EMBASE and MEDLINE searches

The recent article by Wilkins et al on EMBASE versus MEDLINE for family medicine searches¹ clearly illustrates the importance of searching both databases when a comprehensive search is needed, for example, when writing grant proposals or preparing systematic reviews. College of Family Physicians of Canada members who do not have access to EMBASE should be aware that they can request EMBASE searches from the College’s library service. All members are entitled to five free literature searches per year—searches of MEDLINE, EMBASE, or any combination of many databases.

Wilkins et al used Ovid MEDLINE; I usually use PubMed MEDLINE. It is interesting that when the same search strategy Wilkins et al used for their sample search on urinary tract infections is run in PubMed MEDLINE, it yields 73 references, as opposed to the 30 references from the Ovid MEDLINE search. The details appear in Table 1. Seventy-three references retrieved from PubMed MEDLINE compares favourably with the 65 retrieved from EMBASE in Wilkins and colleagues’ article.

Table 1.

Sample search using Ovid MEDLINE and PubMed MEDLINE

*exp—explode command is performed by default in PubMed MEDLINE, so search strategy was Family practice/, Physicians, family/, etc.

**PubMed includes references other than MEDLINE references

I believe that a large part of the difference in number of articles retrieved from the two MEDLINE databases is due to the different ways the “therapy” subheading is handled. PubMed explodes the subheading “therapy” as well as the MeSH heading; Ovid MEDLINE does not seem to do this.

I have not made a detailed comparison of Ovid MEDLINE, PubMed MEDLINE, and EMBASE results, and I continue to believe that it is important to search both MEDLINE and EMBASE when doing a comprehensive literature search.