This guide provides practical guidelines for conducting literature reviews, particularly rapid reviews, for those involved in healthcare policy dialogue. It describes five essential steps: formulate the research question using frameworks such as PESTLE, PICO, SPIDER or SPICE; search for publications in databases and grey sources (PubMed, Google Scholar, Cochrane Library, institutional sites); assess the quality of publications by checking validity, relevance, objectivity, timeliness and methodological rigor; combine publications by extracting data using a structured table (title, authors, year, location, objectives, type, key information) and organizing references with software such as Mendeley, Endnote Basic, Zotero; and finally contextualize and synthesize results to inform context-specific policy options. The document insists on good practices: clarifying the objective, involving knowledge users, searching in at least two databases, defining reproducible approaches for sorting and selection, using a second reviewer for verification and checking at least 10% of extractions. Narrative summaries are recommended, and results should be presented in a format accessible to decision-makers. A checklist lists critical tasks. Examples drawn from the Sub-Saharan African context serve as illustrations. The content is the responsibility of Results for Development, Duke, Feed the Children, Amref, Synergos, RAME, RESADE, CERRHUD and UHF. It also serves as a reminder that the quality of evidence must be assessed independently of the prestige of journals or the notoriety of international researchers.