Conventional cautions regarding online information: Do they predict accuracy of breast cancer websites?

Abstract

6086 Introduction: Cancer patients and oncologists are increasingly using the Internet. Online information is affecting patient decisions regarding treatment. Therefore, multiple organizations published advice on how to screen out problematic information online. In the present study, we sought to determine whether existing quality assessment tools can screen out false or inaccurate breast cancer information online. Methods: Using 15 representative search queries on five popular search engines and only the first page of results, we generated a list of 1,585 Web pages. After eliminating irrelevant, broken and duplicate links 343 unique web pages were downloaded and evaluated for website characteristics, information accuracy and 15 quality criteria: authorship identified, author credentials identified, physician credentials stated, author affiliation stated, sources clear, general disclosures, attribution/references provided, disclosure of ownership, date created displayed, date of last update displayed, date of creation or last updated displayed, editorial process stated, internal search engine present, feedback mechanism provided and copyright notice. Quality criteria were evaluated by an independent evaluator blinded to the accuracy information. Accuracy was determined independently by a pair of clinically trained evaluators. Results: Of 343 pages evaluated, 41 false/misleading statements were displayed on 18 distinct pages. The presence of an internal search engine was correlated most strongly (r = .196, r2 = .038) with health information accuracy. However, all quality criteria (alone or in combinations tested exhaustively) were linearly associated with less than 7% of the variance in number of inaccuracies per website. Conclusion: Currently available technical quality criteria do not identify false or misleading online information about breast cancer. At this time, directing patients to websites with known content appears to be the only way to ensure patients will not encounter false or misleading online information. No significant financial relationships to disclose.