This perspective is written by MDs Pamela Hartzband and Jerome Groopman for the New England Journal of Medicine; a few paragraphs follow:

One woman with recently diagnosed lupus told us, "I really don't want to read what's on the Internet, but I can't help myself." Her condition is currently stable, but she finds herself focusing on the worst possible complications of the disease, such as cerebral vasculitis. Although her doctor gave her detailed information, she cannot resist going on the Web to seek out new data and patients' stories."It's hard to make out what all of this means for my case,"she said. "Half the time, I just end up scaring myself."

Other patients whose diseases have no ready cure are drawn to chat rooms and Web sites that may make unsubstantiated claims — assertions that macrobiotic diets cure aggressive lymphoma, that AIDS can be treated with hyperbaric oxygen, that milk thistle remedies chronic hepatitis, and myriad other fallacious claims. Falsehoods are easily and rapidly propagated on the Internet: once you land on a site that asserts a false rumor as truth, hyperlinks direct you to further sites that reinforce the falsehood. Material is perceived as factual merely because it is on a computer screen. We sometimes find ourselves in the uncomfortable position of trying to dissuade desperate and vulnerable patients from believing false testimonials. Doctors may be perceived as closed-minded, dismissive, or ignorant of "novel therapies" when they challenge such Internet rumors.

Patients also consult the Internet in search of self-diagnosis. Sometimes, doing so leads them to seek medical attention rapidly and to suggest what turns out to be a correct diagnosis. But the Web is perilous for anyone prone to hypochondria. The differential diagnosis for backache, for example, led one of our neighbors to fret that he had an abdominal aortic aneurysm. After a week of self-palpation, he mustered the courage to call his doctor. With a normal exam and repeated reassurance, the backache resolved.

Hospitals may inadvertently contribute to such Web-inspired worry. Many medical centers have secure Web portals that allow patients to view their laboratory, radiology, and pathology results remotely. This technology is efficient, averts the need for multiple phone calls and the mailing of information, and is welcomed by many patients. The benefits, though, must be weighed against the potential negative effects of receiving clinical data without context. Patients and families may be confused by results and worried that minor abnormalities might portend serious consequences. One man saw a report in which his neutrophil count of 78% was highlighted in red; the upper limit of normal was 70%. He took this fluctuation as an indication of possible leukemia and passed a sleepless night before he could contact his physician and learn that the increase was not significant.

In addition to information about their conditions, patients increasingly seek information about their doctors online. Yet search results for doctors vary depending on how the name is entered: "Dr." followed by a name gives a plethora of rating sites, whereas a search for the same name with "M.D." at the end yields scholarly publications as well. Such variation can skew the decision of a patient trying to choose a physician to consult. Furthermore, the information obtained in Web searches of physicians is not always accurate (e.g., one of us was listed on one site as a podiatrist rather than an endocrinologist).

Read the entire perspective at the NEJM