How to Synthesize Evidence in a Narrative Review

Synthesis means bringing findings together to answer a question: comparing studies, explaining agreement and conflict, weighing quality and drawing a conclusion. A summary lists what each study found; a synthesis tells the reader what the body of evidence shows.

A method

1. Extract key details into a table: study, design, population, intervention/exposure, outcome, result, limitations.
2. Group studies by theme or question.
3. Compare within groups: do results agree? If not, what differs (population, dose, design, measurement)?
4. Weigh by quality and relevance: give more weight to studies with lower risk of bias and larger, more precise samples.
5. Draw a statement for each theme with a confidence level.
6. Integrate across themes in the discussion.

Sentence patterns that synthesise

• "Three randomised trials found benefit (refs), whereas two observational studies did not (refs); the differences may reflect population and follow-up."
• "Evidence is limited to small studies with high risk of bias, so conclusions are tentative."

Illustrative example

Invented for teaching. Weak: "Smith found X. Lee found Y. Chen found Z." Better: "Trials in younger adults consistently reported benefit (Smith; Lee), whereas the only trial in older adults found no effect (Chen), which may reflect lower adherence."

Presenting synthesis visually

An evidence map or table that lists themes in rows and sources in columns (with a note on design and quality) makes your synthesis checkable. Readers can see how many studies support each theme and how strong they are. Use the table in the text as a summary, with the details in a supplement, and make sure the narrative explains the pattern rather than repeating every cell.

Common pitfalls

• Study-by-study paragraphs.
• Treating all studies as equal.
• Not explaining conflicting findings.

Key takeaways

• Compare, weigh and explain.
• State your confidence in each conclusion.

This page is general educational information about research methods. It is not medical, legal or regulatory advice; follow your institution's ethics board and your target journal's instructions.