Appraising and interpreting evidence

This branch covers judging the strength and relevance of evidence and interpreting what it supports. Its parts distinguish critical appraisal, bias and confounding, balanced conclusions, developing evidence bases and the particular demands of preclinical and lifestyle research. Choosing study designs, statistical methods, generating real-world evidence and constructing a manuscript discussion belong to neighbouring topics and branches.

• Critical appraisal of evidence — relevance, methodological strength, conduct, outcomes, design-specific checklists and source limitations
• Bias and confounding — validity, entry and retention, measurement, time, and the difference between aggregate and individual associations
• Balanced interpretation of findings — weighing evidence on both sides, reconciling divergent findings, proportionate conclusions and causal inference
• Writing while the evidence is still moving — provisional evidence, monitoring, publication, revision and reporting before a causal link is established
• Quality and reproducibility in preclinical research — replication, sources of irreproducibility, quality practices, transparency, publication bias, registration and exploratory or confirmatory work
• Evidence about diet and lifestyle exposures — exposure and design, converging evidence, causal inference, disagreement between designs and appraisal of lifestyle real-world data