research-synthesis

# Research Synthesis Skill Use this skill when conducting literature reviews, synthesizing medical research, or evaluating clinical evidence. ## When to Use - Systematic literature searches - Evidence synthesis for clinical questions - Quality assessment of studies - Writing literature reviews - Translating research to clinical practice ## Research Synthesis Framework ### 1. Question Formulation (PICO) Structure the clinical question: - **P**opulation: Who is the patient/population? - **I**ntervention: What is the exposure/treatment? - **C**omparison: What is the alternative? - **O**utcome: What outcomes are important? ### 2. Search Strategy - Define search terms - Select databases (PubMed, Cochrane, Embase) - Include/exclude criteria - Document the search process ### 3. Study Evaluation Assess each study for: - Study design - Sample size - Follow-up duration - Bias risk - Funding sources ### 4. Evidence Synthesis Combine findings: - Forest plots when appropriate - Heterogeneity assessment - Sensitivity analyses - Subgroup analyses ### 5. Grading Rate evidence quality: ``` GRADE System: - High: Further research unlikely to change confidence - Moderate: Further research may change confidence - Low: Further research likely to change confidence - Very Low: Any estimate is very uncertain ``` ## Evidence Hierarchy | Level | Quality | Type | |-------|---------|------| | 1a | High | Systematic review of RCTs | | 1b | High | Individual RCT with narrow CI | | 2a | Moderate | Systematic review of cohort | | 2b | Moderate | Individual cohort study | | 3a | Low | Systematic review of case-control | | 3b | Low | Individual case-control | | 4 | Very Low | Case series | | 5 | Very Low | Expert opinion | ## Critical Appraisal ### RCTs (CONSORT) - Randomization method - Allocation concealment - Blinding - Follow-up - ITT analysis ### Observational (STROBE) - Selection bias - Confounding - Exposure measurement - Outcome assessment - Follow-up adequacy ### Systematic Reviews (PRISMA) - Search methodology - Study selection - Quality assessment - Synthesis method ## Output Structure ``` Literature Summary: Clinical Question: [PICO] Search Results: - Databases searched - Studies identified - Studies included Evidence Summary: [Study 1]: Design, n, results, quality [Study 2]: Design, n, results, quality Synthesis: - Overall effect estimate - Heterogeneity (I²) - Publication bias Recommendations: - Strength of recommendation - Clinical applicability - Research gaps ``` ## Reporting Standards - CONSORT: Randomized trials - STROBE: Observational studies - PRISMA: Systematic reviews - MOOSE: Meta-analyses of observational - CARE: Case reports ## Best Practices - Pre-register protocols - Document search strategy - Include flow diagrams - Rate evidence quality - Acknowledge limitations - Provide clinical context - Cite properly (PMID, DOI)