Trip Database Blog

Earlier today we posted some preliminary results of our survey (click here).  I’ve now had a chance to look at the main textual results and – while viewing them – I spotted the overlap to responses to the questions ‘What is bad about TRIP?’ and ‘What features would make you use TRIP more?’.

The response has been great and has been really useful in helping to focus our minds.  Below is a brief overview of the main results.  I’ve given a global result plus broken down into 3 categories based on profession (apologies for lumping in ‘Other health professionals’):

Doctors (top 3 improvements)

• Access to full-text articles
• Improve the advanced search
• Smart phone app

Information specialists

• Improve the advanced search
• Clarity of the results (how the results are arrived at)
• Poor, unrelated articles appearing in the results

Other health professionals

• Refine search – make it easier
• Improve the monthly email
• Improve the Boolean search

Global results for all respondants (top 7)

• Refine search – make it easier
• Improve the advanced search
• Access to full-text articles
• Poor, unrelated articles appearing in the results
• Improve the monthly email
• Improve the Boolean search
• Clarity of the results (how the results are arrived at)

Overall, lots of good ideas.  The one worry for us is the frequent reporting of ‘wrong’ articles appearing in the search results.  Something for us to deal with well before the next upgrade.

As ever, one big thing that was apparent from the users – the love for TRIP 🙂

We posted our latest survey yesterday afternoon and so far we’ve had an amazing 345 responses (if you’ve not done it, it’s not too late https://www.surveymonkey.com/s/WMS38PY).

I’m really keen to get our users input into the next upgrade to TRIP (scheduled for early 2012) and this is the first stage survey.  We asked people to leave their emails if they want to take part in further surveys and we’ve already got 200+.  The first survey asks mainly high-levels questions (e.g. what’s good about TRIP, what’s bad etc).  The second survey will ask more specific questions about functionality, design etc.  I really can’t wait.  But before the second survey I have to finish reading a few books and papers on search and to fully analyses the first round of results. 

A lot of the questions will require some form of textual analyses but other lend themselves to a simple numerical analysis, so some early results below:

Background (top 3)

• Doctors – 44.1%
• Information specialists – 18.8%
• Other health professionals – 10.7%

What do you use TRIP for?

• Answering clinical questions – 69.6%
• Keeping up to date – 50.7%
• Research – 56.5%
• Other – 10.1%

How do you use TRIP?

• A pull way (searching the site) – 55.9%
• A push way (monthly emails, RSS etc) – 8.4%
• Both – 35.7%

One issue we’re looking at is the content offering of TRIP, what best describes it for you?

• Perfect – 59.1%
• Way too much – I get too many results – 24.3%
• Way too few – I often get few and/or poor results – 16.5%

Another area of interest is the design of TRIP (how it looks). We have recently rolled out a new design and we’re not convinced it’s great. What do you think?

• The site looks great – 22.9%
• I’m not bothered how the site looks – 26.1%
• I think it looks ok – 47.5%
• I think it looks awful – 3.5%

If you were required to login to TRIP to use it, how would you feel?

• I’d stop using the site – 5.5%
• I’m happy to do it, but make it easy – 53.0%
• I’m happy to do so, as long as I get a better service – 24.9%
• I’m not too keen on this idea – 16.5%

So, lots of interesting results.  But the biggest ‘steer’ will be the textual analysis of the 3 questions:

• What’s good about TRIP?
• What’s bad about TRIP?
• What feature would make you use TRIP more?

Bing, the world’s 2nd best used general search engine (after Google) has just announced the release of adaptive search.  A brief overview can be seen via this searchengineland blog article.

Adaptive search excites me enormously and it’s something we’re working on at the moment.  I say working on, we’re working on the theory and how it might work.  We need to save up our pennies and test a few assumptions before building it!

The principle is that adaptive search learns the type of person you are and alters the results based on what it has learnt about you.  This means different people would see different results.  For instance, a general practitioner in the UK might search for hypertension and s/he would see different results to a cardiologist based in Canada.  It makes sense that they see different results as their contexts are different.

We’re actually pretty confident with the theory but the biggest assumption is, for this to work, people need to login.  So, we’re exploring that and allowing people to login with their Facebook and Twitter accounts should make this easier. 

This may seem a strange title as many of you will feel that TRIP is an evidence-based tool already.  Well, TRIP helps users find the best available evidence, I’m comfortable with that.  However, there is more to this than that….

Since the new interface/functionality went live I’ve been tidying up loose ends, fixing a few broken bits of functionality and looking forward to the next changes to TRIP.  This has coincided with a wonderful opportunity to get involved with a research grant that will explore creating a search interface based on academic theories around information gathering.  This will involve creating a series of interfaces which will then be tested with a group of clinicians.  So, we’ll be doing research to explore various aspects of search interface design and results display.  We’ll be creating an evidence-base for clinical search.

I’ve no idea if we’ll get funding but the person in charge is hugely influential so I think we’ve got a reasonable chance.  Even if it doesn’t happen I’ve already learnt an awful lot from just reading the background papers. 

Exciting times!

The latest upgrade to TRIP had lots of new features easily visible to users.  However, for all the external changes we have made a number of internal ones.  One of the unhidden wonders has been a broken link reporter!  Basically, if someone clicks on a link and it doesn’t result in a valid webpage being open we now get an email telling us the link isn’t working.  Previously, we had no system for this and it was down to the goodwill of our users to let us know.

When the new feature went live it was truly depressing as I was personally receiving about an email every minute.  However, it soon showed up that a few particular publications were causing the bulk of the issues. Within 4 days these links were fixed – causing a 90% reduction in the error messages.  Now, 3 weeks later and lots of hard work, we’ve fixed approximately 99% of all the broken links.

For most people they will not notice a difference – but it’s great to know our system’s working well and long may it continue. 

I’m very pleased to announce that we have launched the latest version of the TRIP Database a few hours ago.  New features include:

• Translate function.  Users can translate the site and subsequent links into one of six languages (Spanish, Portuguese, Italian, French, German and Welsh).
• We’ve added the ability to restrict results to only those added in the last month.
• We’ve started searching one of the largest clinical trials databases.
• Started extracting and displaying DOIs against each result (if they exist).
• Included a separate section for clinical calculators.
• Included a separate section for social media results.
• The speed of the site has been improved.
• New design.
• Started to include patient decision aids and medical education content.
• Massively increased our medical video offering, now searching over 6,100 videos.

Feel free to let us know what you think!

We’re building a TRIP app.  Initially it’ll be for the Apple iPhone and if that proves successful we’ll create a version for Android.  It’ll actually be multiple apps, but all following the same format – highlighting the latest content added to TRIP for a given area. 

We see there being two main types of app:

• Clinical area – all the new content in areas such as cardiology, oncology etc.  They’ll be around 25 of these.
• Individual clinical conditions – while we could produce unlimited, we’ll start with 20-30 of the biggest e.g. diabetes, asthma, hypertension.

Each month users will be alerted when the new content is available and will be able to select to see content based on our filtering system (e.g. systematic reviews, guidelines) and then simply scroll through the results.  Below is a screenshot from our prototype – and it’s great!

We’ll have to charge for this, but it’ll be suitably cheap!

For a while now we’ve been exploring linking to full-text articles (currently we only link to abstracts of primary research articles).  It’s an area I have little technical experience in and have allowed myself to avoid the issue.  In the latest version of TRIP (out next week) we’re extracting the DOI for journal articles (and Cochrane), but this is only part of the issue.

To link to a full-text requires an additional step – it requires a link resolver! Trying to keep things simple, it tells the system where to point the full-text request to (ie send the person to Wiley, Ebsco etc).  We’re looking to introduce this and I’m fairly confident we can do it!  It should work like this:

• We extract the DOI for an article – which we’ve done already!
• An institution tells TRIP which full text holdings it has (e.g. NEJM, Thorax etc) and also some details of their link resolver.
• A user comes to TRIP from a given institution and carries out a search and we display a full-text link to all articles the user has full-text access to.  They click on the link and they’ve got access to the full-text (based on an authentication system).

So, to my mind, the hard work lies with the librarian who needs to tell us what full-text holdings they have!  The rest seems relatively straightforward – I must be missing something.

But, bottom line, does easy linking out to full-text excite people?

For those of you who follow TRIP you’ll know that I frequently reveal our lack of money.  It’s something I’ve got used to and learnt to live with.  I like to think of myself as an innovator and product developer, as opposed to a businessman.  I’m not great at the latter, I don’t like asking for money!  However, we’ve survived for over ten years and are still financially viable (I was going to say strong, but that’d be a slight exaggeration)!  What we’ve achieved with a small budget has been spectacular and part of me thinks that being kept hungry keeps the pressure on me to innovate.

But, I do occasionally fret over the lack of business development. I do wonder what we could do with a decent budget.  I do have a realistic set of ideas/innovations that would make TRIP significantly better.  I do think I could make TRIP the Google for medicine – the first port of call when clinicians have knowledge needs.

So, when I was approached 4-6 weeks ago by someone interested in helping TRIP on the business side, I was excited (and apprehensive).  We had some great chats and interesting exchanges of emails.  I’m thinking the approach may well have stalled but I’m not dispondent and that’s for the following reasons:

• I learnt an awful lot during the due diligence process.  I received numerous pieces of advice – all free – around various aspects of business, including corporate structure.  I’ve come out of this stronger and more confident than before. 
• As a result of one thread of advice I’ve been introduced to a venture capital firm.  This relationship isn’t necessarily about raising finance, it’s more about exploring options and business models.

With the new version of TRIP due before the end of the month and all this excitement around business development I can’t help but be excited about the future.  I doubt I’ll ever be a businessman but as long as I keep enjoying my business that’s fine by me 🙂