Medical World Search Logo
Medical World Search Banner
Medical Intelligence At Your Fingertips

Goals of Medical World Search

Medical World Search is the first search engine on the World Wide Web especially developed for the medical field. Medical World Search can aid medical practitioners, researchers, or anyone with basic knowledge of medicine, to formulate an optimally precise query to search the World Wide Web and find exactly the information they need. The major goals of Medical World Search are to provide a search engine that operates over a selection of the most high quality medical sites on the Web and to facilitate searching by using a medical thesaurus that understands medical terminology and can thus search for related terms automatically.

The World Wide Web is triggering a large growth in the amount of medical information that is potentially freely available to practicing physicians quickly. This readily available information could be important to help physicians maintain the level of their clinical skills, which tend to decline with time [1,2], and learn more quickly about advances in medical diagnosis and therapy, reducing the long delay that is often observed [3,4,5]. The result could be an improvement in clinical decision-making, which has been found lacking in a number of studies [6,7,8,9,10].

However, medical information is often difficult to find on the World Wide Web, compared to MEDLINE, the National Library of Medicine's premier bibliographic database, standing in contrast with the potentially larger usefulness of the Web, given that it presents full-text information instead of just abstracts.

In MEDLINE, abstracts have been indexed manually with MeSH terms [11], bypassing some of the complexity and richness of medical terminology. A given medical term often has a number of synonyms, as well as potentially a large number of more specific terms, and locating the most relevant document to answer a specific information need should include searching for these related terms. A variety of search systems allow to search MEDLINE records using Boolean operators, explosion to search for broader concepts by retrieving documents that contain narrower terms, and other features. For instance, Haynes et al. [12] listed 27 MEDLINE systems.

Another major problem while searching for medical information on the World Wide Web is the presence of much noise, resulting from the varying values of the published information. Search engines have usually taken one of two approaches. They may index the full text of a large number of sites and rely on the user to filter out noisy results. Altavista, Lycos, and Infoseek are examples of such systems. The major disadvantage is the large result set returned for many queries and the variable quality of these results. Another approach is to choose the major sites manually and search or browse over this limited selection, and instead of searching the full text, index manually generated descriptions of the information resources. An example is Yahoo. For both approaches, clinically useful information is usually lost in these search engines among the consumer- oriented, medically irrelevant, or variable quality information. This observation led to the creation of sites such as Medical Matrix, CliniWeb [13] and MedWeb [14], which takes a similar approach to Yahoo, but focus on medical information.

Medical World Search is a system that allows searching the full text of a large collection of Web pages for medical information and retrieve results that are ranked by relevance to the user's query. For indexing and searching, it uses a word-statistical approach [15], similar to Altavista, Lycos, and Infoseek, and to Knowledge Finder, a MEDLINE system [16]. However, in contrast to the Internet search engines, Medical World Search selectively indexes medical documents, carefully choosing sites that are of high quality and high clinical content to reduce noise in the search results.

In addition, a major feature of Medical World Search is its use of the Unified Medical Language System (UMLS) to allow including synonymous terms in searches and doing explosions. The UMLS is a large project of the U.S. National Library of Medicine (NLM) [17]. The UMLS Metathesaurus contains a representation of many different biomedical vocabularies (such as MeSH, CPT, ICD-9-CM, SNOMED-2, SNOMED International, just to name the major components) that are unified so that a UMLS concept might list the equivalent strings (synonyms) in the various vocabularies. As such, the UMLS is a large source of medical terminology information that can be used for information retrieval. The UMLS can be used by matching medical concepts in user queries and database documents so as to abstract away from the variability of medical language (normalization), similarly to how MeSH is used in the MEDLINE database.

Fifteen percent of existing web pages are related to medical and health; some are of high quality and some are junky. Our goal is to provide the world an intelligent tool and a useful service to find the quality health and medical information on the web. One common problem with web information is that their URLs (locator addresses) frequently change and not always properly maintained. We urge our users to join us to inform the information owner when broken links are discovered.


  1. Ramsey PG, Carline JD, Inui TS, et al. Changes over time in the knowledge base of practicing internists. JAMA, 1991; 266:1103-8.

  2. Leigh TM, Young PR, Haley JV. Performances of family practice diplomates on successive mandatory recertification examinations. Acad Med, 1993; 68:912-21.

  3. Stross JK, Harlan WR. The dissemination of new medical information. JAM, 1979; 241:2622-4.

  4. Stross JK, Harlan WR. Dissemination of relevant information on hypertension. JAMA, 1981; 246:360-2.

  5. Williamson JW, German PS, Weiss R, Skinner EA, Bowes F. Health science information management and continuing education of physicians

  6. Kunin CM, Tupasi T, Craig WA. Use of antibiotics - a brief exposition of the problem and some tentative solutions. Ann Intern Med, 1973; 79:555-60.

  7. Simmons HE, Stolley PD. This is medical progress ? Trends and consequences of antibiotic use in the United States. JAMA, 1974; 227:1023-8.

  8. Bernstein LR, Barriere SL, Conte JE. Utilization of antibiotics: analysis of appropriateness of use. Ann Emerg Med, 1982; 11:21-4.

  9. Weiner JP, Parente ST, Garnick DW, et al. Variation in office-based quality - a claims-based profile of care provided to Medicare patients with diabetes. JAMA, 1995; 273:1503-8.

  10. Ellerbeck EF, Jencks SF, Radford MJ, et al. Quality of care for Medicare patients with acute myocardial infarction: a four-state pilot study from the Cooperative Cardiovascular Project. JAMA, 1995; 273:1509-14.

  11. Bachrach CA, Charen T. Selection of MEDLINE contents, the development of its thesaurus, and the indexing process. Med Inf, 1978; 3(3):237- 54.

  12. Haynes RB, Walker CJ, McKibbon KA, Johnston ME, Willan AR. Performance of 27 MEDLINE systems tested by searches with clinical questions. J Am Med Informatics Assoc, 1994;1:285-95.

  13. Hersh W, Brown K, Donohoe L, Campbell E. Managing distributed information on the WWW: CliniWeb. Abstract Book of the 1996 AMIA Spring Congress, 1996: 91.

  14. Kogelnik AM, Foote S. MedWeb: biomedical Internet resources. Proceedings of the 1996 AMIA Annual Fall Symposium (Formerly SCAMC), Hanley & Belfus, 1996:969.

  15. Hersh WR. Information retrieval: a health care perspective. Springer-Verlag; 1996.

  16. Hersh WR, Hickam D. Use of a multi-application computer workstation in a clinical setting. Bull Med Libr Assoc 82(4), October 1994.

  17. Lindberg DAB, Humphreys BL, McCray AT. The Unified Medical Language System. Methods Inf Med,1993; 32: 281-91.

All documents made publicly available on this server are Copyright © 1997 Medical World Search .
In addition, for materials from the Unified Medical Language System® of the National Library of Medicine , additional copyright restrictions apply. Use of this information service is subject to the disclaimer and the terms and conditions .