Monday, 16 August 2010

Introducing Quertle

Quertle is a new search engine specialising in biomedical literature but also covering health care. It's still very new (and seems mostly to have content which has been culled from BioMed Central and PubMed Central at the moment) but it's worth keeping an eye on. Quertle's unique-selling-point is the way it searches using "conceptual relationships", supposedly making your search results more relevant. A quick search for diabetes retrieves almost 200,000 hits, so there's certainly plenty to read.
Posted by at
Labels: , , ,

8 comments:

  1. Anonymous7 December 2010 at 20:52

    PubMed has some 366,000 hits, PubGet 402,000, Highwire some 655,000. Now what? Why is Quertle better?

    ReplyDelete
  2. Janine8 December 2010 at 00:29

    Hi Anonymous

    Thanks for commenting! I never suggested that Quertle was better than the other sources you mentioned, I merely point out that it arranges search results in a different way - which some users might like, and some might not! As you say, there are many high quality information sources to choose from.

    I'm impressed that you took the time to search all those sources - it's a very useful comparison.

    Best wishes
    Janine

    ReplyDelete
  3. Anonymous9 December 2010 at 08:24

    Hi Janine,

    what I meant is this: if one mentions a new source one should perhaps indicate why it may (may!) be better than standard resources, e.g. PubMed. Quertl is in my view an experimental resource with only sketchy explanations / background on its website. So, what can Quertl do and why / when should one use it? Giving mere quantitatve info about the number of hits for a single term subject is perhaps not very helpful.

    Thanks for the quick reply to my note...

    Regards, Reinhard

    Kind regards,

    Reinhard

    ReplyDelete
  4. Janine9 December 2010 at 08:26

    Hi Reinhard

    Thanks for commenting!

    There is some information about what Quertle does and why you might want to use it here, bearing in mind the product is still very new: http://www.quertle.info/training/Using_Quertle_Flyer.pdf - it seems pretty straightforward to me. Since the content of the database is the same as from 'standard resources' such as PubMed I don't see the problem of recommending Quertle to students who like to search differently - for example the use of Power Terms and concept linking may help students who are beginning research on a dissertation and want a broad overview of a topic and a visual display of links between concepts. Does that give you more of an idea of when one might use it? The content is still reliable, just the information retrieval is different. This blog is not the place for detailed evaluation and comparison of bibliographic databases and information retrieval systems! All bibliographic databases (including PubMed although it is often seen as the 'holy grail' amongst physicians and scientists) have flaws in the way that information is indexed, arranged and retrieved, but that debate is for another place and time. Quertle is a valid choice for anyone who likes the way the search works. PubMed is a valid choice for anyone who doesn't want or like to use Quertle. Google Scholar is a valid choice for anyone who doesn't want or like to use either of the above. Like any information decision we make, it depends on our information needs at that time and the context of our query. Does that help you? I hope so.

    I've contacted Quertle and invited them to respond to your comment. I'll be interested to see if they do.

    ReplyDelete
  5. Anonymous10 December 2010 at 06:30

    Thank you for your comments, Janine. Precisely which information needs are satisfied by Quertle and not the NCBI databases? I asked for functionalities of Quertle which might justify its use instead of e.g. PubMed. Here are a few one could mention:

    - Quertle does cross-database searches of several NCBI / NLM databases in one go.
    - Quertle can searches abbreviations, e.g. NO, which on other resources may be stopwords.
    - Quertle suggests (in some situations) additional key concepts: it may help to integrate these into further expanded searches
    - Quertle searches the full-text of e.g. PMC articles; this may retrieve additional studies

    What are the shortcomings? Let’s try a specific subject: ‘Use of botulinum in writers cramp’. A quick search on PubMed, done with ease, identifies some 30 DIRECTLY relevant studies. A search on Quertle in the form ‘writers cramp botulinum’ finds 23 studies, of which only (on cursory inspection) the first three are directly relevant. The others deal with dystonia affecting other anatomical sites, such as neck or eye. They could be called false-positives and I could explain why Quertle gets this wrong, and not trivially so.

    The concept of ‘semantic searching’ or using ontologies is still immature and may never work in medicine, as opposed to a ‘sharp-edged’ subject such as chemistry. I would guess that Quertle may work better in such desk- and lab-based subjects like biochemistry, genetics, genome research - their choice of ‘Power Terms' confirms that impression. But clinical medicine (that includes nursing) deals with a fuzzy reality which e.g. PubMed copes with excellently.

    In their PR video Quertle gives a few more very questionable example where PubMed ‘fails’, such as (stupid R.) search formulation ‘drug adverse effects’, or ‘caffeine causes what?. Have they never heard of the ‘focused question’?

    Quertle, like all PubMed ‘alternatives’, depends entirely on NCBI databases, carefully developed, organized and structured and accessible for free and EASY TO USE. Why claim otherwise? Why run NCBI / PubMed won? Quertle is one of the worst offenders. I wouldn’t recommend it or leave it to students to find out how much they miss in terms of germane quality database interfaces or relevant studies.

    In conclusion: None of Quertle’s additional functionalities outweighs its conceptual shortcomings. It’s not a PubMed alternative.

    P.S. Quote from Quertle publicity: ‘Traditional search engines … BRAUNLESSLY [my UC, R.] find keywords without considering context or position’. Calumny, calumny!

    Regards, Reinhard

    ReplyDelete
  6. Anonymous10 December 2010 at 06:31

    For BRAUNLESSLY in my previous note read 'brainlessly'

    ReplyDelete
  7. Janine10 December 2010 at 06:35

    Hi Reinhard

    Thanks again for your great comments! I love the fact you are so energised about this issue.

    As I've indicated before, I'm neutral about this issue, but I'm so glad you have put your arguments forward so others can read it.

    Thanks again and I wish you and PubMed a happy life together!

    ReplyDelete
  8. Anonymous10 December 2010 at 06:53

    Janine, thank you for finding such a charming way of ending this particular thread or your blog. Unless somebody else joins I will stay mum.

    ReplyDelete

Subscribe to: Post Comments (Atom)