Evaluating the Quality of Evidence
- Well-written articles must identify a clear and novel hypothesis, a relevant topic, valid study methods, and the overall importance of the research.
- Baseline level of quality is established by the hierarchy of study designs.
- Overall quality of research can be downgraded based on five criteria identified in the GRADE guidelines: imprecision, inconsistency, indirectness, publication bias, and lack of internal validity.
- Overall quality of research can be upgraded based on three criteria identified in the GRADE guidelines: when the size of the effect observed is very large, when a dose-response relationship exists, or when plausible confounders or other biases paradoxically increase confidence in the direction or magnitude of the signal.
- A basic understanding of statistics is helpful for the reader to determine if the author has applied the appropriate statistical tools to interpret and analyze the data correctly. Otherwise, the conclusions may not be valid.