Peer support is sometimes formalized and offered as an intervention in mental health services and organizations. Evidence suggests that empowerment, self-efficacy, and internalized stigma are theoretically linked and implicated in the change processes involved in peer support. This review aimed to synthesize quantitative evidence published in the English language from trials that introduced any type of peer support intervention on the outcomes of empowerment, self-efficacy, and internalized stigma for those in receipt of peer support. Literature searches were conducted between November 2016 and April 2017 on CENTRAL, CINAHL, Clinicaltrials.gov, EMBASE, MEDLINE, PsycINFO, and Web of Science databases. Study quality was appraised. Results were integrated first through narrative synthesis. Where data was available, effect sizes were calculated and meta-analyses conducted when there were at least four randomized trials with similar characteristics. Twenty-three studies met inclusion criteria and could be separated into three broad categories: peer-led group interventions, one-to-one peer support, and peer-run services. Most were moderate to weak in quality. Meta-analyses were conducted for group interventions only. Results suggested that peer-facilitated time-limited group interventions can result in small but significant improvements in empowerment and self-efficacy compared with treatment as usual. Evidence was inconclusive for one-to-one peer support, peer-run services, and for internalized stigma. Areas for future research include: equivalence trials of group interventions with nonpeer facilitators, developing peer-led group interventions specific to the needs of people with particular mental health difficulties, high-quality research on one-to-one peer support and peer-run services, and research to understand the essential components and change mechanisms involved in peer support.