Diagram of reliability and net gain: a new method for summarizing and displaying the results of controlled trials.

Sound evidence-based clinical and policy therapy recommendations depend on a process of acquiring literature, appraising it for study design and quality, and assessing its results in terms of relative benefits and harms. In this article we describe two summary metrics that have been developed to assist guideline developers to assess the importance of the quality/size of trials and the relative benefit and harm of treatments, and a related graphical method for displaying them together. To illustrate this graphic approach, we have used the evidence base for drug therapies for neuropathic pain. "Reliability" is calculated from the sample size and quality of trials, while the second metric, "Net Gain," is calculated from the relative benefit and harm of each trial. The graphical method differentiated the evidence for neuropathic pain drugs. This graphical output provides a simple and relatively easy way to understand summary of information on large amounts of trial data across a number of therapies, which may present an impression of the evidence at a glance even for a non-expert. The usefulness of this graphical approach needs to be tested in the development of future clinical guidelines.