Methodology
Most evidence syntheses are within reach of the researchers who conceived them. These are the working methods Meridian applies, written out in full, so that a review can be carried through without external support.
Available now
7 titles for evidence-synthesis practice
Methodology
Lifetime updates included
Methodology
Lifetime updates included
Complete collection
Lifetime updates included
The Cochrane Handbook tells you what a systematic review must contain and how the analysis is performed. It does not tell you how to run one from a standing start, and that is where the time goes.
You are told to register a protocol. You are not told how to choose between PROSPERO and INPLASY, or what to do when neither fits. You are told to search comprehensively. You are not told that each database has its own syntax, and that a strategy which works in MEDLINE must be rebuilt for Embase. PRISMA specifies what belongs in the flow diagram, but not that deduplicating several thousand records by hand is a week you will not get back, or that EPPI-Reviewer handles it considerably better than Zotero.
The same is true of analysis. Mean difference or standardised mean difference. Frequentist or Bayesian. REML or Hartung-Knapp. Fixed or random effects. MetaInsight, R, RevMan, GRADEpro. Every one of these is documented somewhere, and almost nowhere together.
These guides are that assembly, written from having done it: the sequence, the tool that actually works at each step, what reviewers look for, and the mistakes we have already made.
Systematic review methodology is procedural, not arcane. What separates a defensible synthesis from a rejected one is usually a sequence of decisions made in the right order, and those decisions can be written down.
A review conducted first hand leaves you able to defend every choice at peer review, extend the analysis, and repeat the method on the next question. That capability does not transfer when the work is delegated.
Nobody understands the clinical or theoretical context of a question better than the researcher who framed it. Method can be taught; that context cannot be outsourced.
Some projects exceed what is practical alone: a network meta-analysis under a submission deadline, a Bayesian model outside your group’s expertise, or a review whose scope has grown beyond the time available. Meridian exists for those. Reaching that point is a matter of scope, not of ability.
