COSMOS-E
COSMOS-E is a set of guidelines that helps researchers conduct systematic reviews and meta-analyses of observational studies that look at the causes or risk factors of diseases. It covers all steps, from defining the research question to analyzing results, with special attention to biases common in observational studies.
At a glance
Use when
Conducting systematic reviews and meta-analyses of observational studies that aim to assess causal relationships between exposures and health outcomes.
Avoid when
Reviewing interventions evaluated in randomized controlled trials or when the research question is not focused on etiology or risk factors.
Inputs
Observational studies of etiology (e.g., cohort, case-control studies) relevant to a specific exposure-outcome relationship.
Outputs
A structured systematic review and, optionally, a meta-analysis of observational evidence on etiology, including assessments of bias, heterogeneity, and causality.
How it works
COSMOS-E (Conducting Systematic Reviews and Meta-Analyses of Observational Studies of Etiology) provides comprehensive methodological guidance for systematic reviews of observational studies focused on etiology. It addresses key steps including formulating research questions, defining exposures and outcomes, assessing risk of bias (particularly confounding, selection bias, and information bias), and performing statistical analyses. It emphasizes careful interpretation of meta-analytic results, noting that increased precision does not correct for underlying bias, and highlights the importance of exploring heterogeneity to assess validity and causality. Developed by experts in meta-analysis and observational research, it underwent standard peer review.
- HTA domains
- Clinical Effectiveness, Safety
- Assumptions
- That systematic synthesis of observational data on etiology requires tailored methods distinct from those used for randomized trials, particularly regarding bias assessment and interpretation of causal inference.
- Strengths
- Provides comprehensive, step-by-step guidance specific to etiology-focused observational reviews,Developed by experts and peer-reviewed,Addresses critical methodological issues like confounding and heterogeneity
- Limitations
- Does not eliminate inherent biases in primary observational studies,Greater precision from meta-analysis does not correct for bias,May not be applicable to non-etiological research questions
- Also known as
- Conducting Systematic Reviews and Meta-Analyses of Observational Studies of Etiology
Questions this answers
- › How should research questions be framed in systematic reviews of observational studies of etiology?
- › What are appropriate methods for defining exposures and outcomes in such reviews?
- › How should risk of bias (e.g., confounding, selection bias, information bias) be assessed?
- › What are best practices for statistical analysis and meta-analysis in this context?
- › How should heterogeneity be explored and interpreted?
- › What are the limitations of combining observational studies in meta-analyses?
References & sources
Related methods
Similar by meaning
- ROBINS-E
- ISPOR Prospective Observational Studies to Assess Comparative Effectiveness Good Research Practices Task Force Report
- ISPOR Retrospective Database Analysis Good Research Practices — Part III (Analytic Methods)
- GRACE Checklist (Good ReseArch for Comparative Effectiveness)
- ISPOR Good Research Practices for Retrospective Database Analysis Checklist (Part I)
Beta record. Generated from the primary source via AI extraction and independent audit, pending final human review.