Argumentum Ad Populum: Medical Wikis 101

What exactly is a medical wiki, and for that matter, a wiki in general? Are these resources useful to you and your patients?

Our story begins, as do many great things, with jellybeans.

Suppose the existence of a large jar of jellybeans in a room. Into the room we bring a mechanical engineer and a jellybean executive. We also bring in 1,000 other people (the room is quite large). We ask all in attendance to estimate the number of jellybeans, and offer you a choice: will you bet on the engineer’s estimate, the jellybean expert’s estimate, or the mean of the guesses of the remaining 1,000?

In his 2004 book, The Wisdom of Crowds, James Surowiecki finds that the combined individual efforts of a group outperform any single individual effort over a series of trials. He describes a related experiment, in which a group of volunteers travel through a maze. According to Surowiecki, investigators “figured out what a majority of the group did at each node of the maze, and then plotted a path through the maze based on the majority’s decisions.” This “collective solution” resulted in a path through the maze shorter than the path taken by any individual participant, including those who had already completed the maze once (ie, “maze experts”).

Surowiecki goes on to argue that properly interpreting and harnessing the wisdom of crowds can produce desirable outcomes in fields ranging from economics to sociology. But what about medicine? Can reliance on the collective knowledge of a large group produce value in a discipline in which precision is paramount? Medical Wikis—practical collections of this kind of knowledge—exist and confer advantages that cannot be duplicated by any other media. But there are pitfalls, as well. Remember, Surowiecki’s optimistic title is a play on a much earlier work by Scottish journalist Charles MacKay: Extraordinary Popular Delusions, and the Madness of Crowds.

What on earth is a wiki?

You’ve almost certainly heard of Wikipedia, the largest and most famous collective effort at online information distribution. Wikipedia has become so iconic that it would be easy to assume that the terms Wikipedia and wiki are functionally equivalent, but this is not so. Which brings us to:

KEY POINT #1: Wikipedia is but one example of a wiki, and by no means the definitive one.

Broadly speaking, a wiki (from the Hawaiian word for “quick/fast”) is a group of Web pages on a particular subject that allows open modification of the content, using a simplified interface that requires only a standard Web browser. Most wikis use extensive hyperlinking (again, easily accomplished by all contributors) within each article to connect subjects and enable quick and intuitive research. The concept and the name were both products of Ward Cunningham, whose WikiWikiWeb (“fast fast Web”) went live in 1994.

Ideally, a wiki article on, say, jellybeans, would be written by a contributor with extensive specific knowledge on the delectable treats, with fact-checking by additional experts. A contributor with a n in-depth understanding of the jellybean manufacturing process might add detail on that subject, while another provides a well-researched history of the Jelly Belly Company. An expert on traditional holidays might add a section on Easter, and link to a larger article on the same subject. Contributors with language expertise would polish the grammar. The result: a clean, detailed, referenced article, freely accessible to all, which can be instantly updated with breaking jellybean news. However:

KEY POINT #2: Allowing open contribution to a Web page introduces the inevitability that some contributors will make undesirable changes, additions, or deletions.

These may range from simple vandalism (inserting profanity), spreading of misinformation (Wikipedia carried false news of the death of Ted Kennedy for some time), or even well-intentioned additions of incorrect information. The wiki approach is to make no particular effort to prevent misinformation or other damage; rather, it relies upon a community of users to quickly catch and correct damage as it occurs. In this way, says the Encyclopedia Britannica, the wiki is a rare example of “an Internet counterculture that has a basic assumption of the goodness of people.”

Wikipedia

Wikipedia is the most prominent wiki, with more than 12 million articles covering every subject imaginable. At this point, encounters with Wikipedia are essentially inevitable; if one of your patients decides to Google a condition or medication, a page from Wikipedia will almost surely be one of the top five hits.

KEY POINT #3: Wikipedia, like many wikis, is a truly open reference source.

Contributors need have no particular credentials. As a result, the reliability of the material may be questionable. Malicious users can introduce misinformation; operatives can add political spin or delete disadvantageous material. There is no accountability, because editing can be anonymous.

Thus, Wikipedia has caused a fair amount of hand-wringing. In a 2007 Encyclopedia Britannica blog post, former American Library Association president Michael Gorman wrote: “Here is a present danger that we are ‘educating’ a generation of intellectual sluggards incapable of moving beyond the Internet and of interacting with—and learning from—the myriad texts created by human minds over the millennia.” The irony of sentiments like these appearing on a blog aside, does Gorman have a point? Does the use of Wikipedia erode research standards?

In a 2007 paper, “Creating, Destroying, and Restoring Value in Wikipedia,” a group of University of Minnesota students attempted to quantify some of the problems with Wikipedia. Among their findings:

• The top 0.1% of Wikipedia contributors by volume of edits provides nearly half of the lasting value on Wikipedia.

• Only two of the top 50 editors by value added were bots (automated editing programs), suggesting that human contribution is ongoing and critical.

• About 5% of all revisions “damage” the entry in some way; the overall probability of encountering a “damaged” entry is 0.0037 per page view.

• 42% of damage incidents are repaired within one page view (median persistence of damage is <3 minutes)

• The most common types of damage are “mass delete” (entire entry deleted), “nonsense” (gibberish inserted), or “offensive” (profanity, etc). These are easily recognized by the casual reader, and thus, not likely to spread misinformation.

• About 20% of damage incidents are classified as “misinformation.” Such changes are obviously the most problematic.

Overall, then, introduction of significant misinformation is not common and generally fixed quickly. Wikipedia allows editors to click on version histories of any article (to quickly see any changes made, ever, to the document), revert to previous versions, and maintain a “Watch List,” notifying them of any changes to an article of interest.

Nonetheless, misinformation does exist. A number of potential solutions are available. Some proposals suggest that instant changes be permitted only by individuals who have attained “trusted editor” status through consistently good contributions. Other wikis require all contributors to register, or even to be a credentialed expert in the subject matter. However, such plans have met with objection from those who fear they will reduce participation, closing down a system that derives its value from openness.

The medical wiki

The issue of reliability becomes even more critical in the case of wikis dealing with health and medicine. Medical misinformation can be fatal misinformation. Wikis containing medical information come in two basic flavors: wikis for patients, and wikis for doctors.

KEY POINT #4: From a patient perspective, the wiki is already commonplace. Your Web-savvy patients will almost certainly encounter wikis when they research their own conditions and treatments.

While unbridled acceptance of any wiki-fact as reliable is not to be recommended, neither is alarmism. Patients should be cautioned that wikis, like any reference work, should be used as a first stop and not a final answer. Patients should be advised to check version histories of articles they read to see the edits that may have been made, and to confirm any information with primary sources—and with a physician—before acting on it.

More controversial is the subject of the true medical wiki—a wiki that provides high-level information for doctors and other health professionals. Many have appeared in recent years; among the most significant is AskDrWiki, founded in 2007 by Kenny Civello, MD, and colleagues from the Cleveland Clinic. The site is intended to allow users to share reviews, journal articles, and clinical experiences in a truly peer-reviewed environment.

Bertalan Mesko, a medical student and Wikipedia administrator, notes on his own blog that “[Wikipedia] mustn’t include too deep medical trains of thought, even if [they are] important… AskDrWiki could fill this gap.” For example, says Mesko, Wikipedia wouldn’t include angiography videos, as they would be of little value to the typical Wikipedia visitor; a site like AskDrWiki, on the other hand, can provide them for a more specialized clientele. Other emergent wikis cover drug monographs, specific procedures, or individual conditions.

There are, of course, concerns about the quality of the information provided by medical wikis. David Rothman, an information services specialist at the Community General Hospital Medical Library in Syracuse, NY, is a frequent writer on the subject of medical wikis. He argues that the mere fact that a wiki is a wiki tells the reader precisely nothing about its content. He recommends wikis with clear and detailed editorial processes, which describe what types of material will be accepted and clearly list responsible administrators. “A medical wiki with good editorial policies and vetted contributors may soon contain information of quality similar to an established medical journal or textbook,” he notes on his site. Moreover, a wiki has the advantage of portability, ease of use, and rapid correction of errors.

Of the qualities above, the phrase “vetted contributors” generates perhaps the most controversy. This leads to:

KEY POINT #5: A restricted access wiki is still a wiki, but it’s a very different kind of wiki.

When AskDrWiki was founded, it had an open contribution model similar to that employed by Wikipedia. As it began to grow, its founders sought advice from wiki observers like Rothman; responding to their recommendations, AskDrWiki instituted a credentialing system under which only healthcare professionals may contribute to or edit the site.

As a result, the contributor list contains only those with recognized expertise in the subject area. It also, however, disqualifies both individuals without proper credentials who might nonetheless be able to add value, and genuine experts who are unwilling to take the time to become a formal editorial board member but would be willing to make periodic edits on an ad hoc basis.

Mesko, who is a member of the 70+-member AskDrWiki editorial team, suggests that wikis benefit from expanding the contributor base. He notes on his blog that in the wiki world, “it’s not a question of credentials, but of references.” Wikis, including Wikipedia, tend to frown upon un-sourced material; opinion, speculation, or facts without reference are not allowed to stand regardless of their provenance. Proponents of a more open wiki community argue that if information is properly referenced to reliable primary sources, it doesn’t much matter who places it on the wiki. On the other hand, subject matter experts might be better equipped to determine whether and how a fact properly fits in an article.

Rothman maintains a list of medical wikis at http://davidrothman.net/list-of-medicalwikis. There, he indicates the degree of credentialing required by each, along with other information intended to allow the physician to make informed choices. He also recommends seeking out wikis with active revision histories, writing that frequent changes may “indicate that the information is being kept up-to-date or that there is an active community keeping an eye on the content to ensure accuracy… an unchanging wiki is not a healthy wiki.”

Which brings us around at last to:

KEY POINT #6: The value of a medical wiki is entirely dependent upon the efforts of individual contributors.

Perhaps the most interesting thing about the medical wiki phenomenon, compared to the many technological trends we chart in MDNG, is that it is one of the few areas in which the individual reader can make an immediate and substantive impact at little expense. If any of this sounds interesting, you could be contributing to a medical wiki on a topic of particular interest within minutes (perhaps one in the near future focusing on HIT or EMRs on www.hcplive.com/mdnglive), or editing a Wikipedia entry with useful information on patients. If an appropriate wiki doesn’t exist, you can even create one with minimal effort, using so-called “Wiki Farms” like PBwiki, Wetpaint, or Wikia. The wisdom of crowds exists only in the context of a willingness to share the wisdom of individuals.

Have you contributed knowledge to a medical wiki or similar online information resource?

Frank Ferrara is a freelance medical writer and former MDNG editor.