Print

Print


Yesterday I read Yet-Another-Article (published in an important medical
journal that we all know and get) where someone did a big search on MEDLINE.
The author bemoaned (with statistics and charts and graphs) about how
"Medline missed x- percent" of thus-and-so    -- and this is why a whole new
database of blah blah blah ought to be funded, invented, maintained,
promoted, and made accessible for all the doctors in the world.
 
The search strategy was displayed in the article.  The search was not done on
MEDLARS, DIALOG, BRS or any other large file-vendor where all the necessary
MeSH-explosions and other crucial search-tricks are at the disposal of
experienced, savvy searchers.  It was done mostly with text-words.  There was
no consultation or strategy-making with an experienced searcher.
 
But:  "Medline missed" all sorts of important & relevant references, and so
it is thus hopelessly inadequate to serve these newly defined needs blah blah
blah...
 
(...OK.  That was the background and here comes the flame....)
 
Folks: we must, with renewed diligence, get the under-informed to realize
what MEDLINE is and what it is not.  First & foremost, it is not a living,
sentient being.  MEDLINE doesn't "miss things" nor does my new Subaru station
wagon if I steer it into a tree.  It's not a glowing, infinite, exhaustive,
'last word', medical god-oracle-thingy...   It has come to be known to the
public and the medical establishment as an infinitely large, generic, perfect
and absolutely complete, mysterious, all-knowing entity.  Some of us know
better and have reasonable expectations of this very nice tool.
 
These faulty perceptions out there really need to change before something
happens more awful than just the wasted expenditure of horrendous amounts of
time & energy & dollars.  We need to DEFINE this beast for people and get
them to understand it in its proper context.  (And context &
searchability-factors are DIFFERENT in  different 'mountings' on the various
vendors, university OPACs, CD-ROM products, and cutsie-friendly 'interfaces'
to all these...).
 
We need to DO their searches (on powerful & adequate 'mountings' of the
file!) if we want to prevent authors from publishing wrong facts,
embarrassing themselves and their institutions, and embarking on half-baked
new projects.  We need to discuss, promote, and use alternatives to MEDLINE
(and we need to develop and maintain our competence at using them).  Journal
editors should send these sorts of articles (where MEDLINE retrieval IS the
topic) to medical reference librarians for review.  It took me all of 2 or 3
minutes to see the fatal flaw in yesterday's article.  It didn't matter that
there was all sorts of lengthy & scholarly-looking analysis of what was found
and what was not.  The search (and the 'mounting' of the file and thus the
searchability of the file) was completely inadequate for the particular task.
To me, that rendered useless any subsequent "analysis."
 
We need to get INVOLVED with these "...LET'S REINVENT THE WHEEL AND MAKE
ANOTHER NIFTY DATABASE..." projects.  For many of these proposals, I feel
someone ought to call NLM to see if MEDLINE could be tweaked or enhanced into
doing what they think they need, or to determine whether  NLM or NIH has or
knows-about projects already underway or contemplated to do thus 'n such.
Rarely are these ideas completely "fresh."  They frequently come from a
particular specialty - and the authors figure that nothing at all like what
they need to do has ever been done before.
 
Lastly, when we offer bibliographic instruction (workshops, minicourses,
demos) for the latest internal or cute or "free" MEDLINE or
interface-to-MEDLINE, we MUST give a balanced & careful "pitch" about this
business.  (It isn't easy and it takes time).  Don't fire up the computer &
projector and begin raving happily.  Be instead the picture of careful,
concerned scepticism.  Start with a low-tech, slow, careful soberness.  Tell
them:  "We're gunna put this business into perspective for a moment on the
board here and THEN we'll go online..."  (You'll have plenty of opportunity
in a few moments for fun & amazement & impressing the audience...)   Start
with a black-board presentation where you outline the computer & paper
versions (years-of-coverage, number-of-journals-indexed, format,
availability, etc. etc. as well as reasonings about  "WHEN TO USE..." and
"WHEN TO RUN FOR HELP..." factors) of Index Medicus, Biological Abstracts,
Science Citation Index, Excerpta Medica, assorted other files & resources in
Cancer, Nursing, Pharmacology, Hosp. Administration, Toxicology, Ethics,
Education, Psychology, GovDocs, etc. etc. as well as the "Big Federal
Medline*" and (finally) the little local version or friendly interface of the
miniature & partially disabled "MEDLINE" you're about to demo.  In this way,
workshop attendees will see how & where "MEDLINE" fits in to the total
picture; they'll also see what fraction of total-MEDLINE is really available
to them.They'll start to get an understanding of the SIZES of things.
 
*call it "MOTHER Medline, BIG-BERTHA Medline, Real Medline"     -- do
whatever is necessary (and consistent with your style & lingo)  to
DISTINGUISH it from the 'miniMEDLINE' or the disabled-because-of-friendliness
MEDLINE you're about to launch.
 
Frankly, I come right out and tell them:
 
"Don't be publishin' anything where you say 'WE SEARCHED MEDLINE AND......'
without showing it to ME first...!  I don't trust your MEDLINE and I *really*
don't trust your searchin' "
 
[This does sound a tad egocentric, so I humor-it-up just a little, and
provide examples of search-idiocy and how to "miss" things, but I'm dead
serious. I think they need to know they're just playing around and there's an
expert (who can afford to be cocky) at the end of the phone or e-mail
connection...]
 
Yes, searching is now "easy" and loads of fun and oh so helpful...
 
But the flip side is that it's very easy to get it ALL WRONG and to discover
that "MEDLINE missed things..." which to me is like saying "my SUBARU
crashed..."      Duh...
 
 
=====
 
Thomas Mead, MLS
Biomedical Libraries
Dartmouth College
Hanover, NH 03755
 
[log in to unmask]
(603) 650-1741
 
(I do indeed LIKE my new Subaru and I drive it after consulting the TREES,
etc.   Bye!)