Clinical Evidence Based Information Specialist — Part II

Natasha S. den Dekker
7 min readJan 20, 2019


Last week I spoke about the where’s and why’s of the types of searching that I do in my current role, so this week — we’re going to look at the different types of search strings that can be utilised for clinical and non-clinical questions. I’ve only been in-role for 6 months so I’m still figuring things out! And my colleague who has been in the role for 7/8 years and was awarded ‘Search Professional of the Year ‘ by the BCS is also still learning so…this article is by no means a full guide as to ‘How to Do It’ but more a ‘This is What I’ve Been Doing, and Maybe it Will Help You’.

Let’s take a question that I did last year:

Why tinnitus patients prefer sound generating devices

In order to do this properly, you have to break down the question — and sometimes (often) this will involve you going back to the person that made the request so you can gain a bit more clarity. Things like:

  • Who is it for? (a patient, research)
  • What type of patient is it? (child, adult, etc)
  • How far back do you want me to look? (5 years, 10 years, FIND EVERYTHING POSSIBLE ON THIS TOPIC)

You get the idea. Thing is, the more information you have, the easier it will be for you to do your job. And, if there is a word you don’t understand, for god’s sake look it up or ask the person. OR if you know what the condition is but you aren’t sure about the context, ASK. I cannot stress this enough. Do not dive into a question half-cocked; an extra 10–15 mins spent trying to get your head around a topic is gold in my opinion. Familiarising yourself with the landscape of that condition/disease will make the following search so.much.easier. What next?

Well….I’m a bit old school in this regard and I’ll usually find scrap paper and start mapping out what I’m going to do. I take my question and I break down the different concepts. In this case it’s the following:

I’m not joking when I say ‘map’ — if anyone has studied grammar or taxonomies, that’s exactly what my mapping looks like. I list all the alternative words/terms for each phrase, I look up what each term means in context. I try and be as thorough as possible because it really does help later on. And if you have to re-do your search it’s nice to know what you’ve already done.

I’m building my search strategy

And this should take time. I have had questions where I’ve had to leave it, sleep on it, and come back to it. Because you don’t really want to leave anything to chance — beyond the slim hope that you may pick up on some phrases in the articles that you find. At this point, I haven’t even gone near a database! Honestly — beyond a bit of Wiki, and different governing bodies for the condition via Google I haven’t started trying to answer the question! (Unless if it’s something I know a lot about i.e. strategy, or morale — that stuff is already in my head from previous questions/roles).

NOW YOU CAN BEGIN!!! This is probably my favourite part — the bit where you go into PubMed and start building out all the stuff you’ve written down. PubMed is my favourite mainly because it makes sense, and it has the largest repository of #stuff to find.

Each line is for a concept and then you click the menus to do your standard AND/OR type search. As well as choosing the MeSH heading for specific conditions if you know that a particular word is going to bring up a lot of phrases — with my questions I generally go down the keyword search because PubMed doesn’t index articles straight away, so by using the MeSH you may be missing on recent additions that just haven’t been tagged properly. I appreciate that this may be quite hard to visualise, SO LOOK AT THE ONE I MADE EARLIER:


tinnitus AND combination devices or sound generating devices AND choice or satisfaction or prefer or reference

Tinnitus AND white noise OR sound AND generator OR generators OR generating AND choice OR prefer OR preference OR satisfaction

tinnitus AND white noise OR sound AND generator OR generators OR generating AND choice OR prefer OR preference

tinnitus AND white noise OR sound AND generator OR generators OR generating AND patient satisfaction OR choice OR prefer OR preference

tinnitus AND white noise OR sound AND generator OR generators OR generating

tinnitus AND white noise or sound and generator or generators or generating

tinnitus) AND sound therapy devices OR sound generating devices OR wearable sound generator OR white noise generator

tinnitus AND sound generator

tinnitus AND sound therapy devices OR sound generating devices OR wearable sound generator OR white noise generator AND patients’ choice OR patient centred

tinnitus AND sound generator AND choice

tinnitus and sound generating devices AND prefer or preference

My first attempt is at the bottom and the string that yielded the most useful is at the top. This hopefully helps break down some of what I previously mentioned i.e. breaking down different terms to see what might yield the most results. For such a simple question, this was one of the most detailed searches I had done at that point. It was a topic that the team hadn’t done too much on, so I was essentially in uncharted waters and there was a lot of trial and error. I didn’t just stop at PubMed though (be very wary of ANY research that *just* uses PubMed), below are the other places I looked:

Google Advanced Search

tinnitus and sound generating devices

Google Scholar

tinnitus and sound generator and choice


tinnitus AND white noise or sound AND generator or generators or generating AND outcome or satisfaction


tinnitus and sound generators and choice

tinnitus and sound generators

NICE Evidence

tinnitus and patient satisfaction and combination devices

tinnitus and sound generating devices

Had so much fun with this one — no I didn’t.

BUT I think it was the first time that it all clicked for me. The understanding how to break down terms, how to properly use the strengths of each database to pull out the results I needed. It’s definitely trial and error in some cases — and sometimes you do have to account for different spelling (hey tumour and tumor I’m looking at you!) and proper drug names vs. their components. It’s what makes it easier to search and why I will always advocate going back and asking for more information. I will be the first to say that yes, this can be rather tedious especially if it’s a topic where you’re a bit like NO, I DON’T WANT TO. And when you get a question that is so vague (because the person doesn’t really know what s/he wants) and they’re all ‘just see what’s out there’ (#eyetwitch) and it means that you have to make very prolific use of the Ctrl+F button in every article you find to see where (if any) your search terms may appear. Titles can be so misleading and you do have to read the abstract and sometime the whole sodding article. (I’m still reeling from all the information retained from looking into ‘hand therapy’).

Also, there’s that thing of ‘when do I stop’ and the thing is, it’s not a quantifiable thing. I’m not going to say, oh ok, you found 10 articles that seem legit, stop.

It doesn’t work like that. This is where the independent thinking comes in, and understanding what the person wants. Have you found enough for their needs? If it’s a systematic review — RUN THE SEARCH AS MANY TIMES AS YOU CAN. But for a conference presentation, a bit of CPD research, a bit of curiosity — you have to use your common sense. Logistically, if you’ve built your search properly — you will start to see the same things crop up. Personally, I know when I’ve hit that point when I’m reading articles that referencing articles that I’ve already found (always gratifying) and also when you do a quick double-check and yup, your search string is still pulling up the same results (in whatever database you’re searching). It can take 2 hours, it could 4 hours, it could even take a day or 2, it’s hard to know. *insert sobbing face* BUT THEN you have questions that are so specific and so niche that you know that you aren’t going to find much — when that happens we throw it back to that person and we’re like ‘maybe you could recommend some search terms, if you feel we haven’t been thorough enough’ — highly unlikely though (they rarely come back to us 🤣)

This is what I do — everyday. It’s not for everyone and there are days when I do miss having more interaction with people, but then, I don’t miss frontline librarian-ing at all. My team makes the difference since we do talk about what we’re doing most days and looking at growths on scrotums totally talks to the really macabre and dark side of my personality. If there’s one thing Oxfam gave me, it was an appreciation of being able to talk about really gross bodily functions in a serious way, and now I have a job where I get to research the use of lasers, learn the difference between different cancers and so on. You can’t help but learn stuff as you go and I’m already a better researcher than I was 4 months ago. Please join me in chortling at the irony that post-MS I couldn’t get a single research job, but have somehow walked into one in the NHS (even if it’s a maternity cover contract).

Originally published at on January 20, 2019.



Natasha S. den Dekker

User researcher, Ex-Librarian. Microsoft, Oxfam, NHS. Civica. Hyperlearning AI, Lexis Nexis exercise. Probably drumming, lifting weights or planning a holiday.