It is common for people to search for health information online. Indeed over 60% do so per year, and only 2% of them will use sites requiring payment. Searches range from specific questions about drugs and procedures, to how to interpret test results. More than half state that the information they found influenced a medical decision, and over a third don’t follow up their internet searches by consulting a doctor. The accuracy of free online medical information is therefore pretty important for public health. Of the competing free sources online, traffic to Wikipedia’s heath content is the highest (with only the American NIH coming close). And it’s not only the general public. Unsurprisingly, Wikipedia’s medical pages are used by 95% of medical students, but also over by half of practicing clinicians.

Wikipedia is also available in nearly 300 languages, and its importance can be even more striking in countries where English is not widely spoken. For example, during the ebola outbreak of 2014, the relevant Wikipedia pages were updated and translated into over 100 languages over a matter of weeks, and was read over 100 million times.

Measurements of Wikipedia’s accuracy vary widely for different fields, and depending on what it is being compared to. However, there are several common observations: quality varies massively between articles, most errors are omissions, readability is often low, illustrations are limited.

So, what can be done to ensure medical information on Wikipedia is as accurate as possible? A pair of papers in Science and JECH this month suggest that the key is expert participation. The majority of authors of Wikipedia’s medical articles are medical professionals, with the remainder mostly consisting of medical students and researchers. Yet, the total number of editors is relatively small (only a few hundred), and there are over 30,000 medical articles to improve and keep up to date. Also, the editor community isn’t as diverse as it should be, with a strong skew towards northern hemisphere, white males.

Attracting contribution from busy professionals requires efforts on multiple fronts. Firstly, the editing interface must be as easy as possible to use. The encyclopedia has already come long way over the last couple of years and editing is now like using a Word document.

There are increasing efforts to peer review Wikipedia. One of the earliest examples was by Nature, which organised review of 42 Wikipedia pages and compiled a list of errors which were fixed in the subsequent weeks. This has now been extended to dual-publishing schemes where academic journals put articles through peer review before publishing a citable version in the Journal, and replacing the Wikipedia page. Examples include PLOS, Gene, and WikiJournal of Medicine. Galleries of images or videos by medical professionals can also be uploaded for use in articles.

Most importantly, editing the encyclopedia needs to be seen as a worthwhile use of time. In part, this is progressing as the encyclopedia’s reputation improves. However, formal recognition of work by professional bodies would allow professionals to justify their time by having something to put on their CV.

For those who want to improve the accuracy of medical information online, improving Wikipedia is probably one of the most immediately effective platforms to do so.

Thomas Shafee

Thomas Shafee is a biochemist and bioinformatician researching protein evolution and engineering at the La Trobe Institute for Molecular Science, Melbourne. He is also a Wikipedian, working mainly on improving science articles. He is also on the editorial boards of the WikiJournal of Medicine, and PLOS Genetics to facilitate better interaction between the academic and Wikipedic communities.

Ref: Shafee, Thomas; Mietchen, Daniel; Su, Andrew I. (2017-08-11). “Academics can help shape Wikipedia”. Science. 357 (6351): 557–558. doi:10.1126/science.aao0462.

Ref: Shafee, Thomas; Masukume, Gwinyai; Kipersztok, Lisa; Das, Diptanshu; Häggström, Mikael; Heilman, James (2017-10-29). “The evolution of Wikipedia’s medical content: past, present and future“. Journal of Epidemiology and Community Health. 71 (10). doi:10.1136/jech-2016-208601.

Featured image credit: Laertis333 CC BY-SA 3.0, via Wikimedia Commons