Databases

Usage Stats: Are They a Double Edged Sword?

Back in the olden days a library bought a subscription to a journal and they paid the institutional price which was often listed on the inside cover of the printed issue.  It was always more expensive than the personal subscription, but there wasn’t tiered pricing, FTE pricing, or pricing based on inpatient admissions and number of specialists.  For the most part the price you saw on the inside cover was the price you paid.

Then came the electronic journal.  At first journals weren’t quite sure how they were going to have their articles online.  Some gave it away free, others were free with a print subscription, some charged a nominal upcharge, while some charged a specific online journal price. Ejournals grew in usage and with tightening budgets librarians began dumping the duplicate print. During that time institutional prices evolved to a Ladon of possibilities.

Additionally, the concept of eresources has moved beyond journals. It extends to books, databases, integrated EMR and patient education products, image databases, etc.  As librarians we demand to know our usage statistics for our eresources.  We need to know what our patrons are using so we can get the most bang for our buck.  However, we aren’t the only ones who see our usage statistics.  The vendors that sell us our products run the reports and it isn’t in their best interest for us to get the biggest bang out of our buck.  I am not trying to imply that all of the vendors are nefarious. I am just saying that if they see that your cost per use stats are so phenomenal that they may be looking how to get more money from you.  For example you are paying $50,000 for a product that you use so often that you have $.05 per use but the average library in your tier pays about $.10 per use,  the vendors think you are getting their product for a $50,000 discount compared to others in your tier. 

Prior to eresources, vendors knew very little about the usage of their product in the institution.  The usage of printed journals and books were often only known by the librarian through shelving studies or circulation statistics.  I remember when we had CD Plus and had to load the MEDLINE CDs on a CD tower for people to search.  Despite not having the type of usage data we have to today, librarians still looked at how their databases were used (Volkers AC. Bull Med Libr Assoc. 1995 Oct’ 83(4):436-9.) and even tried to determine journal needs through the database (Dunn, K. Medinfo. 1995;8 Pt 2: 1428-32.) The usage stats were all in house.  So while you might have known what your cost per use was for a journal, book, or database there was no way that a vendor knew, unless you published it in a journal article that they read. 

It seems that with wide scale use of eresources, usage stats have become a double edge sword.  Not only do we still need to know what is being used but vendors now also know what we are using.  They can use this information to their advantage as well.   While neither party wants to have a resource that is a dud, I’ve got to wonder if we are now also victims of our own success.  Many of us have already cut the chaff from the wheat years ago.  All of our eresources are high performers.  Yet because they are high performers are they costing us more than if they were less utilized?  If so isn’t that the exact opposite of what a librarian needs to be thinking about? 

Betsy Kelly, Claire Hamasu, and Barbara Jones wrote an interesting article, “Applying Return on Investment (ROI) Libraries. (Journal of Library Administration. 2012;52(8):656-71.)  Determining the ROI is necessary to measure the value of the library resources to the institution.  Many medical librarians use the NN/LM MCR ROI Calculator to determine the replacement value of services provide by the library.  In addition to quantifying the number of classes, room use, photocopies, and ILL’s the calculator can also factor in the cost of ejournals, databases, ebooks and their usage.  So in order to get a good ROI we want high usage for these electronic resources. 

ROI is what hospital administrators are looking at when it comes to everything.  Hospital administrators are focused on controlling costs and demanding the biggest savings possible.  According to an article from the Daily Beast about the Cleveland Clinic , CEO Dr. Cosgrove is described as something of a “fanatic” regarding controlling costs.

“Our physicians are so engaged in our supply chain that they help negotiate the price down for the things we use,” Cosgrove told me (Daily Beast), and reeled off a list of examples:

• When I was the head of surgery, we needed a new heart-lung machine, and we decided there were three models that could work, so we did a reverse auction to get the lowest price.
• We put price tags on things in the operating room: before you open that $250 set of new sutures, make sure you actually need it.
• We found out that there’s a lot of redundant tests that are done, or tests that won’t be vital to the patient’s care. We know that there are some things that don’t change. For example, the reticulocyte count can’t change but week to week. So if someone’s ordered a reticulocyte count, you can’t ordered another for a week.”

I might be going out on a limb here, but I have to think that all administrators are pretty fanatical about costs and keeping them low.  So how does the idea of keeping costs low factor in with eresources?  Are we at a point with some resources that good usage is actually hurting us, costing us more come negotiation time (if we can even negotiate)? In the spirit of the $250 suture kit, do we start adding a price tag to our eresources before users click on them?  That would be kind of absurd and certainly would drive down our usage stats which in turn would drive up our cost per use. 

In this day and age where we use our usage statistics to drop resources and vendors use them to determine pricing, how are we to come to a even playing field when our budget is shrinking and our administrator wants to see increase cost savings?  We struggle to show our ROI on a smaller and smaller budget as our resources increase in price.  We explain to administration that if they didn’t have us to do what we do it would actually end up costing them a lot more in time and money to provide the same resources and services.  But as Kelly et al mention, the “problem with ROI calculations based on cost avoidance is the underlying assumption that users will look elsewhere to purchase the same services and resources they receive from the library. It is not realistic to assume that users could afford or would make the effort to personally pay for all of the services they receive.”   Hospital administrators are essentially already doing this.  By cutting the library’s budgets to the bone they are forcing librarians to not pay for all of the same services and resources.  When a hospital library closes, the budget for those electronic journals, books, and databases (as well as everything else) is gone.  Almost none of the resources are kept by the institution. When administration closes a hospital library, they are not replacing the same services and resources. 

Usage statistics help librarians determine ROI to hospital administration, but what are we to do when administration wants to see usage and ROI go up but vendors increase the price (thus decreasing our ROI) as a result of our usage stats?  It seems as if librarians are between a rock and hard place. Do we need to look at another method of valuing our services and resources?  If so, what?

2 comments - What do you think?  Posted by KraftyLibrarian - March 19, 2013 at 10:45 am

Categories: Databases, Electronic Access, Journals and Books, Library Products (not Databases), Technology   Tags:

MDConsult and ClinicalKey

Much was posted on Medlib-l when Elsevier announced their decision to drop MDConsult for ClinicalKey.  I can’t say that I am surprised by the decision because I didn’t really see the company keeping both products in tandem especially when it appears that ClinicalKey ate MDConsult’s content (and then some).

In August I wrote a brief review of ClinicalKey on my blog, and a longer more extensive (and more directed at doctors) review on iMedicalApps.com.

A few things have changed and have been updated since August and since the latest news of MDC’s departure I thought I would provide an updated brief review focusing a little bit on the differences between MDC and CK.

MDConsult had a core collection/subscription as well as specialty collections/subscriptions.  I have heard rumors and conjectures that ClinicalKey might split up their ginormous collection according to subjects.  If this were the case I would assume (total speculation!!) that the ClinicalKey core would be the entire kit and caboodle while their specialty parts would broken down by subject.  As I mentioned, this is pure guessing on my part, I have no insider knowledge and the idea that CK might be split up and sold in parts is just rumor.  However it were true it would be very helpful to community and smaller hospitals who might find the sheer quantity of information within CK to be overwhelming.  Really libraries will have to look at their usage stats for MDC to see if it even warrants thinking of CK as an option.  I know of some hospital libraries that just don’t use MDC enough so it would be very tough for them to justify CK.

MDC has many online textbooks, but CK has a lot more (approximately 1000 titles).  As I mentioned in a Medlib-l post there are some titles in MDC that are NOT in CK, but they are fewer than 10.  Diane Bartoli, VP, Global Product Development, responded to my Medlib-l post that there are only 9 titles not in CK.  Since we only subscribed to MDC’s core collection I wasn’t exactly sure how many were not in CK (I mentioned we noticed about 8) so I was hesitant to also include in the Medlib-l email that I noticed some of those titles were old.  That falls in line with Diane’s comment that those titles (ones not in CK) did not meet their standard for current, premium clinical content.
Khatri’s Operative Surgery Manual (2003) being replaced with Khatri’s Atlas of Advanced Operative Surgery (2012) is a great example. The title list of CK books can be found here (click on Master Content List). 

With 500 journal titles, CK carries many more journal titles than MDC. However, the backfiles for CK do not go as far back as MDC.  Most CK titles only go back to 2007 which is a significant backfile loss to any library that planned their journal subscriptions around MDC.  Determining whether you need to go as far back as MDC did and the rest of your current journal subscription needs will require very careful planning and attention to license agreements from your Elsevier journal providers if you decide to move forward with CK. 

The PDF login requirement is one of the more significan’t changes since my last review.  Thankfully Elsevier removed the requirement for full text journal articles and they worked with several libraries to improve HTML access to the book chapters and lessen confusion surrounding logging into the PDF.  However institutional users still are required to login using their personal login to access the PDF of books. 

The PDF login issue, is one that I think librarians and Elsevier will just have to agree to disagree on.  Diane mentions according to their data that the PDF login requirement isn’t a deterrent.  Yet librarians are fielding calls and emails from users on accessing the PDF.  A library would have to look at their MDC usage stats for the books that are both in MDC and CK and compare them to their use in CK.  It does no good to look at the aggregate PDF stats for CK’s books when they have so many more titles than MDC.  To get a true picture of whether the PDF login is a deterrent, librarians also need at least 6 months to 1 years worth of data.  Without having 6 months to 1 years worth of CK data, it is difficult to make a definitive conclusion as to whether it is truly a barrier.  Perhaps somebody will think to look in 6 -12 months and see the usage and report back.  Would be interesting poster for a meeting?

I reported on Medlib-l that the displayed listing of titles across web browsers is erratic. Clearly it was a file upload and certain browsers just didn’t know what to do with apostrophes.  Additionally, some titles include the author in the title while others don’t.  The MDC title list is definitely cleaner and easier to browse through.  However, Diane did report that they will be fixing these formatting issues in the next release.  Equally helpful is that they will be adding an “enhanced author title search function” to make it easier for people to find specific books via title or author.  This feature is planned for the Q2 release. 

Special textbook content is in CK not in MDC.  Recently there was a big discussion on Medlib-l regarding online only content and printed texts. The original poster was complaining about an LWW book.  But this practice is common across several publishers including Elsevier.  We have purchased several printed text books where important parts of the book are “missing” and only available online.  Some examples of missing pieces: chapter missing from printed book is only available online, references online only, updates online only, etc.  In Elsevier books this content was available via a scratch off code for use on StudentConsult, ExpertConsult, etc.  The content was NOT in MDC.  However, in CK all of the “missing” content that was only available via StudentConsult or ExpertConsult is in CK. 

MDC did not have an app and so far there is no CK app. So they are about the same from that perspective.  However as Diane mentions CK will be mobile optimized soon.  That should be very helpful as mobile devices are invading the work place.  For people who have FirstConsult or who have Clinical Key, and want to use FirstConsult as an app.  It is free and just needs your personal login (FirstConsult, ClinicalKey, MDconult valid users).  Curiously if you have a Science Direct login and you have ClinicalKey, your login could be your Science Direct login.  That might be confusing for some people.

I want to say that I really appreciate Diane’s response to my Medlib-l post.  It was very informative.  I just wanted update my brief review and MDC vs CK comparison on my blog because I know some people don’t subscribe to Medlib-l. 

5 comments - What do you think?  Posted by KraftyLibrarian - January 28, 2013 at 2:01 pm

Categories: Databases   Tags:

Internet Explorer, PubMed and the End of the World

If you haven’t heard about the  Mayan civilzation’s calendar predicting the end of the world on December 21, 2012, then you have been living under a rock.  Personally I believe the Mayans were on to something.  Instead, I believe the end of the world will happen on January 1, 2013.  Why?

As of January 1st NCBI will no longer support Internet Explorer 7 and all the hospitals that haven’t upgraded will begin to have problems searching PubMed. You can’t blame the Mayan’s for not warning us.  I think they were pretty close to their prediction considering that browsers and the Internet were not known in AD 250.  I just think all of the doomsday prophets just translated things wrong (wouldn’t be the first time that happened). 

The end may not come as a big bang right on the New Year, but as NLM makes enhancements and changes to NCBI the people in the IE 7 hospitals will begin to have problems with PubMed.  www.ncbi.nlm.nih.gov/guide/browsers

The compatibility issue is just going to continue on.  The newly launched PubReader hasn’t even been tested on Internet Explorer and from the looks of the browser compatibilty chart they aren’t dilly dallying around with IE 8 or 9, if they design for Explorer they are going straight for IE 10. 
www.ncbi.nlm.nih.gov/pmc/about/pr-browsers/

Since PubReader was “designed particularly for enhancing the readability of PMC journal articles on tablet and other small screen devices,” the compatability for desktop and laptop browsers may not be an issue for a while. 

But this brings up the issue of IT departments needing to update the browsers.  Many librarians I have communicated with have expressed how getting IT to upgrade anything (including browsers) is a monumental task.  Just from my average web browsing it seems to me that a lot of web sites are jumping from IE 7,8 to IE 10.  Even more frustrating/interesting for hospital librarians is that there seems to a growing number of people not even designing for Explorer. 

Knowing who is winning the browser wars is tricky and getting good data on browser market share really depends on the site that measures market share.  Network World’s article “Browser battle: Chrome vs. Firefox vs. IE vs. Opera,” says “it’s difficult to say who’s on top in this four-way scrap. For one thing, different methods of measuring market share often provide very different numbers – while data from NetMarketshare.com shows IE in front with 54% of the market for October 2012, StatCounter gives a slight edge to Chrome, about 35% to 32%. W3Schools’ information paints another picture again, showing a big lead for Chrome (44%) over about 32% for Firefox and just 16% for IE.” 

It may just be me and my apocalyptic Mayan frome of mind but I am thinking of the Thunderdome for browsers. Although saying “Four browser enter, one browser leaves” isn’t as cool as Tina Turner’s line, “Two men enter, one man leaves.” 

Basically with the amount of browsers vying for for top spot it makes it difficult for us and IT to keep up.  So it is easy to see how people can be in this predicament.  So instead of stocking up on food and water in anticipation of the end of the world, start working on your IT department to upgrade your browsers.

8 comments - What do you think?  Posted by KraftyLibrarian - December 12, 2012 at 10:19 am

Categories: PubMed, Technology   Tags:

Confusion on NLM’s Policies for PMC, PubMed, and MEDLINE Inclusion

Over at the Scholarly Kitchen, Kent Anderson writes of his frustrations regarding PMC, PubMed and MEDLINE and non indexed journals (particularly the start up journal eLife) in his post, “Something’s Rotten in Bethesda — The Troubling Tale of PubMed Central, PubMed, and eLife.”

I find myself both agreeing and disagreeing with Anderson. 

I agree there is a big problem with the blurring of the lines in the minds of most people (mainly doctors and researchers) regarding PMC, PubMed and MEDLINE.  PubMed houses the citations for PMC articles as well as the citations to articles in journals indexed within MEDLINE.  The problem is that to most normal people PubMed = MEDLINE.  I mention the blurring of the lines between PubMed and MEDLINE in post “Back Door Method to Getting Articles in PubMed: Is Indexing so Important?“  In my post I mention that doctors and researchers think of PubMed and MEDLINE as the same. I likened it to ordering a cola.  ”PubMed and MEDLINE have become the Coke/Pepsi of medical databases.  Two different products but people use the terms interchangeably when ordering a cola soft drink.”  I even posted the email of a researcher friend further illustrating how they don’t distinguish between PubMed and MEDLINE and if the article is PMC it is in PubMed and that in their minds it is in MEDLINE.  At the time of my orginal post I questioned the point of actually indexing journal articles since researchers don’t search by index terms and they erroneously think PubMed is Medline.  All they have to do is get into PMC and it can be found in PubMed via keywords (which is how everybody searches these days).

Anderson’s main argument is NLM is acting as competitor to publishers and technology companies, by allowing certain journal publishers to bypass rules for inclusion into PMC and PubMed.  In his argument he brings up the journal eLife a “fledgling funder-backed journal” that was allowed include articles in PMC despite not having published the required 15 articles, not being indexed in MEDLINE, and PMC acting as the sole provider of the articles.  Not only is NLM circumventing the rules for inclusion to its databases but he believes that NLM is acting as the primary publisher to eLife because their articles can only be found on PMC.  Anderson uses JMLA and Journal of Biomolecular Techniques as other examples of journals that NLM acts as the primary publisher.  I don’t know anything about the Journal of Biomolecular Techniques but I disagree with JMLA serving as an example similar to eLife.  As I mention in my comment to his post on Scholarly Kitchen, JMLA has been around since 1911 so it has fulfilled the 15 article requirement and is published by a publisher (who is not PMC) and sends me the print 4 times a year. The journal is indexed in MEDLINE (since 2002).  Additionally the printed edition clearly states that the digital archives of JMLA are on PMC.  I went to PMC today (October 22, 2012) and the October 2012 issue is not available.  So the most recent issue is not online and PMC is acting as the a digital archive. Therefore NLM is not acting as the publisher of JMLA in the way he describes. In the case of JMLA NLM’s PMC is the secondary publisher that he describes, which is the case of many indexed MEDLINE journals. 

Unfortunatley I think Anderson’s argument misses a bigger issue.  The question of quality within the PubMed database.  As I mentioned there is confusion among PMC, PubMed, and MEDLINE.  People searching PubMed will find an article from the PMC that is in a journal that is not indexed in MEDLINE.  However people will erroneously think the article and and journal are in MEDLINE when in fact they are just in PMC.  By allowing non indexed journals into PMC, NLM is basically allowing a back door into PubMed, and by default into MEDLINE.  Of course NLM doesn’t see it as that, because they are one of the few people who can still see a distinction between PubMed and MEDLINE.  Their users (doctors and researchers) do not see the distinction.  To them PubMed is MEDLINE.  This calls into question the quality of the articles in PMC in journals that are not indexed in MEDLINE.   If the journal isn’t good enought to get into Medline then why is the article good enough to be found in PubMed? 

5 comments - What do you think?  Posted by KraftyLibrarian - October 22, 2012 at 1:18 pm

Categories: Journals and Books, Medline, PubMed   Tags:

PubMed Won’t Support IE7

According to the NCBI Website, the NCBI resources will no longer support IE7 after January 1, 2013.  So medical librarians in Internet Explorer institutions, now is the time to start talking to your IT departments to get IE 8 or greater, or Firefox 4 or greater installed on your computers.

This means that after January 1, 2013 if you are searching PubMed using IE7 and something goes wonky, it is your problem.  They will no longer design PubMed so that it will work using IE7.

For information on other browsers like Chrome, Safari, and Opera check out the NCBI browser check list.

On a medical librarian related note, MLA’s site gets fussy using IE 7 and 8.  Bascially Internet Explorer is a frustrating browser to design for because there is designing for the rest of the web then there is designing for IE.  I am told IE 9 is better, but it is probably worth trying to get Firefox or Chrome added to your computer if you can.  However, it all depends on your IT department and their policies.  So pick your battle to at least make sure whatever browser you are allowed to use is compatible for the NCBI resources.

**Update 8/21/12  10:48am)**

ACGME won’t work with IE7 after September 1, 2012.  For more information on other browsers and versions go to:

www.acgme.org/residentdatacollection/documentation/browser_requirements.asp

Be the first to comment - What do you think?  Posted by KraftyLibrarian - August 21, 2012 at 10:08 am

Categories: Medline, PubMed, Technology   Tags:

ClinicalKey Brief Review

Over on iMedicalApps.com I wrote a brief review on ClinicalKey.  The post is direct more towards doctors rather than librarians but it is worth a read through for librarians. 

Over here I thought I would include a few things that librarians might find interesting about ClinicalKey.

I have to say that the front page of ClinicalKey where it has two large squares “For Institutions” and “For Individuals” kind of confusing.  While testing the product we had a couple of librarians who kept clicking the “For Institutions” square to try and search the product.  But they were confused because they kept getting into information about ClinicalKey for Institutions.  They thought they were supposed to click the “For Insitutions” button to search because they were in an institution.  We had to explain to them you just type in the box and hit the search button.  As a a librarian who often fields questions from users, I have a feeling this will be confusing for our users.  I am not sure which blue box they will click on but the boxes are just so large and so close to the search box it implies there are different accounts for different users.  I understand Elsevier wants to market ClinicalKey to both groups, perhaps they can find a less confusing design to do that.