Actually, PubMed has never been very good…

March 30th, 2008 Stewart Posted in Uncategorized |

I Am Not Yelling. Not Out Loud. - Lab Life - Anna Kushnir’s blog on Nature Network

As the blogger above will testify, PubMed is not at all attuned to how modern searchers search. On the other hand, it really has never been a particularly good search engine, so I don’t know why this is striking so many librarians as a harsh criticism. PubMed isn’t bad for government work, but it has many problems as we all know.

This isn’t an information literacy issue either, I’m sorry to say. Researchers and clinicians shouldn’t have to “learn to search” to be able to run basic searches, and instead they should call the librarian when things get rough. The analogy I like to use is for car owners: Everyone should be able to pump their own gas, some might even change their own oil, but sometimes you need to hire a pro to come in and fix the transmission.

Unfortunately, many of the competing products, including my old standby Ovid, are also just terrible for running simple MEDLINE searches. Partly this is due to legacy problems with MEDLINE, partly it’s due to limitations with the MeSH and the indexing, and partly it’s due to PubMed being coded in the mid-1990s and remaining largely unchanged this whole time. Researchers used to complain about PubMed’s imprecision; clinicians complained about getting too many hits. And now, we’re going to have a new generation of searchers used to very simple, effective techniques that work fine in Google and elsewhere not getting what they need either…