Objective: To determine the sensitivity and precision of existing search strategies for retrieving child health systematic reviews in MEDLINE using PubMed.
Design: Filter (diagnostic) accuracy study. We identified existing search strategies for systematic reviews, combined them with a filter that identifies articles relevant to child health, and applied the combination in MEDLINE to a reference set of child health systematic reviews.
Main outcome measures: Total number of records retrieved, sensitivity, and precision.
Results: We tested 9 search filters. Sensitivity of the systematic review filters combined with the child filter ranged from 68% to 96%; sensitivity of the child filter alone was 98%. The number of records retrieved with PubMed (limited to January 1990-January 2006) by the systematic review filters combined with the child filter ranged from 7861 to 618 053. Precision for the combined filters ranged from 2% to 52%. Because of poor reporting of specific systematic review criteria in both titles and abstracts, in 25% of the records screened we were unsure whether the article concerned a systematic review according to our definition.
Conclusions: The high numbers of records yielded by sensitive search strategies and the low precision threaten the use of systematic reviews for clinical decision making and guideline development. Reporting of specific systematic review criteria in titles and abstracts is poor, and reporting recommendations given by Quality of Reporting of Meta-analyses (QUOROM) should be used more strictly. To make identification using MEDLINE easier, there is an urgent need to set minimal criteria that any review should fulfill for it to be indexed as a systematic review.