For the past few weeks I've been madly busy - as well as my regular day job being quite busy due to a technical error on my part (long story - but in short I forgot that it was the end of the school term this week for 90% of the country) I have been attending some interesting conferences (more to come) and last week I was on a whole weeks course learning all about (and how to do) systematic reviews. Lots and lots of interesting stuff. So I figured I would start backwards.

First up I thought I would share with the world my tips on finding clinically useful stuff on the net. The web is a tricky one to traverse when in your in a hurry - particularly when you want to find stuff FAST and make sure its accurate. Here are some tips:

* _Get your phone all internet-ed up!_ Check your data plan with your mobile operator and then skip the regular browser that comes with your phone by downloading and using [Opera Mini](http://www.operamini.com/ "Opera Mini") (ps: I actually prefer Opera 3.1)
Then, when that odd acronym pops up in conversation, or a rare diagnosis appears in the notes you are reading and you just need to solve that "What the heck is that?" thought, you can google/wikipedia it there and then! Did I mention wikipedia? That brings me on to my next point:

* _Don't trust all sources as the truth!_ OK, thats a bit strong. Use wikipedia for quick guidance to point you in the right direction. Don't however use it to base your entire working life on. Its just plain silly. No, infact its dangerous. I once wrote a quick talk based on something (I have wiped it from my memory now so don't ask me what it was on) off wikipedia. Different sources were conflicting with the facts or just not comprehensive enough. It was, I decided, far easier to use wikipedia but boy did I look a wally when someone says "Hmm are you sure??" Don't be caught out!

* _Use some reliable sources_ So what are they? Well pubmed is a good one - although perhaps somewhat limited. Now here's for the real tips du jour and some of the things you may not quite realise.
* Can't remember the (rather annoyingly long) url? Go to "[pubmed.gov](http://pubmed.gov)" in the URL bar of your web browser and it takes you to the right place
* [Medline is in essence just a component to pubmed](http://www.nlm.nih.gov/pubs/factsheets/dif_med_pub.html). If its in Medline its in pubmed. So why restrict your search? Use pubmed its far easier to remember than always going through your local library web page first
* Want to find clinically useful stuff fast? Try "[Clinical Searches](http://www.ncbi.nlm.nih.gov/entrez/query/static/clinical.shtml)" on the left. Now I say "try". Its not going to be *that* helpful for a lot of things. Two reasons why; 1) Its very US led - the vast majority of journals are from the states 2)Its very medically led - so it only contains a small fraction of OT journals out there. But saying that it is a very handy tool - and very comprehensive. Remember that some of the well recognised are in JAMA (e.g. well-elderly study) or the BMJ.
* Think carefully about your topic and think how a computer may find it. For example if your real question is "I wonder what the best OT interventions are for falls in the elderly population?" Don't write exactly that. Try simply typing in the words "older adults occupational therapy falls". Which brings me on to my next point..
* Pubmed is clever. Sometimes too clever. Lets take the above search "older adults occupational therapy falls" typed into the clinical search box. It actually gets converted to

"(older adults occupational therapy falls) AND (randomized controlled trial[Publication Type] OR (randomized[Title/Abstract] AND controlled[Title/Abstract] AND trial[Title/Abstract]))"

in the search field. So thats all its doing right? Well actually wrong. If you type anything into the search box it gets read by pubmed and then made "correct". To see what its done click on the "details" tab. For this instance the REAL search is this:

"(older[All Fields] AND
(("adult"[TIAB] NOT Medline[SB]) OR "adult"[MeSH Terms] OR adults[Text Word]) AND
("occupational therapy"[MeSH Terms] OR occupational therapy[Text Word]) AND
falls[All Fields]) AND
(randomized controlled trial[Publication Type] OR (randomized[Title/Abstract] AND
controlled[Title/Abstract] AND trial[Title/Abstract]))"

Clever eh? This stuff is really important when you need to do a systematic review because you need to reproduce your search - but equally its important for yourself so you know how to tame the beast and get better results.

* "You can't do subset queries in Pubmed can you?" This will probably be obvious to the rest of the world - Im amazed I didn't realise this. I know I could do this with other databases but figured it was beyond the simple design of pubmed - how silly I could be.. Lets start off with a simple example of how this works. Lets imagine your a OT wanting to investigate interventions for older adults.

1. Type in "Occupational Therapy" (no speech marks) in the search box of pubmed. Press the Go button. Note the number returned (e.g. 103218)
2. In place of your search term "Occupational Therapy" now write "Older Adults" Press the Go button
3. And once more: "intervention" and press Go.
4. Now click on the "History" tab. Here is a history of all your previous searches for this session. (NB: You will need to create an account to get pubmed to remember these searches). You will notice that each search has a "#n" where n is a incremental number of your search. Now lets imagine you want to see all results with all 3 terms in. Simply type the number of your search with the AND key word in between. e.g. for mine I did: "#1 AND #2 AND #6". Now you have a manageable number of articles to peruse.

The thing you will realise from my ranting about pubmed is that its actually not that great for OT. My example in the previous tip only shoes 37 articles for example. So what can we do? Well there is always Google Scholar. Now some warnings:

Google scholar is Good but not great. Some problems:
* You can't do IN/AND searches like the pubmed example (i.e. chaining)
* You have limited control of how the results are sorted.
Remember google lists the "most relevant first" but what "relevant" is, is only what google "thinks" is relevant to you..
that top result is not the holy grail
* Its out of date. Well that has been the argument from some areas of the researching sphere.
I'm not sure how accurate this is still the case given that these criticism were made in 2005..
* You have no idea as to what google scholar is actually searching
* Doing refined searching is hard.

Before you flame me I don't agree fully with the above statements - its just what has been complained about elsewhere (e.g. [the OM blog](http://blog.openmedicine.ca/node/24)). So how can we make use of it the best?

* First up to make it "Useful" make sure you set google scholar to link directly to your local library. It will save your searching to the full-text or even abstract a lot easier. To do this go to the preferences and search/add your local library (see screenshot). Note some other rather funky preferences: e.g. "Bibliography Manager" (I actually use a firefox plugin to make it talk to citeulike a whole lot easier)
* You can do relatively clever searching with google. Here's an example to show you a lot of them:

> paediatric|pediatric|children "occupational therapy" Sitting -standing intervention author:Reid

Lets break that down.
1. The pipe: | . "paediatric|pediatric|children" basically says "find me either paediatric OR pediatric OR children". Its the OR operator
2. The "" . Speech marks = search for this exact phrase. Google will ignore common words such as 'the and or' putting the words in speech marks search for sentences exactly with this phrase. If you just enter it will look for documents with the word 'occupational' and 'therapy' anywhere in its text. Its why many pages pick up random pages about occupational health.
3. The - . Minus sign followed immediately by the word. - = DON'T include. So for the above search where I want to search documents not mentioning standing (i.e. just articles looking at sitting) I just write -standing. Note that this is highly exclusive and not half as friendly as pubmed AND search.
4. author: . This one is the first that is only usable in google scholar. Allows you to search for a specific author. Note this isn't always that great. e.g if it was just author:Sm it searches for any text in the author field *starting* with Sm - capitals or otherwise. So it will pick up initials as well as surnames. Possibly a feature but something to be aware of.
NB: You can perform exactly the same search using the "Advanced" search but hey, you might as well do it the easy way right?!

Some other places to go hunting for clinically useful/research useful information if your in a hurry. (Note: you may need to access these through your institution to access them fully)

* [OTSeeker](http://www.otseeker.com/): "database that contains abstracts of systematic reviews and randomised controlled trials relevant to occupational therapy". NB: You can also do a search on pubmed for all RCT's relating to OT.
* [Cochrane Reviews](http://www3.interscience.wiley.com/cgi-bin/mrwhome/106568753/HOME?CRETRY...). A "small" (but actually quite diverse) database of reviews. Systematic reviews are of course the Gold Standard but few and far between in OT. A whole range can be found here - not just systematic reviews and its worth getting to know. NB: Cochrane reviews are listed in pubmed but may be easier to search on their own site. Just a tip.
* [OTDBase](http://www.otdbase.org/): Its a pain in the backside to search. Its butt ugly. And its slow (It runs off filemaker which is only slightly better than Microsoft access which my explain some of my complaints). But hey it has a lot of data. It is however not free.
* [Embase](http://www.embase.com/). Pubmed covers much of the same ground as Embase but Embase has more European journals and a more non-English journals.
* [CINAHL](http://www.cinahl.com/). Again, a lot of overlap in the Pubmed database (hence why I push pubmed first) but note that pubmed only indexes journals. CINAHL indexes journal articles, books and book chapters, dissertations, audiovisual materials, and other formats. So worth a look
* [Web of Science](http://isiknowledge.com/). Or ISI Web of Knowledge as its now known.
* [PsycInfo](http://psycinfo.apa.org/psycinfo/). Psychology based database.. worth a peek since so much OT covers others ground.
* [Google](http://google.com) main: Or [OT Google search](http://www.google.com/coop/cse?cx=005746190099945023050%3Aggfrkbnswqe). The latter isn't fantastic as I need to constantly add/remove relevant sites but hey, its a good start! (One advantage of the OT only one is No ads!)