The Search Principle blog

The Canadian Medical Association (CMA) has been in the news: first, the announcement that an expert panel has recommended editorial independence for the CMAJ and second, the race for who will be CMA President – the controversial advocate for private healthcare, Dr. Brian Day or Dr. Jack Burak? Oddly, the BC Medical Association endorsed Day. Ah, medicine – can you say politics?

Physicians have a very strong, collegial system where free speech can get quite salty (ever been to a departmental meeting?). It may seem completely boring in comparison, but here is my pre-pub evaluation of the CMA public website, which will be published in the Journal of Consumer Health on the Internet in Fall 2006.

In the meantime, from the article’s conclusion:

“The Canadian Medical Association (CMA) may yet experience a resurgence of positive public opinion in the months ahead as consumer health moves into a post Consumer 2.0 -era. CMA should consider using collaborative blogs, RSS feeds, and an Ask-an-expert feature to connect with health consumers on its public Web site. The editorial scandal may not have registered for most Canadians, and by revamping its public Web site and moving more into providing information to consumers, the CMA could rehabilitate its image. In due course, Canadian consumers might even realize what a rich resource of reliable information they can find at the CMA public Web site”.

It’s not Christmas for months yet, but for many librarians commited to open search – it feels like it after hearing about plans for more access to OCLC’s OpenWorldCat. InfoToday’s announcement suggests that OCLC is poised to make it easier for searchers scattered around the globe to search across the content of libraries. Does this translate into a coherent global search tool for print items held in libraries?:

“Away to my search tools I flew like a flash,
Searched WorldCat and Yahoo and threw open the sash.”

OK, OpenWorldCat is coming soon, in August 2006. In the meantime, you can start to search for these records in search engines like Yahoo this way.

This moves us bibliographically closer to being able to search across the entire corpus of human knowledge in a free and open manner, and not just for materials Google and other search tools deem worth digitizing – or deem worthy merely because they are popular. (You know, PageRank and all that.)

Gunther Eysenbach’s cogent editorial “The Open Access Advantage” over at the open-access Journal of Medical Internet Research hits several points square on the head. The one I find most appealing is that incremental benefits accrue to researchers who publish in open-access journals; the data is clear and unequivocal. Another is the simple truth that open source tools like those developed by UBC’s Public Knowledge Project are better alternatives to the commercial OA publishers.

Eysenbach also articulates an important distinction between OA giants like BioMedCentral (one of the so-called commercial or “for profit” OA publishers) and the leaner, meaner OA pioneers like CMAJ and BMJ. Though, I’d be more inclined to applaud outfits like PubMedCentral than the journals or websites he mentions.

And speaking of PubMed, it truly is a pioneer in open search, something health librarians have been advocating since we “let go” of our paranoia that end-user searching was a threat to our profession. Open search merged with open access.

There is little point in making scientific research open access in the absence of coherent open search. That means, a unified search tool – not fragmentary attempts that we’ve seen in 2006 by Google (sorry – I’m really hammering Google these days) and others.

Some new open, forward-thinking pioneer will find the solution – but let’s hope it’s sooner rather than later.

Before health librarians can lead the charge into the brave new, evidence-based future, we need to build (and maintain) expertise in the principles and practice of health librarianship, as well as tracking and staying current with technological trends.

If the new generation of LIS grads are any indication, the field is in good shape. Recent MLIS grad (and SFU summer student) Jeremiah Saunders and I collaborated on Podcasting: the future of staying current in medicine, and are now looking at setting up a suitable techno-wiki. From Jeremiah’s blog, read his synopsis Your Mission: Choose a Wiki for your Library.

Fraser health librarian Greg Rowell and I are teaching LIBR534 this fall for SLAIS, and are looking into a suitable wiki to build a knowledge-base for health librarians that could be the basis of future collaboration across hospital libraries and LIS schools in Canada, and the U.S. Stay tuned.

I’m beginning to think search companies don’t get it. The medical search space used by doctors is different from the general search space. Google scholar, still in beta, has scaled up in size, but not much in terms of functionality, since November 2004. In fact, this summer I sense physicians moving away from Google scholar. Other missteps: Google Coop is a non-entity, and Adam Bosworth and his team at Google are said to be working on a Google health scrapbook – Google health revisited?

The solution for consolidating medical search? A portal for evidence-based practice. Maybe Jon Brassey will get it right. Physicians and other health professionals need simple tools to find the best appraised evidence, and they need MEDLINE, the open web, other primary research and open access journals. Forget the vendors, the proprietary databases. Open up access to medical evidence, make it freely-searchable for all to see, around the world.

Why is this so difficult? Maybe I should e-mail Warren Buffett.

In this week’s Biomedical Digital Libraries (BDL) – I sit on its advisory a conflict? – there is an interesting article that compares citation counts in Web of Science from Thomson Scientific, Scopus from Elsevier and Google Scholar from Google. Bakkalbasi et al compared citation counts in two specific areas of science: oncology and condensed matter physics. Were these two disciplines representative somehow?

The article is entitled “Three options for citation tracking: Google scholar, Scopus and Web of Science”, and is worth a read.

The authors conclude that “All three [search] tools returned some unique material. Our data indicate that the question of which tool provides the most complete set of citing literature may depend on the subject and publication year of a given article.” That’s self-evident; the dissimilarity of tools is a priori evidence of the authors’ conclusions.

This paper makes some assumptions based on an assessment of two very narrow subjects. In that sense, the paper has limited value. What I like about the paper is that it attempts to compare three dissimilar citation tracking tools, which is much appreciated. The conclusion that all three tools are needed is defensible, though it could be argued that each search tool is used for quite different purposes.

The challenge with comparative studies like this? Study design is difficult. The problem (or challenge) is the issue of coverage in each source, which is markedly different. And, in the case of Google scholar, we don’t know what is covered.

It seems obvious to librarians that Google scholar is not a reliable citation tool (though the cited by feature is generally viewed as useful). “Cited by” in GS is only meant to be an indicator of citedness, and to lead to other scholarship. Results should be contextualized by comparing the same searches in WoS.

Conclusion? Even though it too has limitations, WoS is still the best indicator of citation counts for academic papers, across all disciplines. (I am unable to comment about Scopus as UBC Library does not subscribe to it.)

Today’s post – written midway through 2006 – is a riff on ideas I’ve knocked around the blog for months. As such, it contains more questions than answers, but perhaps readers will enjoy an EBM 2.0 brainstorm over the holiday.

At its essence, medicine is information-intensive. Charts; patient histories; research literature; evidence. Though not a new concept, data distillation in medicine would integrate epidemiological and patient information with EBM. In a sense, data distillation is already a part of the systematic review process, and meta-analyses. Sherrilynne Fuller’s keynote at the CHLA/ABSC 2006 Conference, for example, discussed “integration” of EBM and the electronic patient record.

But let’s look at the integration of knowledge-based systems beyond EBM. Clinical decision-support tools (like DynaMed and UpToDate) assist physicians by synthesizing medical literature – but they are really only reviewing a tiny sliver of available data, and much medical evidence is ultimately lost. Ideally, data distillation integrates all available medical information; epidemiological data is brought together, and pertinent information collated from across time zones, systems and databases.

How might this be useful? Emerging disease warning systems could be developed; patterns in patient records would be more easily seen and described, and specific sub-populations treated, early. Think how this would save lives. (Think of the 1980s, and how lack of pattern recognition in patients led to the spread of AIDS.)

One large question mark is how genetic data fits into the picture? How can the human genome be better integrated with MEDLINE and other databases on the web (including the grey literature in the deep web)? There are some who believe that the U.S. could have averted 9/11 if various security agencies had shared data beforehand. Did a lack of pattern recognition by authorities lead to 9/11?

Recognizing patterns among disparate sources of medical information is critical to the future of human health. New forms of data distillation require better relational databases, massive networks that can process data quickly and efficiently, and data mining tools that recognize patterns across various populations and geographies.

At the other end – output – medical librarians must continue to present information to physicians in usable, digested formats – ie. distilled – to save time. Data distillation is an area librarians can direct their energies as we move further into the information age, and further into digitized biomedical libraries.

Our future as library professionals in medicine should continue to comprise global perspectives beyond what our current, limited databases provide for our users.

Renaissance medicine must have been a nasty, toxic brew – between plagues, bloodletting, religion, wars and quackery. For the psychology of illness (and views of the human condition), the Complete Works of Shakespeare is a rich resource of poetry and insight. Google understands this simple notion by emphasizing Shakespeare through Google Books.

In the Winter’s Tale, Camillo suggests that humans can spread infection, even though they may remain asymptomatic:

“There is a sickness
which puts some of us in distemper, but
I cannot name the disease; and it is caught
Of you, that yet are well.” – ~ Camillo from the Winter’s Tale

Scholars believe that Shakespeare’s knowledge of medicine was a result of knowing John Hall, a physician and herbalist who married Susanna, the Bard’s eldest daughter.

In The Medical Mind of Shakespeare, Aubrey C. Kail describes efforts to contain plague. “Special officials called searchers went to houses to search out victims and were paid more if victims were found dead.”

Searchers provided Shakespeare with a plausible explanation for a significant development in Romeo and Juliet. Friar John, suspected of being in an infected house, was shut in by the searchers and was thus prevented from carrying the all-important message from Friar Lawrence to Romeo. No messenger could be found to return the letter to Friar Lawrence, so afraid were the citizens of Verona of the infection.

Here is a link to a simple PubMed search on medicine and Shakespeare.

Google Scholar has been mysteriously quiet, of late. The post-library conference lull is almost upon us, and yet the article The Depth and Breadth of Google Scholar: An Empirical Study published in portal: Libraries and the Academy, for my tastes, pierces the summer silence. It’s worth a read if your library has it.

Authors Neuhaus et al have examined 47 databases, using a number of tests to compare content, depth, language(s), strengths and weaknesses with Google Scholar. Allan Cho, a UBC library school student at SLAIS, has written a good overview of the piece (even quoting his SFU supervisor). A UBC health librarian-on-the-rise is Eugene Barsky who spoke to a posh group of SLA and CLA librarians about Google Scholar and the World of Scholarly Searching.

The best part about all of this? There are other librarians interested in Google Scholar. I’m getting close to feeling like I need to concentrate on other matters in search – like clinical decision support tools for residents. More about that, in due course, of course! Residents start at the hospital in less than two weeks – Dean

NCBI’s David Lipman mentioned six months ago that PubMed users should watch the UI (user interface) change over the calendar year. Today’s introduction of the display format AbstractPlus is a positive step, and comes after positive changes to the Limits function.

The result? Easier navigability, and quick linkages to related literature (five articles at far right) is a plus.

ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=PureSearch&db=pubmed&details_term=%22 Antibiotic%20Prophylaxis%22%5BMeSH%5D%20AND%20%22ticks%22%5BMeSH%20Terms%
5D

….just use DISPLAY pulldown “AbstractPlus”.