The amazing growth of the Internet has been posing many difficulties for the health student or professional who wishes to locate specific information about a subject. There are hundreds of thousands of documents containing medical information on the Internet, and the monthly growth rate is very high. For example, one of the largest on line catalogs, Yahoo!, adds between 30 to 60 new medical resources every day and this is just a small fraction of those that went "on the air" every day.
The solution is to use services and publications which permit finding the desired information in the fastest and most precise possible way. These catalogs may exist in book form (such as the well-known Internet Yellow Pages) or in CD-ROMs, magazine supplements, etc.; or, for those who have direct access to the Internet, through the miriad of interactive on-line services.
The most efficient way of searching for highly specific information on Internet, in this way, is by using one of the two basic on-line secondary indexing systems: catalogs and indexes (also called search engines).
CatalogsA catalog contains collections of links (a link is the "clickable" address of an Internet resource, i.e., its name and where it can be found), which are usually categorized or subdivided into several topics and subtopics. These links were usually submitted and added to the catalog by the same persons who have placed the site or documents in the Internet, and not all of them are accepted for inclusion.The information which the user wishes to locate may then be searched in to ways:
Some catalog services, like Yahoo!, offer in addition a combination of the two methods, allowing browsing and, within each category, a keyword search restricted to its domain. A large on-line catalog, such as Yahoo! has something like 800,000 links, with a daily growth rate of ca. 1 %.
Some
of the better known catalogs are:
:
| Yahoo! | http://www.yahoo.com |
| InfoSeek | http://www.infoseek.com |
| A2Z | http://a2z.lycos.com/index.html |
| Surf | http://www.surf.com.br |
| Cadê? | http://www.cade.com.br |
IndexesIndexes, on the other hand, are general search services, which usually indexes, word by word, each document in the Internet, particularly in the WWW (World Wide Web). They work similarly to the search mechanisms of catalogs, by using combinations of keywords provided by the user. Indexes have a good response time because they are based on spiders or search sofbots (software robots), which perform the indexing task in an automatic way. These are computer programs which travel continuously through the Internet, jumping from link to link, and recognizing automatically if a page has already been included in the index. In case they were not, the page's address is added to the index and all text words found in it are added to file of keywords. Thus, in opposition to the catalogs, indexes don't have the necessity to be warned about the appearance of a new page on the Internet. If a page has a link to it in any previously indexed public location, it will be found automatically after a time span varying from one to four weeks.
Due to this philosophy, general indexes acumulate a gigantic number of pages and indexed words. For example, HotBot, one of the most complete indexes in existence had about 54 million pages indexed on its files, by May 1997. The obvious disadvantage is that each time that a search is specified by using a rather generic combination of words, the system returns thousands, tens of thousands, or even hundreds of thousands of links, which makes virtually impossible to locate the desired piece of information, in a sort of modern "searching for a pin in the hay". For example, by entering the keyword "health" into HotBot's search mechanism produced recently 2,362,171 links. A more specific combination like "mental health" still generated 242,526 links. Finally a very specific search like "treatment of mental retardation in autism" lead to "only" 1792 links, which is nearly tolerable regarding the possibility of exploring the most likely candidates for good "hits", mainly because this particular index shows the resulting links in inverse order of relevance (a number between 1 and 99%).
Some
of the better known indexes are:
| Altavista | http://www.altavista.com |
| HotBot | http://www.hotbot.com |
| InfoSeek | http://www.infoseek.com |
| WebCrawler | http://www.webcrawl.com |
| Lycos | http://www.lycos.com |
| Excite | http://www.excite.com |
Some of these indexes also offer
catalogs, such as Excite and Infoseek.
Medical
Catalogs and IndexesThe problem of searching for specific information in medicine, biology and health sciences can be solved by using the generic catalogs and indexes mentioned above. Almost all catalogs offer these areas as part of its classification scheme, but the scope, completeness and level of depth, as well as the detail of descriptive material for each link varies a lot from service to service. One of the best health subcatalogs on an international level is Yahoo! Health, which will be reviewed later on. In Brazil, the Surf catalog has one of the best collections of Brazilian links in health sciences. Catalogs have the advantage of allowing the location of a existing information source about a subject in a more general level of coverage, but from there the user will have to explore the links which were obtained in order to find more specific topics. This can take more time, consequently the use of indexes is the best option for those who want to get straight to the job. The advantage of indexes, on the other hand, is to locate quickly a highly specific and complex combination of keywords, but it is less useful when locating more general sites or homepages.
But in the meantime, the best solution for those who wish a more complete and intelligent search is to use the indexes and catalogs which are specialized in medicine and health sciences. They have many more links, and their classification scheme is generally more specialized. There are hundreds of indexes and catalogs of this type, and in this article we attempt to review some of those which are better known or have a higher quality, so that the reader may freely explore them and include his favorite ones to his "bookmarks".
The resources for searching the medical Internet can be classified into
four types:
Below we present briefly the characteristics of some of the best known general sites in the WWW. In futures articles of Intermedic, some of the best specialized medical catalogs will be examined.
http://www.gen.emory.edu/MEDWEB/medweb.html
This is considered the best on-line medicine and health catalog in the WWW. Developed and maintained by the medical library from the University of Emory, in Atlanta, USA, under the coordination of Steve Clancy, the MedWeb contains the largest number of links from all existing catalogs (circa 15.000), categorized into 98 large groups (specialties, kinds of information and resources, etc.) and hundreds of subcategories. The listings are extensive (all listings from one category are kept in a single file), and can be seen in three manners: by alphabetical order, by subcategories or by keyword search. Besides that, there are two aditional axis of classification, by country and by region. Links are not evaluated regarding its quality and are not anotated with extensive comments, but the catalog is extremely complete, being very rich with links outside the USA, which is rare.
This MedWeb must not be confounded with another catalog, called Adam's MedWeb (http://www.mcs.com/~ablock/www/medweb.html), which is much smaller and which has all of its links contained in a single file.
This is an excellent and complete subcatalog, which is part of the general catalog Yahoo!, one of the best known, with more than one million links. It presents a detailed classification tree, with many levels. It is possible to visualize the subcatalog's complete map, through an alphabetical listing. Yahoo! is a spontaneous submission catalog, thus it is not so complete as a specialized index, like MedWeb. But it has many advantages and resources: it is well known, has an excellent search device (like the one from Altavista, which may be used automatically, if the user wishes), and can make searches in all catalog or in any of the subcatalogs. It has also a What's New section, organized by categories, which permits to follow what's new in the health area.
Self-named Yahoo of the medical area, Achoo is an exhaustive catalog, with more than 9.000 links, accessed via an internal search engine, by using keywords. The catalog is subdivided into three general areas:
Inside each area there are many subdvisions. Links are listed with a short description. There is no lqualitiy rating for the listed resources.
Besides that, it offers the following services:
http://www-sci.lib.uci.edu/~martindale/Medical.html
An excellent and extensive catalog, developed and mantained by Jim Martindale (jmartindale@vmsa.oac.uci.edu) from University of California, in Irvine, USA. It is divided in virtual "centers" and "departments", containing selected links, with elucidative comments. It is updated in a daily basis. Many of the categories are rarely found on other catalogs, like Medical Dictionaries & Glossaries, Calculators On-Line, Clinical Toxicology, Evidenced-Based Medicine, Finance, International Health, Biomathematics, Primary Care Center and Reference Desk. It is also the only site to declare the total number of resouces types embraced by the link. In April 97, VMC had approximately 35.300 multimedia learning archives, 803 multimedia tutorials, 1.875 databases, 42.800 multimedia clinical cases, 9.950 videos and 549 courses listed. Thus, it is ideal for educational support. It has no search device, neither site quality rating.
It is a catalagog categorized by
speciality area and kind of information, developed and mantained by a commercial
company. Similar to Achoo and Medical
Matrix. It is one of the largest catalogs in the health area, with
15.000 links. A site quality rating (Reviewed Sites) has great utility,
by using a 1 to 5 stars system, and a keyword search device, that can list
all sites or just those evaluated, in the HealthAtoZ database.
http://www.slackinc.com/matrix
The Medical Matrix Project has as an objective to index clinical resources in Internet, through a high standart selection (it has only 3.500 links, approximately) and quality evaluation by a editorial board (1 to 5 stars). The main criterion for links inclusion in Medical Matrix is that it is a resource possible to be used, during a medical search, for information access. The links are anoted and categorized in large groups:
The site has also keywords search,
restricted to its content. A hard copy of Medical Matrix can be purchased
from Slack Inc., Customer Service, 6900 Grove Road, Thorofare, NJ 08086,
1-800-257-8290, email: dblood@slackinc.com.
Medical Matrix is a é Internet Working Group project from the American
Medical Informatics Association (http://www.amia.org),
under Dr. Gary Malet coordination (gmalet@healthtel.com).
http://www.ghsl.nwu.edu/healthweb/index.html
One of the best non-commercial catalogs
in the health area, developed by health sciences librarians from the Committee
for Institutional Cooperation (CIC) in 1994. CIC is a consortium formed
by the 12 largest American universities. Many institutions participate
cooperatively, with the purpose of offering a interface that promotes organized
access to non-commercial Internet resources. The site has emphasis on educational
material. There is a little identification comment under each listing link.
Links are categorized according to health or medicine area. Links number
is small, when compared with exhaustive sites like MedWeb,
but authors assure that they were selected by specialists, in each area.
There is no links quality rating. Pages are not in the same server, but
there is a What's New section. A disadvantage is the lack of a keywords
search device.
CliniWeb is a clinical information catalog in WWW, and contains circa 10,000 categorized and indexed adresses, according to the National Library of Medicine thesaurus used to classify medical articles for the MEDLINE system (MeSH - Medical Subject Headings, that exists also in Portuguese, in BIREME). Each registered link, besides name and address receives a group of keywords taken form the English MeSH (anatomical and disease terms). Links access can be made by two mechanisms: MeSH disease classificatory tree browsing or through keywords search. CliniWeb was developed by a Oregon Health Sciences University team, directed by Dr. William Hersh. The same organization lodges also one of the most traditional alphabetical WWW resources listings, called WWW Virtual Library in Medicine (address: http://www.ohsu.edu/cliniweb/wwwvl/).
Karolinska
Institute MeSH Classified Resourceshttp://www.mic.ki.se/Diseases/index.html
This is another excellent clinical
specialized catalog, elaborated by project MIC-KIBIC of Karolinska Institute,
in Sweden. Links are categorized by MeSH (Medical Subject Headings) thesaurus
tree, from the National Library of Medicine, USA. Besides this classification,
there are surgery, anestesiology, diagnosis and therapy, odontology, material
and human resoruces categories, each with many classification levels. The
links are finely chosen, but have no commentes, neither a qualitity rating
system.
Hardin
Meta Directory of Internet Health Sourceshttp://www.arcade.uiowa.edu/hardin-www/md.html
This catalogi is very useful and
must be a starting point for any person making a profound research on any
medicine or health area in Internet. The site, developed by Hardin Library
of Universty of Iowa, USA, is called a metacatalog, because it is a list
of lists, that is, contains a list of catalogs and general and specialized
indexes in many health areas (36 areas altogether, with 574 links). Each
speciality has its lists classified by size (large, medium and small lists).
Links are not commented, neither have its quality evaluated. Also there
is no search device. An interesting service is YAWN (Yahoo's What's
New Sources), that lists new links in health area from Yahoo!, from MedWeb
and from Medical
Matrix.
World
Medical SearchThis is the only search device in Internet "intelligent" enough to index only the largest medical on-line catalogs and search through keywords. The complete content of these sites was copied in the Medical World Search server, and consequently all words contained in them are accessible. Besides that the MWS is also able to direct searchs to general search devices, like Yahoo and Altavista, and return results directly to the user. Another MWS advantage is its complex search capacity, using logical expressions, and using a standard medical vocabulary, the UMLS (Unified Medical Language System), developed by the National Library of Medicine (http://www.nlm.nih.gov.br), which allows restriction or expansion of a search, by avoiding synonyms. Finally, the MWS is able of using the same terms used in the Internet search to perform a bibliographical search directly in the MEDLINE system, through HealthGate service (http://www.healthgate.com). It was developed by a PRIDE Institute team, from New York Polytechnics, USA, under Dr. Humbert H. Suarez direction.
http://www.isleuth.com/medi.html
This is another service that allows multiple search by keywords in 44 selected on-line medical databases that include all American Medical Association journals and many others, like the New England Journal of Medicine, many societies sites, MEDLINE bibliographical database, and on-line medical catalogs like Yahoo!. After specifying keywords, the user can ask for a search in even 6 databases, simultaneously, without having to re-type the keywords. It is not possible to search all databases at the same time, like World Medical Search. The service was developed by a commercial company, Internet Sleuth, that has many similar services in other knowledge areas.
http://www.ha.osd.mil/main/mednewss.html
This is a specialized search device, and is extremely useful for those trying to locate information in hundreds of newsgroups (discussion forums in UseNet) with medical and heath subjects. It is simple as entering keywords, and the system will locate and return messages sent to newsgroups that satisfy the search specifications.
Compared
with the exterior, the number of links in Medicine and Health is still
small, in Brazilian servers. Even so, the number is already significative,
to the point of having the necessity of a catalog or index to locate resources.
The two best spontaneous submission catalogs, with very close Yahoo! philosophy
are mantained by commercial sites, Surf
and Cadê?. Both have
medicine and health reasonably complete catalogs, organized in subcategories.
Another rich source of medical links
to be considered is the Brazilian
Virtual Hospital, a metaphorical structure to information access
developed by the Center for Biomedical Informatics of UNICAMP.
Other Health and Medicine Catalogs
The Author
Prof.Dr.
Renato M.E. Sabbatini is the Center
for Biomedical Informatics director of the State
University of Campinas, coordinator of Medical Informatics Discipline
of the College of Medical Sciences
from the same University, and Informatics director from the Brazilian
Medical Association. He is also Chief-Editor from Informédica
and Intermedic Magazines.
Email: renato@sabbatini.com
Home Page: http://www.sabbatini.com.br/renato
Published by:
Center for Biomedical Informatics State University of Campinas © 1997 Renato M.E. Sabbatini |
Sponsored by: Searle Farmacêutica Brasil |