Librarians’ Rx

It’s been almost 5 months since our University has been using the OvidSP interface exclusively. I’d have to say that reaction has been mixed. I find myself waffling over whether I love or hate the basic search more often than I care to admit. I tend to hate it when I’m helping someone on the reference desk, and I see that they’re getting 0 hits because they’re using the Basic Search the same way they use the Advanced Ovid Search. Or even worse, they’re doing parts of the same search in both the Basic Search and the Advanced Ovid Search. This brings up a larger question about whether it’s appropriate to have two search interfaces that physically look the same but do not work the same way. Is it realistic to expect users to work by 2 sets of rules?

Anyway, I’ve come up with some basic hints to understanding the Basic Search:

• It’s not a quick search. You can’t put in a query and have the database search every field looking for your terms.
• You cannot use the searching tricks we normally use. This means:

• No Boolean
• No truncation
• No combining searches
• No parentheses
• Also, searching in to other languages seems to be problematic, which is an issue if you’re doing a search in Pascal.

• It doesn’t search every word in your search string. This makes it better than google in a lot of ways, because it’s smart enough to recognize that only relevant terms should be searched. For example, it will only search the underlined terms in the following query and will ignore the rest:

Is there an effective treatment for hepatitis B in children

To see the terms that are being searched just check the Search Aid box on the left of the results.

This is important because there are noise words that aren’t searched at all. A few of these are: policy, condition, related. Obviously, this could cause some serious problems for certain searches.

• Limited results. The first time you do a search in OVID SP you may get thrown for a loop because there will only be around 500 results, even when you do a search where there is a lot of literature. Don’t worry; the database is designed to retrieve around 500 results since most people don’t bother looking beyond the first few pages of search results.
• Relevancy ranking. OVID SP ranks results based on relevancy. Relevancy is determined by the number of terms found in the record (it’s important to remember that this doesn’t necessarily mean the number of terms in your query, but just the number of terms it actually searched for), where the terms appear in the record, whether the actual query term was found or if a related term was found, number, frequency, and proximity of concepts in the record. Don’t forget that you can re-sort results based on a number of different parameters if so desired.
• Combining searches. Do not do the usual thing where you search for one concept, then search for another concept, and then combine the searches. This does NOT work. Here’s an example of what the problem is:

I searched for diabetes. I retrieved 515 articles.

I searched for retinopathy. I retrieved 565 articles.

I combined the two searches and only retrieved 1 article.

This happens because each search only pull the most relevant articles. So in actual fact, there are around 45 000 articles indexed with the MeSH Diabetes Mellitus, but in the basic search I only retrieved 515 of these. Just like there are a lot more articles on retinopathy than the 565 I retrieved.

So when the searches are combined you’re only combining small portions of citations dealing with your subjects.

The solution is to combine terms in the same query. You need to do a one line search. When I did this, I retrieved 652 articles.

• Related Terms allows for searching of synonymous terms in a query. So instead of searching therapy or cure or treatment the database will search all of those terms at the same time. It works fantastically some of the time, and not that well a lot of the time. According to OVID the related terms are evolved from the Unified Medical Language System meta thesaurus, medical dictionaries & thesauri, medical acronyms, drug and disease names, and American and British dictionaries. So if you put in a term like therapy, you get a lot of relevant related terms. But if you put in smoking, the only related term is smokings (not cigarette or tobacco addiction). If you use the term obese, there are no related terms identified, although there are a lot of related terms for obesity. I wish this worked better. It’s also strange that the related terms don’t usually include relevant MeSH headings such as the ones you’d be mapped to in the Advanced Ovid Search.

So basically if you’re a librarian you may have difficulty with the basic search, because you won’t be able to use you’re normal searching tricks. Also, if you’re a student you may have difficulty because you won’t be able to search for an author’s name or a journal title in the basic search. You’ll need to search in the Find Citation section of the interface or do an Advanced Ovid Search. But overall, the ability to communicate with a database as though you were talking to a colleague is pretty amazing, as long as you know what it can and cannot do.

–Thane