How to Review a Paper

This list is loosely adapted from Trisha Greenhalgh’s excellent series on evidence-based medicine [2].

• Why was the study done? If the aims are unclear, irrelevant or unimportant to you, then you can stop here
• What type of study did they do?
• What was the population in the study (and are they like my patients)?
• Has the test been compared to a true gold standard? Have they used surrogate or composite outcomes?
• Did they include an appropriate spectrum of patients? (Is there any spectrum bias?)
• Is there any work-up bias? (Did everyone who received the test also get the gold standard?)
• Is there any expectation bias (i.e. might knowledge of the patient’s status have influenced interpretation of the test)? How much blinding was there?
• Did the authors lose any patients?
• Is this test repeatable?
• Was the study big enough? (How wide are the confidence intervals?)
• Is it possible that competing interests may have influenced the study?
• What do the results mean? What is the sensitivity and specificity? What are the likelihood ratios?
• Is this study internally valid? (Do you accept the findings from the study?)
• Is this study externally valid? (Can I apply this to my own patients?)
• How much does this cost?