The medical review article revisited: has the science improved?

Abstract

Background: The validity of a review depends on its methodologic quality.

Objective: To determine the methodologic quality of recently published review articles.

Design: Critical appraisal.

Setting: All reviews of clinical topics published in six general medical journals in 1996.

Measurements: Explicit criteria that have been published and validated were used.

Results: Of 158 review articles, only 2 satisfied all 10 methodologic criteria (median number of criteria satisfied, 1). Less than a quarter of the articles described how evidence was identified, evaluated, or integrated; 34% addressed a focused clinical question; and 39% identified gaps in existing knowledge. Of the 111 reviews that made treatment recommendations, 48% provided an estimate of the magnitude of potential benefits (and 34%, the potential adverse effects) of the treatment options, 45% cited randomized clinical trials to support their recommendations, and only 6% made any reference to costs.

Conclusions: The methodologic quality of clinical review articles is highly variable, and many of these articles do not specify systematic methods.