Evidence synthesis — foundations (opens in new tab)
The big picture: what a systematic review is, why pooling, where heterogeneity comes from, what a forest plot means.
Foundations
Start here if evidence synthesis is new to you. Question framing, the synthesis vocabulary, and the methodologist career arc.
Picking a meta-analysis topic (opens in new tab)
PICO and PECO question framing, search-feasibility checks, scope discipline, and avoiding the duplicate-review trap.
Meta-analysis methods (opens in new tab)
Estimators (DL, REML, Paule-Mandel, MH, Peto), the inverse-variance weighting story, and when to use which.
Becoming a methodologist (opens in new tab)
The methodologist's toolkit — running R locally, validating your browser results against canonical CRAN packages, and the career arc from first MA to methods paper.
Workflows
End-to-end paths for specific review types. Pair each course with the matching allmeta workflow.
Rapid reviews (opens in new tab)
Cochrane Rapid Reviews Methods Group (Garritty 2021) — single-reviewer screening with verification, streamlined extraction.
Pairs with:
Rapid review workflow
Prognostic reviews (opens in new tab)
PROGRESS framework, PICOTS, CHARMS extraction, QUIPS / PROBAST appraisal, GRADE-for-prognosis.
Pairs with:
Prognostic review workflow
Umbrella reviews (opens in new tab)
Reviews of systematic reviews: when to do one, AMSTAR-2 appraisal, handling overlap, conflicting reviews.
Pairs with:
AMSTAR-2
Qualitative evidence synthesis (opens in new tab)
Thomas-and-Harden thematic synthesis, GRADE-CERQual confidence, when narrative beats meta-aggregation.
Living reviews (opens in new tab)
Continuously updated evidence syntheses: trigger criteria, version control, sustaining a living review.
Observational evidence (opens in new tab)
Pooling cohorts and case-controls: ROBINS-I, confounding control, when adjusted estimates can be combined.
Pairs with:
ROBINS-I
Methods (deep dives)
Once you have the basics, go deeper into the techniques that turn a competent MA into a defensible one.
Advanced meta-analysis (opens in new tab)
Network meta-analysis, multilevel models, meta-regression, GOSH, Bucher indirect, transitivity and consistency.
IPD meta-analysis (opens in new tab)
One-stage vs two-stage, harmonising covariates, IPD vs aggregate-data trade-offs, K-M reconstruction.
Diagnostic test accuracy — "when the test lies" (opens in new tab)
Bivariate and HSROC pooling, threshold effects, QUADAS-2, Reitsma vs Rutter-Gatsonis.
Publication bias — detective work (opens in new tab)
Funnel asymmetry, Egger and Peters tests, trim-and-fill, PET-PEESE, Copas, p-curve. Pick the right tool for the bias mechanism.
Trial sequential analysis — when certainty kills (opens in new tab)
Sequential monitoring boundaries, O'Brien-Fleming alpha-spending, Required Information Size, futility.
Appraisal & reporting
Risk of bias, certainty of evidence, and how to write up an MA so reviewers don't bounce it.
Risk of bias — mastery (opens in new tab)
RoB 2 (Sterne 2019), ROBINS-I (Sterne 2016), ROBINS-E (Higgins 2024), QUADAS-2, when to use which.
GRADE certainty (opens in new tab)
The five downgrade domains, three upgrade considerations, NMA-specific CINeMA, GRADE-DTA, GRADE-CERQual.
Writing the meta-analysis (opens in new tab)
PRISMA 2020 reporting, PRISMA-NMA, structured abstracts, supplementary materials, what reviewers actually look for.
Applied tracks
Specialised paths: speed, decision-making, AI assistance, and reproducibility receipts.
Meta-sprint (opens in new tab)
How to run a meta-analysis in a week without sacrificing rigour. Templates, checklists, parallelisation.
Health technology assessment (opens in new tab)
Cost-effectiveness, decision analysis, MCID and non-inferiority margins. Built around an HTA Oman case study.
AI-assisted meta-analysis (opens in new tab)
Using LLMs responsibly: extraction, screening, what to verify, where to keep humans in the loop.
TruthCert — reproducibility receipts (opens in new tab)
HMAC-signed bundles for evidence reproducibility: what a receipt covers, what it doesn't, how reviewers verify.
New to evidence synthesis? Start with Foundations (opens in new tab), then pick a workflow. The full multi-language collection lives at synthesis-courses (opens in new tab).