Believe it or not we are into the second week of January. The holidays have come and gone and now we need to turn our bleary eyes to what lays ahead. The switch from regular Ovid to OvidSP in February is right around the corner. If you haven't played with it, start now. I also recommend taking one of their OvidSP classes. OvidSP Basic Search is a totally different beast. Take everything you know about Medline and standard Ovid searching and throw it out the window.

First let me thank the people at Ovid, Julie Quain and Bob Morrisey for setting up a Medline OvidSP webinar for us to address the many issues and questions we had about OvidSP and OvidSP Basic Search. The session helped a lot of us to understand what the man behind the curtain was doing.

Here is what we learned about OvidSP:

The majority of the time, librarians will want to search Ovid Sytnax.

Don't let the name Basic Search fool you. While its name implies that it is the basic version of Ovid Syntax, that is biggest mistake you could make. Librarians (myself included) have been approaching Basic Search as if it is the simpler version of Ovid Syntax. Ovid Basic Search is so different from regular Ovid (or Ovid Sytnax) that it is almost a totally different Medline database.

Basic Search uses natural language processing. Yeah, yeah, we have heard that, big deal. Well the way you search it makes it a big deal! Every term typed into the search box is weighted according to specific algorithm that looks at the amount of times it is used within the Medline record (not the full text article). If you check the box "Search Related Terms," Basic Search is supposed to look for similar terms to your query. The natural language system does have a cut off of approximately 500. (For more information on the NLP 500 cut off click here to view a reprint of Ovid's explanation).

At our library we were concerned because we noticed the NLP was forgetting to include MeSH terms as part of the related terms. We discovered this with quite a few terms. Specifically, ragweed (ambrosia was not a related term), breast cancer (breast neoplasms was not a related term) and when you typed in breast neoplasms (breast cancer was not a related term). While one might make a pretty good case that ambrosia did not make the cut to be included in the 500 results, one would think the terms breast cancer and breast neoplasms are so interchangeable that they easily should have been included in the related terms before less popular terms like carcinoma breast. Because the backbone of MEDLINE (the records) is built around the MeSH vocabulary, this troubled us as librarians. We asked our trainer about the problem and she understood our concerns and told us she would find out about it. (When I find out, I will let you know.)

We thought there was a big problem with AND and OR. This was something that perplexed me. I had no idea why Basic Search handled AND and OR so poorly. I even used the AND and OR buttons at the bottom of the Search History box. The problem is that while in Basic Search, the system is trying to do Boolean searching while also trying to handle the algorithm and weighted concepts. The results are yucky. The bottom line…DO NOT use AND or OR while in Basic Search to combine searches. Only use it to combine results. What I mean is do not combine your breast cancer and tamoxifen search with your other search of breast neoplasms and tamoxifen. Go through both searches and select the articles you like. Once you have a set containing the articles you like, then you can combine those sets of selected articles so that you only have to print, email, or save one set.

They (the Ovid people) keep saying the best way to search is just to type the whole question/query into the search box. They are right. Stop trying to force Ovid Basic to perform like Ovid Syntax and you will be MUCH happier. I like to call them Magic 8 Ball searches. They work surprisingly well on OvidSP Basic.

Is OvidSP Basic the answer to comprehensive Medline searching? Heck no. But it is similar to and better than using Google for searching Medline, which is what a lot of people do (like it or not). OvidSP Basic also offers an alternative method of searching for librarians who want to make sure they aren't missing anything that may have slipped through the indexing cracks. As good as the indexers are, there are those articles and we librarians love to lament about those “perfect” articles that weren't retrieved unless we did some sort of dirty search.

What would be really helpful is if Ovid changed OvidSP Basic's name. Currently as it is named, it implies that it is a simpler version than Ovid Sytnax, and that is just not the case. You are going to have librarians and power users trying to force it to do a "simple" Ovid Syntax search and we now know that isn't pretty. Regular or first time users are going to think searching it is searching all of MEDLINE and we know that isn't the case either. I am not sure what you would name it. "A Few Good Articles," while descriptive, is probably too long for the pretty tabs on the display. I guess that is why I am a librarian and not in marketing.

So, I have changed my stance on OvidSP Basic. Once, I wouldn't recommend it, now I think it could be a good tool if used correctly. Use it to find a few good articles on exactly what you typed in the search box. Know that the results are not the end all be all of medical literature on the topic. Personally, I would list and describe the two search methods separately as two different products on my website. I would still have a link for Ovid Medline which would go to OvidSP Syntax. I would create another link that would say something like A Few Good Articles and it would go to OvidSP Basic. That way the users have the option to search either product when they need just a few good articles or in depth research.