Systematic reviews of diagnostic tests: a guide to methods and application

This chapter describes the basic steps in a systematic review to evaluate test accuracy, and the threats to validity of reviews inherent in each step. First, the problems to be addressed are specified in the form of well-structured questions (Step 1). This is a key step, as all other aspects of the review follow directly from the questions. Second, thorough literature searches are conducted to identify potentially relevant studies that shed light on the questions (Step 2). This is one essential feature that makes a review systematic. Third, the quality of the selected studies is assessed (Step 3). Fourth, the evidence concerning study characteristics and results is summarized, and differences between studies are explored (Step 4). When feasible and appropriate, meta-analysis helps in collating results. Finally, inferences and recommendations for practice are generated from interpretation and exploration of clinical relevance of the findings (Step 5). These steps are illustrated using a published review concerning the cervicovaginal fetal fibronectin test