Review overview — Systematic review with pairwise meta-analysis

Research question

Define the review question, intended decision and eligibility framework. Use PICO for intervention comparisons, PCC for scoping, or test/condition/clinical-role definitions for accuracy reviews.

Protocol

Write the planned eligibility, search, selection, extraction, appraisal and synthesis methods before conducting the review. Record registration where eligible, or a public protocol location, and date subsequent amendments.

Registration where applicable

Check the registry’s current eligibility and register prospectively where applicable. Record the identifier and timing; if ineligible, document a public protocol location and rationale.

Search

Choose information sources and supplementary search routes, retain reproducible strategies and dates, and plan an update search before final interpretation.

Select studies

Pilot eligibility criteria, record reviewer decisions and resolve disagreements. Link multiple reports to their underlying study or evidence source.

Extract data and assess study bias

Pilot fields, definitions and source-location capture. Verify important entries and retain study/report linkage. Use linked detailed flows for extraction and any associated appraisal.

Synthesize

Apply the prespecified synthesis strategy with documented deviations. Check compatibility, dependence and uncertainty; use structured synthesis when the intended pooling is not defensible.

Assess robustness and certainty

Examine justified alternative assumptions and missing evidence. Assess certainty or confidence for the relevant outcome or comparison, explaining every judgment rather than deriving it from a single test.

Report

Report methods, deviations, selection, findings and limitations using the relevant reporting guidance. Separate planned table shells from verified results and archive reproducible materials where permitted.

Review overview

End flow

Protocol and decision register

Record the protocol, approvals and registration where applicable. Preserve planned decisions and amendments separately from results.

Protocol and rationale evidence

Attach the research rationale, protocol, registration and supporting evidence. Leave unverified citations empty.