Synthesis and Export
Stage 7: Synthesis
The Synthesis stage answers the question from the included studies: pooled where their results can be combined, narrative where they cannot, with the certainty of each finding.
Choose how to synthesise, then click Synthesise (or Synthesise again):
| Choice | What it does |
|---|---|
| Decide from the data | Meta-analysis when two or more studies report the same effect with a confidence interval; otherwise narrative. |
| Meta-analysis | Pools effect estimates. Falls back to narrative, and says why, if nothing can be pooled. |
| Narrative | A structured narrative synthesis, without pooling. |
The result is badged Meta-analysis or Narrative synthesis. Up to five syntheses are kept; choose an earlier one from the history list.
Findings by outcome
Each outcome has a card with its effect measure, how many studies reported an estimate, and a certainty badge. Where studies were pooled, it shows the pooled estimate with its 95% confidence interval and p-value, the number of studies, the model (random effects, DerSimonian–Laird, or fixed effect) and I² heterogeneity, with a forest plot. Marker size reflects each study’s weight, and the vertical line marks no effect. Open Why <certainty> certainty to see all five GRADE domains for the outcome.
The report
The written report has these sections: Summary, Methods, Results (Study selection, Included studies, Findings by outcome), Certainty of evidence, Limitations and Conclusion.
To change a section’s wording, click its Refine button and tell Medbot what to change. Its rewrite appears next to the Current wording; keeping it saves a revised synthesis (badged Revised wording, with the section marked Revised and its Previous wording available). Rewrites never change the pooled estimates or certainty: for different numbers, change the extracted data or the grades and synthesise again.
PRISMA flow
The synthesis includes a PRISMA 2020 flow of records: Records identified in PubMed, Duplicates removed, Records screened, Excluded on title and abstract, Assessed for eligibility, Excluded after assessment (broken down by reason) and Studies included, noting any records still awaiting a decision.
Export
Export the report
Export report (Markdown) on the Synthesis stage downloads the synthesis as a Markdown file, headed with the review’s title, question and the date of the synthesis. The report is generated by Medbot; check it before relying on it.
Export articles
More actions → Export articles (CSV) downloads every article in the review with these columns: PMID, Title, Authors, Journal, Publication date, DOI, URL, Screening decision, Screening decided by, Eligibility decision, Eligibility decided by, Exclusion reason, Text assessed, Study design, Sample size, Outcomes and GRADE certainty.