Written and maintained by CASRAI Editorial Board
Last updated
The MOOSE checklist (Meta-analysis Of Observational Studies in Epidemiology) is the reporting guideline for systematic reviews and meta-analyses that pool results from observational studies — cohort, case-control, and cross-sectional designs — rather than randomized controlled trials. It was developed by a working group led by Donna Stroup following an April 1997 workshop in Atlanta, and published as “Meta-analysis of observational studies in epidemiology: a proposal for reporting” in JAMA in 2000 (283(15):2008–2012, PMID 10789670). MOOSE predates PRISMA by two decades and remains the reporting guideline specifically named by many epidemiology and public-health journals for observational meta-analyses, used alongside or instead of PRISMA depending on the journal’s instructions.
Last verified 2026-08-24 against the EQUATOR Network’s MOOSE reporting-guideline page and the original 2000 JAMA publication record.
The MOOSE Checklist: 35 Items, 6 Categories
MOOSE organizes 35 reporting items into six categories that address the specific sources of bias in observational data — confounding, exposure misclassification, and study heterogeneity — that a trial-focused checklist does not need to cover. The table below summarizes what each category covers; complete the numbered checklist itself from a primary source (several university libraries and journal publishers host reproductions of the original table) rather than from this summary.
| Category | What It Covers |
|---|---|
| Reporting of Background | The problem definition, hypothesis, study outcomes, study population(s), exposure/intervention, and study design and rationale for the review. |
| Reporting of Search Strategy | Qualifications of searchers, search strategy including time period covered and databases used, use of hand-searching and non-English-language literature, and contact with authors for additional data. |
| Reporting of Methods | Study inclusion/exclusion criteria, rationale for restricting studies by publication status or language, a data-abstraction process, assessment of confounding and study quality, statistical methods (weighting, handling of heterogeneity, sensitivity analysis) and their rationale. |
| Reporting of Results | A graphical summary of individual study results and the combined estimate, a table describing characteristics of each included study, and sensitivity analyses. |
| Reporting of Discussion | Quantitative assessment of bias (e.g. publication bias), justification for excluding studies, and assessment of the quality of included studies. |
| Reporting of Conclusions | Consideration of alternative explanations for the observed results, generalizability of the findings, guidelines for future research, and disclosure of funding source. |
The checklist is maintained in the EQUATOR Network’s library at its MOOSE reporting-guideline page, which links to the original JAMA article.
MOOSE vs. PRISMA: Which Applies
Both guidelines cover systematic reviews and meta-analyses, but they were built for different evidence bases. PRISMA 2020 is design-agnostic and now the more widely required guideline across medical and health-sciences journals generally, including for reviews that mix or focus on observational studies. MOOSE remains the specifically named requirement at a number of epidemiology, public-health, and specialty-society journals, particularly for reviews built entirely from observational data, because several of its items — assessment of confounding, exposure misclassification, and study-quality heterogeneity across non-randomized designs — have no direct PRISMA equivalent. Some journals ask authors to complete both checklists for an observational meta-analysis; check the specific journal’s Instructions for Authors rather than assuming one supersedes the other. A meta-analysis of observational studies commonly pairs MOOSE’s reporting requirements with the Newcastle-Ottawa Scale for appraising each included study’s methodological quality, since MOOSE specifies what to report about quality assessment without itself being a quality-appraisal tool. For the broader landscape of which guideline applies to which review type, see CASRAI’s EQUATOR Network guide; for reviews that cannot be pooled quantitatively at all, see Narrative Synthesis and SWiM.
Frequently Asked Questions
What does MOOSE stand for?
Meta-analysis Of Observational Studies in Epidemiology. It refers both to the working group that developed the checklist and the checklist itself.
How many items are in the MOOSE checklist?
35 items organized across six categories: reporting of background, search strategy, methods, results, discussion, and conclusions.
Is MOOSE the same as PRISMA?
No. PRISMA is a general reporting guideline for systematic reviews and meta-analyses regardless of the underlying study design; MOOSE was developed specifically for meta-analyses that pool observational (cohort, case-control, cross-sectional) studies and addresses reporting needs, like confounding and exposure misclassification, that are specific to non-randomized data.
When was MOOSE published and by whom?
In 2000, in JAMA (283(15):2008–2012), by a working group led by Donna Stroup following an April 1997 workshop in Atlanta.
Do I need both MOOSE and a quality-appraisal tool?
Yes, typically. MOOSE tells you what to report about how you assessed study quality and confounding; it is not itself a tool for scoring individual studies. The Newcastle-Ottawa Scale is the appraisal tool most commonly paired with a MOOSE-reported meta-analysis of observational studies.
Related reading: The Newcastle-Ottawa Scale (NOS) — how to score the quality of the individual observational studies your MOOSE-reported meta-analysis pools.
See also: Choosing a Critical Appraisal Tool for Your Study Design — a side-by-side comparison of CASP, AMSTAR 2, Newcastle-Ottawa, and RoB 2.
