OBJECTIVES: To compare two approaches for re-running database searches in OVID and assess best practices for citation management and reporting.

METHODS: A search strategy using population terms for a rare disease was executed in MEDLINE and Embase during May 2022. In December 2022, two Cochrane-based approaches for re-running searches in OVID were executed: (1) applying date limits (‘date-limited refresh’), and (2) not applying limits (‘unrestricted refresh’); all retrieved records were matched against the previous yield. Multiple date-limiting fields were applied in each database (Embase: date delivered [dd], revised date [rd]; MEDLINE: create date [dt], rd). Hits from both approaches were matched to the initial search using a matching algorithm implemented in R (v4.2.1). The sensitivity of each approach was compared based on the respective number of total and unique hits identified.

RESULTS: The initial search executed in May 2022 yielded 8,026 hits. The date-limited refresh yielded 592 total hits, with 432 hits unique from the initial search. The unrestricted refresh yielded 445 unique hits (total 8,451); of the 8,006 algorithmic matches, 8,002 were made on year, title and author, and four on title alone. An additional eight title matches were false based on manual inspection. All hits from the date-limited refresh had been captured by the unrestricted refresh. Thirteen hits (2.9%) from the unrestricted refresh were not identified in the date-limited refresh; however, none had been published in the May-December refresh timeframe, nor had been identified in the initial search.

CONCLUSIONS: An unrestricted approach employing an automated matching algorithm is recommended when refreshing systematic literature reviews as it is more sensitive than applying date-limits. If a date-limited approach is preferred, multiple date fields should be included to ensure comprehensive results. Future work could add comparisons with other refresh approaches, such as the Cochrane-proposed method of using automated monthly alerts.