No consensus exists on search reporting methods for systematic reviews.
Sampson M., McGowan J., Tetzlaff J., Cogo E., Moher D.
OBJECTIVES: The reporting of the search methods used in systematic reviews has implications for how systematic reviews are critically appraised, their reproducibility and how easily they may be updated. The objective of this paper was to identify validated or evaluated search reporting instruments used in reporting systematic review searches and to compare reported and recommended searching practices. STUDY DESIGN AND SETTING: This cohort study was a systematic review. The search strategy to identify instruments addressing the reporting of search strategies was developed first in the Library & Information Science Abstracts (LISAs) database and then adapted for MEDLINE and five additional databases. Additional instruments were identified through experts. Current reporting practice data were analyzed from a cohort of 297 recent systematic reviews. RESULTS: Of the 11 instruments examined, 7 cited supporting evidence but only 4 were validated. Eighteen different reporting items were identified but only one item, "databases used," appeared in all instruments. There was a trend toward including more items in more recent instruments (r=0.41). Current search reporting practices ranged from a high of 98.7% for databases used to a low of 11.4% for qualifications of the searcher. CONCLUSIONS: There is no clear consensus regarding optimum reporting of systematic review search methods and commonly recommended items show suboptimal reporting.