Objective: Screening involves the use of a test or procedure to detect the presence of a disease in individuals not already seeking treatment for symptoms and not already diagnosed with the condition. Studies evaluating the accuracy of screening instruments that include already diagnosed or treated patients are known to produce inflated estimates of screening sensitivity and case yield. The objectives of this study were to investigate: (1) the proportion of original studies included in systematic reviews and meta-analyses on the diagnostic accuracy of depression screening tools that appropriately exclude already diagnosed or treated patients; and (2) whether systematic reviews and meta-analyses of the accuracy of depression screening tools evaluate possible bias due to the inclusion of already diagnosed or treated patients.

Design: Systematic review.

Data sources: MEDLINE, PsycINFO, CINAHL, EMBASE, ISI, SCOPUS, and Cochrane databases were searched January 1, 2005 to October 29, 2009. Eligibility criteria for selecting studies: Systematic reviews and meta-analyses in any language that reported on the diagnostic accuracy of depression screening tools.

Results: Only 8 of 197 (4.1%) unique publications from 17 systematic reviews and meta-analyses specifically excluded already diagnosed or treated patients. No systematic reviews or meta-analyses commented on possible bias from the inclusion of already diagnosed or treated patients, even though 10 reviews used quality assessment tools with items to rate risk of bias from patient sample composition.

Conclusions: Studies of the accuracy of depression screening tools rarely exclude already diagnosed or treated patients, a potential bias that is not evaluated in systematic reviews and meta-analyses. This may result in inflated accuracy estimates on which clinical practice and preventive care guidelines are often based, a problem that takes on greater importance as the rate of diagnosed and treated depression in the population increases.