Search

Trip Database Blog

Liberating the literature

Document similarity

I really enjoy it when we can fund a bit of R&D and one such project is starting to bear fruit. The overall aim of the project is to develop a ‘brain’ underlying Trip that can better deliver evidence to users. This can help in a large number of our areas of interest.

This sub-project is based on the notion of document similarity; in other words if you have a document, which other documents are similar.  A clear use case is one where you find a document that is really interesting, which other ones are most similar to it? Typically you’d keep scrolling through the results. If you have a similarity measure, it automatically finds them for you!

But there are many other uses. For instance we can take clinical questions, classify them as documents (I’m using the term document quite broadly – essentially it’s a distinct amount of text) and see what documents are most similar. Why search when the system can do it automatically?

We can also use this intelligence to keep people up to date with the latest evidence. If we know a user likes document A, we can scan new evidence to see those that are similar enough and alert the user to these.

It’s still early days and we’re still developing things (we’re using a variety of techniques including machine learning) but initial results are promising. Below is a list of text documents we uploaded and ‘asked’ the system to arrange them by similarity:

Most dissimilar
– Prostate-Specific Antigen-Based Screening for Prostate Cancer: Evidence Report and Systematic Review for the US Preventive Services Task Force
– The Society for Vascular Surgery Wound, Ischemia, and foot Infection (WIfI) classification independently predicts wound healing in diabetic foot ulcers

– Docetaxel Versus Surveillance After Radical Prostatectomy for High-risk Prostate Cancer: Results from the Prospective Randomised, Open-label Phase 3 Scandinavian Prostate Cancer Group 12 Trial
Disordered Eating Behaviors Are Not Increased by an Intervention to Improve Diet Quality but Are Associated With Poorer Glycemic Control Among Youth With Type 1 Diabetes

– Decreasing Seroprevalence of Measles Antibodies after Vaccination – Possible Gap in Measles Protection in Adults in the Czech Republic
– The Society for Vascular Surgery Wound, Ischemia, and foot Infection (WIfI) classification independently predicts wound healing in diabetic foot ulcers

Most similar
– Results of Prostate Cancer Screening in a Unique Cohort at 19yr of Follow-up
– Prostate cancer screening with prostate-specific antigen (PSA) test: a systematic review and meta-analysis

– Prostate cancer screening with prostate-specific antigen (PSA) test: a systematic review and meta-analysis
– Prostate-Specific Antigen-Based Screening for Prostate Cancer: Evidence Report and Systematic Review for the US Preventive Services Task Force

– Results of Prostate Cancer Screening in a Unique Cohort at 19yr of Follow-up
– Prostate-Specific Antigen-Based Screening for Prostate Cancer: Evidence Report and Systematic Review for the US Preventive Services Task Force

So far, so good!

 

Screen time and children – supporting decision making

We’re currently doing our monthly update highlighting new content and when doing these, some topics stand out; no idea why.  This month there have been a few on screen time in children:

The top entry, from PROSPERO, is a systematic review protocol exploring the association of screen time and sleep. The second (EPPI Centre) is an ‘evidence map’, while the final entry (CPS) is a clinical guideline.

I was interested to see that the guideline has a distinct section on sleep and if you dig deeper through the documents you seen overlap in a number of places.  Is this good or bad? I suspect that for the producers of the reviews there are ‘logical’ reasons for the content of their product. But I can only think, from a decision makers position, it’s unhelpful.  I’m not sure what the answer is, well not entirely.  Throw in to the mix all the other publications on the topic (see this Trip search as an example) and it could be seen as a mess!

At Trip we’re focused on supporting decision makers to take evidence-based/informed decisions. So, is the answer to, somehow, extract the actionable messages from these very long documents? Sounds reasonable to me, but it opens up additional issues – such as who are the stakeholders, what are their decisions and what do they need to make them!?

As seems so often the response – it’s something to ponder (is this a bit like saying ‘more research needed’)!

 

 

What is Trip’s Community Q&A (Beta)

In a nutshell it’s a system to support Trip users to obtain answers to their clinical questions. It’s not designed to replace Trip, more like compliment it. You may not have been able to fuller answer your question, be stuck on the meaning of a passage, the nuance of a paper or the meaning of a statistical test.  We’ve developed this powerful solution to harness the wonderfulness of the Trip Community (over 100,000 registered users).

The steps are as follows:

  • If you can’t find the answer in Trip you record your question via the ‘Ask a question’ link.
  • We send it to members of the Trip community who we feel are best placed to answer it. For instance, you may have a question on heart failure – we’d email it to users with an interest in heart failure and/or cardiology.
  • It is anticipated/hoped that at least one of those emailed will answer your question.
  • Quality is important and users will be able to upvote an answer (if they like it) or downvote it if they don’t! People answering the questions will be encouraged to reference their answers.

By contributing to this service as a member of the Trip community you’ll be sharing your knowledge to support healthcare around the globe and creating a system that you too may use if you find yourself struggling for an answer.

The Q&A system falls within our new Trip Learning Environment (Triple) a concept whereby users share and learn – in a non-critical manner – from each other. The Q&A is the first system we’re introducing but others will follow afterwards.

The system is a beta as we’re wanting people to test it to see that it works well and highlight any problems (a beta signifies a non-finished, test version of a feature). This should not be onerous, we simply want you to join up to the system and have an explore. It would be great if you asked a question or responded to a question should you know the answer. The test will be over a month or so and we are not anticipating lots of activity – just enough for people to get a feel for it.

Email me via jon.brassey@tripdatabase.com if you’re wanting to help test it.

Trip’s community Q&A service – request for volunteers to test the system!

Trip’s main focus is on answering the clinical questions of health professionals. Searching Trip is great – but not perfect. Our community Q&A is there if you can’t find the answer you want/need.

The concept is simple, a user posts a question and we send it to the health professionals – registered on Trip – who we feel are best placed to be able to answer them.  Hopefully one or more will respond.

We’ve previously asked our users about this idea and the vast majority were highly supportive and gave us excellent feedback – which we’ve incorporated into the design of the service.  So, the next step is to test the system we’ve built.

Please let us know if you’re interested in testing the system. Time commitments is up to you and should be minimal. You can ask questions, answer questions or simply observe the system working. Testing should start within a week or two.

If you’re interested please email me jon.brassey@tripdatabase.com

Eight new publications added to Trip

Trip is loved because it’s easy to use and allows users to quickly identify high-quality evidence. A clear prerequisite is to have lots of high-quality evidence! As such we’re always on the look out for such content and this month has been a bumper month – with the inclusion of eight new publishers. Many of these were identified by reviewing the latest CADTH Grey Matters tool.

And, one request for you, if you know of any sites which should be included in Trip, please let us know (via comments or email jon.brassey@tripdatabase.com)

Below is the list of newly added content:

  1. Palliative Care Evidence Review Service (Rapid reviews)
  2. EUnetHTA (HTA)
  3. SickKids Reports (Reviews)
  4. SickKids Supportive Care Guidelines (Guidelines)
  5. Washington Health Care Authority (HTA)
  6. Health Information and Quality Authority (HTA)
  7. European Society of Endocrinology (Guidelines)
  8. Pediatric Endocrine Society (Guidelines)

NOTE: These have only just been uploaded and will take up to 72 hours to be indexed and searchable!

Answering clinical questions – positioning survey

Last week I highlighted our plans for a Q&A system to harness the combined intelligence of the Trip community.  In that, I identified two main challenges:

  • Technology – to make the system work
  • User acceptance – how can we explain the system in a way that users can relate to and make sense of (and ultimately use)?

The latter challenge seems much harder and to help us we’ve created a brief survey which we’re hoping as many of you as possible will complete. It should take around 2 minutes – so hopefully time is not a barrier to use.  To undertake the survey please CLICK HERE.

To repeat, please take the survey (CLICK HERE) – your opinion really does count.

Answering clinical questions

Trip started as a service to support clinical question answering. Since then the site has changed dramatically but we still see clinical question answering as the main drive and focus.

But what happens if Trip lets a user down? In other words the user has a question, uses Trip and can’t find an answer! In the last two days we’ve received two questions from users:

  • EBM causation of intervertebral disc herniation?
  • Dexmedetomidine and brain protection: fact or fiction?

Now, I dare say the answer is somewhere in Trip and I imagine that terms like ‘brain protection’ may need clarifying. But, at the heart of it – for whatever reason – the users are struggling.

So, we return to the notion of using Trip’s community to help answer them. There is no doubt that there are people, who are registered on Trip, who will know about disc herniation and dexmedetomidine. They may well be able to easily answer these questions. So, why not send the Qs to those most likely to know the answer?

Well, you may have guessed it, but that’s the plan, to create a community Q&A system. There are a number of challenges, for instance:

  • Technology: linking the Q to those most likely to be able to answer them. We’ve engaged with an external company to work on this problem and we’re feeling very confident that good progress will be made.
  • User engagement: we know many people love Trip and we know we have many registered users (over 100,000) but will a large enough proportion be willing to help their fellow users? We’re thinking yes! But we need to communicate really well to explain what we’re doing and reduce any barriers to community involvement.

No timeline, but now we’re committed I’m hoping we can motor!

 

New search, new server

We’ve just upgraded to a new and more powerful server system – lots of extra memory and speed.

As part of the new system we’ve also improved two other things:

  • Upgraded the version of our search engine technology to the latest version. This has improved many things including my own bugbear – duplicate terms in the title boosting the result. In the old version if a document had the search term(s) mentioned twice in the title it would get a large boost. It didn’t happen often but it was annoying! But it appears that’s been fixed.
  • Improved our indexing of PubMed.  Previously we’d been grabbing all the content from PubMed via eUtilities and we only realised recently that this included all the additional meta-data including things like related articles and cited articles data.  This created false positive results.  Our new indexing will now only grab content from the abstract.  This means results counts are both reduced and more accurate (less noise).

We’ve tested it extensively but if you spot anything strange please let us know.

 

NOTE: I can see the new server but it might take 24 hours for the internet’s address book to point at the new version. If the ‘Evidence Maps’ link at the top of the page is highlighted yellow, you’re still pointing at the old version!

Include all of PubMed in Trip?

Currently Trip includes a sub-set of PubMed. This includes all RCTs and systematic reviews and the content of around 500 ‘core’ journals. In total this represents around 10-15% of all the PubMed content. Putting it another way we have around 3-4 million PubMed records out of around 28 million.

But, to be clear not all records are equal.  Some will be to do with veterinary medicine (e.g. Serum amyloid A and plasma fibrinogen concentrations in horses following emergency exploratory celiotomy)  and others might be very specialised technical articles (e.g. Phase-matched virtual coil reconstruction for highly accelerated diffusion echo-planar imaging). So, we’ll be potentially including extra noise. We can mitigate against that – to a point.

In summary, searching all of PubMed is potentially very useful but there may be downsides. So with that in mind what do you think?

Blog at WordPress.com.

Up ↑