Wikipedia: proceed with caution.

Despite the study's strengths, a minor obstacle in the methodology should be noted. Hasty et al 9 point out that a reporting bias by the reviewers may have caused a discrepancy in dissimilar assertions that failed to show discordance. What 1 reviewer counted as an assertion, another might not have, and thus this was purely left to the subjec-tivity of the reviewer. Different results might have been found—and a greater or lesser number of assertions or levels of concordance or discordance might have occurred—if expert clinicians were used to review medical topics associated with their specialty. As it was, all reviewers were internal medicine residents or rotating interns; although this demographic allowed for redundancy, their specialization and limited experience may have also influenced the results. Thus the validity of the reviewer assertions might be questioned. Also, the legitimacy the individual, peer-reviewed articles used in the study is in question. Although the databases used to search for peer-reviewed articles were identified (UpToDate, PubMed, and Google Scholar), the articles retrieved from these databases were not verified by the 2 independent reviewers, who were tasked only with checking assertions in the Wikipedia articles. Both PubMed and Google Scholar contain non–peer-reviewed articles, and there is a chance that an article that was used to discredit an assertion was not peer-reviewed. Hasty et al 9 acknowledge that the use of " any " peer-reviewed source was a limitation and that checking the peer-reviewed articles with the assertions would be beneficial for a future study. Some reviews of Wikipedia articles have criticized its open-editing policies, as well as the writing and presentation of the topics themselves. 1,6,10 Because some of the articles are poorly written and hard to understand, the identification of assertions might be difficult. As previously noted, the validity of reviewer assertions might be questioned. P revious studies 1-8 have evaluated the usage rates and contents of the Internet resource Wikipedia. From a medical librarian's perspective , this subject demands not only judicious research, but also continual critical review. In the current issue of The Journal of the American Os-teopathic Association, Hasty et al 9 present a ran-domized, blinded study that singles out the 10 most costly medical conditions and compares articles from Wikipedia vs articles in other peer-reviewed sources. The findings of their study demonstrate that assertions (facts) presented in Wikipedia for these medical conditions were mostly in discordance with peer-reviewed literature. These …