Validating Cochrane Neonatal’s standard search databases: Is it okay to stop searching Embase?

Article type
Authors
Ovelman C1, Eckert C2, Friesen C1
1Cochrane Neonatal
2University of Vermont
Abstract
Background: Methodological Expectations of Cochrane Intervention Reviews (MECIR) standards require Cochrane reviews to search Cochrane’s Centralised Register of Controlled Trials (CENTRAL), MEDLINE and Embase (if Embase is available). In July 2019, Cochrane Neonatal (CN) lost access to Embase. Previously, we searched MEDLINE, Embase, CENTRAL and CINAHL for all intervention reviews.

CENTRAL is comprised of randomized controlled trials (RCT) and controlled clinical trials (CCT) from Embase and MEDLINE, identified via Cochrane’s centralized search and screening processes. Cochrane Information Specialists add included RCTs from their reviews to CENTRAL. CN added included RCTs to CENTRAL through December 2017.

For intervention reviews, CN uses these standard search methods: RCT filters created using Cochrane’s highly sensitive search strategies for identifying RCTs and neonatal filters created and tested for each database by our Information Specialist.

Objectives: to assess the effectiveness of conducting searches for CN intervention reviews in MEDLINE, CENTRAL and CINAHL, without searching Embase.

Methods: We searched for all new and updated CN intervention reviews published from January 2018 through April 2019. For new reviews, we compiled a list of included studies. For updated reviews, we compiled a list of only new studies added in the current update. If an included study had multiple publications, we only included the primary publication.

We then used our standard search methods to search for each study in MEDLINE and CENTRAL. When a study was not identified in MEDLINE or CENTRAL, we searched Embase and CINAHL.

Results: We found 13 new intervention reviews. One new review was outside the neonatal scope and was excluded. The remaining 12 new reviews had 64 included studies. We found 17 updated intervention reviews, with 66 new studies included in the recent updates. The new and updated reviews included only RCTs and CCTs, and represented 12 of 24 active neonatal topic areas.

For the 130 included studies we tested, 4 did not appear in either MEDLINE or CENTRAL. Of these 4 studies: none were found in Embase; 1 study appeared only in CINAHL; 2 studies were conference abstracts and were not identified in any of the databases searched; 1 study was in a journal not indexed by MEDLINE, Embase, or CINAHL. This study had a conference abstract that was identified in CENTRAL.

Conclusions: For new and updated neonatal intervention reviews, which plan to include only RCTs and CCTs, it is adequate to conduct searches in: MEDLINE, CENTRAL and CINAHL, without additionally searching Embase. Authors should search the reference lists of any studies selected for inclusion in their reviews in order to identify additional relevant articles. Searches for conference abstracts could be conducted separately. Authors may consider if searching CINAHL is necessary for their review topic.

Consumers: contribute to the creation of CENTRAL through Cochrane Crowd.