Shail M Govani, Peter D R Higgins

While the number of clinical trials performed yearly is increasing, the application of these results to individual patients is quite difficult. This article reviews key portions of the process of applying research results to clinical practice. The first step involves defining the study population and determining whether these patients are similar to the patients seen in clinical practice in terms of demographics, disease type, and disease severity. The dropout rate should be compared between the different study arms. Design aspects, including randomization and blinding, should be checked for signs of bias. When comparing studies, clinicians should be aware that the outcomes being studied may vary greatly from one study to another, and some outcomes are much more reliable and valuable than others. The definition of clinical response should also be scrutinized, as it may be too lenient. Surrogate outcomes should be viewed cautiously, and their use should be well justified. Clinicians should also note that statistical significance, as defined by a P-value cutoff, may be the result of a large sample size rather than a clinically significant difference. The treatment effect can be estimated by calculating the number needed to treat, which will demonstrate whether changes in clinical practice are worthwhile. Finally, this article discusses some common issues that can arise with figures.