Scope, Completeness, and Accuracy of Drug Information in Wikipedia

Pharmacy Practice, College of Pharmacy-West Palm Beach, Nova Southeastern University, Palm Beach Gardens, FL 33410, USA.
Annals of Pharmacotherapy (Impact Factor: 2.92). 11/2008; 42(12):1814-21. DOI: 10.1345/aph.1L474
Source: PubMed

ABSTRACT With the advent of Web 2.0 technologies, user-edited online resources such as Wikipedia are increasingly tapped for information. However, there is little research on the quality of health information found in Wikipedia.
To compare the scope, completeness, and accuracy of drug information in Wikipedia with that of a free, online, traditionally edited database (Medscape Drug Reference [MDR]).
Wikipedia and MDR were assessed on 8 categories of drug information. Questions were constructed and answers were verified with authoritative resources. Wikipedia and MDR were evaluated according to scope (breadth of coverage) and completeness. Accuracy was tracked by factual errors and errors of omission. Descriptive statistics were used to summarize the components. Fisher's exact test was used to compare scope and paired Student's t-test was used to compare current results in Wikipedia with entries 90 days prior to the current access.
Wikipedia was able to answer significantly fewer drug information questions (40.0%) compared with MDR (82.5%; p < 0.001). Wikipedia performed poorly regarding information on dosing, with a score of 0% versus the MDR score of 90.0%. Answers found in Wikipedia were 76.0% complete, while MDR provided answers that were 95.5% complete; overall, 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).
Wikipedia has a more narrow scope, is less complete, and has more errors of omission than the comparator database. Wikipedia may be a useful point of engagement for consumers, but is not authoritative and should only be a supplemental source of drug information.

  • [Show abstract] [Hide abstract]
    ABSTRACT: Wikis may give clinician communities the opportunity to build knowledge relevant to their practice. The only previous study reviewing a set of health-related wikis, without specification of purpose or audience, globally showed a poor reliability. Our aim was to review medical wiki websites dedicated to clinical practices. We used Google in ten languages, PubMed, Embase, Lilacs, and Web of Science to identify websites. The review included wiki sites, accessible and operating, having a topic relevant for clinical medicine, targeting physicians or medical students. Wikis were described according to their purposes, platform, management, information framework, contributions, content, and activity. Purposes were classified as "encyclopedic" or "non-encyclopedic". The information framework quality was assessed based on the Health On the Net (HONcode) principles for collaborative websites, with additional criteria related to users' transparency and editorial policy. From a sample of five articles per wikis, we assessed the readability using the Flesch test and compared articles according to the wikis' main purpose. Annual editorial activities were estimated using the Google engine. Among 25 wikis included, 11 aimed at building an encyclopedia, five a textbook, three lessons, two oncology protocols, one a single article, and three at reporting clinical cases. Sixteen wikis were specialized with specific themes or disciplines. Fifteen wikis were using MediaWiki software as-is, three were hosted by online wiki farms, and seven were purpose-built. Except for one MediaWiki-based site, only purpose-built platforms managed detailed user disclosures. The owners were ten organizations, six individuals, four private companies, two universities, two scientific societies, and one unknown. Among 21 open communities, 10 required users' credentials to give editing rights. The median information framework quality score was 6 out of 16 (range 0-15). Beyond this score, only one wiki had standardized peer-reviews. Physicians contributed to 22 wikis, medical learners to nine, and lay persons to four. Among 116 sampled articles, those from encyclopedic wikis had more videos, pictures, and external resources, whereas others had more posology details and better readability. The median creation year was 2007 (1997-2011), the median number of content pages was 620.5 (3-98,039), the median of revisions per article was 17.7 (3.6-180.5) and 0.015 of talk pages per article (0-0.42). Five wikis were particularly active, whereas six were declining. Two wikis have been discontinued after the completion of the study. The 25 medical wikis we studied present various limitations in their format, management, and collaborative features. Professional medical wikis may be improved by using clinical cases, developing more detailed transparency and editorial policies, and involving postgraduate and continuing medical education learners.
    Journal of Medical Internet Research 17(2):e48. · 4.67 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Although potential drug-drug interactions (PDDIs) are a significant source of preventable drug-related harm, there is currently no single complete source of PDDI information. In the current study, all publically available sources of PDDI information that could be identified using a comprehensive and broad search were combined into a single dataset. The combined dataset merged fourteen different sources including 5 clinically-oriented information sources, 4 Natural Language Processing (NLP) Corpora, and 5 Bioinformatics/Pharmacovigilance information sources. As a comprehensive PDDI source, the merged dataset might benefit the pharmacovigilance text mining community by making it possible to compare the representativeness of NLP corpora for PDDI text extraction tasks, and specifying elements that can be useful for future PDDI extraction purposes.
    Journal of Biomedical Informatics 04/2015; DOI:10.1016/j.jbi.2015.04.006 · 2.48 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Background Wikipedia gains growing attention as a provider of health information. This study aimed to investigate the use, relevance and challenges of Wikipedia among medical students. Methods An online questionnaire was made accessible to students at five medical universities in Germany, Austria, and Norway. Besides demographical data, the questions covered the role of Wikipedia in the academic life of medical students. The questionnaire investigated if the students had ever found erroneous medical entries and whether they corrected these. Results A frequent use of Wikipedia in general is statistically significant correlated with a frequent use in medical studies (p < 0.001). Information retrieved from Wikipedia is predominantly critically appraised either by comparing it to profound knowledge (79%) and/or to specific literature (75%). Despite most (97%) respondents disclosed that they already had found false information in Wikipedia, recognized errors were seldomly corrected (~20%). Conclusions The information retrieved from Wikipedia is critically appraised. However, we found shortcomings in handling erroneous entries. We argue for professional responsibility among medical students in dealing with this dynamic resource. Moreover, we encourage medical schools to supplement information to Wikipedia to further benefit from the vast possibilities of this platform.
    BMC Medical Education 03/2015; 15(1). DOI:10.1186/s12909-015-0309-2 · 1.41 Impact Factor


Available from