Protocol Harem

I recently put out a request to the Clinical Librarian mailing list for literature search protocols and I promised to share what I found. Life then took over with a pile of literature searches and the small matter of finishing my dissertation for an MSc in Health Services Research.

So, a bit of background. I had noticed that the whole library team were using very different approaches to how we went about doing literature searches, and presenting the results back to the requesters. Don't get me wrong, I think a variety of approaches is absolutely fine, we've spent time all doing the same search before and then discussing how we went about it, and my general anecdotal finding is that everyone finds the major relevant studies no matter how they combine the terms. Scoping something out first may be your thing, or you might, like me, jump straight in at Medline and throw words at it and launch into a full search. I realise my route probably isn't perfect, as with a therapy question you really ought to check for the highest level evidence on the Cochrane Database of Systematic Reviews first (and it was nearly my downfall when I interviewed for this job - luckily I remembered in the nick of time, yet I've still not learned).

I thought we needed a bit of a plan at least, to make sure that people requesting searches get an equitable service. And the response I got to my request for literature search protocols was really quite good, since clin-lib doesn't get a lot of traffic and you can never be sure who's out there in the ether.

Some of the protocols are web-based, which I think is a really good idea. You're advertising exactly what your requester can hope as service when they ask their question. One such protocol is from the Royal College of Obstetricians & Gynaecologists (RCOG). It's simple, succinct and explains clearly what you can expect. Elaine Garrett, from the RCOG told me that it's based on the ATTRACT protocol

A more detailed protocol is actually on the NHS Evidence "For Librarians" section, from the Thames Valley Health Libraries Network (updated link). This is really comprehensive and would definitely be of use to the novice searcher as well as the expert. I also received protocols from the Library & Knowledge Service at Derby Hospitals NHS Foundation Trust, which is very similar in the depth of detail to the Thames Valley version, and also another similar one from the Medical Library at the Royal Free Hospital, so thanks to Lisa Lawrence and Ruth Muscat respectively.

The main points to note from all of these is preparing for the search, making sure the question is understood, documenting the process (standards about the format are an issue we've been trying to iron out), searching the right resources for the question (in the right order? I'm undecided about the order being important), and presenting the results clearly back to the requester. We haven't yet decided whether to adopt a formal protocol at UHL, but we're looking at the way we all do searching, which I hope to share in the future.

There are so many factors at play in every single request for information that I'm not sure a definitive search protocol that could be applied to every single question exists. I view the protocols more as guidelines for searching, or as an aide memoire, but I would be interested to know what colleagues and readers of this blog think.

Evidence summaries

So how do Clinical Librarians present their search results to clinicians? It's something I get asked about a lot. I don't know if my way's the right way, and I'd be interested to know what other people do too.

Generally I consider the level of evidence, and the date published and go from there. I never paraphrase results, I always lift the conclusion from the abstract. Due to time, it's rare that I'd do a full critical appraisal of a full text article, so most decisions of whether to include a study are based on a reading of the abstract.

I'd be really interested to know how others approach this, and what your opinions are. Please comment below!

Getting the best out of search filters, University of York, 26/2/09

I attended this one day course with Louise yesterday, as we both wanted to know more about using search filters in our work.

Carol Lefebvre & Julie Glanville are both highly experienced searchers & developers of search filters and ran the course really well, considering it was a new one, and has quite complex content. Overall, Louise & I really enjoyed it, but felt we needed even more time to spend on looking at appraising filters and learning about translating them between interfaces and databases. Our only complaint was that we didn't spend long enough on the exercise trying out the filters and translating them.

I tend to use quite pragmatic filters in my searching, and do it in a very slapdash way, so to learn that the real way is better was no surprise! But I do wonder how much time I would have to really make use of what I have learned. I do hope so.

We found out about the web site belonging to the InterTASC Information Specialists Sub Group, which aims to identify & share best practice in information retrieval for health technology assessments. The web site is a repository for published & unpublished filters, and Carol & Julie encouraged us to add any we may find in the course of our own searching. It also contains guidance on the critical appraisal of existing filters. It looks like an excellent resource, and I am planning to make use of it in the future.

Sarah is attending the Oxford course next week, and hopefully she'll enjoy it as much as we did.