By Joel Topf, MD
Wing of Zock recently published a post entitled, “No More Wikipedia! Teaching Medical Students to Access Evidence-Based Information.” It has all of the same tropes on Wikipedia that are paraded around every time an academic looks at the encyclopedia:
- Anonymity prevents the reader from vetting the author for biases.
- It is editable by anybody, including people without expertise.
- It is not evidence-based.
- It is not peer reviewed.
The author, Senthil Rajasekaran, MD, then decries how students are using the Internet all the time:
Getting into the habit of accessing those sites may continue beyond medical school into residency and clinical practice. New Scientist cites a report that finds up to half of medical doctors turn to Wikipedia for medical information. It’s important that medical educators intervene early to teach medical students how to access appropriate information.
For an author who purports to value evidence-based medicine, Rajasekaran fails to mention any support for the implication that Wikipedia is full of mistakes and misinformation. “The information can be heavily biased or just anecdotal,” he writes.
It is interesting, in a blog post about evidence-based medicine, how scant that evidence is. But Rajasekaran’s position is not entirely without merit.
In 2011, the Journal of Medical Internet Research found that while the information in Wikipedia was generally accurate, its entries were often brief and contained omissions. Similar findings resulted from two examinations of drug information in Wikipedia, where the data provided was generally accurate but incomplete. (See here and here.) The results of the latter reference are interesting. Wikipedia was pitted against the Medscape Drug Reference (MDR) and compared to authoritative texts:
Wikipedia answers were less complete than those in Medscape (p < 0.001). No factual errors were found in Wikipedia, whereas 4 answers in Medscape conflicted with the answer key; errors of omission were higher in Wikipedia (n = 48) than in MDR (n = 14). There was a marked improvement in Wikipedia over time, as current entries were superior to those 90 days prior (p = 0.024).
I love that Wikipedia was significantly improved only 90 days after the beginning of the study. Also note that there were no errors of fact in Wikipedia, just errors of omission.
In an assessment of patient-oriented cancer information, Wikipedia was as accurate as professional databases, though not as readable. The trend is clear: Wikipedia gets facts right but the story is often incomplete.
Students and residents are clever and they want to be right. The best evidence that Wikipedia is good enough for medical practice is that residents and students keep using it. If Wikipedia failed them, they would stop using it. Students’ and residents’ personal experience is that it is good enough, especially given how easily it is to surface answers. The primary problem with using Wikipedia is not misinformation, but rather the wrath of attendings when residents admit that they used Wikipedia, a problem easily avoided with a quick lie.
Wikipedia does have problems; the incompleteness discovered in trials is noted by students and residents. I ask the following questions to every cohort of residents and students I teach:
- Do you use Wikipedia?
All of them answer “yes.” If they say “no,” I tell them that I use it, and they will then come clean and say “yes.”
- Have you ever found a mistake in Wikipedia?
Almost invariably, the answer is no. Occasionally someone will say yes and in most cases this will be an omission. If they report an actual error, I ask them if they fixed the error and discuss what our responsibility is, as physicians, to fix mistakes on Wikipedia. Students rarely report factual errors.
- Have you ever found an incomplete or poorly written article on Wikipedia?
The answer here is inevitably an emphatic “yes.”
Everyone who doesn’t use Wikipedia says the problem with Wikipedia in medicine is accuracy.
Everyone who uses Wikipedia in medicine knows that the real problem is variability. Too many articles are incomplete or too cursory.
That is a correctable problem, and instead of trying to train medical students to use resources behind paywalls in order to restore the old editorial order, we should be sending our medical students out empowered to fix and complete Wikipedia. There is a reason that the Wikipedia logo is incomplete: The logo encourages people to repair and complete the encyclopedia.
For decades, residents and med students have been assigned to do five-minute presentations on a diagnosis, disease, or treatment. These presentations are a basic teaching tool: Students must learn a topic well enough to teach it. However, the work is transient and has no impact beyond the student and rounding team. What if these assignments were changed from a five-minute presentation on post-transplant lymphoma to an assignment to review, correct, and complete the Wikipedia page on post-transplant lymphoma?
How long before an army of medical students and residents fixed all of the omissions of Wikipedia and made medical information more accessible to all? UCSF is already taking up the fight. They have established an elective in which students edit the medical Wikipedia.
Free, open access medical information is the basis for much of the transformative change medicine is undergoing. This revolution is not going to stop. Leaders of medical education must embrace and extend this change or risk becoming forgotten behind editorial paywalls.
—Joel Topf, MD, is an Assistant Clinical Professor of Medicine at Oakland University William Beaumont School of Medicine, and teaches residents and fellows at St. John Hospital and Medical Center in Detroit. Follow him on Twitter @kidney_boy.