The timing of Nelson and colleagues' article [1] could not have been better. We have recently been observing the bigger thrust toward evidence-based medicine. Although this is a healthy trend, the need to educate the users of the evidence is greater than ever before. The problem with numbers such as sensitivity, specificity, predictive value, and number needed to treat is that they may not bail us out when it comes to the patient sitting across the desk in our clinic. The problem with probabilities is that the final result depends on pretest probability. And pretest probability needs to be assessed by a human being on the basis of his or her knowledge, experience, and clinical sense. This article describes a useful, commonsense way of using the available evidence.

A cautionary note: Managed health care groups are gaining the financial muscle to do more and more studies whose sole objective is to support the focus on cost containment. It is possible that physicians in practice, especially those working for managed care groups, will be whipped with such evidence day in and day out. This may be done with the purpose of shaping physicians' practice style into one that makes better financial sense for the insurance companies. This is another reason for us to know how a given study applies to our patients.