The study, published today in the online Journal of Medical Internet Research, suggests the importance of teaching teens better search strategies as well as encouraging Web site designers to target teens.
"I think a lot of people don't understand how important the Internet has become in terms of health care delivery," says Caroline Richardson, M.D., a study author and physician in the Family Medicine Department at the University of Michigan Health System.
"Web-based health care content is becoming more and more sophisticated every day and physicians are beginning to rely on the Internet to deliver detailed, tailored information to patients about their particular medical problems. Often in my own clinical practice, patients come in to me only after seeking information on the Internet about their particular problem or symptom. It is critically important that we learn more about the barriers patients of all ages are facing in accessing relevant health information so that we can continue to improve access to health content on the Internet."
The study asked 12 middle school and high school students in southeastern Michigan to search the Internet for answers to six health-related questions. Video cameras recorded the computer screen and the students' voices, while tracking software installed on the computer took screen pictures twice per second. Students were asked to talk out loud while they searched so the researchers could get a better idea of the reasons behind their searching behavior.
The six questions were chosen as popular topics for adolescents but avoided any controversial issues. One question asked the students to find information on what food someone with diabetes should or should not eat. In other questions, students had to find a local Alcoholics Anonymous meeting, determine if fatigue could be caused by taking the anti-depressant drug Paxil, and locate a place that provides free and confidential HIV tests.
The questions were written to be applicable to students' lives; for example, one question reads, "You are about to get a tattoo, but a friend warned you that some places spread infections like HIV and hepatitis. Use the Internet to find out if this is true."
"Students had the most trouble with the AA question and the HIV test question. It seemed to be the local element that tripped them up," says Derek Hansen, a doctoral student at the U-M School of Information and lead author of the study. "The easiest question was on diabetes and diet, most likely because there is so much information on the Web about that topic."
Searches were deemed successful, partially successful or unsuccessful. Successful answers were complete, correct and useful. For example, an Alcoholics Anonymous meeting in another state was not useful. Partially successful answers either addressed only part of the question or were correct but not useful.
Students completed a total of 68 searches, averaging 5 minutes, 41 seconds. In 60 of those searches, students started by using a search engine. Search phrases were almost always fewer than four words, and the two most common searches – used five times each – were simply "diabetes" and "Paxil."
"The students tended to use extremely general search terms," Hansen says. "For example, when the question asked if fatigue could be caused by Paxil, many of the students just typed 'Paxil,' resulting in a large number of hits that contained no information about fatigue."
Another roadblock was misspelled words. Despite the written questions being posted next to the computer, 30 of the 132 search phrases entered contained at least one misspelled word. Some search engines offered an alternate search that corrected the spelling error, but students rarely noticed or used it.
Even when students found a Web site that contained the answer to a question, they didn't always find the answer. In the Paxil question, many of the students ended up at the drug company's Web site, a site dense with links and information. They then had difficulty finding the link that describes side effects such as fatigue.
"One of our suggestions is that site designers look at why people are using their site and how they can make it more tailored to the user's needs," Hansen says. "Sites targeted at students need to be especially aware of this and make their sites well-organized, concise and understandable. Long paragraphs, too many links and difficult vocabulary all make it more likely that teens won't find the information they need, even if it's within the site."
Of the 68 searches, seven were abandoned when the students gave up or the class period ended. Forty-seven searches were successful and 14 were either partially successful or completely incorrect. Every student answered at least one question correctly and the older teens were more likely to be successful than the younger ones.
School staff selected students for the study who were strong academically and who were comfortable using computers and searching for information on the Internet. Study authors say this suggests students less familiar with Internet searching would have even more trouble. Each student was tested individually for a maximum of one hour.
The researchers offer several suggestions that both teens and Web designers can take to help improve access to health information. For example, teens can be taught better search strategies. This may mean using directories within search engines that drill down into specific topics, or teens could learn to formulate and refine search terms used on general-purpose search engines.
For search engine and site designers, it's important to understand how teens search and to build sites with that in mind. For example, most of the students relied on only the first few results from search engines, suggesting it's important that health sites appear near the top of results for searches on health terms. Intuitive, clear site navigation is also important.
"As we move toward a health care delivery system that relies more and more on the Internet to deliver high quality health information, we must pay attention to issues of accessibility in cyberspace. There are very specific barriers to finding relevant health information on the Internet. They can be minimized, but we're just starting to get a handle on what these specific barriers are," Richardson says.
"Also, a lot can be done to tailor Web sites to particular groups. In this case we are focusing on adolescents who don't tend to see a doctor very often and who have health issues that are often sensitive and hard to talk about with anyone."
Contact:
Nicole Fawcett, nfawcett@umich.edu, or
Krista Hopson, khopson@umich.edu
734-764-2220
Funding for this study was provided under a contract from the Kaiser Family Foundation. In addition to Hansen and Richardson, study authors include Paul Resnick of the U-M School of Information, and Holly Derry of the U-M Comprehensive Cancer Center.
The paper can be viewed online at the Journal of Medical Internet Research Web site, http://www.jmir.org/2003/4/e25/index.htm.
Journal
Journal of Medical Internet Research