Abstract
Psychotherapies are first‐line treatments for most mental disorders, but their absolute outcomes (i.e., response and remission rates) are not well studied, despite the relevance of such information for health care users, providers and policy makers. We aimed to examine absolute and relative outcomes of psychotherapies across eight mental disorders: major depressive disorder (MDD), social anxiety disorder, panic disorder, generalized anxiety disorder (GAD), specific phobia, post‐traumatic stress disorder (PTSD), obsessive‐compulsive disorder (OCD), and borderline personality disorder (BPD). We used a series of living systematic reviews included in the Metapsy initiative (www.metapsy.org), with a common strategy for literature search, inclusion of studies and extraction of data, and a common format for the analyses. Literature search was conducted in major bibliographical databases (PubMed, PsycINFO, Embase, and the Cochrane Register of Controlled Trials) up to January 1, 2023. We included randomized controlled trials comparing psychotherapies for any of the eight mental disorders, established by a diagnostic interview, with a control group (waitlist, care‐as‐usual, or pill placebo). We conducted random‐effects model pairwise meta‐analyses. The main outcome was the absolute rate of response (at least 50% symptom reduction between baseline and post‐test) in the treatment and control conditions. Secondary outcomes included the relative risk (RR) of response, and the number needed to treat (NNT). Random‐effects meta‐analyses of the included 441 trials (33,881 patients) indicated modest response rates for psychotherapies: 0.42 (95% CI: 0.39‐0.45) for MDD; 0.38 (95% CI: 0.33‐0.43) for PTSD; 0.38 (95% CI: 0.30‐0.47) for OCD; 0.38 (95% CI: 0.33‐0.43) for panic disorder; 0.36 (95% CI: 0.30‐0.42) for GAD; 0.32 (95% CI: 0.29‐0.37) for social anxiety disorder; 0.32 (95% CI: 0.23‐0.42) for specific phobia; and 0.24 (95% CI: 0.15‐0.36) for BPD. Most sensitivity analyses broadly supported these findings. The RRs were significant for all disorders, except BPD. Our conclusion is that most psychotherapies for the eight mental disorders are effective compared with control conditions, but absolute response rates are modest. More effective treatments and interventions for those not responding to a first‐line treatment are needed.
Keywords: Psychotherapies, response rates, depression, social anxiety disorder, panic disorder, generalized anxiety disorder, specific phobia, post‐traumatic stress disorder, obsessive‐compulsive disorder, borderline personality disorder
About 970 million people worldwide suffer from a mental disorder 1 . Depression and anxiety disorders are the most prevalent conditions, but the prevalence rates of other disorders – such as post‐traumatic stress disorder (PTSD), obsessive‐compulsive disorder (OCD) and borderline personality disorder (BPD) – are also remarkably high. For these disorders, evidence‐based treatments are available, with psychotherapies being first‐line interventions 2 .
Treatment outcomes of psychotherapies are usually examined by continuous measures (i.e., mean scores on rating scales). Meta‐analyses typically report outcomes as standardized mean differences (SMDs), which indicate the difference between an intervention and a control condition in terms of standard deviation units. Although this is a correct way of reporting outcomes, these effect sizes are difficult to interpret for patients, providers and policy makers 3 , 4 .
Binary outcomes such as response (defined, in the case of depression, as a 50% symptom reduction between baseline and post‐test) and remission (i.e., a state in which the individual has no more than minimal symptoms) are easier to interpret 5 . However, trials and meta‐analyses typically report these outcomes by comparing the treatment with a control condition in terms of relative risks (RRs), odds ratios (ORs), or number needed to treat (NNT), i.e., how many patients need to be treated to have one more positive outcome compared to controls 6 .
Absolute outcomes, in terms of response or remission rates of treatments, are typically not pooled in meta‐analyses, because this would assume that these rates are comparable across trials, which may not be the case 7 . Such absolute outcomes are, however, essential for clinical decision‐making. In other areas, for example in meta‐analyses of prevalence rates of mental disorders, absolute rates are often pooled 8 , 9 , 10 .
Hundreds of randomized controlled trials have examined the effects of psychotherapies, and meta‐analyses have shown that these therapies are effective across a broad range of mental disorders, including major depressive disorder (MDD) 11 , 12 , PTSD 13 , 14 , OCD 15 , BPD 16 , 17 , and anxiety disorders 18 , 19 , 20 , 21 , 22 . Very few meta‐analyses, however, have focused on more than one mental disorder, while a broader focus on multiple disorders can allow an assessment of the comparative “treatability” of these conditions.
This is also important because psychotherapies are compared with different types of control conditions, which can lead to differential effect sizes. For example, in the field of depression, it is well established that the use of waitlist as the control condition may overestimate the effects of treatments 23 . In anxiety disorders, almost all trials have used waitlist, and there are indications that care‐as‐usual control groups result in considerably smaller effect sizes 22 . In BPD, virtually all trials make use of care‐as‐usual control groups 16 , 17 . One advantage of examining absolute outcomes is that this is done separately for the treatment and control groups, so that a comparison between the relative “treatability” of mental disorders is not complicated by differences in control conditions.
In this meta‐analytic study, we examined the absolute and relative outcomes of psychotherapies for eight major mental disorders: MDD, PTSD, OCD, BPD, panic disorder, generalized anxiety disorder (GAD), social anxiety disorder, and specific phobia. As primary outcome, we considered the absolute measure of response rate 24 . In the field of treatments for depression, response is often defined as a 50% symptom reduction 5 . In other fields, the definition of response is less clear. For example, in studies on anxiety disorders, a broad range of different definitions is used 25 , and the same is true for PTSD 26 . In this meta‐analysis, we used the same definition of response (number of patients with at least 50% symptom reduction between baseline and post‐test, divided by the total number of patients) across all included mental disorders. This allows to have a comparable outcome across these different disorders, providing an indication of their relative “treatability”. To the best of our knowledge, no previous meta‐analysis has reported absolute outcomes across psychotherapies for a range of mental disorders.
METHODS
Search strategy and selection criteria
This meta‐analysis was registered at the Open Science Framework (https://doi.org/10.17605/osf.io/rpw6g). We used a series of living systematic reviews included in the Metapsy project (www.metapsy.org), with a common strategy for search of literature, inclusion of studies and extraction of data, and a common format for the analyses. All included datasets are updated at least once per year (deadline for the current study: January 1, 2023).
Major bibliographical databases – including PubMed, PsycINFO, Embase, and the Cochrane Register of Controlled Trials – were searched by combining terms (index and text words) indicative of each of the mental disorders and psychotherapies, with filters for randomized controlled trials. Separate searches were conducted for MDD, PTSD, OCD and BPD. A combined search was performed for four anxiety disorders: panic disorder, GAD, social anxiety disorder, and specific phobia. Full search strings are presented in the supplementary information.
All identified records were screened by two researchers, and all records that could possibly meet inclusion criteria according to one of the researchers were retrieved in full text. The decision to include a study was also made by two independent researchers, and disagreements were resolved through consensus and, if needed, consultation with a third rater.
For the current meta‐analysis, we included randomized controlled trials in which a psychotherapy 27 for adults with any of the eight mental disorders, as established by a diagnostic interview, was compared with a control group (waitlist, care‐as‐usual, or pill placebo). We excluded studies with other control conditions, those including participants based on self‐report symptom measures; those on inpatients, children or adolescents; those examining unguided self‐help interventions, and those with insufficient data to calculate the response rate.
Quality assessment and data extraction
The quality of the included trials was evaluated using the Cochrane Risk of Bias (RoB) assessment tool version 1 (for the dataset on MDD) 28 or version 2 (for the other seven datasets) 29 . We used four criteria of the RoB tool 1: sequence generation, concealment of allocation to conditions, prevention of knowledge of the allocated intervention, and dealing with incomplete outcome data. RoB tool 2 assessed the same domains in more detail, as well as whether analyses were pre‐specified. We separated the trials in each of the eight datasets into two categories: low RoB and other. The definition of low RoB varied across the datasets (definitions are provided in Table 1). All assessments of RoB were conducted by two independent researchers, and disagreements were resolved through discussion with a third author.
Table 1.
Overview of literature search and characteristics of included studies and interventions
| MDD | PAN | SAD | GAD | PHOB | PTSD | OCD | BPD | Total | |
|---|---|---|---|---|---|---|---|---|---|
| Literature search | |||||||||
| Identified records, n | 33,181 | For all anxiety disorders: 19,535 | 36,108 | 11,235 | 11,827 | 111,886 | |||
| Records after removal of duplicates, n | 23,243 | For all anxiety disorders: 13,328 | 14,369 | 7,752 | 5,867 | 64,559 | |||
| Full texts assessed, n | 3,987 | For all anxiety disorders: 1,378 | 2,138 | 573 | 228 | 8,304 | |||
| Studies and patients included | |||||||||
| Studies, n | 159 | 48 | 52 | 48 | 22 | 69 | 22 | 21 | 441 |
| Comparisons, n | 196 | 71 | 74 | 61 | 32 | 87 | 26 | 22 | 569 |
| Patients, N | 14,908 | 3,559 | 4,053 | 3,773 | 1,189 | 2,986 | 1,011 | 1,525 | 33,881 |
| Patients in therapy, N | 8,362 | 2,236 | 2,593 | 2,162 | 711 | 2,325 | 585 | 795 | 19,769 |
| Patients in control condition, N | 6,546 | 1,323 | 1,460 | 1,611 | 478 | 1,538 | 426 | 730 | 14,112 |
| Characteristics of included studies | |||||||||
| Clinical recruitment, n (%) | 63 (38.9) | 17 (35.4) | 4 (7.7) | 20 (41.7) | 1 (4.2) | 31 (44.9) | 7 (31.8) | 18 (85.7) | 160 (36.3) |
| Age, years (mean±SD) | 42.2±13.1 | 37.3±3.8 | 34.3±6.0 | 41.9±12.1 | 32.8±10.8 | 39.1±7.9 | 34.5±4.6 | 31.9±5.6 | 39.0±10.7 |
| Proportion of women (mean±SD) | 0.72±0.2 | 0.71±0.1 | 0.56±0.1 | 0.72±0.1 | 0.75±0.2 | 0.58±0.4 | 0.57±0.2 | 0.83±0.1 | 0.67±0.2 |
| Type of control, n (%) | |||||||||
| Waitlist | 68 (42.0) | 36 (75.0) | 44 (84.6) | 34 (70.8) | 23 (95.8) | 54 (79.3) | 16 (72.7) | 0 | 270 (61.2) |
| Care‐as‐usual | 89 (54.9) | 6 (12.5) | 4 (7.7) | 11 (22.9) | 1 (4.2) | 13 (18.8) | 4 (18.2) | 21 (100.0) | 149 (33.8) |
| Pill placebo | 5 (3.1) | 6 (12.5) | 4 (7.7) | 3 (6.2) | 0 | 2 (3.9) | 2 (9.1) | 0 | 22 (5.0) |
| Country, n (%) | |||||||||
| North America | 51 (31.5) | 24 (50.0) | 18 (34.6) | 19 (39.6) | 13 (54.2) | 37 (53.6) | 6 (27.3) | 7 (33.3) | 172 (39.0) |
| Europe | 66 (40.7) | 19 (39.6) | 21 (40.4) | 18 (37.5) | 9 (37.5) | 17 (24.6) | 7 (31.8) | 12 (57.1) | 166 (37.6) |
| Other | 45 (27.8) | 5 (10.4) | 13 (25.0) | 11 (22.9) | 2 (8.3) | 15 (21.7) | 9 (40.9) | 2 (9.5) | 103 (23.4) |
| Risk of bias (RoB) | |||||||||
| RoB tool | Version 1 | Version 2 | Version 2 | Version 2 | Version 2 | Version 2 | Version 2 | Version 2 | |
| Definition of low RoB | Low risk for four core items | Overall low RoB | Overall low RoB | Overall low RoB | Overall low RoB | Overall low RoB | Overall low RoB or some concerns | Overall low RoB or some concerns | |
| Low RoB, n (%) | 64 (39.5) | 3 (6.2) | 4 (7.7) | 8 (16.7) | 3 (12.5) | 6 (8.7) | 3 (13.6) | 2 (9.5) | 93 (21.1) |
| Characteristics of interventions | |||||||||
| Format, n (%) | |||||||||
| Individual | 80 (40.8) | 39 (54.9) | 29 (39.2) | 34 (55.7) | 23 (71.9) | 70 (80.5) | 12 (46.2) | 8 (36.4) | 295 (51.8) |
| Group | 60 (30.6) | 15 (21.2) | 25 (33.8) | 14 (23.0) | 0 | 8 (9.2) | 0 | 0 | 122 (21.4) |
| Guided self‐help | 34 (17.3) | 12 (16.9) | 17 (23.0) | 13 (21.3) | 1 (3.1) | 5 (5.7) | 4 (15.4) | 0 | 86 (15.1) |
| Other/mixed | 22 (11.2) | 5 (7.0) | 3 (4.1) | 0 | 8 (25.0) | 4 (4.6) | 10 (38.5) | 14 (63.4) | 66 (11.6) |
| Sessions, n (mean±SD) | 10.6±5.6 | 9.6±3.8 | 11.4±5.3 | 11.3±5.6 | 3.3±3.1 | 10.0±5.1 | 12.4±7.0 | 53.6±42.6 | 11.8±8.5 |
MDD – major depressive disorder, PAN – panic disorder, SAD – social anxiety disorder, GAD – generalized anxiety disorder, PHOB – specific phobia, PTSD – post‐traumatic stress disorder, OCD – obsessive‐compulsive disorder, BPD – borderline personality disorder
We extracted several other characteristics of the trials and interventions: mean age of participants, proportion of women, whether the participants were recruited exclusively from clinical samples, type of psychotherapy (see supplementary information for definitions), treatment format, number of sessions, type of control condition, and country where the study was conducted.
Estimation of response rates
Treatment response was defined as the number of patients with at least 50% symptom reduction between baseline and post‐test, divided by the total number of patients. For the calculation of the response rate, we used one outcome measure in each study. If multiple outcome measures were used, we selected the measure that was most frequently used.
Patients randomized but not included in the analyses of responders in the original reports were assumed to be non‐responders and included in the main analyses to abide by the intention‐to‐treat principle. This was considered the most conservative estimate. We conducted sensitivity analyses in which we only included study completers, or all dropouts were considered responders.
We calculated response rates using a validated method using the baseline and post‐test means and standard deviations of the continuous outcome measures, and the number of patients at post‐test 24 . This method of estimating response rates has been tested in five meta‐analyses of treatments of depression and anxiety 24 , 30 , resulting in correlations between 0.94 and 0.97 between the estimated and true response rates.
Meta‐analyses
We pooled the response rates using the metaprop function of the meta package in R (version 4.2.2), and ran all analyses in RStudio (2023.03.0+386). Binomial proportion data were first transformed to a logit scale, and then random‐effects meta‐analyses for the transformed data were conducted. Finally, the summary results were inverted to the raw binomial proportion scale, and the estimates and their 95% confidence intervals (CIs) were reported. Because we expected considerable heterogeneity, we employed a random‐effects pooling model in all analyses (restricted maximum likelihood method).
After meta‐analyzing the response rates, we pooled the RR of the intervention over the control condition across the included studies using the metabin function. The NNTs were calculated using the pooled RR and the response rate in the control group 31 .
We conducted five sensitivity analyses: two with the less conservative estimates of the response rates (i.e., only including study completers, or considering all dropouts as responders), one in which we excluded outliers (studies in which the 95% CI of the rates did not overlap with the 95% CI of the pooled rates), one including only studies with low RoB, and one adjusting for publication bias. As an indicator of heterogeneity, we calculated the I2 statistic and its 95% CI 32 . We examined potential publication bias using Egger's test of the intercept, and Duval and Tweedie trim‐and‐fill procedure.
We conducted a series of subgroup analyses, using mixed‐effects methods, considering type of psychotherapy, treatment format, whether participants were recruited exclusively from clinical samples, type of control condition, and country where the study was carried out. Where possible, we avoided subgroups with less than five studies, clustering them into larger subgroups.
RESULTS
Selection, inclusion and characteristics of included studies
The searches across all disorders resulted in 111,886 records (64,559 after removal of duplicates), 8,304 full‐text papers retrieved, and 441 studies included (see Table 1 for an overview, and supplementary information for the flow chart concerning each disorder). The number of included studies ranged from 159 (for MDD) to 21 (for BPD). Because several studies compared two or more interventions with one control group, the number of comparisons was larger (total: 569, ranging from 196 for MDD to 22 for BPD). The number of included patients was 33,881 (19,769 in treatment and 14,112 in control conditions), ranging from 1,011 in OCD to 14,908 in MDD dataset.
The proportion of patients exclusively recruited from clinical settings ranged from 4.2% for specific phobia to 85.7% for BPD, and was 36.3% across all disorders. The overall mean age was 39.0±10.7 years (range: 31.9±5.6 for BPD to 42.2±13.1 for MDD). The overall proportion of women was 67% (range: 56% for social anxiety disorder to 83% for BPD).
The type of control condition varied considerably across datasets. Waitlist was used in most trials in the four anxiety disorders (70.8% to 95.8%), PTSD (79.3%) and OCD (72.7%), but less frequently in MDD (42.0%) and not at all in BPD. Pill placebo was used in only a small proportion of studies (5.0% overall), while care‐as‐usual was the most used control condition in MDD (54.9%) and BPD (100%). Most trials were conducted in North America (39.0%) or Europe (37.6%).
Half of the interventions used an individual format (51.8%; range: 36.4% for BPD to 80.5% for PTSD). Group therapies were not used at all in some disorders (BPD, specific phobia and OCD) and ranged from 9.2% for PTSD to 33.8% for social anxiety disorder. Guided self‐help was not used in BPD, and ranged from 3.1% for specific phobia to 23.0% for social anxiety disorder. The mean number of sessions across all disorders was 11.8±8.5 (range: 3.3±3.1 for specific phobia to 53.6±42.6 for BPD). The proportion of studies with low RoB ranged from 6.2% for panic disorder to 39.5% for MDD, but this should be considered with caution, because the definition of low RoB differed across datasets.
Response rates
Response rates in the psychotherapy conditions were 0.42 (95% CI: 0.39‐0.45) for MDD; 0.38 (95% CI: 0.33‐0.43) for PTSD; 0.38 (95% CI: 0.30‐0.47) for OCD; 0.38 (95% CI: 0.33‐0.43) for panic disorder; 0.36 (95% CI: 0.30‐0.42) for GAD; 0.32 (95% CI: 0.29‐0.37) for social anxiety disorder; 0.32 (95% CI: 0.23‐0.42) for specific phobia; and 0.24 (95% CI: 0.15‐0.36) for BPD (see Table 2 and Figure 1). Heterogeneity was moderate to high for all disorders (I2 ranged from 65% for OCD to 82% for MDD and BPD).
Table 2.
Response rates and relative outcomes across all psychotherapies and all disorders
| All psychotherapies | All control groups | Relative outcomes | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n | Response rate | 95% CI | I2 (%) | 95% CI | n | Response rate | 95% CI | I2 (%) | 95% CI | RR | 95% CI | I2 (%) | 95% CI | NNT | 95% CI | |
| Main outcomes | ||||||||||||||||
| MDD | 196 | 0.42 | 0.39‐0.45 | 82 | 79‐84 | 162 | 0.19 | 0.17‐0.21 | 72 | 67‐76 | 2.09 | 1.91‐2.28 | 40 | 29‐50 | 4.8 | 4.1‐5.8 |
| SAD | 74 | 0.32 | 0.29‐0.37 | 66 | 56‐73 | 52 | 0.12 | 0.09‐0.14 | 45 | 23‐60 | 2.74 | 2.36‐3.18 | 0 | 0‐28 | 4.8 | 3.8‐6.1 |
| PAN | 71 | 0.38 | 0.33‐0.43 | 77 | 71‐82 | 48 | 0.16 | 0.13‐0.20 | 61 | 46‐71 | 2.24 | 1.86‐2.69 | 47 | 31‐60 | 5.0 | 3.7‐7.3 |
| GAD | 61 | 0.36 | 0.30‐0.42 | 74 | 67‐80 | 48 | 0.15 | 0.11‐0.19 | 71 | 61‐78 | 2.28 | 1.85‐2.81 | 24 | 0‐45 | 5.2 | 3.7‐7.8 |
| PHOB | 32 | 0.32 | 0.23‐0.42 | 78 | 70‐84 | 22 | 0.09 | 0.06‐0.12 | 0 | 0‐46 | 3.40 | 2.35‐4.92 | 0 | 0‐0.40 | 4.6 | 2.8‐8.2 |
| PTSD | 87 | 0.38 | 0.33‐0.43 | 74 | 68‐79 | 69 | 0.10 | 0.08‐0.13 | 44 | 25‐58 | 3.09 | 2.62‐3.65 | 0 | 0‐26 | 4.8 | 3.8‐6.2 |
| OCD | 26 | 0.38 | 0.30‐0.47 | 65 | 47‐77 | 22 | 0.05 | 0.03‐0.07 | 0 | 0‐46 | 9.28 | 6.40‐13.48 | 0 | 0‐43 | 2.4 | 1.6‐3.7 |
| BPD | 22 | 0.24 | 0.15‐0.36 | 82 | 73‐87 | 21 | 0.15 | 0.10‐0.21 | 75 | 61‐83 | 1.47 | 0.90‐2.39 | 27 | 0‐57 | 14.2 | ‐ |
| All dropouts as responders | ||||||||||||||||
| MDD | 196 | 0.48 | 0.45‐0.51 | 80 | 77‐83 | 162 | 0.23 | 0.20‐0.26 | 78 | 75‐81 | 1.99 | 1.81‐2.18 | 51 | 43‐59 | 4.4 | 3.7‐5.4 |
| SAD | 74 | 0.33 | 0.29‐0.37 | 66 | 56‐73 | 52 | 0.12 | 0.09‐0.14 | 45 | 23‐60 | 2.74 | 2.36‐3.18 | 0 | 0‐28 | 4.8 | 3.8‐6.1 |
| PAN | 71 | 0.42 | 0.37‐0.48 | 80 | 75‐84 | 48 | 0.18 | 0.15‐0.23 | 70 | 60‐78 | 2.19 | 1.87‐2.57 | 49 | 33‐62 | 4.7 | 3.5‐6.4 |
| GAD | 61 | 0.40 | 0.34‐0.46 | 76 | 69‐81 | 48 | 0.17 | 0.13‐0.22 | 69 | 59‐77 | 2.18 | 1.80‐2.65 | 57 | 43‐68 | 5.0 | 3.6‐7.4 |
| PHOB | 32 | 0.35 | 0.25‐0.46 | 79 | 71‐85 | 22 | 0.09 | 0.07‐0.13 | 1 | 0‐46 | 3.40 | 2.35‐4.92 | 0 | 0‐0.40 | 4.6 | 2.8‐8.2 |
| PTSD | 87 | 0.49 | 0.44‐0.54 | 71 | 65‐77 | 69 | 0.17 | 0.14‐0.20 | 55 | 41‐66 | 2.75 | 2.36‐3.20 | 22 | 0‐44 | 3.4 | 2.7‐4.3 |
| OCD | 26 | 0.44 | 0.36‐0.52 | 63 | 44‐76 | 22 | 0.10 | 0.07‐0.15 | 15 | 0‐49 | 4.10 | 2.79‐6.02 | 0 | 0‐43 | 3.2 | 2.0‐5.6 |
| BPD | 22 | 0.35 | 0.25‐0.47 | 78 | 67‐85 | 21 | 0.23 | 0.17‐0.31 | 71 | 55‐81 | 1.48 | 1.09‐2.02 | 23 | 0‐54 | 8.2 | 3.8‐43.5 |
| Completers only | ||||||||||||||||
| MDD | 196 | 0.44 | 0.41‐0.48 | 80 | 77‐82 | 162 | 0.20 | 0.18‐0.22 | 72 | 67‐76 | 2.12 | 1.93‐2.32 | 42 | 31‐51 | 4.5 | 3.8‐5.4 |
| SAD | 74 | 0.34 | 0.30‐0.38 | 68 | 59‐75 | 52 | 0.12 | 0.09‐0.15 | 47 | 27‐62 | 2.73 | 2.34‐3.17 | 0 | 0‐28 | 4.8 | 3.8‐6.2 |
| PAN | 71 | 0.40 | 0.35‐0.45 | 78 | 72‐82 | 48 | 0.16 | 0.13‐0.21 | 64 | 51‐73 | 2.27 | 1.89‐2.72 | 48 | 31‐60 | 4.9 | 3.6‐7.0 |
| GAD | 61 | 0.37 | 0.31‐0.44 | 75 | 68‐80 | 48 | 0.15 | 0.12‐0.20 | 67 | 55‐75 | 2.29 | 1.86‐2.82 | 46 | 26‐60 | 5.2 | 3.7‐7.8 |
| PHOB | 32 | 0.33 | 0.24‐0.43 | 77 | 68‐83 | 32 | 0.09 | 0.06‐0.13 | 0 | 0‐46 | 3.45 | 2.37‐5.01 | 0 | 0‐0.40 | 4.5 | 2.8‐8.1 |
| PTSD | 87 | 0.42 | 0.38‐0.47 | 68 | 60‐74 | 69 | 0.11 | 0.09‐0.14 | 40 | 20‐56 | 3.48 | 2.97‐4.08 | 0 | 0‐26 | 3.7 | 3.0‐4.6 |
| OCD | 26 | 0.40 | 0.32‐0.49 | 64 | 45‐76 | 22 | 0.05 | 0.03‐0.08 | 0 | 0‐46 | 9.28 | 6.42‐13.40 | 0 | 0‐43 | 2.4 | 1.6‐3.7 |
| BPD | 22 | 0.26 | 0.17‐0.39 | 79 | 70‐86 | 21 | 0.16 | 0.11‐0.23 | 72 | 57‐82 | 1.50 | 0.95‐2.37 | 24 | 0‐55 | 12.5 | ‐ |
RR – relative risk, NNT – number needed to treat, MDD – major depressive disorder, SAD – social anxiety disorder, PAN – panic disorder, GAD – generalized anxiety disorder, PHOB – specific phobia, PTSD – post‐traumatic stress disorder, OCD – obsessive‐compulsive disorder, BPD – borderline personality disorder
Figure 1.
Response rates and relative risks for response to psychotherapies (with 95% CIs in brackets). MDD – major depressive disorder, SAD – social anxiety disorder, PAN – panic disorder, GAD – generalized anxiety disorder, PHOB – specific phobia, PTSD – post‐traumatic stress disorder, OCD – obsessive‐compulsive disorder, BPD – borderline personality disorder. Grey boxes indicate control conditions and black boxes indicate psychotherapies.
When response rates were based on the least conservative assumption of all dropouts being responders, the outcomes were more positive (0.48 for MDD, 0.49 for PTSD, 0.44 for OCD, 0.42 for panic disorder, 0.40 for GAD, 0.33 for social anxiety disorder, 0.35 for specific phobia, and 0.35 for BPD). Heterogeneity remained moderate to large (I2 range: 63 to 80%). Response rates based on completers only were 0.44 for MDD, 0.42 for PTSD, 0.40 for OCD, 0.40 for panic disorder, 0.37 for GAD, 0.34 for social anxiety disorder, 0.33 for specific phobia, and 0.26 for BPD (I2 range: 64 to 80%).
Sensitivity analyses indicated that the number of studies with low RoB was very small across all disorders, except for MDD (see Table 3). For this condition, the response rate for psychotherapies in low RoB studies was somewhat smaller than in the main analyses, but not substantially (0.37; 95% CI: 0.33‐0.42). Exclusion of outliers did not result in major changes of the response rates (all differences with respect to main analyses were smaller than 0.06). The same applied to adjustment for publication bias (all differences <0.08), except for BPD, which had a considerably larger response rate after adjustment (0.38 compared to 0.24).
Table 3.
Results of sensitivity analyses
| All psychotherapies | All control groups | Relative outcomes | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n | Response rate | 95% CI | I2 (%) | 95% CI | n | Response rate | 95% CI | I2 (%) | 95% CI | RR | 95% CI | I2 (%) | 95% CI | NNT | 95% CI | |
| Only low RoB | ||||||||||||||||
| MDD | 71 | 0.37 | 0.33‐0.42 | 86 | 83‐88 | 64 | 0.19 | 0.16‐0.22 | 75 | 68‐80 | 1.82 | 1.61‐2.06 | 25 | 0‐45 | 6.4 | 5.0‐8.6 |
| SAD | 5 | 0.76 | 0.32‐0.95 | 89 | 78‐95 | 4 | 0.14 | 0.04‐0.38 | 73 | 23‐90 | 3.50 | 1.23‐9.90 | 10 | 0‐81 | 2.9 | 0.8‐31.1 |
| PAN | 4 | 0.22 | 0.14‐0.32 | 74 | 27‐91 | 3 | 0.12 | 0.02‐0.47 | 88 | 68‐96 | 1.89 | 0.50‐7.05 | 39 | 0‐79 | 9.4 | ‐ |
| GAD | 9 | 0.44 | 0.28‐0.60 | 82 | 67‐90 | 8 | 0.17 | 0.13‐0.23 | 29 | 0‐68 | 2.30 | 1.64‐3.25 | 13 | 0‐54 | 4.5 | 2.6‐9.2 |
| PHOB | 3 | 0.36 | 0.14‐0.66 | 86 | 60‐95 | 3 | 0.08 | 0.03‐0.17 | 9 | 0‐91 | 5.44 | 0.33‐90.19 | 51 | 0‐86 | 2.8 | ‐ |
| PTSD | 7 | 0.43 | 0.31‐0.55 | 66 | 24‐85 | 6 | 0.09 | 0.06‐0.15 | 0 | 0‐75 | 3.62 | 1.71‐7.63 | 18 | 0‐62 | 4.2 | 1.7‐15.6 |
| OCD | 3 | 0.44 | 0.21‐0.71 | 62 | 0‐89 | 3 | 0.09 | 0.03‐0.24 | 0 | 0‐90 | 5.24 | 0.07‐406.51 | 0 | 0‐90 | 2.6 | ‐ |
| BPD | 4 | 0.49 | 0.20‐0.78 | 90 | 77‐96 | 3 | 0.25 | 0.07‐0.59 | 87 | 64‐96 | 1.45 | 0.72‐2.91 | 69 | 10‐89 | 8.9 | ‐ |
| Outliers excluded | ||||||||||||||||
| MDD | 133 | 0.41 | 0.39‐0.43 | 31 | 14‐45 | 131 | 0.18 | 0.17‐0.20 | 29 | 11‐43 | 2.22 | 2.03‐2.42 | 0 | 0‐19 | 4.6 | 3.9‐5.4 |
| SAD | 63 | 0.31 | 0.29‐0.34 | 14 | 0‐38 | 49 | 0.11 | 0.09‐0.13 | 0 | 0‐33 | 2.89 | 2.49‐3.36 | 0 | 0‐28 | 4.8 | 3.9‐6.1 |
| PAN | 55 | 0.35 | 0.32‐0.38 | 37 | 13‐55 | 42 | 0.15 | 0.12‐0.18 | 32 | 0‐53 | 2.31 | 1.92‐2.78 | 0 | 0‐29 | 5.1 | 3.7‐7.2 |
| GAD | 46 | 0.37 | 0.34‐0.41 | 43 | 19‐60 | 45 | 0.14 | 0.11‐0.17 | 37 | 9‐56 | 2.47 | 2.02‐3.03 | 0 | 0‐31 | 4.9 | 3.5‐7.0 |
| PHOB | 26 | 0.24 | 0.19‐0.30 | 39 | 2‐62 | 22 | 0.09 | 0.06‐0.12 | 0 | 0‐46 | 3.40 | 2.35‐4.92 | 0 | 0‐0.40 | 4.6 | 2.8‐8.2 |
| PTSD | 63 | 0.37 | 0.34‐0.41 | 40 | 19‐56 | 66 | 0.10 | 0.09‐0.12 | 0 | 0‐29 | 3.24 | 2.76‐3.80 | 0 | 0‐27 | 4.5 | 3.6‐5.7 |
| OCD | 23 | 0.36 | 0.31‐0.42 | 36 | 0‐62 | 22 | 0.05 | 0.03‐0.07 | 0 | 0‐46 | 9.28 | 6.40‐13.48 | 0 | 0‐43 | 2.4 | 1.6‐3.7 |
| BPD | 17 | 0.20 | 0.14‐0.28 | 69 | 49‐81 | 19 | 0.13 | 0.10‐0.18 | 31 | 0‐61 | 1.61 | 1.07‐2.42 | 14 | 0‐49 | 12.6 | 5.4‐109.9 |
| Adjusted for publication bias | ||||||||||||||||
| MDD | 206 | 0.44 | 0.41‐0.48 | 84 | 82‐85 | 217 | 0.25 | 0.22‐0.28 | 75 | 72‐78 | 1.63 | 1.45‐1.84 | 54 | 47‐60 | 6.3 | 4.8‐8.9 |
| SAD | 90 | 0.37 | 0.33‐0.42 | 70 | 62‐75 | 72 | 0.15 | 0.12‐0.18 | 45 | 27‐58 | NC | NC | ||||
| PAN | 71 | 0.38 | 0.33‐0.44 | 77 | 71‐82 | 66 | 0.23 | 0.18‐0.29 | 64 | 53‐72 | 1.29 | 0.96‐1.73 | 55 | 43‐64 | 15.0 | ‐ |
| GAD | 68 | 0.40 | 0.33‐0.46 | 76 | 70‐81 | 68 | 0.24 | 0.18‐0.31 | 73 | 66‐79 | 1.66 | 1.26‐2.20 | 31 | 9‐47 | 6.3 | 3.5‐16.0 |
| PHOB | 40 | 0.40 | 0.30‐0.51 | 78 | 70‐84 | NC | 2.94 | 1.97‐4.38 | 0 | 0‐36 | 5.7 | 3.3‐11.5 | ||||
| PTSD | 95 | 0.41 | 0.35‐0.46 | 75 | 69‐79 | 98 | 0.15 | 0.12‐0.18 | 49 | 35‐59 | 2.25 | 1.80‐2.82 | 11 | 0‐29 | 5.3 | 3.7‐8.3 |
| OCD | 28 | 0.41 | 0.32‐0.51 | 70 | 55‐79 | 30 | 0.06 | 0.04‐0.08 | 0 | 0‐41 | 7.58 | 5.11‐11.24 | 0 | 0‐37 | 2.5 | 1.6‐4.1 |
| BPD | 29 | 0.38 | 0.24‐0.56 | 84 | 78‐88 | 31 | 0.27 | 0.16‐0.42 | 77 | 68‐84 | 1.21 | 0.67‐2.19 | 25 | 0‐53 | 17.6 | ‐ |
RoB – risk of bias, RR – relative risk, NNT – number needed to treat, MDD – major depressive disorder, SAD – social anxiety disorder, PAN – panic disorder, GAD – generalized anxiety disorder, PHOB – specific phobia, PTSD – post‐traumatic stress disorder, OCD – obsessive‐compulsive disorder, BPD – borderline personality disorder, NC – not calculable
The overall response rates for the control conditions ranged from 0.05 for OCD to 0.19 for MDD, with heterogeneity (I2) ranging from 0% for specific phobia and OCD to 75% for BPD. The rates were higher when all dropouts were considered responders (range: 0.09 to 0.23, I2 range: 1% to 78%), and intermediate when only completers were included (range: 0.05 to 0.20; I2 range: 0% to 72%) (see Table 2 and Figure 1).
The response rates for the control conditions in low RoB studies were comparable with those of the main analyses, except for BPD, which had a substantially larger rate in these studies (0.25 compared to 0.15). However, this was based on only three studies. The exclusion of outliers did not result in major differences with the main analyses. Adjustment for publication bias resulted in rates that were higher than in the main analyses (see Table 3).
Subgroup analyses showed that type of psychotherapy was associated with differential response rates for MDD (p<0.001), panic disorder (p=0.02), specific phobia (p=0.003) and PTSD (p<0.001), but not for social anxiety disorder (p=0.12), GAD (p=0.31), OCD (p=0.65) and BPD (p=0.59) (see Table 4). Significantly different response rates were also found for treatment format in MDD, panic disorder, social anxiety disorder, GAD, specific phobia and PTSD, but not for OCD and BPD (see supplementary information). Recruitment from clinical samples was associated with a significantly smaller response rate in MDD and OCD, but not in the other disorders (see supplementary information). Significant differences between response rates in different countries were found for MDD, panic disorder, specific phobia, and BPD, but not for the other disorders (see supplementary information). Response rates differed significantly across type of control condition for MDD, panic disorder, and PTSD, but not for social anxiety disorder, GAD, and OCD. All trials in BPD had care‐as‐usual as the control condition, and all but one trial on specific phobia used waitlist (see supplementary information).
Table 4.
Response rates and relative outcomes for specific types of psychotherapies
| n | Response rate | 95% CI | p | I2 (%) | RR | 95% CI | I2 (%) | nnt | 95% CI | ||
|---|---|---|---|---|---|---|---|---|---|---|---|
| MDD | CBT | 110 | 0.44 | 0.39‐0.48 | <0.001 | 83 | 2.23 | 1.96‐2.54 | 51 | 4.5 | 3.6‐5.8 |
| Third wave | 17 | 0.40 | 0.31‐0.50 | 68 | 3.57 | 2.63‐4.85 | 0 | 2.2 | 1.4‐3.4 | ||
| BAT | 9 | 0.53 | 0.42‐0.65 | 74 | 2.13 | 1.35‐3.36 | 24 | 4.9 | 2.4‐15.9 | ||
| IPT | 17 | 0.42 | 0.33‐0.51 | 79 | 1.78 | 1.37‐2.31 | 61 | 7.1 | 4.2‐15.0 | ||
| PST | 5 | 0.53 | 0.30‐0.74 | 83 | 2.47 | 0.29‐20.90 | 23 | 3.8 | ‐ | ||
| SUP | 6 | 0.36 | 0.17‐0.60 | 88 | 1.38 | 0.94‐2.04 | 0 | 14.6 | ‐ | ||
| Controls | 157 | 0.19 | 0.17‐0.21 | 73 | |||||||
| PAN | CBT | 42 | 0.41 | 0.35‐0.47 | 0.02 | 80 | 2.42 | 1.90‐3.09 | 57 | 5.0 | 3.4‐7.9 |
| BT | 10 | 0.35 | 0.25‐0.47 | 55 | 4.55 | 2.00‐10.31 | 0 | 2.0 | 0.8‐7.1 | ||
| Relaxation | 5 | 0.41 | 0.23‐0.62 | 74 | 2.33 | 0.75‐7.25 | 29 | 5.4 | ‐ | ||
| Controls | 42 | 0.16 | 0.13‐0.20 | 61 | |||||||
| SAD | CBT | 46 | 0.35 | 0.30‐0.41 | 0.12 | 68 | 3.06 | 2.58‐3.63 | 0 | 4.4 | 3.5‐5.8 |
| Exposure | 11 | 0.28 | 0.20‐0.37 | 46 | 2.82 | 1.62‐4.93 | 0 | 5.0 | 2.3‐14.7 | ||
| Third wave | 6 | 0.29 | 0.12‐0.55 | 84 | 2.95 | 0.96‐9.05 | 11 | 4.7 | ‐ | ||
| Controls | 48 | 0.12 | 0.09‐0.14 | 48 | |||||||
| GAD | C(B)T | 42 | 0.34 | 0.28‐0.41 | 0.31 | 74 | 2.26 | 1.71‐2.97 | 52 | 5.7 | 3.6‐10.1 |
| Third wave | 8 | 0.39 | 0.20‐0.61 | 85 | 2.48 | 1.35‐4.56 | 0 | 4.8 | 2.0‐20.4 | ||
| Controls | 45 | 0.15 | 0.11‐0.19 | 72 | |||||||
| PHOB | Exposure | 20 | 0.35 | 0.25‐0.46 | 0.003 | 77 | 3.41 | 1.99‐5.84 | 0 | 4.6 | 2.3‐11.2 |
| CBT | 7 | 0.39 | 0.15‐0.71 | 87 | 4.61 | 2.14‐9.93 | 0 | 3.1 | 1.2‐9.7 | ||
| Controls | 22 | 0.09 | 0.06‐0.12 | 0 | |||||||
| OCD | CBT | 13 | 0.38 | 0.31‐0.45 | 0.65 | 50 | 8.68 | 5.36‐14.07 | 0 | 2.6 | 1.5‐4.6 |
| ERP | 7 | 0.43 | 0.23‐0.66 | 78 | 7.82 | 2.38‐25.64 | 0 | 2.9 | 0.8‐14.5 | ||
| Controls | 20 | 0.05 | 0.03‐0.07 | 0 | |||||||
| BPD | DBT | 7 | 0.24 | 0.08‐0.55 | 0.59 | 86 | 1.56 | 0.98‐2.46 | 0 | 11.9 | ‐ |
| Controls | 21 | 0.15 | 0.10‐0.21 | 75 | |||||||
| PTSD | TF‐CBT | 22 | 0.35 | 0.28‐0.44 | <0.001 | 69 | 2.57 | 2.12‐3.12 | 0 | 6.4 | 4.7‐8.9 |
| TF‐exposure | 21 | 0.36 | 0.29‐0.44 | 62 | 3.90 | 2.91‐5.22 | 0 | 3.4 | 2.4‐5.2 | ||
| EMDR | 9 | 0.34 | 0.19‐0.53 | 71 | 4.19 | 1.90‐9.21 | 21 | 3.1 | 1.2‐11.1 | ||
| NTF‐CBT | 11 | 0.35 | 0.20‐0.54 | 80 | 2.58 | 1.47‐4.51 | 0 | 6.3 | 2.8‐21.3 | ||
| TF‐CT | 8 | 0.52 | 0.33‐0.70 | 83 | 5.32 | 3.11‐9.08 | 0 | 2.3 | 1.2‐4.7 | ||
| Controls | 67 | 0.10 | 0.08‐0.13 | 36 |
Comparisons with pill placebo are not included, because response rates in this group differed considerably from those in the other control groups. RR – relative risk, NNT – number needed to treat, MDD – major depressive disorder, PAN – panic disorder, SAD – social anxiety disorder, GAD – generalized anxiety disorder, PHOB – specific phobia, PTSD – post‐traumatic stress disorder, OCD – obsessive‐compulsive disorder, BPD – borderline personality disorder, CBT – cognitive behavior therapy, BAT – behavioral activation therapy, IPT – interpersonal psychotherapy, PST – problem‐solving therapy, SUP – supportive therapy, BT – behavior therapy, ERP – exposure and response prevention, DBT – dialectical behavior therapy, TF– trauma focused, EMDR – eye movement desensitization and reprocessing, NTF – non‐trauma focused.
RRs and NNTs
Except for BPD, all RRs for psychotherapies were significant. They were 2.09 (95% CI: 1.91‐2.28) for MDD; 3.09 (95% CI: 2.62‐3.65) for PTSD; 9.28 (95% CI: 6.40‐13.48) for OCD; 2.24 (95% CI: 1.86‐2.69) for panic disorder; 2.28 (95% CI: 1.85‐2.81) for GAD; 2.74 (95% CI: 2.36‐3.18) for social anxiety disorder; 3.40 (95% CI: 2.35‐4.92) for specific phobia; and 1.47 (95% CI: 0.90‐2.39) for BPD (see Table 2 and Figure 1). Heterogeneity was low to moderate across all disorders (I2 range: 0 to 47%).
The NNTs for the significant outcomes were 4.8 (95% CI: 4.1‐5.8) for MDD; 4.8 (95% CI: 3.8‐6.2) for PTSD; 2.4 (95% CI: 1.6‐3.7) for OCD; 5.0 (95% CI: 3.7‐7.3) for panic disorder; 5.2 (95% CI: 3.7‐7.8) for GAD; 4.8 (95% CI: 3.8‐6.1) for social anxiety disorder; and 4.6 (95% CI: 2.8‐8.2) for specific phobia.
The RRs for the studies with low RoB and for the analyses in which the outliers were excluded were very comparable to those of the main analyses (all were significant except for BPD), except that the RR for BPD was also significant when outliers were excluded. In the analyses in which we adjusted for publication bias, all resulting RRs were smaller than in the main analyses, and the RR was no longer significant for panic disorder (see Table 3).
DISCUSSION
We conducted a large meta‐analytic study to assess absolute and relative outcomes of psychotherapies across eight major mental disorders. The response rates in the treatment groups varied from 0.24 for BPD to 0.42 for MDD, while the rates for the other six disorders ranged from 0.32 to 0.39. Most sensitivity analyses broadly supported these findings, although the number of trials with low RoB was small, and in several samples significant indications for publication bias were identified. The pooled response rates across the three control conditions ranged from 0.05 for OCD to 0.19 for MDD. The relative outcomes we found in terms of RR were significant for all disorders, except BPD.
These results indicate that, although most psychotherapies lead to better outcomes compared to control conditions, response rates are modest. Most patients receiving psychotherapy across all disorders do not show at least 50% symptom reduction. This means that clinicians often have to try several interventions or move to pharmacological or combined therapies to treat patients more effectively. Unfortunately, very little research on such sequential treatments has been conducted.
There were significant differences between psychotherapies in terms of response rate in MDD, panic disorder, specific phobia and PTSD. Only one psychotherapy (i.e., dialectical behavior therapy) was tested in BPD, and its response rate was not significantly higher than the care‐as‐usual control condition, typically consisting of intensive treatment, which suggests that specific psychotherapies for BPD may have limited additional benefit. This implication, however, should be considered with caution, because some meta‐analyses exclusively dealing with BPD are broader and may give a better estimate of the effects of psychotherapies 16 , 17 .
There are some limitations of this study that need to be considered. We expected and found high levels of heterogeneity, although most response rates were relatively robust in a series of sensitivity analyses. We found that the number of trials with low RoB was limited, meaning that the findings should be taken with caution. The included meta‐analytic datasets also used different ways of defining low RoB, so that we cannot compare RoB across datasets. The response rates were not directly reported in the studies but were estimated. Although the method we adopted has been well validated, these are still estimates and this may have affected the pooled rates.
We could only include trials that reported enough data to calculate the response rates, and this may have resulted in selection bias. We only examined response rates after treatment and did not take the time to follow‐up into account, which may have introduced additional heterogeneity. Furthermore, we chose a 50% symptom reduction as the main outcome, which is the standard in depression, but may not be the best choice for other disorders, for which no clear standards exist 25 , 26 . Finally, we did not examine differences across studies in patient characteristics that can affect outcomes (e.g., gender, and veteran status in PTSD) 33 .
This is the first meta‐analysis of trials examining the effects of psychotherapies for eight different mental disorders, established with a diagnostic interview. It is also the first to report the absolute measure of response rate next to relative outcomes, using the same definition of response (number of patients with at least 50% symptom reduction between baseline and post‐test, divided by the total number of patients) across the mental disorders. The results of this study can contribute to improved clinical decision‐making and technology‐supported tools 34 .
We conclude that most psychotherapies for the eight mental disorders are effective compared to control conditions, but that the response rates are modest, ranging from 0.24 for BPD to 0.42 for MDD. More effective interventions, as well as therapies for those not responding to a first‐line treatment, are clearly needed.
ACKNOWLEDGEMENT
Supplementary information on this study is available at https://osf.io/2htsv.
REFERENCES
- 1. World Health Organization . World mental health report: transforming mental health for all. Geneva: World Health Organization, 2022. [Google Scholar]
- 2. Leichsenring F, Steinert C, Rabung S et al. The efficacy of psychotherapies and pharmacotherapies for mental disorders in adults: an umbrella review and meta‐analytic evaluation of recent meta‐analyses. World Psychiatry 2022;21:133‐45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Cummings P. Arguments for and against standardized mean differences (effect sizes). Arch Pediatr Adolesc Med 2011;165:592‐6. [DOI] [PubMed] [Google Scholar]
- 4. Hieronymus F, Jauhar S, Østergaard SD. One (effect) size does not fit at all: interpreting clinical significance and effect sizes in depression treatment trials. J Psychopharmacol 2020;34:1074‐8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Cuijpers P, Karyotaki E, Ciharova M et al. The effects of psychotherapies for depression on response, remission, reliable change, and deterioration: a meta‐analysis. Acta Psychiatr Scand 2021;144:288‐99. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Laupacis A, Sackett DL, Roberts RS. An assessment of clinically useful measures of the consequences of treatment. N Engl J Med 1988;318:1728‐33. [DOI] [PubMed] [Google Scholar]
- 7. Furukawa TA, Guyatt GH, Griffith LE. Can we individualize the ‘number needed to treat’? An empirical study of summary effect measures in meta‐analyses. Int J Epidemiol 2002;31:72‐6. [DOI] [PubMed] [Google Scholar]
- 8. Charlson F, van Ommeren M, Flaxman A et al. New WHO prevalence estimates of mental disorders in conflict settings: a systematic review and meta‐analysis. Lancet 2019;394:240‐8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Rotenstein LS, Ramos MA, Torre M et al. Prevalence of depression, depressive symptoms, and suicidal ideation among medical students: a systematic review and meta‐analysis. JAMA 2016;316:2214‐36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Kavousi M, Desai CS, Ayers C et al. Prevalence and prognostic implications of coronary artery calcification in low‐risk women: a meta‐analysis. JAMA 2016;316:2126‐34. [DOI] [PubMed] [Google Scholar]
- 11. Cuijpers P, Quero S, Noma H et al. Psychotherapies for depression: a network meta‐analysis covering efficacy, acceptability and long‐term outcomes of all main treatment types. World Psychiatry 2021;20:283‐93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Furukawa TA, Shinohara K, Sahker E et al. Initial treatment choices to achieve sustained response in major depression: a systematic review and network meta‐analysis. World Psychiatry 2021;20:387‐96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Mavranezouli I, Megnin‐Viggars O, Daly C et al. Psychological treatments for post‐traumatic stress disorder in adults: a network meta‐analysis. Psychol Med 2020;50:542‐55. [DOI] [PubMed] [Google Scholar]
- 14. Hoppen TH, Jehn M, Holling H et al. The efficacy and acceptability of psychological interventions for adult PTSD: a network and pairwise meta‐analysis of randomized controlled trials. J Consult Clin Psychol 2023;91:445‐61. [DOI] [PubMed] [Google Scholar]
- 15. Reid JE, Laws KR, Drummond L et al. Cognitive behavioural therapy with exposure and response prevention in the treatment of obsessive‐compulsive disorder: a systematic review and meta‐analysis of randomised controlled trials. Compr Psychiatry 2021;106:152223. [DOI] [PubMed] [Google Scholar]
- 16. Setkowski K, Palantza C, van Ballegooijen W et al. Which psychotherapy is most effective and acceptable in the treatment of adults with a (sub)clinical borderline personality disorder? A systematic review and network meta‐analysis. Psychol Med 2023;53:3261‐80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Cristea IA, Gentili C, Cotet CD et al. Efficacy of psychotherapy for borderline personality disorder: a systematic review and meta‐analysis. JAMA Psychiatry 2017;74:319‐28. [DOI] [PubMed] [Google Scholar]
- 18. Papola P, Miguel C, Mazzaglia M et al. Psychotherapies for generalized anxiety disorders in adults: systematic review and network meta‐analysis of randomized‐controlled trials. JAMA Psychiatry 2023;81:250‐9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Papola D, Ostuzzi G, Tedeschi F et al. Comparative efficacy and acceptability of psychotherapies for panic disorder with or without agoraphobia: systematic review and network meta‐analysis of randomised controlled trials. Br J Psychiatry 2021;221:507‐19. [DOI] [PubMed] [Google Scholar]
- 20. Odgers K, Kershaw KA, Li SH et al. The relative efficacy and efficiency of single‐ and multi‐session exposure therapies for specific phobia: a meta‐analysis. Behav Res Ther 2022;159:104203. [DOI] [PubMed] [Google Scholar]
- 21. Van Dis EA, Van Veen SC, Hagenaars MA et al. Long‐term outcomes of cognitive behavioral therapy for anxiety‐related disorders: a systematic review and meta‐analysis. JAMA Psychiatry 2020;77:265‐73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Cuijpers P, Cristea IA, Karyotaki E et al. How effective are cognitive behavior therapies for major depression and anxiety disorders? A meta‐analytic update of the evidence. World Psychiatry 2016;15:245‐58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Furukawa TA, Noma H, Caldwell DM et al. Waiting list may be a nocebo condition in psychotherapy trials: a contribution from network meta‐analysis. Acta Psychiatr Scand 2014;130;181‐92. [DOI] [PubMed] [Google Scholar]
- 24. Furukawa TA, Cipriani A, Barbui C et al. Imputing response rates from means and standard deviations in meta‐analyses. Int Clin Psychopharmacol 2005;20:49‐52. [DOI] [PubMed] [Google Scholar]
- 25. Loerinc AG, Meuret AE, Twohig MP et al. Response rates for CBT for anxiety disorders: need for standardized criteria. Clin Psychol Rev 2015;42:72‐82. [DOI] [PubMed] [Google Scholar]
- 26. Varker T, Kartal D, Watson L et al. Defining response and nonresponse to posttraumatic stress disorder treatments: a systematic review. Clin Psychol Sci Pract 2020;27:e12355. [Google Scholar]
- 27. Campbell LF, Norcross JC, Vasquez MJ et al. Recognition of psychotherapy effectiveness: the APA resolution. Psychotherapy 2013;50:98‐101. [DOI] [PubMed] [Google Scholar]
- 28. Higgins JPT, Altman DG, Gøtzsche PC et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ 2011;343:d5928. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Sterne JAC, Savović J, Page MJ et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ 2019;366:l4898. [DOI] [PubMed] [Google Scholar]
- 30. Cuijpers P, Noma H, Karyotaki E et al. A network meta‐analysis of the effects of psychotherapies, pharmacotherapies and their combination in the treatment of adult depression. World Psychiatry 2020;19:92‐107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Higgins JPT, Green S (eds). Cochrane handbook for systematic reviews of interventions. Version 5.1.0. Chichester: Wiley, 2011. [Google Scholar]
- 32. Higgins JPT, Thompson SG, Deeks JJ et al. Measuring inconsistency in meta‐analyses. BMJ 2003;327:557‐60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Watts BV, Schnurr PP, Mayo L et al. Meta‐analysis of the efficacy of treatments for posttraumatic stress disorder. J Clin Psychiatry 2013;74:e541‐50. [DOI] [PubMed] [Google Scholar]
- 34. Hallgren KA, Bauer AM, Atkins DC. Digital technology and clinical decision making in depression treatment: current findings and future opportunities. Depress Anxiety 2017;34:494‐501. [DOI] [PMC free article] [PubMed] [Google Scholar]