Authors: Ibrahim Alkassoum Salifou, Abdel Nacer Amoukou Issaka, Hassane Moussa Djongolé, Amadou Bouba Traore Hassane, Zeinabou Tondi Moussa, Abdou Atté Bachir, Djibrilla Baoua Bonkano, Roukayatou Karimou Issaka
Abstract
Background. Systematic reviews and meta-analyses are time-consuming, particularly at the study selection stage. Rayyan, a web and mobile AI-assisted screening platform, is widely used, including in francophone Africa, often without prior methodological training.
Objective. To provide a didactic review of the use of Rayyan in conducting and reporting a systematic review with meta-analysis, in line with international recommendations on responsible AI use.
Methods. Narrative synthesis of Rayyan’s official documentation (accessed September 2026), studies describing and evaluating the tool, reporting guidelines (PRISMA 2020, PRISMA-S) and the 2025 joint position of Cochrane, the Campbell Collaboration, JBI and the Collaboration for Environmental Evidence; illustration with an entirely fictitious worked example.
Key messages. Rayyan supports reference import, deduplication, blinded multi-reviewer screening and conflict resolution and, depending on the plan, data extraction, risk-of-bias assessment and PRISMA flow diagrams. Its relevance ranking requires at least 50 decisions, including 5 inclusions and 5 exclusions. In one evaluation, sensitivity reached 97-99% at the most cautious threshold, with a specificity of only 19-58%. AI should therefore prioritise, not replace, human decisions, and its use must be reported.
Conclusion. Used appropriately, Rayyan saves time and improves traceability; responsibility remains entirely human.
Keywords: systematic review; meta-analysis; Rayyan; artificial intelligence; study selection; PRISMA 2020; sub-Saharan Africa.