I’m still new to the world of systematic review searching, so I’ve been doing a lot of self-directed learning lately to learn all the ins and outs of this area of librarianship. I’ve recently been exploring the question of whether there is a noticeable difference between running rapid review/systematic review searches in PubMed MEDLINE versus Ovid MEDLINE. The answer led me on an interesting chase, so I thought I’d share what I learned.

Naturally, I started with the literature. The first study I found on this topic was a 2016 JMLA paper that investigated whether conducting a supplementary search of PubMed in addition to a search in Ovid MEDLINE was a useful exercise (Duffy et al.). The authors found that they were able to identify additional unique records in PubMed by limiting the search to records not in Ovid MEDLINE or Ovid MEDLINE In-Process. Their conclusion states that “supplementary searching of PubMed for studies unavailable elsewhere is worthwhile and improves the currency of the systematic reviews”.

This seemed to be further supported by the next study I found, by Katchamart et al. (2011). They found that PubMed has a higher sensitivity than Ovid-MEDLINE but pointed out that the PubMed interface isn’t as supportive of systematic review methodology as the Ovid interface.

However, this conclusion was complicated by an interesting letter to the editor published in 2012, addressing the methodology and results of the Katchamart study (Boeker, Vach, & Motschall). They offered that this study failed to include the Ovid MEDLINE In-Process & Other Non-Indexed Citations (Ovid-MEDLINE-IPONIC), which would have changed the outcome of their search comparison. Ovid allows users to change the subset of records during searches, which is more useful for systematic reviews but also makes the interface more complicated.

A recent study by Burns, Shapiro, Nix and Huber (2019) does an excellent job of describing how the MEDLINE records interact with different platforms, and how these different platforms use their own indexing and search interfaces, resulting in variations in how MEDLINE records are retrieved (I wish I had found this paper first!). They found that five different platforms for accessing MEDLINE records each had variations that make it impossible to substitute one platform for another when conducting or reproducing searches.

This isn’t an exhaustive search, but at this point, I feel like I have enough information to start forming conclusions. What I understand from these articles is that searching PubMed and searching Ovid MEDLINE are definitely not the same thing, but neither one is the “wrong” choice of platform. I think that I will keep using Ovid to search MEDLINE because of its more systematic review-friendly interface, and add a quick PubMed search to make sure I’m finding the most recent articles.

References

Boeker, M., Vach, W., & Motschall, E. (2012). Sematically equivalent PubMed and Ovid-MEDLINE queries: different retrieval results because of database subset inclusion. Journal of Clinical Epidemiology, 65(8), 915-916. doi: 10.1016/j.jclinepi.2012.01.015

Burns, C.S., Shapiro II, R.M., Nix, T., & Huber, J.T. (2019). Search results outliers among MEDLINE platforms. Journal of the Medical Library Association, 107(3), 364-373. doi: 10.5195/jmla.2019.622

Duffy, S., de Kock, S., Misso, K., Noake, C., Ross, J., & Stirk, L. (2016). Supplementary searches of PubMed to improve currency of MEDLINE and MEDLINE In-Process searches via Ovid. Journal of the Medical Library Association, 104(4), 309-312. doi: 10.3163/1536-5050.104.4.011

Katchamart, W., Faulkner, A., Feldman, B., Tomlinson, G., & Bombardier, C. (2011). PubMed had a higher sensitivity than Ovid-MEDLINE in the search for systematic reviews. Journal of Clinical Epidemiology, 64(7), 805-807. doi: 10.1016/j.jclinepi.2010.06.004