Trip Database Blog

The new version of TRIP is very close to launch. Unless we have any major problems highlighted it’ll go live early next week. One of the major new enhancements will be the introduction of a ‘quality slider’. I first raised this problem/solution in 2007 (see this article). It’s amazing that the concept has come back and works very well when in operation:

We have now colour coded all records in TRIP based on the concept of the hierarchy of evidence. When someone carries out a search of TRIP they’ll have all the results presented. But, they may decide to only include the highest quality evidence, in which case they simply move the slider to the left and all the lower quality evidence is removed.

TechCrunch is a technology blog and is respected enough to gain access to te biggest tech players in the world. In this interview with the CEO of Google (Eric Schmidt) they ask about the future of search:

“So I don’t know how to characterize the next 10 years except to say that we’ll get to the point – the long-term goal is to be able to give you one answer, which is exactly the right answer over time. Okay, you know, the question I’ll ask today, how many Americans have – what percentage of Americans have passports?…The Google’s answer was a site, which was somebody who had attempted to answer that question and had multiple answers. It’s quite interesting actually to read…So you go to a very good definitive site. And what I’d like to do is to get to the point where we could read his site and then summarize what it says, and answer the question…Along with the citation and so forth and so on.”

It’s a challenge to do that for fact based questions (such as ‘How many Americans have passports?’) but could they really tackle clinical questions? These sorts of questions are so much harder as there is often not an answer – more the odd bit of research that helps a doctor form an opinion. Perhaps in twenty years time they’ll get close!

Click here for the full article.

Just over a week ago I posted that we’ve managed to grab over 90,000 medical images. That figure has now risen to 284,398. The figure shouldn’t go any higher but will certainly get lower. Our system which grabs the images is fairly sophisticated but it still grabs any image on a page (based on certain parameters). As a result a chunk of those 284,398 images will be ‘junk’. The only way to weed out this junk is for them to be manually reviewed and inappropriate images exluded. If this is completed before the launch (sometime in the next 10-14 days) it’ll be a miracle, but even if it’s not completed by then it’ll gradually improve over time. Even if we end up with ONLY 100,000 images it’ll still be an amazing collection.

6 weeks ago I posted an article about using TRIP to identify content suitable for resource poor settings click here.

Since then I’ve been busy discussing the idea with a significant number of people including Richard Smith (Director of Ovations and ex-editor of the BMJ), David Lipman (Director of the NCBI and responsible for PubMed) and representatives from WHO. I’ve worked up the idea some more – with the help of those I’ve discussed the idea with – and am currently seeking funding. The idea, as it stands, is as follows:

Problem: A clear problem for resource poor settings is identifying high quality evidence upon which to base clinical decisions. Within TRIP, content suitable for these environments, is frequently ‘hidden’ by the volume of material aimed at resource rich environments, making the identification of appropriate evidence difficult. This problem is not restricted to TRIP: there is no way in Medline (or other databases) to select material suitable for resource poor settings.

Proposal: We propose that users of TRIP would be allowed to ‘tag’ (by pressing a single button) an article if it is suitable for resource poor settings. A user coming to TRIP could then carry out a search and decide if they want to restrict the results to just those tagged as being suitable for resource poor settings.

Quality is an important issue, therefore, we propose a quality system whereby an article would be considered ‘pending’ until it has been selected at least two further times. If a user restricts any subsequent search a pending article would be clearly marked as such.

We’ve had medical images on TRIP for many years now. They’ve not worked particularly well and, in a recent survey, we found most people were unaware we had a searchable collection of images. So, in the soon to be released upgrade, we’ve made some significant changes:

• Thumbnails improved. Previously we re-sized all the images ‘on the fly’ which caused all sorts of rendering problems, they tended to look awful. We’re now grabbing the content from the web, re-sizing at our leisure which allows us to significantly enhance the thumbnail quality.
• Quantity. Due to our spidering tool improvements we’ve gone from around 40-50,000 images to over 90,000 with a hope to take the figure to over 100,000 before the end of the year. In addition we’ll also auto-search google images and add those to the main medical image results display.
• Display. We’ll start showing 4 thumbnail images on the main results page of TRIP, see below for how it’ll look.

Nothing to do with TRIP, but an interesting article on the use of technology in medicine:

RXVantage is “Open Table for doctors” – it’s web based scheduling software that doctors and pharma/medical device companies use to calendar those visits, manage appointments, etc. It’s free for all parties to use, and premium features are available to the marketers if they choose.

Click here for the full article.

The site is getting very close to being finished and we’re entering a rigorous testing phase

Internal Testing
I had my initial look at the work on Tuesday and the points raised have now been addressed, so we have a few days of additional internal testing.

Independent Usability Testing
Early next week we will be carrying out some independent usability testing using Minervation. This involves a number of volunteers looking at the new site and being asked to perform a number of tasks. The volunteers will be videoed while performing the tasks and this will be analysed (alongside a post-test questionnaire) to highlight possible weaknesses in the design.

TRIP Advisory Board testing
Once alterations have been made based on the above steps we’ll throw the site open to the members of the TRIP Advisory Board. This should last two weeks, giving longer-term usage feedback.

The above represents the most robust testing that we’ve ever undertaken. It should be worth it and we should be able to launch early September.

While researching a question on oral contraceptives and low BMI I came across a significant number of links to the FSRH clinical enquiries that were leading to ‘file not found’. Due to their set-up they delete files from their servers after about 3 years old. TRIP does not access the files dynamically so ‘dead’ links aren’t removed and people get poor results. Therefore, I’ve manually updated the records, removing 500 (around 30%) of the records.

The issue of longevity of Q&As is problematic and there are no obvious right answers. On TRIP Answers we leave all our answers on the web, but mark when the record is over two years old. We also allow users to comment if they feel something is out of date (or if new research appears) – and this will be further enhanced in the near future. Finally, we will be rolling-out a semi-automatic system to keep records up to date towards the end of the year.

I strongly believe that Q&As (arising from genuine clinical situations) are the most powerful form of evidence. They are nuggets of evidence that clinicians can directly apply to their practice. So, I hate to lose Q&As – they’re too useful. Be transparent, warn people when they’re ‘old’ and let the user always be sceptical.