The concept of “cyberchondria” is not new. My first exposure to the term was through a study by Harris Interactive in 2002 which conducted a global cyberchondria study comparing health citizens online in the France, Germany, Japan and the U.S. The study defined cyberchrondriacs as “people who use the Internet for health purposes.”
Harris’s poll found that these citizens tended to find health information online to be trustworthy, of good quality, easy to understand and easy to find. In 2002, 93% of cyberchondriacs in the U.S. and France found online health information to be trustworthy.
The term cyberchrondria was used as far back as 2001, cited by the BBC
as a new phenomenon that physicians in the National Health Service were managing among Britons.
Fast-forward six years, and the first systematic study on cyberchrondria has been conducted by Microsoft Research. The results, Cyberchondria: Studies of the Escalation of Medical Concerns in Web Search
, have been published by Ryen White and Eric Horvitz (an artificial intelligence expert). These researchers discovered what the NHS physicians found, anecdotally, back in 2001: that searching for health information online can lead people to conclude they’re sick with rare illnesses.
For the study, White and Horvitz define cyberchondria as, “unfounded medical fears or interests based on the review of Web content.”
The study looked at about 250,000 peoples who have conducted at least one online health search. Among this sample, 1 in 3 made the search-leap “escalating” to a very serious illness.
For example, someone with a headache who searches the term online would encounter the phrase “brain tumor” on the first page results of the search. Horvitz pointed out that searching “headache” can result in five hits of “brain tumor” and five hits of “caffeine withdrawal.” The online searcher might conclude from that result that he would have equal risks of both causes.
Microsoft’s objective here is to develop better engines for health searches that will lesson people’s anxieties.
is out warning that using Wikipedia
may be hazardous to one’s health. Researchers at Nova Southeastern University in Florida
reviewed pages on Wikipedia for 80 prescription drugs. The team found no real sins of commission (that is, not many factual errors) but many sins of omission. For example, on the page for the HIV drug Prezista, it was not mentioned that patients taking that drug should avoid St. John’s Wort as the herb can lessen effectiveness of the drug. The Nova SEU team compared the Wikipedia drug resutls to the Medscape Drug Reference site, and found the latter to be far more useful: Wikipedia had 48 “errors of omission,” while Medscape had 14.
Health Populi’s Hot Points:
The “paging Dr. Google” phenomenon has been with us for some time. Based on the Edelman Health Engagement Barometer
, health info-entials (the 22% of citizens who are highly engaged with health information search, sharing, and judging) are now engaged in what Edelman terms “the new second opinion.” This is a phenomenon where the patient-citizen consults with her physician, then checks the doctor’s advice online; and when the citizen finds something about health online they’re interested in, they check that information with their physician.
It is human nature, at least for some people, to jump to the worst-case scenario when they hear or read about health issues. Microsoft’s study demonstrates that behavior.
In their report’s conclusion, White and Horvitz point out the many “algorithmic challenges” and biases that need to be addressed to improve the online health search process. There are miles to go in this effort, but being explicit and transparent about these challenges is an important step toward dealing with the problem.
In the meantime, health site developers need to understand many citizens’ bias toward escalation in health search. And organizations such as the Center for Information Therapy
are helping people get to the most credible and personally relevant health information. Physicians, too — as the NHS doctors found in 2001 — also have a central role to play in this dynamic, as Edelman’s health engagement data point out.