How to make clinical decisions from statistics

Best practice health care depends on clinicians understanding and applying the results of trials in clinical practice. These results are most often presented in the language of statistics, which can be bewildering and even misleading for those of us who have only a limited understanding of the terms used in reporting statistics. This communication provides a simple explanation of some of the more common types of trials and their descriptors, presented in a way that should allow clinicians to understand, assess and apply the results. Confidence intervals are increasingly used in clinical research reports. They can be goldmines of information because they not only provide an estimate of statistical significance but also define the limits of the measured parameters within which the true estimate lies, the clinical importance of the findings and the power of the study to make possible these estimates. The 'number needed' to treat is a readily understood descriptor and if not described in the report it can be calculated simply from the figures already published. These and other statistical descriptors are explained with simple formulae, from which they may be calculated.