|
Searching the Internet |
Searching the Internet for Health-Related Information
by
Allan Barclay, Ebling Library, University of Wisconsin;
Kellie Kaneshiro, Ruth Lilly Medical Library, Indiana
University;
Heidi Marleau, Ebling Library,
University of Wisconsin
Updated: September, 2005
What are Internet search engines?
Internet search engines provide a list of Internet sites that contain a specified word or phrase. While none can claim to index the entire Internet, some index more Web pages than others - increasing the chances of locating obscure or specific information. Different search engines index a different set of Web pages, so if you can't locate information in one search engine it may be a good strategy to try another.
How do Internet search engines work?
The number of Web pages indexed by a search engine does matter for rare or unsual topics. Search engines use special software robots, called spiders, to visit Web sites and collect information. The information is then used to build a searchable word index. Many commercial search engines employ their own proprietary ranking algorithms. These ranking algorithms take various factors into account, such as giving extra weight to words appearing in the title, meta tags, number of times a word appears in the document, or the number of people who link to the page. The drawback is that most designers don’t consider this structure when they design the Web site. In some cases, a Web site owner can pay to have the site ranked higher.
For details on how search engines work look at http://computer.howstuffworks.com/search-engine.htm.
The Invisible Web or “What information might I be missing?”
No search engine can claim to index every single Web page on the Internet. Web page owners may use a robot exclusion protocol that tells spiders to leave the page alone so it doesn’t show up in the results of an Internet search engine, and unlike a human being, a spider cannot register or logon to a Web site. That failure can also prevent a site from being indexed.
A significant amount of information is only available for a fee or to members of a specific group (such as a professional association). How many times have you used a search engine to locate a perfect result or a journal article, only to find that you can’t access it for free? The next time you are about to access a journal article, check to see if the name of your organization is somewhere near the top - if so, the chances are good that your library or institution has paid for access. There is a movement to make some journal articles free usually known as Open Access Initiatives.
For more on the Invisible Web (the part of the Web that isn’t picked up by search engines) see: http://www.lib.berkeley.edu/TeachingLib/Guides/Internet/InvisibleWeb.html#Why
Structured vs. Unstructured Information
When you search the Internet using a general search engine you are usually searching unstructured information. The Web page authors may or may not have used titles on their pages, used meta-tags, used standard terminology, or used a standard classification system. In other words, everyone gets to do their own thing. How organizations use or don’t use unstructured information is a hot topic. They may have presented their questionable information solely with the intent of high search engine rankings or may have presented great content in such a disorganized fashion the search engines don't know what to make of it (automated content management systems excel at this). This usually happens when people code their pages for presentation ("looks") rather than for structure.
What are some of the advantages of using standard terminology, a standard classification system, or organizing information so it fits in databases tables (i.e. structured information)? Structured information can support querying and reporting against predetermined data types and also allows for semantically based relationships (such as subject headings, mapping to terms, or see references). Standards allow for interoperability between different information systems. Structured data also supports far more meaningful filters or limits than are typically available.
Literal Searches, Concept-based Searching and Natural Language Queries
A search engine that looks for words or phrases exactly as entered performs literal searches. This can work well, but problems occur when a word or phrase has multiple meanings, for example: cell, stem cell, cell phone, jail cell, cell in an Excel spreadsheet, etc. Human beings can still make associations between words and phrases that computers cannot.
Concept-based searching involves statistical analysis in order to group like pages. This approach can be computer intensive.
A natural language query system that can handle complex queries, for example, "Is oxybutynin effective in treating urge incontinence?", has been likened to the holy grail of searching for information. AskJeeves.com is a popular natural-language query system that works with simpler queries.
Basic Search Strategies
Typically, searching either yields too many results or too few. If there are too many results, try adding terms or refine the search using the search engine's advanced search interface. If there are too few results, try dropping some of the search terms, or using synonyms. You could also try using broader terms. Another strategy would be to locate an organization or association that might cover your topic. Because each search engine has its own set of protocols for how to enter terms, check the search engine's help section.
Health-Related Search Engines
General search engines such as Google cover many different topics. Specialty search engines are able to focus on different subject areas, such as health and medicine. An advantage to using a specialized search engine is retrieving fewer, but hopefully more relevant results.
Search Engine Description Advanced Search Features? Help
Page?Health On the Net (HON) Foundation’s meta-search engine integrating five information types: MeSH (Medical Subject Headings) terms, authoritative scientific articles, healthcare news as well as other Web sites and multimedia. Search by MeSH terms and their descriptions or by one of four categories: diseases, anatomy, viruses & drugs, psychiatry & psychology. Available in 5 languages.
No (but pretty advanced as is)
No
Offers two interfaces, one for consumers and one for health professionals. Both interfaces require limiting to "consumer information" or "information for patients" in order to retrieve lay information. Limits by type of media or page content (e.g., practice guideline) are available in either interface although although limiting to peer-reviewed or CME sites is only available in the health professional search.
None apparent
No
MedHunt includes two types of documents: 1) annotated Web sites that have been visited and described (but not evaluated) by the Health On the Net (HON) team and 2) auto-indexed sites that have been automatically retrieved by HON's Web spider MARVIN. Some limits available to narrow your search.
No
No
OMNI (Organising Medical Networked Information) is a gateway to evaluated, quality Internet resources in health and medicine produced in the United Kingdom. For audience and scope, see http://biome.ac.uk/guidelines/collection. Browsing by MeSH (Medical Subject Headings) terms or NLM subject headings is also available.
Health-Related Internet Directories
Another way to locate information on the Internet is using an Internet Directory. The main difference between a directory and a search engine is that listings in directories are chosen by humans, whereas search engine listings are created automatically using software. Directories are also arranged by topic. HealthWeb is such a directory, with items chosen by professional librarians. If you don't find what you're looking for here try these:
Directory Description Advanced Search Features? Help
Page?A catalog of health and biomedical Web sites aimed at the academic and research community at Emory University and maintained by the staff of the Robert W. Woodruff Health Sciences Center Library, Emory University. Some sites are of interest to the general public.
A directory of directories. Broadly categorized by topic, and within topic by list size: 1- 20 links = small, 21 - 99 links = medium, 100+ links = large. General medical and disease pictures are highlighted as available.
n/a
General Search Engines
If you can't find what you're looking for at a health-related search engine or directory try a general search engine.
Search
EngineDescription/ Review Advanced
Search?Help
Page ?
Additional General Search Engine resources:
Gregory Notess' Search Engine Showdown has information on general search engines: http://www.searchengineshowdown.com/
Gregory Notess' Search Engine Features Chart has a list of different search engine features: http://www.searchengineshowdown.com/features/
Search Engine Watch is a good source for another perspective on Internet search engines: http://www.searchenginewatch.com/
Questions or comments about this page? Please send email to the Search Page Editors.
