Evaluating the Quality of Evidence

Abstract

One of the most important skills a medical practitioner must develop is the ability to evaluate the evidence as evidence-based practice is the best way to provide patient care. Critical appraisal of an article requires a systematic approach to identify a clear and novel hypothesis, a relevant topic, valid study methods, and the overall importance of the research. This review identifies how to establish a baseline level of quality per the hierarchy of study designs. It then deconstructs each section of the standard IMRAD format article (Introduction, Methods, Results, Analysis, and Discussion), including a brief discussion of statistical methods. Finally, it shows how to apply some of the evaluation methods of the GRADE guidelines that were devised specifically to provide a process for determining quality of evidence through modifiers reflecting priorities in clinical decision making. In particular, the overall quality can be downgraded based on five criteria: imprecision, inconsistency, indirectness, publication bias, and lack of internal validity. In contrast, quality can be upgraded when the size of the effect seen 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. Taken together, a final assessment of quality may be applied, and the practitioner may accept the research for inclusion into practice or reject it as low-quality evidence. Both are examples of appropriate evidence-based practice, and both result in better patient care. Key words: appraisal, evaluate, evidence, grade, statistics