News Release

Study finds cancer information on Wikipedia is accurate, but not very readable

Peer-Reviewed Publication

Thomas Jefferson University

(PHILADELPHIA) It is a commonly held that information on Wikipedia should not be trusted, since it is written and edited by non-experts without professional oversight. But researchers from the Kimmel Cancer Center at Jefferson have found differently, according to data being presented at the 2010 ASCO Annual Meeting in Chicago. (Abstract #6058)

Reassuringly, they found that cancer information found on a wiki was actually similar in accuracy and depth to the information on a peer-reviewed, patient-oriented cancer web site. There is one caveat, however: they found that the information on the peer-reviewed site was written in plainer English.

Researchers lead by Yaacov Lawrence M.D., assistant professor of Radiation Oncology at Jefferson Medical College of Thomas Jefferson University, compared the cancer information found on Wikipedia with the information found on the patient-oriented section of the National Cancer Institute's Physician Data Query (PDQ), a comprehensive peer-reviewed cancer database.

"There are a vast number of web sites where patients can obtain cancer information," Dr. Lawrence said. "The purpose of this study was to answer one question: Is the cancer information on Wikipedia correct? Reassuringly, we found that errors were extremely rare on Wikipedia. But the way information was presented on PDQ is more patient-friendly."

Dr. Lawrence and his colleague Malolan Rajagopalan, a medical student from the University of Pittsburgh, started by choosing ten cancer types and selecting key factual statements for each cancer from standard oncology textbooks. The material covered epidemiology, etiology, symptoms, diagnosis, treatment and controversial topics in cancer care.

Medical student volunteers examined the PDQ and Wikipedia articles against the prepared statements. The web pages were printed out to ensure that each individual looked at the same version of the articles. Standard algorithms were used to calculate readability based upon word and sentence length.

For both web sites, inaccuracies were extremely rare: less than two percent of the information on either site was discordant with that presented in the textbooks. There was no difference between the sites in depth of coverage. Both sites poorly discussed controversial aspects of cancer care. But the PDQ site was notably more readable: whereas PDQ was written at a level suitable for a 9th grader, Wikipedia was written at a level suitable for a college student. This difference was highly statistically significant.

"PDQ's readability is doubtless due to the site's professional editing, whereas Wikipedia's lack of readability may reflect its varied origins and haphazard editing," Dr. Lawrence said. "Overall our results are reassuring: on the one hand Wikipedia appears to be extremely accurate, on the other, the resources invested in the creation and upkeep of the PDQ are clearly justified."

The next step is to repeat the study with cancer patients to truly determine how this difference in readability impacts upon patients' understanding and retention of information, Dr. Lawrence said.

###


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.