Previous PageTable Of ContentsNext Page

Analyzing web sites or the hidden risks of medical web sites

Kathleen Murray

University of Alaska Anchorage, Consortium Library, Health Sciences Information Service
3211 Providence Dr. Anchorage, AK 99508 ph 907.786.1611, fax 907.786.1608
afktm@uaa.alaska.edu

Good afternoon,

Having come from near the North Pole, I thought you might enjoy this image of folks who’ve managed to climb to the top of Mt. McKinley.

“According to a May 2000 report released by the Pew Internet & American Life Project, more than 9 million women have gone online for the first time since January 1, 2000. The number of women who use the Web is now equal to that of men. Looking for health information is the number one activity of women on the Internet.”1

Differentiating between fact and opinion, examining assumptions, being aware of fallacious arguments or manipulative reasoning, and looking for reputable sources are skills many web users have not developed. This skill is especially crucial for health information – both when you are looking for answers and when your patients are surfing the web.

While I don’t agree with the Time article that feels there is an “epidemic of Internet snake oil”2, caution is warranted. Warnings about the lack of reliable information started as early as 1998 when an article published in Pediatrics3 found only 20% of sites they visited listed the current AAP guidelines for childhood diarrhea. Major medical schools and medical institutions, as well as hospitals were included in this study.

In March of this year, two new studies from the American Academy of Orthopedic Surgeons showed 20% to 40% of orthopedic sites they visited contained misleading or grossly wrong information (http://www.aaos.org/wordhtml/press/2001pres/nr1-01.htm).

When reputable sites have outdated information, it is not surprising that the Federal Trade Commission found 1,200 sites last year in just a few hours that touted questionable cures for heart disease and AIDS. All sites were notified that they were in violation of truth-in-advertising laws; fewer than 30% removed the offending cliams.2

While some information is knowingly false, and other sites are just behind in updating their pages, another major problem with web information is that it is consumed out of context – even accurate data needs to be placed within the proper framework to be interpreted correctly. To illustrate this, let me share a true story that happened to a friend of mine. She is a medical librarian with a bright eleven year old son who happened to be feeling quite unwell. One day he called her at work and asked her to make an appointment with his doctor, as he had just looked at the Mayo Clinic site. According to the information he found on their site, he either had a brain tumor, tuberculosis, or a sinus infection. He, of course, figured it was a sinus infection (and it was), but many people would immediately jump to the worst conclusion.

Another challenge is that like the aurora borealis at home or the aurora australis in the southern hemisphere, Internet information has an ephermal nature. Good sites disappear for many reasons. The creator of the pages removes them, the site is updated or revised, the host computer is shut down or has a name change. Often only parts of a site move; shortening the URL on broken links is one trick to finding the sought for page.

Then there are actual hoax sites such as the Clones R Us site. You might enjoy visiting this one at http://www.d-b.net/dti/

Let’s move on ... I doubt there is any disagreement about the need to evaluate web sites before trusting the information they contain. How should sites be evaluated? I have three suggestions.

First, I’ll share my checklist of five criteria. Then I’ll spend a couple of minutes discussing how to analyze web addresses, and finally I’ll give you a short list of reviewed sites to consider.

When the use of the Internet is for medical information, a critical thinking checklist should always be used. The first criteria is Authority. Is the author clearly identified? Are the author's credentials listed? Is there an institutional affiliation? What authority does the author have to write about the subject? Can you easily contact the author for clarification of information?

The second criteria is accuracy. Is the information accurate? Does the site offer a selected list of resources? Is the source of information clearly stated? Is there an explanation of the research method used to gather data?

Third is Objectivity or the Purpose of the site. Is the purpose of the site clearly stated? Is the information impartially presented? Is sponsorship acknowledged? For example, is there drug company involvement.

Fourth, is the site current? Can you tell if the site was updated recently? Are the links kept up-to-date?

Last but very critical in the medical field, does the site have a clear Privacy Policy? How does the site use personal information collected from visitors? Preferably, they do nothing with it -- they don't keep it, don't track it and don't give it to anyone.

After reviewing the checklist, try analyzing the Web Address. The major top level domain names are .gov for government sites (often at the federal or national level), .edu for educational sites (until recently, usually university or college level sites), .com for commercial or for-profit sites, .org for organizational sites, often not-for-profit sites, .mil for military sites, and .net for networks (seen with Internet Service Provider sites). Please note that just because a site is commerical does not mean the information should be ignored … .org sites can contain very biased information, too.

Another hint is looking for country codes. These are two letter designations, such as au for Australia, ca (Canada), ch (Switzerland), or jp for Japan.

Reading the URLs is a first step in deciding to visit a site. www.lib.uaa.alaska.edu/hsis/ is the address for my home page. Web addresses are read in sections with each section or level indicated by a forward slash. Within the first level, I read addresses from right to left. So without knowing anything about this site, I could deduce that it is:

  • an educational site
  • located in Alaska
  • and that it is likely to be a university library

The fact that a second level exists tells me that it is a unit or department of the first level.

The second example, (www.who.dk) is the listing for the Regional Offices of the World Health Organization … located in Denmark. Most sites located around the wold list a country code … the United States is the major exception to this practice.

For our last example I chose the AMA – first of Australia – the Australian Medical Association. (www.ama.com.au) The second AMA (www.ama.org )is for the American Management Association – a business organization. The American version of the AMA (www.ama-assn.org/). Note that in Australia, this group is listed as a .com (commercial group), while in the US, they use the domain name of .org (for organization).

Review Sites

Other recommended sites:

Quackwatch (www.quackwatch.com ) is a bit tough on alternative and comlementary medicine sites. With that caveat, it is worth adding to your bookmarks or favorites list.

Rather than searching using a general search engine, try some medicine-specific tools: Achoo (www.achoo.com) specializes in health sites, while HealthAtoZ (www.healthatoz.com) is a directory or subject hierarchy style search engine. Both have been around for some time.

WebMedLit (www.medlit.com) searches the latest articles from medical journals. Even broader is PubMed – another National Library of Medicine product that gives free access to searching the MEDLINE database, along with an ever growing list of additional NLM databases (such as AIDSline).

The National Electronic Library for Health (pilot site of the NHS) is at www.nelh.nhs.uk/

I’d like to spend a few minutes talking about what your patients are doing … given that more and more of them are using the Internet to find answers to health questions. When your patients are using the Internet seeking answers to health questions, the motto should be “Caveat Lector” – or let the reader beware. What risks do they encounter?

1. Inaccurate or inappropriate health information – which can lead to noncompliance and poor outcomes and also damage patient’s trust in their healthcare providers

2. Deceptive claims for medical products, such as the Vitamin O ad (http://www.aabhealth.com/vospecial2.htm). To follow this further, there is the FTC site with information on the claim settlement. (http://www.ftc.gov/opa/2000/05/rosecreek2.htm)

3. Emerging research shows that people put more credibility in information from computers than from television and other media – the garbage in, gospel out phenomenon

4. Often the first source of information that is received, even if proven to be incorrect, is the one people will believe to be true

5. Most web sites readability level is at the 10th grade level – a level significantly higher than the reading level of much of the US’s patient population.

6. Finally, “Many sites easily accessed by consumers appear to be profit based companies advertising an alternative product claimed to be effective for many conditions. These findings emphasize the need for critical evaluation of Web site contents.”4

The sites your patients are visiting are likely going to be different than the ones you frequent. Nancy Press at the National Network of Libraries of Medicine – Seattle branch, has developed some guidelines that may help your patients.

1. Health news or information is best for you if you are just like the people being talked about in reports of research studies.

2. If the news or information doesn't say who was involved in the study--doesn't say what gender, age, disease, etc., then you need to find out a lot more before believing that it applies to you!

3. News or information about a good therapy should be based on a comparison between that good therapy and other therapies.

4. There is usually new research in health care to help you and your health provider make better decisions about your health. Because of that, don't depend on information that's more than about 3 years old. If you want to use a pamphlet that's older than 3 years, try to find some more recent information to go along with it.

5. A good health study should have involved lots of patients, especially for a common disease.

6. Any article that shows 100% success or 100% failure or 100% of anything, is not an honest article! When dealing with human bodies and minds, nothing is 100%.

7. Read the whole thing.

8. An article about one person and his or her health problem, success, or failure can't be used as good evidence for anyone else's situation.

9. It's good to read opinions on different sides before making a decision. Any one article can often be really far out on one side.

10. The perfect article for you and your situation might not exist. We often have to just work with the best evidence available and considerable ambiguity-as health care professionals do.

11. In complementary and alternative medicine (CAM) some different criteria are necessary. The main principle is that of the accumulation of total evidence. There are difficulties with study design in many traditional medical fields. For example, how does one have a control group in the study of acupuncture?. The mind-body methods, prayer, and distance healing all depend on the orientation and opinion of the patient toward the treatment. The principle of continuity of cumulated evidence can be used. Has the teaching and the understanding stayed the same for hundreds of years?5

12. Obtain information from a minimum of three sources – accurate information will be repeated (this one won’t be on their web page).

Another good starting point is the Agency for Healthcare Research and Quality's Now You Have a Diagnosis: What’s Next? www.ahrq.gov/consumer/diaginfo.htm

I highly recommend the following three sites for consumer information. All three are government initiatives which means no advertising! The Commonwealth Government of Australia’s initiative to “improve the health of Australians by providing easy access to quality information about human health” can be found at www.healthinsite.gov.au. Britain’s nurse-led helpline, available 24 7, for confidential healthcare advise is at www.nhsdirect.uk.main.jhtml . In the US, MedlinePlus is the premier site. It is managed by the National Library of Medicine in conjunction with various public libraries and medical librarians. All three of these sites are as reliable as anything on the Internet can be.

We’ve just skimmed the surface … there is the issue of the “hidden web” – all the information stored in PDF (a.k.a. Adobe files) and databases scattered about the world, and hidden behind firewalls. Or which search engines to use for specific questions. My current favorite is Google.

I hope I’ve given you some useful tips to use for your personal quests on the Internet and tips you can share with your patients.

Thank you.

References

1. MayoClinic.com. “Finding a Balance: Health information on the Net — Women's search for answers.” November 17, 2000. http://www.mayohealth.org/home?id=WO00006

2. Time 2000 Apr 3; 155(13)70-1

3. Pediatrics 1998 101(6):E2 http://www.pediatricds.org/cgi/content/full/101/6/e2

4. Suarez-Almazor ME, Kendall CJ, Dorgan M. Surfing the Net--information on the World Wide Web for persons with arthritis: patient empowerment or patient deceit? J Rheumatol 2001 Jan;28(1):185-91

5. http://www.nnlm.nlm.nih.gov/pnr/hip/criteria.html

About the author

Kathy Murray moved to Alaska in November 1996 to manage the Health Sciences Information Service at the University of Alaska Anchorage, Consortium Library. Prior to going to Alaska, she spent 20+ years as a hospital librarian after obtaining Masters in Library Science in 1976. She has taught healthcare providers Internet and/or medical database searching skills since 1986. This current talk is based on training supported by a grant from the National Library of Medicine awarded to the Health Sciences Information Service in October of 1998 and provided to healthcare providers throughout Alaska.

Previous PageTop Of PageNext Page