Table 1 Characteristics of included studies

Systematic review and meta-analyses

 Burke et al. (2019) [51]

Effectiveness

N = 23 (2007–2017); USA (n = 16), Europe (n = 5), Canada (n = 1)

RCT (n = 15), pre-post (n = 8)

Adults with any mental health problem (and comprised ≥ 50% of sample)

N = 6191

Individual, group, peer-run services

In-person

Settings: Inpatient; Community/ outpatient MH

EPHPP

Weak (10 studies), Moderate (9 studies), strong (4 studies)

Critically low

 Chien et al. (2019) [21]

Effectiveness

N = 13 (2004–2017); USA (n = 8), UK (n = 1), Germany (n = 1), Netherlands (n = 2), China (n = 1)

RCT (all studies)

People (majority between 18 and 65 years) with schizophrenia or similar serious mental illness

N = 2479

Individual, group

Structured (e.g. manualised interventions; psychoeducation) and unstructured (e.g. discussion on set topics)

Settings: Inpatient; Community/outpatient MH

Cochrane RoB Tool

 > half had unclear RoB for the majority of domains; ~ a quarter had high RoB for 4 domains

High

 Fang et al. (2022) [47]

Effectiveness

N = 16 (2000–2020); USA (n = 4), China (n = 4), Canada (n = 2), Pakistan (n = 2), India (n = 1), Zimbabwe (n = 1), Singapore (n = 1), Iran (n = 1)

RCT (all studies)

Pregnant women/women who gave birth within 1 year with diagnosis/risk of perinatal depression N = 3154

Group, individual, combination

In-person; phone; online; combination

Settings: NR

Cochrane RoB Tool

Overall RoB was low. One study had high RoB for random sequence generation, two studies had low RoB for blinding participants/personnel, three studies had high risk of other bias

Critically low

 Fuhr et al. (2014) [37]

Effectiveness

N = 14 (1995–2012); USA (n = 9), Canada (n = 4), Netherlands (n = 1)

RCT (all studies)

Adult service users with an SMI or depression diagnosis N = 3595

Individual, group

In-person; telephone

Structured (manual-based)

Settings: Community/outpatient MH care

Cochrane RoB Tool

6 RCTs had overall high RoB

Low

 Huang et al. (2020) [48]

Effectiveness

N = 10 (2000–2019); USA (n = 1), Canada (n = 3), China (n = 3), India (n = 1), Pakistan (n = 1), South Africa (n = 1)

RCT (all studies)

Pregnant women/women who gave birth within 1 year with diagnosis/risk of perinatal depression N = 3076

Group, individual, combination

In-person; telephone; combination (including internet)

Settings: Inpatient and crisis (incl. Hospital); Community/outpatient MH care; Participant’s home or workplace

Cochrane RoB Tool

65% had low RoB, approximately 35% had unclear RoB

Critically low

 Lloyd-Evans et al. (2014) [22]

Effectiveness

N = 18 (1982–2013); USA (n = 14), Canada (1), UK (n = 1), Netherlands (n = 1), Australia (n = 1)

RCT (all studies)

Adults with SMI diagnoses or those using secondary MH services N = 5597

Individual, group

In-person; online

Structured (manualised), unstructured (e.g. befriending), combination

Settings: Community/outpatient MH care

Cochrane RoB Tool

All but 2 studies exhibited some RoB (unclear or high)

Low

 Lyons et al. (2021) [35]

Effectiveness

N = 8 (2011–2018); USA (n = 7), Switzerland (n = 1)

RCT (all studies)

Adults with any mental health condition (including SMI) N = 2131

Group

Online; in-person

Structured (manualised, classes), unstructured (n = 1; mutual support)

Setting: NR

Cochrane RoB Tool

Low RoB (N = 2 studies); high RoB (N = 2); unclear RoB (N = 4)

Critically low

 Peck et al. (2023) [40]

Effectiveness

N = 17 (2009–2018); USA (n = 12), Canada (n = 1), Australia (n = 1), UK (n = 1), Netherlands (n = 1), Singapore (n = 1)

RCTs (n = 11), non-randomised controlled trial (n = 1), pre-test-post-test (n = 5)

Adults, majority diagnosed with schizophrenia, bipolar disorder or major affective disorders

N = 3189

Individual

Peer-delivered

self-management

interventions or self-management

education programmes

incorporating elements of peer-assisted recovery

Settings: NR

JBI critical appraisal for RCTs and Quasi-experimental trials

RCTs: low-moderate quality;

quasi-experimental trials: moderate quality

Critically low

 Pitt et al. (2013, 2013) [41, 42]

Effectiveness

N = 11 (1979–2011); USA (n = 9), Australia (n = 1), UK (n = 1)

RCT (all studies)

Adults with severe mental health diagnoses

N = 2796

Individual, group

Settings: community/outpatient MH care

Cochrane RoB Tool

Study quality moderate to low. Most studies had: unclear RoB for random sequence generation, allocation concealment, high RoB blinded outcome assessment, selective reporting

Low

 Smit et al. (2022) [52]

Effectiveness

N = 30 (2003–2020); USA (n = 18), UK (n = 2), Canada (n = 3), Netherlands (n = 3), Australia (n = 1), Singapore (n = 1), Switzerland (n = 1), Germany (n = 1)

RCT (all studies)

Adults with any mental health condition

N = 4597

Individual, group

In-person; online

Settings: NR

Cochrane RoB Tool

High RoB (N = 21 studies), some concerns for RoB (N = 6), low RoB (N = 3)

Low

 Sun et al. (2022) [43]

Effectiveness

N = 7 (N papers = 8) (2011–2021); USA (n = 5), Germany (n = 2), Switzerland (n = 1)

RCT (all studies)

People (age unspecified; final sample included adolescents and adults) with any mental health problem including mental health service users without reported diagnoses

N = 763

Group

In-person

Structured (e.g. Open, Honest, Proud classes)

Settings: NR

Cochrane RoB Tool

Studies were generally of low/moderate RoB

Critically low

 Wang et al. (2022) [44]

Effectiveness

N = 28 (2004–2020); USA (n = 18), China (n = 5), Canada (n = 2), Netherlands (n = 2), Germany (n = 1)

RCT (all studies)

Adults: individuals or family members with serious mental illness

N = 806 families

N = 6572 individuals

Individual, family

Settings: inpatient and crisis (incl. Hospitals); community/outpatient MH care; hostels

Cochrane RoB Tool

The majority of studies were at low risk of bias

Critically low

 White et al. (2020) [45]

Effectiveness

N = 19 (N papers = 23) (1995–2018); USA (n = 12), UK (n = 3), Canada (n = 1), Australia (n = 1), Germany (n = 1), Japan (n = 1)

RCT (all studies)

Adults using mental health services with any diagnoses

N = 3329

Individual

In-person; online; combination

Structured (e.g. workbooks) and unstructured (e.g. mentoring)

Settings: inpatient and crisis (incl. Hospital); community/outpatient MH care

Cochrane RoB Tool

Overall quality of trials (compared to previous reviews) is low to moderate

Critically low

Systematic review (without meta-analyses)

 Bassuk et al. (2016) [53]

Effectiveness

N = 9 (2005–2013)

USA

RCT (n = 4), quasi-experimental (n = 3), comparison group (n = 1), programme evaluation (no comparison) (n = 1)

People (age unspecified; final sample all adults) in recovery from addiction from alcohol and/or drugs

N = 6883

NR

Settings: Inpatient; Community/outpatient MH services; third sector

EPHPP

Methodologically strong (n = 2 studies), moderate (n = 2), weak (n = 5)

Critically low

 Chinman et al. (2014) [1]

Effectiveness

N = 20 (N papers = 24) (1995–2013)

International (countries not specified)

RCT (n = 11), quasi-experimental (n = 6), correlational or descriptive (n = 3)

Adults with SMI or co-occurring substance use disorders

N = 40,927

Individual

Peers added to traditional services, peers assuming regular provider positions, peers delivering structured curricula

Settings: Inpatient; Community/outpatient MH care

Criteria developed for the ‘assessing the evidence base series’, that this paper was a part of

Limited (several methodological limitations) (n = 16), adequate (few or minor methodological limitations) (n = 4)

Critically low

 du Plessis et al. (2020) [23]

Experiences

N = 24 (1998 – 2018); USA (n = 7), Canada (n = 2), Australia (n = 7), UK (n = 3), South Korea (n = 1), Hong Kong (n = 1), Unknown (n = 1) Multiple (n = 2)

Qualitative (n = 14), narrative (n = 6), mixed methods (n = 1), meta-synthesis (n = 1), literature review (n = 2)

PSWs (substance abuse or mental health)

N = 307 (1 paper did not report sample size)

NR

Mental health (majority) and substance abuse settings

No quality appraisal

Critically low

 Fortuna et al. (2020) [36]

Effectiveness

N = 30 (2005–2019); USA (n = 22), Australia (n = 5), Italy (n = 1), Japan (n = 1), Denmark (n = 1)

RCT (n = 11), quasi-experimental (n = 3), pre-post designs (n = 10), exploratory (n = 1), mixed methods (n = 1), qualitative (n = 2)

Adults with schizophrenia or bipolar disorder

N = 4642

Individual, group

Digital peer support

MQRS

High methodological quality (n = 6), low quality (n = 4)

Critically low

 Gaiser et al. (2021) [38]

Effectiveness

N = 23 (2013–2020);

USA

RCT (n = 1), quasi-experimental (n = 3), cohort analytic (n = 2), cross-sectional (n = 1), retrospective comparison group (n = 1), survey (n = 1)

Adults with mental health or substance use disorder or those with current or past use of MH or SUD services

N = 14,098

Individual, group

Telephone; In-person

Structured (following a manual/curriculum) and unstructured (without predetermined format – individualised participant needs)

Settings: Community/ outpatient MH care; Inpatient and crisis (including hospitals); Participant’s home, third sector

EPHPP

Weak (n = 12)

Critically low

 Ibrahim et al. (2020) [24]

Implementation

N = 53 (1995–2018); USA (n = 30), UK (n = 7), Australia (n = 5), Canada (n = 3), Republic of Ireland (n = 2), Belgium (n = 1), Germany (n = 1), Hong Kong (n = 1), Japan (n = 1), Netherlands (n = 1), Israel & USA (n = 1)

RCTs (n = 10)

Qualitative (n = 38), Cohort studies (n = 4)

Control studies (n = 1)

PSWs supporting adults with mental illness

N = NR

Individual, group, combination

Excluded online only

Structured (e.g. health coaching, a ‘recovery’ training course), unstructured (e.g. PSWs sharing empathy, insights and skills)

Settings: Community/outpatient MH care; Inpatient and crisis (incl. hospitals); third sector; other rehabilitation services

CASP

Good quality (n = 47), Fair quality (n = 1), Poor quality (n = 5)

Critically low

 Lewis & Foye (2022) [55]

Implementation, experiences

N = 10 (2006–2020); UK (n = 4), USA (n = 2), Australia (n = 2), Netherlands (n = 1), Australia, UK, USA, Canada (n = 1)

RCT (n = 6), quasi experimental (n = 4)

PSWs with lived experience of eating disorders

N = 73 (N not reported in 4 studies)

People with/at risk of eating disorders

N = 4878

NR

In-person; online

Various intervention content, e.g. sharing recovery narratives, providing guidance, deliver lessons in schools

Settings: Inpatient and crisis (incl. Hospitals); community/outpatient MH care; participant’s home; schools, third sector

JBI Checklist for qualitative research

5 qualitative studies assessed: 8/10 (n = 3), 7/10 (n = 1), 6/10 (n = 1)

Critically low

 Miyamoto & Sono (2012) [63]

Effectiveness, experiences

N = 51 (1988–2010)

International (countries not specified)

Qualitative studies (n = 19), reviews (n = 8), other (n = NR)

PSWs supporting adults with mental health difficulties

N = NR

NR

In-person, Setting: NR

No quality appraisal

Critically low

 Mutschler et al. (2022) [46]

Implementation

N = 19 (2007–2019); USA (n = 12), Australia (n = 5), Scotland (n = 1), England (n = 1)

Mixed methods (n = 7), qualitative (n = 10), quantitative (n = 1), discussion paper (n = 1)

Individuals (age unspecified) seeking services for mental health. Diagnoses NR

N = NR

Individual, group

Settings: inpatient and crisis (incl. Hospital); community/outpatient MH care; third sector; primary care

Studies rated on quality in terms of relevance to review, including attention to/use of implementation strategies

Quality rating: high (n = 8), medium (n = 11), low (n = 0)

Critically low

 Pellizzer & Wade (2022) [56]

Effectiveness

N = 11 (2014–2022)

International (countries not specified)

RCTs (n = 4), Case study/series (n = 7)

People (age unspecified; final sample included adults and young people) with eating disorders (N = 1326) or carers (N = 289)

Also included 2 families

Individual, combination

Peer-led or co-led/ adjunct treatment

Structured programmes and unstructured (flexible content)

Settings: NR

Cochrane RoB tool (RCTs)

For all 4 RCTs: low risk rated for most items, all had items rated high risk or unclear risk

Specified sub-selection of 2010 CONSORT guidelines

Mean study quality: 6.45/9

Critically low

 Reif et al. (2014) [54]

Effectiveness

N = 10 (1998–2011)

International (countries not specified)

RCTs (n = 2), Quasi-experimental (n = 4), pre-post (n = 4)

Adults with substance use disorders

N = 7203

Group, individual, combination

In-person, telephone

Structured (e.g. coaching, counselling, activities, informational support, handouts) and unstructured (open-ended contact)

Settings: various

Strength of the evidence (using criteria developed for the publication series)

The evidence for peer support met the minimum criteria for the moderate category

Critically low

 Triece et al. (2022) [50]

Effectiveness, implementation

N = 24 (2008–2021)

LMICs: Uganda (n = 3), Ethiopia (n = 1), Zimbabwe (n = 1), Libya (n = 1), South Africa (n = 1), India (n = 2), Pakistan (n = 2), Philippines (n = 1), China (n = 1)

Qualitative (n = 7), pre-post case series (n = 3), RCTs (n = 7), mixed methods (n = 6), quasi-experimental (n = 1)

Adults with common mental disorders

N = 4694

Individual, group, combined

Psychotherapeutic strategies e.g. psychoeducation, social/emotional support, problem-solving

Settings: Community, clinic based

JBI Checklist for Qualitative Research; MMAT

Case series were mostly low RoB. Mixed methods and qualitative design were overall low risk

Cochrane RoB Tool: RCTs were mostly low RoB

Critically low

 Vandewalle et al. (2016) [62]

Implementation

N = 18 (1998–2015); USA (n = 8), Canada (n = 3), England (n = 3), Australia (n = 2), New Zealand (n = 1), Netherlands (n = 1)

Qualitative (n = 15), mixed (n = 3)

PSWs (adults) employed in mental health services

N = 470

NR

Settings: Inpatient and crisis (incl. hospitals); Community/outpatient MH care

CASP for qualitative studies

Average of 25% of items were evaluated negatively in studies

Critically low

Systematic review and qualitative synthesis

 Bailie & Tickle (2015) [34]

Experiences

N = 8 (N papers = 10) (1996–2013); UK (n = 2), USA (n = 6), Australia (n = 1), Canada (n = 1)

Qualitative (all studies)

PSWs (for mental health)

N = 96

NR

Settings: various including community/outpatient MH services

CASP for qualitative studies

Variable study quality, scores ranged from lowest, 15 (n = 1 paper), to 34 (n = 1 paper)

Critically low

 Jones et al. (2014) [49]

Experiences

N = 5 (1995–2012); England (n = 2), Finland (n = 1), Canada (n = 2)

Qualitative (all studies)

Women with perinatal mental illness

N = 95

Group

NR

CASP for qualitative studies

Overall, studies were of a reasonable quality

Critically low

 Walker & Bryant (2013) [61]

Experiences

N = 25

(N papers = 27) (1994–2010); USA (n = 18), Canada (n = 4), Australia (n = 2), UK (n = 3)

Qualitative (n = 18), mixed methods (n = 5), case study (n = 2)

PSWs (n = 258), non-peer staff (n = 232), service users (n = 88)

Total N = 578

NR

Settings: Statutory mental health settings, settings that share leadership with statutory mental health settings (taken from inclusion criteria)

CASP

NR

Critically low

Scoping review

 Akerblom & Ness (2023) [60]

Effectiveness

N = 172 (2010–2021); USA (n = 75), Oceania (n = 36), GB (n = 24), Canada (n = 15), Europe excluding GB (n = 19), Asia (n = 8), Southern America (n = 1)

NR

Studies of mental health and substance use PSW roles (variety of stakeholders, e.g. PSWs, carers, non-peer staff)

N = 12,044

(N not reported in 45 studies)

NR

Settings: adult mental health and substance use services

No quality appraisal

Critically low

 Bowersox et al. (2021) [58]

Effectiveness

N = 84 (1968–2019);

International (countries not specified)

NR

Articles on mental health peer support for suicide prevention/crisis management with adults

N = NR

NR

Telephone; online; in-person

Setting: inpatient and crisis (incl. hospitals), community/outpatient MH services, third sector

Quality evaluated based on United States Preventative Services Task Force guidelines

3.6% highest quality, 31% lowest level

Critically low

 Ong et al. (2022) [39]

Implementation

N = 16 (2012–2021);

Asia, majority in India (n = 5) and Hong Kong (n = 4)

RCT (n = 1), mixed methods (n = 3), commentary/editorials (n = 4), thesis (n = 1), ethnographic (n = 1), qualitative (n = 5), cross-sectional (n = 1)

People (any age; final sample included adults and children) with any mental health condition (including SMI)

N = 528 participants

N = 82 service providers

Individual, group

In-person; telephone

Individual: unstructured conversations, development of recovery plans, outreach programmes

Group: discussions, structured activities, e.g. role-plays, exercise, homework

Settings: Community/outpatient MH care; inpatient and crisis (incl. Hospitals)

No quality appraisal

Critically low

 Schlichthorst et al. (2020) [57]

Effectiveness

N = 7 (N papers = 8) (2006–2019); USA (n = 3), Germany (n = 1), Germany/Austria (n = 1), China (n = 1), Australia (n = 1)

Cross-sectional (n = 1), qualitative (n = 1), descriptive (n = 3), RCT (n = 1), survey (n = 1)

People (age unspecified) who experience suicidality

N = NR

Individual, group

In-person; online, structured (e.g. PSWs provide training in community) and unstructured (e.g. 1–1 peer support with flexible frequency/duration)

Settings: inpatient and crisis (incl. Hospitals); community/outpatient MH care; schools; online

No quality appraisal

Critically low

 Viking et al. (2022) [64]

Experiences

N = 22 (N papers = 23) (2011–2021); UK (n = 8), Canada (n = 2), Switzerland (n = 1), Australia (n = 3), Belgium (n = 1), USA (n = 3), German (n = 3), Canada & Norway (n = 1), Norway & USA (n = 1)

Qualitative (n = 21), quantitative (n = 1)

Literature concerning PSWs in mental healthcare

PSW (n = 235), service users (n = 18), non-peer staff (n = 191), mixed (n = 247)

NR

Settings: Formal MH care settings

No quality appraisal

Critically low

 Zeng & McNamara (2021) [59]

Implementation

N = 28 (2006–2020); USA (n = 12), UK (n = 7), Australia (n = 8), Canada (n = 1)

Qualitative (n = 25), mixed methods (n = 3)

Mental health PSWs

N = NR

NR

Settings: statutory MH services; third sector

No quality appraisal

Critically low