Adding Psychotherapy to
Pharmacotherapy in the Treatment of
Depressive Disorders in Adults: A Meta-Analysis
Objective: A considerable number of studies has examined whether adding psychotherapy to pharmacotherapy results in stronger effects than pharmacotherapy alone. However, earlier meta-analyses in this field have included only a limited number of available studies and did not conduct extended subgroup analyses to examine possible sources of heterogeneity.
Data Sources: We used a database derived
from a comprehensive literature search in PubMed, PsycINFO, EMBASE, and the Cochrane Central Register of Controlled Trials for studies published from 1966 to January 2008 that examined the psychological treatment of depression. The abstracts
of these studies were identified by combining
terms indicative of psychological treatment
Study Selection: We included randomized trials in which the effects of a pharmacologic treatment were compared to the effects of a combined pharmacologic and psychological treatment in adults with a depressive disorder.
Data Extraction: For each of the studies, we calculated a standardized mean effect size indicating the difference between pharmacotherapy and the combined treatment at posttest. We also coded major characteristics of the population, the interventions, and the quality and design of the study.
Data Synthesis: Twenty-five randomized
trials, with a total of 2,036 patients, were included.
A mean effect size of d = 0.31 (95% CI, 0.20 ~ 0.43) was found for the 25 included studies, indicating a small effect in favor of the combined treatment over pharmacotherapy alone. Studies aimed at patients with dysthymia resulted in significantly lower effect sizes compared to studies aimed at patients with major depression, a finding that suggests that the added value of psychotherapy is less in patients with dysthymia. The dropout rate was significantly lower in the combined treatment group compared to the pharmacotherapy only group (OR = 0.65; 95% CI, 0.50 ~ 0.83).
Conclusions: Psychotherapy seems to have an additional value compared to pharmacotherapy alone for depression.
J Clin Psychiatry 2009;70(9):1219–1229
© Copyright 2009 Physicians Postgraduate Press, Inc.
Submitted: January 7, 2009; accepted February 19, 2009
Corresponding author: Pim Cuijpers, PhD, Department of Clinical Psychology, VU University Amsterdam, Van der Boechorststraat 1, 1081 BT Amsterdam, The Netherlands (P.Cuijpers@psy.vu.nl).
Both pharmacologic1 and psychological treatments2 for depression in adults have been examined in a considerable number of studies. The combination of the 2 treatments has been less well-examined, however. A growing body of evidence that combined treatment is more effective than psychotherapy alone3–5 suggests that adding pharmacotherapy has an independent, cumulative effect on depression. Research is less conclusive, however, on the question of whether combined treatment is more effective than pharmacotherapy alone. Some studies do find support for this,6–8 but others do not.9–11 Because it can be expected that the difference between psychological and combined treatments is small, large sample sizes are required to find significant differences. When small effect sizes are expected in individual studies, meta-analytic techniques can be used to integrate the results of individual studies and to increase the statistical power.12
Although 2 earlier meta-analytic studies4,13 have examined the difference between pharmacologic and combined treatments, these studies suffer from several limitations. Both reviews included only a limited number of currently available studies, included studies in which no strict diagnostic criteria were established, and did not conduct extended subgroup analyses to examine possible sources of heterogeneity. Each of the 2 meta-analyses included less than half of the studies we identified in our searches (the 2 meta-analyses included 10 and 12 studies of the 25 studies we included using the specific inclusion criteria specified below), and our literature searches resulted in 13 studies that were not included in either of the 2 earlier meta-analyses. Furthermore, these earlier meta-analyses did not examine possible sources of heterogeneity. For example, in an earlier meta-analysis,12 we found that pharmacotherapy was significantly more effective than psychotherapy in the treatment of dysthymia and that selective serotonin reuptake inhibitors (SSRIs) were more effective than psychotherapy. Subgroup analyses of this kind were not conducted in the 2 earlier meta-analyses. These 2 earlier meta-analyses did, however, find indications that combined treatment is more effective than pharmacotherapy alone.
We decided to conduct a new comprehensive meta-
analysis of studies in which pharmacotherapy was compared to the combination of pharmacotherapy and psychotherapy. Our hypothesis was that we would confirm that combined treatment is more effective than pharmacotherapy alone. We also wanted to explore whether study characteristics were related to the relative effects of pharmacologic and combined treatment.
Identification and Selection of Studies
First, we used a database of 832 studies on the
psychological treatment of depression in general. This database has been described in detail elsewhere14 and has been used in a series of earlier meta-analyses (www.
evidencebasedpsychotherapies.org). It was developed through a comprehensive literature search (articles published from 1966 to January 2008) in which we examined 6,947 abstracts in PubMed (1,244 abstracts), PsycINFO (1,736), EMBASE (1,911), and the Cochrane Central Register of Controlled Trials (2,056). These abstracts were identified by combining terms indicative of psychological treatment and depression (both MeSH terms and text words). For this database, we also collected the primary studies from 42 meta-analyses on psychological treatment for depression (www.evidencebasedpsychotherapies.org). For the current study, we examined the abstracts of these 832 studies.
We included (1) randomized trials (2) in which the effects of a pharmacologic treatment (3) were compared to the effects of a combined pharmacologic and psychological treatment (4) in adults (5) with a depressive disorder. No language restrictions were applied. Only studies in which the subjects met diagnostic criteria for a depressive disorder (major depression, dysthymia) were included. Studies aimed at subjects with elevated levels of depressive symptoms (as measured by self-report measures) but no indication of diagnosis were excluded, as were studies on inpatients, studies on adolescents or children (below 18 years of age), and studies aimed at relapse prevention or maintenance treatments. Comorbid general medical or psychiatric disorders were not used as an exclusion criterion.
We assessed the validity of included studies by using a number of basic criteria, as suggested in the Cochrane Handbook15: allocation to conditions conducted by an independent (third) party, blinding of assessors to outcomes, and completeness of follow-up data. We did not use the fourth criterion for validity (adequacy of random allocation concealment to respondents) because it was not possible in these studies to conceal the randomization to patients.
For each comparison between pharmacologic and combined treatments, we calculated the effect size (Cohen’s d) indicating the difference between the 2 types of treatment at posttest. We calculated the effect sizes by subtracting (at posttest) the mean score of the combined treatment group from the mean score of the pharmacotherapy group, and dividing the result by the pooled standard deviations of the 2 groups. Effect sizes of 0.8 can be assumed to be large, while effect sizes of 0.5 are moderate, and effect sizes of 0.2 are small.16 When psychological treatments are compared to control groups, effect sizes of 0.6 or larger are usually found.2,17 In our meta-analysis, effect sizes of zero were assumed to indicate that there was no difference between the effects of pharmacotherapy and those of the combined treatment.
In the calculations of effect sizes, we used only those instruments that explicitly measured symptoms of depression. If more than 1 depression measure was used, the mean of the effect sizes was calculated, so that each study (or contrast group) provided only 1 effect size. If means and standard deviations were not reported, we used other statistics that were reported about the test between the 2 conditions at posttest (P or t value). If these were not reported, we used the procedures of the software program Comprehensive Meta-Analysis (version 2.2.021; Biostat, Englewood, New Jersey) to calculate the effect size by using dichotomous outcomes.
We used only the effect sizes indicating the differences between pharmacologic and combined treatments at posttest. We decided not to examine the differential effects at follow-up because the number of effect sizes was relatively low. In addition, the follow-up period differed considerably among these studies, and, in several studies, treatments were continued and others discontinued.
To calculate pooled mean effect sizes, we used the computer program Comprehensive Meta-Analysis (version 2.2.021). As we expected considerable heterogeneity among the studies, we decided to calculate mean effect sizes with the random-effects model. In the random-effects model, it is assumed that the included studies are drawn from populations of studies that differ from each other systematically (heterogeneity). In this model, the effect sizes resulting from included studies differ not only because of the random
error within studies (as in the fixed-effects model) but also because of true variation in effect size from one study to the next.
To indicate homogeneity, we calculated the I2 statistic, which is an indicator of heterogeneity in percentages. A value of 0% indicates no observed heterogeneity, and larger values show increasing heterogeneity, with 25% as low, 50% as moderate, and 75% as high heterogeneity.18 We also calculated the Q statistic but only report whether or not this was statistically significant.
Subgroup analyses were conducted according to the mixed-effects model. In this model, studies within subgroups are pooled with the random-effects model, while tests for statistically significant differences between subgroups are conducted with the fixed-effects model. For continuous variables, we used metaregression analyses to test whether there is a statistically significant relationship between the continuous variable and the effect size, as indicated by a
Z value and an associated P value.
Publication bias was tested by inspecting the funnel plot on primary outcome measures and by using Duval and Tweedie’s trim-and-fill procedure,19 which yields an estimate of the effect size after the publication bias has been taken into account (as implemented in Comprehensive Meta-Analysis, version 2.2.021).
Characteristics of the Included Studies
All inclusion criteria were met by 25 studies,6–11, 20–38 with a total of 2,036 patients (1,018 in the pharmacotherapy conditions and another 1,018 in the combined treatments). The mean number of patients per study was 81 (ranging from 20 to 453), with 15 studies having fewer than 50 patients, 5 having 50 to 100 patients, and 5 having 100 or more patients. Selected characteristics of the included studies are presented in Table 1.
Click figure to enlarge
Sixteen studies were aimed at adults in general, while 9 were aimed at more specific target groups (3 on older adults and 1 each on older adults who lost their spouse, adult women, patients with comorbid borderline personality disorder, patients with chronic depression, patients with coronary artery disease, and women with postpartum depression). Patients were recruited through clinical referrals (18 studies), from the community (5 studies), or through a combination of both (2 studies). Fifteen studies were aimed at patients with a major depressive disorder and 5 at patients with dysthymia, while the remaining 5 studies were aimed at patients with other definitions of depressive disorders (major depression and/or dysthymia, other). In 17 studies, the Hamilton Depression Rating Scale (HDRS) score at pretest was presented (range, 16.7–27.4), while 10 studies reported the pretest Beck Depression Inventory (BDI) score (range, 13.7–36.9).
Eight studies examined cognitive-behavioral therapy, another 8 examined interpersonal psychotherapy, and 9 examined other psychological treatments, such as psychodynamic therapy or problem-solving treatment. Individual psychotherapies were examined in 21 studies, while 4 studies used group therapies (1 used a combined individual and group format). Selective serotonin reuptake inhibitors (SSRIs) were examined in 9 studies, and tricyclic antidepressants (TCAs) were also examined in 9 studies (other medications or a protocol was used in 7 studies).
The quality of the included studies varied. Eight of the 25 studies reported that allocation to conditions was conducted by an independent party. Blinding of assessors was reported in 18 studies. Intention-to-treat analyses were conducted in 16 studies (the other studies were limited to completers-only analyses). Five studies met all 3 quality criteria.
Differences Between Pharmacologic
and Combined Treatments: Overall Effect Sizes
The mean effect size indicating the difference between pharmacotherapy and combined therapy was d = 0.31 (95% CI, 0.20 ~ 0.43; Table 2) for the 25 included studies, indicating a small effect in favor of the combined treatment, which was highly significant (Z = 5.17, P < .001). This effect size corresponds with a number needed to treat (NNT) of 5.75. Heterogeneity was low to moderate (Q = 34.33, P < .1; I2 = 30.08). The effect sizes and 95% confidence intervals of the individual contrast groups are plotted in Figure 1.
Click figure to enlarge
When we limited the analyses to the effect sizes with the HDRS, comparable results were found (d = 0.32; 95% CI, 0.19 ~ 0.44; P < .001; Q = 27.16, not significant [NS]; I2 = 26.37; NNT = 5.56). The same was true when we limited the analyses to the effect sizes found with the BDI (d = 0.28; 95% CI, 0.10 ~ 0.46; P < .01; Q = 5.32, NS; I2 = 0; NNT = 6.41).
Click figure to enlarge
Visual inspection of the funnel plot suggested that the study by Macaskill and Macaskill7 could be an outlier. The resulting effect size was comparable with the effect size of all comparisons (d = 0.30; 95% CI, 0.19 ~ 0.42; Z = 5.22, P < .001) and heterogeneity was somewhat lower (Q = 30.29, NS; I2 = 24.07).
Several of the included studies differed on essential
characteristics from the other studies. For example, 1 study7 examined rational-emotive therapy, another30 examined dialectic behavior therapy, and yet another study36 did not report any details about the pharmacotherapy used. It is possible that 1 or more of such atypical studies had a negative effect on the overall mean effect size. Therefore, we conducted a series of meta-analyses in which we removed the study with the largest impact on the overall effect size. We first removed the study with the largest impact and examined the extent to which the effect size was increased or decreased and repeated this procedure several times. In this way we could examine whether removal of 1 or more studies resulted in important changes to the outcomes.
Removal of the study by Browne and colleagues10 resulted in the largest increase of the effect size (the resulting
effect size was d = 0.37; I2 = 3.53). After the removal of this study, we repeated this procedure and examined which study should be removed in order to realize the next largest increase of the effect size. This study was the one by Bellack and colleagues,9 and the meta-analysis resulted in an effect size of d = 0.39 (I2 = 0). Repeating this procedure a third time (which removed the study by Lesperance and colleagues28) resulted in a mean effect size of d = 0.41 (I2 = 0). A comparable procedure to examine whether individual studies resulted in a decrease of the effect size indicated that removal of the study by Keller and colleagues27 resulted in the largest
decrease (effect size: d = 0.29; I2 = 23.00), followed by the study by de Jonghe22 (effect size: d = 0.26; I2 = 15.52) and the study by Sirey et al36 (effect size: d = 0.22; I2 = 4.04). These analyses suggest that removal of studies did not result in major changes in the effect sizes.
Neither the funnel plot nor Duval and Tweedie’s trim-and-fill procedure pointed to a significant publication bias. The effect size indicating the difference in reduction of depressive symptomatology between combined (including psychological) and pharmacologic treatments did not change significantly after adjustment for possible publication bias (the observed and adjusted effect size did not differ from each other).
Because there was some heterogeneity, we decided to conduct a series of subgroup analyses. The results of these are presented in Table 2. As can be seen, we found no indication of a significant difference between studies aimed at adults in general and studies aimed at more specific target groups, nor did we find a significant association between effect size and type of psychotherapy, between studies
in which individual therapies were used and those in which group therapies were used, between studies in which
intention-to-treat analyses were conducted and studies in which completers-only analyses were conducted, or between studies that met all quality criteria and those that did not. We also grouped the studies into those that did or did not include a separate psychotherapy alone condition (although these psychotherapy-only conditions are not examined in this meta-analysis). However, this subgroup analysis did not indicate that the effect sizes in these groups differed significantly from each other.
However, we did find that studies aimed at patients with dysthymia resulted in significantly lower effect sizes (d = 0.00) compared to studies aimed at patients with major depressive disorder (d = 0.40). We also found that studies in which SSRIs were used resulted in significantly lower effect sizes than studies in which TCAs or other pharmacotherapies were used (P = .004). The studies in which SSRIs were used did not indicate that combined treatment was more effective than treatment with pharmacotherapy alone (d = 0.10). Furthermore, studies in which patients were recruited from clinical samples resulted in higher effect sizes than studies in which patients were recruited in other ways (P < .001).
Because we found that studies in patients with dysthymia resulted in smaller effect sizes than studies in patients with major depression, we conducted some additional subgroup analyses. In these analyses, we removed the studies with dysthymic patients and included only the studies of patients with major depressive disorder. Then we examined whether we still found a difference between studies with TCAs and studies with SSRIs. As can be seen in Table 2, in these analyses, we found that psychotherapy had an additional effect for both TCAs and SSRIs. The difference between the additional effect of psychotherapy in TCAs and in SSRIs was very small and no longer significant. We also examined whether studies in which patients were recruited from clinical samples still resulted in higher effect sizes than studies in which patients were recruited in other ways, but this difference was also no longer significant.
Severity of Depression
We examined in several ways whether the severity of depression at baseline was associated with the effect sizes. First, we conducted a series of metaregression analyses. We selected the 15 studies in which the sample’s mean HDRS (17-item version) score at baseline was presented and examined with a metaregression analysis whether the baseline HDRS was associated with the effect size. These analyses did not indicate that the baseline HDRS score was significantly associated with the effect size (point estimate of slope = −0.04; 95% CI, –0.10 ~ 0.01; Z = −1.48, P = .14). Because our subgroup analyses indicated a difference between studies on dysthymia, we repeated these analyses after removing the studies on dysthymia. Again, no significant association between effect size and baseline HDRS was found. Then we divided the 15 studies that presented the baseline HDRS score into studies in which the HDRS score was 20 or lower and those in which it was higher than 20 (Table 2). A subgroup analysis did not indicate that the studies with lower baseline depression differed significantly from those with more severe depression.
We repeated this procedure on the 10 studies in which the baseline BDI was reported. In a metaregression analysis, we found no indication that the pretest BDI score was significantly associated with the effect size (point estimate of slope = 0.02; 95% CI, –0.01 ~ 0.05; Z = 1.40, P = .16). A subgroup analysis in which we divided the studies into those with mild to moderate depression (BDI score ≤ 18) or moderate to severe depression (BDI score 19–29) also did not find any indications for a significant difference between these groups (Table 2).
Effects on Dropout
We compared the dropout rates of pharmacologic and combined treatments in 19 studies and found that the dropout rate was significantly lower in the combined treatment compared to pharmacotherapy alone (OR = 0.65; 95% CI, 0.50 ~ 0.83; Z = −3.44, P < .01), with almost no heterogeneity (Q = 18.11, NS; I2 = 0.62). We found no indications of significant publication bias when we examined the funnel plot and used Duvall and Tweedie’s19 trim-and-fill method. The adjusted OR was almost the same as the unadjusted OR (adjusted OR: 0.62; 95% CI, 0.49 ~ 0.80).
We conducted the same subgroup and metaregression analyses as we did with the effect sizes (Table 3). As can be seen in Table 3, none of the subgroup analyses indicated that there were significant differences between subgroups. There was a trend (P < .1), however, for the dropout rate to be lower in studies examining adults in general compared to studies in which more specific target groups were examined. Levels of heterogeneity were low or zero in most subgroups.
Click figure to enlarge
We found no indication that severity of depression was associated with dropout in both the subgroup analyses (Table 3) and the metaregression analyses (HDRS: N = 14, point estimate of slope = –0.05, 95% CI, –0.13 ~ 0.22; Z = 0.52, NS; BDI: N = 7, point estimate of slope = –0.01, 95% CI, –0.10 ~ 0.07; Z = –0.28, NS).
We also examined whether adverse effects of medication differed in the pharmacotherapy and combined treatments. However, only 7 studies reported these data, and, because in most of the studies the number of adverse effects was very small (even zero in many cases), we did not find it informative to pool these outcomes into 1 effect size.
We found clear indications that a combined treatment including psychotherapy is more effective than pharmacotherapy alone. Although the effect size indicating the difference between pharmacotherapy and the combined therapy was small, it was highly statistically significant. This suggests that psychotherapy has an additional effect on depression apart from the effects of pharmacotherapy.
However, we also found that in studies aimed at patients with dysthymia, the combined treatment had no additional value compared to pharmacotherapy alone. Our subgroup analyses showed that the studies with dysthymia patients differed significantly from those with major depression
patients. This is in agreement with the results of an earlier meta-analysis of studies in which psychotherapy and pharmacotherapy were directly compared to each other.12 In the earlier meta-analysis, we found that psychotherapy alone was significantly less effective than pharmacotherapy alone in patients with dysthymia. It could be that current psychotherapies do not target a chronic disorder like dysthymia rapidly enough to keep up with pharmacotherapy. In that regard, it is of interest to note that one of the largest effect sizes found in the combinatorial literature was found when an approach developed specifically for use with chronic patients was applied to a sample with MDD.27
We also found that psychotherapy added less to pharmacotherapy in studies in which SSRIs were examined, compared to studies in which TCAs were used. However, this difference was not statistically significant in the sample studies examining patients with major depression, and is strongly influenced by the fact that most studies on dysthymic patients also examined SSRIs.
In addition, we found that the dropout rate was significantly lower in the combined treatment group compared to pharmacotherapy alone group. This finding is in agreement with our earlier research in which we found that dropout rates in psychotherapy are significantly lower than dropout rates in pharmacotherapy12 and is in agreement with several other studies39–41 that have shown that most patients prefer psychotherapy instead of pharmacotherapy.
We found no indication for an association between the effect size and severity of depression, which suggests that combined treatments are superior in both milder and more severe cases of depression. Because most studies examined patients with mild to moderate depression, this finding has to be interpreted with caution.
This study has several important limitations. First, the number of studies we could include was still relatively small, and the quality of several of the included studies was not optimal. Second, it was not possible in any of the included studies to blind patients with respect to treatment assignment. Patients know when they are assigned to the combined treatment because they receive psychotherapy. This may have distorted the results of these studies. Third, only a limited number of the studies conducted intention-to-treat analyses, and, because dropout rates differed significantly between pharmacotherapy and combined treatment, this may have introduced a bias in our outcomes. Fourth, we did not examine the long-term effects of both treatments. It may well be possible that there are major differences between pharmacotherapy and combined treatments in the longer term. Because of these limitations, the results of this meta-analysis should be considered with caution.
From a clinical point of view, our results raise a number of questions and possibilities for future research. First, one interpretation of the results could be that certain patients only respond when given the additional push for change in psychotherapy. On the other hand, we do not know how the patients in the trials perceived the 2 treatments. Were both regarded as equally important? These questions could be addressed in more qualitative research. Second, how should clinicians handle the clinical situation with 2 active treatments, often not provided by the same clinician? It is known that follow-up visits in trials have a therapeutic effect,42 and, potentially, this can cancel out some to the nonspecific
effects of psychotherapy.
More research is also needed to further examine the
effectiveness of psychotherapy in dysthymia, as well as the mechanisms through which psychological and pharmacologic treatment work, and characteristics of patients who respond better to pharmacologic or combined treatment. We can conclude that psychotherapy compared to pharmacotherapy alone seems to have an additional value in the treatment of depression.
Drug names: amitriptyline (Limbitrol and others), citalopram
(Celexa and others), clomipramine (Anafranil and others), desipramine (Norpramin and others), fluoxetine (Prozac, Sarafem, and others), fluvoxamine (Luvox and others), imipramine (Tofranil and others),
nortriptyline (Pamelor, Aventyl, and others), paroxetine (Paxil, Pexeva, and others), sertraline (Zoloft and others).
Author affiliations: Department of Clinical Psychology and EMGO Institute, VU University Amsterdam (Drs Cuijpers and Dekker);
Arkin Institute for Mental Health Care, Amsterdam (Dr Dekker),
The Netherlands; Department of Psychology, Vanderbilt University,
Nashville, Tennessee (Dr Hollon); Department of Behavioural Sciences and Learning, Swedish Institute for Disability Research, Linköping
University and Department of Clinical Neuroscience, Psychiatry
Section, Karolinska Institutet, Stockholm (Dr Andersson), Sweden.
Financial disclosure: None reported.
Funding/support: None reported.
1. Anderson IM. Meta-analytic studies on new antidepressants. Br Med Bull. 2001;57:161–178. PubMed doi:10.1093/bmb/57.1.161
2. Cuijpers P, van Straten A, Warmerdam L, et al. Characteristics of effective psychological treatments of depression; a meta-regression analysis. Psychother Res. 2008;18(2):225–236. PubMed doi:10.1080/10503300701442027
3. Cuijpers P, van Straten A, Warmerdam L, et al. Psychotherapy versus the combination of psychotherapy and pharmacotherapy in the treatment of depression: a meta-analysis. Depress Anxiety. 2009;26(3):279–288. PubMed doi:10.1002/da.20519
4. Friedman MA, Detweiler-Bedell JB, Leventhal HE, et al. Combined psychotherapy and pharmacotherapy for the treatment of major depressive disorder. Clin Psychol Sci Pract. 2004;11:47–68. doi:10.1093/clipsy/bph052
5. De Maat SM, Dekker J, Schoevers RA, et al. Relative efficacy of psychotherapy and combined therapy in the treatment of depression: a meta-analysis. Eur Psychiatry. 2007;22(1):1–8. PubMed doi:10.1016/j.eurpsy.2006.10.008
6. Bellino S, Zizza M, Camilla R, et al. Combined treatment of major depression in patients with borderline personality disorder: a comparison with pharmacotherapy. Can J Psychiatry. 2006;51(7):453–460. PubMed
7. Macaskill ND, Macaskill A. Rational-emotive therapy plus pharmacotherapy versus pharmacotherapy alone in the treatment of high cognitive dysfunction depression. Cognit Ther Res. 1996;20(6):575–592. doi:10.1007/BF02227962
8. Blackburn IM, Bishop S, Glen AI, et al. The efficacy of cognitive therapy in depression: a treatment trial using cognitive therapy and pharmacotherapy, each alone and in combination. Br J Psychiatry. 1981;139:181–189. PubMed doi:10.1192/bjp.139.3.181
9. Bellack AS, Hersen M, Himmelhoch J. Social skills training compared with pharmacotherapy and psychotherapy in the treatment of unipolar depression. Am J Psychiatry. 1981;138(12):1562–1567. PubMed
10. Browne G, Steiner M, Roberts J, et al. Sertraline and/or interpersonal psychotherapy for patients with dysthymic disorder in primary care: 6-month comparison with longitudinal 2-year follow-up of effectiveness and costs. J Affect Disord. 2002;68(2–3):317–330. PubMed doi:10.1016/S0165-0327(01)00343-3
11. Markowitz JC, Kocsis JH, Bleiberg KL, et al. A comparative trial of psychotherapy and pharmacotherapy for “pure” dysthymic patients. J Affect Disord. 2005;89(1–3):167–175. PubMed doi:10.1016/j.jad.2005.10.001
12. Cuijpers P, van Straten A, van Oppen P, et al. Are psychological and pharmacological interventions equally effective in the treatment of adult depressive disorders? A meta-analysis of comparative studies. J Clin Psychiatry. 2008;69(11):1675–1685. PubMed doi:10.4088/JCP.v69n1102
13. Pampallona S, Bollini P, Tibaldi G, et al. Combined pharmacotherapy and psychological treatment for depression: a systematic review. Arch Gen Psychiatry. 2004;61(7):714–719. PubMed doi:10.1001/archpsyc.61.7.714
14. Cuijpers P, van Straten A, Warmerdam L, et al. Psychological treatment of depression: a meta-analytic database of randomized studies. BMC Psychiatry. 2008;8:36. PubMed doi:10.1186/1471-244X-8-36
15. Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of Interventions 4.2.5. In: The Cochrane Library, Issue 3. Chichester, UK: John Wiley & Sons, Ltd; 2005.
16. Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Erlbaum; 1988.
17. Churchill R, Hunot V, Corney R, et al. A systematic review of controlled trials of the effectiveness and cost-effectiveness of brief psychological treatments for depression. Health Technol Assess. 2001;5(35):1–173. PubMed
18. Higgins JP, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ. 2003;327(7414):557–560. PubMed doi:10.1136/bmj.327.7414.557
19. Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics. 2000;56(2):455–463. PubMed doi:10.1111/j.0006-341X.2000.00455.x
20. Blom MBJ, Jonker K, Dusseldorp E, et al. Combination treatment for acute depression is superior only when psychotherapy is added to medication. Psychother Psychosom. 2007;76(5):289–297. PubMed doi:10.1159/000104705
21. Burnand Y, Andreoli A, Kolatte E, et al. Psychodynamic psychotherapy and clomipramine in the treatment of major depression. Psychiatr Serv. 2002;53(5):585–590. PubMed doi:10.1176/appi.ps.53.5.585
22. de Jonghe F, Kool S, van Aalst G, et al. Combining psychotherapy and antidepressants in the treatment of depression. J Affect Disord. 2001;64(2–3):217–229. PubMed doi:10.1016/S0165-0327(00)00259-7
23. De Mello MF, Myczcowisk LM, Menezes PR. A randomized controlled trial comparing moclobemide and moclobemide plus interpersonal psychotherapy in the treatment of dysthymic disorder. J Psychother Pract Res. 2001;10(2):117–123. PubMed
24. Hautzinger M, de Jong-Meyer R, Treiber R, et al. Wirksamkeit Kognitiever Verhaltenstherapie, Pharmakotherapie und deren Kombination bei nicht-endogenen, unipolaren Depressionen. Zeitschr Klin Psychol. 1996;25:130–145.
25. Hellerstein DJ, Little SAS, Samstag LW, et al. Adding group psychotherapy to medication treatment in dysthymia: a randomized prospective pilot study. J Psychother Pract Res. 2001;10(2):93–103. PubMed
26. Hollon SD, DeRubeis RJ, Evans MD, et al. Cognitive therapy and pharmacotherapy for depression: singly and in combination. Arch Gen Psychiatry. 1992;49(10):774–781. PubMed
27. Keller MB, McCullough JP, Klein DN, et al. A comparison of nefazodone, the cognitive behavioral-analysis system of psychotherapy, and their combination for the treatment of chronic depression. N Engl J Med. 2000;342(20):1462–1470. PubMed doi:10.1056/NEJM200005183422001
28. Lespérance F, Frasure-Smith N, Koszycki D, et al. CREATE Investigators. Effects of citalopram and interpersonal psychotherapy on depression in patients with coronary artery disease: the Canadian Cardiac Randomized Evaluation of Antidepressant and Psychotherapy Efficacy (CREATE) trial. JAMA. 2007;297(4):367–379. PubMed doi:10.1001/jama.297.4.367
29. Lopez Rodriguez J, Lopez Butron MA, Vargas Terrez BE, et al. Estudio doble ciego con antidepresivo, psicoterapia breve y placebo en pacientes con depresion leve a moderada. (Double blind study with antidepressant, brief psychotherapy and placebo in patients with mild to moderate depression) Salud Ment. 2004;27:53–61.
30. Lynch TR, Morse JQ, Mendelson T, et al. Dialectical behavior therapy for depressed older adults: a randomized pilot study. Am J Geriatr Psychiatry. 2003;11(1):33–45. PubMed
31. Misri S, Reebye P, Corral M, et al. The use of paroxetine and cognitive-behavioral therapy in postpartum depression and anxiety: a randomized controlled trial. J Clin Psychiatry. 2004;65(9):1236–1241. PubMed
32. Murphy GE, Simons AD, Wetzel RD, et al. Cognitive therapy and pharmacotherapy: singly and together in the treatment of depression. Arch Gen Psychiatry. 1984;41(1):33–41. PubMed
33. Mynors-Wallis LM, Gath DH, Day A, et al. Randomised controlled trial of problem solving treatment, antidepressant medication, and combined treatment for major depression in primary care. BMJ. 2000;320(7226):26–30. PubMed doi:10.1136/bmj.320.7226.26
34. Ravindran AV, Anisman H, Merali Z, et al. Treatment of primary dysthymia with group cognitive therapy and pharmacotherapy: clinical symptoms and functional impairments. Am J Psychiatry. 1999;156(10):1608–1617. PubMed
35. Reynolds CF, Miller MD, Pasternak RE, et al. Treatment of bereavement-related major depressive episodes in later life: a controlled study of acute and continuation treatment with nortriptyline and interpersonal psychotherapy. Am J Psychiatry. 1999;156(2):202–208. PubMed
36. Sirey JA, Bruce ML, Alexopoulos GS. The Treatment Initiation Program: an intervention to improve depression outcomes in older adults. Am J Psychiatry. 2005;162(1):184–186. PubMed doi:10.1176/appi.ajp.162.1.184
37. Thompson LW, Coon DW, Gallagher-Thompson D, et al. Comparison of desipramine and cognitive/behavioral therapy in the treatment of elderly outpatients with mild-to-moderate depression. Am J Geriatr Psychiatry. 2001;9(3):225–240. PubMed
38. Weissman MM, Prusoff BA, Dimascio A, et al. The efficacy of drugs and psychotherapy in the treatment of acute depressive episodes. Am J Psychiatry. 1979;136(14B):555–558. PubMed
39. Dekker JJM, Koelen JA, Van HL, et al. Speed of action: the relative efficacy of short psychodynamic supportive psychotherapy and pharmacotherapy in the first 8 weeks of a treatment algorithm for depression. J Affect Disord. 2008;109 (1–2):183–188. PubMed doi:10.1016/j.jad.2007.10.015
40. Bedi N, Chilvers C, Churchill R, et al. Assessing effectiveness of treatment of depression in primary care; partially randomised preference trial. Br J Psychiatry. 2000;177:312–318. PubMed doi:10.1192/bjp.177.4.312
41. Van Schaik DJ, Klijn AF, van Hout HP, et al. Patients’ preferences in the treatment of depressive disorder in primary care. Gen Hosp Psychiatry. 2004;26(3):184–189. PubMed doi:10.1016/j.genhosppsych.2003.12.001
42. Posternak MA, Zimmerman M. Therapeutic effect of follow-up assessments on antidepressant and placebo response rates in antidepressant efficacy trials: meta-analysis. Br J Psychiatry. 2007;190:287–292. PubMed doi:10.1192/bjp.bp.106.028555