Активирующая поведенческая психотерапия при депрессии у взрослых
Abstract
Background
Behavioural activation is a brief psychotherapeutic approach that seeks to change the way a person interacts with their environment. Behavioural activation is increasingly receiving attention as a potentially cost‐effective intervention for depression, which may require less resources and may be easier to deliver and implement than other types of psychotherapy.
Objectives
To examine the effects of behavioural activation compared with other psychological therapies for depression in adults.
To examine the effects of behavioural activation compared with medication for depression in adults.
To examine the effects of behavioural activation compared with treatment as usual/waiting list/placebo no treatment for depression in adults.
Search methods
We searched CCMD‐CTR (all available years), CENTRAL (current issue), Ovid MEDLINE (1946 onwards), Ovid EMBASE (1980 onwards), and Ovid PsycINFO (1806 onwards) on the 17 January 2020 to identify randomised controlled trials (RCTs) of 'behavioural activation', or the main elements of behavioural activation for depression in participants with clinically diagnosed depression or subthreshold depression. We did not apply any restrictions on date, language or publication status to the searches. We searched international trials registries via the World Health Organization's trials portal (ICTRP) and ClinicalTrials.gov to identify unpublished or ongoing trials.
Selection criteria
We included randomised controlled trials (RCTs) of behavioural activation for the treatment of depression or symptoms of depression in adults aged 18 or over. We excluded RCTs conducted in inpatient settings and with trial participants selected because of a physical comorbidity. Studies were included regardless of reported outcomes.
Data collection and analysis
Two review authors independently screened all titles/abstracts and full‐text manuscripts for inclusion. Data extraction and 'Risk of bias' assessments were also performed by two review authors in duplicate. Where necessary, we contacted study authors for more information.
Main results
Fifty‐three studies with 5495 participants were included; 51 parallel group RCTs and two cluster‐RCTs.
We found moderate‐certainty evidence that behavioural activation had greater short‐term efficacy than treatment as usual (risk ratio (RR) 1.40, 95% confidence interval (CI) 1.10 to 1.78; 7 RCTs, 1533 participants), although this difference was no longer evident in sensitivity analyses using a worst‐case or intention‐to‐treat scenario. Compared with waiting list, behavioural activation may be more effective, but there were fewer data in this comparison and evidence was of low certainty (RR 2.14, 95% CI 0.90 to 5.09; 1 RCT, 26 participants). No evidence on treatment efficacy was available for behavioural activation versus placebo and behavioural activation versus no treatment.
We found moderate‐certainty evidence suggesting no evidence of a difference in short‐term treatment efficacy between behavioural activation and CBT (RR 0.99, 95% CI 0.92 to 1.07; 5 RCTs, 601 participants). Fewer data were available for other comparators. No evidence of a difference in short term‐efficacy was found between behavioural activation and third‐wave CBT (RR 1.10, 95% CI 0.91 to 1.33; 2 RCTs, 98 participants; low certainty), and psychodynamic therapy (RR 1.21, 95% CI 0.74 to 1.99; 1 RCT,60 participants; very low certainty). Behavioural activation was more effective than humanistic therapy (RR 1.84, 95% CI 1.15 to 2.95; 2 RCTs, 46 participants; low certainty) and medication (RR 1.77, 95% CI 1.14 to 2.76; 1 RCT; 141 participants; moderate certainty), but both of these results were based on a small number of trials and participants. No evidence on treatment efficacy was available for comparisons between behavioural activation versus interpersonal, cognitive analytic, and integrative therapies.
There was moderate‐certainty evidence that behavioural activation might have lower treatment acceptability (based on dropout rate) than treatment as usual in the short term, although the data did not confirm a difference and results lacked precision (RR 1.64, 95% CI 0.81 to 3.31; 14 RCTs, 2518 participants). Moderate‐certainty evidence did not suggest any difference in short‐term acceptability between behavioural activation and waiting list (RR 1.17, 95% CI 0.70 to 1.93; 8 RCTs. 359 participants), no treatment (RR 0.97, 95% CI 0.45 to 2.09; 3 RCTs, 187 participants), medication (RR 0.52, 95% CI 0.23 to 1.16; 2 RCTs, 243 participants), or placebo (RR 0.72, 95% CI 0.31 to 1.67; 1 RCT; 96 participants; low‐certainty evidence). No evidence on treatment acceptability was available comparing behavioural activation versus psychodynamic therapy.
Low‐certainty evidence did not show a difference in short‐term treatment acceptability (dropout rate) between behavioural activation and CBT (RR 1.03, 95% CI 0.85 to 1.25; 12 RCTs, 1195 participants), third‐wave CBT (RR 0.84, 95% CI 0.33 to 2.10; 3 RCTs, 147 participants); humanistic therapy (RR 1.06, 95% CI 0.20 to 5.55; 2 RCTs, 96 participants) (very low certainty), and interpersonal, cognitive analytic, and integrative therapy (RR 0.84, 95% CI 0.32 to 2.20; 4 RCTs, 123 participants).
Results from medium‐ and long‐term primary outcomes, secondary outcomes, subgroup analyses, and sensitivity analyses are summarised in the text.
Authors' conclusions
This systematic review suggests that behavioural activation may be more effective than humanistic therapy, medication, and treatment as usual, and that it may be no less effective than CBT, psychodynamic therapy, or being placed on a waiting list. However, our confidence in these findings is limited due to concerns about the certainty of the evidence.
We found no evidence of a difference in short‐term treatment acceptability (based on dropouts) between behavioural activation and most comparison groups (CBT, humanistic therapy, waiting list, placebo, medication, no treatment or treatment as usual). Again, our confidence in all these findings is limited due to concerns about the certainty of the evidence.
No data were available about the efficacy of behavioural activation compared with placebo, or about treatment acceptability comparing behavioural activation and psychodynamic therapy, interpersonal, cognitive analytic and integrative therapies.
The evidence could be strengthened by better reporting and better quality RCTs of behavioural activation and by assessing working mechanisms of behavioural activation.
PICO
Резюме на простом языке
Активирующая поведенческая психотерапия при депрессии у взрослых
Вопрос обзора
В этом Кокрейновском обзоре мы хотели узнать, насколько эффективна активирующая поведенческая психотерапия при депрессии у взрослых.
Почему это важно?
Депрессия ‐ это распространенная проблема психического здоровья, которая может вызвать стойкое чувство печали и потерю интереса к людям, деятельности и вещам, которые когда‐то были приятными. Человек с депрессией может чувствовать большую часть времени плаксивость, раздражительность или усталость, у него могут быть проблемы со сном, концентрацией внимания и памятью. Эти и другие симптомы могут усложнить повседневную жизнь.
Лечение депрессии включает в себя медикаментозное (антидепрессанты) и психотерапевтическое (разговорная терапия) лечение. Поведенческая активация ‐ это вид психологической терапии, который побуждает человека развивать или возвращаться к значимой для него деятельности. Психотерапия включала в себя составление расписания мероприятий и мониторинг поведения, а также рассмотрение конкретных ситуаций, в которых изменение такого поведения и деятельности может быть полезным. Психотерапевт может оказывать помощь во время личных встреч, по телефону или онлайн, как правило, во время нескольких сеансов.
Важно знать, может ли поведенческая активация быть эффективным и приемлемым лечением для людей с депрессией.
Что мы сделали
В январе 2020 года мы провели поиск исследований по поведенческой активирующей психотерапии депрессии у взрослых (старше 18 лет). Мы искали рандомизированные контролируемые испытания, в которых виды лечения были распределены между участниками в случайном порядке; эти исследования дают наиболее надежные доказательства.
Мы включили 53 исследования с 5495 участниками. В исследованиях сравнивали поведенческую активацию с отсутствием лечения, стандартной или обычной помощью, фиктивным лечением (плацебо), приемом лекарств, пребыванием в листе ожидания лечения или другими видами психотерапии (когнитивно‐поведенческой терапией (КПТ), КПТ третьей волны, гуманистической терапией, психодинамической терапией и интегративной терапией).
Исследования проводились в 14 странах; большинство из них ‐ в США (27 исследований). Большинство исследований длились от четырех до 16 недель.
В результате мы сосредоточились на том, насколько хорошо работает лечение и приемлемо ли оно для участников. То, насколько хорошо работало лечение (эффективность), измерялось числом людей, которые хорошо ответили на лечение или перестали соответствовать критериям депрессии по окончании лечения. Приемлемость измеряли путем подсчета, сколько человек выбыло из исследования.
Что мы нашли?
Поведенческая активация может лечить депрессию лучше, чем обычная помощь. Мы не были уверены, работает ли поведенческая активация лучше, чем лекарства или пребывание в листе ожидания, и мы не нашли никаких доказательств этого результата, сравнивая поведенческую активацию с отсутствием лечения или плацебо лечения.
Мы не обнаружили различий между поведенческой активацией и КПТ в лечении депрессии. Несмотря на то, что мы не нашли достаточных доказательств для сравнения поведенческой активации с другими видами психотерапии, она может работать лучше, чем гуманистическая терапия, и мы не обнаружили различий между поведенческой активацией и третьей волной КПТ или психодинамической терапией. Нет было доступных доказательств по сравнению поведенческой активации с интегративной терапией.
Поведенческая активация, вероятно, менее приемлема для людей, чем обычная помощь. Мы не обнаружили различий в приемлемости поведенческой активации по сравнению с пребыванием в листе ожидания, отсутствием лечения, приемом антидепрессантов или плацебо. Мы также не обнаружили различий в приемлемости между поведенческой активацией и другими исследуемыми видами психотерапии (КПТ, КПТ третьей волны, гуманистическая терапия, интегративная терапия). Для сравнения поведенческой активации с психодинамической терапией мы не нашли никаких доказательств по приемлемости лечения.
Выводы
Поведенческая активация может быть эффективным и приемлемым методом лечения депрессии у взрослых. Пациенты с депрессией при назначении этого варианта психотерапии получат больший выбор лечения, а также можно будет изучить различные форматы и типы помощи, чтобы удовлетворить спрос на поддержку психического здоровья. Наше доверие к этим выводам ограничено из‐за опасений относительно определенности доказательств.
Большинство результатов были краткосрочными, а это означает, что мы не можем быть уверены в том, что поведенческая активация будет полезна для людей с депрессией в долгосрочной перспективе.
Определенность доказательств
Наша уверенность в доказательствах в основном низкая или умеренная. Некоторые результаты основаны лишь на нескольких исследованиях с плохо представленными результатами, в которых участники знали, какое лечение они получали. Поэтому мы не уверены, насколько надежны результаты. Наши выводы могут измениться, если будут проведены дополнительные исследования.
Authors' conclusions
Summary of findings
| Behavioural activation compared with CBT for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with CBT | Risk with behavioural activation | |||||
| treatment efficacy | Study population | RR 0.99 | 601 | ⊕⊕⊕⊝ | ||
| 62 per 100 | 61 per 100 | |||||
| treatment acceptability up to 6 months (4‐16 weeks) | Study population | RR 1.03 | 1195 | ⊕⊕⊝⊝ | ||
| 23 per 100 | 24 per 100 | |||||
| depression symptoms (continuous) up to 6 months (4‐16 weeks) | see comment | SMD 0.12 higher | 1205 | ⊕⊕⊕⊝ | Measured with BDI, HRSD, CES‐D, PHQ‐9, HSCL‐25. SMD 0.12 represents a difference between groups of 1.31 points on the BDI and 0.66 points on the HRSD favouring CBT. | |
| quality of life (continuous) | see comment | SMD 0.04 higher (0.20 lower to 0.28 higher) | 268 (2 RCTs) | ⊕⊕⊕⊝ | Measured with SF‐36 physical component and WHOQOL physical component. SMD 0.04 represents a small effect. | |
| social adjustment and functioning (continuous) | see comment | SMD 0.13 lower (0.50 lower to 0.24 higher) | 111 (2 RCTs) | ⊕⊝⊝⊝ | Measured with Social Adjustment Scale and Sheehan Disability Scale. SMD 0.13 represents a small effect. | |
| anxiety symptoms (continuous) up to 6 months (4‐16 weeks) | see comment | SMD 0.03 lower | 646 | ⊕⊕⊕⊝ | Measured with BDI, HSCL‐25, PHQ‐9. SMD 0.03 represents a small effect. | |
| adverse events | 1 study no adverse events, 1 study three serious adverse events in the behavioural activation arm (2 overdose, 1 self‐harm) and eight serious adverse events in the CBT arm (7 overdose, 1 self‐harm). | 398 (2 RCTs) | ⊕⊕⊕⊝ | Any adverse event summarised narratively. | ||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Majority of domains high or unclear risk of bias. High risk for conflict of interest, blinding of participants and personnel, and incomplete outcome data. Downgraded by one level for high risk of bias. 2 No blinding of participants. Reporting bias unclear because protocol or trial registration missing in nine studies and high risk of bias in one study. Potential conflict of interest in four4 studies. High risk of attrition bias in seven studies. Downgraded by one level for high risk of bias (not two levels because trials with higher weight are generally at lower risk of bias). 3 Seven out of 12 studies wide confidence intervals, due to small sample sizes and low rates of dropout in both groups. Downgraded by one level for imprecision. 4 No blinding of participants. 6/15 studies no blinding of outcome assessors. 13/15 selective reporting domain unclear. Downgraded by one level for high risk of bias. 5 Risk of performance and attrition bias and potential conflict of interest. Downgraded by one level for high risk of bias. 6 Two small studies with serious risk of bias across domains (attrition bias, reporting bias, potential conflicts of interest). Downgraded by one level for imprecision and two levels for high risk of bias. 7 One study all domains unclear or high; high risk of bias for randomisation, allocation, and blinding of participants and personnel. One study with risk of performance and attrition bias and potential conflict of interest. Downgraded one level for high risk of bias.Two studies with most domains unclear or high have little weight in the analyses. 8 Various domains high risk of bias in all studies. Attrition bias high in both studies; dropout may be related to adverse events. Downgraded one level for high risk of bias. | ||||||
| Behavioural activation compared with third‐wave CBT for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with third‐wave CBT | Risk with behavioural activation | |||||
| treatment efficacy up to 6 months (4‐8 weeks) | Study population | RR 1.10 (0.91 to 1.33) | 98 | ⊕⊕⊝⊝ | ||
| 74 per 100 | 81 per 100 (67 to 98) | |||||
| treatment acceptability up to 6 months (4‐8 weeks) | Study population | RR 0.84 (0.33 to 2.10) | 147 | ⊕⊕⊝⊝ | ||
| 12 per 100 | 10 per 100 (4 to 25) | |||||
| depression symptoms (continuous) up to 6 months (4‐8 weeks) | see comment | SMD 0.14 lower | 147 | ⊕⊕⊝⊝ | Measured with BDI and HRSD. SMD 0.14 represents a difference between groups of 1.53 points on the BDI and 0.77 points on the HRSD favouring BA. | |
| quality of life (continuous) | mean score 1.13 | MD 0.02 higher (0.96 lower to 1.00 higher) | 81 (1 RCT) | ⊕⊕⊝⊝ | Measured with Quality of Life Inventory. | |
| anxiety symptoms (continuous) | see comment | MD 0.69 higher (0.68 lower to 2.06 higher) | 147 | ⊕⊕⊝⊝ | Measured with BAI. | |
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Evidence of selective reporting and conflict of interest in both trials, in addition to other domains with risk of bias. Downgraded one level for high risk of bias. 2 Two small studies with wide confidence intervals. Downgraded one level for imprecision. 3 Ten domains with high risk of bias across three studies, including blinding, allocation concealment, and selective reporting. Treatment acceptability may be affected by lack of blinding and allocation concealment in particular. Downgraded one level for high risk of bias. 4 Three small studies with wide confidence intervals. Downgraded one level for imprecision. 5 One small study with three domains at high risk of bias. Downgraded one level for imprecision and one level for high risk of bias. Because only one study was included, this outcome could not be assessed for consistency of results. | ||||||
| Behavioural activation compared with humanistic therapy for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with humanistic therapy | Risk with behavioural activation | |||||
| treatment efficacy | Study population | RR 1.84 | 46 | ⊕⊕⊝⊝ | Number needed to treat to achieve one beneficial outcome is 2.5. | |
| 48 per 100 | 88 per 100 | |||||
| treatment acceptability | Study population | RR 1.06 | 96 | ⊕⊝⊝⊝ | ||
| 25 per 100 | 26 per 100 | |||||
| depression symptoms | mean score between 10 and 15 | MD 3.75 lower (6.72 lower to 0.78 lower) | 93 | ⊕⊕⊕⊝ | Measured with BDI. | |
| quality of life (continuous) | mean score 1.2 | MD 0.80 higher (0.12 lower to 1.72 higher) | 50 (1 RCT) | ⊕⊕⊝⊝ | Measured with Quality of Life Inventory. | |
| anxiety symptoms (continuous) | mean score 9.7 | MD 1.30 lower (6.10 lower to 3.50 higher) | 50 (1 RCT) | ⊕⊕⊝⊝ | Measured with BAI. | |
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Both small studies with several domains high risk of bias or unclear. Risk of attrition bias in both studies and reporting bias in one study may affect treatment efficacy outcome. Downgraded one level for high risk of bias and one level for imprecision. 2 Many risk of bias domains unclear in one of the studies. Risk of attrition and reporting bias in the other study. Downgraded one level for high risk of bias and one level for imprecision. 3 One study is religious behavioural activation rather than the conventional behavioural activation intervention. Downgraded one level for indirectness. 4 Two out of three studies mostly high and unclear risk of bias domains. Downgraded one level for high risk of bias. 5 One small study with most domains of the risk of bias tool unclear due to lack of information. Downgraded one level for high risk of bias and one level for imprecision. Because only one study was included, this outcome could not be assessed for consistency of results. | ||||||
| Behavioural activation compared with psychodynamic for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with psychodynamic | Risk with behavioural activation | |||||
| treatment efficacy up to 6 months (12 weeks) | Study population | RR 1.21 | 60 | ⊕⊝⊝⊝ | ||
| 47 per 100 | 56 per 100 | |||||
| depression symptoms | mean score 10 | MD 1.10 lower (4.35 lower to 2.15 higher) | ‐ | 60 | ⊕⊝⊝⊝ | Measured with HRSD |
| social adjustment and functioning | mean score 69 | MD 2.10 higher (4.92 lower to 9.12 higher) | ‐ | 60 | ⊕⊝⊝⊝ | Measured with Global Assessment Scale and Social Adjustment Scale (measures combined) |
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Four 'Risk of bias' domains unclear due to lack of information. Patients excluded from study for lack of adherence and because of dissatisfaction with treatment. This may influence outcomes treatment efficacy, depression symptoms, and social adjustment and functioning. All other 'Risk of bias' domains high or unclear risk. Downgraded two levels for high risk of bias. 2 Only one study with small sample size. Downgraded one level for imprecision. Because only one study was included, this outcome could not be assessed for consistency of results. | ||||||
| Behavioural activation compared with interpersonal, cognitive analytic, integrative for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with interpersonal, cognitive analytic, integrative | Risk with behavioural activation | |||||
| treatment acceptability | Study population | RR 0.84 | 123 | ⊕⊝⊝⊝ | ||
| 16 per 100 | 14 per 100 | |||||
| depression symptoms (continuous) up to 6 months (4‐12 weeks) | see comment | SMD 0.16 lower | 103 | ⊕⊝⊝⊝ | Measured with BDI, Zung rating scale, and HRSD. SMD 0.16 represents a difference between groups of 1.75 points on the BDI and 0.88 points on the HRSD favouring BA. | |
| social adjustment and functioning (continuous) | mean score 79 | MD 3.92 lower (16.78 lower to 8.93 higher) | 39 (1 RCT) | ⊕⊝⊝⊝ | Measured with Global Assessment Scale. | |
| anxiety symptoms (continuous) up to 6 months (4 weeks) | mean score 48 | MD 0.39 lower (11.78 lower to 11.00 higher) | 15 (1 RCT) | ⊕⊝⊝⊝ | Measured with the anxiety scale of the Multiple Affect Adjective Check List. | |
| adverse events | 2 suicide attempts and 1 case of suicidal thoughts in comparator arm; no adverse events in behavioural activation arm. | 24 | ⊕⊕⊝⊝ | Any adverse event summarised narratively. | ||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Only 2 low risk of bias domains across four studies. High risk of bias for randomisation and allocation concealment in 2/4 studies. Downgraded two levels for high risk of bias. Downgraded one level for imprecision because of wide confidence intervals. 2 One small study with high risk of bias across multiple domains. Downgraded two levels for high risk of bias and one level for imprecision. Because only one study was included, this outcome could not be assessed for consistency of results. 3 One very small study, so adverse events reported may not apply to a wider population receiving treatment. High risk of bias included lack of blinding and potential attrition bias and selective reporting may influence reporting of adverse events. Downgraded one level for imprecision and one level for high risk of bias. | ||||||
| Behavioural activation compared with waiting list for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with waiting list | Risk with behavioural activation | |||||
| treatment efficacy | Study population | RR 2.14 (0.90 to 5.09) | 26 (1 RCT) | ⊕⊕⊝⊝ | ||
| 33 per 100 | 71 per 100 (30 to 100) | |||||
| treatment acceptability | Study population | RR 1.17 | 359 | ⊕⊕⊕⊝ | Three studies could not be included in meta‐analyses; no dropouts. | |
| 12 per 100 | 14 per 100 | |||||
| depression symptoms | see comment | SMD 1.04 lower | 619 | ⊕⊕⊝⊝ | Measured with BDI, HRSD, MADRS, PHQ‐9, HSCL‐25. SMD 1.04 represents a difference between groups of 11.37 points on the BDI and 5.75 points on the HRSD favouring BA. | |
| quality of life (continuous) | mean score 0.75 | MD 0.03 higher (0.70 lower to 0.76 higher) | 80 (1 RCT) | ⊕⊕⊝⊝ | Measured with quality of life inventory. | |
| anxiety symptoms | see comment | SMD 0.91 lower | 424 | ⊕⊕⊝⊝ | Measured with BAI, Trait Anxiety Scale, and GAD‐7. SMD 0.91 represents a large effect. | |
| adverse events | see comment | 0 (1 RCT) | see comment | Any adverse event summarised narratively. No adverse events. | ||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 One very small study with high risk of bias for five domains including allocation concealment and selective reporting. Downgraded one level for imprecision and one level for high risk of bias. 2 Most studies high or unclear risk of bias with regard to blinding of participants and outcome assessors, selective reporting, and various other risks of bias: no baseline characteristics reported, potential conflicts of interest. Downgraded one level for high risk of bias. 3 Only domain mostly scoring low risk of bias across studies (7/12) is random sequence generation. Blinding of outcome assessors unclear or high risk of bias in all but two studies. Downgraded one level for high risk of bias. 4 Larger effects reported by smaller studies; smaller studies favouring waiting list are absent. Downgraded one level for risk of publication bias. 5 One study with high risk of bias for three domains including potential conflict of interest. Downgraded one level for high risk of bias and one level for imprecision. Because only one study was included, this outcome could not be assessed for consistency of results. 6 All studies majority of domains unclear or high risk of bias. Some problems with randomisation and allocation concealment. Downgraded one level for high risk of bias. 7 Two studies reporting large effect in favour of behavioural activation while three find no difference between study arms. Downgraded one level for inconsistency. | ||||||
| Behavioural activation compared with placebo for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with behavioural activation | |||||
| treatment acceptability | Study population | RR 0.72 (0.31, 1.67) | 96 (1 RCT) | ⊕⊕⊝⊝ | ||
| 23 per 100 | 16 per 100 (7 to 38) | |||||
| depression symptoms | see comment | SMD 0.18 lower | 108 | ⊕⊕⊝⊝ | Measured with HRSD and Depression Adjective Checklist. SMD 0.18 represents a difference between groups of 1.97 points on the BDI and 1.00 point on the HRSD favouring BA. | |
| adverse events | Various physical side effects from placebo. | 96 | ⊕⊕⊝⊝ | Any adverse event summarised narratively. | ||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Incomplete outcome data influences reporting of dropouts; downgraded one level for high risk of bias. Downgraded one level for imprecision due to large confidence intervals resulting from relatively few dropouts. Because only one study was included, this outcome could not be assessed for consistency of results. 2 One study with poor reporting, which may indicate high risk of bias. Downgraded one level for high risk of bias. 3 Two small studies; one with large confidence intervals. Downgraded one level for imprecision. 4 Incomplete outcome data and potential conflict of interest may have influenced reporting of adverse events. Downgraded one level for high risk of bias. Downgraded one level for imprecision as 96 participants would not be sufficient to measure less frequently occurring side effects. | ||||||
| Behavioural activation compared with medication for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with medication | Risk with behavioural activation | |||||
| Study population | RR 1.77 (1.14 to 2.76) | 141 (1 RCT) | ⊕⊕⊕⊝ | |||
| treatment efficacy up to 6 months (16 weeks) | 28 per 100 | 49 per 100 (31 to 76) | ||||
| treatment acceptability | 34 per 100 | 18 per 100 | RR 0.52 | 243 | ⊕⊕⊕⊝ | |
| depression symptoms | mean change in score between ‐8 and ‐14 | mean difference 1.42 lower (4.80 lower to 1.96 higher) | 180 | ⊕⊕⊝⊝ | Measured with HRSD. | |
| adverse events | Various physical side effects from antidepressant medication. One suicide in antidepressant arm. | 143 | ⊕⊕⊕⊝ | Any adverse event summarised narratively. | ||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 The most concerning issue relating to risk of bias was the large number of dropouts from the medication study arm in particular. Downgraded one level for high risk of bias. Because only one study was included, this outcome could not be assessed for consistency of results. 2 Incomplete outcome data for both studies. No blinding of participants. Potential conflict of interest for one study. Downgraded one level for high risk of bias. 3 Downgraded one level for imprecision; large variation in confidence interval, crossing zero. 4 Incomplete outcome data and potential conflict of interest may have influenced reporting of adverse events. Downgraded one level for high risk of bias. | ||||||
| Behavioural activation compared with no treatment for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with no treatment | Risk with behavioural activation | |||||
| treatment acceptability | Study population | RR 0.97 | 187 | ⊕⊕⊕⊝ | ||
| 9 per 100 | 9 per 100 | |||||
| depression symptoms | see comment | MD 6.10 lower (7.87 lower to 4.33 lower) | 187 | ⊕⊕⊕⊝ | Measured with BDI | |
| quality of life (continuous) | mean score 0.9 | MD 0.07 higher (0.03 higher to 0.11 higher) | 118 (1 RCT) | ⊕⊕⊕⊕ | Measured with EQ‐5D | |
| anxiety symptoms | mean score 11 | MD 5.50 lower (10.01 lower to 0.99 lower) | 30 (1 RCT) | ⊕⊕⊝⊝ | Measured with BAI | |
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 High risk of bias for blinding of participants (3/3), conflict of interest (1/3), and no baseline characteristics reported (1/3). Downgraded one level for high risk of bias. 2 One study mostly low risk of bias, one study mostly unclear risk of bias. Downgraded one level for high risk of bias. 3 Because only one study was included, this outcome could not be assessed for consistency of results. 4 One small study with four domains of bias unclear and two high risk of bias; performance bias and potential conflict of interest. Downgraded one level for high risk of bias and one level for imprecision. Because only one study was included, this outcome could not be assessed for consistency of results. | ||||||
| Behavioural activation compared with treatment as usual for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with treatment as usual | Risk with behavioural activation | |||||
| treatment efficacy | Study population | RR 1.40 | 1533 | ⊕⊕⊕⊝ | Number needed to treat to achieve one beneficial outcome is 4.5 | |
| 53 per 100 | 75 per 100 | |||||
| treatment acceptability | Study population | RR 1.64 | 2518 | ⊕⊕⊕⊝ | ||
| 6 per 100 | 11 per 100 | |||||
| depression symptoms (continuous) up to 6 months (5‐12 weeks) | see comment | SMD 0.78 lower | 2208 | ⊕⊕⊝⊝ | Measured with PHQ‐9, CES‐D, BDI, HRSD, and GDS. SMD 0.78 represents a difference between groups of 8.53 points on the BDI and 4.31 points on the HRSD. | |
| quality of life (continuous) | see comment | SMD 0.97 higher | 1299 | ⊕⊝⊝⊝ | Measured with SF‐12 physical component and WHOQOL. SMD 0.97 represents a large effect. | |
| social adjustment and functioning (continuous) | see comment | SMD 1.27 lower (1.74 lower to 0.84 lower) | 88 (2 RCTs) | ⊕⊕⊝⊝ | Measured with Work and Social Adjustment Scale (WSAS) and Sheehan Disability Scale. SMD 1.27 represents a large effect. | |
| anxiety symptoms | see comment | SMD 0.33 lower | 1063 | ⊕⊕⊕⊝ | Measured with GAD‐7 and BAI. SMD 0.33 represents a small effect. | |
| adverse events | behavioural activation arm: 103 events. treatment as usual arm: 107 events. | 1471 (3 RCTs) | ⊕⊕⊕⊝ | Any adverse event summarised narratively | ||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Mostly low risk of bias for sequence generation, allocation concealment, and selective reporting. Mostly high risk of bias only for blinding of participants and personnel. Some evidence of incomplete outcome data. Downgraded one level for high risk of bias. 2 Several studies with incomplete outcome data and potential conflict of interest. Randomisation and allocation concealment largely low risk of bias. Downgraded one level for high risk of bias. 3 Pooled estimate is influenced by large effect in one small study. Downgraded one level for inconsistency. 4 Pooled estimate is driven by one study with large effect favouring behavioural activation. Wide confidence interval. Downgraded one level for inconsistency and one level for imprecision. 5 One small study with three high risk of bias domains including incomplete outcome data. Other study unclear risk of selection bias, and high risk for attrition bias, reporting bias, and conflict of interest. Downgraded two levels for high risk of bias. Although studies are small, estimates are consistent. 6 No blinding of participants/personnel and outcome assessors in 3/4 studies. Evidence of attrition bias (2/4), performance bias (3/4) and potential conflict of interest (3/4). No evidence of selection bias in 2/4 studies, other two studies some information missing (unclear). Downgraded one level for high risk of bias. 7 One out of three studies with selective reporting, attrition bias, and potential conflict of interest. One study potential conflict of interest. Downgraded one level for high risk of bias. | ||||||
Background
Description of the condition
Depression, when diagnosed in a clinical setting, most often refers to a major depressive disorder. It is characterised by a period of at least two weeks of depressed mood, or a persistent loss of interest or pleasure in activities which were previously considered enjoyable, or both (APA 2013). A range of symptoms may accompany these key features of depression, including weight loss or weight gain, insomnia or hypersomnia (excessive sleeping and/or sleepiness), psychomotor agitation (mental and physical restlessness) or retardation (mental and physical slowness), fatigue, loss of energy, feelings of excessive guilt and worthlessness, diminished concentration, and recurrent thoughts of death (APA 2013).
Depression is the fifth global cause of disease burden in terms of years lived with a disability (YLD), and was ranked in the top 10 of YLD in 191 out of 195 countries worldwide (Vos 2017). In 2014, 7.1% of the population living in the 28 countries of the European Union was estimated to report depression, with higher rates reported by women and by Europeans living in cities. Prevalence rates of self‐reported depression varied from 4% in 15‐ to 24‐year‐olds to 10% in those aged 75 and over (Eurostat 2014).
Depression has a long‐lasting impact on patients, their families, and wider society. It is associated with marked personal and societal economic losses due to healthcare costs for mental and comorbid physical healthcare, reduced productivity in the workplace, and years of life lost (Greenberg 2015). A meta‐analysis of data from 35 countries found a 52% increased risk of mortality, after adjusting for publication bias (Cuijpers 2014).
Description of the intervention
Clinical guidelines recommend pharmacological and psychological interventions, alone or in combination, in the treatment of mild to moderate depression in adults (NICE 2009).
The prescribing of antidepressants has increased dramatically in many Western countries over the past 20 years, mainly with the advent of selective serotonin reuptake inhibitors (SSRIs) and other agents such as serotonin–noradrenaline reuptake inhibitors (SNRIs) and noradrenalinergic and specific serotonergic antidepressants (NaSSAs) (Ilyas 2012). Antidepressants remain the mainstay of treatment for moderate to severe depression in healthcare settings, whereas for subthreshold depressive symptoms (not meeting the threshold for clinical diagnosis) or mild depression, low‐intensity psychosocial therapy and psychological therapies are recommended (NICE 2009).
Whilst antidepressants are proven to be effective for the acute treatment of depression for some people (Arroll 2009; Magni 2013; Cipriani 2009a; Cipriani 2009b; Cipriani 2010; Guaiana 2007), adherence rates remain very low (Hunot 2007; van Geffen 2009), in part because of patients' concerns about side effects and dependency (Hunot 2007; Fawzi 2012). Not adhering to antidepressant medication is related to relapse and/or recurrence, hospital visits and hospitalisation, worsening of depression symptoms, and a lower likelihood of recovery (Ho 2016). Furthermore, surveys consistently demonstrate patients' preference for psychological therapies over antidepressant treatment (Churchill 2000; McHugh 2013; Riedel‐Heller 2005). Therefore, psychological therapies can be an important alternative intervention or an additional treatment for depressive disorders.
A diverse range of psychological therapies is available for the treatment of depression. Psychological therapies may be broadly categorised into four separate philosophical and theoretical schools, comprising psychoanalytic/dynamic (Freud 1949; Jung 1963; Klein 1960), behavioural (Skinner 1953; Watson 1924; Wolpe 1958), humanistic (Maslow 1943; May 1961; Rogers 1951), and cognitive approaches (Beck 1979; Lazarus 1971). Each of these four schools incorporates several different and overlapping psychotherapeutic approaches. Some psychotherapeutic approaches, such as cognitive‐analytic therapy (CAT) (Ryle 1990), explicitly integrate components from several theoretical schools. Other approaches, such as interpersonal therapy (IPT) for depression (Klerman 1984), have been developed to address characteristics considered specific to the disorder of interest.
Behavioural therapy is a term that has been used to describe a broad range of therapies using principles of operant conditioning, in which behaviours are modified through learning. It became a dominant force in the 1950s, drawing on the work of Skinner 1953, Wolpe 1958, and Eysenck 1960. Behavioural therapy emphasises the role of environmental cues in influencing the adoption and maintenance of behaviours (Nelson‐Jones 1990) and, in contrast with psychoanalysis, was developed though experimentally‐ rather than theoretically‐derived principles (Rachman 1997).
With the advent of cognitive therapy in the 1970s, behavioural therapy approaches based purely on operant (learning from the consequences of behaviours) and respondent (responsive behaviour as a result of a stimulus) principles became regarded as insufficient. However, the interest in the feasibility of behavioural treatments for depression has since been renewed (Dimidjian 2011; Ekers 2014; Hopko 2003a). The term behavioural activation appears to have been used for the first time in 1990, as a description of the behavioural components in cognitive therapy (Hollon 1990). Jacobson showed that the behavioural component of cognitive‐behavioural therapy (CBT) was as effective as the full package of CBT, and developed a new and more comprehensive model of behavioural activation that would be amenable to dissemination (Jacobson 1996; Jacobson 2001). It would appear that 'behavioural activation' has now become the commonly adopted description, and we will use this term in the rest of this review to refer to the intervention (Martell 2010).
How the intervention might work
Skinner proposed that depression was associated with an interruption in established sequences of healthy behaviour that were previously positively reinforced by the social environment and were based on operant conditioning principles (in which behaviour patterns are learnt, rather than instinctive) (Skinner 1953). In subsequent expansions of this model, reduction of positively reinforced healthy behaviours has also been attributed to a decrease in the number and range of reinforcing stimuli available to the individual, lack of skill in obtaining positive reinforcement (Lewinsohn 1974), increased frequency of punishment, or a combination of two, or all of these (Lewinsohn 1984).
Behavioural activation can be defined as a brief psychotherapeutic approach that seeks to change the way a person interacts with their environment, aiming to:
-
increase access to positive reinforcers of healthy behaviours;
-
reduce avoidance behaviours that limit access to positive reinforcement;
-
understand and address barriers to activation.
Treatments are collaborative and focused on the present. Many differing techniques are incorporated into treatment; however all use self‐monitoring of a mood‐environment link and scheduling of new or adaptive behaviours to meet targets (Kanter 2010). In doing so, the therapy helps people to make contact with potentially reinforcing experiences (Jacobson 2001).
The original model of behavioural activation, developed by Jacobson, was defined primarily by the elimination of cognitive intervention elements (Dimidjian 2006). On the basis of its original design, behavioural activation model components commonly include developing a shared treatment rationale; promoting access to meaningful events, activities, and consequences; activity scheduling; developing social skills; and self‐monitoring links between behaviour and mood. In some cases the use of some form of problem‐solving or functional analysis are added to understand, consider and overcome any potential barriers to the scheduling of activities. In contrast to CBT, no attempt is made to directly change cognitions. However, behavioural activation commonly involves an exploration of how cognitive processes, such as rumination, can limit access to behaviours and events which give positive reinforcement, for example in stopping people with depression from meeting up with friends or participating in physical exercise.
Why it is important to do this review
According to the clinical guidelines produced by the National Institute for Health and Clinical Excellence (NICE), behavioural activation is one of the recommended treatment options for subthreshold depressive symptoms, mild to moderate depression, and severe depression, along with CBT and IPT. However, the guidelines acknowledge that evidence for behavioural activation is currently less robust than for the other recommended therapies (NICE 2009).
The effects of behavioural therapies for depression versus other psychological therapies were previously examined in a Cochrane Review, which reported that low‐ to moderate‐certainty evidence from 25 trials suggested that behavioural therapies and other psychological therapies were equally effective (Shinohara 2013). This Cochrane Review did not cover trials comparing behavioural therapy to treatment as usual, nor did it include the emerging literature on new treatment models of behavioural activation.
Two Cochrane Reviews of 'third‐wave' cognitive and behavioural therapies, one comparing the intervention to treatment as usual and one comparing to other therapies, identified three trials of behavioural activation for depression (Churchill 2013; Hunot 2013). The small number of trials together with the low certainty of the evidence limited the ability to draw any conclusions on effectiveness. Another systematic review of behavioural activation found evidence from 26 trials, most of them low quality, indicating that behavioural activation is more effective than a wide range of control treatments, including medication (Ekers 2014).
There is no Cochrane Review that includes all behavioural activation therapies currently used for the treatment of depression. Behavioural activation is increasingly receiving attention as a potentially cost‐effective intervention for depression, which may be delivered and implemented in settings with low‐resources or where the demand is greater than the availability of mental health practitioners to deliver more complex treatments (Richards 2016). Given this resurgence of interest, a comprehensive review of the comparative effectiveness and acceptability of behavioural activation interventions for depression is timely to inform and update clinical practice and future clinical guideline development.
Objectives
-
To examine the effects of behavioural activation compared with all other psychological therapies for depression in adults.
-
To examine the effects of behavioural activation compared with all medication for depression in adults.
-
To examine the effects of behavioural activation compared with treatment as usual/ waiting list/placebo/no treatment control conditions for depression in adults.
Methods
Criteria for considering studies for this review
Types of studies
Randomised controlled trials (RCTs) were eligible for inclusion in this review. We included trials employing a cross‐over design (whilst we acknowledge that this design is rarely used in psychological therapy trials), but we only used data from the first active treatment phase. Cluster‐RCTs and pilot RCTs were also eligible for inclusion.
Quasi‐randomised controlled trials, in which treatment assignment is decided through methods such as alternate days of the week, were not eligible for inclusion. We included trials that replaced dropouts without randomisation only when the proportion of replaced participants was less than 20%.
Types of participants
Participant characteristics
Randomised controlled trials of adults aged 18 years and over of any sex or gender were eligible for inclusion. We excluded trials that involved participants under 18 years of age.
Setting
Trials could be conducted in a primary, secondary or community setting. Trials conducted in a hospital clinic were included, but we excluded trials involving inpatients. We included trials that focused on specific populations ‐ nurses, care givers, depressed participants at a specific workplace ‐ if all participants met criteria for depression. Nursing homes in this review were considered outpatient settings, as they are places of residence.
Diagnosis
We included all trials that focused on acute phase treatment of clinically diagnosed depression or subthreshold depression.
-
We included trials adopting any standardised diagnostic criteria to define participants suffering from an acute phase unipolar depressive disorder. Accepted diagnostic criteria include Feighner criteria, Research Diagnostic Criteria and criteria of the Diagnostic and Statistical Manual of Mental Disorders, Third Edition ((DSM‐III); APA 1980), DSM‐III‐Revised ((R); APA 1987), DSM‐Fourth Edition ((IV); APA 1994), DSM‐IV‐Text Revision ((TR); APA 2000), DSM‐5 (APA 2013), and International Classification of Diseases, Tenth Edition ((ICD‐10); WHO 1992).
-
To fully represent the broad spectrum of severity of depressive symptoms encountered by healthcare professionals in primary care, we included trials that used non‐operationalised diagnostic criteria (ICD‐Ninth Edition ((ICD‐9); WHO 1978) or a validated clinician or self‐report depression symptom questionnaire, such as the Hamilton Rating Scale for Depression (HRSD) (Hamilton 1960), or the Beck Depression Inventory (BDI) (Beck 1961), to identify depression cases as based on a recognised threshold.
-
Subthreshold depression, also called subsyndromal, subclinical, or minor depression, in which people experience symptoms of depression but do not meet the threshold for diagnosis. We accepted any trials that established subthreshold depression based on the above diagnostic criteria or validated depression symptom questionnaires.
When possible, we used accepted strategies for classifying mild, moderate and severe depression on the basis of criteria used in the evidence syntheses underpinning the NICE 2009 guidelines for depression. NICE 2009 defines severity of depression in accordance with DSM‐IV (APA 1994) as follows: mild depression: few, if any, symptoms in excess of the five required to make the diagnosis, with symptoms resulting in only minor functional impairment. Moderate depression: symptoms of functional impairment between 'mild' and 'severe'. Severe depression: most symptoms, and marked interference of the symptoms with functioning. Can occur with or without psychotic symptoms.
We excluded trials focusing on chronic depression or treatment‐resistant depression (i.e. trials that list these conditions as inclusion criteria). We also excluded trials in which participants were receiving treatment to prevent relapse after a depressive episode (i.e. where participants did not have symptoms of depression at trial entry). Postnatal depression is considered a separate condition with contributing factors distinct from major depressive disorder, and we therefore excluded it.
If participants met the criteria for depression or subthreshold depression as stated above, we included trials with people described as ‘at risk of suicide’ or with dysthymia (persistent depressive disorder), or other affective disorders such as panic disorder, but otherwise we excluded these trials.
We did not include subgroup analyses of people with depression selected from people with mixed diagnoses because such trials would be susceptible to publication bias (the trial authors reported such subgroup trials because the results were 'interesting'). In other words, we included these trials only if the inclusion criteria for the entire trial satisfied our eligibility criteria.
Comorbidity
Trials involving participants with comorbid physical or common mental disorders were eligible for inclusion as long as the comorbidity was not the focus of the trial. For example, we excluded trials that focused on depression among individuals with Parkinson's disease or after acute myocardial infarction but accepted trials that may have included some participants with Parkinson's disease or with acute myocardial infarction. This decision was made because the intervention and study design may in such cases be adapted for these specific populations. A separate Cochrane Review of behavioural activation for the treatment of depression in people with physical comorbidities is to be published in 2020 (Uphoff 2019b).
Types of interventions
Experimental interventions
A previously published Cochrane Review for behavioural therapy in depression provided a framework for psychological therapies, including behavioural therapy (Shinohara 2013). Given recent developments in literature and practice regarding behavioural activation approaches, we consider behavioural activation as part of behavioural therapies, rather than being classified as a 'third‐wave' therapy. In line with the behavioural therapy review, we created the comparator categories of psychological therapies on the basis of both treatment approach (e.g. their theoretical background and the manuals they used) and content (what therapeutic techniques they mainly used or what was their area of focus). See also Appendix 1.
Behavioural activation
We included trials evaluating treatment approaches for depression that are either explicitly called 'behavioural activation', or treatments that are described using the main elements of behavioural activation for depression, such as pleasant events and activities, activity scheduling, positive reinforcement from the environment, positive interaction or re‐engagement with the environment. This means that we included behavioural therapies in the treatment group as long as they were described using the main elements of behavioural activation. Experimental interventions that contained some elements of behavioural therapy, such as CBT or problem‐solving therapy, were not eligible for inclusion.
Format of psychological therapies
Therapies delivered by therapists of all levels were eligible for inclusion. This includes: 1) psychologists or psychotherapists accredited by a professional body for psychology or psychotherapy, who completed formal training to deliver psychological therapies, 2) those who received substantial training (more than a year) but are not yet qualified, and 3) lay counsellors and non‐specialist therapists who have been specifically trained to deliver treatment according to a behavioural activation protocol.
We included computerised and self‐help interventions if they were facilitated. This means at least some element of interaction with a therapist was required.
Psychological therapies conducted on an individual or group basis were eligible for inclusion.
The number of sessions was not limited, and we accepted psychological therapies delivered in only one session.
Comparators
All comparators were accepted as long as they are not a type of behavioural activation. We categorised psychological therapies as behavioural therapy, social skills training/assertiveness training, relaxation therapy, CBT, third‐wave CBT, psychodynamic, humanistic and integrative approaches.
Behavioural therapy
We planned to include any behavioural therapies that did not contain the main elements of behavioural activation as comparators.
Social skills training/assertiveness training
The social skills training model (SST) proposes that depressed people may have difficulty initiating, maintaining and ending conversations (Jackson 1985). Because of these deficits, the individual is unable to elicit mutually reinforcing behaviour from other people in his or her environment. SST subsumes assertion and conversational skills, together with more specialised subskills such as dating and job interview skills. Different social contexts may be targeted, for example interaction with friends, family members, people at school, or at work, and interventions such as instruction, modelling, rehearsal, feedback and reinforcement are used to enable the development of new responses (Jackson 1985). As assertiveness training represents a key component of SST, we included it in the SST category.
Relaxation therapy
Relaxation training is a behavioural stress management technique that induces a relaxation response, helping to switch off the fight/flight response and causing levels of stress hormones in the bloodstream to fall. A variety of techniques may be used to induce relaxation, the most common of which is Jacobson's progressive muscle relaxation training (Bernstein 1973).
Cognitive‐behavioural therapies (CBTs)
In CBT, therapists aim to work together with people receiving treatment to understand the link between thoughts, feelings and behaviours, and to identify and modify unhelpful thinking patterns and underlying assumptions about the self, others and the world (Beck 1979). Cognitive change methods for depression are targeted at the automatic thought level in the first instance and include thought catching, reality testing and task assigning as well as generating alternative strategies (Williams 1997). Behavioural experiments are then used to re‐evaluate underlying beliefs and assumptions (Bennett‐Levy 2004). We categorised these therapies into six subcategories: cognitive therapy, rational emotive behaviour therapy, problem‐solving therapy, self‐control therapy, a coping with depression course and other CBTs.
'Third‐wave' cognitive and behavioural therapies (third‐wave CBTs)
Third‐wave CBT approaches have been developed more recently and now exist alongside established therapies such as CBT. Rather than focusing on the contents of thoughts, these therapies tend to focus on the process and functions of thoughts and an individual's relationship with thoughts and emotions. This may include suppressing or avoidance of emotions, thoughts, and bodily sensations (Hofmann 2008). Third‐wave approaches use strategies relating to mindfulness, emotions, acceptance, relationships, values, goals, and understanding the thinking process, to bring about changes in thinking (Hayes 2007). Drawing from psychodynamic and humanistic principles, third‐wave CBT approaches place great emphasis on use of the therapeutic relationship. We categorised these therapies into subcategories: acceptance and commitment therapy, compassionate mind training, functional analytic psychotherapy, metacognitive therapy, mindfulness‐based cognitive therapy, dialectical behaviour therapy and other third‐wave CBTs.
Psychodynamic therapies
Grounded in psychoanalytic theory (Freud 1949), psychodynamic therapy (PD) uses the therapeutic relationship to explore and resolve unconscious conflict through transference (projection of feelings on to the therapist) and interpretation, with development of insight and character change (within certain boundaries) as therapeutic goals, and relief of symptoms as an indirect outcome. Brief therapy models have been devised by Malan 1963, Mann 1973 and Strupp 1984. We categorised these therapies into four subcategories: drive/structural model (Freud), relational model (Strupp, Luborsky), integrative analytic model (Mann) and other psychodynamic therapies.
Humanistic therapies
Contemporary models of humanistic therapies differ from one another somewhat in clinical approach, but all focus attention on the therapeutic relationship (Cain 2002), within which therapist ‘core conditions’ of empathy, genuineness, and unconditional acceptance and support (positive regard) (Rogers 1951), are regarded as cornerstones for facilitating insight and change. We categorised these therapies into seven subcategories: person‐centred therapy (Rogerian), gestalt therapy, experiential therapies, transactional analysis, existential therapy, non‐directive/supportive therapies, and other humanistic therapies.
Interpersonal, cognitive analytic and other integrative therapies
Integrative therapies are approaches that combine components of different psychological therapy models. Integrative therapy models include interpersonal therapy (IPT) (Klerman 1984), cognitive analytic therapy (CAT; (Ryle 1990)), and Hobson’s conversational model (Hobson 1985), manualised as psychodynamic interpersonal therapy (Shapiro 1990). With its focus on the interpersonal context, IPT was developed to specify what was thought to be a set of helpful procedures commonly used in psychotherapy for depressed outpatients (Weissman 2007), drawing in part from attachment theory (Bowlby 1980), and cognitive‐behavioural therapy within a set timeframe (time‐limited). CAT, also devised as a time‐limited psychotherapy, integrates components from cognitive and psychodynamic approaches. The conversational model integrates psychodynamic, interpersonal and person‐centred model components.
Counselling interventions traditionally draw from a wide range of psychological therapy models, including person‐centred, psychodynamic and cognitive‐behavioural approaches, applied in combination, according to the theoretical orientation of practitioners (Stiles 2008). Therefore, we usually included trials of counselling with integrative therapies. However, if the counselling intervention consists of a single discrete psychological therapy approach, we categorised it as such, even if the intervention is referred to as 'counselling'. If the intervention was manualised, this would inform our classification.
Motivational interviewing and other forms of integrative therapy approaches are also included in this category.
Waiting list
Participants are randomly assigned to the active intervention group or control group, and they will either receive the intervention first or be assigned to a waiting list until all participants in the intervention group have received the intervention. During the course of the trial, people on the waiting list can receive any appropriate medical care.
Attention placebo
We define this as a control condition that is regarded as inactive by both researchers and participants in a trial.
Psychological placebo
We define this as a control condition in a trial that is regarded by researchers as inactive but is regarded by participants as active (also called placebo therapy or sham treatment).
Medication
All medication prescribed with the goal to treat depression, most commonly antidepressants; any dose, route of administration, duration, and frequency.
Medical placebo
All types of medical placebos or 'sugar pills'.
No treatment
Trial participants not receiving any treatment for depression during the course of the trial.
Treatment as usual
Treatment as usual, standard care, or usual care would be any appropriate medical care during the course of the study. This may for example involve monitoring of the person receiving treatment, regular check‐ups, no treatment, or any type of treatment. What constitutes treatment as usual will depend on the setting and healthcare system in which the study was conducted. If a study arm fitted clearly in any of the above categories, for example 'no treatment' or a type of psychological therapy, we categorised it as such.
Excluded interventions
We excluded from the review trials of long‐term, continuation, or maintenance therapy interventions designed to prevent relapse of depression or to treat chronic depressive disorders. Similarly, we excluded trials of interventions designed to prevent a future episode of depression.
We excluded psychological therapy models based on social constructionist principles (that focus on the ways in which individuals and groups participate in the construction of their perceived social reality), including couples therapy, family therapy, solution‐focused therapy (de Shazer 1988), narrative therapy, personal construct therapy, neuro‐linguistic programming and brief problem‐solving (Watzlavick 1974). These therapies work with patterns and dynamics of relating within and between family, social and cultural systems to create a socially constructed framework of ideas (O'Connell 2007), rather than focusing on an individual's reality. A previously published Cochrane Review on couples therapy for depression has recently been updated (Barbato 2018), and a review of family therapy for depression is to be updated (Henken 2007).
Types of outcome measures
Primary outcomes
-
Treatment efficacy: the number of participants who responded to treatment, as determined by changes in scores for Beck Depression Inventory (BDI; Beck 1961), Hamilton Rating Scale for Depression (HRSD; Hamilton 1960), or Montgomery‐Asberg Depression Rating Scale (MADRS; Montgomery 1979), or in scores from any other validated depression scale. Many trials define response as 50% or greater reduction on BDI, HRSD, etc., with some trials defining response using Jacobson's Reliable Change Index; we accepted the trial authors' original definition. If trials reported multiple measures of treatment efficacy, we prioritised remission over clinically significant improvement, and recovery or remission over response.
-
Treatment acceptability: the number of participants who dropped out of the study for any reason after being randomised and allocated to a study arm.
Secondary outcomes
-
Improvement in depression symptoms, based on a continuous outcome of group mean scores at the end of treatment using BDI, HAM‐D, MADRS or any other validated depression scale
-
Quality of life, as assessed with the use of validated measures such as Short Form (SF)‐36 (Ware 1993), Health of the Nation Outcome Scales (HoNOS; Wing 1994), EuroQol (Brooks 1995), and World Health Organization Quality of Life (WHOQOL; WHOQL 1998).
-
Social adjustment and social functioning, including Global Assessment of Function (GAF) (Luborsky 1962) scores.
-
Improvement in anxiety symptoms, as measured using a validated continuous scale, either assessor‐rated, such as the Hamilton Anxiety Scale (HAM‐A) (Hamilton 1959), or self‐report, including the Trait subscale of the Spielberger State‐Trait Anxiety Inventory (STAI‐T) (Spielberger 1983), and the Beck Anxiety Inventory (BAI) (Beck 1988).
-
Adverse effects, such as counts of completed suicides, attempted suicides, or worsening of symptoms were summarised in narrative form.
Management of time points
We summarised and categorised post‐treatment outcomes and outcomes at each reported follow‐up point as follows: short term (up to six months post‐treatment), medium term (seven to 12 months post‐treatment) and long term (longer than 12 months). If data at multiple time points were available within one of our categories, we used the latest time point.
Search methods for identification of studies
Electronic searches
The Cochrane Common Mental Disorders' Information Specialist conducted searches on 17 January 2020 (and a previous search on 17 January 2019) in the following bibliographic databases using relevant subject headings (controlled vocabularies) and search syntax, appropriate to each resource. The search strategies were designed to identify RCTs of 'behavioural activation', or the main elements of behavioural activation for depression in participants with clinically diagnosed depression or subthreshold depression.
-
Cochrane Common Mental Disorders Trials Register (CCMD‐CTR); all available years (Appendix 2).
-
Cochrane Central Register of Controlled Trials (CENTRAL; current issue).
-
Ovid MEDLINE (1946 onwards; Appendix 3).
-
Ovid Embase (1980 onwards).
-
Ovid PsycINFO (1806 onwards).
We did not apply any restrictions on date, language or publication status to the searches.
We searched international trials registries via the World Health Organization's trials portal (ICTRP) and ClinicalTrials.gov to identify unpublished or ongoing trials.
We also searched for any relevant retraction statements and errata in January 2020.
Searching other resources
Grey literature
We searched the following sources of grey literature (primarily for dissertations and theses) on 17 January 2020:
-
Open Grey (www.opengrey.eu/);
-
ProQuest Dissertations & Theses Global (www.proquest.com/products-services/pqdtglobal.html);
-
DART‐Europe E‐theses Portal (www.dart-europe.eu/);
-
EThOS ‐ the British Libraries e‐theses online service (ethos.bl.uk/);
-
Open Acces Theses and Dissertations (oatd.org).
Reference lists
We checked the reference lists of all included trials and relevant systematic reviews to identify additional trials missed from the original electronic searches (e.g. unpublished or in‐press citations).
Personal communication
We contacted trial authors and subject experts for information on unpublished or ongoing trials, or to request additional trial data.
Data collection and analysis
Selection of studies
Two review authors independently examined each title and abstract obtained through the search strategy (EU, LR, SD, ESo). We then obtained full articles of all trials identified by any one of the review authors and two review authors independently assessed full‐texts according to the criteria relating to characteristics of the studies, participants, and interventions (EU, LR, SD, ESo). We discussed reasons for disagreement with a third reviewer (DE, DR, RC), and contacted external experts or trial authors if necessary in order to reach agreement. We recorded reasons for excluding records at this stage. For all included studies, we linked multiple reports from the same study. We presented a PRISMA flow diagram to show the process of study selection (Moher 2009).
Data extraction and management
Two review authors independently extracted data from each trial (EU, LR, ESa, ESo). These review authors discussed any disagreement with an additional review author (DE, RC), and, when necessary, contacted the authors of the trials for further information.
We extracted and entered information for the following categories into Covidence data extraction forms: trial design, source of funding, study population, interventions and comparators, outcomes and sample size.
Assessment of risk of bias in included studies
We assessed risk of bias for each included trial using the Cochrane Collaboration's 'Risk of bias' tool (Higgins 2016), which considers the following domains.
-
Risk of bias arising from the randomisation process, including allocation and randomisation
-
Risk of bias due to deviations from the intended interventions, including blinding of participants and people delivering the interventions
-
Incomplete outcome data
-
Risk of bias in measurement of the outcome, including blinding of outcome assessors
-
Selective outcome reporting
-
Other bias
In the assessment of risk of attrition bias (domain 5), we considered the amount of missing outcome data in each study arm and judged whether these data were likely to be missing at random.
In the 'other bias' domain we considered any other problems with a study that may lead to bias, including the following items specific to psychological therapy trials.
-
Treatment fidelity: was the therapy monitored against a manual or a scale through audiotapes or videotapes?
-
Researcher allegiance/conflict of interest: did the researcher have a vested interest for or against the therapies under examination?
-
Therapist allegiance/conflict of interest: did the therapist have a vested interest for or against the therapies provided?
For cluster‐RCTs and cross‐over trials, we used the templates specifically designed to assess these types of trials, with the same five domains.
We judged the risk of bias for each domain within and across trials, and categorised this as low, unclear, or high risk of bias.
Two review authors independently assessed the risk of bias in included trials (EU, LR, ESa, ESo) and discussed any disagreements with a third review author (EU, LR, ESa, ESo, RC, DE). Where necessary, we contacted trial authors for further information. We presented all 'Risk of bias' data graphically, and narratively in the text.
Measures of treatment effect
Continuous outcomes
Where trials used the same outcome measure for comparison, we pooled data by calculating the mean difference (MD) and 95% confidence intervals (95% CIs). When trials used different measures to assess the same outcome, we pooled data calculating the standardised mean difference (SMD) and 95% 95% CIs. We used both endpoint data and change from baseline data, depending on availability. If both were available, we used endpoint data. In accordance with the Cochrane Handbook, endpoint and change from baseline data were combined in one meta‐analysis but included in different subgroups (Schünemann 2017a).
An SMD or MD of zero means that the intervention and control groups have equivalent treatment effects. We anticipated that, for most measures, a lower score will indicate greater improvement. For example, a lower score on depression symptom instruments such as the Hamilton Rating Scale for Depression (HRSD), Beck Depression Inventory (BDI) or Patient Health Questionnaire (PHQ‐9) indicates an improvement in symptoms. In these cases, an SMD or MD less than zero indicates that the intervention has a greater effect than the control. An SMD or MD greater than zero indicates that the intervention has a smaller effect than the control. Interpretation of the SMD and MD is reversed in cases where a greater continuous score indicates greater improvement.
To facilitate interpretation of results in terms of their clinical relevance, we expressed SMDs for continuous outcomes in terms of units on a commonly used participant‐rated outcome (BDI) and a commonly used clinician‐rated instrument (HRSD). We calculated these re‐expressed estimates according to guidance in the Cochrane Handbook (Schünemann 2017a).
Dichotomous outcomes
We analysed dichotomous outcomes by calculating risk ratios (RRs) and 95% CIs for each comparison in Review Manager 5 (Review Manager 2014).
In addition, we calculated the number needed to benefit (NNTB) with 95% CIs for all dichotomous outcomes to facilitate interpretation; this is the expected number of people who need to receive the intervention rather than the comparator for one additional person to achieve a beneficial outcome (Schünemann 2017a).
Unit of analysis issues
Cluster‐randomised trials
We included cluster‐randomised trials as long as proper adjustment for the intracluster correlation could be conducted in accordance with the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
Cross‐over trials
We included trials employing a cross‐over design in the review, but we only used data from the first active treatment phase.
Trials with multiple treatment groups
Multiple‐arm trials (those with more than two intervention arms) can pose analytical problems in pair‐wise meta‐analysis. For trials with more than two eligible arms, we managed data in this review as follows.
Multiple experimental intervention groups versus a single control group
If studies compared multiple eligible experimental interventions with a single control group, we split the control group to enable pair‐wise comparisons.
One or more experimental intervention groups versus multiple control groups
-
If studies used multiple 'active' comparator interventions, we combined these comparator groups to compare to the behavioural activation intervention group (objective 1/2).
-
If studies used multiple control groups including treatment as usual/ waiting list/ attention placebo/ psychological placebo, we combined the control groups to compare to the behavioural activation intervention group (objective 3).
Dealing with missing data
We managed missing dichotomous data through intention‐to‐treat (ITT) analysis, in which we assumed that participants who dropped out after randomisation had a negative outcome. We also conducted best/worse case scenarios for the clinical response outcome, in which we assumed that dropouts in the active treatment group had positive outcomes and those in the control group had negative outcomes (best‐case scenario), and that dropouts in the active treatment group had negative outcomes and those in the control group had positive outcomes (worst‐case scenario), thus providing boundaries for the observed treatment effect. We gave these best/worst case scenarios greater emphasis in the presentation of results if a large amount of information proved to be missing.
We analysed missing continuous data on an endpoint basis, including only participants with a final assessment, or by using the last observation carried forward (LOCF) to the final assessment, if trial authors reported LOCF data. When standard deviations (SDs) were missing, we attempted to obtain these data by contacting trial authors. When SDs were not available from trial authors, we calculated them from P values, t values, CIs or standard errors (SEs), if these were reported in the articles (Deeks 1997).
If a vast majority of SDs were available and only a minority of SDs were unavailable or unobtainable, we used the method devised by Furukawa and colleagues to impute SDs and calculate percentage responders (da Costa 2012; Furukawa 2005; Furukawa 2006). We planned to interpret these data with caution and take into account the degree of observed heterogeneity. We would also planned to undertake a sensitivity analysis to examine the effect of the decision to use imputed data. When conducting the review however, this method for imputing data was not used.
If additional figures were not available or obtainable and it was not deemed appropriate to use the Furukawa method as described above, we did not include the trial data in the comparison of interest.
Assessment of heterogeneity
We formally tested statistical heterogeneity using the Chi2 test, which provides evidence of variation in effect estimates beyond that of chance. Because the Chi2 test has low power to assess heterogeneity when a small number of participants or trials are included, we conservatively set the P value at 0.1 (Deeks 2017). We also quantified heterogeneity using the I2 statistic, which calculates the percentage of variability due to heterogeneity rather than to chance (Higgins 2003). We considered I2 statistic values in the range of 50% to 90% to represent substantial statistical heterogeneity and explored these further. However, the importance of the observed I2 statistic depends on the magnitude and direction of treatment effects and the strength of evidence for heterogeneity. Forest plots generated in Review Manager 5 will provide an estimate of tau2, the between‐trial variance in a random‐effects meta‐analysis (Deeks 2017; Review Manager 2014).
Assessment of reporting biases
As far as possible, we minimised the impact of reporting biases by undertaking comprehensive searches of multiple sources (including trials registries), to identify unpublished material and including non–English language publications.
We also tried to identify outcome reporting bias in trials by recording all trial outcomes, planned and reported, and noting where outcomes were missing. If we found evidence of missing outcomes, we attempted to obtain any available data directly from the trial authors.
Where sufficient data were available, we constructed funnel plots to establish the potential influence of reporting biases and small‐trial effects (Sterne 2017).
Data synthesis
We conducted a meta‐analysis of included trials. Given the potential heterogeneity of behavioural activation approaches for inclusion, together with the likelihood of differing secondary comorbid mental disorders in the population of interest, we used a random‐effects model in all analyses.
Subgroup analysis and investigation of heterogeneity
Clinical heterogeneity
We conducted the following subgroup analyses, depending on the availability of sufficient data for each outcome and comparison.
-
Participant age: old age in particular can be expected to relate to treatment effect, as older people are more likely to suffer comorbidities. We conducted subgroup analyses with participants younger than 65 years and those aged 65 years or older.
-
Level of therapist: one of the often mentioned potential benefits of less complex models of behavioural activation is that therapies can be delivered by a therapist with minimal training, or without a relevant accreditation. We expected that this analysis by level of therapist would also account for potential differences by intervention complexities. We conducted subgroup analyses according to the level of therapist delivering behavioural activation, classified as:
-
accredited/received formal training of several years (specialist); or
-
minimal training/lay counsellor (non‐specialist); or
-
specialist in training; received substantial training but not yet an accredited therapist.
-
-
Baseline depression severity: the severity of depression on entry into the trial is expected to have an impact on outcomes. We planned to categorise depression severity as subthreshold depression, mild, moderate, or severe. As this was not possible in practice, we used the categories of subtreshold/ mild depression and moderate to severe depression instead (see Differences between protocol and review).
-
Length of treatment: we categorised treatment into those delivered in one to three sessions and treatments of longer duration. We anticipated that the length of treatment could influence effectiveness.
-
Type of psychological therapy comparison: the type of psychological therapy comparator used is likely to influence the observed effectiveness of the intervention. When possible, comparators were categorised as psychodynamic, behavioural, humanistic, integrative, or cognitive‐behavioural.
-
Type of control comparison: the type of control comparator used is likely to influence the observed effectiveness of the intervention. When possible, comparators were categorised as waiting list, treatment as usual/usual care, no treatment, attention placebo, or psychological placebo.
Sensitivity analysis
-
Trial quality: we excluded low‐ quality trials in a sensitivity analysis, if we identified a number of higher‐quality trials. As a marker of quality, we used the 'allocation concealment' criteria from the 'Risk of bias' assessment.
-
Mode of delivery: we excluded therapies delivered through computer‐based or electronic guidance without a substantial face‐to‐face component.
-
Subthreshold depression: we planned to exclude trials of subthreshold depression to determine whether our decision to include non‐clinical levels of depression had a substantial impact on the results. We did not conduct this analysis, as it would give the same results as subgroup analyses 3 (baseline depression severity).
-
Group therapy: we excluded trials of group therapy for behavioural activation as the mode of delivery of psychotherapy could influence effectiveness of the therapy.
In addition to these planned sensitivity analyses, we performed several sensitivity analyses to further explore findings of the review. We removed one small study with a large weight (Analysis 1.1) and one outlier (Analysis 10.3; Analysis 10.4). We also conducted fixed‐effect rather than random‐effects analyses for comparisons with smaller and larger studies and extreme estimates (Analysis 6.3; Analysis 6.5; Analysis 10.3; Analysis 10.4) (see Differences between protocol and review).
'Summary of findings' tables
We constructed 'Summary of findings' tables to present the main findings of the review. We reported the outcomes listed below, when available, and presented standardised effect size estimates and 95% CIs. Review author EU performed an assessment of the certainty of the evidence for each outcome using the GRADE approach (Schünemann 2017a). We used GRADEproGDT to create our 'Summary of findings' tables (GRADEpro 2015), and followed standard methods as described in the Cochrane Handbook for Systematic Reviews of Interventions to prepare the tables (Schünemann 2017b). Review authors LR and NM checked GRADE assessments and 'Summary of findings' tables and tables were revised to reflect discussion between EU, LR, and NM.
For each of our main comparisons, we included the following outcomes (measured up to 24 months).
-
Treatment efficacy (number of participants responding to treatment).
-
Treatment acceptability (number of participants who dropped out).
-
Improvement in depression outcomes as a continuous score.
-
Quality of life.
-
Social adjustment/ functioning score.
-
Improvement in anxiety symptoms as a continuous score.
The 'Summary of findings' table was created before writing our discussion, abstract, and conclusions, so that the authors could jointly consider the potential impact of the certainty of the evidence for each outcome on the mean treatment effect and our confidence in these findings. Our confidence in the mean treatment effects based on the GRADE assessments was thus reflected in the interpretation of the results, which informed the abstract, lay summary, and discussion sections of the review.
Results
Description of studies
Results of the search
Searches in all pre‐specified databases were performed by the Cochrane Common Mental Disorders' Information Specialist on 17 January 2019 and an update search was performed on 17 January 2020.
Figure 1 shows the selection of studies through screening of abstracts and full‐text papers. After duplicates were removed, EU, SD, and LR screened titles and abstracts of 5823 records in duplicate. For 380 records, full‐texts were obtained and screened in duplicate (EU, LR, ESo). Conflicts were resolved in discussion with DE, DR, and RC. After linking records belonging to the same study, 53 studies were included in the qualitative synthesis and 49 studies in meta‐analyses. EU, LR, ESa, and ESo extracted data and assessed the risk of bias in duplicate.
Study flow diagram.
Included studies
We included 53 studies in this systematic review. Two of these studies were found as a result of the update search in January 2020.
Study design
Fifty‐one studies were parallel group randomised controlled trials (RCTs) and two were cluster‐RCTs (Fleming 1980; Luo 2018). One RCT allowed switching of placebo and medication treatment after eight weeks depending on participant preference (Dimidjian 2006), and in one cross‐over RCT treatments were switched between groups after six weeks (Kelly 1983). For these two trials employing a cross‐over (Kelly 1983) and partial cross‐over design (Dimidjian 2006), only outcome data for the first phase of the study are included in meta‐analyses of this review as per protocol.
Most trials had two study arms, with either two active treatments or an active and a control group (31 studies). The other studies had three arms (13 studies), four arms (seven studies), five arms (two studies), or six arms (two studies). If study arms were variations of the same treatment, for example two types of behavioural activation, these data were combined in the meta‐analyses of this review.
Sample size
The 53 included studies had 5495 participants, ranging from less than six participants per study arm (Skinner 1984) to an average of 352 participants per study arm (Gilbody 2017).
Setting
Many studies did not report on the setting and appear to have been conducted at a university or medical centre. Recruitment settings that were reported included: universities (Armento 2012; Cullen 2003; Gawrysiak 2009; Hammen 1975; Kelly 1983; McCluskey 2018; McIndoo 2016; McNamara 1986; Shaw 1977; Takagaki 2016; Taylor 1977; Weinberg 1978; Wilson 1983; Zeiss 1979; Zemestani 2016) , medical centres including psychiatric outpatient facilities (Toghyani 2018; van den Hout 1995), community mental health services, primary care and community health centres (Bolton 2014; Bosanquet 2017; Bowe 2014; Chang 2018; Chowdhary 2016; Ekers 2011; Gilbody 2017; Nasrin 2017; Kanter 2015; Richards 2017), and nursing homes or facilities for older people (Luo 2020; Meeks 2008; Raue 2019). Several interventions were computer‐based or phone‐based and supported from a distance (Carlbring 2013; Carlbring 2013a; Ly 2014; Stiles‐Shields 2019).
Studies were conducted in the USA (27 studies), UK (five studies), Iran (three studies), Sweden (three studies), Australia (two studies), Canada (two studies), India (two studies), Brazil (one study), China (one study), Hong Kong (one study), Indonesia (one study), Iraq (one study), Japan (one study), the Netherlands (one study), South Korea (one study), and Spain (one study).
Participants
We extracted data on participant age, sex, ethnic group, socioeconomic characteristics (household income, occupation/employment, education level), depression severity, and comorbid anxiety. In this section we briefly summarise the information available in study reports.
Age
Most studies included adult participants of all ages. Four studies included only adults up to 60 years old (Dimidjian 2006; Hemanny 2019; Nasrin 2017; Wilson 1983), seven studies included only participants aged 65 and over (Bosanquet 2017; Chang 2018; Gilbody 2017; Luo 2020; Meeks 2008; Raue 2019; Xie 2019), and in four studies samples were exclusively made up of young adult college/university students with an average age between 18 and 24 (Gawrysiak 2009; McIndoo 2016; Takagaki 2016; Zemestani 2016).
Differences in results for treatment efficacy and treatment acceptability by participant age (under 65 and aged 65 and over) are explored in subgroup analyses (Analysis 11.1; Analysis 11.2).
Sex
Six studies included only women (Fuchs 1977; Kornblith 1980; Padfield 1976; Rehm 1982; Thomas 1987; Toghyani 2018). Two studies included more men than women (39% and 38% women, respectively) (Cullen 2003; Takagaki 2016). In all other studies that reported on the sex of participants (36 studies), women represented between 58% and 93% of the sample.
Ethnicity
Five studies included participants of a specific region or ethnic group: people from various islands in Indonesia (Arjadi 2018), a sample of African American participants (Bowe 2014), Puerto Ricans (Comas Díaz 1981), and Latinos living in the USA (Collado 2016; Kanter 2015).
The other 14 studies reporting on participant ethnicity inlcuded predominantly White American or White British participants (58% to 99%), except for a study from Brazil reporting a mix of participants from three ethnic groups (Hemanny 2019).
Socioeconomic characteristics
Studies collected data on income, level of education, and employment status or occupation. It is difficult to compare study participants as these characteristics are time‐ and place‐dependent. In many studies the sample represented a mix of people with various socioeconomic characteristics.
Some studies predominantly recruited participants of a higher socioeconomic status. For example, in the study by Ly and colleagues 7% of participants were unemployed and 63% had attended university (Ly 2014). Similarly, in the study by Carlbring and colleagues 9% of participants were unemployed and 62% had attended university (Carlbring 2013a), and in Arjadi 2018 unemployment was 8% and university attendance 55%.
Other studies had samples with predominantly people from a lower socioeconomic status, and some of these studies were conducted in low‐ and middle‐income countries, or with people from such countries. For example, in seven studies a majority of participants had completed no more than primary education (Bolton 2014; Chang 2018; Chowdhary 2016; Comas Díaz 1981; Luo 2020; Weobong 2017; Xie 2019). In six studies levels of unemployment ranged from 50% to 100% (Bolton 2014; Chowdhary 2016; Comas Díaz 1981; Kanter 2015; Thomas 1987; Weobong 2017).
Severity of depression
For most studies, inclusion criteria specified a range or lower limit of depression symptoms using a commonly used screening tool such as the Patient Health Questionnaire (PHQ‐9), Beck Depression Inventory (BDI), or Hamilton Rating Scale for Depression (HRSD).
In 25 studies, only people with diagnosed major depressive disorder or moderate to severe depression symptoms were included (Arjadi 2018; Bosanquet 2017; Bowe 2014; Chang 2018; Chowdhary 2016; Collado 2016; Cullen 2003; Dimidjian 2006; Hemanny 2019; McNamara 1986; Meeks 2008; Moradveisi 2015; Nasrin 2017; Padfield 1976; Rehm 1982; Richards 2017; Kanter 2015; Shaw 1977; Stiles‐Shields 2019; Thompson 1987; Toghyani 2018; van den Hout 1995; Weobong 2017; Xie 2019; Zemestani 2016).
In 16 studies, people with various levels of depression severity (mild, moderate, severe) were included (Armento 2012; Carlbring 2013a; Ekers 2011; Fleming 1980; Gawrysiak 2009; Hammen 1975; Kelly 1983; Ly 2014; McCluskey 2018; McIndoo 2016; Raue 2019; Skinner 1984; Taylor 1977; Thomas 1987; Weinberg 1978; Wilson 1983). Participants in 11 of these studies predominantly had major depressive disorder, or moderate to severe symptoms of depression (Armento 2012; Carlbring 2013a; Ekers 2011; Fleming 1980; Gawrysiak 2009; Kelly 1983; Ly 2014; McIndoo 2016; Raue 2019; Taylor 1977; Thomas 1987).
Three studies included only participants with subthreshold depression or minimal to mild symptoms (Gilbody 2017; Takagaki 2016; Vázquez 2014).
Based on this information from eligibility criteria and descriptions of the study samples, we can conclude that, in most studies, the majority of participants were suffering from moderate to severe levels of depression.
Anxiety
Most studies did not report on numbers of participants with comorbid anxiety. Anxiety disorder was an exclusion criteria in six trials (Chang 2018; Jacobson 1996; Kornblith 1980; Rehm 1982; Toghyani 2018; van den Hout 1995). In eight studies that reported on anxiety among participants, levels of anxiety disorder varied from 11% (Moradveisi 2015) to 65% (Collado 2016).
Intervention
Description of intervention
Behavioural activation interventions were described in different ways, and some were specifically designed for the study setting or population. All interventions are described in the Characteristics of included studies tables. Examples include the following.
-
Behavioural Activities Intervention (BE‐ACTIV) for nursing home residents (Luo 2020; Meeks 2008)
-
Healthy Activity Programme (HAP) for treatment of moderate to severe depression in primary care in India (Chowdhary 2016; Weobong 2017)
-
Culturally Enhanced Behavioural Activation (CEBA) for African American communities (Bowe 2014)
-
Behavioural Activation for Latinos (BAL) for a low‐income Spanish speaking Latino community (Kanter 2015)
-
Behavioural Activation of Religious Behaviors (BARB) (Armento 2012)
Others were described as behavioural activation or behavioural therapy based on Lewinsohns' approach (McNamara 1986; Padfield 1976; Shaw 1977; Skinner 1984; Taylor 1977; Thompson 1987; Vázquez 2014; Weinberg 1978), Behavioural Activation Treatment for Depression (BATD) (Bolton 2014; Collado 2016; Gawrysiak 2009; McCluskey 2018; Nasrin 2017), or behavioural activation based on the intervention evaluated by Fuch and Rehm (Fleming 1980; Fuchs 1977; Kornblith 1980; Rehm 1982; Thomas 1987; van den Hout 1995).
Level of therapist
For several trials, behavioural activation was delivered by a specialist in training (Carlbring 2013a; Fleming 1980; Fuchs 1977; Kelly 1983; McNamara 1986; Shaw 1977; Thomas 1987; Thompson 1987; Weinberg 1978; Zeiss 1979; Zemestani 2016), or a non‐specialist (Arjadi 2018; Bolton 2014; Bosanquet 2017; Chang 2018; Chowdhary 2016; Collado 2016; Ekers 2011; Gilbody 2017; Luo 2020; Raue 2019; Richards 2017; Weobong 2017; Xie 2019), rather than a mental health specialist.
Trials published before the 1990s regularly used graduate or doctoral students to deliver interventions within the trial setting, even if the treatment would normally be delivered by accredited mental health specialists who completed formal training. In recent years, several behavioural activation interventions evaluated in trials have been delivered by non‐specialists such as primary care workers or lay health workers, with a view to test an alternative therapy feasible for delivery in settings with limited resources.
Duration and format
Most of the interventions were delivered face‐to‐face. Four studies involved initial or occasional face‐to‐face contact, with most of the intervention delivered via phone calls (Armento 2012; Bosanquet 2017; Chang 2018; Gilbody 2017). One intervention was delivered through a series of conference calls (Vázquez 2014), three were delivered online (Arjadi 2018; Carlbring 2013a; Carlbring 2013), and two used a smartphone app in combination with contact via phone or email (Ly 2014; Stiles‐Shields 2019). The exclusion of studies delivering interventions without a substantial face‐to‐face component is explored in sensitivity analyses (Analysis 20.1; Analysis 20.2; Analysis 21.1; Analysis 21.2).
Most interventions were delivered once or twice a week (Chowdhary 2016; Dimidjian 2006; Fleming 1980; Hemanny 2019; Moradveisi 2015; Richards 2017; Shaw 1977; Thompson 1987; Toghyani 2018), and a few interventions were delivered in only one session (Armento 2012; Gawrysiak 2009; Nasrin 2017). Most interventions were delivered over a period of four to 12 weeks, with the longest duration being 16 weeks (Richards 2017).
Therapy sessions were usually up to an hour in duration, with others lasting between 90 minutes and two hours (Bowe 2014; Comas Díaz 1981; Fleming 1980; Fuchs 1977; Gawrysiak 2009; Kornblith 1980; McCluskey 2018; Nasrin 2017; Rehm 1982; Shaw 1977; Toghyani 2018; van den Hout 1995; Vázquez 2014; Xie 2019; Zemestani 2016).
In most of the studies interventions were delivered to individuals, while 10 were delivered in a group format (Fleming 1980; Kornblith 1980; Rehm 1982; Shaw 1977; Thomas 1987; Toghyani 2018; van den Hout 1995; Vázquez 2014; Xie 2019; Zemestani 2016), and three used a mixed individual/group format (Bowe 2014; Fuchs 1977; Takagaki 2016). We performed sensitivity analyses to explore outcomes for the individual format only (Analysis 22.1; Analysis 22.2; Analysis 23.1; Analysis 23.2).
Comparators
Other psychological therapies
Psychological therapies other than behavioural activation, which were used as comparators, included the following.
-
CBT (Bolton 2014; Dimidjian 2006; Hemanny 2019; Jacobson 1996; McNamara 1986; Rehm 1982; Richards 2017; Stiles‐Shields 2019; Taylor 1977; Thomas 1987; Thompson 1987; Vázquez 2014; Weinberg 1978; Wilson 1983)
-
Third‐wave cognitive and behavioural therapies (Ly 2014; McIndoo 2016; Zemestani 2016)
-
Humanistic therapy (Armento 2012; Collado 2016; McNamara 1986)
-
Psychodynamic therapy (Thompson 1987)
-
Interpersonal, cognitive analytic, and other integrative therapies (Kornblith 1980; Padfield 1976; Toghyani 2018; Weinberg 1978)
We categorised cognitive processing therapy and cognitive therapy as CBT. Emotional awareness training, general counselling, and general psychotherapy were included as an integrative therapies.
Other non‐therapy comparators included the following.
-
Waiting list (Bolton 2014; Carlbring 2013a; Carlbring 2013; Cullen 2003; McIndoo 2016; Nasrin 2017; Taylor 1977; Stiles‐Shields 2019; Weinberg 1978; Zemestani 2016)
-
Placebo; medical placebo (Dimidjian 2006), and attention placebo (Hammen 1975)
-
Medication (Dimidjian 2006; Moradveisi 2015)
-
No treatment (Gawrysiak 2009; Hammen 1975; McCluskey 2018; Takagaki 2016)
-
Treatment as usual (Bosanquet 2017; Ekers 2011; Gilbody 2017; Hemanny 2019; Kanter 2015; Luo 2020; Meeks 2008; Xie 2019)
If treatment as usual comprised medication or therapy it was added to the relevant medication or therapy comparisons. Online 'minimal psychoeducation', referral to mental health services and enhanced usual care (described as 'routine consultation and referral to services) were categorised as treatment as usual. Self‐monitoring, described as 'no change from normal activities', was categorised as no treatment.
Outcomes
Studies reported data on all of the seven outcomes specified for this review. We report meta‐analyses and forest plots for all outcomes at short‐term, medium‐term, and long‐term endpoints, and subgroup analyses for primary outcomes treatment efficacy and treatment acceptability (dropouts) at short‐term time endpoints.
Most data are available for depression symptoms, as measured by commonly used instruments for depression severity such as the BDI or HRSD. Depression symptom outcomes were more commonly reported than treatment efficacy. For treatment efficacy, we encountered multiple measures across studies, and sometimes multiple measures within one study. If a study reported multiple measures of treatment efficacy, we prioritised as follows: remission over clinically significant improvement, and recovery or remission over response.
For six studies data were missing and standard deviations could not be calculated (Bowe 2014; Comas Díaz 1981; Fleming 1980; Kelly 1983; Skinner 1984; Zeiss 1979). These studies are not included in any meta‐analyses. For some studies data were missing but could be calculated (Fuchs 1977; Gardner 1981; McCluskey 2018; Shaw 1977; van den Hout 1995).
Data on adverse events are summarised narratively in Table 1.
| First author | Year of publication | Comparator group(s) | Description of adverse events (at end of study period) |
| Bosanquet | 2017 | Treatment as usual | BA: 47 suspected adverse events. Usual care: 34 suspected adverse events. Elderly sample. |
| Dimidjian | 2006 | CBT, medication, medical placebo | various physical side effects from antidepressant medication and placebo. 1 suicide in antidepressant arm. |
| Gilbody | 2017 | Treatment as usual | BA: 37 events; 35 unrelated to intervention and 2 unlikely to be related to intervention. Usual care: 44 events; 40 unrelated and 4 unlikely to be related to intervention. 18 patients died (elderly sample). |
| Padfield | 1976 | Interpersonal, cognitive analytic, integrative | 2 suicide attempts and 1 case of suicidal thoughts in comparator arm; no adverse events in behavioural activation arm. |
| Richards | 2017 | CBT | 3 serious adverse events in behavioural activation arm (2 overdose, 1 self‐harm) and 8 serious adverse events in comparator arm (7 overdose, 1 self‐harm). |
| Stiles‐Shields | 2018 | CBT and waiting list | No adverse events. |
| Weobong | 2017 | Treatment as usual | 1 suicide attempt and 18 unplanned hospitalisations in behavioural activation arm. 1 suicide attempt, 26 unplanned hospitalisations, and 2 deaths in comparator arm. |
BA: Behavioural activation; CBT: cognitive‐behavioural therapy;
Excluded studies
After obtaining full‐text manuscripts, a total of 256 studies were excluded (259 full‐text records) (Figure 1). Of the studies excluded at this stage, 120 studies were of participants with a physical comorbidity. These studies were included at the title and abstract screening stage as they informed a Cochrane review focussed on this population (Uphoff 2019b). A further 78 studies were excluded because the intervention was not behavioural activation, or behavioural activation was a component but not the key ingredient of the intervention. Another 33 studies were excluded because they were not RCTs. We were unable to exclude these studies at the stage of title and abstract screening because the abstracts did not clearly specified the study design and/or intervention. These records are not listed in this review.
Among the 25 studies (28 references) which are listed as Excluded studies in this review, the most common reasons for exclusion were wrong comparator (k = 10) and interventions with no interaction with a therapist (k = 5). Further reasons for exclusion are listed in the Characteristics of excluded studies table.
In addition to the excluded studies, authors were contacted to check whether studies identified through protocols had been published. Fifteen studies (20 references) are listed as ongoing (Characteristics of ongoing studies) and eight are awaiting classification (Studies awaiting classification).
Risk of bias in included studies
Out of 53 included studies, for 46 studies one or more risk of bias domains were initially rated as 'unclear' because information was missing from the study report and/or trial registration/ protocol. For 25 studies the author could not be contacted; 15 of these studies were published before 1990. Authors of two old studies replied but could no longer provide the requested information (Gardner 1981; Hammen 1975).
Allocation
Eighteen studies were rated as low risk of selection bias (Figure 2; Figure 3). The others, particularly older studies, either did not report sufficient information on randomisation and/or allocation concealment or were found to be at high risk of bias. In several studies randomisation was performed correctly, but a researcher involved in the study was aware of the allocation participant list (McIndoo 2016; Meeks 2008; Zemestani 2016).
'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.
Blinding
Blinding of participants and personnel (performance bias) was not achieved in any of the included studies (Figure 2) and is rarely attempted for psychological therapy interventions. In 18 studies, authors limited the risk of detection bias through blinding of outcome assessors. Where outcome assessors were not blinded this was usually because participants self‐completed questionnaires on symptoms of depression.
Incomplete outcome data
There was evidence of incomplete outcome data in 25 studies, and an assessment of 'unclear' risk of attrition bias in a further 14 studies. Issues included no or unclear reporting of participants who dropped out of the trial, unclear reasons for dropout, the exclusion of participants who dropped out from the analyses, or a substantial difference in dropout rates between the different study arms.
Selective reporting
For the majority of studies, no reference was made to a study protocol or online trial registration. Nine studies were rated as low risk of reporting bias, because a protocol or trial registration was available and no differences with the results were found. For seven studies, there were discrepancies in outcomes reported between the methods section, trial registration, or protocol, and the published study results (Carlbring 2013a; Chang 2018; Collado 2016; Hemanny 2019; Ly 2014; McIndoo 2016; Meeks 2008). For example, reporting of extra outcomes, no reporting of outcomes listed in the protocol, a change in the measures used, or differences in time points reported.
Other potential sources of bias
Other issues that were rated as a high risk of bias were identified for 26 studies. This included issues relevant to the conduct of trials in psychotherapy, such as low treatment fidelity of therapists, researcher or therapist allegiance to one of the interventions, or a conflict of interest from researchers or therapists. For example, several trials were conducted by authors who were involved in the development of the intervention. Another common potential source of bias was any indication of inadequate randomisation, for example with extremely small sample sizes (less than 10 participants per study arm) and substantial differences in participant characteristics between study arms.
Effects of interventions
See: Summary of findings 1 Behavioural activation compared with CBT for depression in adults; Summary of findings 2 Behavioural activation compared with third‐wave CBT for depression in adults; Summary of findings 3 Behavioural activation compared with humanistic therapy for depression in adults; Summary of findings 4 Behavioural activation compared with psychodynamic for depression in adults; Summary of findings 5 Behavioural activation compared with interpersonal, cognitive analytic, integrative for depression in adults; Summary of findings 6 Behavioural activation compared with waiting list for depression in adults; Summary of findings 7 Behavioural activation compared with placebo for depression in adults; Summary of findings 8 Behavioural activation compared with medication for depression in adults; Summary of findings 9 Behavioural activation compared with no treatment for depression in adults; Summary of findings 10 Behavioural activation compared with treatment as usual for depression in adults
Behavioural activation versus psychological therapies
Included studies compared behavioural activation with CBT, third‐wave CBT, humanistic therapy, psychodynamic therapy, and interpersonal, cognitive analytic, and integrative therapy.
Comparison 1. Behavioural activation versus cognitive‐behavioural therapy (CBT)
Short‐term outcomes
Moderate‐ to very low‐certainty evidence from randomised controlled trials (RCTs) showed no statistically significant differences in short‐term (up to six months) outcomes between behavioural activation and CBT in terms of treatment efficacy (risk ratio (RR) 0.99, 95% CI 0.92 to 1.07; 5 RCTs, 601 participants; moderate‐certainty evidence) (Analysis 1.1), treatment acceptability (RR 1.03, 95% CI 0.85 to 1.25; 12 trials, 1195 participants) (Analysis 1.2; low certainty), depression symptoms (standardised mean difference (SMD) 0.12, 95% CI ‐0.08 to 0.32; high heterogeneity I2 52%; 16 RCTS, 1205 participants) (Analysis 1.3), quality of life (SMD 0.04, 95% CI ‐0.20 to 0.28; 2 RCTs, 268 participants; moderate‐certainty evidence) (Analysis 1.4), social adjustment and functioning (SMD ‐0.13, 95% CI ‐0.50 to 0.24; 2 RCTs, 111 participants; very low‐certainty evidence) (Analysis 1.5), and anxiety symptoms (SMD ‐0.03, 95% CI ‐0.18 to 0.13; 4 RCTs, 646 participants; moderate‐certainty evidence) (Analysis 1.6).
One small study (Vázquez 2014) has a high weight in the analyses comparing the treatment efficacy of behavioural activation and CBT because none of the participants had depression at follow‐up. Removing this study made the pooled estimate less precise but did not substantially change the estimate (RR 0.94, 95% CI 0.81 to 1.10).
Medium‐ and long‐term outcomes
One study (Richards 2017) compared outcomes between a behavioural activation and a CBT group in the medium term (seven to 12 months) and long term (>12 months), and found no evidence of a difference in treatment efficacy (medium term: RR 1.00, 95% CI 0.86 to 1.16; 364 participants, long term: (RR 0.93, 95% CI 0.81 to 1.08;356 participants)), treatment acceptability (medium term: RR 1.25, 95% CI 0.97 to 1.62; 440 participants, long term: RR 1.16, 95% CI 0.90 to 1.49; 440 participants), depression symptoms (medium term: SMD ‐0.18, 95% CI ‐0.38 to 0.02; 380 participants, long term: (SMD 0.00, 95% CI ‐0.21 to 0.21; 364 participants), quality of life (medium term: SMD 0.15, 95% CI ‐0.07 to 0.37, long term: SMD 0.06, 95% CI ‐0.15 to 0.28), and anxiety symptoms (medium term: SMD 0.02, 95% CI ‐0.20 to 0.23; 337 participants, long term: SMD ‐0.10, 95% CI ‐0.31 to 0.12; 332 participants).
Adverse events
Adverse events was included as an outcome in two trials; one reported no adverse events (Stiles‐Shields 2019), and the other reported three serious adverse events in the behavioural activation arm (two overdose, one self‐harm) and eight serious adverse events in the CBT arm (seven overdose, one self‐harm) (Richards 2017) (Table 1).
Funnel plots revealed no indications of publication bias for treatment acceptability and depression symptoms.
Comparison 2. Behavioural activation versus third‐wave CBT
Short‐term outcomes
Three RCTs contributed low‐certainty evidence comparing behavioural activation with third‐wave CBT.
Low‐certainty evidence showed no statistically significant differences for short‐term (up to six months) outcomes between behavioural activation and third‐wave CBT in terms of treatment efficacy (RR 1.10, 95% CI 0.91 to 1.33; 2 RCTs, 98 participants) (Analysis 2.1), treatment acceptability (RR 0.84, 95% CI 0.33 to 2.10; 3 RCTs, 147 participants) (Analysis 2.2), depression symptoms (SMD ‐0.14, 95% CI ‐0.47 to 0.18; 3 RCTs, 147 participants) (Analysis 2.3), quality of life (MD 0.02, 95% CI ‐0.96 to 1.00; 1 RCT, 81 participants) (Analysis 2.4), and anxiety symptoms (MD 0.69, 95% CI ‐0.68 to 2.06; 3 RCTs, 147 participants) (Analysis 2.5).
Data on social adjustment and functioning and adverse events were not reported.
Comparison 3. Behavioural activation versus humanistic therapy
Short‐term outcomes
Three RCTs contributed moderate‐ to very low‐certainty evidence on the comparison of short‐term (up to six months) outcomes between behavioural activation and humanistic therapy.
Low‐certainty evidence showed greater treatment efficacy for behavioural activation compared with humanistic therapy (RR 1.84, 95% CI 1.15 to 2.95; 2 RCTs, 46 participants) (Analysis 3.1). Three people would need to receive treatment for one person with depression to benefit.
Low‐ to very low‐certainty evidence showed no statistically significant difference in treatment acceptability (RR 1.06, 95% CI 0.20 to 5.55; 2 RCTs, 96 participants, very low certainty) (Analysis 3.2), quality of life (MD 0.80, 95% CI ‐0.12 to 1.72; 1 RCT, 50 participants; low certainty) (Analysis 3.4), or anxiety symptoms (MD ‐1.30, 95% CI ‐6.10 to 3.50; 1 RCT, 50 participants; low certainty) (Analysis 3.5).
Moderate‐certainty evidence indicated that depression symptoms improved more in those assigned to behavioural activation compared with those assigned to humantistic therapy (MD ‐3.75, 95% CI ‐6.72 to ‐0.78; 3 RCTs, 93 participants) (Analysis 3.3).
Data on social adjustment and functioning and adverse events were not reported.
Comparison 4. Behavioural activation versus psychodynamic therapy
Short‐term outcomes
Very low‐certainty evidence from one RCT (60 participants) showed no difference between behavioural activation and psychodynamic therapy for short‐term outcomes treatment efficacy (RR 1.21, 95% CI 0.74 to 1.99) (Analysis 4.1), depression symptoms (MD ‐1.10, 95% CI ‐4.35 to 2.15) (Analysis 4.2), and social adjustment and functioning (MD 2.10, 95% CI ‐4.92 to 9.12) (Analysis 4.3).
Data on treatment acceptability, quality of life, anxiety symptoms, and adverse events were not reported.
Comparison 5. Behavioural activation versus interpersonal, cognitive analytic, and integrative therapy
Short‐term outcomes
Very low‐certainty evidence showed no difference between behavioural activation and interpersonal, cognitive analytic, and integrative therapies for short‐term outcomes treatment acceptability (RR 0.84, 95% CI 0.32 to 2.20; 4 RCTs, 123 participants) (Analysis 5.1), depression symptoms (SMD ‐0.16, 95% CI ‐0.59 to 0.28; 4 RCTs, 103 participants) (Analysis 5.2), social adjustment and functioning (MD ‐3.92, 95% CI ‐16.78 to 8.93; 1 RCT, 39 participants) (Analysis 5.3), and anxiety symptoms (MD ‐0.39, 95% CI ‐11.78 to 11.00; 1 RCT, 15 participants) (Analysis 5.4).
Data on treatment efficacy and quality of life were not reported.
Adverse events
Padfield 1976 reported there were no adverse events in the behavioural activation study arm and two suicide attempts and one case of suicidal thoughts in the comparator arm (low‐certainty evidence) (Table 1).
Behavioural activation versus other comparators
Included studies compared behavioural activation with being on a waiting list, receiving a placebo, medication (anti‐depressants), no treatment, or treatment as usual.
Comparison 6. Behavioural activation versus waiting list
Short‐term outcomes
Moderate‐ to low‐certainty evidence suggested there is no difference in treatment efficacy (RR 2.14, 95% CI 0.90 to 5.09; 1 RCT, 26 participants; low certainty) (Analysis 6.1) and treatment acceptability (RR 1.17, 95% CI 0.70 to 1.93; 8 RCTs, 359 participants; moderate certainty) (Analysis 6.2) in the short term (up to six months) between behavioural activation and waiting list.
Low‐certainty evidence showed that those who received behavioural activation had a greater short‐term reduction in depression symptoms than those on a waiting list (SMD ‐1.04, 95% CI ‐1.44 to ‐0.63; 12 RCTs, 619 participants) (Analysis 6.3). The funnel plot indicates that smaller studies with results favouring waiting list may be missing from these data (Figure 4).
Funnel plot of comparison: 6 behavioural activation vversus waiting list, outcome: 6.3 depression symptoms.
Low‐certainty evidence indicated benefits of behavioural activation compared with waiting list for anxiety symptoms (SMD ‐0.91, 95% CI ‐1.59 to ‐0.23; 5 RCTs, 424 participants) (Analysis 6.5), but not for quality of life (MD 0.03, 95% CI ‐0.70 to 0.76; 1 RCT, 80 participants) (Analysis 6.4).
No data were reported on social adjustment and functioning.
Estimates for depression symptoms and anxiety symptoms suggested a large effect and high level of heterogeneity (I2 75% and 87%, respectively). To test how robust these findings are, we conducted sensitivity analyses with fixed‐effect instead of random‐effects models. The pooled estimates were reduced to SMD ‐0.72 (95% CI ‐0.89 to ‐0.55) for depression symptoms (Analysis 24.1) and SMD ‐0.54 (95% CI ‐0.74 to ‐0.33) for anxiety symptoms (Analysis 24.2).
Adverse events
The authors of a trial comparing behavioural activation to CBT and waiting list reported that no adverse events took place (Stiles‐Shields 2019)(Table 1).
Comparison 7. Behavioural activation versus placebo
Short‐term outcomes
Two RCTs contributed low‐certainty evidence comparing behavioural activation to a placebo. One study used a medical placebo (Dimidjian 2006) and one used an attention placebo (Hammen 1975).
No difference between behavioural activation and placebo was found for treatment acceptability (RR 0.72, 95% CI 0.31 to 1.67; 1 RCT; 96 participants) (Analysis 7.1) and depression symptoms (SMD ‐0.18, 95% CI ‐0.57 to 0.20; 2 RCTs, 108 participants) (Analysis 7.2).
No data were reported for treatment efficacy, quality of life, anxiety symptoms, and social adjustment and functioning.
Adverse events
One RCT reported various physical side effects from the medication placebo, and no adverse events for the behavioural activation group (Dimidjian 2006) (Table 1).
Comparison 8. Behavioural activation versus medication
Short‐term outcomes
One RCT (141 participants) on treatment efficacy, with treatment efficacy being higher for behavioural activation than medication (RR 1.77, 95% CI 1.14 to 2.76; 1 RCT, 141 participants) (Analysis 8.1). We judged this evidence to be of moderate certainty following the GRADE approach. However, because this evidence is based on data from one trial only, we were not able to assess inconsistency in results between trials.
Moderate‐certainty evidence showed no difference between behavioural activation and medication in short‐term treatment acceptability (RR 0.52, 95% CI 0.23 to 1.16; 2 RCTs, 243 participants) (Analysis 8.2). Low‐certainty evidence showed no difference in short‐term symptoms of depression (MD ‐1.42, 95% CI ‐4.80 to 1.96; 2 RCTs, 180 participants; I2 83% indicating high heterogeneity) (Analysis 8.3).
No data were reported on quality of life, anxiety symptoms, and social adjustment and functioning.
Medium‐ and long‐term outcomes
One RCT (reported medium‐term (seven to 12 months) outcomes comparing behavioural activation to medication. There was no difference in treatment acceptability between the groups (RR 0.86, 95% CI 0.31 to 2.37; 100 participants) (Analysis 8.2). Symptoms of depression decreased more in the behavioural activation than the medication group (MD ‐2.34, 95% CI ‐3.84 to ‐0.84; 100 participants) (Analysis 8.3).
Adverse events
In one RCT comparing CBT, medication, and medical placebo to behavioural activation, a range of physical side effects were reported for the medication and placebo study arms (Dimidjian 2006). There was one case of suicide in the medication arm. No adverse events of behavioural activation were reported (Table 1). This information is also included in the behavioural activation and placebo comparison in summary of findings Table 7.
Comparison 9. Behavioural activation versus no treatment
Short‐term outcomes
Three RCTs contributed data on short‐term (up to six months) outcomes to the comparison of behavioural activation versus no treatment.
Moderate‐certainty evidence indicated there was no difference between behavioural activation and no treatment in treatment acceptability (RR 0.97, 95% CI 0.45 to 2.09; 3 RCTs, 187 participants) (Analysis 9.1).
Moderate‐certainty evidence showed a benefit of behavioural activation in improvement in depression symptoms (MD ‐6.10, 95% CI ‐7.87 to ‐4.33; 3 RCTs, 187 participants) (Analysis 9.2), and high‐certainty evidence showed greater improvements in quality of life for behavioural activation compared with no treatment (MD 0.07, 95% CI 0.03 to 0.11; 1 RCT, 118 participants) (Analysis 9.3). Low‐certainty evidence indicated a greater reduction in anxiety symptoms for behavioural activation compared with no treatment (MD ‐5.50, 95% CI ‐10.01 to ‐0.99; 1 RCT, 30 participants) (Analysis 9.4).
No data were reported on treatment efficacy, social adjustment and functioning, and adverse events.
Medium‐ and long‐term outcomes
One RCT reported on medium‐term outcomes comparing behavioural activation to no treatment. Results showed no difference in treatment acceptability (RR 1.57, 95% CI 0.65 to 3.79; 124 participants)(Analysis 9.1). There was a greater reduction in depression symptoms for the behavioural activation group compared with the no treatment group (MD ‐2.83, 95% CI ‐5.32 to ‐0.34; 118 participants) (Analysis 9.2).
Comparison 10. Behavioural activation versus treatment as usual
Short‐term outcomes
Fifteen RCTs contributed very low‐ to moderate‐certainty evidence on short term (up to six months) outcomes for behavioural activation versus treatment as usual.
Moderate‐certainty evidence indicated greater treatment efficacy for behavioural activation compared with treatment as usual (RR 1.40, 95% CI 1.10 to 1.78; 7 RCTs, 1533 participants) (Analysis 10.1), although this difference was not found in sensitivity analyses using a worst‐case or intention‐to‐treat scenario (Analysis 26.1; Analysis 28.1). Moderate‐certainty evidence suggested greater treatment acceptability, as indicated by dropouts, for treatment as usual, although results lacked precision (RR 1.64, 95% CI 0.81 to 3.31; 14 RCTs, 2518 participants) (Analysis 10.2).
Low‐certainty evidence suggested a benefit of behavioural activation in terms of depression symptoms (SMD ‐0.78, 95% CI ‐1.05 to ‐0.51; 15 RCTs, 2208 participants) (Analysis 10.3), and social adjustment and functioning (SMD ‐1.27, 95% CI ‐1.74 to ‐0.81; 2 RCTs, 88 participants) (Analysis 10.5). Very low‐certainty evidence suggested a greater improvement in quality of life for behavioural activation compared with treatment as usual (SMD 0.97, 95% CI 0.38 to 1.57; 6 RCTs, 1299 participants) (Analysis 10.4). Moderate‐certainty evidence showed a greater improvement in anxiety symptoms for behavioural activation compared with treatment as usual (SMD ‐0.33, 95% CI ‐0.45 to ‐0.21; 4 RCTs, 1063 participants) (Analysis 10.6).
In the study by Luo and colleagues (Luo 2020), large effects were reported favouring behavioural activation over treatment as usual for depressions symptoms and quality of life. Removing this study from the analyses changed the pooled estimate of depression symptoms (SMD ‐0.57, 95% CI ‐0.75 to ‐0.39) and quality of life (SMD 0.34, 95% CI 0.03 to 0.66).
Estimates for depression symptoms and quality of life suggested a large effect. To test how robust these findings are, we conducted sensitivity analyses with fixed‐effect instead of random‐effects models. The pooled estimates changed to SMD ‐0.48 (95% CI ‐0.57 to ‐0.39) for depression symptoms (Analysis 25.1) and SMD 0.25 (95% CI 0.14 to 0.37) for quality of life (Analysis 25.2).
Funnel plots reveal no indication of publication bias for treatment efficacy, treatment acceptability, and depression symptoms. The I2 test statistic suggests high levels of statistical heterogeneity for short‐term outcomes treatment efficacy (84%), acceptability (85%), depression symptoms (85%), and quality of life (95%), and medium‐term outcomes treatment acceptability (94%) and quality of life (72%).
Medium‐ and long‐term outcomes
Five RCTs reported medium term (seven to 12 months) and/or long term (> 12 months) estimates comparing behavioural activation to treatment as usual.
There was evidence of a difference in medium term treatment efficacy (RR 1.23, 95% CI 1.07 to 1.42; 2 RCTs, 1012 participants) (Analysis 10.1), but not for treatment acceptability (RR 2.84, 95% CI 0.92 to 8.75; 4 RCTs, 1726 participants) (Analysis 10.2). Long‐term treatment acceptability favoured treatment as usual (RR 2.17, 95% CI 1.39 to 3.39; 1 RCT, 485 participants).
Medium‐term depression symptoms showed greater improvement for behavioural activation than treatment as usual (SMD ‐0.23, 95% CI ‐0.38 to ‐0.08; 4 RCTs, 1381 participants), while no difference was found for long term depression symptoms (SMD 0.02, 95% CI ‐0.19 to 0.23; 1 RCT, 343 participants).
There was no difference in quality of life in the medium term (SMD 0.14, 95% CI ‐0.12 to 0.40; 2 RCTs, 879 participants) and long term (SMD ‐0.09, 95% CI ‐0.30 to 0.13; 1 RCT, 325 participants).
A greater reduction in anxiety symptoms was found for behavioural activation compared with treatment as usual in the medium term (SMD ‐0.27, 95% CI ‐0.41 to ‐0.12; 2 RCTs, 851 participants), but not in the long term (SMD ‐0.08, 95% CI ‐0.29 to 0.14; 1 RCT; 332 participants).
Adverse events
Three studies with a 'treatment as usual' comparator reported on adverse events (Table 1). Two studies of participants aged 65 and older reported a large number of suspected or potential adverse events. In Bosanquet 2017 there were 47 suspected adverse events in the behavioural activation group, and 34 in the treatment as usual group. In Gilbody 2017 there were 37 adverse events in the behavioural activation group and 44 in the treatment as usual group, but none of these were thought to be related to the interventions. In a study of people with severe depression, there was one suicide attempt and 18 unplanned hospitalisations in the behavioural activation arm and one suicide attempt, 26 unplanned hospitalisations, and two deaths in the comparator arm (Weobong 2017).
Subgroup analyses
Inclusion of studies in subgroup analyses was dependent on the information provided within those studies. For some studies, we could not obtain the data needed to correctly categorise the study. Subgroup analyses were performed for the primary outcomes (up to six months) for the following comparisons with data available from more than one study.
-
Behavioural activation versus other control groups (other than psychological therapies) by age (under 65 and 65 and over) (Analysis 11.1Analysis 11.2)
-
Behavioural activation versus other psychological therapies by type of therapist delivering behavioural activation (specialist, specialist in training, non‐specialist) (Analysis 12.1; Analysis 12.2)
-
Behavioural activation versus other control groups by type of therapist (specialist, specialist in training, non‐specialist) (Analysis 13.1; Analysis 13.2)
-
Behavioural activation versus other control groups by severity of depression symptoms (subthreshold/ moderate to severe depression) (Analysis 14.1; Analysis 14.2)
-
Behavioural activation versus other control groups by length of therapy (one to three and more than three sessions) (Analysis 15.1)
-
Behavioural activation versus other psychological therapies by type of comparator therapy (CBT, third‐wave CBT, psychodynamic/humanist/integrative) (Analysis 16.1; Analysis 16.2)
-
Behavioural activation versus other control groups by type of other control group (treatment as usual, waiting list, no treatment, placebo, other comparator) (Analysis 17.1; Analysis 17.2)
Comparison 11. Age
There was no difference between age groups in treatment efficacy (under 65: RR 2.03, 95% CI 1.49 to 2.75, 65 and over: RR 3.32, 95% CI 0.20 to 54.59) and treatment acceptability (under 65: RR 0.83, 95% CI 0.49 to 1.40, 65 and over: RR 1.30, 95% CI 0.26 to 6.38 for behavioural activation versus comparators other than psychological therapies.
Comparisons 12 and 13. Type of therapist
There was no difference between types of therapists in treatment efficacy (specialist: RR 1.11, 95% CI 0.93 to 1.32, specialist in training: RR 1.13, 95% CI 0.85 to 1.49, non‐specialist: RR 1.30, 95% CI 0.86 to 1.98) and treatment acceptability for behavioural activation versus other psychological therapies (specialist: RR 0.88, 95% CI 0.62 to 1.25, specialist in training: RR 0.83, 95% CI 0.31 to 2.25, non‐specialist: RR 1.05, 95% CI 0.84 to 1.31).
There was no difference between types of therapists in treatment efficacy (specialist: RR 1.71, 95% CI 1.08 to 2.70, specialist in training no data, non‐specialist: RR 1.49, 95% CI 1.13 to 1.97) for behavioural activation versus comparators other than psychological therapies. Behavioural activation was less acceptable (higher percentage of dropouts) than other comparators for interventions delivered by non‐specialists (RR 2.20, 95% CI 1.06 to 4.57), and more acceptable than other comparators for interventions delivered by specialists (RR 0.65, 95% CI 0.47 to 0.89). There was no difference with behavioural activation delivered by specialists in training (RR 1.35, 95% CI 0.42 to 4.35).
Comparison 14. Severity of depressions symptoms
Behavioural activation showed greater treatment efficacy than comparators other than psychological therapies for participants with moderate to severe depression (RR 1.62, 95% CI 1.41 to 1.85), than for those with subthreshold depression (RR 1.09, 95% CI 1.01 to 1.17).
There was no difference in the treatment acceptability of behavioural activation and other comparators between participants with subthreshold depression (RR 4.30, 95% CI 0.46 to 40.44) and those with moderate to severe depression (RR 1.04, 95% CI 0.55 to 1.97).
Comparison 15. Length of therapy
No data were available to compare the treatment efficacy of behavioural activation versus comparators other than psychological therapies between short and longer length therapy.
There was no difference for treatment acceptability of behavioural activation versus comparators other than psychological therapies between a short (RR 1.03, 95% CI 0.53 to 2.03) and a longer length of therapy (RR 1.35, 95% CI 0.76 to 2.37).
Comparison 16. Type of comparator therapy
Behavioural activation showed a greater treatment efficacy when compared with psychodynamic, humanist, or integrative therapies (RR 1.50, 95% CI 1.24 to 1.81) than when compared with CBT (RR 0.99, 95% CI 0.91 to 1.07), but there was no difference for behavioural activation versus third‐wave CBT (RR 1.08, 95% CI 0.91 to 1.29).
There was no difference in treatment acceptability when comparing behavioural activation to CBT (RR 1.04, 95% CI 0.85 to 1.28), third‐wave CBT (RR 0.86, 95% CI 0.54 to 1.36), or psychodynamic, humanist, or integrative therapies (RR 0.77, 95% CI 0.44 to 1.33).
Comparison 17. Other control groups
There was no difference in treatment efficacy when comparing behavioural activation to treatment as usual (RR 1.17, 95% CI 0.95 to 1.45), waiting list (RR 1.17, 95% CI 0.70 to 1.93), no treatment (RR 2.97, 95% CI 1.42 to 6.24), medication (RR 1.77, 95% CI 1.14 to 2.76), or another comparator (RR 1.59, 95% CI 1.38 to 1.83).
There was no difference in treatment acceptability when comparing behavioural activation to treatment as usual (RR 1.50, 95% CI 0.56 to 3.99), waiting list (RR 1.17, 95% CI 0.70 to 1.93), no treatment (RR 0.97, 95% CI 0.45 to 2.09), placebo (RR 0.72, 95% CI 0.31 to 1.67), medication (RR 0.37, 95% CI 0.18 to 0.75), or another comparator (RR 2.17, 95% CI 1.04 to 4.53).
Sensitivity analyses
Comparisons 18 and 19. Study quality
Sensitivity analyses were carried out removing studies which scored 'unclear' or 'high' risk of bias for allocation concealment for the primary outcomes (Analysis 18.1; Analysis 18.2; Analysis 19.1; Analysis 19.2).
Behavioural activation was no more or less effective than other psychological therapies (RR 1.20, 95% CI 0.95 to 1.51) and there was no difference in treatment acceptability (RR 1.04, 95% CI 0.84 to 1.29).
Behavioural activation was more effective than comparators other than psychological therapies (RR 1.49, 95% CI 1.16 to 1.90) and had lower treatment acceptability than other comparators (RR 2.22, 95% CI 1.00 to 4.95).
Comparisons 20 and 21. Mode of delivery
We analysed studies which were predominantly delivered face‐to‐face, removing 10 studies which evaluated interventions mostly delivered online, over the phone, or email (Analysis 20.1; Analysis 20.2; Analysis 21.1; Analysis 21.2).
There was no difference in treatment efficacy (RR 1.09, 95% CI 0.92 to 1.29) or acceptability (RR 1.00, 95% CI 0.83 to 1.20) between behavioural activation and other psychological therapies.
Behavioural activation was more effective than other comparators (RR 1.76, 95% CI 1.50 to 2.05) and there was no difference in treatment acceptability (RR 0.85, 95% CI 0.67 to 1.08).
Severity of depression
We did not perform sensitivity analyses excluding studies with participants with subthreshold depression, because these analyses would have included the same study as those in the subgroup analyses for participants diagnosed with major depressive disorder or moderate to severe symptoms of depression (Analysis 14.1; Analysis 14.2).
Comparisons 22 and 23. Individual therapy
We performed sensitivity analyses for the primary outcomes excluding nine studies delivered in a group format and three studies delivered in a mixed individual/ group format (Analysis 22.1; Analysis 22.2; Analysis 23.1; Analysis 23.2).
There was no difference in treatment efficacy (RR 1.17, 95% CI 1.00 to 1.37) or acceptability (RR 1.05, 95% CI 0.91 to 1.22) between behavioural activation and other psychological therapies.
Behavioural activation was more effective than other comparators (RR 1.61, 95% CI 1.26 to 2.05) and there was no difference in treatment acceptability (RR 1.55, 95% CI 0.85 to 2.79).
Missing data
The impact of missing data on treatment efficacy was explored in three scenarios: intention‐to‐treat analysis, best‐case scenario, and worst‐case scenario. Not all studies contributed data to these analyses, as dropout rates were not consistently reported across trials.
Comparison 24. Intention‐to‐treat
In this scenario, we assumed that treatment was not effective for participants who dropped out after randomisation (Analysis 26.1).
There was no difference in treatment efficacy between behavioural activation and CBT (RR 0.93, 95% CI 0.83 to 1.05), behavioural activation and third‐wave CBT (RR 1.17, 95% CI 0.91 to 1.52), and behavioural activation and treatment as usual (RR 1.29, 95% CI 0.99 to 1.68).
Behavioural activation was more effective than humanistic therapy (RR 2.33, 95% CI 1.09 to 5.00).
Comparison 25. Best‐case scenario
In this scenario we assumed that treatment was effective for participants who dropped out of the behavioural activation study arm and that treatment was not effective for participants who dropped out of the comparator study arm (Analysis 27.1).
There was no difference in treatment efficacy between behavioural activation and CBT (RR 1.17, 95% CI 0.90 to 1.52). Behavioural activation was more effective than third‐wave CBT (RR 1.41, 95% CI 1.12 to 1.76), humanistic therapy (RR 3.67, 95% CI 1.83 to 7.34), and treatment as usual (RR 1.63, 95% CI 1.29 to 2.04).
Comparison 26. Worst‐case scenario
In this scenario we assumed that treatment was not effective for participants who dropped out of the behavioural activation study arm and that treatment was effective for participants who dropped out of the comparator study arm (Analysis 28.1).
There was no difference in treatment efficacy between behavioural activation and CBT (RR 0.82, 95% CI 0.58 to 1.17), third‐wave CBT (RR 0.89, 95% CI 0.73 to 1.09), humanistic therapy (RR 0.78, 95% CI 0.53 to 1.15), and treatment as usual (RR 1.14, 95% CI 0.89 to 1.46).
Discussion
Summary of main results
Results for each of the 10 comparisons are summarised in the 'Summary of findings' tables.
This review comprised 53 studies, including 26 studies published after previous reviews on this topic were conducted (Churchill 2013; Ekers 2014; Hunot 2013).
The objectives of this reveiw were to examine the effects of behavioural activation for depression in adults compared with 1) all other psychological therapies, 2) medication, and 3) other comparators (treatment as usual, waiting list, placebo, no treatment).
Behavioural activation versus psychological therapies
Trials included in this review compared behavioural activation with cognitive‐behavioural therapy (CBT), third‐wave CBT, humanistic therapy, psychodynamic therapy, and interpersonal/ cognitive analytic/ integrative therapy for the treatment of depression. Most trials reported data on short‐term outcomes only.
Primary outcomes
Moderate‐ to very low‐certainty evidence showed no difference in treatment efficacy and acceptability (dropouts) between behavioural activation and other psychological therapies, except that behavioural activation may be more effective than humanistic therapy (low‐certainty evidence).
Subgroup and sensitivity analyses
Subgroup analyses showed no difference in efficacy and acceptability when comparing behavioural activation to other psychological therapies by participant and treatment characteristics. Efficacy of behavioural activation was greater compared with psychodynamic, humanistic, and integrative therapies than it was for CBT. No difference by type of comparator therapy was found for treatment acceptability.
When considering high‐quality studies only, behavioural activation was no more or less effective or acceptable than other psychological therapies.
When considering only face‐to‐face and only individual therapies, there was no difference between behavioural activation and other therapies in terms of treatment efficacy and treatment acceptability.
When using an intention‐to‐treat approach to missing data, treatment efficacy remained higher for behavioural activation than for humanistic therapy. This was no longer the case when a worst‐case scenario was used. In a, more realistic, intention‐to‐treat analysis, behavioural activation showed no difference in effectiveness compared with CBT and third‐wave CBT.
We performed an unplanned sensitivity analysis removing one small study with a large weighting from comparison 1.1. This reduced the precision of the estimate, but did not change the finding that CBT is no more effective than behavioural activation in the treatment of depression.
Secondary outcomes
There was no evidence of a difference in our secondary outcomes between behavioural activation and other psychological therapies (moderate‐ to very low‐certainty evidence), except for depression symptoms being reduced to a greater extent with behavioural activation compared with humanistic therapy (moderate‐certainty evidence).
Adverse events were reported in various studies for participants receiving behavioural activation, CBT, general counselling, medication placebo, medication, and treatment as usual. These events included serious adverse events such as hospitalisation, suicide attempt, and suicide. Authors of various trials reported that adverse events were not thought to be related to the treatment received.
Behavioural activation versus medication
Two trials compared behavioural activation with medication for depression.
Primary outcomes
Moderate‐certainty evidence from one study suggests that behavioural activation is probably more efficacious than medication. There was moderate‐certainty evidence that behavioural activation and medication probably do not differ in terms of treatment acceptability.
Secondary outcomes
Low‐certainty evidence suggests reduction in depression symptoms did not differ between behavioural activation and medication in the short term, but favoured behavioural activation in the medium term.
Behavioural activation versus other comparators
Included trials compared behavioural activation with waiting list, placebo, no treatment, or treatment as usual.
Primary outcomes
Moderate‐certainty evidence showed better treatment efficacy for behavioural activation compared with treatment as usual in the short term and medium term. Low‐certainty evidence on the treatment efficacy of behavioural activation compared with waiting list favours behavioural activation but lacks precision.
Moderate‐ to low‐certainty evidence showed no difference in treatment acceptability between behavioural activation and comparison groups in the short term (waiting list, placebo, no treatment, treatment as usual). Treatment acceptability was higher for treatment as usual than for behavioural activation, although the short‐ and medium term estimates lacked precision.
Planned subgroup and sensitivity analyses
Subgroup analyses for participant age, length of therapy, and type of comparator showed no differences in treatment efficacy and acceptability (dropouts).
Treatment efficacy was greater for behavioural activation versus other comparators for participants with moderate to severe depression compared with those with mild or subthreshold depression. No difference by severity of depression was found for treatment acceptability.
Treatment acceptability was higher for behavioural activation than for other comparators for interventions delivered by a specialist, but lower for interventions delivered by a non‐specialist. This finding was driven by three trials of non‐specialist interventions, for which the comparator group consisted of treatment as usual by a general practitioner or minimal psychoeducation.
Sensitivity analyses of high‐quality studies, face‐to‐face therapy, and individual therapy showed benefits of behavioural activation versus other comparators for treatment efficacy, but not for treatment acceptability. When considering high‐quality studies only, behavioural activation had a lower treatment acceptability than comparators.
When re‐analysing data using three different approaches for missing data, behavioural activation was more effective than treatment as usual only when a best‐case‐scenario was used.
Secondary outcomes
Moderate‐ to low‐certainty evidence suggested that depression symptoms were reduced more for behavioural activation when compared with waiting list, treatment as usual (short and medium term but not long term), and no treatment (short and medium term), but not when compared with a placebo.
Very low‐ to high‐certainty evidence showed benefits of behavioural activation for short‐term quality of life when compared with treatment as usual and no treatment, but not when compared with waiting list. Anxiety symptoms were reduced more for behavioural activation than waiting list, treatment as usual, and no treatment (low‐ to moderate‐certainty evidence).
Low‐certainty evidence showed a benefit of behavioural activation compared with treatment as usual for short‐term social adjustment and functioning.
Unplanned sensitivity analyses
We removed one outlier in analyses of behavioural activation versus treatment as usual for depression symptoms and quality of life. The estimates for depression symptoms and quality of life were reduced as a result of this, but still showed a benefit of behavioural activation.
Analyses of behavioural activation versus waiting list and versus treatment as usual showed large beneficial effects of behavioural activation, based on a mix of studies including those with small sample sizes. We conducted fixed‐effect analyses in addition to random‐effects analyses to investigate the impact of small studies on the results of two comparisons. The estimates of behavioural activation versus waiting list for depression symptoms (Analysis 24.1) and anxiety symptoms (Analysis 24.2) were reduced, but still favoured behavioural activation. The estimates of depression symptoms (Analysis 25.1), and quality of life (Analysis 25.2) for behavioural activation versus waiting list were reduced, but still showed a benefit of behavioural activation.
Overall completeness and applicability of evidence
Most of the evidence came from studies conducted in high‐income countries, and from the USA in particular. This may make evidence from this review less applicable to Low and Middle Income Countries, where the majority of people with depression live.
In settings with less resources to deliver mental health interventions, behavioural activation may be delivered in a format which does not require a specialist, for example using lay health workers or community workers. Our subgroup analyses did not show a difference in treatment efficacy between behavioural activation delivered by specialists, specialists‐in‐training, or non‐specialists. When comparing behaviour activation with other comparators, comparisons in which behavioural activation was delivered by specialists favoured behavioural activation, while in comparisons with behavioural activation delivered by non‐specialists treatment acceptability was higher for other comparators.
We included studies of participants with moderate and severe depression, as well as subthreshold or mild symptoms of depression, to reflect variation in severity of symptoms found in clinical practice and in the general population. Subgroup analyses suggested that behavioural activation may be more effective than non‐therapy comparators for people with moderate to severe depression rather than subthreshold or mild depression.
Trial participants were not necessarily representative of the population of people with depression. This makes it difficult to apply evidence from this review to clinical practice. People with mental health problems in addition to depression, such as anxiety disorder or substance abuse, were excluded from participating in some trials. This is problematic if trial participants are more amenable to treatment than people with depression not included in these trials. As for other participant characteristics, several population groups may be overrepresented. For example, some studies included only young adults attending college or university, while others included only women. Ethnicity of trial participants was not usually reported, and for most studies in which it was reported, the majority of participants were White American or White British. Socioeconomic characteristics were also poorly reported. For trials conducted in high‐income countries there was mostly a mix of participants with different socioeconomic status, although it is difficult to assess to what extent these participant characteristics are representative of the population eligible for inclusion in trials of behavioural activation for depression.
Included trials were published between 1977 and 2020. Eligibility for inclusion was not based on date of publication in our review, in order to capture the entire evidence base. However, there may be differences in the way behavioural activation would be evaluated and offered in practice nowadays and in the past. Whereas behavioural therapy was initially based on the extraction of the behavioural component from CBT, more recently, behavioural activation has been integrated into multidisciplinary treatments or collaborative care. In our review, we only included such trials if behavioural activation was clearly specified as the main component of the intervention (Bosanquet 2017; Gilbody 2017), rather than being only one of several elements of the treatment (Richards 2013). In a future update of this review, it would be worth reconsidering selection criteria, including the publication date and scope of the intervention.
Most studies reported short‐term outcomes, within six months of starting treatment. We cannot be sure that any benefits of behavioural activation reported shortly after the treatment ends would continue over time.
Quality of the evidence
The certainty of the evidence was mostly low to moderate. This means that the effect sizes calculated in our review may deviate from the true effects of behavioural activation for depression in adults. For several comparisons, evidence for some outcomes was based on data from one trial only. This means we could not assess inconsistencies in the results between trials. For the comparison 'behavioural activation versus medication', this means the evidence for treatment efficacy was based on only one study but judged to be of moderate certainty.
The quality of the trials was limited by risk of bias relating to lack of blinding of participants and personnel, no published study protocol or trial report, missing information on incomplete outcome data, and potential conflicts of interest relating to the study authors being involved in the development of the intervention. We judged some of the estimates to be imprecise due to the limited availability of data for these outcomes.
Incomplete outcome data may have resulted in overestimation of the efficacy of behavioural activation compared with treatment as usual. In sensitivity analyses using a worst‐case scenario or intention‐to‐treat scenario, the benefit effect of behavioural activation over treatment as usual was no longer clearly observed.
Sensitivity analyses of high‐quality studies suggested that there was no difference in treatment efficacy and acceptability between behavioural activation and other psychological therapies. In these analyses, behavioural activation was more effective than non‐therapy comparators and had lower treatment acceptability. However, we used allocation concealment as a crude proxy for quality in these analyses, and all studies included in the sensitivity analyses as 'high quality' had other domains for which risk of bias was assessed to be high.
Findings were frequently found to be imprecise due to a small number of studies per comparison, particularly for the primary outcomes, and a small number of participants per study. The majority of studies had less than 20 participants per study arm. This makes it difficult to determine whether behavioural activation performs as well as other psychological therapies.
Searches
Although we are confident that our search of the literature included the most important databases and sources of clinical trials on behavioural activation for depression, we cannot rule out the possibility that relevant data were missed. For example, our search did not include databases from low‐ and middle‐income countries. In years to come, behavioural activation may be rolled out in these countries as a feasible intervention to treat depression and other mental health conditions in settings where resources are limited, and an update of this review should therefore consider a broader search including such databases.
Missing data
We contacted authors of 44 included studies for information required to complete the extraction of key data and the 'Risk of bias' assessment. Authors of 23 studies could not be contacted and authors of two studies replied, but could no longer provide the requested information. Many of these studies were published more than 20 years ago; some nearly 40 years ago. This hindered our ability to retrieve all missing data, and as a consequence many 'Risk of bias' domains remained 'unclear'. For nine studies standard deviations or sample sizes required for meta‐analyses were missing, and for four studies published between 1979 and 1983 this information could not be estimated or obtained.
Publication bias
The small number of studies for most comparisons made it difficult to assess the possibility of publication bias. There was an indication of publication bias in the comparison of behavioural activation versus waiting list for depressive symptoms, with small studies favouring waiting list missing from the review (Figure 4). This information was taken into account when assessing the certainty of the evidence, as summarised in summary of findings Table 6. Funnel plots did not indicate publication bias for treatment acceptability or depression symptoms in the behavioural activation versus CBT comparison, nor for the treatment efficacy, acceptability, and depression symptoms in the behavioural activation versus treatment as usual comparison.
Primary outcome
Our primary outcome was treatment efficacy measured by the number of people who responded to treatment. We accepted trial authors' definitions of 'treatment response'. For some of the included trials a 50% or greater reduction in symptom severity measured on a validated depressions scale was defined as response or clinically significant improvement, while other trials used recovery or remission (symptom level below the cut‐off for clinically diagnosed depression). This may have led to heterogeneity in the results. However, because effect estimates are based on comparisons between intervention and control groups in each trial, we do not expect this to substantially bias the results.
Conflict of interest
Two authors on this review (DE, DR) have been involved in multiple included trials of behavioural activation. We have reported this potential conflict of interest in the 'Risk of bias' assessments pertaining to these studies in accordance with the Conflicts of interest and Cochrane Reviews policy. DE and DR were not involved in the data extraction, 'Risk of bias' assessments, and GRADE assessments of the certainty of the evidence for this review.
Agreements and disagreements with other studies or reviews
Conclusions regarding the effectiveness of behavioural activation have been limited in previous systematic reviews by the absence of substantive, high‐certainty evidence (Churchill 2013; Hunot 2013; Shinohara 2013). No difference had previously been found in effectiveness between behavioural activation and other psychological therapies (Shinohara 2013). Our review suggested similar efficacy between behavioural activation and other psychological therapies, although our confidence in these findings is limited due to concerns about the certainty of the evidence. Moderate‐ to low‐certainty evidence suggested that behavioural activation was more effective than humanistic therapy, both in terms of depression as a binary outcome and symptoms of depression.
The most recent systematic review of behavioural activation for depression versus comparators other than psychological therapy found mostly low‐quality evidence indicating that behavioural activation was superior to a wide range of control treatments, including medication (Ekers 2014). Our review also suggests a benefit of behavioural activation in terms of treatment efficacy or depression symptoms when compared with treatment as usual or no treatment. Our review also suggested that, compared with being on a waiting list, behavioural activation improved depression symptoms, but was not necessarily better in terms of treatment efficacy (although we found only one trial that looked at this). We found no difference between behavioural activation and placebo in terms of depression symptoms, although no data were available on treatment efficacy for this comparison. Behavioural activation performed better than medication in terms of treatment efficacy, but this was based on only one trial, and we found no difference between the two interventions in relation to decreasing symptoms of depression. Any differences between reviews are most likely due to the addition of new studies, minor differences in the selection criteria, and the choice of comparisons and outcomes.
Study flow diagram.
'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.
Funnel plot of comparison: 6 behavioural activation vversus waiting list, outcome: 6.3 depression symptoms.
Comparison 1: behavioural activation vs CBT, Outcome 1: treatment efficacy
Comparison 1: behavioural activation vs CBT, Outcome 2: treatment acceptability (dropouts)
Comparison 1: behavioural activation vs CBT, Outcome 3: depression symptoms
Comparison 1: behavioural activation vs CBT, Outcome 4: quality of life
Comparison 1: behavioural activation vs CBT, Outcome 5: social adjustment and functioning
Comparison 1: behavioural activation vs CBT, Outcome 6: anxiety symptoms
Comparison 2: behavioural activation vs third‐wave CBT, Outcome 1: treatment efficacy
Comparison 2: behavioural activation vs third‐wave CBT, Outcome 2: treatment acceptability (dropouts)
Comparison 2: behavioural activation vs third‐wave CBT, Outcome 3: depression symptoms
Comparison 2: behavioural activation vs third‐wave CBT, Outcome 4: quality of life
Comparison 2: behavioural activation vs third‐wave CBT, Outcome 5: anxiety symptoms
Comparison 3: behavioural activation vs humanistic therapy, Outcome 1: treatment efficacy
Comparison 3: behavioural activation vs humanistic therapy, Outcome 2: treatment acceptability (dropouts)
Comparison 3: behavioural activation vs humanistic therapy, Outcome 3: depression symptoms
Comparison 3: behavioural activation vs humanistic therapy, Outcome 4: quality of life
Comparison 3: behavioural activation vs humanistic therapy, Outcome 5: anxiety symptoms
Comparison 4: behavioural activation vs psychodynamic, Outcome 1: treatment efficacy
Comparison 4: behavioural activation vs psychodynamic, Outcome 2: depression symptoms
Comparison 4: behavioural activation vs psychodynamic, Outcome 3: social adjustment and functioning
Comparison 5: behavioural activation vs interpersonal, cognitive analytic, integrative, Outcome 1: treatment acceptability (dropouts)
Comparison 5: behavioural activation vs interpersonal, cognitive analytic, integrative, Outcome 2: depression symptoms
Comparison 5: behavioural activation vs interpersonal, cognitive analytic, integrative, Outcome 3: social adjustment and functioning
Comparison 5: behavioural activation vs interpersonal, cognitive analytic, integrative, Outcome 4: anxiety symptoms
Comparison 6: behavioural activation vs waiting list, Outcome 1: treatment efficacy
Comparison 6: behavioural activation vs waiting list, Outcome 2: treatment acceptability (dropouts)
Comparison 6: behavioural activation vs waiting list, Outcome 3: depression symptoms
Comparison 6: behavioural activation vs waiting list, Outcome 4: quality of life
Comparison 6: behavioural activation vs waiting list, Outcome 5: anxiety symptoms
Comparison 7: behavioural activation vs placebo, Outcome 1: treatment acceptability (dropouts)
Comparison 7: behavioural activation vs placebo, Outcome 2: depression symptoms
Comparison 8: behavioural activation vs medication, Outcome 1: treatment efficacy
Comparison 8: behavioural activation vs medication, Outcome 2: treatment acceptability (dropouts)
Comparison 8: behavioural activation vs medication, Outcome 3: depression symptoms
Comparison 9: behavioural activation vs no treatment, Outcome 1: treatment acceptability (dropouts)
Comparison 9: behavioural activation vs no treatment, Outcome 2: depression symptoms
Comparison 9: behavioural activation vs no treatment, Outcome 3: quality of life
Comparison 9: behavioural activation vs no treatment, Outcome 4: anxiety symptoms
Comparison 10: behavioural activation vs treatment as usual, Outcome 1: treatment efficacy
Comparison 10: behavioural activation vs treatment as usual, Outcome 2: treatment acceptability (dropouts)
Comparison 10: behavioural activation vs treatment as usual, Outcome 3: depression symptoms
Comparison 10: behavioural activation vs treatment as usual, Outcome 4: quality of life
Comparison 10: behavioural activation vs treatment as usual, Outcome 5: social adjustment and functioning
Comparison 10: behavioural activation vs treatment as usual, Outcome 6: anxiety symptoms
Comparison 11: SUBGROUP 1 AGE behavioural activation vs other controls (up to 6 months), Outcome 1: treatment efficacy
Comparison 11: SUBGROUP 1 AGE behavioural activation vs other controls (up to 6 months), Outcome 2: treatment acceptability (dropouts)
Comparison 12: SUBGROUP 2 THERAPIST behavioural activation vs other psychological therapies (up to 6 months), Outcome 1: treatment efficacy
Comparison 12: SUBGROUP 2 THERAPIST behavioural activation vs other psychological therapies (up to 6 months), Outcome 2: treatment acceptability (dropouts)
Comparison 13: SUBGROUP 2 THERAPIST behavioural activation vs other controls (up to 6 months), Outcome 1: treatment efficacy
Comparison 13: SUBGROUP 2 THERAPIST behavioural activation vs other controls (up to 6 months), Outcome 2: treatment acceptability (dropouts)
Comparison 14: SUBGROUP 3 SEVERITY behavioural activation vs other controls (up to 6 months), Outcome 1: treatment efficacy
Comparison 14: SUBGROUP 3 SEVERITY behavioural activation vs other controls (up to 6 months), Outcome 2: treatment acceptability (dropouts)
Comparison 15: SUBGROUP 4 LENGTH behavioural activation vs other controls (up to 6 months), Outcome 1: treatment acceptability (dropouts)
Comparison 16: SUBGROUP 5 THERAPY behavioural activation vs other psychological therapies (up to 6 months), Outcome 1: treatment efficacy
Comparison 16: SUBGROUP 5 THERAPY behavioural activation vs other psychological therapies (up to 6 months), Outcome 2: treatment acceptability (dropouts)
Comparison 17: SUBGROUP 6 CONTROL behavioural activation vs other controls (up to 6 months), Outcome 1: treatment efficacy
Comparison 17: SUBGROUP 6 CONTROL behavioural activation vs other controls (up to 6 months), Outcome 2: treatment acceptability (dropouts)
Comparison 18: SENSITIVITY 1 HIGH QUALITY STUDIES behavioural activation versus other psychological therapies (up to 6 months), Outcome 1: treatment efficacy
Comparison 18: SENSITIVITY 1 HIGH QUALITY STUDIES behavioural activation versus other psychological therapies (up to 6 months), Outcome 2: treatment acceptability (dropouts)
Comparison 19: SENSITIVITY 2 HIGH QUALITY STUDIES behavioural activation versus other controls (up to 6 months), Outcome 1: treatment efficacy
Comparison 19: SENSITIVITY 2 HIGH QUALITY STUDIES behavioural activation versus other controls (up to 6 months), Outcome 2: treatment acceptability (dropouts)
Comparison 20: SENSITIVITY 3 FACE‐TO‐FACE behavioural activation vs other psychological therapies (up to 6 months), Outcome 1: treatment efficacy
Comparison 20: SENSITIVITY 3 FACE‐TO‐FACE behavioural activation vs other psychological therapies (up to 6 months), Outcome 2: treatment acceptability (dropouts)
Comparison 21: SENSITIVITY 4 FACE‐TO‐FACE behavioural activation vs other controls (up to 6 months), Outcome 1: treatment efficacy
Comparison 21: SENSITIVITY 4 FACE‐TO‐FACE behavioural activation vs other controls (up to 6 months), Outcome 2: treatment acceptability (dropouts)
Comparison 22: SENSITIVITY 5 INDIVIDUAL behavioural activation versus other psychological therapies (up to 6 months), Outcome 1: treatment efficacy
Comparison 22: SENSITIVITY 5 INDIVIDUAL behavioural activation versus other psychological therapies (up to 6 months), Outcome 2: treatment acceptability (dropouts)
Comparison 23: SENSITIVITY 5 INDIVIDUAL behavioural activation versus other controls (up to 6 months), Outcome 1: treatment efficacy
Comparison 23: SENSITIVITY 5 INDIVIDUAL behavioural activation versus other controls (up to 6 months), Outcome 2: treatment acceptability (dropouts)
Comparison 24: SENSITIVITY 6 fixed effects BA vs waiting list, Outcome 1: depression symptoms
Comparison 24: SENSITIVITY 6 fixed effects BA vs waiting list, Outcome 2: anxiety symptoms
Comparison 25: SENSITIVITY 7 fixed effects BA vs treatment as usual, Outcome 1: depression symptoms
Comparison 25: SENSITIVITY 7 fixed effects BA vs treatment as usual, Outcome 2: quality of life
Comparison 26: MISSING DATA ITT (up to 6 months), Outcome 1: treatment efficacy
Comparison 27: MISSING DATA BEST CASE (up to 6 months), Outcome 1: treatment efficacy
Comparison 28: MISSING DATA WORST CASE (up to 6 months), Outcome 1: treatment efficacy
| Behavioural activation compared with CBT for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with CBT | Risk with behavioural activation | |||||
| treatment efficacy | Study population | RR 0.99 | 601 | ⊕⊕⊕⊝ | ||
| 62 per 100 | 61 per 100 | |||||
| treatment acceptability up to 6 months (4‐16 weeks) | Study population | RR 1.03 | 1195 | ⊕⊕⊝⊝ | ||
| 23 per 100 | 24 per 100 | |||||
| depression symptoms (continuous) up to 6 months (4‐16 weeks) | see comment | SMD 0.12 higher | 1205 | ⊕⊕⊕⊝ | Measured with BDI, HRSD, CES‐D, PHQ‐9, HSCL‐25. SMD 0.12 represents a difference between groups of 1.31 points on the BDI and 0.66 points on the HRSD favouring CBT. | |
| quality of life (continuous) | see comment | SMD 0.04 higher (0.20 lower to 0.28 higher) | 268 (2 RCTs) | ⊕⊕⊕⊝ | Measured with SF‐36 physical component and WHOQOL physical component. SMD 0.04 represents a small effect. | |
| social adjustment and functioning (continuous) | see comment | SMD 0.13 lower (0.50 lower to 0.24 higher) | 111 (2 RCTs) | ⊕⊝⊝⊝ | Measured with Social Adjustment Scale and Sheehan Disability Scale. SMD 0.13 represents a small effect. | |
| anxiety symptoms (continuous) up to 6 months (4‐16 weeks) | see comment | SMD 0.03 lower | 646 | ⊕⊕⊕⊝ | Measured with BDI, HSCL‐25, PHQ‐9. SMD 0.03 represents a small effect. | |
| adverse events | 1 study no adverse events, 1 study three serious adverse events in the behavioural activation arm (2 overdose, 1 self‐harm) and eight serious adverse events in the CBT arm (7 overdose, 1 self‐harm). | 398 (2 RCTs) | ⊕⊕⊕⊝ | Any adverse event summarised narratively. | ||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Majority of domains high or unclear risk of bias. High risk for conflict of interest, blinding of participants and personnel, and incomplete outcome data. Downgraded by one level for high risk of bias. 2 No blinding of participants. Reporting bias unclear because protocol or trial registration missing in nine studies and high risk of bias in one study. Potential conflict of interest in four4 studies. High risk of attrition bias in seven studies. Downgraded by one level for high risk of bias (not two levels because trials with higher weight are generally at lower risk of bias). 3 Seven out of 12 studies wide confidence intervals, due to small sample sizes and low rates of dropout in both groups. Downgraded by one level for imprecision. 4 No blinding of participants. 6/15 studies no blinding of outcome assessors. 13/15 selective reporting domain unclear. Downgraded by one level for high risk of bias. 5 Risk of performance and attrition bias and potential conflict of interest. Downgraded by one level for high risk of bias. 6 Two small studies with serious risk of bias across domains (attrition bias, reporting bias, potential conflicts of interest). Downgraded by one level for imprecision and two levels for high risk of bias. 7 One study all domains unclear or high; high risk of bias for randomisation, allocation, and blinding of participants and personnel. One study with risk of performance and attrition bias and potential conflict of interest. Downgraded one level for high risk of bias.Two studies with most domains unclear or high have little weight in the analyses. 8 Various domains high risk of bias in all studies. Attrition bias high in both studies; dropout may be related to adverse events. Downgraded one level for high risk of bias. | ||||||
| Behavioural activation compared with third‐wave CBT for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with third‐wave CBT | Risk with behavioural activation | |||||
| treatment efficacy up to 6 months (4‐8 weeks) | Study population | RR 1.10 (0.91 to 1.33) | 98 | ⊕⊕⊝⊝ | ||
| 74 per 100 | 81 per 100 (67 to 98) | |||||
| treatment acceptability up to 6 months (4‐8 weeks) | Study population | RR 0.84 (0.33 to 2.10) | 147 | ⊕⊕⊝⊝ | ||
| 12 per 100 | 10 per 100 (4 to 25) | |||||
| depression symptoms (continuous) up to 6 months (4‐8 weeks) | see comment | SMD 0.14 lower | 147 | ⊕⊕⊝⊝ | Measured with BDI and HRSD. SMD 0.14 represents a difference between groups of 1.53 points on the BDI and 0.77 points on the HRSD favouring BA. | |
| quality of life (continuous) | mean score 1.13 | MD 0.02 higher (0.96 lower to 1.00 higher) | 81 (1 RCT) | ⊕⊕⊝⊝ | Measured with Quality of Life Inventory. | |
| anxiety symptoms (continuous) | see comment | MD 0.69 higher (0.68 lower to 2.06 higher) | 147 | ⊕⊕⊝⊝ | Measured with BAI. | |
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Evidence of selective reporting and conflict of interest in both trials, in addition to other domains with risk of bias. Downgraded one level for high risk of bias. 2 Two small studies with wide confidence intervals. Downgraded one level for imprecision. 3 Ten domains with high risk of bias across three studies, including blinding, allocation concealment, and selective reporting. Treatment acceptability may be affected by lack of blinding and allocation concealment in particular. Downgraded one level for high risk of bias. 4 Three small studies with wide confidence intervals. Downgraded one level for imprecision. 5 One small study with three domains at high risk of bias. Downgraded one level for imprecision and one level for high risk of bias. Because only one study was included, this outcome could not be assessed for consistency of results. | ||||||
| Behavioural activation compared with humanistic therapy for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with humanistic therapy | Risk with behavioural activation | |||||
| treatment efficacy | Study population | RR 1.84 | 46 | ⊕⊕⊝⊝ | Number needed to treat to achieve one beneficial outcome is 2.5. | |
| 48 per 100 | 88 per 100 | |||||
| treatment acceptability | Study population | RR 1.06 | 96 | ⊕⊝⊝⊝ | ||
| 25 per 100 | 26 per 100 | |||||
| depression symptoms | mean score between 10 and 15 | MD 3.75 lower (6.72 lower to 0.78 lower) | 93 | ⊕⊕⊕⊝ | Measured with BDI. | |
| quality of life (continuous) | mean score 1.2 | MD 0.80 higher (0.12 lower to 1.72 higher) | 50 (1 RCT) | ⊕⊕⊝⊝ | Measured with Quality of Life Inventory. | |
| anxiety symptoms (continuous) | mean score 9.7 | MD 1.30 lower (6.10 lower to 3.50 higher) | 50 (1 RCT) | ⊕⊕⊝⊝ | Measured with BAI. | |
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Both small studies with several domains high risk of bias or unclear. Risk of attrition bias in both studies and reporting bias in one study may affect treatment efficacy outcome. Downgraded one level for high risk of bias and one level for imprecision. 2 Many risk of bias domains unclear in one of the studies. Risk of attrition and reporting bias in the other study. Downgraded one level for high risk of bias and one level for imprecision. 3 One study is religious behavioural activation rather than the conventional behavioural activation intervention. Downgraded one level for indirectness. 4 Two out of three studies mostly high and unclear risk of bias domains. Downgraded one level for high risk of bias. 5 One small study with most domains of the risk of bias tool unclear due to lack of information. Downgraded one level for high risk of bias and one level for imprecision. Because only one study was included, this outcome could not be assessed for consistency of results. | ||||||
| Behavioural activation compared with psychodynamic for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with psychodynamic | Risk with behavioural activation | |||||
| treatment efficacy up to 6 months (12 weeks) | Study population | RR 1.21 | 60 | ⊕⊝⊝⊝ | ||
| 47 per 100 | 56 per 100 | |||||
| depression symptoms | mean score 10 | MD 1.10 lower (4.35 lower to 2.15 higher) | ‐ | 60 | ⊕⊝⊝⊝ | Measured with HRSD |
| social adjustment and functioning | mean score 69 | MD 2.10 higher (4.92 lower to 9.12 higher) | ‐ | 60 | ⊕⊝⊝⊝ | Measured with Global Assessment Scale and Social Adjustment Scale (measures combined) |
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Four 'Risk of bias' domains unclear due to lack of information. Patients excluded from study for lack of adherence and because of dissatisfaction with treatment. This may influence outcomes treatment efficacy, depression symptoms, and social adjustment and functioning. All other 'Risk of bias' domains high or unclear risk. Downgraded two levels for high risk of bias. 2 Only one study with small sample size. Downgraded one level for imprecision. Because only one study was included, this outcome could not be assessed for consistency of results. | ||||||
| Behavioural activation compared with interpersonal, cognitive analytic, integrative for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with interpersonal, cognitive analytic, integrative | Risk with behavioural activation | |||||
| treatment acceptability | Study population | RR 0.84 | 123 | ⊕⊝⊝⊝ | ||
| 16 per 100 | 14 per 100 | |||||
| depression symptoms (continuous) up to 6 months (4‐12 weeks) | see comment | SMD 0.16 lower | 103 | ⊕⊝⊝⊝ | Measured with BDI, Zung rating scale, and HRSD. SMD 0.16 represents a difference between groups of 1.75 points on the BDI and 0.88 points on the HRSD favouring BA. | |
| social adjustment and functioning (continuous) | mean score 79 | MD 3.92 lower (16.78 lower to 8.93 higher) | 39 (1 RCT) | ⊕⊝⊝⊝ | Measured with Global Assessment Scale. | |
| anxiety symptoms (continuous) up to 6 months (4 weeks) | mean score 48 | MD 0.39 lower (11.78 lower to 11.00 higher) | 15 (1 RCT) | ⊕⊝⊝⊝ | Measured with the anxiety scale of the Multiple Affect Adjective Check List. | |
| adverse events | 2 suicide attempts and 1 case of suicidal thoughts in comparator arm; no adverse events in behavioural activation arm. | 24 | ⊕⊕⊝⊝ | Any adverse event summarised narratively. | ||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Only 2 low risk of bias domains across four studies. High risk of bias for randomisation and allocation concealment in 2/4 studies. Downgraded two levels for high risk of bias. Downgraded one level for imprecision because of wide confidence intervals. 2 One small study with high risk of bias across multiple domains. Downgraded two levels for high risk of bias and one level for imprecision. Because only one study was included, this outcome could not be assessed for consistency of results. 3 One very small study, so adverse events reported may not apply to a wider population receiving treatment. High risk of bias included lack of blinding and potential attrition bias and selective reporting may influence reporting of adverse events. Downgraded one level for imprecision and one level for high risk of bias. | ||||||
| Behavioural activation compared with waiting list for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with waiting list | Risk with behavioural activation | |||||
| treatment efficacy | Study population | RR 2.14 (0.90 to 5.09) | 26 (1 RCT) | ⊕⊕⊝⊝ | ||
| 33 per 100 | 71 per 100 (30 to 100) | |||||
| treatment acceptability | Study population | RR 1.17 | 359 | ⊕⊕⊕⊝ | Three studies could not be included in meta‐analyses; no dropouts. | |
| 12 per 100 | 14 per 100 | |||||
| depression symptoms | see comment | SMD 1.04 lower | 619 | ⊕⊕⊝⊝ | Measured with BDI, HRSD, MADRS, PHQ‐9, HSCL‐25. SMD 1.04 represents a difference between groups of 11.37 points on the BDI and 5.75 points on the HRSD favouring BA. | |
| quality of life (continuous) | mean score 0.75 | MD 0.03 higher (0.70 lower to 0.76 higher) | 80 (1 RCT) | ⊕⊕⊝⊝ | Measured with quality of life inventory. | |
| anxiety symptoms | see comment | SMD 0.91 lower | 424 | ⊕⊕⊝⊝ | Measured with BAI, Trait Anxiety Scale, and GAD‐7. SMD 0.91 represents a large effect. | |
| adverse events | see comment | 0 (1 RCT) | see comment | Any adverse event summarised narratively. No adverse events. | ||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 One very small study with high risk of bias for five domains including allocation concealment and selective reporting. Downgraded one level for imprecision and one level for high risk of bias. 2 Most studies high or unclear risk of bias with regard to blinding of participants and outcome assessors, selective reporting, and various other risks of bias: no baseline characteristics reported, potential conflicts of interest. Downgraded one level for high risk of bias. 3 Only domain mostly scoring low risk of bias across studies (7/12) is random sequence generation. Blinding of outcome assessors unclear or high risk of bias in all but two studies. Downgraded one level for high risk of bias. 4 Larger effects reported by smaller studies; smaller studies favouring waiting list are absent. Downgraded one level for risk of publication bias. 5 One study with high risk of bias for three domains including potential conflict of interest. Downgraded one level for high risk of bias and one level for imprecision. Because only one study was included, this outcome could not be assessed for consistency of results. 6 All studies majority of domains unclear or high risk of bias. Some problems with randomisation and allocation concealment. Downgraded one level for high risk of bias. 7 Two studies reporting large effect in favour of behavioural activation while three find no difference between study arms. Downgraded one level for inconsistency. | ||||||
| Behavioural activation compared with placebo for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with behavioural activation | |||||
| treatment acceptability | Study population | RR 0.72 (0.31, 1.67) | 96 (1 RCT) | ⊕⊕⊝⊝ | ||
| 23 per 100 | 16 per 100 (7 to 38) | |||||
| depression symptoms | see comment | SMD 0.18 lower | 108 | ⊕⊕⊝⊝ | Measured with HRSD and Depression Adjective Checklist. SMD 0.18 represents a difference between groups of 1.97 points on the BDI and 1.00 point on the HRSD favouring BA. | |
| adverse events | Various physical side effects from placebo. | 96 | ⊕⊕⊝⊝ | Any adverse event summarised narratively. | ||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Incomplete outcome data influences reporting of dropouts; downgraded one level for high risk of bias. Downgraded one level for imprecision due to large confidence intervals resulting from relatively few dropouts. Because only one study was included, this outcome could not be assessed for consistency of results. 2 One study with poor reporting, which may indicate high risk of bias. Downgraded one level for high risk of bias. 3 Two small studies; one with large confidence intervals. Downgraded one level for imprecision. 4 Incomplete outcome data and potential conflict of interest may have influenced reporting of adverse events. Downgraded one level for high risk of bias. Downgraded one level for imprecision as 96 participants would not be sufficient to measure less frequently occurring side effects. | ||||||
| Behavioural activation compared with medication for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with medication | Risk with behavioural activation | |||||
| Study population | RR 1.77 (1.14 to 2.76) | 141 (1 RCT) | ⊕⊕⊕⊝ | |||
| treatment efficacy up to 6 months (16 weeks) | 28 per 100 | 49 per 100 (31 to 76) | ||||
| treatment acceptability | 34 per 100 | 18 per 100 | RR 0.52 | 243 | ⊕⊕⊕⊝ | |
| depression symptoms | mean change in score between ‐8 and ‐14 | mean difference 1.42 lower (4.80 lower to 1.96 higher) | 180 | ⊕⊕⊝⊝ | Measured with HRSD. | |
| adverse events | Various physical side effects from antidepressant medication. One suicide in antidepressant arm. | 143 | ⊕⊕⊕⊝ | Any adverse event summarised narratively. | ||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 The most concerning issue relating to risk of bias was the large number of dropouts from the medication study arm in particular. Downgraded one level for high risk of bias. Because only one study was included, this outcome could not be assessed for consistency of results. 2 Incomplete outcome data for both studies. No blinding of participants. Potential conflict of interest for one study. Downgraded one level for high risk of bias. 3 Downgraded one level for imprecision; large variation in confidence interval, crossing zero. 4 Incomplete outcome data and potential conflict of interest may have influenced reporting of adverse events. Downgraded one level for high risk of bias. | ||||||
| Behavioural activation compared with no treatment for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with no treatment | Risk with behavioural activation | |||||
| treatment acceptability | Study population | RR 0.97 | 187 | ⊕⊕⊕⊝ | ||
| 9 per 100 | 9 per 100 | |||||
| depression symptoms | see comment | MD 6.10 lower (7.87 lower to 4.33 lower) | 187 | ⊕⊕⊕⊝ | Measured with BDI | |
| quality of life (continuous) | mean score 0.9 | MD 0.07 higher (0.03 higher to 0.11 higher) | 118 (1 RCT) | ⊕⊕⊕⊕ | Measured with EQ‐5D | |
| anxiety symptoms | mean score 11 | MD 5.50 lower (10.01 lower to 0.99 lower) | 30 (1 RCT) | ⊕⊕⊝⊝ | Measured with BAI | |
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 High risk of bias for blinding of participants (3/3), conflict of interest (1/3), and no baseline characteristics reported (1/3). Downgraded one level for high risk of bias. 2 One study mostly low risk of bias, one study mostly unclear risk of bias. Downgraded one level for high risk of bias. 3 Because only one study was included, this outcome could not be assessed for consistency of results. 4 One small study with four domains of bias unclear and two high risk of bias; performance bias and potential conflict of interest. Downgraded one level for high risk of bias and one level for imprecision. Because only one study was included, this outcome could not be assessed for consistency of results. | ||||||
| Behavioural activation compared with treatment as usual for depression in adults | ||||||
| Patient or population: depression in adults | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with treatment as usual | Risk with behavioural activation | |||||
| treatment efficacy | Study population | RR 1.40 | 1533 | ⊕⊕⊕⊝ | Number needed to treat to achieve one beneficial outcome is 4.5 | |
| 53 per 100 | 75 per 100 | |||||
| treatment acceptability | Study population | RR 1.64 | 2518 | ⊕⊕⊕⊝ | ||
| 6 per 100 | 11 per 100 | |||||
| depression symptoms (continuous) up to 6 months (5‐12 weeks) | see comment | SMD 0.78 lower | 2208 | ⊕⊕⊝⊝ | Measured with PHQ‐9, CES‐D, BDI, HRSD, and GDS. SMD 0.78 represents a difference between groups of 8.53 points on the BDI and 4.31 points on the HRSD. | |
| quality of life (continuous) | see comment | SMD 0.97 higher | 1299 | ⊕⊝⊝⊝ | Measured with SF‐12 physical component and WHOQOL. SMD 0.97 represents a large effect. | |
| social adjustment and functioning (continuous) | see comment | SMD 1.27 lower (1.74 lower to 0.84 lower) | 88 (2 RCTs) | ⊕⊕⊝⊝ | Measured with Work and Social Adjustment Scale (WSAS) and Sheehan Disability Scale. SMD 1.27 represents a large effect. | |
| anxiety symptoms | see comment | SMD 0.33 lower | 1063 | ⊕⊕⊕⊝ | Measured with GAD‐7 and BAI. SMD 0.33 represents a small effect. | |
| adverse events | behavioural activation arm: 103 events. treatment as usual arm: 107 events. | 1471 (3 RCTs) | ⊕⊕⊕⊝ | Any adverse event summarised narratively | ||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1 Mostly low risk of bias for sequence generation, allocation concealment, and selective reporting. Mostly high risk of bias only for blinding of participants and personnel. Some evidence of incomplete outcome data. Downgraded one level for high risk of bias. 2 Several studies with incomplete outcome data and potential conflict of interest. Randomisation and allocation concealment largely low risk of bias. Downgraded one level for high risk of bias. 3 Pooled estimate is influenced by large effect in one small study. Downgraded one level for inconsistency. 4 Pooled estimate is driven by one study with large effect favouring behavioural activation. Wide confidence interval. Downgraded one level for inconsistency and one level for imprecision. 5 One small study with three high risk of bias domains including incomplete outcome data. Other study unclear risk of selection bias, and high risk for attrition bias, reporting bias, and conflict of interest. Downgraded two levels for high risk of bias. Although studies are small, estimates are consistent. 6 No blinding of participants/personnel and outcome assessors in 3/4 studies. Evidence of attrition bias (2/4), performance bias (3/4) and potential conflict of interest (3/4). No evidence of selection bias in 2/4 studies, other two studies some information missing (unclear). Downgraded one level for high risk of bias. 7 One out of three studies with selective reporting, attrition bias, and potential conflict of interest. One study potential conflict of interest. Downgraded one level for high risk of bias. | ||||||
| First author | Year of publication | Comparator group(s) | Description of adverse events (at end of study period) |
| Bosanquet | 2017 | Treatment as usual | BA: 47 suspected adverse events. Usual care: 34 suspected adverse events. Elderly sample. |
| Dimidjian | 2006 | CBT, medication, medical placebo | various physical side effects from antidepressant medication and placebo. 1 suicide in antidepressant arm. |
| Gilbody | 2017 | Treatment as usual | BA: 37 events; 35 unrelated to intervention and 2 unlikely to be related to intervention. Usual care: 44 events; 40 unrelated and 4 unlikely to be related to intervention. 18 patients died (elderly sample). |
| Padfield | 1976 | Interpersonal, cognitive analytic, integrative | 2 suicide attempts and 1 case of suicidal thoughts in comparator arm; no adverse events in behavioural activation arm. |
| Richards | 2017 | CBT | 3 serious adverse events in behavioural activation arm (2 overdose, 1 self‐harm) and 8 serious adverse events in comparator arm (7 overdose, 1 self‐harm). |
| Stiles‐Shields | 2018 | CBT and waiting list | No adverse events. |
| Weobong | 2017 | Treatment as usual | 1 suicide attempt and 18 unplanned hospitalisations in behavioural activation arm. 1 suicide attempt, 26 unplanned hospitalisations, and 2 deaths in comparator arm. |
| BA: Behavioural activation; CBT: cognitive‐behavioural therapy; | |||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1.1 treatment efficacy Show forest plot | 5 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 1.1.1 Short‐term (up to 6 months) | 5 | 601 | Risk Ratio (IV, Random, 95% CI) | 0.99 [0.92, 1.07] |
| 1.1.2 Medium‐term (7‐12 months) | 1 | 364 | Risk Ratio (IV, Random, 95% CI) | 1.00 [0.86, 1.16] |
| 1.1.3 Long‐term (>12 months) | 1 | 356 | Risk Ratio (IV, Random, 95% CI) | 0.93 [0.81, 1.08] |
| 1.2 treatment acceptability (dropouts) Show forest plot | 12 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 1.2.1 Short‐term (up to 6 months) | 12 | 1195 | Risk Ratio (IV, Random, 95% CI) | 1.03 [0.85, 1.25] |
| 1.2.2 Medium‐term (7‐12 months) | 1 | 440 | Risk Ratio (IV, Random, 95% CI) | 1.25 [0.97, 1.62] |
| 1.2.3 Long‐term (>12 months) | 1 | 440 | Risk Ratio (IV, Random, 95% CI) | 1.16 [0.90, 1.49] |
| 1.3 depression symptoms Show forest plot | 16 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 1.3.1 Short‐term (up to 6 months) | 16 | 1205 | Std. Mean Difference (IV, Random, 95% CI) | 0.12 [‐0.08, 0.32] |
| 1.3.2 Medium‐term (7‐12 months) | 1 | 380 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.18 [‐0.38, 0.02] |
| 1.3.3 Long‐term (>12 months) | 1 | 364 | Std. Mean Difference (IV, Random, 95% CI) | 0.00 [‐0.21, 0.21] |
| 1.4 quality of life Show forest plot | 2 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 1.4.1 Short‐term (up to 6 months) | 2 | 268 | Std. Mean Difference (IV, Random, 95% CI) | 0.04 [‐0.20, 0.28] |
| 1.4.2 Medium‐term (7‐12 months) | 1 | 318 | Std. Mean Difference (IV, Random, 95% CI) | 0.15 [‐0.07, 0.37] |
| 1.4.3 Long‐term (>12 months) | 1 | 327 | Std. Mean Difference (IV, Random, 95% CI) | 0.06 [‐0.15, 0.28] |
| 1.5 social adjustment and functioning Show forest plot | 2 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 1.5.1 Short‐term (up to 6 months) | 2 | 111 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.13 [‐0.50, 0.24] |
| 1.6 anxiety symptoms Show forest plot | 4 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 1.6.1 Short‐term (up to 6 months) | 4 | 646 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.03 [‐0.18, 0.13] |
| 1.6.2 Medium‐term (7‐12 months) | 1 | 337 | Std. Mean Difference (IV, Random, 95% CI) | 0.02 [‐0.20, 0.23] |
| 1.6.3 Long‐term (>12 months) | 1 | 332 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.10 [‐0.31, 0.12] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 2.1 treatment efficacy Show forest plot | 2 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 2.1.1 Short‐term (up to 6 months) | 2 | 98 | Risk Ratio (IV, Random, 95% CI) | 1.10 [0.91, 1.33] |
| 2.2 treatment acceptability (dropouts) Show forest plot | 3 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 2.2.1 Short‐term (up to 6 months) | 3 | 147 | Risk Ratio (IV, Random, 95% CI) | 0.84 [0.33, 2.10] |
| 2.3 depression symptoms Show forest plot | 3 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 2.3.1 Short‐term (up to 6 months) | 3 | 147 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.14 [‐0.47, 0.18] |
| 2.4 quality of life Show forest plot | 1 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 2.4.1 Short‐term (up to 6 months) | 1 | 81 | Mean Difference (IV, Random, 95% CI) | 0.02 [‐0.96, 1.00] |
| 2.5 anxiety symptoms Show forest plot | 3 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 2.5.1 Short‐term (up to 6 months) | 3 | 147 | Mean Difference (IV, Random, 95% CI) | 0.69 [‐0.68, 2.06] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 3.1 treatment efficacy Show forest plot | 2 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 3.1.1 Short‐term (up to 6 months) | 2 | 46 | Risk Ratio (IV, Random, 95% CI) | 1.84 [1.15, 2.95] |
| 3.2 treatment acceptability (dropouts) Show forest plot | 2 | 96 | Risk Ratio (IV, Random, 95% CI) | 1.06 [0.20, 5.55] |
| 3.2.1 Short‐term (up to 6 months) | 2 | 96 | Risk Ratio (IV, Random, 95% CI) | 1.06 [0.20, 5.55] |
| 3.3 depression symptoms Show forest plot | 3 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 3.3.1 Short‐term (up to 6 months) | 3 | 93 | Mean Difference (IV, Random, 95% CI) | ‐3.75 [‐6.72, ‐0.78] |
| 3.4 quality of life Show forest plot | 1 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 3.4.1 Short‐term (up to 6 months) | 1 | 50 | Mean Difference (IV, Random, 95% CI) | 0.80 [‐0.12, 1.72] |
| 3.5 anxiety symptoms Show forest plot | 1 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 3.5.1 Short‐term (up to 6 months) | 1 | 50 | Mean Difference (IV, Random, 95% CI) | ‐1.30 [‐6.10, 3.50] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 4.1 treatment efficacy Show forest plot | 1 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 4.1.1 Short‐term (up to 6 months) | 1 | 60 | Risk Ratio (IV, Random, 95% CI) | 1.21 [0.74, 1.99] |
| 4.2 depression symptoms Show forest plot | 1 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 4.2.1 Short‐term (up to 6 months) | 1 | 60 | Mean Difference (IV, Random, 95% CI) | ‐1.10 [‐4.35, 2.15] |
| 4.3 social adjustment and functioning Show forest plot | 1 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 4.3.1 Short‐term (up to 6 months) | 1 | 60 | Mean Difference (IV, Random, 95% CI) | 2.10 [‐4.92, 9.12] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 5.1 treatment acceptability (dropouts) Show forest plot | 4 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 5.1.1 Short‐term (up to 6 months) | 4 | 123 | Risk Ratio (IV, Random, 95% CI) | 0.84 [0.32, 2.20] |
| 5.2 depression symptoms Show forest plot | 4 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 5.2.1 Short‐term (up to 6 months) | 4 | 103 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.16 [‐0.59, 0.28] |
| 5.3 social adjustment and functioning Show forest plot | 1 | 39 | Mean Difference (IV, Random, 95% CI) | ‐3.92 [‐16.78, 8.93] |
| 5.3.1 Short‐term (up to 6 months) | 1 | 39 | Mean Difference (IV, Random, 95% CI) | ‐3.92 [‐16.78, 8.93] |
| 5.4 anxiety symptoms Show forest plot | 1 | 15 | Mean Difference (IV, Fixed, 95% CI) | ‐0.39 [‐11.78, 11.00] |
| 5.4.1 Short‐term (up to 6 months) | 1 | 15 | Mean Difference (IV, Fixed, 95% CI) | ‐0.39 [‐11.78, 11.00] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 6.1 treatment efficacy Show forest plot | 1 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 6.1.1 Short‐term (up to 6 months) | 1 | 26 | Risk Ratio (IV, Random, 95% CI) | 2.14 [0.90, 5.09] |
| 6.2 treatment acceptability (dropouts) Show forest plot | 8 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 6.2.1 Short‐term (up to 6 months) | 8 | 359 | Risk Ratio (IV, Random, 95% CI) | 1.17 [0.70, 1.93] |
| 6.3 depression symptoms Show forest plot | 12 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 6.3.1 Short‐term (up to 6 months) | 12 | 619 | Std. Mean Difference (IV, Random, 95% CI) | ‐1.04 [‐1.44, ‐0.63] |
| 6.4 quality of life Show forest plot | 1 | 80 | Mean Difference (IV, Random, 95% CI) | 0.03 [‐0.70, 0.76] |
| 6.4.1 Short‐term (up to 6 months) | 1 | 80 | Mean Difference (IV, Random, 95% CI) | 0.03 [‐0.70, 0.76] |
| 6.5 anxiety symptoms Show forest plot | 5 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 6.5.1 Short‐term (up to 6 months) | 5 | 424 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.91 [‐1.59, ‐0.23] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 7.1 treatment acceptability (dropouts) Show forest plot | 1 | 96 | Risk Ratio (M‐H, Random, 95% CI) | 0.72 [0.31, 1.67] |
| 7.1.1 Short‐term (up to 6 months) | 1 | 96 | Risk Ratio (M‐H, Random, 95% CI) | 0.72 [0.31, 1.67] |
| 7.2 depression symptoms Show forest plot | 2 | 108 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.18 [‐0.57, 0.20] |
| 7.2.1 Short‐term (up to 6 months) | 2 | 108 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.18 [‐0.57, 0.20] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 8.1 treatment efficacy Show forest plot | 1 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 8.1.1 Short‐term (up to 6 months) | 1 | 141 | Risk Ratio (IV, Random, 95% CI) | 1.77 [1.14, 2.76] |
| 8.2 treatment acceptability (dropouts) Show forest plot | 2 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 8.2.1 Short‐term (up to 6 months) | 2 | 243 | Risk Ratio (IV, Random, 95% CI) | 0.52 [0.23, 1.16] |
| 8.2.2 Medium‐term (7‐12 months) | 1 | 100 | Risk Ratio (IV, Random, 95% CI) | 0.86 [0.31, 2.37] |
| 8.3 depression symptoms Show forest plot | 2 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 8.3.1 Short‐term change from baseline (up to 6 months) | 2 | 180 | Mean Difference (IV, Random, 95% CI) | ‐1.42 [‐4.80, 1.96] |
| 8.3.2 Medium‐term change from baseline (7‐12 months) | 1 | 100 | Mean Difference (IV, Random, 95% CI) | ‐2.34 [‐3.84, ‐0.84] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 9.1 treatment acceptability (dropouts) Show forest plot | 3 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 9.1.1 Short‐term (up to 6 months) | 3 | 187 | Risk Ratio (IV, Random, 95% CI) | 0.97 [0.45, 2.09] |
| 9.1.2 Medium‐term (7‐12 months) | 1 | 124 | Risk Ratio (IV, Random, 95% CI) | 1.57 [0.65, 3.79] |
| 9.2 depression symptoms Show forest plot | 3 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 9.2.1 Short‐term (up to 6 months) | 3 | 187 | Mean Difference (IV, Random, 95% CI) | ‐6.10 [‐7.87, ‐4.33] |
| 9.2.2 Medium‐term (7‐12 months) | 1 | 118 | Mean Difference (IV, Random, 95% CI) | ‐2.83 [‐5.32, ‐0.34] |
| 9.3 quality of life Show forest plot | 1 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 9.3.1 Short‐term (up to 6 months) | 1 | 118 | Mean Difference (IV, Random, 95% CI) | 0.07 [0.03, 0.11] |
| 9.4 anxiety symptoms Show forest plot | 1 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 9.4.1 Short‐term (up to 6 months) | 1 | 30 | Mean Difference (IV, Random, 95% CI) | ‐5.50 [‐10.01, ‐0.99] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 10.1 treatment efficacy Show forest plot | 7 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 10.1.1 Short‐term (up to 6 months) | 7 | 1533 | Risk Ratio (M‐H, Random, 95% CI) | 1.40 [1.10, 1.78] |
| 10.1.2 Medium‐term (7‐12 months) | 2 | 1012 | Risk Ratio (M‐H, Random, 95% CI) | 1.23 [1.07, 1.42] |
| 10.2 treatment acceptability (dropouts) Show forest plot | 14 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 10.2.1 Short‐term (up to 6 months) | 14 | 2518 | Risk Ratio (IV, Random, 95% CI) | 1.64 [0.81, 3.31] |
| 10.2.2 Medium‐term (7‐12 months) | 4 | 1726 | Risk Ratio (IV, Random, 95% CI) | 2.84 [0.92, 8.75] |
| 10.2.3 Long‐term (>12 months) | 1 | 485 | Risk Ratio (IV, Random, 95% CI) | 2.17 [1.39, 3.39] |
| 10.3 depression symptoms Show forest plot | 15 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 10.3.1 Short‐term (up to 6 months) | 15 | 2208 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.78 [‐1.05, ‐0.51] |
| 10.3.2 Medium‐term (7‐12 months) | 4 | 1381 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.23 [‐0.38, ‐0.08] |
| 10.3.3 Long‐term (>12 months) | 1 | 343 | Std. Mean Difference (IV, Random, 95% CI) | 0.02 [‐0.19, 0.23] |
| 10.4 quality of life Show forest plot | 6 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 10.4.1 short‐term (up to 6 months) | 6 | 1299 | Std. Mean Difference (IV, Random, 95% CI) | 0.97 [0.38, 1.57] |
| 10.4.2 medium‐term (7‐12 months) | 2 | 879 | Std. Mean Difference (IV, Random, 95% CI) | 0.14 [‐0.12, 0.40] |
| 10.4.3 long‐term (>12 months) | 1 | 325 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.09 [‐0.30, 0.13] |
| 10.5 social adjustment and functioning Show forest plot | 2 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 10.5.1 Short‐term (up to 6 months) | 2 | 88 | Std. Mean Difference (IV, Random, 95% CI) | ‐1.27 [‐1.74, ‐0.81] |
| 10.6 anxiety symptoms Show forest plot | 4 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 10.6.1 Short‐term (up to 6 months) | 4 | 1063 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.33 [‐0.45, ‐0.21] |
| 10.6.2 Medium‐term (7‐12 months) | 2 | 851 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.27 [‐0.41, ‐0.12] |
| 10.6.3 Long‐term (>12 months) | 1 | 332 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.08 [‐0.29, 0.14] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 11.1 treatment efficacy Show forest plot | 6 | 903 | Risk Ratio (IV, Random, 95% CI) | 1.87 [1.18, 2.95] |
| 11.1.1 under 65 | 4 | 244 | Risk Ratio (IV, Random, 95% CI) | 2.03 [1.49, 2.75] |
| 11.1.2 65 and over | 2 | 659 | Risk Ratio (IV, Random, 95% CI) | 3.32 [0.20, 54.59] |
| 11.2 treatment acceptability (dropouts) Show forest plot | 15 | 1550 | Risk Ratio (IV, Random, 95% CI) | 1.20 [0.54, 2.67] |
| 11.2.1 under 65 | 9 | 566 | Risk Ratio (IV, Random, 95% CI) | 0.83 [0.49, 1.40] |
| 11.2.2 65 and over | 6 | 984 | Risk Ratio (IV, Random, 95% CI) | 1.30 [0.26, 6.38] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 12.1 treatment efficacy Show forest plot | 8 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 12.1.1 specialist | 3 | 186 | Risk Ratio (IV, Random, 95% CI) | 1.11 [0.93, 1.32] |
| 12.1.2 specialist in training | 2 | 130 | Risk Ratio (IV, Random, 95% CI) | 1.13 [0.85, 1.49] |
| 12.1.3 non‐specialist | 3 | 672 | Risk Ratio (IV, Random, 95% CI) | 1.30 [0.86, 1.98] |
| 12.2 treatment acceptability (dropouts) Show forest plot | 17 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 12.2.1 specialist | 11 | 593 | Risk Ratio (IV, Random, 95% CI) | 0.88 [0.62, 1.25] |
| 12.2.2 specialist in training | 3 | 90 | Risk Ratio (IV, Random, 95% CI) | 0.83 [0.31, 2.25] |
| 12.2.3 non‐specialist | 3 | 701 | Risk Ratio (IV, Random, 95% CI) | 1.05 [0.84, 1.31] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 13.1 treatment efficacy Show forest plot | 10 | 1730 | Risk Ratio (IV, Random, 95% CI) | 1.51 [1.24, 1.85] |
| 13.1.1 specialist | 4 | 238 | Risk Ratio (IV, Random, 95% CI) | 1.71 [1.08, 2.70] |
| 13.1.2 non‐specialist | 6 | 1492 | Risk Ratio (IV, Random, 95% CI) | 1.49 [1.13, 1.97] |
| 13.2 treatment acceptability (dropouts) Show forest plot | 26 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 13.2.1 specialist | 12 | 618 | Risk Ratio (IV, Random, 95% CI) | 0.65 [0.47, 0.89] |
| 13.2.2 specialist in training | 3 | 98 | Risk Ratio (IV, Random, 95% CI) | 1.35 [0.42, 4.35] |
| 13.2.3 non‐specialist | 11 | 2544 | Risk Ratio (IV, Random, 95% CI) | 2.20 [1.06, 4.57] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 14.1 treatment efficacy Show forest plot | 7 | 1627 | Risk Ratio (IV, Random, 95% CI) | 1.38 [1.13, 1.70] |
| 14.1.1 subthreshold depression | 2 | 627 | Risk Ratio (IV, Random, 95% CI) | 1.09 [1.01, 1.17] |
| 14.1.2 moderate/ severe depression | 5 | 1000 | Risk Ratio (IV, Random, 95% CI) | 1.62 [1.41, 1.85] |
| 14.2 treatment acceptability (dropouts) Show forest plot | 15 | 2278 | Risk Ratio (IV, Random, 95% CI) | 1.45 [0.65, 3.25] |
| 14.2.1 subthreshold depression | 3 | 864 | Risk Ratio (IV, Random, 95% CI) | 4.30 [0.46, 40.44] |
| 14.2.2 moderate/ severe depression | 12 | 1414 | Risk Ratio (IV, Random, 95% CI) | 1.04 [0.55, 1.97] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 15.1 treatment acceptability (dropouts) Show forest plot | 25 | 2947 | Risk Ratio (IV, Random, 95% CI) | 1.31 [0.79, 2.17] |
| 15.1.1 1‐3 sessions | 3 | 117 | Risk Ratio (IV, Random, 95% CI) | 1.03 [0.53, 2.03] |
| 15.1.2 >3 sessions | 22 | 2830 | Risk Ratio (IV, Random, 95% CI) | 1.35 [0.76, 2.37] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 16.1 treatment efficacy Show forest plot | 9 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 16.1.1 CBT comparator | 5 | 591 | Risk Ratio (IV, Random, 95% CI) | 0.99 [0.91, 1.07] |
| 16.1.2 Third‐wave CBT comparator | 3 | 118 | Risk Ratio (IV, Random, 95% CI) | 1.08 [0.91, 1.29] |
| 16.1.3 Psychodynamic/ humanist/ integrative comparator | 4 | 371 | Risk Ratio (IV, Random, 95% CI) | 1.50 [1.24, 1.81] |
| 16.2 treatment acceptability (dropouts) Show forest plot | 20 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 16.2.1 CBT comparator | 8 | 1017 | Risk Ratio (IV, Random, 95% CI) | 1.04 [0.85, 1.28] |
| 16.2.2 Third‐wave CBT comparator | 9 | 393 | Risk Ratio (IV, Random, 95% CI) | 0.86 [0.54, 1.36] |
| 16.2.3 Psychodynamic/ humanist/ integrative comparator | 7 | 249 | Risk Ratio (IV, Random, 95% CI) | 0.77 [0.44, 1.33] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 17.1 treatment efficacy Show forest plot | 10 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 17.1.1 treatment as usual | 4 | 747 | Risk Ratio (IV, Random, 95% CI) | 1.17 [0.95, 1.45] |
| 17.1.2 waiting list | 1 | 26 | Risk Ratio (IV, Random, 95% CI) | 2.14 [0.90, 5.09] |
| 17.1.3 no treatment | 1 | 30 | Risk Ratio (IV, Random, 95% CI) | 2.97 [1.42, 6.24] |
| 17.1.4 medication | 1 | 141 | Risk Ratio (IV, Random, 95% CI) | 1.77 [1.14, 2.76] |
| 17.1.5 other comparator (enhanced usual care, mental health referral, psychoeducation) | 3 | 786 | Risk Ratio (IV, Random, 95% CI) | 1.59 [1.38, 1.83] |
| 17.2 treatment acceptability (dropouts) Show forest plot | 26 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 17.2.1 treatment as usual | 10 | 1632 | Risk Ratio (IV, Random, 95% CI) | 1.50 [0.56, 3.99] |
| 17.2.2 waiting list | 8 | 359 | Risk Ratio (IV, Random, 95% CI) | 1.17 [0.70, 1.93] |
| 17.2.3 no treatment | 3 | 187 | Risk Ratio (IV, Random, 95% CI) | 0.97 [0.45, 2.09] |
| 17.2.4 medication placebo | 1 | 96 | Risk Ratio (IV, Random, 95% CI) | 0.72 [0.31, 1.67] |
| 17.2.5 medication | 1 | 143 | Risk Ratio (IV, Random, 95% CI) | 0.37 [0.18, 0.75] |
| 17.2.6 other comparator (enhanced usual care, mental health referral, psychoeducation) | 4 | 886 | Risk Ratio (IV, Random, 95% CI) | 2.17 [1.04, 4.53] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 18.1 treatment efficacy Show forest plot | 5 | 826 | Risk Ratio (IV, Random, 95% CI) | 1.20 [0.95, 1.51] |
| 18.2 treatment acceptability (dropouts) Show forest plot | 7 | 1039 | Risk Ratio (IV, Random, 95% CI) | 1.04 [0.84, 1.29] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 19.1 treatment efficacy Show forest plot | 6 | 1560 | Risk Ratio (IV, Random, 95% CI) | 1.49 [1.16, 1.90] |
| 19.2 treatment acceptability (dropouts) Show forest plot | 12 | 2753 | Risk Ratio (IV, Random, 95% CI) | 2.22 [1.00, 4.95] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 20.1 treatment efficacy Show forest plot | 6 | 657 | Risk Ratio (IV, Random, 95% CI) | 1.09 [0.92, 1.29] |
| 20.2 treatment acceptability (dropouts) Show forest plot | 16 | 1351 | Risk Ratio (IV, Random, 95% CI) | 1.00 [0.83, 1.20] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 21.1 treatment efficacy Show forest plot | 7 | 838 | Risk Ratio (IV, Random, 95% CI) | 1.76 [1.50, 2.05] |
| 21.2 treatment acceptability (dropouts) Show forest plot | 20 | 1603 | Risk Ratio (IV, Random, 95% CI) | 0.85 [0.67, 1.08] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 22.1 treatment efficacy Show forest plot | 8 | 988 | Risk Ratio (IV, Random, 95% CI) | 1.17 [1.00, 1.37] |
| 22.2 treatment acceptability (dropouts) Show forest plot | 14 | 1251 | Risk Ratio (IV, Random, 95% CI) | 1.05 [0.91, 1.22] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 23.1 treatment efficacy Show forest plot | 8 | 1616 | Risk Ratio (IV, Random, 95% CI) | 1.61 [1.26, 2.05] |
| 23.2 treatment acceptability (dropouts) Show forest plot | 21 | 2811 | Risk Ratio (IV, Random, 95% CI) | 1.55 [0.85, 2.79] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 24.1 depression symptoms Show forest plot | 12 | Std. Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 24.1.1 Short‐term (up to 6 months) | 12 | 619 | Std. Mean Difference (IV, Fixed, 95% CI) | ‐0.72 [‐0.89, ‐0.55] |
| 24.2 anxiety symptoms Show forest plot | 5 | Std. Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 24.2.1 Short‐term (up to 6 months) | 5 | 424 | Std. Mean Difference (IV, Fixed, 95% CI) | ‐0.54 [‐0.74, ‐0.33] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 25.1 depression symptoms Show forest plot | 14 | Std. Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 25.1.1 Short‐term (up to 6 months) | 14 | 2158 | Std. Mean Difference (IV, Fixed, 95% CI) | ‐0.48 [‐0.57, ‐0.39] |
| 25.1.2 Medium‐term (7‐12 months) | 4 | 1381 | Std. Mean Difference (IV, Fixed, 95% CI) | ‐0.23 [‐0.34, ‐0.13] |
| 25.1.3 Long‐term (>12 months) | 1 | 343 | Std. Mean Difference (IV, Fixed, 95% CI) | 0.02 [‐0.19, 0.23] |
| 25.2 quality of life Show forest plot | 5 | Std. Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 25.2.1 short‐term (up to 6 months) | 5 | 1249 | Std. Mean Difference (IV, Fixed, 95% CI) | 0.25 [0.14, 0.37] |
| 25.2.2 medium‐term (7‐12 months) | 2 | 879 | Std. Mean Difference (IV, Fixed, 95% CI) | 0.16 [0.03, 0.29] |
| 25.2.3 long‐term (>12 months) | 1 | 325 | Std. Mean Difference (IV, Fixed, 95% CI) | ‐0.09 [‐0.30, 0.13] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 26.1 treatment efficacy Show forest plot | 13 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 26.1.1 CBT | 4 | 573 | Risk Ratio (IV, Random, 95% CI) | 0.93 [0.83, 1.05] |
| 26.1.2 third‐wave CBT | 2 | 117 | Risk Ratio (IV, Random, 95% CI) | 1.17 [0.91, 1.52] |
| 26.1.3 humanistic | 1 | 46 | Risk Ratio (IV, Random, 95% CI) | 2.33 [1.09, 5.00] |
| 26.1.4 treatment as usual | 7 | 1743 | Risk Ratio (IV, Random, 95% CI) | 1.29 [0.99, 1.68] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 27.1 treatment efficacy Show forest plot | 13 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 27.1.1 CBT | 4 | 573 | Risk Ratio (IV, Random, 95% CI) | 1.17 [0.90, 1.52] |
| 27.1.2 third‐wave CBT | 2 | 117 | Risk Ratio (IV, Random, 95% CI) | 1.41 [1.12, 1.76] |
| 27.1.3 humanistic | 1 | 46 | Risk Ratio (IV, Random, 95% CI) | 3.67 [1.83, 7.34] |
| 27.1.4 treatment as usual | 7 | 1743 | Risk Ratio (IV, Random, 95% CI) | 1.63 [1.29, 2.04] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 28.1 treatment efficacy Show forest plot | 13 | Risk Ratio (IV, Random, 95% CI) | Subtotals only | |
| 28.1.1 CBT | 4 | 573 | Risk Ratio (IV, Random, 95% CI) | 0.82 [0.58, 1.17] |
| 28.1.2 third‐wave CBT | 2 | 117 | Risk Ratio (IV, Random, 95% CI) | 0.89 [0.73, 1.09] |
| 28.1.3 humanistic | 1 | 46 | Risk Ratio (IV, Random, 95% CI) | 0.78 [0.53, 1.15] |
| 28.1.4 treatment as usual | 7 | 1743 | Risk Ratio (IV, Random, 95% CI) | 1.14 [0.89, 1.46] |
