ABSTRACT
OBJECTIVE: To investigate if Cochrane reviews that assess screening interventions address their major harms. STUDY DESIGN AND SETTING: A systematic search for Cochrane reviews that assess screening interventions was performed. Two authors independently screened abstracts, assessed full-texts, and extracted data from included reviews. For each review, two authors judged whether each predefined harm was relevant. When the harm was judged as of questionable relevance, the review was excluded from the denominator in our calculations. RESULTS: Forty-seven reviews were included. Overdiagnosis was addressed in 6 of 39 (15%), overtreatment in 7 of 43 (16%), and psychosocial consequences in 30 of 47 (64%) of reviews where this was judged relevant. When data on harms were included, they were generally not treated with the same methodological rigor as the benefits, with no assessment of the risk of bias or certainty of the evidence. About half of the Abstracts, Plain Language Summaries, and Summary of Findings tables did not include any harms. CONCLUSION: The underreporting of harms of screening in Cochrane reviews likely reflects primary research and is problematic. We call for broad collaboration to develop reporting guidelines and core outcome sets for studies of screening interventions.
