Step-by-step · ~14 min

Quickstart: from research question to report

Use this guide to set up a review, add studies, retrieve full text, screen records, extract outcomes, run an analysis, and generate a first report. It is intentionally practical: enough to get moving, with links to the deeper methodology pages when you need them.

Step 1 of 6 · ~2 min

Configure your protocol

Start with the question and eligibility criteria that every downstream step should respect. Pick a framework, define the population and outcomes, and add inclusion or exclusion rules before launching search.

  • Use PICO for intervention reviews and PEO for exposure, prevalence, and many qualitative questions.
  • Use PECO, PICOC, or PO for ecology and evolution reviews when those frameworks fit the question better.
  • Describe qualitative or mixed-methods questions in plain language so the review can use the right screening and synthesis path.
  • Lock major protocol decisions before formal screening whenever your review needs audit-ready reproducibility.
www.axelium.ai · Configure
Configure tab showing review question and outcomes
Fig 1.1Configure the review question, outcomes, and eligibility criteria before search.
Learn moreConcepts · questions and frameworks

Step 2 of 6 · ~2 min

Add studies and publications

Add records by identifier, import a reference-manager file, or let the search assistant draft and test a strategy from your protocol. Known relevant papers can be used as pearls to improve recall.

  • Add individual trials or publications by NCT ID, PMID, DOI, or keyword search.
  • Import RIS or BibTeX from EndNote, Zotero, Mendeley, PubMed, Embase, Web of Science, Ovid, or similar tools.
  • Use the search assistant for PubMed and ClinicalTrials.gov, with optional extra source coverage where available.
  • Use Discover for citation chasing once you have seed papers that should anchor the evidence base.
  • Expect new records to enter screening first; inclusion is decided by screening, not by search alone.
www.axelium.ai · Search
Search step showing ways to add studies
Fig 2.1Add records directly, import a library, or generate a search strategy from the protocol.

Step 3 of 6 · ~2 min

Retrieve full text and supplements

Move to the Full Text tab to collect PDFs, supplements, and other source documents. Axelium tries open-access retrieval first; when a document is unavailable, upload your own copy or use the browser companion for articles you can legally access.

  • Parsed documents are marked Parsed and become available for extraction or qualitative coding.
  • Supplements such as appendices, spreadsheets, and Word documents can be attached to the same study.
  • Publisher-blocked or paywalled records remain visible so you can upload the missing document manually.
  • Use bulk upload when a folder of PDFs needs to be matched back to several studies.
www.axelium.ai · Full text
Full Text tab showing document status
Fig 3.1Full Text tracks retrieval, parsing, supplements, and manual-upload needs for each study.
Learn moreBrowser companion

Step 4 of 6 · ~2 min

Screen records and resolve uncertainty

Screening applies your eligibility criteria to titles, abstracts, and full text. AI suggestions can speed up triage, but formal inclusion and exclusion decisions should be reviewed according to your protocol.

  • Capture exclusion reasons as you screen so PRISMA reporting is easier later.
  • Use custom screening guidance when repeated edge cases appear.
  • Use dual-review or two-human modes when your review requires independent decisions.
  • Resolve conflicts before treating the included set as final.
www.axelium.ai · Screening
Screening tab showing AI-assisted decisions
Fig 4.1Screening keeps suggested decisions, reviewer decisions, rationales, and exclusion reasons together.
Learn moreCollaboration guide

Step 5 of 6 · ~3 min

Extract data and run analysis

Once studies have eligible documents, extract outcome data and inspect the evidence trail. Then run the appropriate synthesis for the review: pairwise meta-analysis, qualitative synthesis, network meta-analysis, or ecology/evolution models when enabled and appropriate.

  • Review low-confidence values before pooling them.
  • Check that the effect measure matches the endpoint and protocol.
  • Use the analysis results, plots, and diagnostics as decision support, not a substitute for methodological review.
  • For NMA and ecology/evolution reviews, read the current-scope notes before interpreting results.
www.axelium.ai · Extraction
Extraction tab showing evidence-backed values
Fig 5.1Extraction keeps values connected to source evidence so they can be checked before analysis.
www.axelium.ai · Analysis
Forest plot showing pooled results
Fig 5.2The Analysis tab shows pooled effects, study-level estimates, and model summaries for reviewer interpretation.

Step 6 of 6 · ~3 min

Pin evidence and generate a report

Use the Evidence Board and Reports area to collect the plots, tables, decisions, and narrative notes that should support the final write-up. Choose the intended audience, generate a draft, then edit it like any scientific manuscript.

  • Pin plots, tables, and important explanations before generating a draft.
  • Check that the report reflects only reviewed evidence and approved decisions.
  • Export to DOCX or XLSX for editing and downstream sharing.
  • Use RevMan export for supported quantitative outcomes when handing off to Cochrane authoring tools.
www.axelium.ai · Reports
Report creation dialog
Fig 6.1Generate a report from selected evidence, then review and edit before sharing.
Learn moreHow it works · reports

Next: keep the review current

For living reviews, lock the protocol and search settings before enabling monitoring. New studies and evidence changes should flow back through screening, extraction, analysis, and reviewer approval rather than silently rewriting the review.

Read about living reviews