2009; epub ahead of print: doi:10.1017/S0033291709006114.

19 Cohen J. Statistical Power Analysis for the Behavioral Sciences. AcademicPress, 1977.

20 Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews ofInterventions 4.2.6. Wiley, 2006.

21 Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistencyin meta-analyses. BMJ 2003; 327: 557–60.

22 Light R, Pillemer D. Summing Up: The Science of Reviewing Research.Harvard University Press, 1984.

23 Borenstein M. Software for publication bias. In Publication Bias in Meta-analysis: Prevention, Assessment and Adjustments (eds H Rothstein,AJ Sutton, M Borenstein): 391–403. Wiley, 2005.

24 Tang JL, Liu JL. Misleading funnel plot for detection of bias in meta-analysis.J Clin Epidemiol 2000; 53: 477–84.

25 Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based method oftesting and adjusting for publication bias in meta-analysis. Biometrics 2000;56: 455–63.

26 Begg CB, Mazumdar M. Operating characteristics of a rank correlation testfor publication bias. Biometrics 1994; 50: 1088–101.

27 Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysisdetected by a simple, graphical test. BMJ 1997; 315: 629–34.

28 Sterne JA, Gavaghan D, Egger M. Publication and related bias in meta-analysis: power of statistical tests and prevalence in the literature. J ClinEpidemiol 2000; 53: 1119–29.

29 Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improvingthe quality of reports of meta-analyses of randomised controlled trials: theQUOROM statement. Lancet 1999; 354: 1896–900.

30 Kraemer HC, Kupfer DJ. Size of treatment effects and their importance toclinical research and practice. Biol Psychiatry 2006; 59: 990–6.

31 Sterne JA, Gavaghan D, Egger M. Publication and related bias in meta-analysis: power of statistical tests and prevalence in the literature.J Clin Epidemiol 2000; 53: 1119–29.

32 Chan AW, Hrobjartsson A, Haahr MT, Gotzsche PC, Altman DG. Empiricalevidence for selective reporting of outcomes in randomized trials:comparison of protocols to published articles. JAMA 2004; 291: 2457–65.

33 Furukawa TA, Watanabe N, Omori IM, Montori VM, Guyatt GH. Associationbetween unreported outcomes and effect size estimates in Cochrane meta-analyses. JAMA 2007; 297: 468–70.

34 Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. Introduction to Meta-Analysis. Wiley, 2009.

35 Watanabe N, Hunot V, Omori IM, Churchill R, Furukawa TA. Psychotherapyfor depression among children and adolescents: a systematic review. ActaPsychiatr Scand 2007; 116: 84–95.

36 Ioannidis JP. Why most published research findings are false. PLoS Med 2005;2: e124.

37 Rosenthal R. The ‘file drawer problem’ and tolerance for null results. PsycholBull 1979; 86: 638–41.

178

1

The British Journal of Psychiatry (2010)196, 173–178. doi: 10.1192/bjp.bp.109.066001

Data supplement

Table DS1 Selected characteristics of trials examining psychotherapy for adult depression

Participants, n

Study

(first author, date) Recruitment Diagnosis Target group Type of therapy Format

No. of

sessions

Control

group EXP CTR ITT Measures

Allart, 200738 Comm Other Adults CBT (other) Grp 12 CAU 61 41 + BDI

Andersson, 200539 Comm Other Adults CBT (other) GSH 5 Other 36 49 - BDI; MADRS–9

Arean, 199340 Comm MDD Older PST Grp 12 WL 28 10 - HRSD; BDI; GDS

Other Grp 12 WL 27 10 -

Ayen, 200441 Comm Mood Other CBT (other) Grp 12 WL 11 10 - BDI; IDS

SUP Grp 12 WL 20 10 -

Barrera, 197942 Comm Other Adults BAT Grp 8 WL 10 10 - BDI; MMPI–D

Barrett, 200143 Clin Mood Adults PST Ind 6 PLA 80 81 + HSCL–D–20

Beach, 199244 Comm Mood Adults CBT (Beck) Ind 18 WL 15 15 - BDI

Other Ind 18 WL 15 15 -

Bolton, 200345 Other Mood Adults IPT Grp 16 CAU 107 117 + HSC

Bowman, 199546 Comm Other Adults CBT (other) GSH 4 WL 10 10 - HRSD; BDI

PST GSH 4 WL 10 10 -

Brown, 198447 Comm Mood Adults CBT (other) Grp 12 WL 25 11 - BDI; CES–D

CBT (other) Ind 12 WL 13 11 -

CBT (other) GSH 12 WL 14 11 -

Burns, 200748 Other Other Older Other Ind 6 CAU 50 54 + GDS; HADS

Carpenter, 200849 Clin Mood Other BAT Ind 24 Other 18 20 + HRSD; BDI–II

Castonguay, 200450 Comm MDD Adults Other Ind 16 Other 11 10 - HRSD; BDI

Chen, 200051 Other Other PPD SUP Grp 4 CAU 30 30 - BDI

Chesney, 200352 Comm Other Gen med Other Grp 10 WL 46 38 - CES–D

Cho, 200853 Other Mood PPD CBT (other) Ind 9 CAU 10 12 - BDI

Christensen, 200454 Other Other Adults CBT (other) GSH 6 Other 182 178 + CES–D

Other GSH 6 Other 165 178 +

Clark, 200855 Clin MDD PPD Other Grp 12 WL 18 14 - BDI; PSI–d

Comas-Diaz, 198156 Clin Other Other CBT (Beck) Grp 5 WL 8 10 - HRSD; BDI

BAT Grp 5 WL 8 10 -

Conoley, 198557 Other Other Student Other Ind 2 WL 19 19 - BDI

Other Ind 2 WL 19 19 -

Conradi, 200758 Clin MDD Adults CBT (other) Ind 14 CAU 44 72 + BDI

Cooper, 200359 Other MDD PPD CBT (other) Ind 10 CAU 43 52 + EPDS

SUP Ind 10 CAU 48 52 +

Other Ind 10 CAU 50 52 +

DeRubeis, 200560 Comm MDD Adults CBT (Beck) Ind 14 PLA 60 60 + HRSD

Dimidjian, 200661 Comm MDD Adults BAT Ind 16 PLA 37 41 + HRSD; BDI

CBT (Beck) Ind 16 PLA 38 41 +

Dowrick, 200062 Other Mood Adults PST Ind 6 CAU 128 189 + BDI

CBT (other) Grp 8 CAU 108 189 +

Dunn, 200763 Clin Mood Other Other Grp 14 CAU 51 50 + HRSD; BDI–II

Elkin, 198964 Clin MDD Adults CBT (Beck) Ind 16 PLA 59 59 + HRSD; BDI

IPT Ind 16 PLA 59 59 +

Evans, 199565 Other Other Gen med CBT (other) Grp 8 Other 27 24 - CES–D; SCL–90–D

SUP Grp 8 Other 21 24 -

Faramarzi, 200866 Other MDD Other CBT (other) Grp 10 CAU 29 30 - BDI

Floyd, 200467 Comm Mood Older CBT (other) GSH 4 WL 13 7 - HRSD; GDS

CBT (Beck) Ind 16 WL 8 7 -

Fry, 198368 Other Other Older Other Ind 5 Other 54 27 - MMPI–D

Other Ind 5 Other 54 27 -

Fry, 198469 Other Other Older CBT (other) Ind 12 WL 15 10 - MMPI–D

Fuchs, 197770 Comm Other Adults Other Grp 6 WL 8 10 - BDI; MMPI–D

SUP Grp 6 WL 8 10 -

Hamamci, 200671 Comm Other Student CBT (other) Grp 11 CAU 8 11 - BDI

CBT (Beck) Grp 11 CAU 8 11 -

Haringsma, 200672 Comm Other Older CBT (other) Grp 10 WL 52 58 + CES–D; HADS–D

Harley, 200873 Clin MDD Adults Other Grp 16 WL 10 9 - HRSD; BDI–II

Hautzinger, 200474 Comm Mood Older CBT (other) Grp 12 WL 55 30 + GDS; IDS; SCL–90–D

Hayman, 198075 Clin Other Adults Other Grp 8 WL 16 12 - BDI

Holden, 198976 Other Mood PPD SUP Ind 8 CAU 26 24 - EPDS

Honey, 200277 Other Other PPD CBT (other) Grp 8 CAU 23 22 + EPDS

Jamison, 199578 Comm MDD Adults CBT (other) GSH 4 WL 33 39 - HRSD; BDI

Jarrett, 199979 Clin MDD Adults CBT (Beck) Ind 20 PLA 36 36 + HRSD; BDI

2

Table DS1 (continued)

Participants, n

Study

(first author, date) Recruitment Diagnosis Target group Type of therapy Format

No. of

sessions

Control

group EXP CTR ITT Measures

King, 200080 Clin Other Adults CBT (other) Ind 6 CAU 63 67 + BDI

SUP Ind 6 CAU 67 67 +

Klein, 198581 Comm Mood Adults CBT (other) Grp 12 Other 14 8 - SCL–90–D; CMI–D

Krampen, 199782 Clin Mood Adults CBT (other) Ind 20 WL 15 14 - BDI

Other Ind 20 WL 19 14 -

Laidlaw, 200883 Clin MDD Older CBT (Beck) Ind 8 CAU 20 20 - HRSD; SDS; BDI–II

Landreville, 199784 Comm Other Older CBT (other) GSH 4 WL 10 13 - BDI; GDS; IDD

Lincoln, 200385 Other Other Gen Med CBT (other) Ind 10 CAU 39 41 - BDI; WDI

Lopez, 200486 Clin Mood Adults Other Ind 12 PLA 10 10 - HRSD

Lustman, 199887 Comm MDD Gen Med CBT (other) Ind 10 CAU 20 22 + BDI

Lynch, 200488 Clin Other Adults PST GSH 6 CAU 9 13 - BDI; DHP–D

Other GSH 6 CAU 9 13 -

Lynch, 199789 Clin Other Adults PST GSH 6 CAU 11 13 - HRSD; BDI

Maina, 200590 Clin Mood Adults Other Ind 20 WL 10 10 + HRSD

SUP Ind 19 WL 10 10 +

Malouff, 198891 Comm Other Other CBT (other) Grp 4 WL 14 13 - BDI; POMS–D; BDRS

PST Grp 4 WL 14 13 -

Manson, 199592 Other Other Gen Med CBT (other) Grp 16 WL 22 26 - BDI

McLean, 197993 Comm MDD Adults Other Ind 10 Other 38 40 - BDI; Mood scr

Other Ind 10 Other 38 37 -

Meager, 199694 Comm Other PPD CBT (other) Grp 10 WL 6 6 - BDI; EPDS; POMS

Milgrom, 200595 Other Mood PPD CBT (other) Grp 9 CAU 46 33 + BDI

SUP Grp 9 CAU 47 33 +

SUP Ind 9 CAU 66 33 +

Miller, 200296 Other Other Adults IPT Ind 12 CAU 15 15 + HRSD

Miranda, 200397 Other MDD Other CBT (other) . 8 CAU 90 89 + HRSD

Mohr, 200098 Other Other Gen Med CBT (other) Ind 8 CAU 16 16 + POMS

Mossey, 199699 Other Other Older IPT Ind 10 CAU 31 38 - GDS

Murphy, 1995100 Comm MDD Adults CBT (Beck) Ind 20 Other 11 13 - HRSD; BDI

Mynors-Wallis, 1995101Clin MDD Adults PST Ind 6 PLA 30 30 - HRSD; BDI

Neugebauer, 2006102 Other Other Other IPT Ind 6 CAU 10 9 + HRSD

Nezu, 1986103 Comm Mood Adults PST Grp 8 WL 11 6 - BDI; MMPI–D

Other Grp 8 WL 9 6 -

Nezu, 1989104 Comm MDD Adults PST Grp 10 WL 14 6 - HRSD; BDI

PST Grp 10 WL 14 6 -

O’Hara, 2000105 Other MDD PPD IPT Ind 12 WL 48 51 + HRSD; BDI

Pace, 1993106 Other Other Student CBT (Beck) Ind 7 WL 31 43 - BDI

Pecheur, 1984107 Comm MDD Student CBT (Beck) Ind 8 WL 7 7 + HRSD; BDI

CBT (Beck) Ind 8 WL 7 7 +

Peden, 2000108 Comm Other Student CBT (other) Grp 6 CAU 46 46 - BDI; CES–D

Prendergast, 2001109 Other Mood PPD CBT (other) Ind 6 Other 17 20 - MADRS; EPDS

Propst, 1992110 Comm Other Adults CBT (other) Ind 19 WL 10 11 - HRSD; BDI

CBT (other) Ind 19 WL 9 11 -

CBT (other) Ind 19 WL 9 11 -

CBT (other) Ind 19 WL 10 11 -

Rahman, 2008111 Other MDD PPD CBT (other) Grp 8 Other 418 400 + HRSD

Ransom, 2008112 Comm Mood Gen med IPT Ind 6 CAU 41 38 + BDI–II

Rehm, 1981113 Comm Other Adults Other Grp 7 WL 9 15 - HRSD; BDI; MMPI–D;

Raskin DS

Robinson, 2007114 Comm Other Other Other Grp 8 CAU 42 54 - BDI–II

Rohan, 2007115 Comm MDD Adults CBT (other) Grp 12 WL 15 15 + HRSD; BDI–II; SIGH

Ross, 1985116 Clin MDD Adults CBT (Beck) Grp 12 WL 9 21 - BDI; MADRS

CBT (Beck) Ind 12 WL 21 21 -

Rude, 1986117 Comm Other Adults CBT (other) Grp 12 WL 14 16 - BDI

Other Grp 12 WL 14 16 -

Savard, 2006118 Comm Other Gen Med CBT (Beck) Ind 8 WL 21 16 - HRSD; BDI; HADS–D

Schmidt, 1983119 Comm Other Adults CBT (other) Ind 8 WL 12 10 - BDI; SDS; MMPI–D;

POMS–D

CBT (other) Grp 8 WL 11 10 -

CBT (other) Grp 8 WL 11 10 -

CBT (other) GSH 2 WL 12 10 -

Schulberg, 1996120 Clin MDD Adults IPT Ind 16 CAU 93 92 + HRSD

Scogin, 1987121 Comm Other Older CBT (other) GSH 4 WL 9 8 - HRSD; BDI; GDS

3

Table DS1 (continued)

Participants, n

Study

(first author, date) Recruitment Diagnosis Target group Type of therapy Format

No. of

sessions

Control

group EXP CTR ITT Measures

Scogin, 1989122 Comm Other Older CBT (other) GSH 4 WL 19 11 - HRSD; GDS

CBT (other) GSH 4 WL 21 11 -

Scott, 1992123 Clin MDD Adults CBT (Beck) Ind 16 CAU 29 29 + HRSD

SUP Ind 16 CAU 29 29 +

Scott, 1997124 Clin MDD Adults CBT (other) Ind 6 CAU 18 16 - HRSD; BDI

Scott, 1990125 Clin Mood Adults CBT (other) Grp 12 WL 17 23 + BDI

CBT (other) Ind 12 WL 27 23 +

Selmi, 1990126 Comm Mood Adults CBT (other) GSH 6 WL 12 12 + HRSD; BDI; SCL–90–D

CBT (other) Ind 6 WL 12 12 +

Serrano, 2004127 Other Other Older Other Ind 4 CAU 20 23 - CES–D

Shaw, 1977128 Comm Other Student CBT (Beck) Grp 8 WL 8 8 - HRSD; BDI

BAT Grp 8 WL 8 8 -

SUP Grp 8 WL 8 8 -

Simpson, 2003129 Clin Other Adults Other Ind 5 CAU 73 72 + BDI

Simson, 2008130 Comm Other Gen Med SUP Ind 5 CAU 15 15 + HADS–D

Sloane, 1985131 Other MDD Older IPT Ind 6 PLA 19 14 - HRSD; BDI

Spek, 2007132 Comm Other Older CBT (other) Grp 10 WL 99 100 + HRSD

Spinelli, 2003133 Clin MDD Other IPT Ind 16 Other 21 17 + HRSD; BDI; EPDS

Swartz, 2008134 Comm MDD Other IPT Ind 8 CAU 23 17 + HRSD; BDI

Taylor, 1977135 Comm Other Student CBT (Beck) Ind 6 WL 7 7 - BDI; D–30

BAT Ind 6 WL 7 7 -

CBT (other) Ind 6 WL 7 7 -

Teasdale, 1984136 Clin MDD Adults CBT (Beck) Ind 15 CAU 17 17 + HRSD; BDI; MADRS

Teichmann, 1995137 Clin Mood Adults CBT (other) Ind 13 WL 15 15 - BDI

CBT (Beck) Ind 13 WL 15 15 -

Teri, 1997138 Other Mood Gen Med BAT Ind 9 WL 23 20 - HRSD; BDI; CSDD

PST Ind 9 WL 19 20 -

Turner, 1979139 Comm Other Adults BAT Ind 5 Other 14 14 - DACL

Tyson, 1987140 Other Other Adults Other Grp 4 Other 11 11 - MMPI–D

Other Grp 4 Other 11 11 -

Other Grp 4 Other 11 11 -

Usaf, 1990141 Comm MDD Adults CBT (other) Grp 10 WL 25 15 - BDI; SDS

Van Schaik, 2006142 Clin MDD Older IPT Ind 10 CAU 69 74 + MADRS; GDS

Verduyn, 2003143 Other Other Other CBT (other) Grp 16 CAU 30 13 + HRSD; BDI

Warmerdam, 2008144 Comm Other Adults CBT (Beck) GSH 8 WL 88 87 + CES–D

PST GSH 5 WL 88 87 +

Watt, 2000145 Comm Other Older Other Grp 6 Other 9 4 - GDS

Other Grp 6 Other 9 4 -

Weissman, 1979146 Clin MDD Adults IPT Ind 16 Other 17 21 - HRSD; Raskin DS

Wickberg, 1996147 Other MDD PPD SUP Ind 6 CAU 20 21 - MADRS

Williams, 2000148 Comm Mood Older PST Ind 6 PLA 80 81 + HSCL–D–20

Wilson, 1983149 Comm Other Adults CBT (Beck) Ind 8 WL 8 9 - HRSD; BDI

BAT Ind 8 WL 8 9 -

Wollersheim, 1991150 Comm MDD Adults CBT (other) Grp 10 WL 8 8 - BDI; MMPI–D; SR;

CR; SO

SUP Grp 10 WL 8 8 -

CBT (other) GSH 5 WL 8 8 -

Wong, 2008151 Comm MDD Adults CBT (other) Grp 10 WL 48 40 - BDI

Wong, 2008152 Comm MDD Adults CBT (other) Grp 10 WL 163 159 + BDI

Wright, 2005153 Clin MDD Adults CBT (Beck) Ind 9 WL 15 15 + HRSD; BDI

CBT (Beck) GSH 9 WL 15 15 +

Zeiss, 1979154 Comm Other Adults CBT (other) Ind 12 WL 7 7 - MMPI–D

BAT Ind 12 WL 7 7 -

Other Ind 12 WL 7 7 -

BAT, behavioural activation treatment; BDI, Beck Depression Inventory; BDRS, Behavioral Depression Rating Scale; CAU, care as usual; CBT, cognitive–behavioural therapy; CES–D,Center for Epidemiological Studies – Depression scale; Clin, clinical; CMI–D, Cornell Medical Index – Depression scale; Comm, community; CR, clinical rating of depression; CSDD,Cornell Scale for Depression in Dementia; CTR, control; D–30, Depression 30 scale; DACL, Depression Adjective Checklist; DHP–D, Duke Health Profile – Depression scale; EPDS,Edinburgh Postnatal Depression Scale; EXP, experimental group; GDS, Geriatric Depression Scale; Gen med, general medical; Grp, group; GSH, guided self-help; HADS, HospitalAnxiety and Depression Scale; HADS–D, HADS Depression scale; HRSD, Hamilton Rating Scale for Depression; HSC, Hopkins Symptom Checklist; HSCL–D–20: HSCL Depression scale;IDS, Inventory of Depressive Symptoms; Ind, individual; IPT, interpersonal psychotherapy; ITT, intention to treat; MADRS, Montgomery–Asberg Depression Rating Scale; MDD, majordepressive disorder; MMPI–D, Minnesota Multiphasic Personality Inventory – Depression scale; Mood scr, mood screener; PLA, placebo; POMS, Profile of Mood States; POMS–D,POMS Depression scale; PPD, postpartum depression; PSI–d, Parenting Stress Index – depression scale; PST, problem-solving therapy; Raskin DS, Raskin Depression Scale;SCL–90–D, Symptom Checklist – 90 – Depression scale; SDS, Self-rating Depression Scale; SIGH, Structured Interview Guide for the HRSD plus eight-item Atypical Symptom subscale;SO, significant other rating of depression; SR, self-rating of depression scale; SUP, non-directive supportive therapy; WDI, Wakefield Self-Assessment of Depression Inventory;WL, waiting list.

References of trials listed in Table DS1

38 Allart-van Dam E, Hosman CM, Hoogduin CA, Schaap CP. Prevention ofdepression in subclinically depressed adults: follow-up effects on the’Coping with Depression’ course. J Affect Disord 2007; 97: 219–28.

39 Andersson G, Bergstrom J, Hollandare F, Carlbring P, Kaldo V, Ekselius L.Internet-based self-help for depression: randomised controlled trial. Br JPsychiatry 2005; 187: 456–61.

40 Arean PA, Perri MG, Nezu AM, Schein RL, Christopher F, Joseph TX.Comparative effectiveness of social problem-solving therapy andreminiscence therapy as treatments for depression in older adults. J ConsultClinical Psychol 1993; 61: 1003–10.

41 Ayen I, Hautzinger M. Kognitive Verhaltenstherapie bei Depressionen imKlimakterium: eine kontrollierte, randomisierte Interventionsstudie[Cognitive behaviour therapy for depression in menopausal women: acontrolled, randomised treatment study]. Zeitschr Klin Psychol Psychother2004; 33: 290–9.

42 Barrera M. An evaluation of a brief group therapy for depression. J ConsultClin Psychol 1979; 47: 413–5.

43 Barrett JE, Williams JW, Oxman TE, Frank E, Katon W, Sullivan M, et al.Treatment of dysthymia and minor depression in primary care: arandomized trial in patients aged 18 to 59 years. J Fam Pract 2001; 50:405–12.

44 Beach SR, O’Leary KD. Treating depression in the context of marital discord:outcome and predictors of response of marital therapy versus cognitivetherapy. Behav Ther 1992; 23: 507–28.

45 Bolton P, Bass J, Neugebauer R, Verdeli H, Clougherty KF, Wickramaratne P,et al. Group interpersonal psychotherapy for depression in rural Uganda: arandomized controlled trial. JAMA 2003; 289: 3117–24.

46 Bowman D, Scogin F, Lyrene B. The efficacy of self-examination therapy andcognitive bibliotherapy in the treatment of mild to moderate depression.Psychother Res 1995; 5: 131–40.

47 Brown RA, Lewinsohn PM. A psychoeducational approach to the treatmentof depression: comparison of group, individual and minimal contactprocedures. J Consult Clin Psychol 1984; 52: 774–83.

48 Burns A, Banerjee S, Morris J, Woodward Y, Baldwin R, Proctor R, et al.Treatment and prevention of depression after surgery for hip fracture inolder people: randomized, controlled trials. J Am Geriatr Soc 2007; 55:75–80.

49 Carpenter KM, Smith JL, Aharonovich E, Nunes EV. Developing therapiesfor depression in drug dependence: results of a stage 1 therapy study.Am J Drug Alcohol Abuse 2008; 34: 642–52.

50 Castonguay LG, Schutt AJ, Aikins DE, Constantino MJ, Laurenceau JP,Bologh L, et al. Integrative cognitive therapy for depression: a preliminaryinvestigation. J Psychother Integr 2004; 14: 4–20.

51 Chen CH, Tseng YF, Chou FH, Wang SY. Effects of support groupintervention in postnatally distressed women. A controlled study in Taiwan.J Psychosom Res 2000; 49: 395–9.

52 Chesney MA, Chambers DB, Taylor JM, Johnson LM, Folkman S. Copingeffectiveness training for men living with HIV: results from a randomizedclinical trial testing a group-based intervention. Psychosom Med 2003; 65:1038–46.

53 Cho HJ, Kwon JH, Lee JJ. Antenatal cognitive–behavioral therapy forprevention of postpartum depression: a pilot study. Yonsei Med J 2008; 49:553–62.

54 Christensen H. Delivering interventions for depression by using the internet:randomised controlled trial. BMJ 2004; 328: 265.

55 Clark R, Tluczek A, Brown R. A mother–infant therapy group model forpostpartum depression. Infant Ment Health J 2008; 29: 514–36.

56 Comas-Diaz L. Effects of cognitive and behavioral group treatment on thedepressive symptomatology of Puerto Rican women. J Consult Clin Psychol1981; 49: 627–32.

57 Conoley CW, Garber RA. Effects of reframing and self-control directiveson loneliness, depression and controllability. J Counsel Psychol 1985; 32:139–42.

58 Conradi HJ, de Jonge P, Kluiter H, Smit A, van der Meer K, Jenner JA, et al.Enhanced treatment for depression in primary care: long-term outcomes ofa psycho-educational prevention program alone and enriched withpsychiatric consultation or cognitive behavioral therapy. Psychol Med 2007;37: 849–62.

59 Cooper PJ, Murray L, Wilson A, Romaniuk H. Controlled trial of the short- andlong-term effect of psychological treatment of post-partum depression. 1.Impact on maternal mood. Br J Psychiatry 2003; 182: 412–9.

60 DeRubeis RJ, Hollon SD, Amsterdam JD, Shelton RC, Young PR, Salomon RM,et al. Cognitive therapy vs medications in the treatment of moderate tosevere depression. Arch Gen Psychiatry 2005; 62: 409–16.

61 Dimidjian S, Hollon SD, Dobson KS, Schmaling KB, Kohlenberg RJ, Addis ME,et al. Randomized trial of behavioral activation, cognitive therapy, andantidepressant medication in the acute treatment of adults with majordepression. J Consult Clin Psychol 2006; 74: 658–70.

62 Dowrick C, Dunn G, Ayuso-Mateos JL, Dalgard OS, Page H, Lehtinen V, et al.Problem solving treatment and group psychoeducation for depression:multicentre randomised controlled trial. BMJ 2000; 321: 1450–4.

63 Dunn NJ, Rehm LP, Schillaci J, Souchek J, Mehta P, Ashton CM, et al.A randomized trial of self-management and psychoeducational grouptherapies for comorbid chronic posttraumatic stress disorder anddepressive disorder. J Trauma Stress 2007; 20: 221–37.

64 Elkin I, Shea MT, Watkins JT, Imber SD, Sotsky SM, Collins JF, et al. NationalInstitute of Mental Health Treatment of Depression Collaborative ResearchProgram. General effectiveness of treatments. Arch Gen Psychiatry 1989;46: 971–82.

65 Evans RL, Connis RT. Comparison of brief group therapies for depressedcancer patients receiving radiation treatment. Publ Health Rep 1995; 110:306–11.

66 Faramarzi M, Alipor A, Esmaelzadeh S, Kheirkhah F, Poladi K, Pash H.Treatment of depression and anxiety in infertile women: cognitivebehavioral therapy versus fluoxetine. J Affect Disord 2008; 108: 159–64.

67 Floyd M, Scogin F, McKendree-Smith NL, Floyd DL, Rokke PD. Cognitivetherapy for depression: a comparison of individual psychotherapy andbibliotherapy for depressed older adults. Behav Modif 2004; 28: 297–318.

68 Fry PS. Structured and unstructured reminiscence training and depressionamong the elderly. Clin Gerontol 1983; 1: 15–37.

69 Fry PS. Cognitive training and cognitive–behavioral variables in thetreatment of depression in the elderly. Clin Gerontol 1984; 3: 25–45.

70 Fuchs CZ, Rehm LP. A self-control behavior therapy program for depression.J Consult Clin Psychol 1977; 45: 206–15.

71 Hamamci Z. Integrating psychodrama and cognitive behavioral therapy totreat moderate depression. Arts Psychother 2006; 33: 199–207.

72 Haringsma R, Engels GI, Cuijpers P, Spinhoven P. Effectiveness of the CopingWith Depression (CWD) course for older adults provided by the community-based mental health care system in the Netherlands: a randomizedcontrolled field trial. Int Psychogeriatr 2006; 18: 307–25.

73 Harley R, Sprich S, Safren S, Jacobo M, Fava M. Adaptation of dialecticalbehavior therapy skills training group for treatment-resistant depression.J Nerv Ment Dis 2008; 196: 136–43.

74 Hautzinger M, Welz S. Verhaltenstherapie bei Depressionen im Alter:Ergebnisse einer kontrollierten Vergleichsstudie unter AmbulantenBedingungen an Depressionen Mittleren Schweregrads [Cognitive–behavioural therapy for depressed older outpatients – a controlled,randomised trial]. Zeitschr Gerontol Geriatr 2004; 37: 427–35.

75 Hayman PM, Cope CS. Effects of assertion training on depression. J ClinPsychol 1980; 36: 534–43.

76 Holden JM, Sagovsky R, Cox JL. Counselling in a general practice setting:controlled study of health visitor intervention in treatment of postnataldepression. BMJ 1989; 298: 223–6.

77 Honey KL, Bennett P, Morgan M. A brief psycho-educational groupintervention for postnatal depression. Br J Clin Psychol 2002; 41: 405–9.

78 Jamison C, Scogin F. The outcome of cognitive bibliotherapy with depressedadults. J Consult Clin Psychol 1995; 63: 644–50.

79 Jarrett RB, Schaffer M, McIntire D, Witt-Browder A, Kraft D, Risser RC.Treatment of atypical depression with cognitive therapy or phenelzine: adouble-blind, placebo-controlled trial. Arch Gen Psychiatry 1999; 56: 431–7.

80 King M, Sibbald B, Ward E, Bower P, Lloyd M, Gabbay M, et al. Randomisedcontrolled trial of non-directive counselling, cognitive-behaviour therapy andusual general practitioner care in the management of depression as well asmixed anxiety and depression in primary care. Health Technol Assess 2000;4: 1–83.

81 Klein M, Greist J, Gurman A, Neimeyer R, Lesser D, Bushnell N, et al.Comparative outcome study of group psychotherapy vs. exercisetreatments for depression. Int J Ment Health 1985; 13: 148–77.

82 Krampen G. Autogenes Training vor und begleitend zurMethodenubergreifenden Einzelpsychotherapie bei Depressiven Storungen[Application of autogenic training before and in addition to integrativepsychotherapy of depressive disorders]. Zeitschr Klin Psychol PsychiatrPsychother 1997; 45: 214–32.

83 Laidlaw K, Davidson K, Toner H, Jackson G, Clark S, Law J, et al. Arandomised controlled trial of cognitive behaviour therapy vs treatment as

4

usual in the treatment of mild to moderate late life depression. Int J GeriatrPsychiatry 2008; 23: 843–50.

84 Landreville P, Bissonnette L. Effects of cognitive bibliotherapy for depressedolder adults with a disability. Clin Gerontol 1997; 17: 35–55.

85 Lincoln NB, Flannaghan T. Cognitive behavioral psychotherapy fordepression following stroke: a randomized controlled trial. Stroke 2003; 34:111–5.

86 Lopez Rodriguez J, Lopez Butron MA, Vargas Terrez BE, Villamil Salcedo V.Estudio doble ciego con antidepresivo, psicoterapia breve y placebo enpacientes con depresion leve a moderada [Double blind study withantidepressant, brief psychotherapy and placebo in patients with mild tomoderate depression]. Salud Mental 2004; 27: 53–61.

87 Lustman PJ, Griffith LS, Freedland KE, Kissel SS, Clouse RE. Cognitivebehavior therapy for depression in type 2 diabetes mellitus. A randomized,controlled trial. Ann Intern Med 1998; 129: 613–21.

88 Lynch DJ, Tamburrino MB, Nagel R. Telephone counseling for patients withminor depression: preliminary findings in a family practice setting. J FamPract 1997; 44: 293–8.

89 Lynch D, Tamburrino M, Nagel R, Smith MK. Telephone-based treatmentfor family practice patients with mild depression. Psychol Rep 2004; 94:785–92.

90 Maina G, Forner F, Bogetto F. Randomized controlled trial comparing briefdynamic and supportive therapy with waiting list condition in minordepressive disorders. Psychother Psychosom 2005; 74: 43–50.

91 Malouff JM, Lanyon RI, Schutte NS. Effectiveness of a brief group treatmentfor divorce-related dysphoria. J Ration Emot Cogn Behav Ther 1988; 6:162–71.

92 Manson SM, Brenneman DL. Chronic disease among older American Indians:preventing depressive symptoms and related problems of coping. InHandbook on Ethnicity, Aging, and Mental Health (ed DK Pagett): 284–303.Greenwood Press, 1995.

93 McLean PD, Hakstian AR. Clinical depression: comparative efficacy ofoutpatient treatments. J Consult Clin Psychol 1979; 47: 818–36.

94 Meager I, Milgrom J. Group treatment for postpartum depression: a pilotstudy. Aust N Z J Psychiatry 1996; 30: 852–60.

95 Milgrom J, Negri LM, Gemmill AW, McNeil M, Martin PR. A randomizedcontrolled trial of psychological interventions for postnatal depression.Br J Clin Psychol 2005; 44: 529–42.

96 Miller L, Weissman M. Interpersonal psychotherapy delivered over thetelephone to recurrent depressives. A pilot study. Depress Anx 2002; 16:114–7.

97 Miranda J, Chung JY, Green BL, Krupnick J, Siddique J, Revicki DA, et al.Treating depression in predominantly low-income young minority women: arandomized controlled trial. JAMA 2003; 290: 57–65.

98 Mohr DC, Likosky W, Bertagnolli A, Goodkin DE, Van Der Wende J, Dwyer P,et al. Telephone-administered cognitive–behavioral therapy for thetreatment of depressive symptoms in multiple sclerosis. J Consult ClinPsychol 2000; 68: 356–61.

99 Mossey JM, Knott KA, Higgins M, Talerico K. Effectiveness of a psychosocialintervention, interpersonal counseling, for subdysthymic depression inmedically ill elderly. J Gerontol A Biol Sci Med Sci 1996; 51: M172–8.

100 Murphy GE, Carney RM, Knesevich MA, Wetsel RD, Whitworth P. Cognitivebehavior therapy, relaxation training and tricyclic antidepressant medicationin the treatment of depression. Psychol Rep 1995; 77: 403–20.

101 Mynors-Wallis LM, Gath DH, Lloyd-Thomas AR, Tomlinson D. Randomisedcontrolled trial comparing problem solving treatment with amitriptyline andplacebo for major depression in primary care. BMJ 1995; 310: 441–5.

102 Neugebauer R, Kline J, Markowitz JC, Bleiberg KL, Baxi L, Rosing MA, et al.Pilot randomized controlled trial of interpersonal counseling forsubsyndromal depression following miscarriage. J Clin Psychiatry 2006; 67:1299–304.

103 Nezu A. Efficacy of social problem-solving therapy approach for unipolardepression. J Consult Clin Psychol 1986; 54: 196–202.

104 Nezu AM, Perri MG. Social problem-solving therapy for unipolar depression:an initial dismantling investigation. J Consult Clin Psychol 1989; 57: 408–13.

105 O’Hara MW, Stuart S, Gorman LL, Wenzel A. Efficacy of interpersonaltherapy for postpartum depression. Arch Gen Psychiatry 2000; 57: 1039–45.

106 Pace TM, Dixon DN. Changes in depressive self-schemata and depressivesymptoms following cognitive therapy. J Counsel Psychol 1993; 40: 288–94.

107 Pecheur DR, Edwards KJ. A comparison of secular and religious versions ofcognitive therapy with depressed Christian college students. J Psychol Theol1984; 12: 45–54.

108 Peden AR, Hall LA, Rayens MK, Beebe LL. Reducing negative thinking anddepressive symptoms in college women. J Nurs Scholarsh 2000; 32: 145–51.

109 Prendergast J, Austin MP. Early childhood nurse-delivered cognitivebehavioural counselling for post-natal depression. Australas Psychiatry2001; 9: 255–9.

110 Propst LR, Ostrom R, Watkins P, Dean T, Mashburn D. Comparative efficacyof religious and nonreligious cognitive–behavioral therapy for the treatmentof clinical depression in religious individuals. J Consult Clin Psychol 1992;60: 94–103.

111 Rahman A, Malik A, Sikander S, Roberts C, Creed F. Cognitive behaviourtherapy-based intervention by community health workers for mothers withdepression and their infants in rural Pakistan: a cluster-randomisedcontrolled trial. Lancet 2008; 372: 902–9.

112 Ransom D, Heckman TG, Anderson T, Garske J, Holroyd K, Basta T.Telephone-delivered, interpersonal psychotherapy for HIV-infected ruralpersons with depression: a pilot trial. Psychiatr Serv 2008; 59: 871–7.

113 Rehm LP, Kornblith SJ, O’Hara MW, Lamparski DM, Romano JM, Vokin JI.An evaluation of major components in a self-control therapy program fordepression. Behav Mod 1981; 5: 459–89.

114 Robinson-Whelen S, Hughes RB, Taylor HB, Hall JW, Rehm LP. Depressionself-management program for rural women with physical disabilities. RehabPsychol 2007; 52: 254–62.

115 Rohan KJ, Roecklein KA, Tierney Lindsey K, Johnson LG, Lippy RD, Lacy TJ,et al. A randomized controlled trial of cognitive-behavioral therapy, lighttherapy, and their combination for seasonal affective disorder. J Consult ClinPsychol 2007; 75: 489–500.

116 Ross M, Scott M. An evaluation of the effectiveness of individual and groupcognitive therapy in the treatment of depressed patients in an inner cityhealth centre. J R Coll Gen Pract 1985; 35: 239–42.

117 Rude SS. Relative benefits of assertion or cognitive self-control treatmentfor depression as a function of proficiency in each domain. J Consult ClinPsychol 1986; 54: 390–4.

118 Savard J, Simard S, Giguere I, Ivers H, Morin CM, Maunsell E, et al.Randomized clinical trial on cognitive therapy for depression in women withmetastatic breast cancer: psychological and immunological effects. PalliatSupport Care 2006; 4: 219–37.

119 Schmidt MM, Miller WR. Amount of therapist contact and outcome in amultidimensional depression treatment program. Acta Psychiatr Scand1983; 67: 319–32.

120 Schulberg HC, Block MR, Madonia MJ, Scott CP, Rodriguez E, Imber SD, et al.Treating major depression in primary care practice. Eight-month clinicaloutcomes. Arch Gen Psychiatry 1996; 53: 913–9.

121 Scogin F, Hamblin D, Beutler L. Bibliotherapy for depressed older adults:a self-help alternative. Gerontologist 1987; 27: 383–7.

122 Scogin F, Jamison C, Gochneaur K. Comparative efficacy of cognitive andbehavioral bibliotherapy for mildly and moderately depressed older adults.J Consult Clin Psychol 1989; 57: 403–7.

123 Scott AI, Freeman CP. Edinburgh primary care depression study: treatmentoutcome, patient satisfaction, and cost after 16 weeks. BMJ 1992; 304:883–7.

124 Scott C, Tacchi MJ, Jones R, Scott J. Acute and one-year outcome of arandomised controlled trial of brief cognitive therapy for major depressivedisorder in primary care. Br J Psychiatry 1997; 171: 131–4.

125 Scott MJ, Stradling SG. Group cognitive therapy for depression producesclinically significant reliable change in community-based settings. BehavPsychother 1990; 18: 1–19.

126 Selmi PM, Klein MH, Greist JH, Sorrell S, Erdman HP. Computer-administeredcognitive–behavioral therapy for depression. Am J Psychiatry 1990; 147:51–6.

127 Serrano JP, Latorre JM, Gatz M, Montanes J. Life review therapy usingautobiographical retrieval practice for older adults with depressivesymptomatology. Psychol Aging 2004; 19: 270–7.

128 Shaw BF. Comparison of cognitive therapy and behavior therapy in thetreatment of depression. J Consult Clin Psychol 1977; 45: 543–51.

129 Simpson S, Corney R, Fitzgerald P, Beecham J. A randomized controlled trialto evaluate the effectiveness and cost-effectiveness of psychodynamiccounselling for general practice patients with chronic depression. PsycholMed 2003; 33: 229–39.

130 Simson U, Nawarotzky U, Friese G, Porck W, Schottenfeld-Naor Y, Hahn S,et al. Psychotherapy intervention to reduce depressive symptoms inpatients with diabetic foot syndrome. Diabet Med 2008; 25: 206–12.

131 Sloane RB, Staples FR, Schneider LS. Interpersonal therapy vs. nortriptylinefor depression in the elderly. In Clinical and Pharmacological Studies inPsychiatric Disorders (eds GD Burrows, TR Norman, L Dennerstein): 344–6.John Libbey, 1985.

132 Spek V, Nyklicek I, Smits N, Cuijpers P, Riper H, Keyzer J, et al. Internet-based cognitive behavioural therapy for subthreshold depression in people

5

over 50 years old: a randomized controlled clinical trial. Psychol Med 2007;37: 1797–806.

133 Spinelli MG, Endicott J. Controlled clinical trial of interpersonalpsychotherapy versus parenting education program for depressed pregnantwomen. Am J Psychiatry 2003; 160: 555–62.

134 Swartz HA, Frank E, Zuckoff A, Cyranowski JM, Houck PR, Cheng Y, et al.Brief interpersonal psychotherapy for depressed mothers whose childrenare receiving psychiatric treatment. Am J Psychiatry 2008; 165: 1155–62.

135 Taylor FG, Marshall WL. Experimental analysis of a cognitive–behavioraltherapy for depression. Cogn Ther Res 1977; 1: 59–72.

136 Teasdale JD, Fennell MJ, Hibbert GA, Amies PL. Cognitive therapy for majordepressive disorder in primary care. Br J Psychiatry 1984; 144: 400–6.

137 Teichman Y, Bar-el Z, Shor H, Sirota P, Elizur A. A comparison of twomodalities of cognitive therapy (individual and marital) in treatingdepression. Psychiatry 1995; 58: 136–48.

138 Teri L, Logsdon RG, Uomoto J, McCurry SM. Behavioral treatment ofdepression in dementia patients: a controlled clinical trial. J Gerontol BPsychol Sci Soc Sci 1997; 52: P159–66.

139 Turner RW, Ward MF, Turner DJ. Behavioral treatment for depression: anevaluation of therapeutic components. J Clin Psychol 1979; 35: 166–75.

140 Tyson GM, Range LM. Gestalt dialogues as a treatment for mild depression:time works just as well. J Clin Psychol 1987; 43: 227–31.

141 Usaf SO, Kavanagh DJ. Mechanisms of improvement in treatment fordepression: test of a self-efficacy and performance model. J CognPsychother 1990; 4: 51–70.

142 Van Schaik A, van Marwijk H, Ader H, van Dyck R, de Haan M, Penninx B,et al. Interpersonal psychotherapy for elderly patients in primary care.Am J Geriatr Psychiatry 2006; 14: 777–86.

143 Verduyn C, Barrowclough C, Roberts J, Tarrier N, Harrington R. Maternaldepression and child behaviour problems. Randomised placebo-controlledtrial of a cognitive–behavioural group intervention. Br J Psychiatry 2003;183: 342–8.

144 Warmerdam L, van Straten A, Twisk J, Riper H, Cuijpers P. Internet-basedtreatment for adults with depressive symptoms: randomized controlled trial.J Med Internet Res 2008; 10: e44.

145 Watt LM, Cappeliez P. Integrative and instrumental reminiscence therapiesfor depression in older adults: intervention strategies and treatmenteffectiveness Aging Ment Health 2000; 4: 166–77.

146 Weissman MM, Prusoff BA, Dimascio A, Neu C, Goklaney M, Klerman GL.The efficacy of drugs and psychotherapy in the treatment of acutedepressive episodes. Am J Psychiatry 1979; 136: 555–8.

147 Wickberg B, Hwang CP. Counselling of postnatal depression: a controlledstudy on a population based Swedish sample. J Affect Disord 1996; 39:209–16.

148 Williams JW, Barrett J, Oxman T, Frank E, Katon W, Sullivan M, et al.Treatment of dysthymia and minor depression in primary care: arandomized controlled trial in older adults. JAMA 2000; 284: 1519–26.

149 Wilson PH, Goldin JC, Charbonneau Powis M. Comparative efficacy ofbehavioral and cognitive treatments of depression. Cogn Ther Res 1983; 7:111–24.

150 Wollersheim JP, Wilson GL. Group treatment of unipolar depression: acomparison of coping, supportive, bibliotherapy, and delayed treatmentgroups. Prof Psychol Res Pract 1991; 22: 496–502.

151 Wong DFK. Cognitive behavioral treatment groups for people with chronicdepression in Hong Kong: a randomized wait-list control design. DepressAnx 2008; 25: 142–8.

152 Wong DF. Cognitive and health-related outcomes of group cognitivebehavioural treatment for people with depressive symptoms in Hong Kong:randomized wait-list control study. Aust N Z J Psychiatry 2008; 42: 702–11.

153 Wright JH, Wright AS, Albano AM, Basco MR, Goldsmith LJ, Raffield T, et al.Computer-assisted cognitive therapy for depression: maintaining efficacywhile reducing therapist time. Am J Psychiatry 2005; 162: 1158–64.

154 Zeiss AM, Lewinsohn PM, Munoz RF. Nonspecific improvement effects indepression using interpersonal skills training, pleasant activity schedules orcognitive training. J Consult Clin Psychol 1979; 47: 427–39.

6

7

Table DS2 Mean effect sizes for psychological treatments of adult depression from published studies: tests for publication bias

in subgroups

Unadjusted values Adjusted valuesa Egger’s test

Nc d 95% CI I2 d 95% CI Nimpa Taub I c 95% CI P

Recruitment

Community 90 0.81 0.68–0.93 63.65*** 0.50 0.37–0.63 30 0.42*** 2.00 1.37–2.64 ***

Clinical 41 0.51 0.38–0.64 52.88*** 0.32 0.18–0.46 14 0.49*** 2.51 1.59–3.44 ***

Other 44 0.59 0.44–0.75 81.99*** 0.59 0.44–0.75 0 0.12 (NS) 0.76 70.74 to 2.26 NS

Diagnosis

Mood disorder 92 0.67 0.55–0.78 75.74*** 0.41 0.28–0.53 26 0.40*** 1.60 0.77–2.42 ***

Self-report/other 83 0.69 0.57–0.80 60.92*** 0.43 0.31–0.56 26 0.37*** 1.62 0.97–2.28 ***

Target group

Adults 95 0.63 0.52–0.74 69.48*** 0.38 0.26–0.50 29 0.33*** 1.42 0.72–2.12 ***

Older adults 23 0.75 0.47–1.02 80.82*** 0.30 0.00–0.60 10 0.33* 2.58 0.44–4.72 *

Student population 14 1.04 0.72–1.36 47.84* 0.76 0.41–1.11 5 0.74*** 2.43 0.95–3.90 **

Women with PPD 16 0.53 0.40–0.67 31.47 (NS) 0.49 0.34–0.65 1 0.21 (NS) 70.10 71.39 to 1.18 NS

General medical 12 0.56 0.33–0.80 49.32* 0.30 0.04–0.56 5 0.35o 5.49 2.30–8.69 **

Other 15 0.96 0.57–1.36 82.52*** 0.38 70.05 to 0.82 6 0.57** 3.95 1.58–6.31 **

Type of therapy

CBT 89 0.69 0.59–0.79 60.50*** 0.49 0.38–0.60 26 0.42*** 1.48 0.90–2.05 ***

Interpersonal 13 0.57 0.22–0.92 86.38*** 0.57 0.22–0.92 0 0.06 (NS) 71.65 70.74 to 3.44 NS

Problem-solving 14 0.80 0.46–1.14 82.04*** 0.28 70.08 to 0.64 6 0.51** 3.45 1.82–5.07 ***

Supportive 15 0.55 0.36–0.74 33.79o 0.55 0.36–0.74 0 0.32* 1.80 70.06 to 3.67 *

Behavioural

activation 10 0.70 0.25–1.14 66.49** 0.35 70.14 to 0.84 3 0.53* 3.07 70.16 to 6.30 *

Other 34 0.65 0.44–0.86 76.58*** 0.31 0.07–0.55 11 0.24* 1.83 0.30–3.35 *

Format

Individual 89 0.63 0.52–0.74 68.84*** 0.32 0.20–0.44 33 0.48*** 2.83 2.08–3.59 ***

Group 64 0.75 0.61–0.89 73.10*** 0.52 0.36–0.68 14 0.31*** 0.93 0.07–1.80 *

Guided self-help 21 0.68 0.49–0.87 56.33** 0.47 0.26–0.68 6 0.20o 1.31 0.18–2.45 *

Control group

Waiting list 93 0.90 0.79–1.01 49.05*** 0.68 0.56–0.80 27 0.38*** 1.71 1.13–2.30 ***

Care as usual 48 0.48 0.35–0.61 71.71*** 0.29 0.14–0.44 13 0.31** 1.39 70.10 to 2.89 *

Pill placebo 11 0.29 0.09–0.49 58.09** 0.19 70.04 to 0.42 2 0.47* 3.92 1.29–6.55 **

Other control 23 0.55 0.31–0.79 82.73*** 0.55 0.31–0.79 0 70.03 (NS) 0.18 71.61 to 1.98 NS

Analyses

Intention to treat 62 0.47 0.38–0.57 72.37*** 0.37 0.26–0.49 10 0.36*** 0.91 70.21 to 2.02 o

Completers only 113 0.84 0.72–0.96 63.91*** 0.55 0.42–0.69 30 0.30*** 2.12 1.16–3.08 ***

CBT, cognitive–behavioural therapy; Ncomp, number of comparisons; Nimp, number of imputed studies; NS, not significant; PPD, postpartum depression.a. Based on Duval & Tweedie trim and fill procedure.b. Kendall’s tau based on Begg & Mazumbar rank correlation.c. Intercept based on Egger’s test.oP50.1, *P50.05, **P50.01, ***P50.001.

8

Table

DS3

Me

an

eff

ec

tsiz

es

of

co

gn

itiv

e–

be

ha

vio

ura

lth

era

py

for

ad

ult

de

pre

ssio

nfr

om

pu

bli

sh

ed

stu

die

s:

tests

for

pu

bli

ca

tio

nb

ias

insu

bg

rou

ps

Un

ad

just

ed

valu

es

Ad

just

ed

valu

esa

Egger’

ste

st

Nc

d95%

CI

I2d

95%

CI

Nim

pTa

ub

Ic

95%

CI

P

All

com

par

iso

ns

890.

690.

59–0

.79

60.5

0***

0.49

0.38

–0.6

026

0.42

***

1.48

0.90

–2.0

5**

*

Ou

tlie

rse

xclu

de

dc

780.

570.

49–0

.65

41.7

6***

0.46

0.38

–0.5

519

0.31

***

0.93

0.37

–1.4

8**

Mu

ltip

leco

mp

aris

on

se

xclu

din

glo

we

st70

0.61

0.51

–0.7

261

.09*

**0.

470.

35–0

.58

160.

34**

*1.

160.

45–1

.87

**

Mu

ltip

leco

mp

aris

on

se

xclu

din

gh

igh

est

700.

680.

57–0

.78

65.5

9***

0.48

0.36

–0.6

020

0.41

***

1.53

0.82

–2.2

4**

*

BD

Io

nly

660.

820.

68–0

.95

63.1

7***

0.52

0.38

–0.6

624

0.43

***

2.26

1.51

–3.0

2**

*

HR

SDo

nly

350.

750.

59–0

.91

57.5

2***

0.53

0.36

–0.7

012

0.62

***

1.42

0.64

–2.2

0**

*

Subgroupan

alyses

Re

cru

itme

nt

Co

mm

un

ity54

0.85

0.71

–0.9

957

.51*

**0.

620.

47–0

.78

160.

43**

*1.

630.

84–2

.41

***

Clin

ical

180.

570.

39–0

.74

36.4

9o0.

380.

19–0

.57

70.

63**

*3.

131.

84–4

.41

***

Oth

er

170.

460.

28–0

.63

67.4

3***

0.46

0.28

–0.6

30

0.24

o0.

227

1.56

to2.

00N

S

Dia

gno

sis

Mo

od

dis

ord

er

480.

650.

51–0

.80

68.5

6***

0.48

0.33

–0.6

410

0.28

**1.

080.

13–2

.03

*

Self-

rep

ort

/oth

er

410.

730.

59–0

.86

45.1

2**

0.51

0.36

–0.6

516

0.56

***

2.07

1.47

–2.6

6**

*

Targ

et

gro

up

Ad

ults

500.

630.

51–0

.75

54.6

9***

0.44

0.31

–0.5

715

0.27

**1.

390.

57–2

.21

***

Spe

cific

gro

up

390.

790.

63–0

.95

66.1

8***

0.55

0.37

–0.7

312

0.54

***

1.62

0.77

–2.4

7**

*

Typ

eo

fth

era

py

Be

ck’s

CB

T25

0.77

0.58

–0.9

651

.27*

*0.

570.

36–0

.78

80.

66**

*2.

771.

90–3

.64

***

Oth

er

CB

T64

0.67

0.55

–0.7

863

.46*

**0.

480.

35–0

.60

170.

30**

*1.

200.

48–1

.93

***

Form

at

Ind

ivid

ual

410.

660.

52–0

.81

50.2

6***

0.41

0.25

–0.5

815

0.58

***

3.00

2.24

–3.7

7**

*

Gro

up

320.

690.

53–0

.86

68.6

9***

0.53

0.35

–0.7

17

0.30

*1.

087

0.01

to2.

17o

Gu

ide

dse

lf-h

elp

150.

840.

59–1

.09

58.7

0**

0.53

0.27

–0.8

06

0.23

(NS)

1.66

0.28

–3.0

4*

Co

ntr

ol

gro

up

Wai

ting

list

570.

880.

75–1

.02

46.6

2***

0.66

0.52

–0.8

118

0.43

***

1.62

0.93

–2.3

2**

*

Car

eas

usu

al21

0.50

0.31

–0.6

864

.62*

**0.

250.

04–0

.45

80.

49**

3.05

1.50

–4.6

0**

*

Oth

er

con

tro

l11

0.42

0.25

–0.5

950

.94*

0.42

0.25

–0.5

90

70.

15(N

S)7

1.30

72.

96to

0.37

NS

An

alys

es

Inte

ntio

nto

tre

at33

0.51

0.40

–0.6

256

.71*

**0.

440.

32–0

.55

70.

43**

*1.

067

0.01

to2.

13o

Co

mp

lete

rso

nly

560.

850.

69–1

.00

54.9

2***

0.64

0.47

–0.8

113

0.34

***

1.89

0.64

–3.1

5**

BD

I,B

eck

De

pre

ssio

nIn

ven

tory

;C

BT,

cogn

itive

–be

hav

iou

ral

the

rap

y;H

RSD

,H

amilt

on

Rat

ing

Scal

efo

rD

ep

ress

ion

;N

c,

nu

mb

er

of

com

par

iso

ns;

Nim

p,

nu

mb

er

of

imp

ute

dst

ud

ies;

NS,

no

tsi

gnifi

can

t.a.

Bas

ed

on

Du

val

&Tw

ee

die

’str

im&

fill

pro

ced

ure

.b

.K

en

dal

l’sta

ub

ase

do

nB

egg

&M

azu

mb

arra

nk

corr

ela

tion

.c.

Inte

rce

pt

bas

ed

on

Egg

er’

ste

st.

oP5

0.1,

*P5

0.05

,**

P5

0.01

,**

*P5

0.00

1.

10.1192/bjp.bp.109.066001Access the most recent version at DOI: 2010, 196:173-178.BJP 

Pim Cuijpers, Filip Smit, Ernst Bohlmeijer, Steven D. Hollon and Gerhard Anderssonstudy of publication biaspsychological treatments for adult depression: meta-analytic

behavioural therapy and other−Efficacy of cognitive

MaterialSupplementary

http://bjp.rcpsych.org/content/suppl/2010/03/03/196.3.173.DC1.htmlSupplementary material can be found at:

Referenceshttp://bjp.rcpsych.org/content/196/3/173#BIBLThis article cites 29 articles, 2 of which you can access for free at:

permissionsReprints/

permissions@rcpsych.ac.ukwrite to To obtain reprints or permission to reproduce material from this paper, please

to this article atYou can respond /letters/submit/bjprcpsych;196/3/173

from Downloaded

The Royal College of PsychiatristsPublished by on April 23, 2016http://bjp.rcpsych.org/

http://bjp.rcpsych.org/site/subscriptions/ go to: The British Journal of PsychiatryTo subscribe to