3
Improving the TOXNET Web Site

In addition to the short-term improvements to the TOXNET Web site discussed in Chapter 2, the committee focused on long-term goals for TOXNET. This chapter discusses the committee's recommendations for integrating TOXNET into the overall plan for NLM's databases on the Internet and then looks more specifically at the four major components of the TOXNET Web site: databases, search interface, educational components (i.e., help screens, online tutorials), and added enhancements to assist the user (e.g., thesaurus, links between databases). Beyond improving the Web site, the committee believes it is crucial for NLM to increase awareness of the TOXNET resources on the Internet, and this issue is discussed in Chapter 4.

INTEGRATING TOXNET INTO THE OVERALL NLM WEB SITE

To fully utilize the information contained in the TOXNET databases, it is critically important to integrate TOXNET with other NLM databases, particularly MEDLINE and TOXLINE. Since MEDLINE is a standard source of biomedical information for health professionals, any links (or other methods) connecting a bibliographic citation to the relevant information in the TOXNET databases would greatly enhance the use of the TOXNET information and would provide an added service to the MEDLINE searcher.

The incredible pace of change in technology makes it potentially imprudent to speculate about distant goals for the NLM Web site. However, the committee believes it is necessary to include the TOXNET databases in any long-term planning. One potential goal may be to provide a single entry point



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 17
Internet Access to the National Library of Medicine's Toxicology and Environmental Health Databases 3 Improving the TOXNET Web Site In addition to the short-term improvements to the TOXNET Web site discussed in Chapter 2, the committee focused on long-term goals for TOXNET. This chapter discusses the committee's recommendations for integrating TOXNET into the overall plan for NLM's databases on the Internet and then looks more specifically at the four major components of the TOXNET Web site: databases, search interface, educational components (i.e., help screens, online tutorials), and added enhancements to assist the user (e.g., thesaurus, links between databases). Beyond improving the Web site, the committee believes it is crucial for NLM to increase awareness of the TOXNET resources on the Internet, and this issue is discussed in Chapter 4. INTEGRATING TOXNET INTO THE OVERALL NLM WEB SITE To fully utilize the information contained in the TOXNET databases, it is critically important to integrate TOXNET with other NLM databases, particularly MEDLINE and TOXLINE. Since MEDLINE is a standard source of biomedical information for health professionals, any links (or other methods) connecting a bibliographic citation to the relevant information in the TOXNET databases would greatly enhance the use of the TOXNET information and would provide an added service to the MEDLINE searcher. The incredible pace of change in technology makes it potentially imprudent to speculate about distant goals for the NLM Web site. However, the committee believes it is necessary to include the TOXNET databases in any long-term planning. One potential goal may be to provide a single entry point

OCR for page 17
Internet Access to the National Library of Medicine's Toxicology and Environmental Health Databases for searching all of the NLM databases. Searchers would not be required to know the specifics of each database to determine which one would best answer their in formation queries; they would only need to become proficient in using one interface. The committee did not explore this option in depth but discussed it as a distant goal that may be reached after a number of transitional steps are completed. In the near future, the committee urges NLM to utilize both existing technologies and those in development to integrate the TOXNET databases into the broader context of other NLM databases. This can be accomplished in the short-term by adding links from PubMed and Internet Grateful Med to the TOXNET Web site. A more long-term goal would be to integrate links to TOXNET information directly into the bibliographic records of MEDLINE and TOXLINE. For example, MEDLINE abstracts on the health effects of vinyl chloride would have links to the factual information on the chemical in the TOXNET databases. Such a system of links could be similar to the extensive linkages between MEDLINE and the National Center for Biotechnology Information's ENTREZ* databases now available in PubMed. Additionally, future potential linkages could include links among the factual data in TOXNET, the bibliographic information in TOXLINE and MEDLINE, and chemical structure images, as well as the chemical identification files which include full records on the chemical synonyms (e.g., the ChemID database). DATABASES TOXNET is not a uniform set of databases because it has evolved, over time, from multiple sources and for a variety of purposes. This complexity is a challenge to the designers of the TOXNET Web site. They need to create an interface that masks differences and gives searchers both consistency and continuity in searching while allowing accessibility to the unique information in each database. There are factors inherent in the structure of the TOXNET databases that currently limit the options for the Internet interface. One example is the display option. It would be preferable for the displayed results to state exactly how many pages are in the entire record and which page the display is currently on (e.g., "page 2 out of 6"). However, this change will require alterations to the database file structure and is considered a long-term goal. *   The National Center for Biotechnology Information's integrated database system, ENTREZ, allows retrieval of molecular biology data and bibliographic citations through the linkages of databases including PubMed, DNA sequences, protein sequences, genome and chromosome mapping data, and three-dimensional protein structures.

OCR for page 17
Internet Access to the National Library of Medicine's Toxicology and Environmental Health Databases The committee is concerned in the long-term with the subject coverage of the TOXNET system. Although the committee is aware that some of the TOXNET databases are also available through commercial vendors, the committee urges the relevant federal agencies to continue the availability of their databases through TOXNET. The free access available through the TOXNET Web site will provide numerous audiences (e.g., students, community organizations) with access to this information that may not otherwise be available. If databases, such as NIOSHTIC, become closed files and do not add new information, coverage of topics such as industrial hygiene and exposure assessment need to be expanded in HSDB and potentially in TOXLINE as well. SEARCH INTERFACE Chapter 2 noted a number of specific improvements to the current search interface. However, it is difficult to have a one-size-fits-all approach to database searching, and as expressed in the following quote, different types of search interfaces are needed to assist individuals with varying levels of expertise. The Great Law of Usability: A user with appropriate knowledge or expertise in the area of application but having no experience with the system should be able to use it effectively without assistance or instruction. The Lesser Law of Usability: The system should not substantially impede or interfere with efficient and sophisticated use by experienced users (Constantine and Lockwood, 1997). NLM's staff has recognized this need and has indicated that plans for additional search interfaces are under consideration. The committee has discussed the varying audiences that could potentially use the TOXNET databases and recommends the development of three levels of search interfaces. The audiences served may vary in both the depth of knowledge of the subject and depth of expertise in database searching. The committee acknowledges that each Web search engine (e.g., PubMed, Yahoo, Alta Vista) utilizes a different combination of search features and urges NLM to examine the trade-offs of each feature. As noted earlier, an implicit criteria for selecting search features should be the aligning of the TOXNET interface with other NLM Internet search interfaces so that the searcher can easily search all NLM databases on the Internet. Quick Search The most basic interface would be a Simple Search or Quick Search option, which would consist of one input box for entering all search terms. Examples include the initial screens for Alta Vista and PubMed. This option could serve the needs of a number of different types of searchers, including new searchers

OCR for page 17
Internet Access to the National Library of Medicine's Toxicology and Environmental Health Databases who want to initially explore the TOXNET databases and students or others who have a simple query. The Quick Search option should support simple Boolean searching. The subsequent screens that display the search results should provide simplified instructions and links to the other levels of search interfaces for more in depth searches. This option would work best as a one-search that searches all of the TOXNET databases at one time. The committee envisioned the Quick Search as being directly displayed on the TOXNET homepage. The other two interfaces described below would be briefly described on the homepage with links to the search screens for each interface. Step-by-Step Search A second type of search interface would be the Step-by-Step Search, which would have detailed instructions to "walk" the searcher through each step of the search process. This search level would be geared to the novice or infrequent searcher who wants to have substantive retrieval but is not familiar with the databases. The first step of selecting the databases would clearly explain the strengths and limitations of each database, the types of information available, and the types of queries best answered by each database. At the next step of entering the search terms, the screens would explain all of the options for Boolean searching and narrowing the search strategy (such as through the use of the customized display option). Explanations would be provided for each of the different display, printing, and downloading options. Explanatory pages could be customized to each of the TOXNET databases to allow for the most precise searching possible. As the technological capabilities to handle natural language searching evolve, it will be important to add this component to the interface. Advanced Search The third level interface would be for the advanced search and would offer the complexity needed to equal the specificity and the power of the command line interface. In presentations to the committee regarding the Web site, NLM's staff indicated that future plans include an advanced search screen. This interface may be structured like the Internet Grateful Med interface, with multiple input boxes for complex searching, or like the PubMed Advanced Search option, with simple but powerful capabilities. Examples of search features to incorporate into this interface include: simple and nested Boolean searching; chemical name fragment searching; field searching;

OCR for page 17
Internet Access to the National Library of Medicine's Toxicology and Environmental Health Databases fuzzy searching to compensate for typing and spelling errors; range and ordering options for numeric and date fields; truncation, including wild card, left, and right truncation to allow for variations of a term, prefixes, suffixes, and characters within a word; phrase searching; proximity searching; and numeric operators. EDUCATIONAL COMPONENTS The use of the Internet opens up multiple opportunities for providing the educational information that searchers need. In the case of the TOXNET databases, two types of educational components are needed: those that provide further details on searching the TOXNET databases to allow searchers to improve their search skills and those that provide further information on the subject of toxicology and environmental health to allow searchers to better understand the information they are retrieving. Online Help Although there is a wealth of useful and relevant material in the current Online Help screens of the TOXNET Web site, it does not appear that they will be fully utilized in their current format. One barrier for using the help information on this Web site may be the term "Online Help." Searchers frequently did not click the Online Help button because they thought it would have information that would be useful only for trouble-shooting problems rather than for proactively assisting the searcher in planning and executing a search. The most useful format for help information would be context-sensitive help that would provide detailed assistance on the specific problem encountered. This approach may be considered a long-term goal because it will require detailed programming. A short-term approach to solving this problem would be to integrate links to the help information directly into the search screens. For example, the phrase "Display Results As" or "Look For" could become a link to the help information on the display options or Boolean operators already available. Creative use of links to the help information is vital to assist searchers in navigating through the TOXNET search process. Tutorials The Specialized Information Services Division has invested in the development of Toxicology Tutor (http://sis.nlm.nih.gov/toxtutrl). This web-based tutorial offers an introductory college-level approach to the basic principles of

OCR for page 17
Internet Access to the National Library of Medicine's Toxicology and Environmental Health Databases toxicology. The committee believes that this tutorial serves an important purpose and urges that it be maintained and updated. Additionally, the committee believes that an investment should be made in a tutorial that will introduce the TOXNET databases and explain the steps needed to search the Web site. Although one of the goals of developing multiple levels of search interfaces, as described previously, is to assist the searcher at the time of the search process, a tutorial is frequently the best approach for providing additional search assistance and serving outreach purposes (Chapter 4). The tutorial can serve as a ready-made demonstration tool for instructors, librarians, community leaders, and others who want to increase awareness of the TOXNET databases and the information that is available. Examples used in a search tutorial should feature realistic scenarios and demonstrate searching both by chemical name and by health effect. Links The committee believes that hypertext links on the TOXNET Web site to Web sites outside NLM should be used primarily to provide the searcher with access to other environmental health databases that are comparable to the TOXNET databases (e.g., the Agency for Toxic Substances and Disease Registry's HAZDAT [Hazardous Substance Release/Health Effects], the Environmental Protection Agency's Envirofacts, and the Environmental Defense Fund's Chemical Scorecard). Additionally, links to Web sites that provide educational information on environmental health issues should be considered and evaluated. Chapter 4 discusses links from other Web sites to TOXNET. ADDED ENHANCEMENTS Once a number of the short-term improvements to the TOXNET Web site are implemented, several enhancements should be considered to improve the quality and efficiency of the search process. As Internet capabilities and technologies improve, the committee hopes that the TOXNET databases will be included in all enhancements. Cross-Database Searching in TOXNET As described previously as a distant goal for all the NLM databases, it would be preferable for the user not to face the dilemma of choosing a database,

OCR for page 17
Internet Access to the National Library of Medicine's Toxicology and Environmental Health Databases but, rather, have the capability to search all of the databases at once. This same goal of cross-database searching may be achieved first across the TOXNET system. Since the TOXNET databases contain the common field of the chemical Registry Number, it has been possible in the menu-searching mode to search several databases at once. It is important for cross-database searching to incorporate features to allow the searcher to eliminate duplicate records and to select only the records that are unique (e.g., choose those records in DART that are not in MEDLINE). Cross-database searching is under consideration by the NLM staff as a future step for the TOXNET Web site, and the committee urges efforts towards this goal. Thesaurus and Chemical Dictionary Inherent difficulties in searches involving chemical names include the complexity of chemical terms and the numerous synonyms involved. The committee encourages all attempts to utilize Internet technology to minimize these complexities. NLM's ChemID database links the Registry Number to the multiple synonyms for that chemical and will be a useful feature when incorporated into the TOXNET Web site. To enhance TOXNET searching to meet the needs of health professionals and others, it will be important to include as many common term and brand name synonyms as possible. The committee urges the exploration of ways to utilize the UMLS (Unified Medical Language System) Metathesaurus for searching the TOXNET databases. This may involve expansion of the Metathesaurus to further include chemical and environmental health vocabularies. Since the TOXNET records do not contain MeSH (NLM Medical Subject Headings) terms, it is particularly important for the TOXNET searcher to be able to locate synonymous terms, identify hierarchical relationships among chemical terms, and see relationships among subject terms. Incorporation of the Metathesaurus in TOXNET, as currently available in Internet Grateful Med and PubMed, would be an eventual goal. Search Analysis A further enhancement to the TOXNET Web site would be a feature, similar to the PubMed and Internet Grateful Med Details Screens, that provides an analysis of the search conducted. Information is included on the number of hits or citations found for each query term, and details are given on how the system translated the search terms. An analysis may offer suggestions for expanding or narrowing the search results and may feature an option for weighting the various search terms and synonyms (i.e., allowing the searcher to specify that one term is more important than another) such as that available in Alta Vista's Refine option.

OCR for page 17
Internet Access to the National Library of Medicine's Toxicology and Environmental Health Databases Hypertext Links Among Records The use of hypertext links among MEDLINE, TOXLINE, TOXNET, and other records can provide the searcher with additional information and paths to pursue to locate related information without constructing a new search strategy. This use of linkages was discussed above as a way to integrate TOXNET into the overall NLM Web site. Linkages to consider include: links from the citations referenced in the TOXNET database records to the bibliographic records in MEDLINE or TOXLINE; links from MEDLINE or TOXLINE abstracts on specific chemicals to the TOXNET records on those chemicals; links among records for related chemicals; and links from MEDLINE, TOXLINE, or TOXNET records to related chemical dictionary records. The committee believes that the use of the Internet as a platform for the TOXNET databases opens up numerous possibilities for enhancing their accessibility and utility. It is hoped that resources will be made available to the Specialized Information Services Division so that these opportunities can be explored and implemented. Further, as technology evolves, it is hoped that TOXNET will be integrated into plans for the overall NLM Web site and included in the use of new Internet technologies at the Library.