Eighty percent of Internet users seek out health information on the web, according to a survey released by the Pew Internet & American Life Project at the beginning of February. It’s the third most common online activity after sending e-mail and using a search engine.

So which among the myriad health information sites are the most trustworthy? It can be hard to tell, especially since there is almost a complete lack of third-party evaluations online—or in any of the news media—at a time when they would seem to be more useful than ever.

That is why it was so refreshing to see The New York Times Magazine’s Virginia Heffernan cause a stir with a February 4 column headlined “A Prescription for Fear.” In it, she blasted one of the Web’s most popular for-profit health information sites, WebMD, arguing that it is “permeated with pseudomedicine and subtle misinformation.” Calling it “synonymous with Big Pharma Shilling,” she urged readers to block the site “because of the way WebMD frames health information commercially, using the meretricious voice of a pharmaceutical rep…”

By way of comparison, Heffernan plugged MayoClinic.com, run by the world-renowned nonprofit research hospital in Rochester, Minnesota. There, she promised, readers would find: “No hysteria. No drug peddling. Good medicine. Good ideas.”

In order to illustrate her point about commercial framing, Heffernan searched Google for “headache” and “Mayo Clinic,” and “headache” and “WebMD.” The latter leads to WebMD’s “Migraines and Headaches Health Center,” which features a bunch of scary looking photos of people in significant pain and prominently recommends medication. The former search leads to Mayo’s much more sedate page for a “Tension Headache,” which is photo-less and recommends non-drug treatments before pain relievers. It’s a strong example of the differences between the two sites—but many people quickly pushed back in social media and on blogs.

Adam Grossberg, the senior vice president of corporate communications at WebMD, quickly responded via Twitter. In a string of tweets, Grossberg told his followers, “Opinion or not - would have expected a call,” and that the Times should have disclosed “its ownership of an ad-supported site that competes with WebMD,” referring to About.com Health. He also wrote, “WebMD’s content undergoes a rigorous medical review to ensure it’s credible and accurate, and we clearly distinguish editorial from ads.”

At the blog eHealth, John W. Sharp applied Heffernan’s search method for “headache” to “fibromyalgia.” He found that the resulting pages at Mayo and WebMD both display a Pfizer ad touting “a treatment option” for fibromyalgia, but the former also has ads from Google while the latter does not. (Sharp pointed out that Mayo and WebMD each have policies laying out fairly strict guidelines for accepting ads, and state that ads should not be considered endorsements of their products and/or services. He also noted that Mayo and WebMD are both HON Code certified and that WebMD also has eTrust and URAC shields. These are fairly well known and accepted standards, but a 2009 paper in the journal Informatics for health & Social Care argued that the “validity and reliability of these quality tools and their applicability to different health websites varies.”)

In the actual text of the “fibromyalgia” pages, however, Mayo recommends “lifestyle and home remedies,” “alternative medicine,” and “copying and support,” whereas WebMD cuts straight to a list of medications. This led Sharp to conclude, “there may be something WebMD suggesting drug treatment more often than Mayo Clinic, but [Heffernan’s] article fails to mention that MayoClinic.com is supported by advertising and in some cases has more ad per page than WebMD.” (Actually, Heffernan does mention that Mayo takes drug ads, but only says so in a discreet sidebar.)

An informal poll of six health reporters revealed they’d all recommend the Mayo Clinic site over WebMD, partly because of the commercial veneer at WebMD, but mostly because of the world-class research center behind Mayo’s site (reflecting reporters’ general preference for primary sources of information). None said they’d go as far as Heffernan in terms of advising people to block WebMD, however. Moreover, they would all recommend a number of sites other than or in addition to Mayo, depending on a person’s particular needs. Indeed, in terms of the wide variety of sources of medical information available online, Heffernan’s article was very limited in scope, and the point of this review is not to settle the debate she started, but rather to expand upon it.

The public badly needs more frequent and detailed analyses the broad spectrum of health information websites. Sixty-six percent of people who go looking for health information online start at a search engine, and while seventy-two percent visit two or more sites during their session, only fifteen percent say they “always” check the source and date of what they’ve found, according to a 2006 survey by the Pew Internet & American Life Project. “Fully three-quarters of health seekers say they check the source and date ‘only sometimes,’ ‘hardly ever,’ or ‘never,’ which translates to about 85 million Americans gathering health advice online without consistently examining the quality indicators of the information they find.” (The survey Pew released this month focuses on who those people are, while one it released in 2009 focused on what they’re searching for.)

Articles like Heffernan’s might encourage some Americans to change their ways, but they are few and far between—on the order of about one article per year, if that. In October 2010, for instance, The New York Times, carried a nice, critical article about a new health Q&A site called Sharecare.com under the headline, “Web Site to Offer Health Advice, Some of It From Marketers,” but buried it on page four of the Business section. In January 2009, The Wall Street Journal carried a good explainer on the front of its Personal Journal section headlined, “What’s New (Or Improved) In Health Sites.” In a July 2007 piece headlined, “Need a New Knee? Click Here,” The Washington Post “test drove some the online tools provided by area health insurers to see how they performed for three common medical needs - total knee replacement, mammography, and pricing prescription drugs.”

Go back even farther and you’ll only find a handful of similar articles published in the last ten years. The most impressive and comprehensive of them appeared in the Post in May 2000 under the headline, “Behind the Screens; Who are the companies behind the Web sites competing to provide you with health information? Can you trust them with your life?” Craig Stoltz, the paper’s health editor at the time, analyzed eight commercial information sites for the piece, which weighed in at an impressive 6,233 words (it’s unavailable online, unfortunately). Stoltz, who is now the managing editor of Healthcare.gov—a website created to help Americans understand the healthcare reform legislation signed into law last year—said in an interview that it seemed like an obvious story to do at the time.

“It was just inherent skepticism because so many people were flocking online,” he said. “I felt it was important for people to have their eyes open to the different quality of information out there and what sort of conflicts to look out for. And I was the section editor, so I didn’t have to sell it very hard.”

Stoltz thought Heffernan’s analysis was “dead on,” and argued that there is a fundamental difference between for-profit sites like WebMD and nonprofits like Mayo. “Many of these sites in the category of commercial enterprises are really designed to attract eyeballs on behalf of pharmaceutical advertisers, and no matter how much they deny it, it infects what they do,” he said, adding that he can speak with some authority on the matter. After leaving the Post in 2006, he worked for eight months as the editorial director of Revolution Health.

“One of my first jobs was to license all of this information provided by places like Mayo and the Cleveland Clinic [also world-renowned non-profit research hospital],” he said. “So our content was built around the core of this very credible, dispassionate stuff, but it tended to be framed more in response to commercial concerns … it’s very hard to live in that environment, where you have advertisers who want to appear directly across from a specific piece of content.”

The landscape is rapidly changing, however. Stoltz was surprised to see the Pfizer ad on the Mayo Clinic’s fibromyalgia page, for example. Ginger Plumbo, a spokesperson for Mayo, says that the site has taken medical ads since in launched in 1995, but there have been other changes. In 2009, Mayo signed a deal with the Everyday Health Network to expand its advertising to include general consumer products like cars and cosmetics (however, a MedCity News article about the partnership reported that ad revenue is a secondary breadwinner for the site, which relies mostly on creating personalized health portals for private organizations).

There also used to be two separate Mayo sites—MayoClinic.com, which focused on providing health information, and MayoClinic.org, which focused on the clinic and its services. They merged into a single site in January, Plumbo said, and there is now a higher level of integration between general health information and information about treatments and services offered by the clinic. Other non-profits are pursuing similar strategies. The Cleveland Clinic’s main website does not accept advertising, but it unveiled a sister site in July 2009, called Cleveland Clinic Health, which does. One health reporter contacted for this review thinks these developments at Mayo and Cleveland Clinic are an indication that nonprofits are not any more “dispassionate” about pushing services than for-profits.

In light of such changes and the many subtle factors that affect a site’s credibility, Stoltz agreed it would be very useful to have some sort of regularly updated database that people could use to evaluate different sites. The closest anyone has come to building such a thing was HealthRatings.org, a creation of Consumer Reports’s WebWatch project that launched in 2005 and evaluated the top twenty health information sites. WebWatch shutdown in 2009 when its grant-derived funding ran out, however, and the ratings are no longer available online.

Thomas Goetz, the executive editor of Wired magazine, who has written extensively about personal health and information technology, said that “it’s easy to scapegoat the pharmaceutical industry” when evaluating sites, arguing that drugs often do help people. He stressed, however, that all information providers should be sensitive to the general public’s propensity for “cyberchondria.” Outlined in a survey performed by Microsoft in 2008, the term describes people’s tendency to gravitate toward the worst-case—and most unlikely—explanations for symptoms (a brain tumor, rather than caffeine withdrawal, in relation to a headache, for example).

Goetz doesn’t think there is an inherent difference between for-profit sites like WebMD and non-profits like Mayo Clinic when it comes to “poking the fear button.” Like former journalist turned PR rep and blogger Brian Reid, Goetz noted that the real fear-mongering comes from places like anti-vaccine sites, which push egregious pseudoscience. The existence of those sites does not negate the need to evaluate more serious and professional health sites, however.

Stoltz pointed out that general information sites like WebMD and Mayo Clinic are often just starting points for health searches, and that sites devoted to specific conditions, such as ACOR.org (the Association of Cancer Online Resources), can be even more helpful. And Goetz, like a number of other journalists contacted for this review, recommended government sites like the National Institutes of Health’s MedlinePlus or the Center for Disease Control’s website, which he said often go overlooked.

“As taxpayers, we’ve spent billions of dollars on the research that informs those sites,” he said, “and it’s unfortunate that not many people got to government sites right off the bat.”

According to Goetz, the takeaway message from a chapter about information technology in his 2010 book, The Decision Tree—Taking Control of Your Health in the New Era of Personalized Medicine, is: “The problem isn’t a lack of information; it’s a lack of a way to process it. The more we understand why we’re likely to go astray, the more we can stop treating our case of the sniffles like it’s the swine flu.”

Yet he doubts that a database like the one created by Consumer Reports would help readers to better filter information. “There is virtually no third-party evaluation of any topic on the Internet,” he said. “Even if someone did a diligent meta-analysis of various health websites, the odds are” that people won’t find it through a run-of-the-mill Google search. Pew’s surveys, which found that people don’t systemically research health, support that notion.

“Most people are not habitual health searchers - a symptom pops up, they search on it; a loved one gets a new diagnosis, they search on it; the First Lady holds a press conference about childhood obesity… you get the picture,” Pew’s Susannah Fox, who helped produce the surveys, wrote in an e-mail. “Online health information is a ‘break glass in emergency’ activity for most people.”

As for the news media, while Goetz thinks that more coverage drawing attention to the role of the reader in evaluating the validity and quality of health information is “absolutely needed,” it’s not something journalists tend to do.

“Reviewing websites or critiquing a class of websites has almost disappeared, as far as I can tell, from journalism, especially print journalism,” he said. “What you do see is people critiquing a class of apps—iPhone apps or Android apps—and ranking those, mainly because there’s a novelty to that right now. But I don’t see anybody reviewing websites. That was what Yahoo! Internet Life did ten years ago. Maybe that means there’s a market opportunity for somebody. It’s an interesting challenge in part because I think so many informational websites are just kind of there. It’s like doing an article critiquing the best newspapers in the country. Who does that?”

Almost nobody. The Medical Library Association offers “A User’s Guide to finding and Evaluating Health Information on the Web,” and CNN carried a similar buyer-beware article about “Tips for savvy medical Web surfing” back in 2008. WebMD’s Editorial Policy tells readers right away, “Be aware, however, that information on WebMD or any other web site should not be used as a substitute for professional healthcare. You should always consult your health professional before acting on any information seen on WebMD or any other web site.” Mayo’s Plumbo said the clinic feels the same way about its site. “We feel the best sources of credible online health information are academic medical organizations, government entities, and reputable non-profit organizations,” she said.

Perhaps such general pointers are all that’s needed. “Frankly,” said Goetz, “the ultimate guide here isn’t going to be a newspaper article; it isn’t’ going to be some kind of external resource. It’s going to be our own faculties that help us decide if something is ringing true and whether information is helping us or hurting us.”

News outlets could surely help people develop those faculties through regular coverage or the creation of an online database that provide site-specific evaluations, however. Heffernan’s analysis may not have been perfect, but it provoked a badly needed conversation about the relative merits of health information sites, and it should inspire other journalists to provoke people, in a similar fashion, more often.

Curtis Brainard is the editor of The Observatory, CJR's online critique of science and environment reporting. Follow him on Twitter @cbrainard.