Untangling the Web: Secrets of a Drug Information Specialist Revealed

The March 1999 issue (vol.14) of The Consultant Pharmacist¹ introduced readers to the basic art of surfing the Internet. But the Internet can create frustration when the need for information is more urgent. Of the World Wide Web (the Web) users surveyed, 75% are not surfing casually but searching for specific information. The same percentage surveyed say finding information is a primary difficulty in using the Web.²

Pinpointing exactly what one needs on the Internet can be tedious because the Web has no systematic way of indexing itself. Unlike a traditional library, countless pages are added or altered every day in less time than it takes a computer screen to flicker. And unlike the traditional handbooks, textbooks, or subscription services, information found on the Web may not be the perfect fit for certain demands. Furthermore, today’s database technology enables the creation of unique, dynamic pages that exist one moment and disappear the next. With the exponential growth of content, combined with the need for improved productivity, one’s ability to efficiently find information on the Web becomes more critical.

This in-depth article focuses on the art of surfing the Web to retrieve drug information (DI). Informal sample cases demonstrate ways the Web can aid (and hinder) gathering relevant information. Please note that brand names and Web sites mentioned in this article are used as examples and do not represent any endorsement by the author.

Pointer 1: Domainia

The Web has become the fastest-growing media for the marketing of products, including pharmaceuticals. The easiest way to draw the targeted audience is to create a Web site using the product’s name as the domain name (the familiar .com, .org, .edu nomenclature that identifies a Web site). Many drug companies will register a domain name for their product simply to protect a brand. Where a "ProductQ.com" Web promotional site can be found, the prescribing information should be available. (The Food and Drug Administration [FDA] requires labeling information for fair balance.) Exploring Web sites such as these can produce an abundance of product information and special programs designed to increase product awareness.

Case 1: Finding a monograph. “I have a doctor’s order for ‘Abonix.’ It says this product has been preauthorized for use. I have no other information about this drug and it is not in Facts and Comparisons. Could you send me some general information?”

Although no indication was provided with the order, a follow-up question revealed that the patient had a diagnosis of multiple sclerosis. Bingo! The DI specialist remembered from a Wall Street Journal advertisement that Biogen released Avonex, indicated for multiple sclerosis, in 1997. With that information, the DI specialist launched a Web browser and typed in “www.avonex.com” in the location line, and sure enough, Biogen had created a marketing site complete with prescribing information.

Epilogue: This case illustrates several important points to remember. First, the likelihood that a branded product is registered as a domain depends on how long it has been on the market and the perceived importance of the product to the consumer. Taxol, while an important cancer drug, does not have a Web site of its own. However, Claritin does, because allergic rhinitis is a “consumer-friendly” condition.

Second, even a minor misspelling during a Web search can produce a mountain of frustrations: abonix.com would have gone nowhere. If a drug, disease, or laboratory test is guessed incorrectly, results will appear unpromising. This applies to anything typed and submitted to the Web. Do not rely on the Internet to understand incorrect spellings (or physicians’ handwriting).

Third, using drug company names in an initial Web search can help significantly in this domain-naming strategy. Typically, these are easy to find (e.g., lilly.com, pfizer.com), but some are harder to guess (e.g., pnu.com belongs to Pharmacia and Upjohn). The manufacturer’s site will provide links to information on numerous products.

Pointer 2: Use a Few Search Tools to Increase Proficiency

In general, there are two types of search tools: subject directories and key word indexes. Subject directories are selected and compiled by people and then categorized into stratified links on a Web site. As a user clicks into a broad topic, multiple subtopics and cross-references will appear as further links. Ultimately, the user will arrive at a list of links that should most closely match the search topic. Key word databases use software applications (called “spiders” or “robots”) to scan (“crawl”) the Internet. Massive, searchable databases index text from entire pages. There are many of these database applications, called “search engines”, and each is unique enough to return varying results from the same key word search. Because spiders catalog words found in every Web page, search engines are much larger than subject directories.

Results of key word searches are listed as links to particular Web pages. Unlike subject directories, which tend to provide links to Web site home pages, search engine results provide links to just about every page containing the queried key word or phrase. This is a common obstacle, since this often produces multiple links to pages from the same site. Many users simply don’t have the time to click through dozens of pages of search results to locate information.

Another hindrance is that clever developers will “stack” their Web pages with hidden key words coded to increase the probability that their site or page will be among the first listed in query results, thereby cluttering the search process.

In general, choose subject directories to locate sites that may logically contain specific information. Attempting to browse a subject directory to pinpoint a piece of data wastes time, and because consultant pharmacists favor lists, subject directories also are useful for compiling lists of resources on the Web. Users should remember that while these directories are extensive, they are not encompassing. Sometimes useful Web sites are excluded from these lists and others are overlooked because of the size of the list.

Search engines are most useful when specific groups of key words can be combined to identify Web pages containing them. Because they index so many text elements of a Web page, general key word searches return numerous irrelevant pages.

Results also vary widely among search engines. Listing priority becomes critical in these instances because thousands of Web pages can be returned from a single search. Webmasters shrewdly manipulate site text and other hidden coding to increase the chances of having their site listed when certain key words are entered. Others simply pay for a higher listing.

Adding to the confusion, many subject directories and search engines are joining together as Internet technologies and corporations merge and align. Running a key word search in some directories can yield a list of possible matches from that directory followed by a list of Web pages returned from a key word database powered by a search engine.

Case 2: Product information beyond the PI. “Could you send me information on the stability and potency limits of insulin after the vial has been punctured?” A facility was being cited for keeping insulin beyond the 30-day recommended “in-use” limit. Insulin was being discarded without being reimbursed by Medicare.

For a DI specialist, an easy point of entry to begin a search in this case might be the American Diabetes Association. The direct approach—typing in diabetes.org (“org” designates a nonprofit organization)—worked. The Yahoo! subject directory search returned the diabetes.org URL under the listing “Home > Health > Diseases and Conditions > Diabetes > Organizations >.” Navigating to the “Practice Recommendations” section and then to the “Insulin Administration” document uncovered a statement suggesting that insulin loses its potency when used beyond 30 days. However, the statement was not referenced. Helpful, but not helpful enough.

Lilly.com offered not only the standard labeling, but also the phone number for the medical information department. To continue the search, more traditional media replaced the Internet. Communication documents faxed from Eli Lilly stated that according to data on file, stability and sterility are compromised after 32 days or 100 punctures of the vial. The document cited testing guidelines set by the International Conference on Harmonisation (ICH). Once the DI specialist noted the unique title, the Web was back in use.

Alta Vista, an indexing search engine, was the next step. It claims to have one of the largest databases on the Web. A command string was entered as + “International Conference on Harmonisation,” which retrieved 372 Web pages referring to the ICH. The first 40 addresses had no relevant information, so the search shifted. A second popular search engine, HotBot, returned only 96 items; however, the fifth item listed was the ICH home page. Navigating to ICH.org and searching the site uncovered a document with recommendations on stability testing of pharmaceuticals. This was retrieved and included in the DI specialist’s response.

Epilogue. This case demonstrates that there is no single tool to meet the needs of a Web user. It is important to assess the resources available before searching for the data. As illustrated, several steps were needed, and even backtracking became necessary. When search terms fail to produce relevant sources, try using alternative spellings, synonyms, and phrase arrangements. The key word(s) selected must match those chosen by writers and developers; therefore some forethought can save time. The Web often provides more clues than answers, and one must not forget to use other modes of communication. In case 2, “data on file” could not have been accessed without the telephone request.

The Yahoo! subject directory lists no fewer than 179 search engines. They each have unique characteristics and cover overlapping amounts of the Internet. So how does one choose? The most efficient approach is to become proficient at using several of the larger databases, including understanding the query-refining tools.

The option of refining searches beyond matching key words is one of the most powerful features that search engines provide. The more popular search engines enable phrase searching and use of the Boolean expressions “AND”, “OR”, and “NOT”. Other advanced features include search refinements such as date isolation, “NEAR” commands (i.e., enforcing proximity words), host isolation (e.g., to search only the fda.gov domain), text branching (e.g., magnet* = magnetism, magnetics, magnetotherapy, etc.), and proper names.

Because each search engine’s method of issuing these commands varies, a specific tutorial is beyond the scope of this article. The “Help” section of each Web site contains ample instruction with easily understood examples. In general, simply typing in two or more key words will default to one of two Boolean command “AND” or “OR”. Knowing which command an engine uses will save time and improve results.

Be aware that some sites may interpret a series of keywords separated by spaces as a phrase; commas may be needed. Others distinguish a series as discrete words and require phrases within quotation marks. Using the “+” and “–” symbols before any key word will replace the AND and NOT commands, respectively.

Several engines have simple drop-down lists that help refine searches without forcing the user to remember unique lexicon. Other search sites have abandoned esoteric command strings in favor of natural language (e.g., where can I find the International Conference on Harmonisation?) The most notable of these is Ask Jeeves. The query above proved more difficult than Jeeves could handle, although “he” does suggest several other indexing sources, including Alta Vista. Some sites index key words but base their list priority on the number of outside links that reference that Web page. In essence, this creates a “popularity search engine” whose logic is based partially on the hypothesis that if many sites are placing links to a page, it must be more relevant.

The development of search engines and directories is continuously altered and tuned to the precise needs of the Web user. Novices should begin by practicing with the two types of search tools, querying on topics of varying breadth and depth. With regular use, a preference will develop based on the success of previous searches and the user’s competency on the application. Secundum artem.

Pointer 3: Get the News from the Source

Case 3: Getting the facts on a news event. “Hey, I heard something happened with trovafloxacin. Did somebody die from taking it? Was it withdrawn? What happened?”

An initial, blind guess proved faulty (a search for ‘trovan.com’ turned up a site for a passive transponder system), but this was not a time-intensive error.

If such a story was released on the day of inquiry, it often can be found in the news site’s health section. In this case, a glance at one widely viewed news site, CNN.com/health, found no headlines on trovafloxacin. Using CNN’s internal search engine at the bottom of the page, the first step was to click the “Help” link. There, it is discovered that comma separators are required for multiple keywords.

So “trovafloxacin, Trovan” was entered, indicating a query for one word or the other. The CNN search returned one relevant story from several days before. It turns out that after 14 reported cases of liver failure, 6 resulting in death, the FDA had issued strengthened warnings on the use of the drug. Because it was an FDA warning, the agency’s site was searched for relevant documents from the Center for Drug Evaluation and Research. Three pages of matches were found, but because the FDA search engine does not include a date filter, the dates of each link had to be checked. From the results, the actual FDA Talk Paper was downloaded and distributed to interested parties. New guidelines for use were quickly developed for this drug from the information gathered.

Epilogue. Colleagues and other caregivers often direct such questions to consultant pharmacists, providing an excellent opportunity to supply the facts, explain the implications of the events, and dispel any myths that exist. Hundreds of sites compete to be the Web surfer’s source for health news and information. But closer examination reveals that many sites republish health news content from the same two or three major sources. CNN, for example, uses reports from Reuters Health, the Associated Press, WebMD.com, and its own financial Web site.

When searching for recent news items, a good approach is to get as close to the source as possible. Manufacturers, universities, and government agencies are practical first targets. General news sites sometimes publish press releases alongside independent reports; pharmacists should carefully read a report before judging expert opinions quoted within the piece. The closing paragraph of each PRNewswire press release acknowledges the inclusion of forward-looking statements. These statements must be considered.

Pointers 4–9: The Lightning Round!

Where can I get a list of reputable sites that deal with Sjögren’s syndrome?

Browse or search the subject directories. Using Yahoo.com, browse: Health > Diseases and Conditions > Sjögren’s syndrome, or simply type the word Sjögren into the Yahoo! search box.

Where can I get a list of pharmacy associations?

Use the same approach as in the first example.

When I search for gingko biloba or other herbal products, how do I avoid the commercial sites that try to sell me their products (e.g., Herbals4All.com)?

Using Alta Vista, for example: +“gingko biloba” -“shopping cart” -”order here” -herbals4all.com Or you may choose a subject directory site with a community leader such as About.com and browse: Home > Health/Fitness > Alternative Medicine > Herbs For Health >

Where can I find the compounding formula for topical morphine?

This is a tough one to find for free on the Internet. Place a “Help Me!” post to a pharmacy news group in dejanews.com. Use the same site to search for earlier posts on the topic. Alta Vista has a Usenet search, also. If you find a relevant thread (a dialog sequence), e-mail the author. Or try Alta Vista: +“compounding pharmacies,” and call one of them.

I’m only looking for evaluated health care sites.

Medmatrix.org, HealthAtoZ.com, and Stayhealthy.com are examples of sites that compile, review, and rate health care sites. Each site has its own local search engines and browsable lists.

Where are all the sites that offer online continuing education?

Try HotBot.com: +ACPE +credits +“continuing education”

I’m looking for pharmacy graphics to spruce up my in-service presentation.

How about searching the Lycos Image Gallery (www.lycos.com/picturethis/), where images can be downloaded without restrictions. Image keyword searches can be performed at both Alta Vista and HotBot.

Finally, Some caveats (“I know what you’re going to say”)

Bao-Anh Nguyen-Khoa, PharmD
Center for Health Information
Chesapeake, Virginia

References 1. Twersky O. The art of surfing. Consult Pharm 1999;14;311–3.
   Georgia Tech Web site. Georgia Tech’s 8th Graphic, Visualization and Usability Survey. Available at: http://www.gvu.gatech.edu/user_surveys/survey-1997-10. Accessed June 21, 1999.