INTERNET GRATEFUL MED:
MAKING A GOOD THING BETTER

by Lawrence Kingsland, Ph.D., NLM, and Dale Graham, Ph.D., DCRT


Access to online information is extraordinarily useful to biomedical researchers - and thanks to a new development, it's now even simpler to access one of the world's largest and most-used scientific databases, the National Library of Medicine's Medline.

Grateful Med, the software currently used by more than 90,000 subscribers to reach Medline and other National Library of Medicine (NLM) databases, is gaining a sidekick on the World Wide Web, appropriately called Internet Grateful Med. Set for production release in April, Internet Grateful Med is already netting positive reviews from NIH's WorldWideWeb Interest Group [see box below] and from intramural researchers who were given early access to the final testing, or "beta," version.


Original vs. Internet

Internet Grateful Med is both easier to use and more powerful than the original Grateful Med program. Unlike its forerunner, the initial version of Internet Grateful Med conducts searches only in Medline. However, NLM is already testing a follow-on version that searches several additional databases. Internet Grateful Med has better capabilities for picking and choosing records for output. You can print, save, or e-mail results using normal Web browser functions. Context-sensitive online help is available throughout the program. The original Grateful Med came in PC (DOS) and Macintosh versions. In contrast, Internet Grateful Med can be used from any computer with Internet access and a compatible Web browser, including Unix workstations. You can even access it using the Lynx character-mode browser from a dumb terminal!

In its current form, Internet Grateful Med helps the user create, submit, and refine a search in Medline. The user can search by subject, by text word in the title, or by author name. Searches can be limited by language, publication type, study group, age group, or range of years back to 1966. Internet Grateful Med offers direct links to the full text of Clinical Practice Guidelines supported by the Agency for Health Care Policy and Research and to nearly 60,000 online images from NLM's History of Medicine Division. Searching in additional databases will be available soon.

As does the original software, Internet Grateful Med includes the capability, called Loansome Doc, for requesting a hard copy of documents through the Interlibrary Loan process. However, unlike the original Grateful Med, it is not yet set up to download results in the tagged Medline format that some researchers use to import citations into bibliography programs such as Endnote or Reference Manager.

Another key difference is that the original Grateful Med allows the user to save search strategies for later reuse, whereas Internet Grateful Med does not. As a Web-based application, Internet Grateful Med cannot read from the users' local computer disk. Small applications, or "applets," written in the Java programming language may provide a solution to this dilemma in the future.

Getting Started

Ready and raring to go? Here's a checklist of what you need to operate Internet Grateful Med.

  1. A Medical Literature Analysis and Retrieval System (MEDLARS) account from the National Library of Medicine (NLM). This account will give you a user identification and password for searching Medline and other NLM files. If you have an NIH Library Card or are qualified to get one, you are also qualified for a MEDLARS account. See the Main Desk at the NIH Library in Building 10 for an application form for the Library Card, or bring your Library Card to the desk.

  2. A Web browser program. Of the graphical browsers suggested for use at NIH, the developers of Internet Grateful Med strongly recommend Netscape Navigator version 2.0 or higher for Macintosh computers or PCs running Microsoft Windows and NCSA X Mosaic version 2.6 or higher for UNIX workstations. One important note: the Windows and Macintosh versions of NCSA Mosaic do not work properly with Internet Grateful Med. If you have an account with NIH's Helix mainframe computer, you can use the Lynx command-line browser program by typing in "lynx" when you see the prompt "helix%".

  3. An Internet connection. This can be either a direct connection via NIHnet or a remote connection via modem through a local Internet service provider (ISP) or through DCRT's Parachute system.

  4. The correct Web address, or Uniform Resource Locator (URL), to access Internet Grateful Med: http://igm.nlm.nih.gov/


How It Works

Internet Grateful Med has two major sets of assisted Medline search functions: "just do it" and "user invoked." "Just do it" functions are performed automatically in the background. "User invoked" functions involve situations in which the user is asked to clarify his or her search or to choose among suggested options. One of the "user invoked" functions offers the opportunity to restrict retrieval to articles in which a given term is a central concept of the article, and it offers guidance in adding subheading qualifiers to help focus a search.

WIGging Out

The WorldWideWeb Interest Group (WIG), just one of the dozens of interinstitute interest groups at NIH, is open to anyone interested in Internet issues. In addition to offering talks and discussions on topics of general interest, the group also features presentations on specific topics of interest to scientists, information providers, and technical users. Meetings are held on the second Tuesday of each month at 2:30 p.m. in Building 10, Lipsett Auditorium. For more information, see WIG's home page on the Web, located at http://mantis.dcrt.nih.gov/WIG/

Sample screen shot
Internet Grateful Med also provides a sophisticated "analyze search" function, which offers users the opportunity to substitute terms or add related terms to augment a search. This function also helps to clarify ambiguous terms such as "management," which to one user might mean "organization and administration" and to another, "therapy." Judicious use of the "analyze search" function can sometimes dramatically improve a disappointing initial search.

Although the graphical nature of most Web browsers makes Internet Grateful Med a snap for researchers to use [see "Getting Started"], another feature of the World Wide Web posed major headaches for Internet Grateful Med's developers at NLM. The standard Web interaction - in which a user sends a request to a remote computer server, gets back a response, and the connection is broken - leaves the computer server with no history of the user's prior requests. NLM's solution for the "stateless" nature of normal Web interactions was to develop an Expert State Engine at the heart of the Internet Grateful Med gateway. This program has two parts: a "listener" that talks to a user's computer and an "expert state maintainer" that remembers what users have done and has rules for mapping terms and creating and refining searches.

For more information on this latest member of the Grateful Med family of programs or other aspects of accessing online databases via the Web, contact NLM's Internet Grateful Med development team (e-mail: access@nlm.nih.gov).


The Numbers Say It All

More than 125,000 individuals and institutions currently have accounts for searching the National Library of Medicine's 40 online databases. These users made more than 7.5 million searches in 1995. Users of the original Grateful Med do 90% of their searching in Medline, which contains more than 8 million citations. More than 30,000 new citations are being added each month.

Metathesaurus Muscle

The National Library of Medicine's Unified Medical Language System® (UMLS®) Metathesaurus® Unified Medical Language System (UMLS) Metathesaurus is an electronic Rosetta stone containing 589,000 names for 253,000 concepts in 30 biomedical vocabularies, thesauri, or classifications. Users of Internet Grateful Med draw upon this deep and carefully organized reservoir of medical terms when they employ the "find related" function, which compares the user's search terms with all terms in the Metathesaurus to produce a ranked concept "hit list," concept definitions, and Medical Subject Heading (MeSH) notes. Many of the concepts will be underlined, indicating that they are hyper links that the user can click on to ask Internet Grateful Med to create a graphic tree display of related MeSH terms. In another striking feature made possible by the Metathesaurus, Internet Grateful Med offers direct access to millions of pairs of co-terms, which are concepts that appear as "major topic" index terms for the same citation in Medline. With the single click of a mouse, users can include a concept and co-term - or even a triad of a concept, qualifier, and co-term - to improve their searches.



Return to the Table of Contents