Background: Co-production is promoted as an effective way of improving the quality of health and social care but the diversity of measures used in individual studies makes their outcomes difficult to interpret.

Objective: The objective is to explore how empirical studies in health and social care have described the outcomes of co-production projects and how those outcomes were measured.

Design and methods: A scoping review forms the basis for this systematic review. Search terms for the concepts (co-produc* OR coproduc* OR co-design* OR codesign*) and contexts (health OR 'public service* OR "public sector") were used in: CINAHL with Full Text (EBSCOHost), Cochrane Central Register of Controlled trials (Wiley), MEDLINE (EBSCOHost), PsycINFO (ProQuest), PubMed (legacy) and Scopus (Elsevier). There was no date limit. Papers describing the process, original data and outcomes of co-production were included. Protocols, reviews and theoretical, conceptual and psychometric papers were excluded. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses guideline was followed. The Mixed Methods Appraisal Tool underpinned the quality of included papers.

Results: 43 empirical studies were included. They were conducted in 12 countries, with the UK representing >50% of all papers. No paper was excluded due to the Mixed Methods Quality Appraisal screening and 60% of included papers were mixed methods studies. The extensive use of self-developed study-specific measures hampered comparisons and cumulative knowledge-building. Overall, the studies reported positive outcomes. Co-production was reported to be positively experienced and provided important learning.

Conclusions: The lack of common approaches to measuring co-production is more problematic than the plurality of measurements itself. Co-production should be measured from three perspectives: outputs of co-production processes, the experiences of participating in co-production processes and outcomes of co-production. Both self-developed study-specific measures and established measures should be used. The maturity of this research field would benefit from the development and use of reporting guidelines.