Primary‐level and community worker interventions for the prevention of mental disorders and the promotion of well‐being in low‐ and middle‐income countries
Abstract
Background
There is a significant research gap in the field of universal, selective, and indicated prevention interventions for mental health promotion and the prevention of mental disorders. Barriers to closing the research gap include scarcity of skilled human resources, large inequities in resource distribution and utilization, and stigma.
Objectives
To assess the effectiveness of delivery by primary workers of interventions for the promotion of mental health and universal prevention, and for the selective and indicated prevention of mental disorders or symptoms of mental illness in low‐ and middle‐income countries (LMICs). To examine the impact of intervention delivery by primary workers on resource use and costs.
Search methods
We searched CENTRAL, MEDLINE, Embase, CINAHL, Global Index Medicus, PsycInfo, WHO ICTRP, and ClinicalTrials.gov from inception to 29 November 2021.
Selection criteria
Randomized controlled trials (RCTs) of primary‐level and/or community health worker interventions for promoting mental health and/or preventing mental disorders versus any control conditions in adults and children in LMICs.
Data collection and analysis
Standardized mean differences (SMD) or mean differences (MD) were used for continuous outcomes, and risk ratios (RR) for dichotomous data, using a random‐effects model. We analyzed data at 0 to 1, 1 to 6, and 7 to 24 months post‐intervention. For SMDs, 0.20 to 0.49 represented small, 0.50 to 0.79 moderate, and ≥ 0.80 large clinical effects. We evaluated the risk of bias (RoB) using Cochrane RoB2.
Main results
Description of studies
We identified 113 studies with 32,992 participants (97 RCTs, 19,570 participants in meta‐analyses) for inclusion. Nineteen RCTs were conducted in low‐income countries, 27 in low‐middle‐income countries, 2 in middle‐income countries, 58 in upper‐middle‐income countries and 7 in mixed settings. Eighty‐three RCTs included adults and 30 RCTs included children. Cadres of primary‐level workers employed primary care health workers (38 studies), community workers (71 studies), both (2 studies), and not reported (2 studies). Interventions were universal prevention/promotion in 22 studies, selective in 36, and indicated prevention in 55 RCTs.
Risk of bias
The most common concerns over risk of bias were performance bias, attrition bias, and reporting bias.
Intervention effects
'Probably', 'may', or 'uncertain' indicates 'moderate‐', 'low‐', or 'very low‐'certainty evidence.
*Certainty of the evidence (using GRADE) was assessed at 0 to 1 month post‐intervention as specified in the review protocol. In the abstract, we did not report results for outcomes for which evidence was missing or very uncertain.
Adults
Promotion/universal prevention, compared to usual care:
‐ probably slightly reduced anxiety symptoms (MD ‐0.14, 95% confidence interval (CI) ‐0.27 to ‐0.01; 1 trial, 158 participants)
‐ may slightly reduce distress/PTSD symptoms (SMD ‐0.24, 95% CI ‐0.41 to ‐0.08; 4 trials, 722 participants)
Selective prevention, compared to usual care:
‐ probably slightly reduced depressive symptoms (SMD ‐0.69, 95% CI ‐1.08 to ‐0.30; 4 trials, 223 participants)
Indicated prevention, compared to usual care:
‐ may reduce adverse events (1 trial, 547 participants)
‐ probably slightly reduced functional impairment (SMD ‐0.12, 95% CI ‐0.39 to ‐0.15; 4 trials, 663 participants)
Children
Promotion/universal prevention, compared to usual care:
‐ may improve the quality of life (SMD ‐0.25, 95% CI ‐0.39 to ‐0.11; 2 trials, 803 participants)
‐ may reduce adverse events (1 trial, 694 participants)
‐ may slightly reduce depressive symptoms (MD ‐3.04, 95% CI ‐6 to ‐0.08; 1 trial, 160 participants)
‐ may slightly reduce anxiety symptoms (MD ‐2.27, 95% CI ‐3.13 to ‐1.41; 1 trial, 183 participants)
Selective prevention, compared to usual care:
‐ probably slightly reduced depressive symptoms (SMD 0, 95% CI ‐0.16 to ‐0.15; 2 trials, 638 participants)
‐ may slightly reduce anxiety symptoms (MD 4.50, 95% CI ‐12.05 to 21.05; 1 trial, 28 participants)
‐ probably slightly reduced distress/PTSD symptoms (MD ‐2.14, 95% CI ‐3.77 to ‐0.51; 1 trial, 159 participants)
Indicated prevention, compared to usual care:
‐ decreased slightly functional impairment (SMD ‐0.29, 95% CI ‐0.47 to ‐0.10; 2 trials, 448 participants)
‐ decreased slightly depressive symptoms (SMD ‐0.18, 95% CI ‐0.32 to ‐0.04; 4 trials, 771 participants)
‐ may slightly reduce distress/PTSD symptoms (SMD 0.24, 95% CI ‐1.28 to 1.76; 2 trials, 448 participants).
Authors' conclusions
The evidence indicated that prevention interventions delivered through primary workers ‐ a form of task‐shifting ‐ may improve mental health outcomes. Certainty in the evidence was influenced by the risk of bias and by substantial levels of heterogeneity. A supportive network of infrastructure and research would enhance and reinforce this delivery modality across LMICs.
PICOs
Plain language summary
Primary‐level and community worker interventions for the prevention of mental disorders and the promotion of well‐being in low‐ and middle‐income countries
What is the main aim of this review?
The aim of this Cochrane Review was to assess the effects of involving people in primary services and the community, such as nurses, midwives, teachers or caregivers, to promote mental health. The review focused on children and adults living in low‐ and middle‐income countries.
Key messages
The employment use of primary‐level and community workers may improve the mental health of adults and children living in low‐ and middle‐income countries. However, more evidence is needed.
What was studied in this review?
Many people who would benefit from mental health support cannot access these services. One reason for this is a lack of specialized mental healthcare staff. This is especially true in low‐ and middle‐income countries. To overcome this barrier, people without a professional background in mental health, such as nurses or teachers, can be trained to deliver some mental health services. In our review, we investigated whether this strategy helps to promote mental health and prevent mental disorders amongst adults and children. We also assessed its costs.
What are the main results of this review?
We included 113 studies from a range of low‐ and middle‐income countries.
The studies assessed the effects of services carried out by primary‐level and community workers on people's mental health, quality of life, and social outcomes.
We grouped interventions depending on their overall objectives. Specifically, we refer to those targeting the whole population as 'promotion/universal prevention', those targeting people at risk for developing a mental disorder as 'selective prevention', and those designed for already presenting some sign of mental disorders as 'indicated prevention'. Below we report evidence of the results of low to moderate‐certainty, directly after the intervention. We did not present results for outcomes for which there was no or very uncertain evidence.
Promotion/universal prevention interventions, compared to usual care:
‐ probably slightly reduced anxiety symptoms in adults
‐ may slightly reduce distress/PTSD symptoms in adults
‐ may improve the quality of life of children
‐ may reduce adverse events in children
‐ may slightly reduce depression symptoms in children
‐ may slightly reduce anxiety symptoms in children
Selective prevention interventions, compared to usual care:
‐ probably slightly reduced depressive symptoms in adults
‐ may slightly reduce functional impairment in children
‐ probably slightly reduced depressive symptoms in children
‐ may slightly reduce anxiety symptoms in children
‐ probably slightly reduced distress/PTSD symptoms in children
Indicated prevention interventions,compared to usual care:
‐ may reduce adverse events in adults
‐ probably slightly reduced functional impairment in adults
‐ decreased slightly functional impairment in children
‐ decreased slightly depressive symptoms in children
‐ may slightly reduce distress/PTSD symptoms in children
Indicated prevention interventions delivered through task‐shifting may improve mental health outcomes.
What are the limitations of the evidence?
The limitations of the evidence in this review stem from the absence of assessments related to the reduction in the incidence of mental disorders in the prevention studies, and the lack of discernible differences in acceptability. Furthermore, the limited number of randomized controlled trials reporting our secondary outcomes, and their low quality, failed to demonstrate clinically significant advantageous effects of the studied prevention interventions for some outcomes in both child and adult populations.
How up‐to‐date is the review?
Review authors searched databases up to November 2021 to find and include all relevant published and unpublished trials.
Authors' conclusions
Summary of findings
| Promotion/universal prevention interventions compared to control group in preventing mental disorders in adults | ||||||
| Patient or population: preventing mental disorders | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with control group | Risk with promotion/universal prevention interventions | |||||
| Diagnosis of mental disorders at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Quality of life at study endpoint (higher score = better quality of life) | ‐ | SMD 0.23 SD lower | ‐ | 684 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.23 (95% CI ‐0.51 to 0.04). It is uncertain whether promotion/universal prevention interventions have any effect on quality of life among adults without risk factors for mental disorders (at post‐intervention) compared with usual care [there is a small effect according to Cohen 1992].1 |
| Adverse events at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Psychological functioning and impairment at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Depressive symptoms at study endpoint (higher score = higher severity ) | ‐ | SMD 0.31 SD lower | ‐ | 349 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.31 (95% CI ‐0.78 to 0.15). It is uncertain whether promotion/universal prevention interventions have any effect on depressive symptoms in adults without risk factors for mental disorders (at post‐intervention) compared to usual care [this is a small effect according to Cohen 1992].1 |
| Anxiety symptoms at study endpoint (higher score = higher severity) | The mean anxiety symptoms at study endpoint was 0 | MD 0.14 lower | ‐ | 158 | ⊕⊕⊕⊝ | Promotion/universal prevention interventions for adults without risk factors for mental disorders probably slightly reduce anxiety symptoms (at post‐intervention) compared to usual care [there is a small effect according to Cohen 1992].1 |
| Distress/PTSD symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.24 SD lower | ‐ | 722 | ⊕⊕⊝⊝ | Scores estimated based on an SMD of ‐0.24 (95% CI ‐0.41 to ‐0.08). Promotion/universal prevention interventions may slightly reduce distress/PTSD symptoms in adults without risk factors for mental disorders (at post‐intervention) comparedto usual care [there is a small effect according to Cohen 1992].1 |
| *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 | ||||||
| See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_429913845297007331. | ||||||
| a Downgraded 2 levels owing to study limitations (over 30% of RCTs have overall high risk of bias) | ||||||
| Selective prevention interventions compared to control group in preventing mental disorders in adults | ||||||
| Patient or population: preventing mental disorders | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with control group | Risk with selective prevention interventions | |||||
| Diagnosis of mental disorders at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Quality of life at study endpoint (higher score = better quality of life) | ‐ | SMD 1.64 lower | ‐ | 229 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐1.64 (95% CI ‐2.97 to ‐0.31). It is uncertain whether selective prevention interventions have any effect on quality of life among adults with risk factors for mental disorders/lack of protective factors (at post‐intervention) compared with usual care. [There is a large effect according to Cohen 1992]1 |
| Adverse events at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Psychological functioning and impairment at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Depressive symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.69 lower | ‐ | 223 | ⊕⊕⊕⊝ | Scores estimated based on an SMD of ‐0.69 (95% CI ‐1.08 to ‐0.3). Selective prevention interventions for adults with risk factors for mental disorders/lack of protective factors probably slightly reduce depressive symptoms (at post‐intervention)compared to usual care. [There is a medium effect according to Cohen 1992]1 |
| Anxiety symptoms at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Distress/PTSD symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.9 lower | ‐ | 535 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.90 (95% CI ‐1.44 to ‐0.36). It is uncertain whether selective prevention interventions have any effect on distress/PTSD symptoms in adults with risk factors for mental disorders/lack of protective factors (at post‐intervention) compared to usual care. [There is a large effect according to Cohen 1992]1 |
| *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 | ||||||
| See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_429913907887557578. | ||||||
| a Downgraded 2 level owing to inconsistency (I2 was higher than 75%, P < 0.00001) | ||||||
| Indicated prevention interventions compared to control group in preventing mental disorders in adults | ||||||
| Patient or population: preventing mental disorders | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with control group | Risk with indicated prevention interventions | |||||
| Diagnosis of mental disorders at study endpoint (RR < 1 denotes lower risk of mental diagnosis) | 170 per 1000 | 51 per 1000 | RR 0.30 | 843 | ⊕⊝⊝⊝ | It is uncertain whether indicated prevention interventions have any effect on the risk of mental disorders in adults with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. |
| Quality of life at study endpoint (higher score = better quality of life) | ‐ | SMD 0.36 lower | ‐ | 1136 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.36 (95% CI ‐0.61 to ‐0.12). It is uncertain whether indicated prevention interventions have any effect on quality of life among adults with a high vulnerability to develop mental disorders (at post‐intervention) compared with usual care. [There is a small effect according to Cohen 1992]1 |
| Adverse events at study endpoint | Not pooled | Not pooled | Not pooled | (1 RCT) | ⊕⊕⊝⊝ | Indicated prevention interventions may reduce adverse events in adults with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. |
| Psychological functioning and impairment at study endpoint (higher score = higher disability) | ‐ | SMD 0.12 lower | ‐ | 663 | ⊕⊕⊕⊝ | Scores estimated based on an SMD of ‐0.12 (95% CI ‐0.39 to 0.15). Indicated prevention interventions for adults with a high vulnerability to develop mental disorders probably slightly reduce functional impairment (at post‐intervention)compared to usual care. [There is a small effect according to Cohen 1992]1 |
| Depressive symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.16 lower | ‐ | 2341 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.16 (95% CI ‐0.3 to ‐0.03). It is uncertain whether indicated prevention interventions have any effect on depressive symptoms in adults with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. [There is a small effect according to Cohen 1992]1 |
| Anxiety symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 1.19 lower | ‐ | 250 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐1.19 (95% CI ‐2.02 to ‐0.035). It is uncertain whether indicated prevention interventions have any effect on depressive symptoms in adults with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. [There is a large effect according to Cohen 1992]1 |
| Distress/PTSD symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.54 lower | ‐ | 2536 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.54 (95% CI ‐0.95 to ‐0.14). It is uncertain whether indicated prevention interventions have any effect on distress/PTSD symptoms in adults with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. [There is a medium effect according to Cohen 1992]1 |
| *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 | ||||||
| See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_429913973563280333. | ||||||
| a Downgraded 2 levels owing to study limitations (over 30% of RCTs had high risk of bias due to missing outcome data and in selection of the reported result) | ||||||
| Promotion/universal prevention interventions compared to control group in preventing mental disorders in children | ||||||
| Patient or population: preventing mental disorders | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with control group | Risk with promotion/universal prevention interventions | |||||
| Diagnosis of mental disorders at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Quality of life at study endpoint (higher score = better quality of life) | ‐ | SMD 0.25 SD lower | ‐ | 803 | ⊕⊕⊝⊝ | Scores estimated based on an SMD of ‐0.25 (95% CI ‐0.39 to ‐0.11). Promotion/universal prevention interventions may improve the quality of life of children without risk factors for mental disorders (at post‐intervention) compared to usual care. [There is a small effect according to Cohen 1992]1 |
| Adverse events at study endpoint (RR < 1 indicates lower risk of adverse events) | Not pooled | Not pooled | Not pooled | (1 RCT) | ⊕⊕⊝⊝ | Promotion/universal prevention interventions may reduce adverse events in children without risk factors for mental disorders (at post‐intervention) compared to usual care |
| Psychological functioning and impairment at study endpoint (higher score = higher disability) | ‐ | SMD 0.04 higher | ‐ | 212 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of 0.04 (95% CI ‐0.9 to 0.98). It is uncertain whether promotion/universal prevention interventions have any effect on functional impairment in children without risk factors for mental disorders (at post‐intervention) compared to usual care. [There is a small effect according to Cohen 1992]1 |
| Depressive symptoms at study endpoint (higher score = higher severity) | MD 3.04 SD lower | ‐ | 160 | ⊕⊕⊝⊝ | Promotion/universal prevention interventions may slightly reduce depression symptoms in children without risk factors for mental disorders (at post‐intervention) compared to usual care. [There is a large effect according to Cohen 1992]1 | |
| Anxiety symptoms at study endpoint (higher score = higher severity) | MD 2.77 higher | ‐ | 183 | ⊕⊕⊝⊝ | Promotion/universal prevention interventions may slightly reduce anxiety symptoms in children without risk factors for mental disorders (at post‐intervention) compared to usual care. [There is a medium effect according to Cohen 1992]1 | |
| Distress/PTSD symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.83 SD lower | ‐ | 800 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.83 (95% CI ‐2.48 to 0.82). It is uncertain whether promotion/universal prevention interventions have any effect on distress/PTSD symptoms in children without risk factors for mental disorders (at post‐intervention) compared to usual care. [There is a large effect according to Cohen 1992]1 |
| *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 | ||||||
| See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_429913294102284283. | ||||||
| a Downgraded 1 level owing to study limitations (all RCTs had some concerns for the deviations from the intended interventions and in measurement of the outcome) | ||||||
| Selective prevention interventions compared to control group in preventing mental disorders in children | ||||||
| Patient or population: preventing mental disorders | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with control group | Risk with selective prevention interventions | |||||
| Diagnosis of mental disorders at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Quality of life at study endpoint (higher score = better quality of life) | MD 1.1 higher | ‐ | 115 | ⊕⊝⊝⊝ | It is uncertain whether selective prevention interventions have any effect on quality of life among children with risk factors for mental disorders/lack of protective factors (at post‐intervention) compared with usual care. [There is a small effect according to Cohen 1992]1 | |
| Adverse events at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Psychological functioning and impairment at study endpoint (higher score = higher disability) | MD 0.02 higher | ‐ | 479 | ⊕⊕⊝⊝ | There is no evidence that selective prevention interventions improve functional impairment in children with risk factors for mental disorders/lack of protective factors (at post‐intervention) compared to usual care. 1 | |
| Depressive symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0 SD | ‐ | 638 | ⊕⊕⊕⊝ | Scores estimated based on an SMD of 0.0 (95% CI ‐0.16 to 0.15). Selective prevention interventions for children with risk factors for mental disorders/lack of protective factors probably slightly reduce depressive symptoms (at post‐intervention)compared to usual care. [There is a small effect according to Cohen 1992]1 |
| Anxiety symptoms at study endpoint (higher score = higher severity) | MD 4.5 higher | ‐ | 28 | ⊕⊕⊝⊝ | Selective prevention interventions may make little or no difference to anxiety symptoms in children with risk factors for mental disorders/lack of protective factors (at post‐intervention) compared to usual care. 1 | |
| Distress/PTSD symptoms at study endpoint (higher score = higher severity) | MD 2.14 lower | ‐ | 159 | ⊕⊕⊕⊝ | Selective prevention interventions for children with risk factors for mental disorders/ lack of protective factors probably slightly reduce distress/PTSD symptoms (at post‐intervention) compared to usual care. [There is a small effect according to Cohen 1992]1 | |
| *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 | ||||||
| See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_429913697396673091. | ||||||
| a Downgraded 2 levels owing to study limitations (over 30% of RCTs had high risk of bias due to deviations from intended interventions and missing outcome data, in measurement of the outcome, and in selection of the reported result) | ||||||
| Indicated prevention interventions compared to control group in preventing mental disorders in children | ||||||
| Patient or population: preventing mental disorders | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with control group | Risk with indicated prevention interventions | |||||
| Diagnosis of mental disorders at study endpoint (RR < 1 denotes lower risk of mental diagnosis) | 336 per 1000 | 259 per 1000 | RR 0.77 | 220 | ⊕⊝⊝⊝ | It is uncertain whether indicated prevention interventions have any effect on the risk of mental disorders in children with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. |
| Quality of life at study endpoint (higher score = better quality of life) | ‐ | SMD 0.65 SD lower | ‐ | 152 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.65 (95% CI ‐2.09 to 0.79). It is uncertain whether indicated prevention interventions have any effect on quality of life among children with a high vulnerability to develop mental disorders (at post‐intervention) compared with usual care. [There is a small effect according to Cohen 1992]1 |
| Adverse events at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Psychological functioning and impairment at study endpoint (higher score = higher disability) | ‐ | SMD 0.29 SD lower | ‐ | 448 | ⊕⊕⊕⊕ | Scores estimated based on an SMD of ‐0.29 (95% CI ‐0.47 to ‐0.1). Indicated prevention interventions decrease slightly functional impairment in children with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. [There is [a small effect according to Cohen 1992]1 |
| Depressive symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.18 SD lower | ‐ | 771 | ⊕⊕⊕⊕ | Scores estimated based on an SMD of ‐0.18 (95% CI ‐0.32 to ‐0.04). Indicated prevention interventions decrease slightly depressive symptoms in children with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. [There is a small effect according to Cohen 1992]1 |
| Anxiety symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.09 lower | ‐ | 888 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.09 (95% CI ‐0.22 to 0.04). It is uncertain whether indicated prevention interventions have any effect on anxiety symptoms in children with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. [There is a small effect according to Cohen 1992]1 |
| Distress/PTSD symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.24 SD higher | ‐ | 448 | ⊕⊕⊝⊝ | Scores estimated based on an SMD of 0.24 (95% CI ‐1.28 to 1.76). Indicated prevention interventions may slightlyreduce distress/PTSD symptoms in children with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. [There is a small effect according to Cohen 1992]1 |
| *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 | ||||||
| See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_429913780370754267. | ||||||
| a Downgraded 1 level owing to study limitations (RCT did not provided information about allocation concealment, and outcome assessment was not described as masked) | ||||||
Background
Description of the condition
Worldwide, the global burden of mental, neurological, and substance use disorders is high. The latest global burden of disease studies estimated that mental, behavioural, and neuropsychiatric disorders are amongst the top 30 causes of all years lived with disability; the highest contributors are anxiety and depressive disorders, drug use disorders, and alcohol use disorders (Kyu 2018; Murray 2020; Santomauro 2021). Mental health and behavioural disorders contribute 7.4% of the global burden of disease in the world ‐ more than, for example, tuberculosis (2%), HIV/AIDS (3.3%), or malaria (4.6%) (Whiteford 2013). The contribution of major depressive disorders alone to worldwide disability‐adjusted life‐years (DALYs) increased by 37% between 1990 and 2010 and is predicted to rise further (Murray 2012; Prince 2007). In addition, self‐inflicted injuries and alcohol‐related disorders are likely to increase in the ranking of disease burden due to the decline in communicable diseases and because of a predicted increase in war and violence. The disease burden due to Alzheimer’s disease is also increasing; this is linked to the demographic transition towards an ageing population (Vos 2012). Despite these figures, levels of public expenditure on mental health are still low (a global median of 2.1% of government health expenditure) and particularly meagre in lower resource settings. Data from official WHO reports highlight persistent gaps in the availability of mental health resources, significant variation between high‐ and low‐income countries and across regions, and the need for continued country‐level investment in policies, plans, health services, and monitoring systems for mental health care (WHO 2021a).
Political instability, war, and natural disasters disproportionally affect populations living in low‐ and middle‐income countries (LMICs) (Guha‐Sapir 2014). In addition, people living in these contexts, are more prone to experiencing extreme chronic poverty (less than 1.90 international dollars per day), which causes about 385 million of children under the age of 18 years to experience severe deprivation of basic human needs, including food, water, sanitation, health, and education. The World Health Organization (WHO) estimates that the prevalence of post‐traumatic stress disorder (PTSD), depression, and mixed anxiety disorders in conflict settings is at 22% (Charlson 2019), which is, conservatively, around five times higher than the general population. The COVID‐19 pandemic is an additional public health emergency that occurred against an already complex backdrop of psychological suffering. It further affected social determinants of mental health, also rasing concerns about the deprioritization of the psychological needs of people in LMICs. Risk factors for mental disorders and increased symptomatology can include community‐level risk factors (e.g. neighbourhood disadvantages, high levels of community violence, community‐level gender inequitable norms), family‐level risk factors (e.g. intimate partner violence, harsh parenting), or individual‐level risk factors (e.g. low self‐esteem, maladaptive coping strategies). Similarly, protective factors can operate at multiple levels of the social environment. Promotion interventions may target promotive factors (i.e. factors associated with an increased chance of achieving positive mental health states).
These illnesses also imply substantial economic costs. One recent report on the global economic burden of non‐communicable diseases (NCDs) suggests that, by the early 2030s, mental health conditions alone will account for the loss of an additional USD 16.1 trillion, with a dramatic impact on productivity and quality of life (Bloom 2011). Data on the macroeconomic costs for LMIC settings remain uncertain (Hu 2006). However, the economic and social costs for individuals and families are substantial. High direct costs are incurred in countries where health spending is met largely through private, as opposed to public, spending, and where health insurance and employer‐met health payments are insubstantial (Patel 2007b). High indirect costs are incurred as the result of informal caregiving and lost work opportunities, as well as untreated disorders and their associated disabilities (Chisholm 2000).
All over the world, the gap between individuals in need of mental health interventions and those who actually receive such care is very large (WHO 2018). Previous estimates suggest that more than 56% of people with depression (De Silva 2014; Kohn 2004; Lora 2012; Patel 2010), along with 78% of persons with alcohol use disorders (Kohn 2004), do not receive care. A study of 21 countries via the WHO Mental Health Surveys found that 52.6% of persons with depressive disorders in LMICs had received no treatment in the past 12 months, and only 20.5% of persons with depressive disorders had received minimally adequate treatment (Thronicroft 2016). Major barriers to closing the treatment gap include scarcity of skilled human resources, large inequities and inefficiencies in resource distribution and utilization, and the significant stigma associated with mental health conditions (Barber 2019). Recent studies report the existence of a range of cost‐effective interventions in mental health care available that can be implemented in LMICs (Barbui 2020; Purgato 2018a). The global mental health community, therefore, advocates for the scaling‐up of these evidence‐based services, that focus on task‐shifting as a mean to bridge the treatment gap (Patel 2018).
Description of the intervention
A recent Lancet Commission sought to align global mental health efforts with sustainable development goals, and emphasized the importance of efforts to prevent mental health disorders and promote mental health, in addition to scaling‐up treatments (Patel 2018). Prevention and promotion of mental health have previously been advocated as critical by the WHO (WHO 2002), and prevention is part of the WHO Mental Health Action Plan (WHO 2013). Prevention and promotion efforts are an important complementary focus, in addition to the treatment of mental disorders, given that (1) many mental disorders have risk factors in the social environment (e.g. gender‐based violence, poverty) that can be effectively addressed; (2) there are limitations to the extent that treatments alone can reduce the burden associated with mental disorders (Freeman 2016); and (3) cost‐effective prevention and promotion interventions are available (Knapp 2011).
In the present review, we followed the classification of promotion and prevention interventions described by the Institute of Medicine (IOM) report on preventing mental disorders (Institute of Medicine 1994; Institute of Medicine 2009; National Academies of Sciences 2019).
Promotion is an approach aimed at strengthening positive aspects of mental health and psychosocial well‐being; it includes, for example, intervention components that foster pro‐social behaviour, self‐esteem, positive coping with stress, and decision‐making capacity (Table 1) (WHO 2014). The definition of promotion has been recently updated to include a wider set of interventions provided at societal, community, individual, and family levels. These updates reflect important trends in research in the field of public mental health and reveal the enduring importance of a spectrum of key tools for fostering mental health (National Academies of Sciences 2019).
| Adults | Participants who were ≥ 18 years old. If some studies had an age range from, for example, 16 years upwards, and a majority of participants (≥ 80%) were over 18 years of age, we included these study participants as adults. |
| Children and adolescents | Children (from birth to 18 years) were considered as a separate group of participants, as they have: • different patterns of psychopathology/mental disorders; and • different help‐seeking behaviours that would, therefore, require different interventions, in different settings (e.g. schools), and a different approach to interventions (e.g. worker interventions such as teacher‐led interventions). |
| Promotion | Promotion is an approach aimed at strengthening positive aspects of mental health and psychosocial well‐being; it includes, for example, components to foster prosocial behaviour, self‐esteem, positive coping with stress, and decision‐making capacity (National Academies of Sciences 2019; WHO 2014). Prevention is an approach aimed at reducing the likelihood of future disorder within the general population or amongst people who are identified as being at risk for developing a full‐blown disorder (Eaton 2012; Tol 2015a). |
| Universal prevention | Universal prevention includes strategies that can be offered to the whole population, based on evidence that prevention strategies are likely to provide some benefit to all (i.e. reduce the probability of a disorder), which clearly outweighs the costs and risks of negative consequences. Examples of common universal prevention interventions include:
|
| Selective prevention | Selective prevention refers to strategies that are targeted to subpopulations identified as being at elevated risk for a disorder; it includes:
|
| Indicated prevention | Indicated prevention includes strategies that are targeted to individuals who are identified (or individually screened) as having increased vulnerability for a disorder based on some individual assessment. These interventions include:
|
| First‐level care, primary care, and community care | First‐level contact with formal health services consists of community‐based interventions or primary care interventions (or both), on their own or attached to hospital settings, provided they had no specialist input apart from supervision (modified from Wiley‐Exley 2007). This would include promotion or prevention programmes in outpatient clinics or primary care practices. This would not include programmes in hospitals unless these programmes were providing prevention interventions to outpatients. Community programmes involve detection of mental disorders in all age groups, often done outside the health facility, for example, through school, training, and other community settings. |
| Low‐ and middle‐income countries (LMICs) | Any country that has ever been an LMIC, as defined by the World Bank lists of LMICs |
| Primary care health workers (PHWs) | Health workers who are not specializing in mental disorders or have not received in‐depth professional specialist training in this clinical area. They work in primary care centres or in the community. These individuals include doctors, nurses, auxiliary nurses, lay health workers, and allied health personnel such as social workers and occupational therapists. This category does not include professional specialist health workers such as psychiatrists, psychiatric nurses, or mental health social workers. For inclusion, PHWs received some training in mental conditions (in the control group or in the intervention group), but this would not constitute a professional category. Study authors made a judgement of what constitutes ‘some training’. Examples of ‘some training’ may include an undergraduate module or a short course in mental health. |
| Community workers (CWs) | People involved as community‐level workers but who are not within the health sector, as many people, particularly adolescents and young adults, have limited contact with health workers. This category includes teachers/trainers/support workers from schools and colleges, along with other volunteers or workers within community‐based networks or nongovernmental organisations. These CWs have an important role, particularly in promotion of mental health and detection of mental disorders (Patel 2007b; Patel 2008). We excluded from this review studies that looked at informal care provided by family members or that extended care only to members of their own family (i.e. who were unavailable to other members of the community). As was previously highlighted in Lewin’s Cochrane Review, “these interventions are qualitatively different from other LHW [lay health worker] interventions included in this review given that parents or spouses have an established close relationship with those receiving care, which could affect the process and effects of the intervention” (Lewin 2010). |
| Primary‐level workers (PWs) | Broad term to encompass both CWs and PHWs |
CW: community worker
LMIC: low‐ and middle‐income country
PHW: primary‐level health worker
PW: primary‐level worker
Prevention is an approach aimed at reducing the likelihood of future mental disorders in the general population or amongst people who are identified as being at risk for developing a full‐blown mental disorder (Eaton 2012; Purgato 2020; Tol 2015a). Prevention is further subdivided on the basis of the targeted population.
-
Universal prevention includes strategies that can be offered to the whole population based on evidence that prevention strategies are likely to provide some benefit to all (i.e. reduce the probability of disorder), which clearly outweighs the costs and risks of negative consequences. Examples of common universal prevention interventions include:
-
community‐wide provision of information on the negative effects of specific behaviours;
-
protection against human rights violations in the whole population (e.g. community mobilization to reduce gender‐based violence); and
-
community‐wide efforts to improve livelihood as a key protective factor for mental health (e.g. working on lifting restrictions on movement and employment for everyone in a refugee camp).
-
-
Selective prevention refers to strategies that target subpopulations identified as being at elevated risk for a disorder because they have known risk factors or lack protective factors. Examples include:
-
support for children whose parents have a mental illness;
-
strengthening of community networks for vulnerable families by activating social networks and supportive communication; and
-
stress management training in communities affected by chronic poverty.
-
-
Indicated prevention includes strategies that are targeted to individuals who are identified (or individually screened) as having increased vulnerability for a disorder based on some individual assessment of symptoms experienced but not meeting criteria for a full‐blown mental disorder. These interventions include but are not limited to:
-
mentoring programmes aimed at teachers and caregivers of children with behavioural problems; and
-
prevention of postnatal depression amongst women with heightened levels of prenatal symptoms (Institute of Medicine 2009) or interventions to prevent intimate partner violence (Turner 2020).
-
These interventions may be delivered at an individual level or at a group level and include antenatal and postnatal classes, parenthood classes, group interventions for managing distress, and continuity of care (home visits, follow‐ups) (O'Connor 2019; US Preventive Services Task Force 2019).
-
-
Primary healthcare workers (PHWs) are first‐level providers who have received general health rather than specialist training in mental health and can be based in a primary care clinic or in the community. It has been suggested that PWs (primary‐level workers) could deliver general and mental health interventions that are at least as effective and acceptable as those delivered by specialist health workers (Chatterjee 2003). In addition, PW interventions often have lower up‐front costs compared with those provided by professional specialist health workers. However, it is possible that these savings may be cancelled out by higher downstream resource use (Chisholm 2000). To address this concern, we aimed to include data on the costs and cost‐effectiveness of PW interventions. PHWs include professionals (mainly nurses, midwives, and other general health professionals) and para‐professionals (such as trained lay health providers, e.g. traditional birth attendants). PHWs do not include those with specialist mental health training, for example, psychiatrists, psychologists, psychiatric nurses, or mental health social workers with formal training. Community workers (CWs) are individuals who work in the community but not within the health sector. These might include teachers, trainers, support workers at schools and colleges, and other volunteers or workers within community‐based networks or non‐governmental organizations (NGOs). CWs are an additional human resource that can be widely deployed in delivering promotion and prevention interventions (Patel 2007a). In this review, both these categories of providers (PHWs and CWs) are referred to under the umbrella heading of primary‐level workers (PWs).
PWs have been deployed for a variety of services, including those delivered in governmental organizations, private clinics, halfway homes, schools, and other community settings. For example, PWs have been involved in supporting and befriending carers and in ensuring intervention adherence (Tol 2020). Nurses, social workers, and CWs may also take on follow‐up or educational/promotive roles (Araya 2003; Chatterjee 2003; Patel 2008). In addition, doctors with general mental health training have been involved in the identification, diagnosis, treatment, and referral of complex cases (Patel 2008).
A summary of the main definitions adopted in this review is provided in Table 1.
How the intervention might work
Prevention interventions commonly target known modifiable risk and protective factors for mental disorders.
Although a vast array of interventions can be implemented to meet a population's psychosocial needs, there are some common elements, especially when interventions target smaller groups or families. Many interventions include techniques from cognitive‐behavioural therapy (CBT) which work at stopping negative cycles, breaking down feelings and problems that generate psychological suffering, anxiety or depression into separate parts. This mechanism facilitates problem management and changes negative thought patterns, improving the way people feel (Hofmann 2017). CBT may comprise, for example, facilitated discussion, strengthening of social networks, space provided for sharing personal experiences and exchange of peer support, and opportunities to practice coping skills to manage adversity. Sessions on problem‐solving skills and emotional support Interventions may work by challenging participants to replace negative or critical self‐talk with compassionate, more constructive self‐talk, and to consider different viewpoints for managing problems (Hofmann 2017). Interventions may also consist of sessions with psychoeducational contents, strategies for stress management, enhanced insight and relationship/rapport building, networking support, communication skills, self‐help interventions, and motivational enhancement (Buntrock 2016; Panter‐Brick 2018).
In many LMICs, training and retaining sufficient numbers of mental health specialists to meet current needs is not feasible. Therefore, it is important in these settings to consider options for expanding access to mental health promotion and disorder prevention services. Given that they are far more numerous and often more accessible than mental health specialists, the deployment of PWs for this purpose is one option that could prove to be of value in LMICs. This review, therefore, focused on the evaluation of a task‐shifting model as a possible implementation modality for prevention and promotive interventions in LMICs.
Why it is important to do this review
This review is limited to LMICs, where the need for PWs is greater than in high‐income settings. Far fewer mental health professionals are present in LMICs (the median number of psychiatrists is 172 times lower in low‐income countries (LICs) than in high‐income countries (HICs)) (Kakuma 2011; WHO 2011b). These differences in the organization of mental health services between LMICs and HICs, with poorer countries having few or no mental health service structures in primary care or in the community, mean that the problem of providing mental health interventions, especially in preventing mental disorders or promoting psychological well‐being, is different in such settings. PWs may need to work with little or no support from specialist mental health services. Consequently, PW interventions might be expected to function differently in LMICs as compared to HICs (Cuijpers 2018; Purgato 2019a).
The paucity of mental health professionals and the abundance of challenges for mental health systems in LMICs make it imperative to focus attention on prevention and promotion strategies via a public health approach (Tol 2015a). To address current shortages of mental health workers, prevention interventions in LMICs have been conceived as short, simple, and mostly delivered through a task‐shifting approach that includes different forms of intervention delivery. Delivery strategies range from informal delivery of simple interventions to more complex and longer strategies, as stepped‐care models. Task‐shifting is increasingly emphasized in global mental health and holds promise for improving access to mental health interventions (Jensen 2018; Patel 2018).
However, reviews on the task‐shifting approach to mental health interventions in LMICs have tended to focus more on treatment interventions (Akhtar 2022; Purgato 2018a; Singla 2017; Van Ginneken 2021). Evidence on the effectiveness of mental health prevention and promotion interventions in LMICs is scarce. Systematic reviews have focused mainly on populations living in high‐income settings, raising applicability concerns related to contextual factors and resource availability, including, for example, the lack of professionals in low‐resource contexts (i.e. psychiatrists, and psychologists) (Barbui 2020). In addition, LMICs differ from HICs with regard to social determinants of mental health, e.g. exposure to conflicts and wars, poverty, and gender‐based violence may be more frequent in LMICs (Lund 2018).
Populations in LMICs can conceptualize and seek assistance for mental health problems and mental health promotion in a wide variety of ways; these approaches may differ from conceptualizations and help‐seeking patterns seen in high‐income industrialized countries. Thus, evidence regarding the effectiveness of interventions implemented in HICs may not directly apply or be relevant to LMICs. For the reasons mentioned above, we expect that interventions might be applied differently in LMICs (Abdulmalik 2019).
The aim of this review is therefore to bridge this gap by assessing the effectiveness of delivery by primary workers of interventions for the promotion of mental health and for prevention of mental disorders or symptoms of mental illness in LMICs. Through this work, we aim to provide a picture of the characteristics and quality of the promotive and preventative research initiatives that have been carried out in lower resource settings. In addition, we strive to provide insights on what type of promotion strategies (universal, selective, and indicated) work best for whom (children and adults). This is overall in line with the WHO principle of integrating mental health into primary care and with the specific prevention objective of the WHO Action Plan for global mental health (WHO 2008; WHO 2021b).
This review has been conducted in parallel with an update of the Cochrane Review focused on treatment interventions in LMICs (Van Ginneken 2013; Van Ginneken 2021).
Objectives
-
To assess the effectiveness of delivery by primary workers of interventions for the promotion of mental health and for prevention of mental disorders or symptoms of mental illness in LMICs.
-
To examine the impact of intervention delivery by primary workers on resource use and costs associated with the provision of mental health care in LMICs.
Methods
Criteria for considering studies for this review
Types of studies
We included RCTs. We included trials that employed a cross‐over design ‐ whilst we acknowledge that this design is rarely used in intervention studies ‐ and we used data from the first randomized stage only. We excluded quasi‐randomized trials, such as those that allocate participants by using alternate days of the week. We considered both individual and cluster‐randomized trials as eligible for inclusion.
We included economic studies conducted as part of included effectiveness studies. We considered full economic evaluations (cost‐effectiveness analyses, cost‐utility analyses, or cost‐benefit analyses), cost analyses, and comparative resource utilization studies. We planned to report only cost and resource usage outcomes from these studies.
Types of participants
Participants
We included participants of any age, gender, ethnicity, and religion. We conducted two separate meta‐analyses on the different outcomes ‐ one for children and adolescents (younger than 18 years) and one for adults (18 years of age and older). Studies with mixed population groups (children and adolescents; adults) were allocated according to the proportion of participants belonging to the child and adolescent age range, or to the adult age range. When the intervention was designed and directed to a specific population group (i.e. children and adolescents or adults), we entered outcome data on efficacy for the specific target group for which the intervention was designed and delivered. For each of these two populations, we conducted meta‐analyses by different outcomes. We included studies focused on the prevention of any common mental disorder. Carers of study participants were included in meta‐analyses on adults, as some interventions may be directed at carers rather than at participants themselves, and consequently, mental health outcomes were measured in carers.
Settings
We considered studies conducted in LMICs. We used the World Bank criteria for categorizing a country as low‐ or middle‐income (World Bank 2020); these criteria provide a historical date of when countries were LMICs. If a country was an LMIC at some point during the recruitment of study participants, we included the study. We excluded studies undertaken in high‐income countries (at the time of study recruitment).
We included mental health promotion and/or prevention interventions delivered in primary care or community settings, refugee camps, schools, communities, survivors’ homes, and detention facilities. We excluded studies evaluating mental health promotion and/or prevention interventions undertaken outside of primary or community settings.
Diagnoses
Given the focus on interventions for the promotion of mental health and prevention of mental disorders, we restricted this review to participants without any formal diagnosis at the time the trial was undertaken. However, because many studies screened on the basis of a risk factor or heightened symptoms (without excluding people with diagnosed mental disorders), we could not exclude trial participants who might have fulfilled the criteria for an actual psychiatric diagnosis that remained unobserved because it was not investigated when the trial was undertaken. For example, we included populations who left their homes due to a sudden impact, threat, or conflict; populations exposed to political violence/armed conflicts/natural and industrial disasters; those with major losses or in poverty; and those belonging to a group (i.e. discriminated against or marginalized) experiencing political oppression, family separation, disruption of social networks, destruction of community structures and resources and trust, increased gender‐based violence, and undermined community structures or traditional support mechanisms (IASC 2007).
Comorbidities
We included studies that recruited participants with physical disorders and studies that focused on the prevention of multiple mental disorders.
Types of interventions
Included interventions
We included trials of primary‐level and/or community health worker interventions for promoting mental health and/or preventing mental disorders. Included mental health promotion or prevention interventions were delivered by primary‐level and/or community workers. Primary‐level health workers (PHWs) are first‐level providers who have received general health rather than specialist mental health training and can be based in a primary care clinic or in the community. PHWs include professionals (doctors, nurses, midwives, and other general health professionals) and para‐professionals (such as trained lay health providers, e.g. traditional birth attendants). PHWs do not include those with specialist mental health training, for example, psychiatrists, psychologists, psychiatric nurses, or mental health social workers. Community workers (CWs) are individuals who work in the community but not within the health sector. These might include teachers, trainers, support workers from schools and colleges, and other volunteers or workers within community‐based networks or non‐governmental organizations (NGOs). These CWs are not trained health workers but have a mental health role. They represent a further human resource employed in the delivery of promotion and prevention interventions (Patel 2007a). In this review, both of these categories of providers (PHWs and CWs) were referred to under the umbrella heading of 'primary‐level workers' (PWs).
This review aimed to include the following comparisons.
-
Provision of promotion and/or prevention interventions by primary‐level health workers and/or community workers versus usual care (little prevention or promotion strategy).
-
Provision of promotion and/or prevention interventions by primary‐level health workers and/or community workers versus no prevention or promotion strategy.
-
Provision of promotion and/or prevention interventions by primary‐level health workers and/or community workers versus interventions delivered by professionals with specialist mental health training.
We grouped the interventions as follows.
Promotion of mental health (e.g. interventions with a mental health or psychological component aimed at creating living conditions and environments that support mental health and encourage healthy lifestyles). We included any types of promotion interventions with a mental health component, delivered at individual, group, family, community, and/or societal levels (National Academies of Sciences 2019).
Universal prevention (e.g. community‐wide provision of information on positive coping methods to help people feel safe and hopeful; to protect against human rights violations; and to support community‐wide efforts to reduce poverty as a key risk for mental illness) (IASC 2007).
Selective prevention (e.g. psychological first aid for people with heightened levels of psychological distress after exposure to severe stressors, loss, or bereavement). These interventions involve human, supportive, and practical help covering both a social and a psychological dimension. They work through communication (asking about people's needs and concerns; listening to people; and helping them to feel calm), practical support (i.e. providing meals or water); a psychological approach (including teaching stress management skills and helping people cope with problems) (WHO 2011a); facilitation of community support for vulnerable individuals by activating social networks and communication; and structured cultural and recreational activities supporting the development of resilience (Institute of Medicine 2009), such as traditional dancing, artwork, sports, and puppetry (Tol 2011).
Indicated prevention (e.g. mentoring programmes aimed at children with behavioural problems; psychosocial support for school children with subclinical levels of post‐traumatic stress disorder (PTSD), anxiety, depression, or somatic symptoms and related disorders; prevention of postnatal depression in women with heightened levels of prenatal symptoms) (Institute of Medicine 2009).
Interventions were delivered through any means, including, for example, face‐to‐face meetings, digital tools, radio, telephone, or self‐help booklets, between participants and PHWs. Both individual and group interventions were included, with no limit on the number of sessions.
As this review was conducted in parallel with the update of the Cochrane Review on treatment interventions (Van Ginneken 2013; Van Ginneken 2021), we looked at the aim of the study and decided whether the aim was prevention or treatment, and we looked at the inclusion criteria for participants (these criteria must include specific mental distress/prodromal symptoms or a diagnosable disorder). When there was no clear distinction between prevention and treatment groups, we made a pragmatic decision on whether these trials were primarily about well‐being/prevention or about treatment and then allocated them to the appropriate review, or included them in both reviews.
Excluded interventions
We excluded interventions aimed at treating people with a diagnosed mental disorder. We also excluded studies that included participants on the basis of scoring above a cut‐off on a symptom checklist, with the explicit authors' stated intention to identify people with mental disorders. We excluded interventions aimed at treating people with a diagnosed disorder (Van Ginneken 2021).
Types of outcome measures
Primary outcomes
-
Diagnosis (or a proxy thereof, as assessed by scoring above a cut‐off for a screening tool) of mental disorders at study endpoint, determined according to the Diagnostic and Statistical Manual of Mental Disorders (DSM) III (APA 1980), DSM‐III‐R (APA 1987), DSM‐IV‐TR (APA 2000), DSM‐V (APA 2013), International Classification of Diseases (ICD)‐10 (WHO 1992), or any other standardized criteria.
-
Diagnosis (or a proxy thereof, as assessed by scoring above a cut‐off for a screening tool) of mental disorders at 1 to 6 months post‐intervention.
-
Diagnosis (or a proxy thereof, as assessed by scoring above a cut‐off for a screening tool) of mental disorders at 7 to 24 months post‐intervention.
-
Quality of life.
-
Adverse events experienced during the intervention.
Secondary outcomes
-
Psychological functioning and impairment.
-
Changes in service utilisation and contact coverage, including admission rates to hospital whether related to mental disorder or not; attendance rates in regard to utilization of primary or community services; or increased demand and/or referral rates from the primary/community setting to a mental health specialist.
-
Changes in mental health symptoms captured on rating scales (i.e. depressive symptoms, anxiety symptoms, distress/PTSD symptoms).
-
Social outcomes (e.g. perception of social inclusion).
-
Resource use (for health services: personnel time allocated/number of consultations, other opportunity costs of the intervention, or other aspects of the health service; for participants: extra costs of travel, lost productivity, employment status, income, work absenteeism, retention, educational attainment).
-
Carer mental health.
We grouped primary and secondary outcomes into three sets of time points.
-
Post‐intervention (0 to 1 month after the intervention) (to detect incidence/symptoms, reduction of the intervention).
-
1 to 6 months post‐intervention (to detect sustained incidence/symptom reduction).
-
7 to 24 months post‐intervention (indicating medium‐ to long‐term reduction).
We chose the latest time point within that category if several time points fitted within a category. We however included a time point that correlated best with other studies being compared within each outcome.
We included studies that reported only secondary outcomes of the review.
Search methods for identification of studies
Electronic searches
The EPOC (Effective Practice and Organisation of Care) Information Specialist developed search strategies in consultation with the review authors.
We searched the following databases for primary studies, from inception to date of search on 29 November 2021.
-
Cochrane Central Register of Controlled Trials (CENTRAL; 2021, Issue 11) in the Cochrane Library.
-
MEDLINE and Epub Ahead of Print, In‐Process, In‐Data‐Review & Other Non‐Indexed Citations, Daily and Versions: 1946 to 24 November 2021, Ovid.
-
Embase: 1974 to 24 November 2021, Ovid.
-
CINAHL (Cumulative Index to Nursing and Allied Health Literature): 1980 to date searched, EBSCO.
-
Global Index Medicus, World Health Organization (WHO) (www.globalindexmedicus.net/).
-
APA PsycINFO: 1806 to November Week 3 2021, Ovid.
Search strategies comprised natural language and controlled vocabulary terms. We did not apply any limits on the language of publication, and we searched all databases from inception to the date of search, except for MEDLINE and Embase. MEDLINE and Embase were limited to records from the last few months to find those not yet included in CENTRAL. Searches were limited by the use of study design filters appropriate to the stated inclusion criteria. See Appendix 1 for all strategies used.
For this prevention review, we used the search strategies developed for the complementary Cochrane Review on treatment (Van Ginneken 2013; Van Ginneken 2021), with appropriate adaptation.
Searching other resources
Trial registries
-
WHO ICTRP (World Health Organization International Clinical Trials Registry Platform) (www.who.int/ictrp) (searched on 29 November 2021).
-
ClinicalTrials.gov, US National Institutes of Health (www.clinicaltrials.gov) (searched on 29 November 2021).
Grey literature
We conducted a grey literature search to identify studies not indexed in the databases listed above.
In particular, we:
-
reviewed reference lists of all included studies and relevant systematic reviews for additional potentially eligible primary studies;
-
contacted authors of included studies/reviews to clarify reported published information and to seek unpublished results/data;
-
contacted researchers with expertise relevant to the review topic/EPOC interventions.
Data collection and analysis
Selection of studies
We downloaded all titles and abstracts retrieved by electronic searching to a reference management database and removed duplicates. Review authors (NvG, MP, EU, EP, CC) independently screened titles and abstracts for inclusion. We (NvG, MP, CC, EU, EP, CCad) retrieved the full‐text study report/publication and independently screened the full text and identified studies for inclusion; we identified and recorded reasons for exclusion of ineligible studies. We resolved disagreements through discussion, or, if required, we consulted a third review author (CB, WT, DP).
We listed in the Characteristics of excluded studies table studies that initially appeared to meet the inclusion criteria but that we later excluded. We collated multiple reports of the same study so that each study rather than each report was the unit of interest in the review. We also provided any information we were able to obtain about ongoing studies. We recorded the selection process in sufficient detail to complete a PRISMA flow diagram (Liberati 2009).
Data extraction and management
We extracted descriptive and outcome data for each study using an adapted version of the EPOC standard data collection form (EPOC 2017a). We piloted the form on at least one study in the review. One review author (MP, DP, CC, EP, CCad, LA) independently extracted descriptive data consecutively, and these were cross‐checked by a second review author (EP, EU, JA). We noted in the Characteristics of included studies table if outcome data were reported in an unusable way. We extracted the following study characteristics from the included studies, and we entered the data into RevMan Web.
-
Methods: study design; number of study centres and locations; study settings; withdrawals; dates of study; follow‐up.
-
Participants: number; mean age; age range; gender; clinical conditions; inclusion criteria; exclusion criteria; other relevant characteristics such as ethnicity and socioeconomic status.
-
Interventions: type and length of intervention; theory of change (hypothesized risk, protective, promotive factors); full description of cadre(s) of primary‐level health and/or community workers including details on supervision, training, and length, frequency, and type of experience; intervention components; comparison; fidelity assessment.
-
Setting: country; type of health and/or community service (e.g. NGO, government‐funded).
-
Outcomes: main and other outcomes specified and collected; time points reported.
-
Notes: funding for the trial; notable conflicts of interest of trial authors; ethical approval.
For economic data, we developed a specific data extraction form based on the format and guidelines used to produce structured abstracts of economic evaluations for inclusion in the National Health Service (NHS) Economic Evaluation Database (NHS EED) (University of York 2002), which we adapted to the specific requirements of this review.
We contacted the authors of included studies via email to collect key unpublished information.
Review authors who were authors of included studies were not involved in the following steps for those studies: study selection, data extraction, risk of bias assessment, and GRADE assessment.
Assessment of risk of bias in included studies
Two review authors (EP, CC) independently assessed risk of bias applying Cochrane Risk of Bias 2 tool and using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions Chapter 8 (Higgins 2019), along with guidance from the EPOC Group (EPOC 2017b). We resolved disagreements by discussion or by consultation with a third review author (CB, MP). We assessed risk of bias according to the following domains.
-
Bias arising from the randomization process.
-
Bias due to deviations from the intended interventions.
-
Bias due to missing outcome data.
-
Bias in measurement of the outcome.
-
Bias in selection of the reported result.
We answered signalling questions as: yes, probably yes, probably no, no, no information. We summarized risk of bias judgements for study results for each of the domains listed. We assigned a risk of bias assessment (low, some concerns, high) to results of the included studies using the approach suggested in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2019). We considered results of studies that had low risk of bias for all key domains, or for which it seemed unlikely for bias to seriously alter the results, to have low risk of bias. We considered results of studies to have some concerns for their risk of bias when risk of bias in at least one domain indicated some concern, or when results were judged to have some bias that could plausibly raise doubts about the conclusions. We considered results of studies to have high risk of bias when they had high risk of bias in at least one domain, or when we judged that they had serious bias that decreased the certainty of the conclusions.
We considered blinding separately for all outcomes (e.g. for unblinded outcome assessment, risk of bias for all‐cause mortality may be very different than for a patient‐reported rating scale). When information on risk of bias was related to unpublished data or correspondence with a trialist, we noted this in the risk of bias table. We did not exclude studies on the grounds of their risk of bias, but we clearly reported the risk of bias when presenting study results.
When considering interventions' effects, we took into account the risk of bias for studies that contributed to that outcome.
We evaluated cluster‐randomized controlled trials using the specific section in the Risk of Bias 2 tool (Higgins 2019).
We conducted the review according to the published protocol and reported any deviations from it in the Differences between protocol and review section of the systematic review.
Measures of treatment effect
We estimated the effect of the intervention by using risk ratio (RR), together with the appropriate associated 95% confidence interval (CI), for dichotomous data; and mean difference (MD) or standardized mean difference (SMD), together with the 95% appropriate associated confidence interval, for continuous data (Higgins 2019). We ensured that an increase in scores for continuous outcomes could be interpreted in the same way for each outcome, and we explained the direction to the reader, and reported when the directions were reversed, if this was necessary. For SMDs, we used the Cochrane Handbook for Systematic Reviews of Interventions to interpret their clinical relevance: 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect (Cohen 1988; Higgins 2011). We attempted to establish minimally important differences per outcome (as suggested in Guyatt 2013).
Unit of analysis issues
We included cluster‐RCTs when primary healthcare facilities, schools, or classes within schools rather than single individuals were the unit of allocation. Because variation in response to psychological or social intervention between clusters may be influenced by cluster membership, we included, when possible, data adjusted with an intra‐cluster correlation coefficient (ICC).
We adjusted the results for clustering by multiplying standard errors of the estimates by the square root of the design effect when the design effect was calculated as DEff = 1 + (M ‐ 1) ICC, where M is the mean cluster size and ICC is the intra‐cluster correlation coefficient. When included studies did not report ICCs for respective outcome measures, we derived ICCs from a different outcome from the same study, or from a different study included in the same meta‐analysis. If the ICC value was not reported or was not available from trial authors directly, we assumed it to be 0.05 (Higgins 2011; Ukoumunne 1999). We combined adjusted measures of effects of cluster‐randomized trials with results of non‐cluster‐randomized trials when it was possible to adjust adequately the results of cluster trials.
Dealing with missing data
We contacted investigators to verify key study characteristics and to obtain missing outcome data when possible (e.g. when a study was identified as abstract only). We tried to compute missing summary data from other reported statistics. We documented all correspondence with trial authors and reported in the Results and Acknowledgements sections details from trial authors who responded. For cluster‐RCTs, we contacted study authors for an ICC value when data were not adjusted and could not be identified from the trial report. As mentioned above, when the ICC was neither available from the trial reports nor directly available from the trial authors, we assumed it to be 0.05 (Ukoumunne 1999).
For continuous data, we applied a looser form of intention‐to‐treat (ITT) analysis, whereby all participants with at least one post‐baseline measurement were represented by their last‐observation‐carried‐forward (LOCF). If the authors of included RCTs stated that they used an LOCF approach, we checked details on LOCF strategy and used data as reported by the study authors. When study authors reported only the standard error (SE) or t statistics or P values, we calculated standard deviations (SDs) according to Altman 1996.
For dichotomous data, we applied a modified ITT analysis without imputation of outcomes where missing, as all analyses with dichotomous outcomes were negative (diagnoses; adverse events).
When it was not possible to obtain data, we reported the level of missingness and considered how that might have impacted the certainty of evidence.
Assessment of heterogeneity
If we found a sufficient number of studies for which we judged participants, interventions, comparisons, and outcomes to be sufficiently similar, we conducted a meta‐analysis (Borenstein 2009). We obtained an initial visual overview of statistical heterogeneity by scrutinizing the forest plots while looking at the overlap between CIs around the estimate for each included study. To quantify the impact of heterogeneity on each meta‐analysis, we used the I² statistic, and we considered the following ranges, according to the Cochrane Handbook for Systematic Reviews of interventions (Higgins 2019).
-
0% to 40%: might not be important.
-
30% to 60%: may represent moderate heterogeneity.
-
50% to 90%: may represent substantial heterogeneity.
-
75% to 100%: represents considerable heterogeneity.
The importance of the observed I² statistic was related to the magnitude and direction of intervention effects and the strength of evidence for heterogeneity (Higgins 2011; Purgato 2012). If we identified substantial heterogeneity, we explored this through prespecified subgroup analysis.
Assessment of reporting biases
If we were able to pool more than 10 trials in a meta‐analysis, we created and examined a funnel plot to explore possible publication biases and interpreted the results with caution (Sterne 2011).
Data synthesis
We undertook meta‐analyses only when this was meaningful (i.e. when the population, intervention, comparison, outcome, and underlying intervention question and the theory of change were similar enough for pooling to make sense) (Borenstein 2009). A common way that trialists indicate when they have skewed data is by reporting medians and interquartile ranges. When we encountered this, we noted that the data were skewed and considered the implications of this. When multiple trial arms were reported in a single trial, we included only the relevant arms. If two comparisons (e.g. intervention A versus usual care and intervention B versus usual care) had to be entered into the same meta‐analysis, we halved the control group to avoid double‐counting. Whenever sufficient data were available, we grouped studies for comparison by type of provider (e.g. primary‐health workers‐led, community workers‐led, collaborative), type of intervention (promotion, universal, selective, indicated prevention), and particular risk, protective, or promotive factors targeted (Eaton 2012; Tol 2015a).
Given the potential heterogeneity of mental health promotion and prevention interventions, we used a random‐effects model in all analyses. The random‐effects model has the highest generalisability in empirical examinations of summary effect measures for meta‐analyses (Furukawa 2002). We examined the robustness of this summary measure by checking the results under a fixed‐effect model. We reported material differences between models.
Specifically, for dichotomous data, we used the Mantel‐Haenszel method, as this is preferable in Cochrane Reviews given its better statistical properties when there are few events (Higgins 2011). We adopted the inverse variance method for continuous data: this method minimizes the imprecision of the pooled effect estimate, as the weight given to each study is chosen to be the inverse of the variance of the effect estimate (Higgins 2011).
Economic data
We aimed to conduct all elements of the economics component of this review according to current guidance on the use of economics methods in the preparation and maintenance of Cochrane Reviews (Shemilt 2006). We classified the included economic evaluations based on an established system (Drummond 2005; Trautmann 2016). We summarized the characteristics and results of included economic evaluations by using additional tables. We displayed resource use and cost data in a table, along with unit cost data (when available). A unit cost is defined as the cost of each specific resource input calculated by multiplying the measured number of units (quantities) of an item of resource use (e.g. the number of hours of time provided by a senior teacher) by an applicable unit cost (e.g. the salary cost of one hour of senior teacher time). We reported the currency and price year applicable to measures of costs and unit costs in each original study. Measures of costs are highly likely to vary across and within study settings and over time. This is the product of variations in underlying quantities of resource use and variations in underlying unit costs. This approach is consistent with that used in the parallel review on treatment that is being updated (Van Ginneken 2013).
Subgroup analysis and investigation of heterogeneity
Within each comparison, we planned the following subgroup analyses.
-
Category of health worker (e.g. professionals, health workers, non‐professional health workers, community workers).
-
Setting of care.
-
Community settings, camps, schools.
-
Chronic or acute humanitarian versus non‐humanitarian settings.
-
-
Type of prevention intervention (universal, selective, indicated).
-
Type of promotion intervention (individual, group).
-
Specific risk, protective, or promotive factor targeted.
We planned to use the following outcomes in subgroup analysis.
-
Proportion of individuals developing new mental distress or mental disorders (incidence).
-
Quality of life outcomes.
-
Harmful outcomes: number of people experiencing harm during the intervention.
-
Change from baseline in average rating scale scores (e.g. psychological symptoms) for the study population.
If the number of included studies for each comparison was sufficient, we performed subgroup analyses to check whether the intervention effect varied with different population characteristics. When applicable, or when subgroup analysis was not possible, we described subgroup differences narratively in the Results section.
For random‐effects meta‐analyses, we used the formal Chi² test and the I² statistic for subgroup differences in RevMan 5, to detect statistically significant subgroup differences.
Sensitivity analysis
We planned to perform sensitivity analyses defined a priori to assess the robustness of our conclusions and to explore its impact on effect sizes. This involved the following.
-
Restricting analysis to published studies.
-
Restricting analysis to studies measuring the incidence of mental disorders (i.e. studies in which all participants at baseline scored below defined symptom thresholds on rating scales).
-
Restricting analysis to studies with low risk of bias, as specified in incomplete outcome data and selective reporting.
-
Excluding trials with methodological characteristics that might generate the highest heterogeneity in a meta‐analysis (when a meta‐analysis has I² > 75%).
Stakeholder consultation and involvement
Consultation with stakeholders was conducted by authors Nadja van Ginneken, Simon Lewin, and Paul Garner of the parallel review focused on treatment of mental disorders in LMICs (Van Ginneken 2021). Consultation was organized as follows.
-
A group face‐to‐face consultation with seven LMIC clinicians who were mature students/master's students or PhD students at the Liverpool School of Tropical Medicine (December 2018).
-
An online consultation with seven implementers, academics, and policy‐makers from LMICs, and a further four written answers from further stakeholders received by email (February to April 2019).
-
An updated literature review of mental health terminology and descriptions.
The overall message that emanated from this consultation was that we should consider the spectrum of mental illness as broader than encompassing only diagnostic categories (Patel 2018). According to this framework, the current review complemented the parallel review on treatment interventions in LMICs.
Summary of findings and assessment of the certainty of the evidence
Two review authors (EP, CC) independently assessed the certainty of the evidence (high, moderate, low, and very low) using the five GRADE considerations (risk of bias, consistency of effect, imprecision, indirectness, and publication bias) (Guyatt 2008). We used methods and recommendations described in Chapter 14 of the Cochrane Handbook for Systematic Reviews of interventions (Higgins 2019), and in the EPOC Worksheets (EPOC 2017c), and we used GRADEpro software (GRADEpro GDT). We resolved disagreements on certainty ratings by discussion and planned to provide justification for decisions to downgrade or upgrade ratings by using footnotes in the table and making comments to aid readers' understanding of the review when necessary. We used plain language statements to report these findings in the review (EPOC 2017c).
We summarized review findings in a summary of findings table(s) for the main intervention comparison(s) and included the following outcomes.
-
Diagnosis (or a proxy thereof, as assessed by scoring above a cut‐off of a screening tool) of mental disorders at study endpoint, determined according to the Diagnostic and Statistical Manual of Mental Disorders (DSM) III (APA 1980), DSM‐III‐R (APA 1987), DSM‐IV‐TR (APA 2000), DSM‐V (APA 2013), International Classification of Diseases (ICD)‐10 (WHO 1992), or any other standardized criteria.
-
Quality of life.
-
Adverse events experienced during the intervention.
-
Change in mental health symptoms seen on rating scales (i.e. depressive symptoms, anxiety symptoms, distress/PTSD symptoms).
-
Resource use (for health services: personnel time allocated/number of consultations, other opportunity costs of the intervention, or other aspects of the health service; for participants: extra cost of travel, lost productivity, employment status, income, work absenteeism, retention, educational attainment).
-
Psychological functioning and impairment.
If during the review process, we became aware of an important outcome that we failed to list in our planned summary of findings tables, we planned to include the relevant outcome and explain the reasons for this in the section Differences between protocol and review. We considered whether there was any additional outcome information that could not be incorporated into meta‐analyses and noted this in the comments; we stated if this supported or contradicted information derived from meta‐analyses. If it was not possible to meta‐analyze the data, we summarized the results in the text. Only the post‐intervention time point was presented for each outcome in the summary of findings tables.
Results
Description of studies
Detailed descriptions of all studies are found in Characteristics of included studies. This section contains detailed characteristics of each included study.
Results of the search
We screened 11,938 records and included 113 randomized controlled trials with 32,992 participants and 97 in the quantitative synthesis of this review (19,570 participants). We identified 83 ongoing studies, and 25 studies awaiting classification. From our search of all the above databases and trial registries on 20 August 2020, we screened 9437 titles and abstracts (Figure 1). We performed a second search on 29 November 2021, we screened 2501 titles and abstracts and identified 27 studies that are included (Figure 1). We contacted 76 trial authors for additional information and/or data, and 33 of them replied to our requests.
Flow chart of studies
Included studies
Study design
Of the 113 included studies, 67 were randomized trials and 46 were cluster‐randomized trials.
Setting
Of the 113 included studies, 19 were conducted in low‐income countries, 27 in low‐middle‐income countries, 2 studies in middle‐income countries, and 58 studies in upper‐middle‐income countries. Seven studies were conducted in mixed settings, i.e. low to lower‐middle income or upper‐middle to lower‐middle‐income. A map of countries where the RCTs were conducted is available in Figure 2.
Map of countries from included studies
There were 23 studies from rural, 63 studies from urban and 11 studies in both rural and urban contexts. For 16 studies, the setting was not specified. Some interventions were delivered in freestanding primary care facilities (12 studies) or in hospital‐based primary care facilities (17 studies). In 19 studies, the interventions were delivered in community groups, in 26 studies at school, and in one study in the workplace. Interventions were delivered at home or in other settings in 36 studies, while in two studies the setting was not specified.
Participants
There were 83 studies that included adults and 30 that included children (up to 18 years).
Conditions
Most studies (n = 65) targeted the prevention of common mental disorders (39 studies of which 34 were included in meta‐analyses) and/or psychological distress (26 studies of which 21 were included in meta‐analyses). Amongst the other included studies, 18 covered perinatal depression or mental disorders and 30 targeted child mental disorders (24 studies included in meta‐analyses). See Effects of interventions for details on the analyses.
Interventionists
Various cadres of primary‐level workers were employed: primary care health workers (38 studies), community workers (71 studies), or both (2 studies). In two studies, the intervention was delivered online and the specific cadres were not described. The remuneration of interventionists was generally not reported or poorly described. The background of these interventionists is reported in detail under Characteristics of included studies.
Interventions
In 22 studies, the prevention interventions delivered were categorized as universal prevention or promotion of well‐being, targeting the general public or the whole population group, regardless of whether people were at higher risk or not of developing a disorder. For example, we included in this category programmes for all caregiver‐child dyads of students attending primary schools and designed to develop social skills in elementary‐age children using a universal, low‐cost approach. In 36 studies, the delivered interventions were categorized as selective prevention, targeting participants whose risk of developing a mental disorder was significantly higher than that of the rest of the population (e.g. exposed to specific risk factors), for example, support interventions for caregivers of patients with Alzheimers with distress levels below the cut‐off. In 55 studies, interventions were categorized as indicated prevention, as they were directed to smaller groups of participants at high risk for mental disorders and that might already present with some sign of disorder like psychological symptoms, without meeting the criteria for a formal diagnosis according to diagnostic tools or cut‐offs in rating scales. For example, we included in this category the intervention Self‐Help Plus, a pre‐recorded audio course, delivered by trained facilitators in a group setting and complemented with an illustrated self‐help book adapted for the target cultural group. The intervention is based on acceptance and commitment therapy and was delivered to participants with increased psychological distress (GHQ ≥ 3) and without a formal diagnosis according to the MINI. Intervention categorization is summarized in Figure 3.
Economic studies
Of the included studies, three (Chang 2015; Osborn 2020; Tol 2012) reported economic evaluations which are summarized in Table 2 with costs reported in dollars. These studies were not included in meta‐analyses due to the heterogeneity across measures and economic outcomes.
| Author year | Country | Type of economic analysis | Study population | Interventions | Intervention‐specific costs and cost‐effectiveness | Resources (i.e. costs to health services other than intervention costs; patient/society costs and productivity) |
|---|---|---|---|---|---|---|
| Sri Lanka | Cost analysis | Children | Classroom‐based intervention (CBI) vs waiting list | Costs: cost analyses for intervention group demonstrated mean cost per service user was USD 8.85 (56% of which is human resources cost) | Health service costs: ‘costs’ included broader package | |
| ‐ | ‐ | Cost data: calculated | ‐ | Delivered by LHWs (paraprofessional | Cost‐effectiveness: cost analyses | Patient cost: none reported |
| Jamaica | Cost analysis | Adults (female, age not specified) | Parenting intervention with routine primary health care vs usual care | Costs: the cost per child was USD 100.9 for 1 year of intervention | Health service cost: USD 100.9 per child including equipment purchases, materials, training, and wages | |
| ‐ | ‐ | ‐ | ‐ | Delivered by community health workers and nurses; training: 3‐day workshops with viewing of films and role plays | Cost‐effectiveness: not reported | Patient cost: none reported |
| Kenya | Cost analysis | Adolescents (both male and females, 13 to 18 years of age) | Shamiri‐Digital Wellness vs active control | Cost: USD 3.57 per student to deliver Shamiri‐Digital | Health service cost: health service costs included equipment (computers, desks, chairs) with an hourly cost of USD 0.97, totalling USD 104.65 for the 9 months of the intervention | |
| ‐ | ‐ | ‐ | ‐ | Self‐help digital‐based intervention | Cost‐effectiveness: depending on the definition of clinically meaningful improvement, 7.1 to 9.7 students needed to receive the intervention for 1 student to experience a clinically meaningful improvement, which translated to a cost of USD 25.35 to USD 34.62 per student | Patient cost: none reported |
CBI: classroom‐based intervention
LHW: lay health workers
USD: US dollar
vs: versus
Excluded studies
We excluded 218 studies. These studies were of interest but were excluded due to specific criteria for participants, interventions, comparators, and outcomes (PICOs) not being met (see Characteristics of excluded studies). The most common reason for exclusion was ineligible population. The other reasons for exclusion were, in descending order, the following: ineligible intervention, ineligible study design, ineligible outcome, ineligible setting, and duplicates.
Risk of bias in included studies
We applied the Cochrane Risk of Bias 2 tool to individual and cluster‐randomized trials (Higgins 2019). The most often identified biases across studies were performance bias ('risk of bias due to deviations from the intended interventions' ‐ 'effect of assignment to intervention and effect of adhering to intervention'), attrition bias ('risk of bias due to missing outcome data'), and reporting bias ('risk of bias in the selection of the reported result').
Adults
1. Risk of bias arising from the randomization process (Domain 1‐A)
Two studies (Jewkes 2008; Sanfilippo 2020) had a high risk of bias with regard to Domain 1‐A. In both cluster‐randomized controlled trials, the reported high RoB, was determined by the timing of identification or recruitment of participants. This is because individual participants were not identified and recruited before randomization and, for both, it was likely that the selection of individual participants was affected by knowledge of the intervention assigned to the cluster as assessed by items 1b.1 and 1b.2.
Ten studies presented a risk of bias classified as having 'some concerns'. This is because, in most cases, studies did not provide enough information on allocation concealment, and 62 had a low risk.
2. Risk of bias due to deviations from the intended interventions (effect of assignment to intervention and effect of adhering to intervention) (Domain 2‐B)
A high risk of bias for the Domain 2‐B was reported for 13 outcomes from six studies (Chang 2015; Hirani 2018; Ozcan 2020; Rachasrimuang 2018; Skar 2021; Ward 2020) in the adult population.
For Chang 2015(indicated prevention; depressive symptoms at endpoint), Ward 2020(indicated prevention; depressive symptoms, distress/PTSD symptoms and social outcomes, all at endpoint and 7‐24 months), Hirani 2018(universal prevention; quality of life at endpoint), and Rachasrimuang 2018(selective prevention; quality of life and depressive symptoms, both at 1 to 6 months) this is due to deviations from the intended intervention that arose because of the trial context, as assessed by item 2.3. These deviations were deemed to be likely affecting the outcome and to be not balanced between groups (items 2.4, 2.5). Skar 2021(indicated prevention; diagnosis of mental disorders at 1 to 6 months) and Ozcan 2020(selective prevention; quality of life and social outcomes, both at 1 to 6 months) instead present a high risk of bias related to the analyses used to estimate the effect of assignment to intervention (items 2.6 and 2.7). For Skar 2021, this was because not enough information was provided: a) to evaluate if an appropriate analysis was used to estimate the effect of assignment to intervention, and b) to evaluate the potential for a substantial impact (on the result) of the failure to analyze participants in the group to which they were randomized. For Ozcan 2020 instead, there was a chance that an appropriate analysis was not used to estimate the effect of assignment to intervention, as indicated by the quote: "The exclusion criteria were as follows: (i) being lost to follow‐up (…)".
For Domain 2‐B, 57 out of 175 outcomes from 26 included studies, presented a risk of bias classified as having 'some concerns'. Most of these studies did not provide enough information on whether the analysis used to estimate the effect of adhering was appropriate. Nevertheless, they held no potential for a substantial impact (on the result) of the failure to analyze participants in the group to which they were randomized.
A total of 105 out of the 175 outcomes from 46 included studies had a low risk of bias.
3. Risk of bias due to missing outcome data (Domain 3‐C)
A high risk of bias for the Domain 3‐C was reported for 12 outcomes from five included studies (Dybdahl 2001, indicated prevention; quality of life, distress/PTSD symptoms and social outcomes, all at endpoint; Rachasrimuang 2018, selective prevention; quality of life and depressive symptoms, both at 1‐6 months; Rodriguez 2021, indicated prevention; depressive symptoms, anxiety symptoms and distress/PTSD symptoms, all at endpoint); Rotheram‐Borus 2014a, indicated prevention; diagnosis of mental disorder at endpoint, 1 to 6 months and 7 to 24 months; Skar 2021, indicated prevention; diagnosis of mental disorders at 1 to 6 months) for the adult population. For all, data for this outcome were not available for all randomized participants/clusters that recruited participants/individual participants within clusters, or information on this (item 3.1) was not reported. In addition, there was no evidence that the result was not biassed by missing data (3.2) and it was likely that 'missingness' in the outcome depended on its true value (3.3, 3.4). Alternatively, not enough information was provided to give a judgement on the item.
For Domain‐3‐C, 9 out of 175 outcomes from four included studies, presented a RoB classified as having 'some concerns'. For these outcomes, data were not available for all randomized participants/clusters that recruited participants/individual participants within clusters, and there was no evidence that results were not biassed by missingness.
A total of 154 out of the 175 outcomes from 65 included studies had a low risk of bias.
4. Risk of bias in measurement of the outcome (Domain 4‐D)
A high risk of bias for Domain 4‐D was reported for three outcomes from one included study (Duan 2019; promotion/universal prevention; quality of life, depressive symptoms and distress/PTSD symptoms, all at endpoint). This was due to the fact that a different method for assessment was used at post‐intervention for the control and intervention groups.
For Domain 4‐D, 151 out of 175 outcomes from 72 included studies had a RoB classified as having 'some concerns'. For the great majority of these, outcome assessors (which were participants for self‐reported outcome measures) were most likely aware of intervention allocation across these non‐blinded designs. Across these studies, knowledge of the assigned intervention could have influenced participant‐reported outcomes, but there was no reason to believe that it did.
A total of 21 out of the 175 outcomes from five included studies had a low risk of bias.
5. Risk of bias in selection of the reported result (Domain 5‐E)
A high risk of bias for Domain 5‐E was reported for five outcomes from three included studies (Latina 2019; Rodriguez 2021; Rotheram‐Borus 2014a). For Rotheram‐Borus 2014a(indicated prevention; diagnosis of mental disorder at endpoint, 1 to 6 months and 7 to 24 months) and Latina 2019(promotion/universal prevention; quality of life at endpoint), this was due to the fact that one outcome mentioned in the protocol was not reported in the results, suggesting outcome selection (item 5.2). Rodriguez 2021(indicated prevention; distress/PTSD symptoms at endpoint), instead presented a high risk of bias related to both outcome selection and analyses selection.
For Domain 5‐E, 15 out of 175 outcomes from six included studies, had a RoB classified as having 'some concerns'. Most of these studies did not provide enough information to assess if a trial was analyzed in accordance with a prespecified plan or selection from multiple outcomes or analyses occurred.
A total of 155 out of the 175 outcomes from 67 included studies had a low risk of bias.
6. Overall risk of bias
An overall high risk of bias was reported for 31 outcomes from 13 included studies (Chang 2015; Duan 2019; Dybdahl 2001; Hirani 2018; Jewkes 2008; Latina 2019; Ozcan 2020; Rachasrimuang 2018; Rodriguez 2021; Rotheram‐Borus 2014a; Sanfilippo 2020; Skar 2021; Ward 2020) analyzing promotion/prevention interventions amongst adults. This overall judgement was determined by a high risk of bias reported in domain 1‐A for two studies (Jewkes 2008 and Sanfilippo 2020), domain 2‐B for six studies (Chang 2015; Hirani 2018; Ozcan 2020; Rachasrimuang 2018; Skar 2021; Ward 2020), domain 3‐C for five studies (Dybdahl 2001; Rachasrimuang 2018; Rodriguez 2021; Rotheram‐Borus 2014a; Skar 2021), domain 4‐D for one study (Duan 2019), and finally domain 5‐E for three studies (Latina 2019; Rodriguez 2021; Rotheram‐Borus 2014a). Amongst these, four studies had a high risk of bias for more than one domain (Rachasrimuang 2018; Rodriguez 2021; Rotheram‐Borus 2014a; Skar 2021).
A total of 127 out of the 175 outcomes from 59 included studies were reported to have 'some concerns', 17 outcomes from four studies had instead a low risk of bias.
Children
1. Risk of bias arising from the randomization process (Domain 1‐A)
No study had a high risk of bias with regard to the randomization process. A total of four studies were reported to have 'some concerns' in Domain 1‐A. Amongst these, most did not provide enough information on allocation concealment. A total of 20 studies had a low risk of bias.
2. Risk of bias due to deviations from the intended interventions (effect of assignment to intervention and effect of adhering to intervention) (Domain 2‐B)
A high risk of bias for Domain 2‐B was reported for eight outcomes from four studies (Ager 2011; Hull 2021; Mohamadi 2021; Velásquez 2015).
For Velásquez 2015(promotion/universal prevention; depressive symptoms, anxiety symptoms, and social outcomes, all at 1 to 6 months) and Hull 2021(indicated prevention; anxiety symptoms, and social outcomes, both at endpoint), this was due to deviations from the intended intervention that arose because of the trial context, as assessed by item 2.3. These deviations were deemed to be likely affecting the outcome and to be not balanced between groups (items 2.4, 2.5).
The two aforementioned studies as well as Ager 2011(selective prevention; quality of life at endpoint) and Mohamadi 2021(promotion/universal prevention; distress/PTSD symptoms, at endpoint and 1‐6 months), presented a high risk of bias related to the analyses used to estimate the effect of assignment to intervention (items 2.6 and 2.7). This was because not enough information was provided to evaluate: a) if an appropriate analysis was used to estimate the effect of assignment to intervention, and b) the potential for a substantial impact (on the result) of the failure to analyze participants in the group to which they were randomized.
Nineteen out of 71 outcomes from seven included studies, were reported as having 'some concerns' in Domain 2‐B. Amongst these, most did not provide enough information on whether the analysis used to estimate the effect of adhering was appropriate. Nevertheless, they held no potential for a substantial impact (on the result) of the failure to analyze participants in the group to which they were randomized.
A total of 44 out of the 71 outcomes from 13 included studies had a low risk of bias.
3. Risk of bias due to missing outcome data (Domain 3‐C)
A high risk of bias for Domain 3‐C was reported for five outcomes from three included studies (Ager 2011; Hull 2021; Mohamadi 2021). For Ager 2011(selective prevention; quality of life at endpoint), data were not available for all randomized participants (3.1), there was no evidence that the result was not biassed by missing data (3.2). It was likely that missingness in the outcome data depended on its true value (3.3, 3.4). For Hull 2021(indicated prevention; anxiety symptoms, and social outcomes, both at endpoint) and Mohamadi 2021(promotion/universal prevention; distress/PTSD symptoms, at endpoint and 1 to 6 months), not enough information was provided to provide a judgement on any of these items.
Two out of 71 outcomes from one included study were reported to have 'some concerns' in Domain 3‐C. Amongst these, data were not available for all randomized participants, clusters, or individual participants within a cluster, and there was no evidence that results were not biassed by missingness.
A total of 64 out of the 71 outcomes from 20 included studies had a low risk of bias.
4. Risk of bias in measurement of the outcome (Domain 4‐D)
No study had a high risk of bias due to the measurement of the outcome. All outcomes from all 24 included studies were reported to have 'some concerns' in Domain 4‐D. Outcome assessors (which were participants for self‐reported outcome measures) were most likely aware of intervention allocation across these non‐blinded designs. Across these studies, knowledge of the assigned intervention could have influenced participant's reported outcomes, but there was no reason to believe that it did.
5. Risk of bias in selection of the reported result (Domain 5‐E)
A high risk of bias for the Domain 5‐E was reported for six outcomes from three included studies (Ager 2011; Dhital 2019; Osborn 2020). For Ager 2011(selective prevention; quality of life at endpoint) and Osborn 2020(indicated prevention; quality of life, depressive symptoms and anxiety symptoms, all at endpoint), this was due to the fact that one outcome mentioned in the protocol was not reported in the results, suggesting outcome selection. Dhital 2019(selective prevention; depressive symptoms and distress/PTSD symptoms, both at 1 to 6 months), instead presented a high risk of bias related to the fact that not all planned data collection was reported in the results.
Two out of 71 outcomes from one included study, were reported to have 'some concerns' in Domain 5‐E. This was because not enough information was provided to assess if a trial was analysed in accordance with a prespecified plan or selection from multiple outcomes or analyses that occurred was provided.
A total of 63 out of the 71 outcomes from 21 included studies had a low risk of bias.
6. Other potential sources of bias
We visually inspected funnel plots to identify asymmetry in any of the comparisons between psychological treatments and comparators when we identified 10 or more studies. We did not identify any asymmetry in the distribution of studies (Figure 4; Figure 5).
7. Overall risk of bias
An overall high risk of bias was reported for 13 outcomes from six studies (Ager 2011; Dhital 2019; Hull 2021; Mohamadi 2021; Osborn 2020; Velásquez 2015) analyzing promotion/prevention interventions amongst children. This overall judgement was determined by a high risk of bias reported in three domains (2, 3 and 5) for one study (Ager 2011), in two domains (2 and 3) for two studies (Mohamadi 2021 and Hull 2021), and in one domain for three studies (Velásquez 2015 (2) Dhital 2019 and Osborn 2020 (5)).
For the remaining 58 out of 71 outcomes from 18 included studies, the risk of bias was classified as having 'some concerns'. No outcome or study had an overall low risk of bias.
Effects of interventions
See: Summary of findings 1 Summary of findings table ‐ Promotion/universal prevention interventions compared to control group in preventing mental disorders in adults; Summary of findings 2 Summary of findings table ‐ Selective prevention interventions compared to control group in preventing mental disorders in adults; Summary of findings 3 Summary of findings table ‐ Indicated prevention interventions compared to control group in preventing mental disorders in adults; Summary of findings 4 Summary of findings table ‐ Promotion/universal prevention interventions compared to control group in preventing mental disorders in children; Summary of findings 5 Summary of findings table ‐ Selective prevention interventions compared to control group in preventing mental disorders in children; Summary of findings 6 Summary of findings table ‐ Indicated prevention interventions compared to control group in preventing mental disorders in children
Results of meta‐analyses (n = 97 studies) of outcomes for the following comparisons were described:
-
Primary‐level health worker and/or community worker‐led universal prevention or promotion interventions for adults (n = 11 studies, 11.3%);
-
Primary‐level health worker and/or community worker‐led selective prevention interventions for adults (n = 20 studies, 20.6%);
-
Primary‐level health worker and/or community worker‐led indicated prevention interventions for adults (n = 42 studies, 43.3%);
-
Primary‐level health worker and/or community worker‐led universal prevention or promotion interventions for children (n = 8 studies, 8.2%);
-
Primary‐level health worker and/or community worker‐led selective prevention interventions for children (n = 7 studies, 7.2%);
-
Primary‐level health worker and/or community worker‐led indicated prevention interventions for children (n = 9 studies, 9.3%).
We were not able to group studies by type of provider (e.g. primary health workers‐led, community workers‐led, collaborative), and by risk, protective, or promotive factors targeted due to the small number of studies for each subgroup, and to the lack of information on risk, protective, and promotive factors. We were not able to report the comparison of promotion or prevention interventions versus interventions delivered by professionals with specialist mental health training, as no data were available for this comparison.
Outcome data on efficacy were extracted for the specific population group for which the intervention was designed and delivered. For example, when the intervention was designed and directed to adults, we entered efficacy data on adults (even though in some RCTs data on children were also available to measure potential indirect benefits). The same applied for interventions directed to children and adolescents.
Comparison 1. Primary‐level health worker and/or community worker‐led universal prevention or promotion interventions for adults (n = 11 studies)
We identified 11 RCTs (6 cluster‐RCTs (54.5%) and five individual RCTs (45.5%)) contributing to meta‐analyses in this comparison. The total sample size was n = 3099 (1567 participants (50.6%) for the intervention group and 1532 participants (49.4%) for the control group).
As for the type of PWs, the majority of interventions (n = 7 studies, 63.6%) were delivered by community workers (CWs) whereas, in four studies (36.4%), interventions were delivered by primary health workers (PWs) (see Appendix 2 for a complete list of studies regarding the classification of PWs). CWs included, for instance, trained facilitators, caregivers, coaches, and peer leaders; PWs included social workers, community‐based health workers, medical teachers, and village doctors. For most of the studies, participants in the control group were allocated to usual care (n = 9 studies, 81.8%); for the remaining two studies (18.2%), the control group was guided by a PW without training and/or supervision or was allocated to active control. Two studies (18.2%) had been conducted in schools, two (18.2%) in a community setting, one (9.1%) at home, one (9.1%) at a workplace and the remaining five (45.5%) in other settings, such as universities, or local parishes. From a wider geographical perspective, four studies (36.4%) were conducted in an urban context, four (36.4%) in a rural one and, for the remaining three studies (27.3%), it was not specified. Seven studies (63.6%) were undertaken in upper‐middle income countries, three (27.3%) in low‐middle‐income countries and for one study (9.1%) income level changed from low‐middle to upper‐middle. As for the condition, most studies addressed psychological distress (n = 8 studies, 72.7%), two studies (18.2%) targeted the prevention of common mental disorders, and one (9.1%) addressed perinatal mental disorders.
Primary outcomes
1. Diagnosis (or a proxy thereof, as assessed by scoring above a cut‐off for a screening tool) of mental disorders
At 0 to 1 month post‐intervention
No studies reported on this outcome.
At 1 to 6 months post‐intervention
One study with 137 participants (George 2020) identified that universal prevention/promotion interventions may lead to a decrease in the incidence of mental disorders at 1 to 6 months, although the actual effect range indicates it may have little or no difference (RR 0.57, 95% CI 0.28 to 1.18; I2 = NA; P = 0.13) (Analysis 1.1).
At 7 to 24 months post‐intervention
One study with 323 participants (Rockers 2018) identified that universal prevention/promotion interventions may lead to a decrease in the incidence of mental disorders at 7 to 24 months, although the actual effect range indicates it may have little or no difference (RR 0.67, 95% CI 0.43 to 1.05; I2 = NA; P = 0.08) (Analysis 1.2).
2. Quality of life
At 0 to 1 month post‐intervention
At 0 to 1 month post‐intervention, it is uncertain whether universal prevention/promotion interventions improve quality of life (SMD ‐0.23, 95% CI ‐0.51 to 0.04; I2 = 61%; P = 0.10; 4 studies, 684 participants; very low‐certainty evidence due to study limitations, inconsistency, and indirectness) (Analysis 1.3).
At 1 to 6 months post‐intervention
No studies reported on this outcome.
At 7 to 24 months post‐intervention
One study (Zhou 2010) identified a positive effect of universal prevention/promotion interventions over the control group (MD ‐7.30, 95% CI ‐12.28 to ‐2.32; I2 = NA; P = 0.004; 222 participants) (Analysis 1.4).
3. Adverse events experienced during the intervention
No studies reported on this outcome.
Secondary outcomes
1. Psychological functioning and impairment
No studies reported on this outcome.
2. Changes in service utilization and contact coverage
No studies reported on this outcome.
3. Changes in mental health symptoms captured on rating scales (i.e. depressive symptoms, anxiety symptoms, distress/PTSD symptoms)
(Analysis 1.5; Analysis 1.6; Analysis 1.7; Analysis 1.8; Analysis 1.9)
Three studies (349 participants) (Duan 2019; Hendriks 2019; Yusoff 2015) reported data on depressive symptoms at 0 to 1 month post‐intervention, indicating uncertainty of beneficial effects for universal prevention/promotion interventions versus control (SMD ‐0.31, 95% CI ‐0.78 to 0.15; I2 = 61%; P = 0.19; 3 studies, 349 participants; very low confidence due to study limitations, inconsistency and indirectness) (Analysis 1.5). At 1 to 6 months, only one study with 153 participants provided data (Yusoff 2015) indicating that a universal prevention/promotion intervention may improve depressive symptoms, although the actual effect range suggests it may have little or no difference (MD ‐1.25, 95% CI ‐3.28 to 0.78; I2 = NA; P = 0.23) (Analysis 1.6). At 7 to 24 months, two studies with 1207 participants (Jewkes 2008; Yusoff 2015) identified a positive effect of universal prevention/promotion interventions over control (SMD ‐0.13, 95% CI ‐0.24 to ‐0.02; I2 = 0%; P = 0.02) (Analysis 1.7).
For anxiety symptoms at 0 to 1 month post‐intervention, one study with 158 participants (Hendriks 2019) found a probable difference in favour of the intervention versus control (MD ‐0.14, 95% CI ‐0.27 to ‐0.01; I2 = NA; P = 0.04; moderate certainty due to imprecision) (Analysis 1.8). At 1 to 6 and 7 to 24 months, no studies reported on this outcome.
For distress/PTSD symptoms, four studies with 722 participants (Baker‐Henningham 2019; Bell 2008; Duan 2019; Hendriks 2019) identified that interventions may have a slightly beneficial effect at 0 to 1 month post‐intervention (SMD ‐0.24, 95% CI ‐0.41 to ‐0.08; I2 = 8 %; P = 0.04; 722 participants; low certainty due to study limitations and indirectness) (Analysis 1.9). At 1 to 6 and 7 to 24 months, no studies reported on this outcome.
4. Social outcomes (e.g. perception of social inclusion)
One study with 158 participants (Hendriks 2019) identified a positive effect of interventions versus control condition in improving social outcomes at 0 to 1 month post‐intervention (MD ‐0.45, 95% CI ‐0.82 to ‐0.08; I2 = NA; P = 0.02; 158 participants). At 1 to 6 and 7 to 24 months, no studies reported on this outcome.
5. Carer mental health
No studies reported on this outcome.
Comparison 2. Primary‐level health worker and/or community worker‐led selective prevention interventions for adults (n = 20 studies)
We identified 20 RCTs (8 cluster‐RCTs (40%) and 12 individual RCTs (60%)) contributing to meta‐analyses in this comparison. The total sample size was 4160 (2087 participants (50.2%) for the intervention group and 2073 participants (49.8%) for the control group).
As for the type of PWs, in nine studies (45%), interventions were delivered by community workers (CWs) whereas in 10 studies (55%) by primary health workers (PWs), and in one study it was a self‐help intervention (see Appendix 2 for a complete list of studies regarding the classification of PWs). CWs included, for instance, community health aiders, trained yoga instructors, family volunteers, peer counsellors, and local groups; PWs included physiotherapists, nursing assistants, healthcare assistants, antenatal care staff team, and midwives. For most of the studies, participants in the control group were allocated to usual care (n = 18 studies, 90%); for the remaining two studies (10%), the control group was guided by a PW without training and/or supervision or was allocated to a waiting list. Six studies (30%) had been conducted in PC facilities (hospitals) and two (10%) in PC facilities (freestanding), eight (40%) at home, three (15%) in a community setting, two (10%) in other settings (yoga university and offices). From a wider geographical perspective, 13 studies (65%) were conducted in an urban context, four (20%) in a rural one, two (10%) both in an urban and rural context, and one in a slum. Ten studies (50%) were undertaken in upper‐middle‐income countries, four (20%) in low‐middle income countries, three (15%) in low‐income countries and, for one study (5%), income level changed from low‐middle to upper‐middle whereas for the remaining two studies (10%), income level changed from low to low‐middle. As for the condition, most studies addressed psychological distress (n = 12 studies, 60%) and eight studies (40%) targeted the prevention of perinatal mental disorders.
Primary outcomes
1. Diagnosis (or a proxy thereof, as assessed by scoring above a cut‐off for a screening tool) of mental disorders
At 0 to 1 month post‐intervention
No studies reported on this outcome.
At 1 to 6 months post‐intervention
No studies reported on this outcome.
At 7 to 24 months post‐intervention
One study with 349 participants (Tripathy 2010) identified an uncertain effect of selective prevention interventions at 7 to 24 months post‐intervention (RR 1.81, 95% CI 0.17 to 19.82; I2 = NA; P = 0.63) (Analysis 2.1).
2. Quality of life
At 0 to 1 month post‐intervention
Three studies (McCann 2015; Miller 2020; Rahimi 2021a) (229 participants) showed that it is uncertain whether selective prevention interventions were more effective than control in improving quality of life at 0 to 1 month post‐intervention (SMD ‐1.64, 95% CI ‐2.97 to ‐0.31; 229 participants, I2 = 92%; P = 0.02; very low‐certainty evidence due to inconsistency, indirectness, and publication bias) (Analysis 2.2).
At 1 to 6 months post‐intervention
Five studies (Hamdani 2021b; McCann 2015; Ozcan 2020; Rachasrimuang 2018; Yang 2022) (798 participants) provided data on this outcome. We identified a positive effect of selective prevention interventions versus control group in improving quality of life (SMD ‐1.05, 95% CI ‐1.84 to ‐0.26; I2 = 96%; P = 0.009; 798 participants) (Analysis 2.3).
At 7 to 24 months post‐intervention
No studies reported on this outcome.
3. Adverse events experienced during the intervention
No studies reported on this outcome.
Secondary outcomes
1. Psychological functioning and impairment
No studies reported on this outcome.
2. Changes in service utilization and contact coverage
No studies reported on this outcome.
3. Changes in mental health symptoms captured on rating scales (i.e. depressive symptoms, anxiety symptoms, distress/PTSD symptoms)
(Analysis 2.4; Analysis 2.5; Analysis 2.6; Analysis 2.7; Analysis 2.8; Analysis 2.9)
For depressive symptoms (Analysis 2.4) at 0 to 1 month post‐intervention, we collected four studies with 223 participants (Baker‐Henningham 2005; Hirani 2010; Ramezani 2017; Sanfilippo 2020). The studies probably identified an important effect of selective preventive interventions versus control in reducing depressive symptoms (SMD ‐0.69, 95% CI ‐1.08 to ‐0.30; I2 = 39%; P = 0.0005; 223 participants; moderate certainty due to study limitations). At 1 to 6 months (Analysis 2.5), three studies (Barnes 2019; Rachasrimuang 2018; Sanfilippo 2020) provided data in favour of selective prevention interventions (SMD ‐0.60, 95% CI ‐1.00 to ‐0.21; I2 = 32%; P = 0.002; 186 participants). At 7 to 24 months, no studies reported on this outcome.
For anxiety symptoms at 1 to 6 months (Analysis 2.6), we collected one study with 2026 participants (Langer 1996) showing that selective prevention interventions may lead to decreased symptoms, although the actual effect range indicates it may have little or no difference (MD ‐0.80, 95% CI ‐1.87 to 0.27; I2 = NA; P = 0.14). At endpoint and 7 to 24 months, no studies reported on this outcome.
For distress/PTSD symptoms at 0 to 1 month post‐intervention (Analysis 2.7), we collected seven studies (Cerquera Córdoba 2021; Chattha 2008; Li 2019; McCann 2015; Miller 2020; Ramezani 2017; Sanfilippo 2020) and found that selective prevention interventions may have a positive effect over controls but the evidence is very uncertain (SMD ‐0.90, 95% CI ‐1.44 to ‐0.36; I2 = 88%; P = 0.001; 535 participants; very low‐certainty evidence due to inconsistency and indirectness). At 1 to 6 months (Analysis 2.8), we collected five studies (Dias 2008; McCann 2015; Rahman 2009; Sanfilippo 2020; Yang 2022) and found a difference in favour of selective prevention interventions (SMD ‐0.67, 95% CI ‐1.21 to ‐0.12; 464 participants, I2 = 86%; P = 0.02). At 7 to 24 months, we did not identify any differences between study arms (Analysis 2.9).
4. Social outcomes (e.g. perception of social inclusion)
(Analysis 2.10; Analysis 2.11; Analysis 2.12)
At 0 to 1 month post‐intervention, we found three studies (Cerquera Córdoba 2021; Rahimi 2021a; Vargas‐Porras 2021). Selective prevention interventions improved social outcomes compared with the control group (SMD ‐9.50, 95% CI ‐15.29 to ‐3.70; I2 = 81%; P = 0.001; 121 participants) (Analysis 2.10). At 1 to 6 months, selective prevention interventions may lead to decrease symptoms, although the actual effect range indicates it may have little or no difference (SMD ‐0.88, 95% CI ‐2.56 to 0.80; 2 studies, 236 participants; I2 = NA; P = 0.30) (Analysis 2.11). At 7 to 24 months, one study (Cerquera Córdoba 2021) with 27 participants showed a positive effect of selective prevention over control in improving social outcomes (MD ‐15.70, 95% CI ‐28.35 to ‐3.05; I2 = NA; P = 0.02; 27 participants) (Analysis 2.12).
5. Resource use
No studies reported on this outcome.
6. Carer mental health
No studies reported on this outcome.
Comparison 3. Primary‐level health worker and/or community worker‐led indicated prevention interventions for adults (n = 42 studies)
We identified 42 RCTs (11 cluster‐RCTs (26.2%) and 31 individual RCTs (73.8%)) contributing to meta‐analyses in this comparison. Total sample size was 6214 (3100 participants (49.9%) for the intervention group and 3114 participants (50.1%) for the control group).
As for the type of PWs, in 25 studies (59.5%), interventions were delivered by community workers (CWs) whereas in 18 studies (42.9%) interventions were delivered by primary health workers (PWs) (see Appendix 2 for a complete list of studies regarding the classification of PWs). CWs included, for instance, lay community workers, group leaders, yoga instructors, teachers, and peer counsellors; PWs included midwives, nurses, physicians, and trained general practitioners. For most of the studies, participants in the control group were allocated to usual care (n = 35 studies, 83.3%); for one study (2.4%), the control group was guided by a PW without training and/or supervision and for the remaining six studies (14.3%), the control group was active control or waiting list. Nine studies (21.4%) had been conducted in a hospital and 6 (14.3%) in a freestanding PC facility, 13 (31%) in a community setting, seven (16.7%) at home, two (4.8%) in schools, one at the workplace and the remaining four (9.5%) in other settings, such as churches, mixed settings (homes, social centres). From a wider geographical perspective, 26 studies (61.9%) were conducted in an urban context, seven (16.7%) in a rural one, five (11.9%) both in urban and rural contexts, and for the remaining four studies (9.5%) it was not specified. Twenty‐five studies (59.5%) were undertaken in upper‐middle income countries, 11 (26.2%) in low‐middle income countries, two (4.8%) in middle‐income countries, two (4.8%) in low‐income countries, and for two studies (4.8%) income level changed from low‐middle to upper‐middle. As for the condition, the majority of studies (n = 32, 76.2%) targeted the prevention of common mental disorders, nine studies (21.4%) addressed perinatal mental disorders and the remaining one (2.4%) targeted psychological distress.
Primary outcomes
1. Diagnosis (or a proxy thereof, as assessed by scoring above a cut‐off for a screening tool) of mental disorders
(Analysis 3.1; Analysis 3.2; Analysis 3.3)
At 0 to 1 month post‐intervention
At 0 to 1 month post‐intervention, we found three studies with 843 participants. It is uncertain whether indicated prevention interventions reduce the incidence of mental health diagnosis (RR 0.30, 95% CI 0.06 to 1.57; I2 = 87%; P = 0.15; 3 studies, 843 participants; very low‐certainty evidence due to study limitations, inconsistency, indirectness, and imprecision) (Analysis 3.1).
At 1 to 6 months post‐intervention
At 1 to 6 months, we found six studies with 1352 participants (Acarturk 2022; Cooper 2009; Rong 2021b; Rotheram‐Borus 2014a; Rotheram‐Borus 2014b; Skar 2021). We found a difference in favour of interventions over controls (RR 0.65, 95% CI 0.50 to 0.84; I2 = 21%; P = 0.001; 1352 participants) (Analysis 3.2).
At 7 to 24 months post‐intervention
At 7 to 24 months, we found two studies with 380 participants. Indicated prevention interventions may lead to a decrease in the incidence of mental disorders at 7 to 24 months, although the actual effect range indicates it may have little or no difference (RR 0.69, 95% CI 0.41 to 1.19; I2 = 0%; P = 0.18) (Analysis 3.3).
2. Quality of life
(Analysis 3.4; Analysis 3.5; Analysis 3.6)
At 0 to 1 month post‐intervention
At 0 to 1 month post‐intervention, we found eight studies (Acarturk 2022; Bernardi 2020; Cheng 2021; Chomat 2019; Dybdahl 2001; Escolar 2014; Novelli 2018; Song 2019). Indicated prevention interventions may improve quality of life but the evidence is very uncertain (SMD ‐0.36, 95% CI ‐0.61 to ‐0.12; I2 = 68%; P = 0.004; 1136 participants; very low‐certainty evidence due to inconsistency, publication bias, and indirectness) (Analysis 3.4).
At 1 to 6 months post‐intervention
At 1 to 6 months, indicated prevention interventions may have little or no effect on improving quality of life (SMD ‐0.04, 95% CI ‐0.23 to 0.16; I2 = 34%; P = 0.70; 4 studies, 847 participants) (Analysis 3.5).
At 7 to 24 months post‐intervention
At 7 to 24 months, indicated prevention interventions may lead to a decrease in the incidence of mental disorders, although the actual effect range indicates it may have little or no difference (SMD ‐0.80, 95% CI ‐3.53 to 1.93; I2 = NA; P = 0.57; 1 study, 94 participants) (Analysis 3.6).
3. Adverse events experienced during the intervention
At 0 to 1 month post‐intervention, only Acarturk 2022 reported that no participants experienced adverse events in the intervention and control groups.
Secondary outcomes
1. Psychological functioning and impairment
(Analysis 3.8; Analysis 3.9; Analysis 3.10)
Indicated prevention interventions probably make little or no difference at 0 to 1 month post‐intervention (SMD ‐0.12, 95% CI ‐0.39 to 0.15; I2 = 38%; P = 0.39; 4 studies, 663 participants; moderate certainty due to indirectness) (Analysis 3.8). At 1 to 6 months (Analysis 3.9), the actual effect range indicates it may make little or no difference (SMD ‐0.10, 95% CI ‐0.60 to 0.41; I2 = 77%; P = 0.71; 2 studies, 594 participants), as well as at 7 to 24 months (SMD ‐0.21, 95% CI ‐0.47 to 0.04; I2 = 38%; P = 0.10; 2 studies, 241 participants) (Analysis 3.10).
2. Changes in service utilization and contact coverage
No studies reported on this outcome.
3. Changes in mental health symptoms captured on rating scales (i.e. depressive symptoms, anxiety symptoms, distress/PTSD symptoms)
(Analysis 3.11; Analysis 3.12; Analysis 3.13; Analysis 3.17; Analysis 3.18; Analysis 3.19; Analysis 3.14; Analysis 3.15; Analysis 3.16; Analysis 3.17; Analysis 3.18; Analysis 3.19)
At 0 to 1 month post‐intervention, we found 18 studies evaluating depressive symptoms. Indicated prevention interventions may have little to no effect in reducing depressive symptoms compared to control groups but the evidence is very uncertain (SMD ‐0.16, 95% CI ‐0.30 to ‐0.03; I2 = 56%; P = 0.02; 2341 participants; very low‐certainty evidence due to inconsistency, publication bias, and study limitations) (Analysis 3.11). At 1 to 6 months, we collected 11 studies confirming that the effect of the intervention was maintained (SMD ‐0.34, 95% CI ‐0.58 to ‐0.10; I2 = 88%; P = 0.005; 2609 participants) (Analysis 3.12). At 7 to 24 months, we found eight studies that suggested that indicated prevention interventions may lead to slightly decreased depressive symptoms, although the actual effect range indicates it may have little or no difference (SMD ‐0.10, 95% CI ‐0.22 to 0.01; I2 = 39%; P = 0.08; 2149 participants) (Analysis 3.13).
For anxiety symptoms at 0 to 1 month post‐intervention, the collected five studies (Bernardi 2020; Ferreira‐Vorkapic 2018; Hajarian Abhari 2021; Rao 2017; Rodriguez 2021) identified that indicated prevention interventions may have a positive effect over controls but the evidence is very uncertain (SMD ‐1.19, 95% CI ‐2.02 to ‐0.35; I2 = 89%; P = 0.005; 250 participants; very low certainty due to study limitations and inconsistency) (Analysis 3.14). At 1 to 6 months, we collected four studies (Dayhimi 2020; Rong 2021a; Srisuwan 2020; Xu 2021) suggesting that the beneficial effect of the intervention was maintained (SMD ‐0.23, 95% CI ‐0.37 to ‐0.09; I2 = 84%; P = 0.002; 771 participants) (Analysis 3.15). At 7 to 24 months, indicated prevention interventions may slightly decrease anxiety symptoms, although the actual effect range indicates the effect of the intervention may have little or no difference (SMD ‐0.12, 95% CI ‐0.29 to 0.05; I2 = 92%; P = 0.17; 2 studies, 549 participants) (Analysis 3.16).
For distress/PTSD symptoms at 0 to 1 month post‐intervention, indicated prevention interventions may have a positive effect but the evidence is very uncertain (SMD ‐0.54, 95% CI ‐0.95 to ‐0.14; I2 = 95%; P = 0.009; 19 studies, 2536 participants; very low certainty due to inconsistency and publication bias) (Analysis 3.17). At 1 to 6 months, the effect of the intervention was maintained (SMD ‐0.29, 95% CI ‐0.51 to ‐0.07; I2 = 75%; P = 0.008; 9 studies, 1702 participants) (Analysis 3.18). At 7 to 24 months, the actual effect range indicates the effect of the intervention may have little or no difference (SMD ‐0.04, 95% CI ‐0.45 to 0.38; I2 = 91%; P = 0.86; 5 studies, 1081 participants) (Analysis 3.19).
4. Social outcomes (e.g. perception of social inclusion)
(Analysis 3.20; Analysis 3.21)
At 0 to 1 month and at 7 to 24 months post‐intervention, the actual effect range indicates the effect of the intervention may make little or no difference on social outcomes. At 1 to 6 months, no studies reported on this outcome.
5. Resource use
No studies reported on this outcome.
6. Carer mental health
No studies reported on this outcome.
Comparison 4. Primary‐level health worker and/or community worker‐led universal prevention or promotion interventions for children (n = 8 studies)
We identified eight RCTs (six cluster‐RCTs (75%) and two individual RCTs (25%)) contributing to meta‐analyses in this comparison. Total sample size was 1496 (760 participants (50.8%) for the intervention group and 736 participants (49.2%) for the control group).
As for the type of PWs, all interventions (n = 8 studies, 100%) were delivered by community workers (CWs) but, in one study (12.5%), the intervention was also delivered by primary health workers (PWs) (see Appendix 2 for a complete list of studies regarding the classification of PWs). CWs included, for instance, school teachers, yoga teachers, staff, and student facilitators, whereas PWs included midwives. For most of the studies, participants in the control group were allocated to usual care (n = 5 studies, 62.5%); for one study (12.5%) the control group was guided by a PW without training and/or supervision and, for the remaining two studies (25%), the control group was active control or waiting list. All eight studies (100%) had been conducted in schools. From a wider geographical perspective, seven studies (87.5%) were conducted in an urban context and one in a rural context. Five studies (62.5%) were undertaken in upper‐middle‐income countries and three (37.5%) in low‐income countries. As for the condition, most included studies (n = 5, 62.5%) targeted the prevention of child mental disorders and three studies (37.5%) aimed at improving psychological well‐being.
Primary outcomes
1. Diagnosis (or a proxy thereof, as assessed by scoring above a cut‐off for a screening tool) of mental disorders
At 0 to 1 month post‐intervention
No studies reported on this outcome.
At 1 to 6 months post‐intervention
No studies reported on this outcome.
At 7 to 24 months post‐intervention
No studies reported on this outcome.
2. Quality of life
At 0 to 1 month post‐intervention
At 0 to 1 month post‐intervention, we found two studies (Barbosa Filho 2017; Devries 2015) showing that universal prevention/promotion interventions may show a positive effect of the interventions versus control in improving the quality of life (SMD ‐0.25, 95% CI ‐0.39 to ‐0.11; I2 = 0%; P = 0.0003; 2 studies, 803 participants; low certainty due to study limitations and indirectness) (Analysis 4.1).
At 1 to 6 months post‐intervention
No studies reported on this outcome.
At 7 to 24 months post‐intervention
No studies reported on this outcome.
3. Adverse events experienced during the intervention
One study (Devries 2015) provided information on this outcome at 0 to 1 month reporting that no participants experienced adverse events in either study arm.
Secondary outcomes
1. Psychological functioning and impairment
(Analysis 4.3; Analysis 4.4; Analysis 4.5)
At 0 to 1 month post‐intervention (Analysis 4.3), it is uncertain whether universal prevention/promotion interventions improve psychological functioning (SMD 0.04, 95% CI ‐0.90 to 0.98; I2 = 82%; P = 0.93; 2 studies, 212 participants; low certainty due to study limitations, inconsistency, indirectness and imprecision). At 1 to 6 months follow‐up (Analysis 4.4), universal prevention/promotion interventions may lead to increased psychological functioning, although the actual effect range indicates it may have little or no difference (MD ‐0.29, 95% CI ‐0.93 to 0.35; I2 = NA; P = 0.37; 1 study, 90 participants). At 7 to 24 months (Analysis 4.5), one study (Berger 2018) identified a difference in favour of the intervention over the control condition (MD ‐3.33, 95% CI ‐5.03 to ‐2.63; I2 = NA; P = 0.0001; 183 participants).
2. Changes in service utilization and contact coverage
No studies reported on this outcome.
3. Changes in mental health symptoms captured on rating scales (i.e. depressive symptoms, anxiety symptoms, distress/PTSD symptoms)
(Analysis 4.6; Analysis 4.7; Analysis 4.8; Analysis 4.9; Analysis 4.10; Analysis 4.11; Analysis 4.12)
For depressive symptoms, we found one study (Rivet‐Duval 2011) reporting that interventions may improve symptoms at 0 to 1 month post‐intervention (SMD ‐3.04, 95% CI ‐6.00 to ‐0.08; I2 = NA%; P = 0.04; 160 participants; low certainty due to study limitations and imprecision) (Analysis 4.6), but the effect was not maintained at 1 to 6 months (SMD ‐0.00, 95% CI ‐0.20 to 0.20; I2 = 0%; P = 0.98; 3 studies, 385 participants) (Analysis 4.7). At 7 to 24 months, no studies reported on this outcome.
For anxiety symptoms at 0 to 1 month post‐intervention, we found one study (Berger 2018) reporting that universal prevention/promotion interventions may reduce anxiety symptoms (MD ‐2.27, 95% CI ‐3.13 to ‐1.41; I2 = NA; P < 0.00001; 183 participants; low certainty due to study limitations and imprecision) (Analysis 4.8). At 1 to 6 months, the difference was not important (MD ‐0.13, 95% CI ‐0.41 to 0.15; I2 = NA; P = 0.37; 1 study, 125 participants) (Analysis 4.9) while, at 7 to 24 months, the study of Berger 2018 confirmed a positive effect of the intervention (SMD ‐2.27, 95% CI ‐3.10 to ‐1.44; I2 = NA; P < 0.00001; 183 participants) (Analysis 4.10).
For distress/PTSD symptoms at 0 to 1 month post‐intervention, it is uncertain whether interventions improve distress (SMD ‐0.83, 95% CI ‐2.48 to 0.82; I2 = 98%; P = 0.33; 2 studies, 800 participants) (Analysis 4.11) while, at 1 to 6 months, we found one study (Mohamadi 2021) reporting that universal prevention/promotion interventions improved symptoms of distress/PTSD (MD ‐4.51, 95% CI ‐5.86 to ‐3.16; I2 = NA; P < 0.00001; 106 participants) (Analysis 4.12). No data were available on this outcome at 7 to 24 months follow‐up.
4. Social outcomes (e.g. perception of social inclusion)
(Analysis 4.13; Analysis 4.14; Analysis 4.15)
At 0 to 1 month post‐intervention (Analysis 4.13), universal prevention/promotion interventions may lead to increased psychological functioning, although the actual effect range indicates it may have little or no difference (SMD ‐0.32, 95% CI ‐0.76 to 0.12; I2 = 67%; P = 0.15; 3 studies, 321 participants) and at 1 to 6 months (SMD ‐0.27, 95% CI ‐0.82 to 0.28; I2 = 75%; P = 0.34; 2 studies, 215 participants) (Analysis 4.14). At 7 to 24 months, we collected one study (Berger 2018) reporting a positive effect of the intervention over the control group (MD ‐0.70, 95% CI ‐1.07 to ‐0.33; I2 = NA; P = 0.0002; 183 participants) (Analysis 4.15).
5. Resource use
No studies reported data on this outcome.
6. Carer mental health
No studies reported data on this outcome.
Comparison 5. Primary‐level health worker and/or community worker‐led selective prevention interventions for children (n = 7 studies)
We identified seven RCTs (four cluster‐RCTs (57.1%) and three individual RCTs (42.9%)) contributing to meta‐analyses in this comparison. The total sample size was 1572 (800 participants (50.9%) for the intervention group and 772 participants (49.1%) for the control group).
As for the type of PWs, most of the interventions (n = 6 studies, 85.7%) were delivered by community workers (CWs) whereas, in only one study (14.3%), the intervention was delivered by primary health workers (PWs) (see Appendix 2 for a complete list of studies regarding the classification of PWs). CWs included, for instance, teachers, facilitators, and coaches, whereas PWs included nursing students. For most of the studies, participants in the control group were allocated to usual care (n = 6 studies, 85.7%); for the remaining study (14.3%), the control group was a waiting‐list condition. Four studies (57.1%) had been conducted in schools, one (14.3%) in hospitals, and the remaining two (28.6%) in other settings, such as local churches and multiple community spaces. From a wider geographical perspective, three studies (42.9%) were conducted in an urban context, one (14.3%) in a rural context, one (14.3%) in both urban and rural contexts and, for the remaining two studies (28.6%), it was not specified. One study (14.3%) was undertaken in an upper‐middle income country, one (14.3%) in a low‐middle income country, and the remaining five studies (71.4%) in low‐income countries. As for the condition, the majority of studies (n = 6 studies, 85.7%) targeted the prevention of child mental disorders and one study (14.3%) aimed at improving psychological well‐being.
Primary outcomes
1. Diagnosis (or a proxy thereof, as assessed by scoring above a cut‐off for a screening tool) of mental disorders
At 0 to 1 month post‐intervention
No studies reported on this outcome.
At 1 to 6 months post‐intervention
No studies reported on this outcome.
At 7 to 24 months post‐intervention
No studies reported on this outcome.
2. Quality of life
At 0 to 1 month post‐intervention
At 0 to 1 month post‐intervention, we found one study (Ager 2011) indicating that it is uncertain whether selective prevention interventions improve quality of life (MD ‐1.10, 95% CI ‐3.32 to 1.12; I2 = NA; P = 0.33; 115 participants; very low certainty due to study limitations and imprecision) (Analysis 5.1).
At 1 to 6 months post‐intervention
No studies reported on this outcome.
At 7 to 24 months post‐intervention
No studies reported on this outcome.
3. Adverse events experienced during the intervention
No studies reported data on this outcome.
Secondary outcomes
1. Psychological functioning and impairment
At 0 to 1 month post‐intervention, there is no evidence that selective prevention interventions improves functional impairment (MD ‐0.02, 95% CI ‐0.09 to 0.05; I2 = NA; P = 0.57; 1 study, 479 participants; low certainty due to indirectness and imprecision). At 1 to 6 and 7 to 24 months, no studies reported on this outcome.
2. Changes in service utilization and contact coverage
No studies reported on this outcome.
3. Changes in mental health symptoms captured on rating scales (i.e. depressive symptoms, anxiety symptoms, distress/PTSD symptoms)
(Analysis 5.3; Analysis 5.4; Analysis 5.5; Analysis 5.6; Analysis 5.7; Analysis 5.8; Analysis 5.9)
For depressive symptoms, selective prevention interventions probably make little or no difference to symptoms at 0 to 1 month post‐intervention (SMD 0.00, 95% CI ‐0.16 to 0.15; I2 = 0%; P = 0.96; 2 studies, 638 participants; moderate certainty due to imprecision) (Analysis 5.3). We did not identify any important difference between the intervention and control group at 1 to 6 months and 7 to 24 months (Analysis 5.4; Analysis 5.5).
For anxiety symptoms, selective prevention interventions may make little or no difference to symptoms at 0 to 1 month post‐intervention (MD 4.50, 95% CI ‐12.05 to 21.05; I2 = NA; P = 0.59; 1 study, 28 participants; low certainty due to imprecision) (Analysis 5.6). At 1 to 6 months, one study reported a worsening of anxiety symptoms (MD 1.42, 95% CI ‐0.00 to 2.84; I2 = NA; P = 0.05; 1 study, 143 participants) (Analysis 5.7). At 7 to 24 months, no studies provided information on this outcome.
For distress/PTSD symptoms at 0 to 1 month post‐intervention, we found one study (O’Callaghan 2014) reporting that interventions probably improve symptoms of distress/PTSD (MD ‐2.14, 95% CI ‐3.77 to ‐0.51; I2 = NA; P = 0.01; 159 participants; moderate certainty due to imprecision) (Analysis 5.8). Selective prevention interventions have an uncertain effect at 1 to 6 months (SMD ‐0.40, 95% CI ‐2.49 to 1.69; I2 = NA; P = 0.71; 1 study, 213 participants) (Analysis 5.9). No data were available on this outcome at 7 to 24 months follow‐up.
4. Social outcomes (e.g. perception of social inclusion)
At 0 to 1 month post‐intervention, we did not find any differences between the intervention and control group. No data were available for the other time points.
5. Resource use
No studies reported data on this outcome.
6. Carer mental health
No studies reported data on this outcome.
Comparison 6. Primary‐level health worker and/or community worker‐led indicated prevention interventions for children (n = 9 studies)
We identified nine RCTs (five cluster‐RCTs (55.6%) and four individual RCTs (44.4%)) contributing to meta‐analyses in this comparison. The total sample size was 3029 (1684 participants for the intervention group and 1345 participants for the control group).
As for the type of PWs, most of the interventions (n = 7 studies, 77.8%) were delivered by community workers (CWs) whereas, in one study (11.1%), the intervention was delivered by primary health workers (PHWs) and, in another study (11.1%) there was a self‐help intervention (see Appendix 2 for a complete list of studies regarding the classification of PWs). CWs included, for instance, adult refugee facilitators, teachers, paraprofessional interventionists, lay counsellors, whereas PWs included social workers. For most of the studies, participants in the control group were allocated to usual care (n = 8 studies, 88.9%); for the remaining study (11.1%), the control group was allocated to active control. Eight studies (88.9%) had been conducted in schools and one (11.1%) in a refugee camp. From a wider geographical perspective, one study (11.1%) was conducted in an urban context, three (33.3%) in a rural context, one (11.1%) in both urban and rural contexts and, for the remaining four studies (44.4%), it was not specified. Three studies (33.3%) were undertaken in upper‐middle income countries, four (44.4%) in low‐middle income countries and two studies (22.2%) in low‐income countries. As for the condition, all included studies (n = 9 studies, 100%) targeted the prevention of child mental disorders.
Primary outcomes
1. Diagnosis (or a proxy thereof, as assessed by scoring above a cut‐off for a screening tool) of mental disorders
At 0 to 1 month post‐intervention
At 0 to 1 month post‐intervention, we found one study (Yu 2002) with 220 participants. It is uncertain whether indicated prevention interventions reduce the incidence of mental health diagnosis (RR 0.77, 95% CI 0.51 to 1.17; I2 = NA; P = 0.22; 1 study, 220 participants; very low certainty due to study limitations indirectness and imprecision) (Analysis 6.1).
At 1 to 6 months post‐intervention
At 1 to 6 months, we found one study (Yu 2002) that indicated prevention interventions may lead to decreased incidence, although the actual effect range indicates the effect of the intervention may have little or no difference (RR 0.77, 95% CI 0.51 to 1.17; I2 = NA; P = 0.19; 220 participants) (Analysis 6.2).
At 7 to 24 months post‐intervention
No studies reported on this outcome.
2. Quality of life
At 0 to 1 month post‐intervention
It is uncertain whether indicated prevention interventions improve quality of life at 0 to 1 month post‐intervention (SMD ‐0.65, 95% CI ‐2.09 to 0.79; I2 = 0%; P = 0.38; 2 studies, 152 participants; very low certainty due to study limitations, indirectness, and imprecision) (Analysis 6.3).
At 1 to 6 months post‐intervention
No studies reported on this outcome.
At 7 to 24 months post‐intervention
No studies reported on this outcome.
3. Adverse events experienced during the intervention
No studies reported data on this outcome.
Secondary outcomes
1. Psychological functioning and impairment
At 0 to 1 month post‐intervention, we collected two studies (Fine 2021; Tol 2012) reporting a difference in favour of the intervention over the control group (SMD ‐0.29, 95% CI ‐0.47 to ‐0.10; I2 = 0%; P = 0.003; 448 participants; high certainty) (Analysis 6.4). At 1 to 6 months, we failed to identify any difference between intervention and control groups (Analysis 6.5). At 7 to 24 months, no studies reported on this outcome.
2. Changes in service utilization and contact coverage
No studies reported on this outcome.
3. Changes in mental health symptoms captured on rating scales (i.e. depressive symptoms, anxiety symptoms, distress/PTSD symptoms)
(Analysis 6.6; Analysis 6.7; Analysis 6.8; Analysis 6.9; Analysis 6.10; Analysis 6.11; Analysis 6.12)
For depressive symptoms, we found four studies (Fine 2021; Osborn 2020; Tol 2012; Yu 2002) reporting a beneficial effect of the intervention over control at 0 to 1 month post‐intervention (SMD ‐0.18, 95% CI ‐0.32 to ‐0.04; I2 = 0; P = 0.01; 771 participants; high certainty) (Analysis 6.6). At 1 to 6 months, this effect was not maintained while, at 7 to 24 months, we found one study by Shinde 2018 that indicated prevention interventions were more effective than controls in reducing depressive symptoms (MD ‐1.27 95% CI ‐1.90 to ‐0.64; I2 = NA; P < 0.0001; 904 participants) (Analysis 6.8).
For anxiety symptoms, it is uncertain whether indicated prevention interventions improve symptoms at 0 to 1 month post‐intervention (SMD ‐0.09, 95% CI ‐0.22 to 0.04; I2 = 0%; P = 0.19; 3 studies, 888 participants; very low certainty due to study limitations and indirectness) (Analysis 6.9). At 1 to 6 months follow‐up, we did not find any important difference between intervention and control groups (Analysis 6.10). At 7 to 24 months, no studies reported data on this outcome.
For distress/PTSD symptoms at 0 to 1 month post‐intervention, prevention interventions may make little or no difference to symptoms (SMD 0.24, 95% CI ‐1.28 to 1.76; I2 = 0%; P = 0.76; 2 studies, 448 participants; low certainty due to inconsistency and imprecision) (Analysis 6.11). At 1 to 6 months, we failed to identify any important difference between intervention and control groups (Analysis 6.12). No study provided data at 7 to 24 months.
4. Social outcomes (e.g. perception of social inclusion)
(Analysis 6.13; Analysis 6.14)
At 0 to 1 month post‐intervention, we failed to identify any difference between intervention and control groups (Analysis 6.13) while at 1 to 6 months, we collected two studies (Jordans 2010; Thurman 2017) that indicated prevention interventions were more effective than control groups in improving social outcomes (SMD ‐0.22, 95% CI ‐0.41 to ‐0.03; I2 = 0%; P = 0.02; 421 participants) (Analysis 6.14). No study provided data at 7 to 24 months.
5. Resource use
No studies reported data on this outcome.
6. Carer mental health
No studies reported data on this outcome.
Subgroup and sensitivity analyses
We presented all the meta‐analyses according to the type of prevention intervention (universal, selective, indicated).
Due to the low number of studies available for each type of prevention intervention, we were not able to perform all the planned subgroup and sensitivity analyses.
-
Category of health worker (e.g. primary care professionals, non‐professional health workers, community workers). This analysis was not performed as most interventions were delivered by community and para‐professional workers.
-
Setting of care.
-
Community settings, camps, schools: this analysis was not performed given the small number of RCTs for each category.
-
Chronic or acute humanitarian versus non‐humanitarian settings: the number of RCTs conducted in humanitarian crises was very low compared to those in LMICs not exposed to the acute phase of a humanitarian crisis.
-
-
Type of promotion intervention (individual, group): it was not possible to conduct this analysis as most of the interventions were delivered in a group setting.
-
Specific risk, protective, or promotive factor targeted; RCT reports did not provide specific information on the targeted risk, protective, or promotive factors to conduct this analysis.
Subgroup analyses
All the analyses were conducted according to the type of prevention interventions and, for this reason, no subgroup analyses were needed on this variable. We were able to conduct subgroup analyses considering the category of health workers and the setting of care. These analyses did not identify any differences across subgroups in relation to the assessed outcomes (P > 0.05). We were not able to conduct subgroup analyses considering chronic or acute humanitarian versus non‐humanitarian settings, the type of promotion intervention in terms of individual versus group modality, and the risk, protective, or promotive factor targeted. No sufficient information was reported in the included studies to clearly identify the variables listed above.
Sensitivity analyses
We conducted sensitivity analyses including only studies with low risk of bias, as specified in incomplete outcome data and selective reporting. The effect of indicated prevention interventions for adults on psychological functioning and impairment was statistically significant at 0 to 1 month post‐intervention after excluding studies with high risk of bias and those having some concerns. For mental health symptoms, both for depressive and PTSD symptoms, the effect of indicated prevention interventions was no longer statistically significant at 0 to 1 month post‐intervention, after excluding studies with high risk of bias and those having some concerns.
We also conducted analyses excluding trials with methodological characteristics that might generate the highest heterogeneity in meta‐analysis (I² > 75%). The effect of indicated prevention interventions on the diagnosis of mental disorders was statistically significant after excluding studies generating the highest heterogeneity, and I2 was lowered to 0%. For quality of life, no differences were detected at 0 to 1 month post‐intervention study endpoint, while the effect of the intervention was not statistically significant at 1 to 6 months.
We were not able to conduct sensitivity analysis considering only published studies, and sensitivity analysis including only studies measuring the incidence of mental disorders (i.e. studies in which all participants at baseline scored below defined symptom thresholds on rating scales).
Discussion
This review included 113 randomized trials evaluating the effectiveness of promotion and prevention interventions delivered through a task‐shifting approach in 39 LMICs. The majority of studies were focused on adult populations (83 RCTs), while only a minority assessed promotion and prevention interventions for children (30 RCTs).
Summary of main results
This review aimed to assess the effectiveness of promotion and prevention interventions delivered through a task‐shifting approach by primary level and community workers to the child and adolescent, and adult populations in LMICs.
With regard to primary outcomes in adults, promotion/universal prevention and selective prevention interventions were not effective in reducing the likelihood of receiving a diagnosis of mental disorders at any time points. Indicated prevention interventions had a significant effect in reducing the frequency of diagnosed mental disorders at 1 to 6 months while, at other time points, we failed to identify a significant effect. For quality of life, promotion/universal prevention interventions were effective at 7 to 24 months follow‐up, while selective and indicated prevention interventions had a beneficial effect at the study endpoint, that in the selective prevention interventions was maintained at 1 to 6 months. In respect to adverse events, the outcome was poorly reported and no inference could be drawn. With regard to secondary outcomes, we found a small‐to‐moderate effect of interventions in reducing psychological symptoms. Functioning was assessed only in indicated prevention studies without detecting any significant difference at any time point. For social outcomes, promotion/universal prevention and selective interventions were more effective than controls at the study endpoint, and the positive effect was maintained at 7 to 24 months for selective prevention interventions. Indicated prevention interventions did not provide any beneficial effect for this outcome.
In children, only studies on indicated prevention interventions provided data on the diagnosis of mental disorders and failed to identify a beneficial effect of prevention interventions over controls. For the quality of life outcome, promotion/universal prevention interventions significantly improved this outcome at the study endpoint, while selective and indicated prevention studies failed to show a beneficial effect. The outcome of adverse events was reported only in one study in the promotion/universal prevention category without showing any difference against the control condition.
With regard to secondary outcomes, we found that promotion/universal prevention interventions improved functioning at 7 to 24 months, while indicated prevention interventions had a moderate effect in improving functioning and impairment at the study endpoint. Promotion/universal prevention interventions were more effective than controls in improving symptoms of depression and anxiety at the study endpoint, and in improving anxiety and distress at follow‐up (7 to 24 months and 1 to 6 months, respectively). Selective prevention interventions provided a beneficial effect on distress symptoms at the study endpoint. Indicated prevention interventions were more effective than controls in improving depressive symptoms at the study endpoint and at 7 to 24 months while, for anxiety and distress, we failed to identify a significant effect. For social outcomes, promotion/universal prevention interventions were significantly more effective than controls at 7 to 24 months, while indicated prevention was more effective at 1 to 6 months.
Though we could not do subgroup analyses per health worker group, those represented were community and social workers, teachers, nurses, lay counsellors, and facilitators receiving specific training for intervention delivery. This task‐shifting approach allows a broader scope of interventions in primary care when specialists are unavailable, and this review confirms that intervention delivery through PWs may improve many relevant clinical outcomes, as already highlighted in previous systematic reviews on treatment interventions (Purgato 2018a; Van Ginneken 2021).
Overall, these findings indicate the benefit of specific types of prevention interventions for reducing the likelihood of being diagnosed with any mental disorders, and improving the quality of life for adults living in LMICs, but show a lack of effect for specific types of prevention interventions and over longer follow‐up. This lack of effect could be related to different factors. Firstly, to the characteristics of interventions, which were sometimes short in duration and focused not only on mental health but also on other areas (e.g. home‐care practices and health‐seeking behaviours). Secondly, to the social determinants of mental health, like exposure to ongoing and chronic adversities that participants had to manage in low‐resource settings, exposure to traumatic events, and unsecured food, water, and housing (Lund 2018). Available data failed to show a strong and beneficial effect in reducing the likelihood of being diagnosed with any mental disorders or improving the quality of life. There were only a few included studies reporting on this outcome, i.e. diagnosis of mental disorders, as the majority of promotion/universal and selective prevention interventions focused on the reduction of symptoms. Additionally and as expected, we identified heterogeneity in outcome measure tools and population groups. An in‐depth mediation analysis would help to shed light on the mechanisms of action of interventions, and the complementary appraisal of moderators would contribute to optimizing the delivery modality, and would allow matching participants with the most appropriate interventions (Cuijpers 2022; Purgato 2021b).
Based on the review results, we identified a significant gap in the scientific literature on intervention for promoting mental health or preventing mental health conditions, and also a lack of tailored task‐shifted interventions for specific conditions with proven efficacy in treatment settings. Future research trajectories might include the design and implementation of RCTs with a clear prevention design, i.e. aimed at reducing the incidence of mental disorders as a primary outcome. Only a minority of the trials included in this review were primarily designed with this goal. Finally, data suggest that research on children should be expanded as, in comparison with adults, little evidence was available to draw firm conclusions on this population.
Overall completeness and applicability of evidence
Interventions were categorized according to the characteristics of their target population group following a public mental health approach (Purgato 2020; Tol 2015a). Promotion and universal prevention interventions were categorized under the same label, as it was often complex to distinguish multi‐component interventions aimed at strengthening well‐being and improving mental health, which fall into the promotion category (Eaton 2012; Papola 2022; WHO 2014), and interventions to universally prevent mental health conditions (Ceccarelli 2022; Tol 2015a; Tol 2015b; WHO 2019). For example, promotion/universal prevention included social skills programmes for all the students attending selected grades, interventions for building self‐esteem, keeping calm, thinking resourcefully, identifying and accessing support networks, considering the perspective of others, and keeping the peace (Castillo 2019; O'Reilly 2018). Examples of selective prevention interventions included psychosocial structured activities delivered in school settings and designed for trauma recovery, interventions with physical exercises, facilitation of self‐efficacy and decision‐making skills, problem‐solving, and maintaining a balanced lifestyle. Indicated prevention interventions, delivered to those already showing some signs of disorders (i.e. psychological symptoms below cut‐offs of rating scales) without meeting the criteria for a psychiatric diagnosis (Wahlbeck 2015), included counselling based on the principles of cognitive‐behavioural therapy, mindfulness, and yoga‐inspired techniques. Some of the indicated prevention interventions included in this review belong to the range of scalable psychological interventions developed by the WHO for use in settings affected by adversity (Sijbrandij 2017; WHO 2018). Amongst these, we find the WHO Self‐Help Plus intervention, a short strategy based on Acceptance and Commitment Therapy that was tested in a large sample of adult refugees resettled in Turkey (Acarturk 2022; Epping‐Jordan 2016). Overall, all analyzed interventions aimed at enhancing protective factors (e.g. increased social skills, coping strategies to manage stress, and problem‐solving ability; building social connectedness and community relationships; and encouraging family support and positive community network) for the prevention of mental health disorders. This is in line with the promotion and prevention component of the WHO community‐based rehabilitation guidelines (WHO 2010). Almost all included interventions adopted a community‐based approach, were implemented in the life context of the participants, and were undertaken together with a significant person for the participant (i.e. caregiver, parent, partner). This explains the increased number of interventions addressing the everyday social impacts on mental health in LMICs, considering the interaction between the local social environment and psychosocial well‐being. In line with this, the majority of interventions assessed in the included RCTs were designed and delivered for preventing common mental health conditions, like depression, PTSD, and anxiety, which are the most prevalent conditions in low‐resource settings affected by humanitarian crises (Charlson 2019).
Within low‐resource settings, the task‐shifting approach has been developed and implemented for improving the efficacy and effectiveness of mental health services, starting with extending their accessibility (Hoeft 2018; Purgato 2020). The intervention gap between people in need of mental health interventions and those who actually access services is constantly increasing, thus requiring the adoption of innovative strategies, such as task‐shifting (Hoeft 2018; Mendenhall 2014; Soltan 2022). In line with this consideration, a recent systematic review with individual patient data meta‐analysis focused on psychological interventions in low‐ and middle‐income countries, and found an association between the task‐shifting approach and an increased reduction in the severity of depressive symptoms (Karyotaki 2022).
The ways in which the organization of society, social interactions, and relationships affect the risk of, and protection from, mental disorders is called 'social domain'. There is widespread global evidence that mental disorders in populations are strongly socially determined, and our findings suggest a growth focus on the 'social domain' as a crucial social determinant of mental health, as reported in the UN Sustainable Development Goals (Burgess 2020; Lund 2018; Patel 2018).
The included studies employed different types of primary workers, in line with the scientific literature on psychological and social interventions (Shahmalak 2019; Van Ginneken 2021). Some workers were existing cadres within the health sector (i.e. nurses, midwives), while others were additionally trained resources. Resources included for example teachers, peer mentors, lay counsellors, peer educators, social workers, field workers, and caregivers. Primary workers received short training before the intervention delivery and, in most cases, received supervision during the course of the study. The training was usually manualized and focused strictly on the intervention's content and delivery. While many studies presented this basic description of training and supervision procedures, many failed to report detailed information on these aspects. Nevertheless, it would be of great value for authors of RCTs to provide detailed descriptions of training contents and duration, to improve readers' understanding of the interventions (Shahmalak 2019).
Regarding the control conditions, we found that the waiting list, no treatment, and treatment‐as‐usual were the most reported comparators, despite the methodological recommendation of using active psychological control groups in psychotherapy research (Guidi 2018). The use of 'inactive' controls does not allow establishing whether any significant difference is related to some specific treatment ingredients, introduced by the experimental procedure, or to nonspecific factors, such as attention and opportunity for disclosure (Guidi 2018; Mulder 2017; Riello 2021). Regardless, inactive controls remain widely used (Barbui 2020; Faltinsen 2022). As a consequence, we were not able to compare active prevention interventions against each other, although this would be vital to increase our understanding of the mechanism of action of prevention interventions (Purgato 2019b; Purgato 2021a).
Data on children and adolescents were not available for some important outcomes or were not available at medium‐ and long‐term follow‐up. This paucity of research on children and adolescents is in line with what was highlighted in the WHO 2020 guidelines on mental health promotive and preventive interventions for adolescents, which indicated a crucial need for research on the topic, especially in low‐resource or high‐adversity settings (WHO 2020).
Promotion and prevention interventions were delivered in variable conditions of cultural contexts, which were not accounted for in our analyses. Factors that can facilitate or impede the implementation of mental health interventions vary across contexts and cultures and examination prior to, during, and after the process of implementation is relevant (Faregh 2019).
Finally, we were not able to perform all the planned subgroup and sensitivity analyses. This was due to the low number of studies available for each type of prevention intervention.
Quality of the evidence
Even though the RCT is the design of choice for evaluating the efficacy and acceptability of healthcare interventions (Jüni 2001; Purgato 2010), the level of evidence generated by our review is poor as evaluated with the Cochrane risk of bias tool 2. This is consistent with our grading within the summary of findings tables (from very low to moderate) (summary of findings Table 1; summary of findings Table 2; summary of findings Table 3; summary of findings Table 4; summary of findings Table 5; summary of findings Table 6).
This review included 113 trials, all of which were randomized and covered a wide range of interventions and settings. For studies included in the meta‐analyses, evidence for most outcomes was of low‐ to moderate‐certainty. The most often identified biases across studies in both children and adolescents and adults were risk of bias due to deviations from the intended interventions (effect of assignment to intervention and effect of adhering to intervention). This is due to the fact that participants randomized to psychological interventions typically know whether they have been randomized to the active intervention or to the control group, and the same is true for people delivering the interventions (Cuijpers 2015). These biases are likely to be especially large in studies with waiting lists or treatment‐as‐usual controls (Cuijpers 2016; Mohr 2014). Other identified biases were related to the risk of bias due to missing outcome data, and the risk of bias in the selection of the reported results, which has been associated with the risk of inflation of effect sizes, contributing to the current uncertainties in assessing the outcomes of psychological interventions (Miguel 2021). On the contrary, the domain related to the randomization process was evaluated as being at low risk of bias in the majority of studies on adults, and in all the studies focused on children and adolescents.
For the adult population, only four studies had an overall low risk of bias. The remaining studies had a high risk of bias or some concerns. Although the number of trials available on the child and adolescent population was smaller, the risk of bias identified was consistent with RCTs for adults.
Additional possible sources of bias not included in the risk of bias assessment were those specifically related to the topic of this review: socio‐cultural differences in relation to psychological suffering across countries; transposition of mental health concepts from western to non‐western cultures (Kaiser 2015), with very different understandings and ways of dealing with psychological distress; and social norms and ways of discussing distress with strangers (Barbui 2017). Moreover, researchers did not always report details on language and nationality, social/economic class, education, geography, age, and background of the people delivering the interventions. These characteristics might have an influence on the establishment of relationships and trust, and thus on some study outcomes. Finally, even though we did not explore the risk of bias related to research allegiance, that is a specific intellectual conflict of interest that is consistent with one’s professional or personal commitment to one type of intervention and may consequently distort the outcomes (Leykin 2009), the use of the task‐shifting approach in which those people delivering the interventions were not involved in its development should lower the probability of this risk.
Potential biases in the review process
Although we tried to include any primary level and community worker categories, it is possible that some studies have been missed. Moreover, although it is unlikely that RCTs would be conducted and would not be publicly accessible, not all those conducting research may necessarily value academic publications, so work may be disseminated through other channels. In addition, primary level and community workers do not have standard widely accepted definitions yet, so some readers may disagree with the definitions that we used or with how this review has aggregated interventionists.
In addition, when we were unable to collect data on our primary outcome, in some cases, we used the number of participants with symptom levels below cut‐offs as a proxy of the cases diagnosed with formal diagnostic instruments, such as the MINI Neuropsychiatric Interview (Sheehan 1998). In the present review, we included interventions with broadly similar aims and methods in countries with similar incomes. However, the diversity of prevention approaches and the different sociocultural and healthcare system contexts in which these interventions were delivered might explain the identified heterogeneity in results for some outcomes.
In general, only a small proportion of RCTs focused on children and adolescents, even though it is known that psychological suffering and exposure to traumatic events during childhood and adolescence can negatively impact future achievements (also at academic and work levels) and raise serious risks for health, such as substance abuse, psychological symptoms and suicidal ideation (Betancourt 2020; Fergusson 2005). A Lancet Commission on Adolescent Health and Well‐being suggested investing in child mental health in light of triple dividends of benefits: now, into future adult life, and for the next generation of children (Patton 2016).
Agreements and disagreements with other studies or reviews
To the best of our knowledge, this is the first systematic review assessing the effects of prevention interventions delivered through a task‐shifting approach for reducing the frequency of mental disorders and improving psychological outcomes.
Our findings indicated that there is a scarcity of evidence on child and adolescent populations compared to adults, and are consistent with those of a Cochrane Review on prevention interventions in LMICs affected by humanitarian crises that collected seven trials with 2398 participants (Papola 2020). The review of Papola and colleagues is the only review we are aware of with a primary outcome on the incidence of mental disorders. This work did not identify any RCT providing data on this outcome at any time point. In relation to the improvement of psychological symptoms in adults, our findings align with those of a Cochrane Review of psychological therapies in humanitarian settings in LMICs (Purgato 2018a), and with a systematic review carried out by Tol and colleagues in 2011 (Tol 2011), which identified substantial beneficial effects of psychological interventions versus control conditions for adults with symptoms of PTSD. Our results of a positive effect of preventive psychological interventions in reducing depressive symptoms in adults are in line with the results of a systematic review and Individual Participant Data (IPD) meta‐analysis collecting 13 RCTs with IPD from 11 RCTs with 4145 participants. Included interventions were CBT, behavioural activation, problem‐solving therapy, and supportive therapy, and all of them were delivered using a task‐shifting approach (Karyotaki 2022). Authors found that task‐shared psychological interventions were associated with a larger reduction in depressive symptom severity and a greater chance of response and remission than control measures, confirming the results of other reviews conducted in LMIC settings (Cuijpers 2018; Singla 2017).
For children and adolescents, the review by Tol and colleagues detected a non‐substantial trend in favour of interventions versus control conditions for PTSD symptoms, along with a substantial effect for internalizing symptoms. A systematic review with IPD meta‐analysis collected data on 3143 children from 11 trials in humanitarian settings in LMICs. This review collected evidence‐based psychotherapeutic interventions within the focused psychosocial support layer of the IASC pyramid (IASC 2007). These interventions involved techniques inspired by psychotherapeutic approaches such as cognitive‐behavioural therapy, but not following complete standard treatment protocols, and the inclusion of additional techniques aimed at establishing strengths, such as creative expressive techniques (e.g. drama, dance, music, art, and games), social support‐building activities (e.g. cooperative games, trust‐focused activities, sharing difficulties, and coping methods), or mind–body‐oriented skills (e.g. meditation and breathing exercises). Authors identified a beneficial effect of interventions in reducing PTSD symptoms and functional impairment, and in increasing hope, coping, and social support (Purgato 2018b).
Our results on secondary outcomes align also with those of a systematic review on the effectiveness of mental health promotion interventions for young people in LMICs, which collected 22 studies (RCTs and quasi‐randomized studies) on 20 school‐based and community interventions (Barry 2013). The review found significant positive effects on students’ emotional and behavioural well‐being, including reduced depression and anxiety and improved coping skills (Barry 2013). Taking into account primary‐level worker interventions for people with mental disorders and distress in LMICs, Van Ginneken 2021 found a difference in the effectiveness of interventions between adult and child populations. As a matter of fact, for adult populations (adults with depression and anxiety, women with depression related to pregnancy and childbirth, and adults in humanitarian settings with post‐traumatic stress or depression and anxiety), treatments from lay health workers may reduce symptoms of depression, whereas, for children in humanitarian settings with post‐traumatic stress or depression and anxiety, LHW‐led interventions may have little to no effect on post‐traumatic stress and depressive symptoms.
Map of countries from included studies
Comparison 1: Promotion/universal prevention interventions versus control group in preventing mental disorders in adults, Outcome 1: Diagnosis of mental disorders at 1‐6 months
Comparison 1: Promotion/universal prevention interventions versus control group in preventing mental disorders in adults, Outcome 2: Diagnosis of mental disorders at 7‐24 months
Comparison 1: Promotion/universal prevention interventions versus control group in preventing mental disorders in adults, Outcome 3: Quality of life at 0‐1 months
Comparison 1: Promotion/universal prevention interventions versus control group in preventing mental disorders in adults, Outcome 4: Quality of life at 7‐24 months
Comparison 1: Promotion/universal prevention interventions versus control group in preventing mental disorders in adults, Outcome 5: Depressive symptoms at 0‐1 months
Comparison 1: Promotion/universal prevention interventions versus control group in preventing mental disorders in adults, Outcome 6: Depressive symptoms at 1‐6 months
Comparison 1: Promotion/universal prevention interventions versus control group in preventing mental disorders in adults, Outcome 7: Depressive symptoms at 7‐24 months
Comparison 1: Promotion/universal prevention interventions versus control group in preventing mental disorders in adults, Outcome 8: Anxiety symptoms at 0‐1 months
Comparison 1: Promotion/universal prevention interventions versus control group in preventing mental disorders in adults, Outcome 9: Distress/PTSD symptoms at 0‐1 months
Comparison 1: Promotion/universal prevention interventions versus control group in preventing mental disorders in adults, Outcome 10: Social outcomes at 0‐1 months
Comparison 2: Selective prevention intervention versus control group in preventing mental disorders in adults, Outcome 1: Diagnosis of mental disorders at 7‐24 months
Comparison 2: Selective prevention intervention versus control group in preventing mental disorders in adults, Outcome 2: Quality of life at 0‐1 months
Comparison 2: Selective prevention intervention versus control group in preventing mental disorders in adults, Outcome 3: Quality of life at 1‐6 months
Comparison 2: Selective prevention intervention versus control group in preventing mental disorders in adults, Outcome 4: Depressive symptoms at 0‐1 months
Comparison 2: Selective prevention intervention versus control group in preventing mental disorders in adults, Outcome 5: Depressive symptoms at 1‐6 months
Comparison 2: Selective prevention intervention versus control group in preventing mental disorders in adults, Outcome 6: Anxiety symptoms at 1‐6 months
Comparison 2: Selective prevention intervention versus control group in preventing mental disorders in adults, Outcome 7: Distress/PTSD symptoms at 0‐1 months
Comparison 2: Selective prevention intervention versus control group in preventing mental disorders in adults, Outcome 8: Distress/PTSD symptoms at 1‐6 months
Comparison 2: Selective prevention intervention versus control group in preventing mental disorders in adults, Outcome 9: Distress/PTSD symptoms at 7‐24 months
Comparison 2: Selective prevention intervention versus control group in preventing mental disorders in adults, Outcome 10: Social outcomes at 0‐1 months
Comparison 2: Selective prevention intervention versus control group in preventing mental disorders in adults, Outcome 11: Social outcomes at 1‐6 months
Comparison 2: Selective prevention intervention versus control group in preventing mental disorders in adults, Outcome 12: Social outcomes at 7‐24 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 1: Diagnosis of mental disorders at 0‐1 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 2: Diagnosis of mental disorders at 1‐6 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 3: Diagnosis of mental disorders at 7‐24 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 4: Quality of life at 0‐1 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 5: Quality of life at 1‐6 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 6: Quality of life at 7‐24 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 7: Adverse events at 0‐1 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 8: Psychological functioning and impairment at 0‐1 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 9: Psychological functioning and impairment at 1‐6 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 10: Psychological functioning and impairment at 7‐24 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 11: Depressive symptoms at 0‐1 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 12: Depressive symptoms at 1‐6 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 13: Depressive symptoms at 7‐24 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 14: Anxiety symptoms at 0‐1 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 15: Anxiety symptoms at 1‐6 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 16: Anxiety symptoms at 7‐24 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 17: Distress/PTSD symptoms at 0‐1 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 18: Distress/PTSD symptoms at 1‐6 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 19: Distress/PTSD symptoms at 7‐24 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 20: Social outcomes at 0‐1 months
Comparison 3: Indicated prevention intervention versus control group in preventing mental disorders in adults, Outcome 21: Social outcomes at 7‐24 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 1: Quality of life at 0‐1 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 2: Adverse events at 0‐1 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 3: Psychological functioning and impairment at 0‐1 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 4: Psychological functioning and impairment at 1‐6 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 5: Psychological functioning and impairment at 7‐24 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 6: Depressive symptoms at 0‐1 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 7: Depressive symptoms at 1‐6 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 8: Anxiety symptoms at 0‐1 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 9: Anxiety symptoms at 1‐6 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 10: Anxiety symptoms at 7‐24 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 11: Distress/PTSD symptoms at 0‐1 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 12: Distress/PTSD symptoms at 1‐6 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 13: Social outcomes at 0‐1 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 14: Social outcomes at 1‐6 months
Comparison 4: Promotion/universal prevention interventions versus control group in preventing mental disorders in children, Outcome 15: Social outcomes at 7‐24 months
Comparison 5: Selective prevention intervention versus control group in preventing mental disorders in children, Outcome 1: Quality of life at 0‐1 months
Comparison 5: Selective prevention intervention versus control group in preventing mental disorders in children, Outcome 2: Psychological functioning and impairment at 0‐1 months
Comparison 5: Selective prevention intervention versus control group in preventing mental disorders in children, Outcome 3: Depressive symptoms at 0‐1 months
Comparison 5: Selective prevention intervention versus control group in preventing mental disorders in children, Outcome 4: Depressive symptoms at 1‐6 months
Comparison 5: Selective prevention intervention versus control group in preventing mental disorders in children, Outcome 5: Depressive symptoms at 7‐24 months
Comparison 5: Selective prevention intervention versus control group in preventing mental disorders in children, Outcome 6: Anxiety symptoms at 0‐1 months
Comparison 5: Selective prevention intervention versus control group in preventing mental disorders in children, Outcome 7: Anxiety symptoms at 1‐6 months
Comparison 5: Selective prevention intervention versus control group in preventing mental disorders in children, Outcome 8: Distress/PTSD symptoms at 0‐1 months
Comparison 5: Selective prevention intervention versus control group in preventing mental disorders in children, Outcome 9: Distress/PTSD symptoms at 1‐6 months
Comparison 5: Selective prevention intervention versus control group in preventing mental disorders in children, Outcome 10: Social outcomes at 0‐1 months
Comparison 6: Indicated prevention intervention versus control group in preventing mental disorders in children, Outcome 1: Diagnosis of mental disorders at 0‐1 months
Comparison 6: Indicated prevention intervention versus control group in preventing mental disorders in children, Outcome 2: Diagnosis of mental disorders at 1 to 6 months
Comparison 6: Indicated prevention intervention versus control group in preventing mental disorders in children, Outcome 3: Quality of life at 0‐1 months
Comparison 6: Indicated prevention intervention versus control group in preventing mental disorders in children, Outcome 4: Psychological functioning and impairment at 0‐1 months
Comparison 6: Indicated prevention intervention versus control group in preventing mental disorders in children, Outcome 5: Psychological functioning and impairment at 1‐6 months
Comparison 6: Indicated prevention intervention versus control group in preventing mental disorders in children, Outcome 6: Depressive symptoms at 0‐1 months
Comparison 6: Indicated prevention intervention versus control group in preventing mental disorders in children, Outcome 7: Depressive symptoms at 1‐6 months
Comparison 6: Indicated prevention intervention versus control group in preventing mental disorders in children, Outcome 8: Depressive symptoms at 7‐24 months
Comparison 6: Indicated prevention intervention versus control group in preventing mental disorders in children, Outcome 9: Anxiety symptoms at 0‐1 months
Comparison 6: Indicated prevention intervention versus control group in preventing mental disorders in children, Outcome 10: Anxiety symptoms at 1‐6 months
Comparison 6: Indicated prevention intervention versus control group in preventing mental disorders in children, Outcome 11: Distress/PTSD symptoms at 0‐1 months
Comparison 6: Indicated prevention intervention versus control group in preventing mental disorders in children, Outcome 12: Distress/PTSD symptoms at 1‐6 months
Comparison 6: Indicated prevention intervention versus control group in preventing mental disorders in children, Outcome 13: Social outcomes at 0‐1 months
Comparison 6: Indicated prevention intervention versus control group in preventing mental disorders in children, Outcome 14: Social outcomes at 1‐6 months
Comparison 7: Sensitivity analysis—excluding studies that might generate high heterogeneity (I² > 75%), Outcome 1: Diagnosis of mental disorders at 0‐1 months—indicated prevention adults
Comparison 7: Sensitivity analysis—excluding studies that might generate high heterogeneity (I² > 75%), Outcome 2: Quality of life at 0‐1 months—selective prevention adults
Comparison 7: Sensitivity analysis—excluding studies that might generate high heterogeneity (I² > 75%), Outcome 3: Quality of life at 1‐6 months—selective prevention adults
Comparison 8: Sensitivity analysis—excluding studies with high risk of bias and some concerns, Outcome 1: Psychological functioning and impairment at 0‐1 months—indicated prevention adults
Comparison 8: Sensitivity analysis—excluding studies with high risk of bias and some concerns, Outcome 2: Depressive symptoms at 0‐1 months—indicated prevention adults
Comparison 8: Sensitivity analysis—excluding studies with high risk of bias and some concerns, Outcome 3: Distress/PTSD symptoms at 0‐1 months—indicated prevention adults
Comparison 9: Subgroup analysis—category of health worker, Outcome 1: Depressive symptoms at 0‐1 months—indicated prevention adults
Comparison 9: Subgroup analysis—category of health worker, Outcome 2: Depressive symptoms at 1‐6 months—indicated prevention adults
Comparison 9: Subgroup analysis—category of health worker, Outcome 3: Distress/PTSD symptoms at 0‐1 months—indicated prevention adults
Comparison 10: Subgroup analysis—setting, Outcome 1: Depressive symptoms at 0‐1 months—indicated prevention adults
Comparison 10: Subgroup analysis—setting, Outcome 2: Depressive symptoms at 1‐6 months—indicated prevention adults
Comparison 10: Subgroup analysis—setting, Outcome 3: Distress/PTSD symptoms at 0‐1 months—indicated prevention adults
| Promotion/universal prevention interventions compared to control group in preventing mental disorders in adults | ||||||
| Patient or population: preventing mental disorders | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with control group | Risk with promotion/universal prevention interventions | |||||
| Diagnosis of mental disorders at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Quality of life at study endpoint (higher score = better quality of life) | ‐ | SMD 0.23 SD lower | ‐ | 684 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.23 (95% CI ‐0.51 to 0.04). It is uncertain whether promotion/universal prevention interventions have any effect on quality of life among adults without risk factors for mental disorders (at post‐intervention) compared with usual care [there is a small effect according to Cohen 1992].1 |
| Adverse events at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Psychological functioning and impairment at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Depressive symptoms at study endpoint (higher score = higher severity ) | ‐ | SMD 0.31 SD lower | ‐ | 349 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.31 (95% CI ‐0.78 to 0.15). It is uncertain whether promotion/universal prevention interventions have any effect on depressive symptoms in adults without risk factors for mental disorders (at post‐intervention) compared to usual care [this is a small effect according to Cohen 1992].1 |
| Anxiety symptoms at study endpoint (higher score = higher severity) | The mean anxiety symptoms at study endpoint was 0 | MD 0.14 lower | ‐ | 158 | ⊕⊕⊕⊝ | Promotion/universal prevention interventions for adults without risk factors for mental disorders probably slightly reduce anxiety symptoms (at post‐intervention) compared to usual care [there is a small effect according to Cohen 1992].1 |
| Distress/PTSD symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.24 SD lower | ‐ | 722 | ⊕⊕⊝⊝ | Scores estimated based on an SMD of ‐0.24 (95% CI ‐0.41 to ‐0.08). Promotion/universal prevention interventions may slightly reduce distress/PTSD symptoms in adults without risk factors for mental disorders (at post‐intervention) comparedto usual care [there is a small effect according to Cohen 1992].1 |
| *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 | ||||||
| See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_429913845297007331. | ||||||
| a Downgraded 2 levels owing to study limitations (over 30% of RCTs have overall high risk of bias) | ||||||
| Selective prevention interventions compared to control group in preventing mental disorders in adults | ||||||
| Patient or population: preventing mental disorders | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with control group | Risk with selective prevention interventions | |||||
| Diagnosis of mental disorders at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Quality of life at study endpoint (higher score = better quality of life) | ‐ | SMD 1.64 lower | ‐ | 229 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐1.64 (95% CI ‐2.97 to ‐0.31). It is uncertain whether selective prevention interventions have any effect on quality of life among adults with risk factors for mental disorders/lack of protective factors (at post‐intervention) compared with usual care. [There is a large effect according to Cohen 1992]1 |
| Adverse events at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Psychological functioning and impairment at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Depressive symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.69 lower | ‐ | 223 | ⊕⊕⊕⊝ | Scores estimated based on an SMD of ‐0.69 (95% CI ‐1.08 to ‐0.3). Selective prevention interventions for adults with risk factors for mental disorders/lack of protective factors probably slightly reduce depressive symptoms (at post‐intervention)compared to usual care. [There is a medium effect according to Cohen 1992]1 |
| Anxiety symptoms at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Distress/PTSD symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.9 lower | ‐ | 535 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.90 (95% CI ‐1.44 to ‐0.36). It is uncertain whether selective prevention interventions have any effect on distress/PTSD symptoms in adults with risk factors for mental disorders/lack of protective factors (at post‐intervention) compared to usual care. [There is a large effect according to Cohen 1992]1 |
| *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 | ||||||
| See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_429913907887557578. | ||||||
| a Downgraded 2 level owing to inconsistency (I2 was higher than 75%, P < 0.00001) | ||||||
| Indicated prevention interventions compared to control group in preventing mental disorders in adults | ||||||
| Patient or population: preventing mental disorders | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with control group | Risk with indicated prevention interventions | |||||
| Diagnosis of mental disorders at study endpoint (RR < 1 denotes lower risk of mental diagnosis) | 170 per 1000 | 51 per 1000 | RR 0.30 | 843 | ⊕⊝⊝⊝ | It is uncertain whether indicated prevention interventions have any effect on the risk of mental disorders in adults with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. |
| Quality of life at study endpoint (higher score = better quality of life) | ‐ | SMD 0.36 lower | ‐ | 1136 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.36 (95% CI ‐0.61 to ‐0.12). It is uncertain whether indicated prevention interventions have any effect on quality of life among adults with a high vulnerability to develop mental disorders (at post‐intervention) compared with usual care. [There is a small effect according to Cohen 1992]1 |
| Adverse events at study endpoint | Not pooled | Not pooled | Not pooled | (1 RCT) | ⊕⊕⊝⊝ | Indicated prevention interventions may reduce adverse events in adults with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. |
| Psychological functioning and impairment at study endpoint (higher score = higher disability) | ‐ | SMD 0.12 lower | ‐ | 663 | ⊕⊕⊕⊝ | Scores estimated based on an SMD of ‐0.12 (95% CI ‐0.39 to 0.15). Indicated prevention interventions for adults with a high vulnerability to develop mental disorders probably slightly reduce functional impairment (at post‐intervention)compared to usual care. [There is a small effect according to Cohen 1992]1 |
| Depressive symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.16 lower | ‐ | 2341 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.16 (95% CI ‐0.3 to ‐0.03). It is uncertain whether indicated prevention interventions have any effect on depressive symptoms in adults with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. [There is a small effect according to Cohen 1992]1 |
| Anxiety symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 1.19 lower | ‐ | 250 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐1.19 (95% CI ‐2.02 to ‐0.035). It is uncertain whether indicated prevention interventions have any effect on depressive symptoms in adults with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. [There is a large effect according to Cohen 1992]1 |
| Distress/PTSD symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.54 lower | ‐ | 2536 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.54 (95% CI ‐0.95 to ‐0.14). It is uncertain whether indicated prevention interventions have any effect on distress/PTSD symptoms in adults with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. [There is a medium effect according to Cohen 1992]1 |
| *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 | ||||||
| See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_429913973563280333. | ||||||
| a Downgraded 2 levels owing to study limitations (over 30% of RCTs had high risk of bias due to missing outcome data and in selection of the reported result) | ||||||
| Promotion/universal prevention interventions compared to control group in preventing mental disorders in children | ||||||
| Patient or population: preventing mental disorders | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with control group | Risk with promotion/universal prevention interventions | |||||
| Diagnosis of mental disorders at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Quality of life at study endpoint (higher score = better quality of life) | ‐ | SMD 0.25 SD lower | ‐ | 803 | ⊕⊕⊝⊝ | Scores estimated based on an SMD of ‐0.25 (95% CI ‐0.39 to ‐0.11). Promotion/universal prevention interventions may improve the quality of life of children without risk factors for mental disorders (at post‐intervention) compared to usual care. [There is a small effect according to Cohen 1992]1 |
| Adverse events at study endpoint (RR < 1 indicates lower risk of adverse events) | Not pooled | Not pooled | Not pooled | (1 RCT) | ⊕⊕⊝⊝ | Promotion/universal prevention interventions may reduce adverse events in children without risk factors for mental disorders (at post‐intervention) compared to usual care |
| Psychological functioning and impairment at study endpoint (higher score = higher disability) | ‐ | SMD 0.04 higher | ‐ | 212 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of 0.04 (95% CI ‐0.9 to 0.98). It is uncertain whether promotion/universal prevention interventions have any effect on functional impairment in children without risk factors for mental disorders (at post‐intervention) compared to usual care. [There is a small effect according to Cohen 1992]1 |
| Depressive symptoms at study endpoint (higher score = higher severity) | MD 3.04 SD lower | ‐ | 160 | ⊕⊕⊝⊝ | Promotion/universal prevention interventions may slightly reduce depression symptoms in children without risk factors for mental disorders (at post‐intervention) compared to usual care. [There is a large effect according to Cohen 1992]1 | |
| Anxiety symptoms at study endpoint (higher score = higher severity) | MD 2.77 higher | ‐ | 183 | ⊕⊕⊝⊝ | Promotion/universal prevention interventions may slightly reduce anxiety symptoms in children without risk factors for mental disorders (at post‐intervention) compared to usual care. [There is a medium effect according to Cohen 1992]1 | |
| Distress/PTSD symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.83 SD lower | ‐ | 800 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.83 (95% CI ‐2.48 to 0.82). It is uncertain whether promotion/universal prevention interventions have any effect on distress/PTSD symptoms in children without risk factors for mental disorders (at post‐intervention) compared to usual care. [There is a large effect according to Cohen 1992]1 |
| *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 | ||||||
| See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_429913294102284283. | ||||||
| a Downgraded 1 level owing to study limitations (all RCTs had some concerns for the deviations from the intended interventions and in measurement of the outcome) | ||||||
| Selective prevention interventions compared to control group in preventing mental disorders in children | ||||||
| Patient or population: preventing mental disorders | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with control group | Risk with selective prevention interventions | |||||
| Diagnosis of mental disorders at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Quality of life at study endpoint (higher score = better quality of life) | MD 1.1 higher | ‐ | 115 | ⊕⊝⊝⊝ | It is uncertain whether selective prevention interventions have any effect on quality of life among children with risk factors for mental disorders/lack of protective factors (at post‐intervention) compared with usual care. [There is a small effect according to Cohen 1992]1 | |
| Adverse events at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Psychological functioning and impairment at study endpoint (higher score = higher disability) | MD 0.02 higher | ‐ | 479 | ⊕⊕⊝⊝ | There is no evidence that selective prevention interventions improve functional impairment in children with risk factors for mental disorders/lack of protective factors (at post‐intervention) compared to usual care. 1 | |
| Depressive symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0 SD | ‐ | 638 | ⊕⊕⊕⊝ | Scores estimated based on an SMD of 0.0 (95% CI ‐0.16 to 0.15). Selective prevention interventions for children with risk factors for mental disorders/lack of protective factors probably slightly reduce depressive symptoms (at post‐intervention)compared to usual care. [There is a small effect according to Cohen 1992]1 |
| Anxiety symptoms at study endpoint (higher score = higher severity) | MD 4.5 higher | ‐ | 28 | ⊕⊕⊝⊝ | Selective prevention interventions may make little or no difference to anxiety symptoms in children with risk factors for mental disorders/lack of protective factors (at post‐intervention) compared to usual care. 1 | |
| Distress/PTSD symptoms at study endpoint (higher score = higher severity) | MD 2.14 lower | ‐ | 159 | ⊕⊕⊕⊝ | Selective prevention interventions for children with risk factors for mental disorders/ lack of protective factors probably slightly reduce distress/PTSD symptoms (at post‐intervention) compared to usual care. [There is a small effect according to Cohen 1992]1 | |
| *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 | ||||||
| See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_429913697396673091. | ||||||
| a Downgraded 2 levels owing to study limitations (over 30% of RCTs had high risk of bias due to deviations from intended interventions and missing outcome data, in measurement of the outcome, and in selection of the reported result) | ||||||
| Indicated prevention interventions compared to control group in preventing mental disorders in children | ||||||
| Patient or population: preventing mental disorders | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with control group | Risk with indicated prevention interventions | |||||
| Diagnosis of mental disorders at study endpoint (RR < 1 denotes lower risk of mental diagnosis) | 336 per 1000 | 259 per 1000 | RR 0.77 | 220 | ⊕⊝⊝⊝ | It is uncertain whether indicated prevention interventions have any effect on the risk of mental disorders in children with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. |
| Quality of life at study endpoint (higher score = better quality of life) | ‐ | SMD 0.65 SD lower | ‐ | 152 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.65 (95% CI ‐2.09 to 0.79). It is uncertain whether indicated prevention interventions have any effect on quality of life among children with a high vulnerability to develop mental disorders (at post‐intervention) compared with usual care. [There is a small effect according to Cohen 1992]1 |
| Adverse events at study endpoint | No studies that measured this outcome were identified. | (0 studies) | ‐ | |||
| Psychological functioning and impairment at study endpoint (higher score = higher disability) | ‐ | SMD 0.29 SD lower | ‐ | 448 | ⊕⊕⊕⊕ | Scores estimated based on an SMD of ‐0.29 (95% CI ‐0.47 to ‐0.1). Indicated prevention interventions decrease slightly functional impairment in children with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. [There is [a small effect according to Cohen 1992]1 |
| Depressive symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.18 SD lower | ‐ | 771 | ⊕⊕⊕⊕ | Scores estimated based on an SMD of ‐0.18 (95% CI ‐0.32 to ‐0.04). Indicated prevention interventions decrease slightly depressive symptoms in children with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. [There is a small effect according to Cohen 1992]1 |
| Anxiety symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.09 lower | ‐ | 888 | ⊕⊝⊝⊝ | Scores estimated based on an SMD of ‐0.09 (95% CI ‐0.22 to 0.04). It is uncertain whether indicated prevention interventions have any effect on anxiety symptoms in children with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. [There is a small effect according to Cohen 1992]1 |
| Distress/PTSD symptoms at study endpoint (higher score = higher severity) | ‐ | SMD 0.24 SD higher | ‐ | 448 | ⊕⊕⊝⊝ | Scores estimated based on an SMD of 0.24 (95% CI ‐1.28 to 1.76). Indicated prevention interventions may slightlyreduce distress/PTSD symptoms in children with a high vulnerability to develop mental disorders (at post‐intervention) compared to usual care. [There is a small effect according to Cohen 1992]1 |
| *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 | ||||||
| See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_429913780370754267. | ||||||
| a Downgraded 1 level owing to study limitations (RCT did not provided information about allocation concealment, and outcome assessment was not described as masked) | ||||||
| Adults | Participants who were ≥ 18 years old. If some studies had an age range from, for example, 16 years upwards, and a majority of participants (≥ 80%) were over 18 years of age, we included these study participants as adults. |
| Children and adolescents | Children (from birth to 18 years) were considered as a separate group of participants, as they have: • different patterns of psychopathology/mental disorders; and • different help‐seeking behaviours that would, therefore, require different interventions, in different settings (e.g. schools), and a different approach to interventions (e.g. worker interventions such as teacher‐led interventions). |
| Promotion | Promotion is an approach aimed at strengthening positive aspects of mental health and psychosocial well‐being; it includes, for example, components to foster prosocial behaviour, self‐esteem, positive coping with stress, and decision‐making capacity (National Academies of Sciences 2019; WHO 2014). Prevention is an approach aimed at reducing the likelihood of future disorder within the general population or amongst people who are identified as being at risk for developing a full‐blown disorder (Eaton 2012; Tol 2015a). |
| Universal prevention | Universal prevention includes strategies that can be offered to the whole population, based on evidence that prevention strategies are likely to provide some benefit to all (i.e. reduce the probability of a disorder), which clearly outweighs the costs and risks of negative consequences. Examples of common universal prevention interventions include:
|
| Selective prevention | Selective prevention refers to strategies that are targeted to subpopulations identified as being at elevated risk for a disorder; it includes:
|
| Indicated prevention | Indicated prevention includes strategies that are targeted to individuals who are identified (or individually screened) as having increased vulnerability for a disorder based on some individual assessment. These interventions include:
|
| First‐level care, primary care, and community care | First‐level contact with formal health services consists of community‐based interventions or primary care interventions (or both), on their own or attached to hospital settings, provided they had no specialist input apart from supervision (modified from Wiley‐Exley 2007). This would include promotion or prevention programmes in outpatient clinics or primary care practices. This would not include programmes in hospitals unless these programmes were providing prevention interventions to outpatients. Community programmes involve detection of mental disorders in all age groups, often done outside the health facility, for example, through school, training, and other community settings. |
| Low‐ and middle‐income countries (LMICs) | Any country that has ever been an LMIC, as defined by the World Bank lists of LMICs |
| Primary care health workers (PHWs) | Health workers who are not specializing in mental disorders or have not received in‐depth professional specialist training in this clinical area. They work in primary care centres or in the community. These individuals include doctors, nurses, auxiliary nurses, lay health workers, and allied health personnel such as social workers and occupational therapists. This category does not include professional specialist health workers such as psychiatrists, psychiatric nurses, or mental health social workers. For inclusion, PHWs received some training in mental conditions (in the control group or in the intervention group), but this would not constitute a professional category. Study authors made a judgement of what constitutes ‘some training’. Examples of ‘some training’ may include an undergraduate module or a short course in mental health. |
| Community workers (CWs) | People involved as community‐level workers but who are not within the health sector, as many people, particularly adolescents and young adults, have limited contact with health workers. This category includes teachers/trainers/support workers from schools and colleges, along with other volunteers or workers within community‐based networks or nongovernmental organisations. These CWs have an important role, particularly in promotion of mental health and detection of mental disorders (Patel 2007b; Patel 2008). We excluded from this review studies that looked at informal care provided by family members or that extended care only to members of their own family (i.e. who were unavailable to other members of the community). As was previously highlighted in Lewin’s Cochrane Review, “these interventions are qualitatively different from other LHW [lay health worker] interventions included in this review given that parents or spouses have an established close relationship with those receiving care, which could affect the process and effects of the intervention” (Lewin 2010). |
| Primary‐level workers (PWs) | Broad term to encompass both CWs and PHWs |
| CW: community worker | |
| Author year | Country | Type of economic analysis | Study population | Interventions | Intervention‐specific costs and cost‐effectiveness | Resources (i.e. costs to health services other than intervention costs; patient/society costs and productivity) |
|---|---|---|---|---|---|---|
| Sri Lanka | Cost analysis | Children | Classroom‐based intervention (CBI) vs waiting list | Costs: cost analyses for intervention group demonstrated mean cost per service user was USD 8.85 (56% of which is human resources cost) | Health service costs: ‘costs’ included broader package | |
| ‐ | ‐ | Cost data: calculated | ‐ | Delivered by LHWs (paraprofessional | Cost‐effectiveness: cost analyses | Patient cost: none reported |
| Jamaica | Cost analysis | Adults (female, age not specified) | Parenting intervention with routine primary health care vs usual care | Costs: the cost per child was USD 100.9 for 1 year of intervention | Health service cost: USD 100.9 per child including equipment purchases, materials, training, and wages | |
| ‐ | ‐ | ‐ | ‐ | Delivered by community health workers and nurses; training: 3‐day workshops with viewing of films and role plays | Cost‐effectiveness: not reported | Patient cost: none reported |
| Kenya | Cost analysis | Adolescents (both male and females, 13 to 18 years of age) | Shamiri‐Digital Wellness vs active control | Cost: USD 3.57 per student to deliver Shamiri‐Digital | Health service cost: health service costs included equipment (computers, desks, chairs) with an hourly cost of USD 0.97, totalling USD 104.65 for the 9 months of the intervention | |
| ‐ | ‐ | ‐ | ‐ | Self‐help digital‐based intervention | Cost‐effectiveness: depending on the definition of clinically meaningful improvement, 7.1 to 9.7 students needed to receive the intervention for 1 student to experience a clinically meaningful improvement, which translated to a cost of USD 25.35 to USD 34.62 per student | Patient cost: none reported |
| CBI: classroom‐based intervention | ||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1.1 Diagnosis of mental disorders at 1‐6 months Show forest plot | 1 | 137 | Risk Ratio (M‐H, Random, 95% CI) | 0.57 [0.28, 1.18] |
| 1.2 Diagnosis of mental disorders at 7‐24 months Show forest plot | 1 | 323 | Risk Ratio (M‐H, Random, 95% CI) | 0.67 [0.43, 1.05] |
| 1.3 Quality of life at 0‐1 months Show forest plot | 4 | 684 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.23 [‐0.51, 0.04] |
| 1.4 Quality of life at 7‐24 months Show forest plot | 1 | 222 | Mean Difference (IV, Random, 95% CI) | ‐7.30 [‐12.28, ‐2.32] |
| 1.5 Depressive symptoms at 0‐1 months Show forest plot | 3 | 349 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.31 [‐0.78, 0.15] |
| 1.6 Depressive symptoms at 1‐6 months Show forest plot | 1 | 153 | Mean Difference (IV, Random, 95% CI) | ‐1.25 [‐3.28, 0.78] |
| 1.7 Depressive symptoms at 7‐24 months Show forest plot | 2 | 1207 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.13 [‐0.24, ‐0.02] |
| 1.8 Anxiety symptoms at 0‐1 months Show forest plot | 1 | 158 | Mean Difference (IV, Random, 95% CI) | ‐0.14 [‐0.27, ‐0.01] |
| 1.9 Distress/PTSD symptoms at 0‐1 months Show forest plot | 4 | 722 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.24 [‐0.41, ‐0.08] |
| 1.10 Social outcomes at 0‐1 months Show forest plot | 1 | 158 | Mean Difference (IV, Random, 95% CI) | ‐0.45 [‐0.82, ‐0.08] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 2.1 Diagnosis of mental disorders at 7‐24 months Show forest plot | 1 | 349 | Risk Ratio (M‐H, Random, 95% CI) | 1.81 [0.17, 19.82] |
| 2.2 Quality of life at 0‐1 months Show forest plot | 3 | 229 | Std. Mean Difference (IV, Random, 95% CI) | ‐1.64 [‐2.97, ‐0.31] |
| 2.3 Quality of life at 1‐6 months Show forest plot | 5 | 798 | Std. Mean Difference (IV, Random, 95% CI) | ‐1.05 [‐1.84, ‐0.26] |
| 2.4 Depressive symptoms at 0‐1 months Show forest plot | 4 | 223 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.69 [‐1.08, ‐0.30] |
| 2.5 Depressive symptoms at 1‐6 months Show forest plot | 3 | 186 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.60 [‐1.00, ‐0.21] |
| 2.6 Anxiety symptoms at 1‐6 months Show forest plot | 1 | 2026 | Mean Difference (IV, Random, 95% CI) | ‐0.80 [‐1.87, 0.27] |
| 2.7 Distress/PTSD symptoms at 0‐1 months Show forest plot | 7 | 535 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.90 [‐1.44, ‐0.36] |
| 2.8 Distress/PTSD symptoms at 1‐6 months Show forest plot | 5 | 464 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.67 [‐1.21, ‐0.12] |
| 2.9 Distress/PTSD symptoms at 7‐24 months Show forest plot | 1 | 27 | Mean Difference (IV, Random, 95% CI) | 8.80 [‐3.55, 21.15] |
| 2.10 Social outcomes at 0‐1 months Show forest plot | 3 | 121 | Mean Difference (IV, Random, 95% CI) | ‐9.50 [‐15.29, ‐3.70] |
| 2.11 Social outcomes at 1‐6 months Show forest plot | 2 | 236 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.88 [‐2.56, 0.80] |
| 2.12 Social outcomes at 7‐24 months Show forest plot | 1 | 27 | Mean Difference (IV, Random, 95% CI) | ‐15.70 [‐28.35, ‐3.05] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 3.1 Diagnosis of mental disorders at 0‐1 months Show forest plot | 3 | 843 | Risk Ratio (M‐H, Random, 95% CI) | 0.30 [0.06, 1.57] |
| 3.2 Diagnosis of mental disorders at 1‐6 months Show forest plot | 6 | 1352 | Risk Ratio (M‐H, Random, 95% CI) | 0.65 [0.50, 0.84] |
| 3.3 Diagnosis of mental disorders at 7‐24 months Show forest plot | 2 | 380 | Risk Ratio (M‐H, Random, 95% CI) | 0.69 [0.41, 1.19] |
| 3.4 Quality of life at 0‐1 months Show forest plot | 8 | 1136 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.36 [‐0.61, ‐0.12] |
| 3.5 Quality of life at 1‐6 months Show forest plot | 4 | 847 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.04 [‐0.23, 0.16] |
| 3.6 Quality of life at 7‐24 months Show forest plot | 1 | 94 | Mean Difference (IV, Random, 95% CI) | ‐0.80 [‐3.53, 1.93] |
| 3.7 Adverse events at 0‐1 months Show forest plot | 1 | 547 | Risk Ratio (M‐H, Random, 95% CI) | Not estimable |
| 3.8 Psychological functioning and impairment at 0‐1 months Show forest plot | 4 | 663 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.12 [‐0.39, 0.15] |
| 3.9 Psychological functioning and impairment at 1‐6 months Show forest plot | 2 | 594 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.10 [‐0.60, 0.41] |
| 3.10 Psychological functioning and impairment at 7‐24 months Show forest plot | 2 | 241 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.21 [‐0.47, 0.04] |
| 3.11 Depressive symptoms at 0‐1 months Show forest plot | 18 | 2341 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.16 [‐0.30, ‐0.03] |
| 3.12 Depressive symptoms at 1‐6 months Show forest plot | 11 | 2609 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.34 [‐0.58, ‐0.10] |
| 3.13 Depressive symptoms at 7‐24 months Show forest plot | 8 | 2149 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.10 [‐0.22, 0.01] |
| 3.14 Anxiety symptoms at 0‐1 months Show forest plot | 5 | 250 | Std. Mean Difference (IV, Random, 95% CI) | ‐1.19 [‐2.02, ‐0.35] |
| 3.15 Anxiety symptoms at 1‐6 months Show forest plot | 4 | 771 | Std. Mean Difference (IV, Fixed, 95% CI) | ‐0.23 [‐0.37, ‐0.09] |
| 3.16 Anxiety symptoms at 7‐24 months Show forest plot | 2 | 549 | Std. Mean Difference (IV, Fixed, 95% CI) | ‐0.12 [‐0.29, 0.05] |
| 3.17 Distress/PTSD symptoms at 0‐1 months Show forest plot | 19 | 2536 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.54 [‐0.95, ‐0.14] |
| 3.18 Distress/PTSD symptoms at 1‐6 months Show forest plot | 9 | 1702 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.29 [‐0.51, ‐0.07] |
| 3.19 Distress/PTSD symptoms at 7‐24 months Show forest plot | 5 | 1081 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.04 [‐0.45, 0.38] |
| 3.20 Social outcomes at 0‐1 months Show forest plot | 5 | 932 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.08 [‐0.26, 0.09] |
| 3.21 Social outcomes at 7‐24 months Show forest plot | 1 | 241 | Mean Difference (IV, Random, 95% CI) | ‐0.19 [‐1.88, 1.50] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 4.1 Quality of life at 0‐1 months Show forest plot | 2 | 803 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.25 [‐0.39, ‐0.11] |
| 4.2 Adverse events at 0‐1 months Show forest plot | 1 | 694 | Risk Ratio (M‐H, Random, 95% CI) | Not estimable |
| 4.3 Psychological functioning and impairment at 0‐1 months Show forest plot | 2 | 212 | Std. Mean Difference (IV, Random, 95% CI) | 0.04 [‐0.90, 0.98] |
| 4.4 Psychological functioning and impairment at 1‐6 months Show forest plot | 1 | 90 | Mean Difference (IV, Random, 95% CI) | ‐0.29 [‐0.93, 0.35] |
| 4.5 Psychological functioning and impairment at 7‐24 months Show forest plot | 1 | 183 | Mean Difference (IV, Random, 95% CI) | ‐3.33 [‐5.03, ‐1.63] |
| 4.6 Depressive symptoms at 0‐1 months Show forest plot | 1 | 160 | Mean Difference (IV, Random, 95% CI) | ‐3.04 [‐6.00, ‐0.08] |
| 4.7 Depressive symptoms at 1‐6 months Show forest plot | 3 | 385 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.00 [‐0.20, 0.20] |
| 4.8 Anxiety symptoms at 0‐1 months Show forest plot | 1 | 183 | Mean Difference (IV, Random, 95% CI) | ‐2.27 [‐3.13, ‐1.41] |
| 4.9 Anxiety symptoms at 1‐6 months Show forest plot | 1 | 125 | Mean Difference (IV, Random, 95% CI) | ‐0.13 [‐0.41, 0.15] |
| 4.10 Anxiety symptoms at 7‐24 months Show forest plot | 1 | 183 | Mean Difference (IV, Random, 95% CI) | ‐2.27 [‐3.10, ‐1.44] |
| 4.11 Distress/PTSD symptoms at 0‐1 months Show forest plot | 2 | 800 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.83 [‐2.48, 0.82] |
| 4.12 Distress/PTSD symptoms at 1‐6 months Show forest plot | 1 | 106 | Mean Difference (IV, Random, 95% CI) | ‐4.51 [‐5.86, ‐3.16] |
| 4.13 Social outcomes at 0‐1 months Show forest plot | 3 | 321 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.32 [‐0.76, 0.12] |
| 4.14 Social outcomes at 1‐6 months Show forest plot | 2 | 215 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.27 [‐0.82, 0.28] |
| 4.15 Social outcomes at 7‐24 months Show forest plot | 1 | 183 | Mean Difference (IV, Random, 95% CI) | ‐0.70 [‐1.07, ‐0.33] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 5.1 Quality of life at 0‐1 months Show forest plot | 1 | 115 | Mean Difference (IV, Random, 95% CI) | ‐1.10 [‐3.32, 1.12] |
| 5.2 Psychological functioning and impairment at 0‐1 months Show forest plot | 1 | 479 | Mean Difference (IV, Random, 95% CI) | ‐0.02 [‐0.09, 0.05] |
| 5.3 Depressive symptoms at 0‐1 months Show forest plot | 2 | 638 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.00 [‐0.16, 0.15] |
| 5.4 Depressive symptoms at 1‐6 months Show forest plot | 3 | 791 | Std. Mean Difference (IV, Random, 95% CI) | 0.14 [‐0.12, 0.39] |
| 5.5 Depressive symptoms at 7‐24 months Show forest plot | 1 | 435 | Mean Difference (IV, Random, 95% CI) | 0.22 [‐0.76, 1.20] |
| 5.6 Anxiety symptoms at 0‐1 months Show forest plot | 1 | 28 | Mean Difference (IV, Random, 95% CI) | 4.50 [‐12.05, 21.05] |
| 5.7 Anxiety symptoms at 1‐6 months Show forest plot | 1 | 143 | Mean Difference (IV, Random, 95% CI) | 1.42 [‐0.00, 2.84] |
| 5.8 Distress/PTSD symptoms at 0‐1 months Show forest plot | 1 | 159 | Mean Difference (IV, Random, 95% CI) | ‐2.14 [‐3.77, ‐0.51] |
| 5.9 Distress/PTSD symptoms at 1‐6 months Show forest plot | 1 | 213 | Mean Difference (IV, Random, 95% CI) | ‐0.40 [‐2.49, 1.69] |
| 5.10 Social outcomes at 0‐1 months Show forest plot | 2 | 638 | Std. Mean Difference (IV, Random, 95% CI) | 0.03 [‐0.42, 0.48] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 6.1 Diagnosis of mental disorders at 0‐1 months Show forest plot | 1 | 220 | Risk Ratio (M‐H, Random, 95% CI) | 0.77 [0.51, 1.17] |
| 6.2 Diagnosis of mental disorders at 1 to 6 months Show forest plot | 1 | 220 | Risk Ratio (M‐H, Random, 95% CI) | 0.77 [0.52, 1.14] |
| 6.3 Quality of life at 0‐1 months Show forest plot | 2 | 152 | Mean Difference (IV, Random, 95% CI) | ‐0.65 [‐2.09, 0.79] |
| 6.4 Psychological functioning and impairment at 0‐1 months Show forest plot | 2 | 448 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.29 [‐0.47, ‐0.10] |
| 6.5 Psychological functioning and impairment at 1‐6 months Show forest plot | 3 | 813 | Std. Mean Difference (IV, Random, 95% CI) | 0.07 [‐0.13, 0.27] |
| 6.6 Depressive symptoms at 0‐1 months Show forest plot | 4 | 771 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.18 [‐0.32, ‐0.04] |
| 6.7 Depressive symptoms at 1‐6 months Show forest plot | 6 | 2483 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.07 [‐0.17, 0.03] |
| 6.8 Depressive symptoms at 7‐24 months Show forest plot | 1 | 904 | Mean Difference (IV, Random, 95% CI) | ‐1.27 [‐1.90, ‐0.64] |
| 6.9 Anxiety symptoms at 0‐1 months Show forest plot | 3 | 888 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.09 [‐0.22, 0.04] |
| 6.10 Anxiety symptoms at 1‐6 months Show forest plot | 3 | 1362 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.03 [‐0.15, 0.08] |
| 6.11 Distress/PTSD symptoms at 0‐1 months Show forest plot | 2 | 448 | Mean Difference (IV, Random, 95% CI) | 0.24 [‐1.28, 1.76] |
| 6.12 Distress/PTSD symptoms at 1‐6 months Show forest plot | 2 | 417 | Mean Difference (IV, Random, 95% CI) | 0.96 [‐0.55, 2.47] |
| 6.13 Social outcomes at 0‐1 months Show forest plot | 2 | 435 | Std. Mean Difference (IV, Random, 95% CI) | 0.15 [‐0.15, 0.46] |
| 6.14 Social outcomes at 1‐6 months Show forest plot | 2 | 421 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.22 [‐0.41, ‐0.03] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 7.1 Diagnosis of mental disorders at 0‐1 months—indicated prevention adults Show forest plot | 2 | 339 | Risk Ratio (M‐H, Random, 95% CI) | 0.13 [0.05, 0.33] |
| 7.2 Quality of life at 0‐1 months—selective prevention adults Show forest plot | 2 | 205 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.66 [‐1.26, ‐0.06] |
| 7.3 Quality of life at 1‐6 months—selective prevention adults Show forest plot | 3 | 634 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.43 [‐0.92, 0.05] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 8.1 Psychological functioning and impairment at 0‐1 months—indicated prevention adults Show forest plot | 2 | 126 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.41 [‐0.77, ‐0.05] |
| 8.2 Depressive symptoms at 0‐1 months—indicated prevention adults Show forest plot | 2 | 124 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.19 [‐0.62, 0.23] |
| 8.3 Distress/PTSD symptoms at 0‐1 months—indicated prevention adults Show forest plot | 2 | 112 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.15 [‐1.13, 0.84] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 9.1 Depressive symptoms at 0‐1 months—indicated prevention adults Show forest plot | 18 | 2341 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.16 [‐0.30, ‐0.03] |
| 9.1.1 Community workers | 12 | 1775 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.07 [‐0.21, 0.08] |
| 9.1.2 Primary healthcare workers | 6 | 566 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.35 [‐0.62, ‐0.08] |
| 9.2 Depressive symptoms at 1‐6 months—indicated prevention adults Show forest plot | 11 | 2609 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.34 [‐0.58, ‐0.10] |
| 9.2.1 Community workers | 5 | 1463 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.24 [‐0.34, ‐0.14] |
| 9.2.2 Primary healthcare workers | 6 | 1146 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.44 [‐0.95, 0.08] |
| 9.3 Distress/PTSD symptoms at 0‐1 months—indicated prevention adults Show forest plot | 19 | 2536 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.54 [‐0.95, ‐0.14] |
| 9.3.1 Community workers | 14 | 1906 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.19 [‐0.39, 0.00] |
| 9.3.2 Primary healthcare workers | 5 | 630 | Std. Mean Difference (IV, Random, 95% CI) | ‐1.34 [‐2.90, 0.23] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 10.1 Depressive symptoms at 0‐1 months—indicated prevention adults Show forest plot | 18 | 2341 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.16 [‐0.30, ‐0.03] |
| 10.1.1 Community setting | 8 | 1278 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.17 [‐0.43, 0.09] |
| 10.1.2 Other | 9 | 987 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.17 [‐0.31, ‐0.03] |
| 10.1.3 Refugee camp | 1 | 76 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.11 [‐0.56, 0.34] |
| 10.2 Depressive symptoms at 1‐6 months—indicated prevention adults Show forest plot | 11 | 2609 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.34 [‐0.58, ‐0.10] |
| 10.2.1 Community setting | 4 | 1458 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.27 [‐0.37, ‐0.16] |
| 10.2.2 Other | 6 | 1005 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.40 [‐0.94, 0.14] |
| 10.2.3 School | 1 | 146 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.39 [‐0.72, ‐0.06] |
| 10.3 Distress/PTSD symptoms at 0‐1 months—indicated prevention adults Show forest plot | 19 | 2536 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.54 [‐0.95, ‐0.14] |
| 10.3.1 Community setting | 6 | 1033 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.03 [‐0.23, 0.16] |
| 10.3.2 Other | 11 | 1367 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.82 [‐1.57, ‐0.08] |
| 10.3.3 Refugee camp | 1 | 76 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.28 [‐0.73, 0.17] |
| 10.3.4 School | 1 | 60 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.52 [‐1.04, ‐0.01] |
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Baker‐Henningham 2019 |  | | |  | | | ||||||
| 1a.1: PY; 1a.2: Y. Quote: "Schools were randomised to intervention or control group in the summer preceding the intervention 1a.3: NI. Note: No useful information is reported to evaluate this element. 1b.1: PY; Note: schools and teachers were identified and rectruited before the randomization of clusters, children from the teacher's classes were randomly selected at a later stage (outcome of interest refers to teachers); 1b.2 NA; 1b.3 NA; | ||||||||||||
| 2.1a: NI; 2.1b: PY; 2.2: PY. Note: No information provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were most likely aware of intervention allocation. 2.3: PN. Note: No evidence to suggest deviations from the intended intervention that arose because of the trial context 2.4: NA. 2.5: NA. 2.6: Y. Note: Data was analysed on an intention‐to‐treat basis. 2.7: NA. | ||||||||||||
| 3.1a PY. Note: all 14 clusters, schools, that recruited participants were analyzed; 3.1b: Y. Note:there was 1 loss among teachers (out of 54), no reported loss among children . 3.2: NA. 3.3: NA; 3.4: NA. | ||||||||||||
| 4.1: PN. The outcome measure (CES‐D) is widely established and used across contexts. 4.2: PN. Note: no evidence to suggest differences in measurement between intervention and control groups. 4.3a: NI; 4.3b: PY. Note: No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were most likely aware of the group allocation due to the nature of the intervention. Research assistants that were involved in data collection were blind to allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned interventio could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note:No evidence to suggest otherwise, protocol and paper do not show discrepancies. 5.2: PN. Note:No evidence to suggest outcome selection. All outcomes mentioned in the protocol, methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the protocol, methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Bell 2008 | | | | | | | ||||||
| 1a.1: Y;1.2: PY. Title mentions RCT design. 1.3: NI. Note: No information provided about baseline characterists. 1b.1: Y. Quote: "CHAMPSA staff would explain the CHAMPSA program and chools were then requested to indicate if they wished to participate." Then, the schools were randomized for intervention or control group.Overall comment: Low risk. | ||||||||||||
| 2.1a: PY; 2.1b: PY. Due to the nature of the intervention, participants were most likely aware that they were in a trial and were aware of their assigned intevrention.Quote: "Active caregiver consent and student assent were obtained before study participation". 2.2: Y. School teachers, health educators and community caregivers were trained as facilitators for the program. 2.3: NI. Note: No information provided. 2.4: NA. 2.5: NA. 2.6: NI. Note: No explicit mention of the analyses used to estimate the effect pf assignment to intervention. 2.7: PN. Note: No evidence to suggest that this had a substantial impact on the results. | ||||||||||||
| 3.1a: NI. No information provided. 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note:The selected outcome method (Social Capital) is validated and established. 4.2: PN. Quote: "Measures were obtained before and once after participation in the intervention". No evidence to suggest differences in measurement between intervention and control. 4.3a: Y. The outcome assessors (caregiver) were aware that the trial was taking place. 4.3b: Y. The outcome assessors (caregiver) were aware of the intervention received by study participation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PN. Note: The data were collected n additional sessions before and after the program session due the loe‐literacy levels of by reading aloud the items by adult participants. 5.2: PN. Note: no protocol available.No evidence to suggest outcome selection. All outcome mentioned in the methods and results were reported. 5.3: PN. No protocol available. No evdience to suggest analyses selection. All outcomes mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in three domains (2, 4, 5), but not to be at high risk of bias for any domain. | ||||||||||||
| Duan 2019 | | | |  | | | ||||||
| 1.1: PY; 1.2: NI. Quote: "Randomization 1.3: PN. Quote: "no significant difference | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention participants and personnel were most likely aware of allocation 2.3: PN. Note: No evidence to suggest issues to implement intervention 2.4: NA. 2.5: NA. 2.6: NI. Note: No explicit mention of the analysis used to estimate the effect of assignment to intervention. Data was analyzed for participants whose follow‐up data was available (73%) 2.7: PN. No evidence to suggest analysis in the wrong group or exclusion of participants that could have had a substantial impact on the results | ||||||||||||
| 3.1: PN. Note: Out of the 52 randomized participants, 14 (27%) were lost at follow‐up (1 week after intervention). At post‐intervention, the rate was of 21%. 3.2: PN. Note: No information to suggest that results were not biased by missing outcome data. 3.3: PY. Note: Missing data occurred for reasons that could be related to at least one of the assessed outcomes.The dropouts were balanced across comparison groups. Quote: "the dropout rate was 27% because 3.4: PN. | ||||||||||||
| 4.1: N. Note: The selected outcome method (BIT) is established, validated for the local context, and assessed in its psychometric properties for this sample. 4.2: PY. Note: It appears that different data collection methods were used at post‐intervention for the control and intervention group. Quote: "At the end of the lesson, the intervention group was 4.3: NA. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: NA. 4.5: NA. | ||||||||||||
| 5.1: NI. Note: Information provided is insufficient to make a judgment on this item as a specific method section for statistical procedures is missing. 5.2: NI. Note: Information provided is insufficient to make a judgment on this item as a specific method section for statistical procedures is missing. 5.3: NI. Note: Information provided is insufficient to make a judgment on this item as a specific method section for statistical procedures is missing. | ||||||||||||
| The study is judged to be at high risk of bias in one domain (4). | ||||||||||||
| Hendriks 2019 | | | | | | | ||||||
| 1.1: Y; 1.2: PY. Note: Quote: "Participants were randomly assigned to either the intervention 1.3: PN. Quote: "No | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention participants and personnel were most likely aware of allocation. 2.3: PN. Note: No evidence to suggest issues to implement intervention. 2.4: NA. 2.5: NA. 2.6: Y. Quote: "Of the 173 randomized participants, 15 did not attend the 2.7: NA. | ||||||||||||
| 3.1: PY. Quote: "In total, 158 participants completed the baseline assessment 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. Quote: "Cronbach’s alpha in the present study was 4.2: PN. Note: no evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. The investigators in the trial were blinded. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PN. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: no evidence to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Tripathy 2010 | | | | | | | ||||||
| 1a.1: PY; 1a.2: Y. Note: Randomization was obtained through the drawing of folded papers in the presence of extrnal observers in order to convince local communities to participate. 1.3: NI. Note: No useful information is reported to evaluate this element 1b.1: N: particpants were recruited after randomization; 1b.2: PN. Note: there is no evidence to suggest that selection of individual participants was affected by knoweldge of the intervention assigned to the cluster; 1b.3:PN. Note: groups were comparable at baseline, the observed differences are compatible with chance in a cluster design. | ||||||||||||
| 2.1a: NI; 2.1b: PY; 2.2: PY. Note: No information provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were aware of intervention allocation. 2.3: PN. Note: No evidence to suggest deviations from the intended intervention that arose because of the trial context 2.4: NA. 2.5: NA. 2.6: Y. Note: Data was analysed on an intention‐to‐treat basis. 2.7: NA. | ||||||||||||
| 3.1a: PY. Note: data was available for all clusters; 3.1b: Y. Note: data was available for nearly all participants within clusters, Quote: "All 18 selected clusters 3.2: NA. 3.3: NA. | ||||||||||||
| 4.1: PN. Note: The outcome measure is widely established and validated for use across different contexts, included the one of the trial. 4.2: PN. No evidence to suggest that. 4.3a: NI; 4.3b: PY. Note: No information provided on whether participants knew that they were in a trial, participants were most likely aware of the group allocation due to the nature of the intervention 4.4: PY, 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, protocol and final article not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. Further outcomes than those mentioned in the protocol were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Acarturk 2022 | | | | | | | ||||||
| 1.1: Y. Quote: "Participants were randomly assigned either to the Self‐Help Plus arm or to ECAU only, in a 1:1 ratio." 1.2: Y. Quote "The randomization schedule was generated by Castor Electronic Data Capture (EDC). (..) Research team members involved in 1.3: N. Note: No significant differences at baseline. | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention participants and the interventionists were most likely aware of the allocation. 2.3: PN. Note: No evidence to suggest issues to implement intervention. 2.4: NA. 2.5: NA. 2.6: Y. Quote: "We followed an intent‐to‐treat approach for analysis of primary and secondary outcomes" 2.7: NA. | ||||||||||||
| 3.1: PY. Quote: " The distribution of participants lost to follow‐up was similar between the study groups (15.53% vs. 14.06%)" 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome method (PHQ‐9) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: no evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Asnani 2021 | | | | | | | ||||||
| 1.1: Y; 1.2: Y. Note: Quote: "A random numbers 1.3: PN. Note: Baseline characteristics were overall comparable across groups, significant differences were reported for 2 variables (compatible with chances) | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention participants and personnel were most likely aware of their assigned intervention 2.3: PN. Note: No evidence to suggest issues to implement intervention. Some changes were made to the protocol before recruitment. These are reported in the manuscript. 2.4: NA. 2.5: NA. 2.6: Y. Quote: "Analyses were performed based on intention to treat principles" 2.7: NA. | ||||||||||||
| 3.1: Y. Note: Data was availalble for all 64 randomized participants, there were no losses at the 6 months follow‐up. 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. Note: The selected outcome method (CES‐D) is widely validated and established and demonstrated good inter‐item reliability in the study (0.9) 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. Staff involved in data collection was blinded to the allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. Authors mention that the CES‐D was not initially planned to be an outcome of interest in the study, this was added in a second moment, but still before the beginning of recruitment. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Baumgartner 2021 | | | | | | | ||||||
| 1a.1: PY; 1a.2: PY. Researchers performed stratified constrained randomization. This was carried out through software. Quote: "Constrained randomization was implemented using the cvcrand Stata package by the US‐based team". 1a.3: NI. No useful information is reported to evaluate this element. 1b.1: PN. It appears the randomization was conducted prior to the recruitment of participants: 1b.2: PN. It is unlikely that the selection of individual participants was affected by knowledge of the intervention assigned to the cluster. 1b.3: PN. There were some baseline imbalances that were compatible with chance, there is no evidence to indicate differential identification or recruitment of individual participants. | ||||||||||||
| 2.1 a: NI; 2.1b: PY; 2.2: PY. No information was provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were most likely aware of intervention allocation. 2.3: PY. The low literacy of the lay providers could have impacted implementation fidelity. Quote: "This strategy supported the potential sustainability of the intervention but came with literacy challenges for delivering the iMBC content with full fidelity". 2.4: PN. No evidence to suggest deviations from the intended intervention would have influenced the outcomes. 2.5: NA. 2.6: PY. Data were analysed for most randomized participants. 2.7: NA. | ||||||||||||
| 3.1a: PY. All included clusters were analyzed. 3.1b: PY. Data were available for nearly all participants within the cluster. Quote: "303 (81%), 313 (83.7%), and 266 (71.1%) participants were followed up at the mini‐survey, imme‐ diate post‐intervention survey (follow‐up 1), and the 8‐month post‐ intervention survey (follow‐up 2) time points, respectively ". 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. The outcome measure is well established and reported good psychometric properties in this sample. 4.2: PN. No evidence to suggest differences in measurement between intervention and control. 4.3A: NI; 4.3b: PY. No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were most likely aware of the group allocation due to the nature of the intervention. 4.4: PY; 4.5: PN. Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. No evidence to suggest outcomes selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. No evidence to suggest analyses selection. All outcomes mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in two domains (2 and 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Chaharrahifard 2021 | | | | | | | ||||||
| 1.1: Y. Note: Title mentions RCT design; Quote: "the samples were randomly assigned to two groups of received counseling (intervention) and non‐received counseling (control) by four‐sized randomized blocks." 1.2: Y. Quote: "The details of blocks were contained in asset of sealed. 1.3: N. Quote: "There was not significant difference between the demographic characteristics in the control and intervention group." | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention, participants and the interventionists were most likely aware of the allocation. 2.3: PN. No evidence to suggest issues to implement intervention. 2.4: NA 2.5: NA 2.5: NI. Note: No explicit mention of the analysis used to estimate the effect of assignment to intervention. 2.6: N. Note: There was not a substantial impact of the failure to analyse participants in the group to which they were randomized. | ||||||||||||
| 3.1: PY. Note: Drop‐out rate: 21%. 3.2: NA 3.3: NA 3.4: NA | ||||||||||||
| 4.1: N. Note: The selected outcome method (EPDS) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All outcomes mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in two domains (2 and 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Chang 2015 | | | | | | | ||||||
| 1.1: Y; 1.2: PY. Note: Title mentions RCT design. Quote: "The evaluation was a cluster randomized trial conducted in Jamaica, Antigua, and St Lucia with public health center as the unit of randomization." 1.3: N. Note: Quote: "There were no significant differences between groups in enrollment characteristics" | ||||||||||||
| 2.1a: NI; 2.1b: PY; 2.2: PY. Note: No information provided on wheter participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were most likely aware of intervention allocation. 2.3: PY. Note: Quote: "Clinics were often noisy and crowded, and some mothers would have had 2.4: NA. 2.5: NA. 2.6: Y. Quote: "Analyses were by intention to treat" 2.7: NA. | ||||||||||||
| 3.1a: NI. No information on wheter data was available for all clusters. 3.2: PY. Quote: "Loss did not differ by group (16% of children were lost to follow‐up in the control and 14% were lost to follow‐up in the intevrention)" 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome measure (center for Epidemiologic Studies‐depression Scale) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurements between intevrention and control. 4.3a: PY; 4.3b: PY. Note: Interviewers and testers were blind to center assignment but participants (mothers) probably knew that they were in a trial and participants were most likely aware of the group allocation due to the nature of the intervention. Self‐assessment scale. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcome, but there is no reason to believe that it did | ||||||||||||
| 5.1: PY. Note: no evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to be at high risk of bias in one domain (2). | ||||||||||||
| Chew 2018 | | | | | | | ||||||
| 1a.1: Y; 1a.2: PY. Note: Title mentions RCT design. Quote: "A member of the Data Management Services team at the University Medical center, Ultrecht will carry out the randomisation." 1.3: NI. Note: No useful information is reported to evaluate this element 1b.1: PY. Seems that participants were recruited after randomization. | ||||||||||||
| 2.1a: PY. 2.1b: PN. Quote: "We use a modified informed consent procedure to ensure that patients are unware of the two different intervention programmes". 2.3: PN. Note: No evidence to suggest deviations from the intended intervention that arose because of the trial context. 2.4: NA. 2.5: NA. 2.6: Y. Note: The analysis was carried out on an intention to treat basis. 2.7: NA. | ||||||||||||
| 3.1a: PY. All 10 clusters that recruited participants were analyzed. 3.2: PN. No evidence that the result was not biased by missing data. 3.3: PN; 3.4: NA. Note: Missingness in the outcome do not depend on its true value. | ||||||||||||
| 4.1: N. Note: The selected outcome method (Malay Brief Illness Perception Quetsionnaire, MBIPQ) is validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control 4.3a: PY.Outcome assessors (participants) were most likeley aware that a trial was taking a place. 4.4: NA. 4.5: NA. | ||||||||||||
| 5.1: PY. Note: Protocol is available. 5.2: PN. Note: No evidence to suggest outcome selection. Almost all outcomes mentioned in the methods and protocol were reported (It is not reported "heath‐care utilization hospital" that is a secondary outcome). 5.3: PN. Note: No evidence to suggest analyses selection. All outcomes mentioned in the methdis and protocol were reported. | ||||||||||||
| The study is judged to be at low risk of bias for all domains. | ||||||||||||
| Escolar 2014 | | | | | | | ||||||
| 1.1: Y; 1.2: NI. Quote: "The respondents were then randomly assigned to either the experimental or control group" No information provided about allocation concealment. 1.3: PN. Note: No significant differences at baseline. | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention participants and the interventionists were most likely aware of the allocation. 2.3: NI. Note: No information provided. 2.4: NA. 2.5: NA. 2.6: NI. Note: No infomation provided. 2.7: N. Note: No drop‐out. | ||||||||||||
| 3.1: Y. Note: No missing data. 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome method (PHQ‐9) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: Y. Note: Only self‐reported measure. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No protocol available. No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in three domains (1, 2, 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Ferreira‐Vorkapic 2018 | | | | | | | ||||||
| 1.1: Y; 1.2: Y. Quote: "(...) volunteers were randomly allocated to one of the three experimental groups. A separate examiner coordinated the group distribution (and all statistical analysis was performed by an expert from another 1.3: N. Quote: "the baseline data showing no significant differences between the groups, except for education years, where the relaxation group showed a higher | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention participants and personnel were most likely aware of their assigned intervention 2.3: PN. Note: No evidence to suggest issues to implement intervention 2.4: NA. 2.5: NA. 2.6: NI. Note: No explicit mention of the analysis used to estimate the effect of assignment to intervention 2.7: NA. Note: There was not a substantial impact of the failure to analyse participants in the group to which they were randomized | ||||||||||||
| 3.1: Y. Note: Drop‐out: 0 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome method (BAI) is widely validated and established across multiple languages.ility in the study (0.9) 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in two domains (2 and 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Gao 2015 | | | | | | | ||||||
| 1.1: Y. Quote: "The randomization sequence was generated using a computerized random number generator and the allocation was kept in sealed opaque consecutively numbered envelopes." 1.2.: Y. Quote: "This simple randomization scheme was independently prepared by a research assistant who was not involved in determining eligibility, providing care, or assessing outcome." 1.3: N. Quote: "There were no significant differences between the two groups in their demographic, obstetric and related characteristics (...) There were also no significant differences between the two groups in their baseline measures" | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention, participants and the interventionists were most likely aware of the allocation. 2.3: PN. Note: No evidence to suggest issues to implement intervention. 2.4: NA 2.5: BA 2.6: Y. Note: Intention to treat analysis done. 2.7: NA | ||||||||||||
| 3.1: PY. Note: Drop‐out: 14 of 180. 3.2: NA 3.3: NA 3.4: NA | ||||||||||||
| 4.1: N. Note: The selected outcome method (EPDS) is widely validated in Chinese mothers and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4.: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All outcomes mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Hinton 2021 | | | | | | | ||||||
| 1a.1: PY. 1a.2: Y. Quote: "Randomizationwas at the level of clusters, defined as geographic areas (ie, communes) served by commune health stations which have a population 1a.3: NI. No useful information is reported to evaluate this element 1b.1: N. Note: particpants were recruited after randomization; 1b.2: PN. Note: there is no evidence to suggest that selection of individual participants was affected by knoweldge of the intervention assigned to the cluster; 1b.3:N. Note: no significant differences at baseline were identified. | ||||||||||||
| 2.1a: NI; 2.1b: PY; 2.2: Y. Note: No information provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were most likely aware of intervention allocation. 2.3: PN. Note: No evidence to suggest deviations from the intended intervention that arose because of the trial context 2.4: NA. 2.5: NA. 2.6: Y. Note: Data was analysed on an intention‐to‐treat basis. 2.7: NA. | ||||||||||||
| 3.1a PY. Note: no evidence to suggest failure in analyzing any clyster; 3.1b: PY. Note: data was avaiable for 85% of participants at follow‐up. 3.2: NA. 3.3: NA; 3.4: NA. | ||||||||||||
| 4.1: PN. The outcome measure (CES‐D) is widely established and used across contexts. 4.2: PN. Note: no evidence to suggest that. 4.3a: NI; 4.3b: PY. Note: No information provided on whether participants knew that they were in a trial, participants were most likely aware of the group allocation due to the nature of the intervention. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned interventio could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: Y. Note: A pre‐determined statistical plan is avaiable and in line with procedures reported in the final manuscript 5.2: PN. Note:No evidence to suggest outcome selection. All outcomes mentioned in the protocol, methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the protocol, methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Lachman 2017 | | | | | | | ||||||
| 1.1: Y; 1.2:Y. Quote: "An external researcher not directly involved in the study conducted the randomization procedures remotely in Oxford, United 1.3:PN. | ||||||||||||
| 2.1:Y; 2.2:Y. Quote: "Although program implementers and participants were aware of their allocation status, researchers 2.3: PN. Note: No evidence to suggest issues to implement intervention 2.4:NA. 2.5:NA. 2.6:Y. Quote: "Data analyses were conducted with an intention‐to‐treat design" 2.7:NA. | ||||||||||||
| 3.1:Y. Note: 3 out of the 68 participants randomized were lost at follow‐up (2.9%). 3.2:NA. 3.3:NA. 3.4:NA. | ||||||||||||
| 4.1:PN. Note: The selected outcome method is widely validated and established, appropriate for the local context, and assessed in its psychometric properties for this sample. 4.2:PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3:Y. Note: Outcome assessors (participants) were most likely aware of intervention allocation. The research assistants administering the questionnaires were blinded. 4.4: PY; 4.5:PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did | ||||||||||||
| 5.1:PY. Note: No evidence to suggest otherwise, the protocol and manuscript do not show discrepancies. 5.2: PY. Note: A few of the secondary outcomes mentioned in the protocol (e.g. ASSIST, Grover‐Counter Scale of Cognitive Development) were not reported in this manuscript. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the protocol and methods were reported in the results. | ||||||||||||
| The study is judged to raise some concerns in two domains (4 and 5), but not to be at high risk of bias for any domain. | ||||||||||||
| Lachman 2020 | | | | | | | ||||||
| 1a.1: Y; 1a.2: Y. Quote: "Cluster randomisation was conducted at village level prior to baseline assessments, to reduce the likelihood of contamination between arms. An external researcher used concealed computer‐generated codes to randomly allocate eight villages into three treatment arms and a control arm "; "The implementing partner notified the participating families of their allocation status after baseline data collection." 1a.3: NI. No useful information is reported to evaluate this element. 1b.1: PN. Note: randomization was conducted prior to the recruitment of participants 1b.2: PN. Note: groups were comparable at baseline, see Table 1. | ||||||||||||
| 2.1a: NI; 2.1b: PY; 2.2: PY. Note: No information provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were aware of intervention allocation. 2.3: PN. Note: No evidence to indicate deviations from protocol. 2.4: NA. 2.5: NA. 2.6: Y. Note: Data was analysed on an intention‐to‐treat basis. 2.7: NA. | ||||||||||||
| 3.1a.: Y. Note: all included clusters were analyzed; 3.1b: PY. Note: data were available for nearly all participants within the cluster. Quote:"Study retention was considerably higher than anticipated with 94.8% adults (n=235/248), 87.5% children (n=154/176) and 87.8% in their early childhood (n=122/139) assessments completed at post‐treatment, with no differences between arms " 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. Note: The outcome measure is widely established and validated for use across different contexts. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3a: NI; 4.3b: PY. Note: No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were most likely aware of the group allocation due to the nature of the intervention. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcomes selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All outcomes mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Luoto 2020 | | | | | | | ||||||
| 1a.1: PY; 1a.2: PY. Quote: "Using a random 1a.3: NI. No useful information is reported to evaluate this element. 1b.1: Y. Note: from protocol it is clear that recruitment of both adults and children as well as baseline assessments, took place before randomization to study arms | ||||||||||||
| 2.1 a: NI; 2.1b:PY; 2.2: PY. Note: No information provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were most likely aware of intervention allocation. Quote: "Due to the nature of the intervention, masking of 2.3: PN. Note: No evidence to suggest deviations from the intended intervention that arose because of the trial context. 2.4: NA. 2.5: NA. 2.6: Y. Note: Data was analysed on an intention‐to‐treat basis. 2.7: NA. | ||||||||||||
| 3.1a PY. Note: all 60 clusters that recruited participants were analyzed; 3.1b: PY. Note: 92 participants out of 1152 were lost at followup (8%). 3.2: NA. 3.3: NA; 3.4: NA. | ||||||||||||
| 4.1: PN. The outcome measure (CES‐D) is widely established and used across contexts. 4.2: PN. Note: No evidence to suggest that measurement or ascertainment of the outcome differed between intervention groups. 4.3a: NI; 4.3b: PY. Note: No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were most likely aware of the group allocation due to the nature of the intervention. Efforts were made to keep the data collection teams as masked to group status as possible. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned interventio could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note:No evidence to suggest otherwise, protocol and paper do not show discrepancies. 5.2: PN. Note:No evidence to suggest outcome selection. All outcomes mentioned in the protocol, methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the protocol, methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Rodriguez 2021 | | | | | | | ||||||
| 1.1: Y. Note: Title mentions RCT design. Quote: "Students who completed the web‐based baseline assessment measures were randomly assigned to a brief, 4‐week internet‐based mindfulness intervention (MIND), or to the intervention plus peer counselor support (MIND+)." 1.2: Y. Quote: "The randomization sequence was sourced through random.org [49], an automated, web‐based randomization service that generates randomness using atmospheric noise. Using random.org, two 25‐person blocks were used to randomize the participants into 2 equally‐sized groups." 1.3: PN. Note: No significant differences at baseline. | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention, participants and the interventionists were most likely aware of the allocation. 2.3: PN. Note: No evidence to suggest issues to implement intervention. 2.4: NA 2.5: NA 2.6: NI. Note: No explicit mention of the analysis used to estimate the effect of assignment to intervention. 2.7: N. Note: There was not a substantial impact of the failure to analyze participants in the group to which they were randomized. | ||||||||||||
| 3.1: PN. Note: Drop‐out: 31 of 54. 3.2: PN. Note: Proportions of lost to follow‐up between control and intervention are not comparable (n= 20 and n= 11). 3.3: NI; 3.4: NI. Note: No information about documented reasons that withdrawal are related to the outcome. | ||||||||||||
| 4.1: N. Note: The selected outcome method (PHQ‐9) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. 5.3: PN. Note: No evidence to suggest analyses selection. | ||||||||||||
| The study is judged to be at high risk of bias in one domain (3). | ||||||||||||
| Sangraula 2020 | | | | | | | ||||||
| 1a.1: PY. 1a.2: PY. Quote: "A randomization procedure was used, in which names of the two VDCs were written on cards and placed into a hat. The District Public Health Officer (DPHO) was to draw one card from the hat which would be allocated as the intervention arm VDC and the VDC in the remaining card would be allocated to the control arm. Program staff, including community psychosocial workers (CPSWs) and research assistants (RAs) were only assigned to either VDC after the random drawing to reduce risk of unblinding. 1a.3: PN. There is no evidencde to suggest problems in the randomization process 1b.1: PN. it appears the randomization was conducted prior to the recruitment of participants. 1b.2: PN. It is unlikely that the selection of inidividual participants was affected by knowledge of the intervention assigned to the cluster. 1b.3: PN. there is no evidence to indicate imbalances in baseline values to indicate differentail identification or recruitment of individual participants. | ||||||||||||
| 2.1 a:NI. 2.1b: PN. 2.2: PN. Quote: "CPSWs, RAs, trial participants, and local mhGAP trained health workers were blinded to the allocation of the study conditions. The VDCs of the two arms were separated by another VDC which worked as a buffer and physical barrier against contamination between the two arms. CPSWs in both arms were instructed not to disclose the treatment that any participants received except with their clinical supervisors. The trial statistician was blinded to treatment arm during analysis. " 2.3: NA. 2.4: NA. 2.5: NA. 2.6: PY. Data was analysed for most participants randomized. | ||||||||||||
| 3.1a: Y. all included clusters were analyzed; 3.1b: PY. data were available for nearly all participants within the cluster 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. The outcome measure (PHQ‐9) is widely established and used across contexts, included the one of the trial 4.2: PN. No evidence to suggest that measurement or ascertainment of the outcome differed between intervention groups. 4.3a:NI. 4.3b: PN. No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were blinded | ||||||||||||
| 5.1: PY. No evidence to suggest otherwise, protocol and paper do not show discrepancies. 5.2: PN. No evidence to suggest outcome selection. All outcomes mentioned in the protocol, methods and results were reported. 5.3: PN. No evidence to suggest analyses selection. All analyses mentioned in the protocol, methods and results were reported. | ||||||||||||
| The study is judged to be at low risk of bias for all domains. | ||||||||||||
| Song 2019 | | | | | | | ||||||
| 1.1: Y; 1.2: PY. Note: Block randomization, computer‐generated random list. Using opaque and sealed envelopes for group allocation code keeping. This was done by a person not involved in data collection. 1.3: N. Quote: "No considerable differences were observed between the intervention and control groups in terms of demographic characteristics, baseline outcome and mediating variables." | ||||||||||||
| 2.1: Y; 2.2: PY. Note: Due to the nature of the intervention participants were most likely aware of their allocation. 2.3: PN. Note: No evidence to suggest issues in recruitment and engagement nor failures to implement intervention 2.4: NA. 2.5: NA. 2.6: Y. Note: An ITT analysis was conducted. 2.7: NA. | ||||||||||||
| 3.1: PY. Note: Data at end line was missing, due to withdrawal or loss to follow‐up, for 29 participants (circa 24%). 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. Quote: "GDS is easy to administer and is a valid tool for assessing depressive symptoms in individuals with mild cognitive impairment"; The Cronbach’s alpha in this study was 0.784. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. Researchers involved in data collection were blinded to group allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: no evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Ward 2020 | | | | | | | ||||||
| 1.1: PY; 1.2: PY. Quote: "Randomization was conducted after data collection by an off‐site statistician with no other contact with the trial"." 1.3: PN. Baseline characteristics were overall comparable across groups as demonstrated by Table 1. | ||||||||||||
| 2.1: PY; 2.2: PY. Due to the nature of the intervention participants and personnel were moslt likely aware of their assigned intervention. 2.3: PY. There is some evidence to suggest possible contamination between the intervention and control group. Quote: "In addition, in the dense living environments of informal settlements, it is possible that there was contamination between intervention and control groups." 2.4: PY. Contamination could have affected the outcomes. 2.5: PY. Contamination is more liekly to affect the participants in the control group. 2.6: PY. No explicit mention of the analysis used to estimate the effect of assignment to intervention, but data was analyzed for almost all randomized participants. Quote: "For caregiver self‐report, the follow‐up rate was 97.0% at t1 and 91.9% at t". 2.7: NA. | ||||||||||||
| 3.1: PY. Data was availalble for almost all randomized participants. Quote: "For caregiver self‐report, the follow‐up rate was 97.0% at t1 and 91.9% at t" 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. The selected outcome method is widely validated and established across contexts and demonstrated good psychometric properties in the study 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY. Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did 4.5: PN. | ||||||||||||
| 5.1: PY. No evidence to suggest otherwise, methods and results do not show discrepancies.. 5.2: PN. No evidence to suggest outcome selection.. 5.3: PN. No evidence to suggest analyses selection. All analyses mentioned in the methods were reported.. | ||||||||||||
| The study is judged to be at high risk of bias in one domain (2). | ||||||||||||
| Yeomans 2010 | | | | | | | ||||||
| 1.1: Y; 1.2: NI.Note: Title mentions RCT design. No further information. 1.3: N. Quote: "There were no significant baseline differences between the three treatment groups across age, gender, ethnicity, symptoms, education level, traumatic events experienced, or on prior exposure to trauma discourse" | ||||||||||||
| 2.1: Y; 2.2: PY. Quote: "Participants (...) were informed of random allocation procedures". Due to the nature of the intervention, the interventionists were most likely aware of the allocation. 2.3: PN. Note: No evidence to suggest issues to implement intervention 2.4: NA. 2.5: NA. 2.6: NI. Note: No explicit mention of the analysis used to estimate the effect of assignment to intervention 2.7: NA. Note: There was not a substantial impact of the failure to analyse participants in the group to which they were randomized | ||||||||||||
| 3.1: Y. Note: Lost to follow‐up: 7 of 120 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome method (HSCL‐25) is widely validated and established across multiple languages. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in two domains (2 and 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Amador Buenabad 2020 | | | | | | | ||||||
| 1a.1: Y; 1b.2: Y. Note: Title mentions cluster RCT design. 1.3: PN. Note: Not significant differences at baseline 1b.1: PN. the participants were recruited after school‐randomization. Quote: ". Once schools agreed to engage in the research 1b.2: PN there is no evidence to suggest that selection was affected by knowledge of the intervention assigned to the cluster; 1b.3: PN. Some baseline imbalances are presents but compatible with chance | ||||||||||||
| 2.1a: PY. Quote: "written informed consent was obtained from all adult participants and parents of all nonadult participants. Written informed assent was obtained from the children" 2.2: PY. People delivering the intervention (teachers) were trained for doing that. They were most likely aware of their assigned intervention. 2.3: PN. Note: No evidence to suggest deviations from the intended intervention taht arose because of the trial context. 2.4: NA. 2.5: NA. 2.6: NI. Note: No information provided. 2.7: PN. Note: There is no reason to believe that it did. | ||||||||||||
| 3.1a: PY. 3.1b: PY. Dropout: 2 for Huellitas intervention, 0 for control group. 3.2: NA. 3.3: NA; 3.4: NA. | ||||||||||||
| 4.1: N. Quote: "The reliability for the present sample was.83" 4.2: PN. Note: no evidence to suggest differences in measurement between intervention and control groups. 4.3a: Y. Outcome assessors (participants) were most likeley aware that a trial was taking a place. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned interventio could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PN. Note: Protocol is available. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods and protocol were reported 5.3: PN. Note: No evidence to suggest analyses selection. All outcomes mentioned in the methods and protocol were reported. | ||||||||||||
| The study is judged to raise some concerns in two domains (2 and 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Berger 2018 | | | | | | | ||||||
| 1a.1: Y; 1a.2: PY. Quote: "The principal of this school randomply chose (by tossing a coin) to implement the ESPS (experimental group) in only one of the two classes in each of the 4th‐6th grades, while the other classes participated in Social Studies (SS, control group)." 1.3: N. Quote: "No significant differences were found at baseline (prior to the intervention) between the intervention and control group for any of the outcome measures" 1b.1: Y. Quote: "The ESPS program was presented to all the principals from the primary schools in the Meru district. (...). Since we had limited funds, we randomly chose (by tossic dice) one school for the study." | ||||||||||||
| 2.1a: PY; 2.1b: PY. Quote: "All the parents or guardians whose children were assigned to the ESPS consented to their children's participation in the study and expressed support for the program" 2.3: PN. Note: No evidence to suggest deviations from the intended intervention that arose because of the trial context. 2.4: NA. 2.5: NA. 2.6: NI. Note: Not enough information provided to give a judgemnet on this item. 2.7: PN. Note: No evidence to suggest that failure to analyzed participants in randomized groups could a substantial impact on the results. | ||||||||||||
| 3.1a:PY. No evidence to suggest cluster loss. 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome method (Spence Anxiety scale for children‐SCAS) is validated and established. Quote: "Cronbach alpha in current sample was 0.74)" 4.2: PN. No evidence to suggest that measurement or ascertainment of the outcome differed between intervention groups. 4.3a: NI; 4.3b: PY. No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were most likely aware of the group allocation due to the nature of the intervention. 4.4: PY; 4.5: PN. Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No protocol available. 5.2: PN. Note: No evdience to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: no evidence to suggest analyses selection. All outcomes mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in two domains (2 and 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Huang 2017 | | | | | | | ||||||
| 1a.1: PY; 1a.2: PY. 1.1 Quote: "the study of effectiveness outcomes applies a cluster randomized wait‐list controlled design in 10 Ugandan schools". 1.2: Randomization was done by a researcher that was unfamiliar with the study schools 1.3: NI. Note: No useful information is reported to evaluate this element 1b.1: N: participants were recruited after cluster allocation to intervention or control; 1b.2: NI ‐ No evidence to suggest that selection was influenced by knowledge of the assigned intervention (93% of elegible teachers participated) for teachers, but teachers selected children for inclusion in a non‐random fashion; 1b.3: PN: there were no significant differences for intervention groups. Overall: Some concerns | ||||||||||||
| 2.1a: NI; 2.1b: PY; 2.2: PY. Note: No information provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were most likely aware of intervention allocation. 2.3: PN. Note: No evidence to suggest deviations from the intended intervention that arose because of the trial context 2.4: NA. 2.5: NA. 2.6: Y. Note: Data was analysed on an intention‐to‐treat basis. 2.7: NA. | ||||||||||||
| 3.1a: PY ‐ all clusters were analyzed; 3.1b: PN: data was available for almost all teachers (84/86%), loss at follow‐up was greater for families/children (62‐69%) 3.2: PN. No evidence that the result was not biased by missing data 3.3: PN; 3.4: NA. Quote: "the followed and | ||||||||||||
| 4.1: PN. Note: The outcome measure is widely established and adapted for use in Nepal. 4.2: PN. Note: No evidence to suggest that. 4.3a: NI; 4.3b: PY. Note: no information provided on whether participants knew that they were in a trial, participants were most likely aware of the group allocation due to the nature of the intervention 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Chew 2018 | | | | | | | ||||||
| 1a.1: Y; 1a.2: PY. Note: Title mentions RCT design. Quote: "A member of the Data Management Services team at the University Medical center, Ultrecht will carry out the randomisation." 1.3: NI. Note: No useful information is reported to evaluate this element 1b.1: PY. Seems that participants were recruited after randomization. | ||||||||||||
| 2.1a: PY. 2.1b: PN. Quote: "We use a modified informed consent procedure to ensure that patients are unware of the two different intervention programmes". 2.3: PN. Note: No evidence to suggest deviations from the intended intervention that arose because of the trial context. 2.4: NA. 2.5: NA. 2.6: Y. Note: The analysis was carried out on an intention to treat basis. 2.7: NA. | ||||||||||||
| 3.1a: PY. All 10 clusters that recruited participants were analyzed. 3.2: PN. No evidence that the result was not biased by missing data. 3.3: PN; 3.4: NA. Note: Missingness in the outcome do not depend on its true value. | ||||||||||||
| 4.1: N. Note: The selected outcome method (Malay Brief Illness Perception Quetsionnaire, MBIPQ) is validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control 4.3a: PY.Outcome assessors (participants) were most likeley aware that a trial was taking a place. 4.4: NA. 4.5: NA. | ||||||||||||
| 5.1: PY. Note: Protocol is available. 5.2: PN. Note: No evidence to suggest outcome selection. Almost all outcomes mentioned in the methods and protocol were reported (It is not reported "heath‐care utilization hospital" that is a secondary outcome). 5.3: PN. Note: No evidence to suggest analyses selection. All outcomes mentioned in the methdis and protocol were reported. | ||||||||||||
| The study is judged to be at low risk of bias for all domains. | ||||||||||||
| Sangraula 2020 | | | | | | | ||||||
| 1a.1: PY. 1a.2: PY. Quote: "A randomization procedure was used, in which names of the two VDCs were written on cards and placed into a hat. The District Public Health Officer (DPHO) was to draw one card from the hat which would be allocated as the intervention arm VDC and the VDC in the remaining card would be allocated to the control arm. Program staff, including community psychosocial workers (CPSWs) and research assistants (RAs) were only assigned to either VDC after the random drawing to reduce risk of unblinding. 1a.3: PN. There is no evidencde to suggest problems in the randomization process 1b.1: PN. it appears the randomization was conducted prior to the recruitment of participants. 1b.2: PN. It is unlikely that the selection of inidividual participants was affected by knowledge of the intervention assigned to the cluster. 1b.3: PN. there is no evidence to indicate imbalances in baseline values to indicate differentail identification or recruitment of individual participants. | ||||||||||||
| 2.1 a:NI. 2.1b: PN. 2.2: PN. Quote: "CPSWs, RAs, trial participants, and local mhGAP trained health workers were blinded to the allocation of the study conditions. The VDCs of the two arms were separated by another VDC which worked as a buffer and physical barrier against contamination between the two arms. CPSWs in both arms were instructed not to disclose the treatment that any participants received except with their clinical supervisors. The trial statistician was blinded to treatment arm during analysis. " 2.3: NA. 2.4: NA. 2.5: NA. 2.6: PY. Data was analysed for most participants randomized. | ||||||||||||
| 3.1a: Y. all included clusters were analyzed; 3.1b: PY. data were available for nearly all participants within the cluster 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. The outcome measure (PHQ‐9) is widely established and used across contexts, included the one of the trial 4.2: PN. No evidence to suggest that measurement or ascertainment of the outcome differed between intervention groups. 4.3a:NI. 4.3b: PN. No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were blinded | ||||||||||||
| 5.1: PY. No evidence to suggest otherwise, protocol and paper do not show discrepancies. 5.2: PN. No evidence to suggest outcome selection. All outcomes mentioned in the protocol, methods and results were reported. 5.3: PN. No evidence to suggest analyses selection. All analyses mentioned in the protocol, methods and results were reported. | ||||||||||||
| The study is judged to be at low risk of bias for all domains. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Subgroup 9.1.1 Community workers | ||||||||||||
| Acarturk 2022 | | | | | | | ||||||
| 1.1: Y. Quote: "Participants were randomly assigned either to the Self‐Help Plus arm or to ECAU only, in a 1:1 ratio." 1.2: Y. Quote "The randomization schedule was generated by Castor Electronic Data Capture (EDC). (..) Research team members involved in 1.3: N. Note: No significant differences at baseline. | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention participants and the interventionists were most likely aware of the allocation. 2.3: PN. Note: No evidence to suggest issues to implement intervention. 2.4: NA. 2.5: NA. 2.6: Y. Quote: "We followed an intent‐to‐treat approach for analysis of primary and secondary outcomes" 2.7: NA. | ||||||||||||
| 3.1: PY. Quote: " The distribution of participants lost to follow‐up was similar between the study groups (15.53% vs. 14.06%)" 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome method (PHQ‐9) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: no evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Baumgartner 2021 | | | | | | | ||||||
| 1a.1: PY; 1a.2: PY. Researchers performed stratified constrained randomization. This was carried out through software. Quote: "Constrained randomization was implemented using the cvcrand Stata package by the US‐based team". 1a.3: NI. No useful information is reported to evaluate this element. 1b.1: PN. It appears the randomization was conducted prior to the recruitment of participants: 1b.2: PN. It is unlikely that the selection of individual participants was affected by knowledge of the intervention assigned to the cluster. 1b.3: PN. There were some baseline imbalances that were compatible with chance, there is no evidence to indicate differential identification or recruitment of individual participants. | ||||||||||||
| 2.1 a: NI; 2.1b: PY; 2.2: PY. No information was provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were most likely aware of intervention allocation. 2.3: PY. The low literacy of the lay providers could have impacted implementation fidelity. Quote: "This strategy supported the potential sustainability of the intervention but came with literacy challenges for delivering the iMBC content with full fidelity". 2.4: PN. No evidence to suggest deviations from the intended intervention would have influenced the outcomes. 2.5: NA. 2.6: PY. Data were analysed for most randomized participants. 2.7: NA. | ||||||||||||
| 3.1a: PY. All included clusters were analyzed. 3.1b: PY. Data were available for nearly all participants within the cluster. Quote: "303 (81%), 313 (83.7%), and 266 (71.1%) participants were followed up at the mini‐survey, imme‐ diate post‐intervention survey (follow‐up 1), and the 8‐month post‐ intervention survey (follow‐up 2) time points, respectively ". 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. The outcome measure is well established and reported good psychometric properties in this sample. 4.2: PN. No evidence to suggest differences in measurement between intervention and control. 4.3A: NI; 4.3b: PY. No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were most likely aware of the group allocation due to the nature of the intervention. 4.4: PY; 4.5: PN. Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. No evidence to suggest outcomes selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. No evidence to suggest analyses selection. All outcomes mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in two domains (2 and 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Chang 2015 | | | | | | | ||||||
| 1.1: Y; 1.2: PY. Note: Title mentions RCT design. Quote: "The evaluation was a cluster randomized trial conducted in Jamaica, Antigua, and St Lucia with public health center as the unit of randomization." 1.3: N. Note: Quote: "There were no significant differences between groups in enrollment characteristics" | ||||||||||||
| 2.1a: NI; 2.1b: PY; 2.2: PY. Note: No information provided on wheter participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were most likely aware of intervention allocation. 2.3: PY. Note: Quote: "Clinics were often noisy and crowded, and some mothers would have had 2.4: NA. 2.5: NA. 2.6: Y. Quote: "Analyses were by intention to treat" 2.7: NA. | ||||||||||||
| 3.1a: NI. No information on wheter data was available for all clusters. 3.2: PY. Quote: "Loss did not differ by group (16% of children were lost to follow‐up in the control and 14% were lost to follow‐up in the intevrention)" 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome measure (center for Epidemiologic Studies‐depression Scale) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurements between intevrention and control. 4.3a: PY; 4.3b: PY. Note: Interviewers and testers were blind to center assignment but participants (mothers) probably knew that they were in a trial and participants were most likely aware of the group allocation due to the nature of the intervention. Self‐assessment scale. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcome, but there is no reason to believe that it did | ||||||||||||
| 5.1: PY. Note: no evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to be at high risk of bias in one domain (2). | ||||||||||||
| Ferreira‐Vorkapic 2018 | | | | | | | ||||||
| 1.1: Y; 1.2: Y. Quote: "(...) volunteers were randomly allocated to one of the three experimental groups. A separate examiner coordinated the group distribution (and all statistical analysis was performed by an expert from another 1.3: N. Quote: "the baseline data showing no significant differences between the groups, except for education years, where the relaxation group showed a higher | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention participants and personnel were most likely aware of their assigned intervention 2.3: PN. Note: No evidence to suggest issues to implement intervention 2.4: NA. 2.5: NA. 2.6: NI. Note: No explicit mention of the analysis used to estimate the effect of assignment to intervention 2.7: NA. Note: There was not a substantial impact of the failure to analyse participants in the group to which they were randomized | ||||||||||||
| 3.1: Y. Note: Drop‐out: 0 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome method (BAI) is widely validated and established across multiple languages.ility in the study (0.9) 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in two domains (2 and 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Hinton 2021 | | | | | | | ||||||
| 1a.1: PY. 1a.2: Y. Quote: "Randomizationwas at the level of clusters, defined as geographic areas (ie, communes) served by commune health stations which have a population 1a.3: NI. No useful information is reported to evaluate this element 1b.1: N. Note: particpants were recruited after randomization; 1b.2: PN. Note: there is no evidence to suggest that selection of individual participants was affected by knoweldge of the intervention assigned to the cluster; 1b.3:N. Note: no significant differences at baseline were identified. | ||||||||||||
| 2.1a: NI; 2.1b: PY; 2.2: Y. Note: No information provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were most likely aware of intervention allocation. 2.3: PN. Note: No evidence to suggest deviations from the intended intervention that arose because of the trial context 2.4: NA. 2.5: NA. 2.6: Y. Note: Data was analysed on an intention‐to‐treat basis. 2.7: NA. | ||||||||||||
| 3.1a PY. Note: no evidence to suggest failure in analyzing any clyster; 3.1b: PY. Note: data was avaiable for 85% of participants at follow‐up. 3.2: NA. 3.3: NA; 3.4: NA. | ||||||||||||
| 4.1: PN. The outcome measure (CES‐D) is widely established and used across contexts. 4.2: PN. Note: no evidence to suggest that. 4.3a: NI; 4.3b: PY. Note: No information provided on whether participants knew that they were in a trial, participants were most likely aware of the group allocation due to the nature of the intervention. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned interventio could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: Y. Note: A pre‐determined statistical plan is avaiable and in line with procedures reported in the final manuscript 5.2: PN. Note:No evidence to suggest outcome selection. All outcomes mentioned in the protocol, methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the protocol, methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Lachman 2017 | | | | | | | ||||||
| 1.1: Y; 1.2:Y. Quote: "An external researcher not directly involved in the study conducted the randomization procedures remotely in Oxford, United 1.3:PN. | ||||||||||||
| 2.1:Y; 2.2:Y. Quote: "Although program implementers and participants were aware of their allocation status, researchers 2.3: PN. Note: No evidence to suggest issues to implement intervention 2.4:NA. 2.5:NA. 2.6:Y. Quote: "Data analyses were conducted with an intention‐to‐treat design" 2.7:NA. | ||||||||||||
| 3.1:Y. Note: 3 out of the 68 participants randomized were lost at follow‐up (2.9%). 3.2:NA. 3.3:NA. 3.4:NA. | ||||||||||||
| 4.1:PN. Note: The selected outcome method is widely validated and established, appropriate for the local context, and assessed in its psychometric properties for this sample. 4.2:PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3:Y. Note: Outcome assessors (participants) were most likely aware of intervention allocation. The research assistants administering the questionnaires were blinded. 4.4: PY; 4.5:PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did | ||||||||||||
| 5.1:PY. Note: No evidence to suggest otherwise, the protocol and manuscript do not show discrepancies. 5.2: PY. Note: A few of the secondary outcomes mentioned in the protocol (e.g. ASSIST, Grover‐Counter Scale of Cognitive Development) were not reported in this manuscript. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the protocol and methods were reported in the results. | ||||||||||||
| The study is judged to raise some concerns in two domains (4 and 5), but not to be at high risk of bias for any domain. | ||||||||||||
| Lachman 2020 | | | | | | | ||||||
| 1a.1: Y; 1a.2: Y. Quote: "Cluster randomisation was conducted at village level prior to baseline assessments, to reduce the likelihood of contamination between arms. An external researcher used concealed computer‐generated codes to randomly allocate eight villages into three treatment arms and a control arm "; "The implementing partner notified the participating families of their allocation status after baseline data collection." 1a.3: NI. No useful information is reported to evaluate this element. 1b.1: PN. Note: randomization was conducted prior to the recruitment of participants 1b.2: PN. Note: groups were comparable at baseline, see Table 1. | ||||||||||||
| 2.1a: NI; 2.1b: PY; 2.2: PY. Note: No information provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were aware of intervention allocation. 2.3: PN. Note: No evidence to indicate deviations from protocol. 2.4: NA. 2.5: NA. 2.6: Y. Note: Data was analysed on an intention‐to‐treat basis. 2.7: NA. | ||||||||||||
| 3.1a.: Y. Note: all included clusters were analyzed; 3.1b: PY. Note: data were available for nearly all participants within the cluster. Quote:"Study retention was considerably higher than anticipated with 94.8% adults (n=235/248), 87.5% children (n=154/176) and 87.8% in their early childhood (n=122/139) assessments completed at post‐treatment, with no differences between arms " 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. Note: The outcome measure is widely established and validated for use across different contexts. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3a: NI; 4.3b: PY. Note: No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were most likely aware of the group allocation due to the nature of the intervention. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcomes selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All outcomes mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Luoto 2020 | | | | | | | ||||||
| 1a.1: PY; 1a.2: PY. Quote: "Using a random 1a.3: NI. No useful information is reported to evaluate this element. 1b.1: Y. Note: from protocol it is clear that recruitment of both adults and children as well as baseline assessments, took place before randomization to study arms | ||||||||||||
| 2.1 a: NI; 2.1b:PY; 2.2: PY. Note: No information provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were most likely aware of intervention allocation. Quote: "Due to the nature of the intervention, masking of 2.3: PN. Note: No evidence to suggest deviations from the intended intervention that arose because of the trial context. 2.4: NA. 2.5: NA. 2.6: Y. Note: Data was analysed on an intention‐to‐treat basis. 2.7: NA. | ||||||||||||
| 3.1a PY. Note: all 60 clusters that recruited participants were analyzed; 3.1b: PY. Note: 92 participants out of 1152 were lost at followup (8%). 3.2: NA. 3.3: NA; 3.4: NA. | ||||||||||||
| 4.1: PN. The outcome measure (CES‐D) is widely established and used across contexts. 4.2: PN. Note: No evidence to suggest that measurement or ascertainment of the outcome differed between intervention groups. 4.3a: NI; 4.3b: PY. Note: No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were most likely aware of the group allocation due to the nature of the intervention. Efforts were made to keep the data collection teams as masked to group status as possible. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned interventio could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note:No evidence to suggest otherwise, protocol and paper do not show discrepancies. 5.2: PN. Note:No evidence to suggest outcome selection. All outcomes mentioned in the protocol, methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the protocol, methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Rodriguez 2021 | | | | | | | ||||||
| 1.1: Y. Note: Title mentions RCT design. Quote: "Students who completed the web‐based baseline assessment measures were randomly assigned to a brief, 4‐week internet‐based mindfulness intervention (MIND), or to the intervention plus peer counselor support (MIND+)." 1.2: Y. Quote: "The randomization sequence was sourced through random.org [49], an automated, web‐based randomization service that generates randomness using atmospheric noise. Using random.org, two 25‐person blocks were used to randomize the participants into 2 equally‐sized groups." 1.3: PN. Note: No significant differences at baseline. | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention, participants and the interventionists were most likely aware of the allocation. 2.3: PN. Note: No evidence to suggest issues to implement intervention. 2.4: NA 2.5: NA 2.6: NI. Note: No explicit mention of the analysis used to estimate the effect of assignment to intervention. 2.7: N. Note: There was not a substantial impact of the failure to analyze participants in the group to which they were randomized. | ||||||||||||
| 3.1: PN. Note: Drop‐out: 31 of 54. 3.2: PN. Note: Proportions of lost to follow‐up between control and intervention are not comparable (n= 20 and n= 11). 3.3: NI; 3.4: NI. Note: No information about documented reasons that withdrawal are related to the outcome. | ||||||||||||
| 4.1: N. Note: The selected outcome method (PHQ‐9) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. 5.3: PN. Note: No evidence to suggest analyses selection. | ||||||||||||
| The study is judged to be at high risk of bias in one domain (3). | ||||||||||||
| Sangraula 2020 | | | | | | | ||||||
| 1a.1: PY. 1a.2: PY. Quote: "A randomization procedure was used, in which names of the two VDCs were written on cards and placed into a hat. The District Public Health Officer (DPHO) was to draw one card from the hat which would be allocated as the intervention arm VDC and the VDC in the remaining card would be allocated to the control arm. Program staff, including community psychosocial workers (CPSWs) and research assistants (RAs) were only assigned to either VDC after the random drawing to reduce risk of unblinding. 1a.3: PN. There is no evidencde to suggest problems in the randomization process 1b.1: PN. it appears the randomization was conducted prior to the recruitment of participants. 1b.2: PN. It is unlikely that the selection of inidividual participants was affected by knowledge of the intervention assigned to the cluster. 1b.3: PN. there is no evidence to indicate imbalances in baseline values to indicate differentail identification or recruitment of individual participants. | ||||||||||||
| 2.1 a:NI. 2.1b: PN. 2.2: PN. Quote: "CPSWs, RAs, trial participants, and local mhGAP trained health workers were blinded to the allocation of the study conditions. The VDCs of the two arms were separated by another VDC which worked as a buffer and physical barrier against contamination between the two arms. CPSWs in both arms were instructed not to disclose the treatment that any participants received except with their clinical supervisors. The trial statistician was blinded to treatment arm during analysis. " 2.3: NA. 2.4: NA. 2.5: NA. 2.6: PY. Data was analysed for most participants randomized. | ||||||||||||
| 3.1a: Y. all included clusters were analyzed; 3.1b: PY. data were available for nearly all participants within the cluster 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. The outcome measure (PHQ‐9) is widely established and used across contexts, included the one of the trial 4.2: PN. No evidence to suggest that measurement or ascertainment of the outcome differed between intervention groups. 4.3a:NI. 4.3b: PN. No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were blinded | ||||||||||||
| 5.1: PY. No evidence to suggest otherwise, protocol and paper do not show discrepancies. 5.2: PN. No evidence to suggest outcome selection. All outcomes mentioned in the protocol, methods and results were reported. 5.3: PN. No evidence to suggest analyses selection. All analyses mentioned in the protocol, methods and results were reported. | ||||||||||||
| The study is judged to be at low risk of bias for all domains. | ||||||||||||
| Ward 2020 | | | | | | | ||||||
| 1.1: PY; 1.2: PY. Quote: "Randomization was conducted after data collection by an off‐site statistician with no other contact with the trial"." 1.3: PN. Baseline characteristics were overall comparable across groups as demonstrated by Table 1. | ||||||||||||
| 2.1: PY; 2.2: PY. Due to the nature of the intervention participants and personnel were moslt likely aware of their assigned intervention. 2.3: PY. There is some evidence to suggest possible contamination between the intervention and control group. Quote: "In addition, in the dense living environments of informal settlements, it is possible that there was contamination between intervention and control groups." 2.4: PY. Contamination could have affected the outcomes. 2.5: PY. Contamination is more liekly to affect the participants in the control group. 2.6: PY. No explicit mention of the analysis used to estimate the effect of assignment to intervention, but data was analyzed for almost all randomized participants. Quote: "For caregiver self‐report, the follow‐up rate was 97.0% at t1 and 91.9% at t". 2.7: NA. | ||||||||||||
| 3.1: PY. Data was availalble for almost all randomized participants. Quote: "For caregiver self‐report, the follow‐up rate was 97.0% at t1 and 91.9% at t" 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. The selected outcome method is widely validated and established across contexts and demonstrated good psychometric properties in the study 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY. Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did 4.5: PN. | ||||||||||||
| 5.1: PY. No evidence to suggest otherwise, methods and results do not show discrepancies.. 5.2: PN. No evidence to suggest outcome selection.. 5.3: PN. No evidence to suggest analyses selection. All analyses mentioned in the methods were reported.. | ||||||||||||
| The study is judged to be at high risk of bias in one domain (2). | ||||||||||||
| Yeomans 2010 | | | | | | | ||||||
| 1.1: Y; 1.2: NI.Note: Title mentions RCT design. No further information. 1.3: N. Quote: "There were no significant baseline differences between the three treatment groups across age, gender, ethnicity, symptoms, education level, traumatic events experienced, or on prior exposure to trauma discourse" | ||||||||||||
| 2.1: Y; 2.2: PY. Quote: "Participants (...) were informed of random allocation procedures". Due to the nature of the intervention, the interventionists were most likely aware of the allocation. 2.3: PN. Note: No evidence to suggest issues to implement intervention 2.4: NA. 2.5: NA. 2.6: NI. Note: No explicit mention of the analysis used to estimate the effect of assignment to intervention 2.7: NA. Note: There was not a substantial impact of the failure to analyse participants in the group to which they were randomized | ||||||||||||
| 3.1: Y. Note: Lost to follow‐up: 7 of 120 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome method (HSCL‐25) is widely validated and established across multiple languages. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in two domains (2 and 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Subgroup 9.1.2 Primary healthcare workers | ||||||||||||
| Asnani 2021 | | | | | | | ||||||
| 1.1: Y; 1.2: Y. Note: Quote: "A random numbers 1.3: PN. Note: Baseline characteristics were overall comparable across groups, significant differences were reported for 2 variables (compatible with chances) | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention participants and personnel were most likely aware of their assigned intervention 2.3: PN. Note: No evidence to suggest issues to implement intervention. Some changes were made to the protocol before recruitment. These are reported in the manuscript. 2.4: NA. 2.5: NA. 2.6: Y. Quote: "Analyses were performed based on intention to treat principles" 2.7: NA. | ||||||||||||
| 3.1: Y. Note: Data was availalble for all 64 randomized participants, there were no losses at the 6 months follow‐up. 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. Note: The selected outcome method (CES‐D) is widely validated and established and demonstrated good inter‐item reliability in the study (0.9) 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. Staff involved in data collection was blinded to the allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. Authors mention that the CES‐D was not initially planned to be an outcome of interest in the study, this was added in a second moment, but still before the beginning of recruitment. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Chaharrahifard 2021 | | | | | | | ||||||
| 1.1: Y. Note: Title mentions RCT design; Quote: "the samples were randomly assigned to two groups of received counseling (intervention) and non‐received counseling (control) by four‐sized randomized blocks." 1.2: Y. Quote: "The details of blocks were contained in asset of sealed. 1.3: N. Quote: "There was not significant difference between the demographic characteristics in the control and intervention group." | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention, participants and the interventionists were most likely aware of the allocation. 2.3: PN. No evidence to suggest issues to implement intervention. 2.4: NA 2.5: NA 2.5: NI. Note: No explicit mention of the analysis used to estimate the effect of assignment to intervention. 2.6: N. Note: There was not a substantial impact of the failure to analyse participants in the group to which they were randomized. | ||||||||||||
| 3.1: PY. Note: Drop‐out rate: 21%. 3.2: NA 3.3: NA 3.4: NA | ||||||||||||
| 4.1: N. Note: The selected outcome method (EPDS) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All outcomes mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in two domains (2 and 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Chew 2018 | | | | | | | ||||||
| 1a.1: Y; 1a.2: PY. Note: Title mentions RCT design. Quote: "A member of the Data Management Services team at the University Medical center, Ultrecht will carry out the randomisation." 1.3: NI. Note: No useful information is reported to evaluate this element 1b.1: PY. Seems that participants were recruited after randomization. | ||||||||||||
| 2.1a: PY. 2.1b: PN. Quote: "We use a modified informed consent procedure to ensure that patients are unware of the two different intervention programmes". 2.3: PN. Note: No evidence to suggest deviations from the intended intervention that arose because of the trial context. 2.4: NA. 2.5: NA. 2.6: Y. Note: The analysis was carried out on an intention to treat basis. 2.7: NA. | ||||||||||||
| 3.1a: PY. All 10 clusters that recruited participants were analyzed. 3.2: PN. No evidence that the result was not biased by missing data. 3.3: PN; 3.4: NA. Note: Missingness in the outcome do not depend on its true value. | ||||||||||||
| 4.1: N. Note: The selected outcome method (Malay Brief Illness Perception Quetsionnaire, MBIPQ) is validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control 4.3a: PY.Outcome assessors (participants) were most likeley aware that a trial was taking a place. 4.4: NA. 4.5: NA. | ||||||||||||
| 5.1: PY. Note: Protocol is available. 5.2: PN. Note: No evidence to suggest outcome selection. Almost all outcomes mentioned in the methods and protocol were reported (It is not reported "heath‐care utilization hospital" that is a secondary outcome). 5.3: PN. Note: No evidence to suggest analyses selection. All outcomes mentioned in the methdis and protocol were reported. | ||||||||||||
| The study is judged to be at low risk of bias for all domains. | ||||||||||||
| Escolar 2014 | | | | | | | ||||||
| 1.1: Y; 1.2: NI. Quote: "The respondents were then randomly assigned to either the experimental or control group" No information provided about allocation concealment. 1.3: PN. Note: No significant differences at baseline. | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention participants and the interventionists were most likely aware of the allocation. 2.3: NI. Note: No information provided. 2.4: NA. 2.5: NA. 2.6: NI. Note: No infomation provided. 2.7: N. Note: No drop‐out. | ||||||||||||
| 3.1: Y. Note: No missing data. 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome method (PHQ‐9) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: Y. Note: Only self‐reported measure. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No protocol available. No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in three domains (1, 2, 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Gao 2015 | | | | | | | ||||||
| 1.1: Y. Quote: "The randomization sequence was generated using a computerized random number generator and the allocation was kept in sealed opaque consecutively numbered envelopes." 1.2.: Y. Quote: "This simple randomization scheme was independently prepared by a research assistant who was not involved in determining eligibility, providing care, or assessing outcome." 1.3: N. Quote: "There were no significant differences between the two groups in their demographic, obstetric and related characteristics (...) There were also no significant differences between the two groups in their baseline measures" | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention, participants and the interventionists were most likely aware of the allocation. 2.3: PN. Note: No evidence to suggest issues to implement intervention. 2.4: NA 2.5: BA 2.6: Y. Note: Intention to treat analysis done. 2.7: NA | ||||||||||||
| 3.1: PY. Note: Drop‐out: 14 of 180. 3.2: NA 3.3: NA 3.4: NA | ||||||||||||
| 4.1: N. Note: The selected outcome method (EPDS) is widely validated in Chinese mothers and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4.: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All outcomes mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Song 2019 | | | | | | | ||||||
| 1.1: Y; 1.2: PY. Note: Block randomization, computer‐generated random list. Using opaque and sealed envelopes for group allocation code keeping. This was done by a person not involved in data collection. 1.3: N. Quote: "No considerable differences were observed between the intervention and control groups in terms of demographic characteristics, baseline outcome and mediating variables." | ||||||||||||
| 2.1: Y; 2.2: PY. Note: Due to the nature of the intervention participants were most likely aware of their allocation. 2.3: PN. Note: No evidence to suggest issues in recruitment and engagement nor failures to implement intervention 2.4: NA. 2.5: NA. 2.6: Y. Note: An ITT analysis was conducted. 2.7: NA. | ||||||||||||
| 3.1: PY. Note: Data at end line was missing, due to withdrawal or loss to follow‐up, for 29 participants (circa 24%). 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. Quote: "GDS is easy to administer and is a valid tool for assessing depressive symptoms in individuals with mild cognitive impairment"; The Cronbach’s alpha in this study was 0.784. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. Researchers involved in data collection were blinded to group allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: no evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Bias | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Study | Randomisation process | Deviations from intended interventions | Missing outcome data | Measurement of the outcome | Selection of the reported results | Overall | ||||||
| Subgroup 10.1.1 Community setting | ||||||||||||
| Acarturk 2022 | | | | | | | ||||||
| 1.1: Y. Quote: "Participants were randomly assigned either to the Self‐Help Plus arm or to ECAU only, in a 1:1 ratio." 1.2: Y. Quote "The randomization schedule was generated by Castor Electronic Data Capture (EDC). (..) Research team members involved in 1.3: N. Note: No significant differences at baseline. | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention participants and the interventionists were most likely aware of the allocation. 2.3: PN. Note: No evidence to suggest issues to implement intervention. 2.4: NA. 2.5: NA. 2.6: Y. Quote: "We followed an intent‐to‐treat approach for analysis of primary and secondary outcomes" 2.7: NA. | ||||||||||||
| 3.1: PY. Quote: " The distribution of participants lost to follow‐up was similar between the study groups (15.53% vs. 14.06%)" 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome method (PHQ‐9) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: no evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Baumgartner 2021 | | | | | | | ||||||
| 1a.1: PY; 1a.2: PY. Researchers performed stratified constrained randomization. This was carried out through software. Quote: "Constrained randomization was implemented using the cvcrand Stata package by the US‐based team". 1a.3: NI. No useful information is reported to evaluate this element. 1b.1: PN. It appears the randomization was conducted prior to the recruitment of participants: 1b.2: PN. It is unlikely that the selection of individual participants was affected by knowledge of the intervention assigned to the cluster. 1b.3: PN. There were some baseline imbalances that were compatible with chance, there is no evidence to indicate differential identification or recruitment of individual participants. | ||||||||||||
| 2.1 a: NI; 2.1b: PY; 2.2: PY. No information was provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were most likely aware of intervention allocation. 2.3: PY. The low literacy of the lay providers could have impacted implementation fidelity. Quote: "This strategy supported the potential sustainability of the intervention but came with literacy challenges for delivering the iMBC content with full fidelity". 2.4: PN. No evidence to suggest deviations from the intended intervention would have influenced the outcomes. 2.5: NA. 2.6: PY. Data were analysed for most randomized participants. 2.7: NA. | ||||||||||||
| 3.1a: PY. All included clusters were analyzed. 3.1b: PY. Data were available for nearly all participants within the cluster. Quote: "303 (81%), 313 (83.7%), and 266 (71.1%) participants were followed up at the mini‐survey, imme‐ diate post‐intervention survey (follow‐up 1), and the 8‐month post‐ intervention survey (follow‐up 2) time points, respectively ". 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. The outcome measure is well established and reported good psychometric properties in this sample. 4.2: PN. No evidence to suggest differences in measurement between intervention and control. 4.3A: NI; 4.3b: PY. No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were most likely aware of the group allocation due to the nature of the intervention. 4.4: PY; 4.5: PN. Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. No evidence to suggest outcomes selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. No evidence to suggest analyses selection. All outcomes mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in two domains (2 and 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Escolar 2014 | | | | | | | ||||||
| 1.1: Y; 1.2: NI. Quote: "The respondents were then randomly assigned to either the experimental or control group" No information provided about allocation concealment. 1.3: PN. Note: No significant differences at baseline. | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention participants and the interventionists were most likely aware of the allocation. 2.3: NI. Note: No information provided. 2.4: NA. 2.5: NA. 2.6: NI. Note: No infomation provided. 2.7: N. Note: No drop‐out. | ||||||||||||
| 3.1: Y. Note: No missing data. 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome method (PHQ‐9) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: Y. Note: Only self‐reported measure. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No protocol available. No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in three domains (1, 2, 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Lachman 2017 | | | | | | | ||||||
| 1.1: Y; 1.2:Y. Quote: "An external researcher not directly involved in the study conducted the randomization procedures remotely in Oxford, United 1.3:PN. | ||||||||||||
| 2.1:Y; 2.2:Y. Quote: "Although program implementers and participants were aware of their allocation status, researchers 2.3: PN. Note: No evidence to suggest issues to implement intervention 2.4:NA. 2.5:NA. 2.6:Y. Quote: "Data analyses were conducted with an intention‐to‐treat design" 2.7:NA. | ||||||||||||
| 3.1:Y. Note: 3 out of the 68 participants randomized were lost at follow‐up (2.9%). 3.2:NA. 3.3:NA. 3.4:NA. | ||||||||||||
| 4.1:PN. Note: The selected outcome method is widely validated and established, appropriate for the local context, and assessed in its psychometric properties for this sample. 4.2:PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3:Y. Note: Outcome assessors (participants) were most likely aware of intervention allocation. The research assistants administering the questionnaires were blinded. 4.4: PY; 4.5:PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did | ||||||||||||
| 5.1:PY. Note: No evidence to suggest otherwise, the protocol and manuscript do not show discrepancies. 5.2: PY. Note: A few of the secondary outcomes mentioned in the protocol (e.g. ASSIST, Grover‐Counter Scale of Cognitive Development) were not reported in this manuscript. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the protocol and methods were reported in the results. | ||||||||||||
| The study is judged to raise some concerns in two domains (4 and 5), but not to be at high risk of bias for any domain. | ||||||||||||
| Lachman 2020 | | | | | | | ||||||
| 1a.1: Y; 1a.2: Y. Quote: "Cluster randomisation was conducted at village level prior to baseline assessments, to reduce the likelihood of contamination between arms. An external researcher used concealed computer‐generated codes to randomly allocate eight villages into three treatment arms and a control arm "; "The implementing partner notified the participating families of their allocation status after baseline data collection." 1a.3: NI. No useful information is reported to evaluate this element. 1b.1: PN. Note: randomization was conducted prior to the recruitment of participants 1b.2: PN. Note: groups were comparable at baseline, see Table 1. | ||||||||||||
| 2.1a: NI; 2.1b: PY; 2.2: PY. Note: No information provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were aware of intervention allocation. 2.3: PN. Note: No evidence to indicate deviations from protocol. 2.4: NA. 2.5: NA. 2.6: Y. Note: Data was analysed on an intention‐to‐treat basis. 2.7: NA. | ||||||||||||
| 3.1a.: Y. Note: all included clusters were analyzed; 3.1b: PY. Note: data were available for nearly all participants within the cluster. Quote:"Study retention was considerably higher than anticipated with 94.8% adults (n=235/248), 87.5% children (n=154/176) and 87.8% in their early childhood (n=122/139) assessments completed at post‐treatment, with no differences between arms " 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. Note: The outcome measure is widely established and validated for use across different contexts. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3a: NI; 4.3b: PY. Note: No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were most likely aware of the group allocation due to the nature of the intervention. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcomes selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All outcomes mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Sangraula 2020 | | | | | | | ||||||
| 1a.1: PY. 1a.2: PY. Quote: "A randomization procedure was used, in which names of the two VDCs were written on cards and placed into a hat. The District Public Health Officer (DPHO) was to draw one card from the hat which would be allocated as the intervention arm VDC and the VDC in the remaining card would be allocated to the control arm. Program staff, including community psychosocial workers (CPSWs) and research assistants (RAs) were only assigned to either VDC after the random drawing to reduce risk of unblinding. 1a.3: PN. There is no evidencde to suggest problems in the randomization process 1b.1: PN. it appears the randomization was conducted prior to the recruitment of participants. 1b.2: PN. It is unlikely that the selection of inidividual participants was affected by knowledge of the intervention assigned to the cluster. 1b.3: PN. there is no evidence to indicate imbalances in baseline values to indicate differentail identification or recruitment of individual participants. | ||||||||||||
| 2.1 a:NI. 2.1b: PN. 2.2: PN. Quote: "CPSWs, RAs, trial participants, and local mhGAP trained health workers were blinded to the allocation of the study conditions. The VDCs of the two arms were separated by another VDC which worked as a buffer and physical barrier against contamination between the two arms. CPSWs in both arms were instructed not to disclose the treatment that any participants received except with their clinical supervisors. The trial statistician was blinded to treatment arm during analysis. " 2.3: NA. 2.4: NA. 2.5: NA. 2.6: PY. Data was analysed for most participants randomized. | ||||||||||||
| 3.1a: Y. all included clusters were analyzed; 3.1b: PY. data were available for nearly all participants within the cluster 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. The outcome measure (PHQ‐9) is widely established and used across contexts, included the one of the trial 4.2: PN. No evidence to suggest that measurement or ascertainment of the outcome differed between intervention groups. 4.3a:NI. 4.3b: PN. No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were blinded | ||||||||||||
| 5.1: PY. No evidence to suggest otherwise, protocol and paper do not show discrepancies. 5.2: PN. No evidence to suggest outcome selection. All outcomes mentioned in the protocol, methods and results were reported. 5.3: PN. No evidence to suggest analyses selection. All analyses mentioned in the protocol, methods and results were reported. | ||||||||||||
| The study is judged to be at low risk of bias for all domains. | ||||||||||||
| Song 2019 | | | | | | | ||||||
| 1.1: Y; 1.2: PY. Note: Block randomization, computer‐generated random list. Using opaque and sealed envelopes for group allocation code keeping. This was done by a person not involved in data collection. 1.3: N. Quote: "No considerable differences were observed between the intervention and control groups in terms of demographic characteristics, baseline outcome and mediating variables." | ||||||||||||
| 2.1: Y; 2.2: PY. Note: Due to the nature of the intervention participants were most likely aware of their allocation. 2.3: PN. Note: No evidence to suggest issues in recruitment and engagement nor failures to implement intervention 2.4: NA. 2.5: NA. 2.6: Y. Note: An ITT analysis was conducted. 2.7: NA. | ||||||||||||
| 3.1: PY. Note: Data at end line was missing, due to withdrawal or loss to follow‐up, for 29 participants (circa 24%). 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. Quote: "GDS is easy to administer and is a valid tool for assessing depressive symptoms in individuals with mild cognitive impairment"; The Cronbach’s alpha in this study was 0.784. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. Researchers involved in data collection were blinded to group allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: no evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Ward 2020 | | | | | | | ||||||
| 1.1: PY; 1.2: PY. Quote: "Randomization was conducted after data collection by an off‐site statistician with no other contact with the trial"." 1.3: PN. Baseline characteristics were overall comparable across groups as demonstrated by Table 1. | ||||||||||||
| 2.1: PY; 2.2: PY. Due to the nature of the intervention participants and personnel were moslt likely aware of their assigned intervention. 2.3: PY. There is some evidence to suggest possible contamination between the intervention and control group. Quote: "In addition, in the dense living environments of informal settlements, it is possible that there was contamination between intervention and control groups." 2.4: PY. Contamination could have affected the outcomes. 2.5: PY. Contamination is more liekly to affect the participants in the control group. 2.6: PY. No explicit mention of the analysis used to estimate the effect of assignment to intervention, but data was analyzed for almost all randomized participants. Quote: "For caregiver self‐report, the follow‐up rate was 97.0% at t1 and 91.9% at t". 2.7: NA. | ||||||||||||
| 3.1: PY. Data was availalble for almost all randomized participants. Quote: "For caregiver self‐report, the follow‐up rate was 97.0% at t1 and 91.9% at t" 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. The selected outcome method is widely validated and established across contexts and demonstrated good psychometric properties in the study 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY. Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did 4.5: PN. | ||||||||||||
| 5.1: PY. No evidence to suggest otherwise, methods and results do not show discrepancies.. 5.2: PN. No evidence to suggest outcome selection.. 5.3: PN. No evidence to suggest analyses selection. All analyses mentioned in the methods were reported.. | ||||||||||||
| The study is judged to be at high risk of bias in one domain (2). | ||||||||||||
| Subgroup 10.1.2 Other | ||||||||||||
| Asnani 2021 | | | | | | | ||||||
| 1.1: Y; 1.2: Y. Note: Quote: "A random numbers 1.3: PN. Note: Baseline characteristics were overall comparable across groups, significant differences were reported for 2 variables (compatible with chances) | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention participants and personnel were most likely aware of their assigned intervention 2.3: PN. Note: No evidence to suggest issues to implement intervention. Some changes were made to the protocol before recruitment. These are reported in the manuscript. 2.4: NA. 2.5: NA. 2.6: Y. Quote: "Analyses were performed based on intention to treat principles" 2.7: NA. | ||||||||||||
| 3.1: Y. Note: Data was availalble for all 64 randomized participants, there were no losses at the 6 months follow‐up. 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: PN. Note: The selected outcome method (CES‐D) is widely validated and established and demonstrated good inter‐item reliability in the study (0.9) 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. Staff involved in data collection was blinded to the allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. Authors mention that the CES‐D was not initially planned to be an outcome of interest in the study, this was added in a second moment, but still before the beginning of recruitment. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Chaharrahifard 2021 | | | | | | | ||||||
| 1.1: Y. Note: Title mentions RCT design; Quote: "the samples were randomly assigned to two groups of received counseling (intervention) and non‐received counseling (control) by four‐sized randomized blocks." 1.2: Y. Quote: "The details of blocks were contained in asset of sealed. 1.3: N. Quote: "There was not significant difference between the demographic characteristics in the control and intervention group." | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention, participants and the interventionists were most likely aware of the allocation. 2.3: PN. No evidence to suggest issues to implement intervention. 2.4: NA 2.5: NA 2.5: NI. Note: No explicit mention of the analysis used to estimate the effect of assignment to intervention. 2.6: N. Note: There was not a substantial impact of the failure to analyse participants in the group to which they were randomized. | ||||||||||||
| 3.1: PY. Note: Drop‐out rate: 21%. 3.2: NA 3.3: NA 3.4: NA | ||||||||||||
| 4.1: N. Note: The selected outcome method (EPDS) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All outcomes mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in two domains (2 and 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Chang 2015 | | | | | | | ||||||
| 1.1: Y; 1.2: PY. Note: Title mentions RCT design. Quote: "The evaluation was a cluster randomized trial conducted in Jamaica, Antigua, and St Lucia with public health center as the unit of randomization." 1.3: N. Note: Quote: "There were no significant differences between groups in enrollment characteristics" | ||||||||||||
| 2.1a: NI; 2.1b: PY; 2.2: PY. Note: No information provided on wheter participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were most likely aware of intervention allocation. 2.3: PY. Note: Quote: "Clinics were often noisy and crowded, and some mothers would have had 2.4: NA. 2.5: NA. 2.6: Y. Quote: "Analyses were by intention to treat" 2.7: NA. | ||||||||||||
| 3.1a: NI. No information on wheter data was available for all clusters. 3.2: PY. Quote: "Loss did not differ by group (16% of children were lost to follow‐up in the control and 14% were lost to follow‐up in the intevrention)" 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome measure (center for Epidemiologic Studies‐depression Scale) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurements between intevrention and control. 4.3a: PY; 4.3b: PY. Note: Interviewers and testers were blind to center assignment but participants (mothers) probably knew that they were in a trial and participants were most likely aware of the group allocation due to the nature of the intervention. Self‐assessment scale. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcome, but there is no reason to believe that it did | ||||||||||||
| 5.1: PY. Note: no evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods are reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to be at high risk of bias in one domain (2). | ||||||||||||
| Chew 2018 | | | | | | | ||||||
| 1a.1: Y; 1a.2: PY. Note: Title mentions RCT design. Quote: "A member of the Data Management Services team at the University Medical center, Ultrecht will carry out the randomisation." 1.3: NI. Note: No useful information is reported to evaluate this element 1b.1: PY. Seems that participants were recruited after randomization. | ||||||||||||
| 2.1a: PY. 2.1b: PN. Quote: "We use a modified informed consent procedure to ensure that patients are unware of the two different intervention programmes". 2.3: PN. Note: No evidence to suggest deviations from the intended intervention that arose because of the trial context. 2.4: NA. 2.5: NA. 2.6: Y. Note: The analysis was carried out on an intention to treat basis. 2.7: NA. | ||||||||||||
| 3.1a: PY. All 10 clusters that recruited participants were analyzed. 3.2: PN. No evidence that the result was not biased by missing data. 3.3: PN; 3.4: NA. Note: Missingness in the outcome do not depend on its true value. | ||||||||||||
| 4.1: N. Note: The selected outcome method (Malay Brief Illness Perception Quetsionnaire, MBIPQ) is validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control 4.3a: PY.Outcome assessors (participants) were most likeley aware that a trial was taking a place. 4.4: NA. 4.5: NA. | ||||||||||||
| 5.1: PY. Note: Protocol is available. 5.2: PN. Note: No evidence to suggest outcome selection. Almost all outcomes mentioned in the methods and protocol were reported (It is not reported "heath‐care utilization hospital" that is a secondary outcome). 5.3: PN. Note: No evidence to suggest analyses selection. All outcomes mentioned in the methdis and protocol were reported. | ||||||||||||
| The study is judged to be at low risk of bias for all domains. | ||||||||||||
| Ferreira‐Vorkapic 2018 | | | | | | | ||||||
| 1.1: Y; 1.2: Y. Quote: "(...) volunteers were randomly allocated to one of the three experimental groups. A separate examiner coordinated the group distribution (and all statistical analysis was performed by an expert from another 1.3: N. Quote: "the baseline data showing no significant differences between the groups, except for education years, where the relaxation group showed a higher | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention participants and personnel were most likely aware of their assigned intervention 2.3: PN. Note: No evidence to suggest issues to implement intervention 2.4: NA. 2.5: NA. 2.6: NI. Note: No explicit mention of the analysis used to estimate the effect of assignment to intervention 2.7: NA. Note: There was not a substantial impact of the failure to analyse participants in the group to which they were randomized | ||||||||||||
| 3.1: Y. Note: Drop‐out: 0 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome method (BAI) is widely validated and established across multiple languages.ility in the study (0.9) 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in two domains (2 and 4), but not to be at high risk of bias for any domain. | ||||||||||||
| Gao 2015 | | | | | | | ||||||
| 1.1: Y. Quote: "The randomization sequence was generated using a computerized random number generator and the allocation was kept in sealed opaque consecutively numbered envelopes." 1.2.: Y. Quote: "This simple randomization scheme was independently prepared by a research assistant who was not involved in determining eligibility, providing care, or assessing outcome." 1.3: N. Quote: "There were no significant differences between the two groups in their demographic, obstetric and related characteristics (...) There were also no significant differences between the two groups in their baseline measures" | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention, participants and the interventionists were most likely aware of the allocation. 2.3: PN. Note: No evidence to suggest issues to implement intervention. 2.4: NA 2.5: BA 2.6: Y. Note: Intention to treat analysis done. 2.7: NA | ||||||||||||
| 3.1: PY. Note: Drop‐out: 14 of 180. 3.2: NA 3.3: NA 3.4: NA | ||||||||||||
| 4.1: N. Note: The selected outcome method (EPDS) is widely validated in Chinese mothers and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4.: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All outcomes mentioned in the methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Hinton 2021 | | | | | | | ||||||
| 1a.1: PY. 1a.2: Y. Quote: "Randomizationwas at the level of clusters, defined as geographic areas (ie, communes) served by commune health stations which have a population 1a.3: NI. No useful information is reported to evaluate this element 1b.1: N. Note: particpants were recruited after randomization; 1b.2: PN. Note: there is no evidence to suggest that selection of individual participants was affected by knoweldge of the intervention assigned to the cluster; 1b.3:N. Note: no significant differences at baseline were identified. | ||||||||||||
| 2.1a: NI; 2.1b: PY; 2.2: Y. Note: No information provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were most likely aware of intervention allocation. 2.3: PN. Note: No evidence to suggest deviations from the intended intervention that arose because of the trial context 2.4: NA. 2.5: NA. 2.6: Y. Note: Data was analysed on an intention‐to‐treat basis. 2.7: NA. | ||||||||||||
| 3.1a PY. Note: no evidence to suggest failure in analyzing any clyster; 3.1b: PY. Note: data was avaiable for 85% of participants at follow‐up. 3.2: NA. 3.3: NA; 3.4: NA. | ||||||||||||
| 4.1: PN. The outcome measure (CES‐D) is widely established and used across contexts. 4.2: PN. Note: no evidence to suggest that. 4.3a: NI; 4.3b: PY. Note: No information provided on whether participants knew that they were in a trial, participants were most likely aware of the group allocation due to the nature of the intervention. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned interventio could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: Y. Note: A pre‐determined statistical plan is avaiable and in line with procedures reported in the final manuscript 5.2: PN. Note:No evidence to suggest outcome selection. All outcomes mentioned in the protocol, methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the protocol, methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Luoto 2020 | | | | | | | ||||||
| 1a.1: PY; 1a.2: PY. Quote: "Using a random 1a.3: NI. No useful information is reported to evaluate this element. 1b.1: Y. Note: from protocol it is clear that recruitment of both adults and children as well as baseline assessments, took place before randomization to study arms | ||||||||||||
| 2.1 a: NI; 2.1b:PY; 2.2: PY. Note: No information provided on whether participants knew that they were in a trial, due to the nature of the intervention they (as those delivering the intervention) were most likely aware of intervention allocation. Quote: "Due to the nature of the intervention, masking of 2.3: PN. Note: No evidence to suggest deviations from the intended intervention that arose because of the trial context. 2.4: NA. 2.5: NA. 2.6: Y. Note: Data was analysed on an intention‐to‐treat basis. 2.7: NA. | ||||||||||||
| 3.1a PY. Note: all 60 clusters that recruited participants were analyzed; 3.1b: PY. Note: 92 participants out of 1152 were lost at followup (8%). 3.2: NA. 3.3: NA; 3.4: NA. | ||||||||||||
| 4.1: PN. The outcome measure (CES‐D) is widely established and used across contexts. 4.2: PN. Note: No evidence to suggest that measurement or ascertainment of the outcome differed between intervention groups. 4.3a: NI; 4.3b: PY. Note: No information provided on whether participants knew that they were in a trial, participants (outcome assessors) were most likely aware of the group allocation due to the nature of the intervention. Efforts were made to keep the data collection teams as masked to group status as possible. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned interventio could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note:No evidence to suggest otherwise, protocol and paper do not show discrepancies. 5.2: PN. Note:No evidence to suggest outcome selection. All outcomes mentioned in the protocol, methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the protocol, methods and results were reported. | ||||||||||||
| The study is judged to raise some concerns in one domain (4), but not to be at high risk of bias for any domain. | ||||||||||||
| Rodriguez 2021 | | | | | | | ||||||
| 1.1: Y. Note: Title mentions RCT design. Quote: "Students who completed the web‐based baseline assessment measures were randomly assigned to a brief, 4‐week internet‐based mindfulness intervention (MIND), or to the intervention plus peer counselor support (MIND+)." 1.2: Y. Quote: "The randomization sequence was sourced through random.org [49], an automated, web‐based randomization service that generates randomness using atmospheric noise. Using random.org, two 25‐person blocks were used to randomize the participants into 2 equally‐sized groups." 1.3: PN. Note: No significant differences at baseline. | ||||||||||||
| 2.1: PY; 2.2: PY. Note: Due to the nature of the intervention, participants and the interventionists were most likely aware of the allocation. 2.3: PN. Note: No evidence to suggest issues to implement intervention. 2.4: NA 2.5: NA 2.6: NI. Note: No explicit mention of the analysis used to estimate the effect of assignment to intervention. 2.7: N. Note: There was not a substantial impact of the failure to analyze participants in the group to which they were randomized. | ||||||||||||
| 3.1: PN. Note: Drop‐out: 31 of 54. 3.2: PN. Note: Proportions of lost to follow‐up between control and intervention are not comparable (n= 20 and n= 11). 3.3: NI; 3.4: NI. Note: No information about documented reasons that withdrawal are related to the outcome. | ||||||||||||
| 4.1: N. Note: The selected outcome method (PHQ‐9) is widely validated and established. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. 5.3: PN. Note: No evidence to suggest analyses selection. | ||||||||||||
| The study is judged to be at high risk of bias in one domain (3). | ||||||||||||
| Subgroup 10.1.3 Refugee camp | ||||||||||||
| Yeomans 2010 | | | | | | | ||||||
| 1.1: Y; 1.2: NI.Note: Title mentions RCT design. No further information. 1.3: N. Quote: "There were no significant baseline differences between the three treatment groups across age, gender, ethnicity, symptoms, education level, traumatic events experienced, or on prior exposure to trauma discourse" | ||||||||||||
| 2.1: Y; 2.2: PY. Quote: "Participants (...) were informed of random allocation procedures". Due to the nature of the intervention, the interventionists were most likely aware of the allocation. 2.3: PN. Note: No evidence to suggest issues to implement intervention 2.4: NA. 2.5: NA. 2.6: NI. Note: No explicit mention of the analysis used to estimate the effect of assignment to intervention 2.7: NA. Note: There was not a substantial impact of the failure to analyse participants in the group to which they were randomized | ||||||||||||
| 3.1: Y. Note: Lost to follow‐up: 7 of 120 3.2: NA. 3.3: NA. 3.4: NA. | ||||||||||||
| 4.1: N. Note: The selected outcome method (HSCL‐25) is widely validated and established across multiple languages. 4.2: PN. Note: No evidence to suggest differences in measurement between intervention and control. 4.3: PY. Note: Outcome assessors (participants) were most likely aware of intervention allocation. 4.4: PY; 4.5: PN. Note: Knowledge of the assigned intervention could influence participant‐reported outcomes, but there is no reason to believe that it did. | ||||||||||||
| 5.1: PY. Note: No evidence to suggest otherwise, methods and results do not show discrepancies. 5.2: PN. Note: No evidence to suggest outcome selection. All outcomes mentioned in the methods and results were reported. 5.3: PN. Note: No evidence to suggest analyses selection. All analyses mentioned in the methods were reported. | ||||||||||||
| The study is judged to raise some concerns in two domains (2 and 4), but not to be at high risk of bias for any domain. | ||||||||||||
