The purpose of this study was to utilize sentiment analysis to describe online opinions toward vaginal mesh. We hypothesized that sentiment in legal Web sites would be more negative than that in medical and reference Web sites.
We generated a list of relevant key words related to vaginal mesh and searched Web sites using the Google search engine. Each unique uniform resource locator (URL) was sorted into 1 of 6 categories: “medical”, “legal”, “news/media”, “patient generated”, “reference”, or “unrelated”. Sentiment of relevant Web sites, the primary outcome, was scored on a scale of −1 to +1, and mean sentiment was compared across all categories using 1-way analysis of variance. Tukey test evaluated differences between category pairs.
Google searches of 464 unique key words resulted in 11,405 URLs. Sentiment analysis was performed on 8029 relevant URLs (3472 legal, 1625 “medical”, 1774 “reference”, 666 “news media”, 492 “patient generated”). The mean sentiment for all relevant Web sites was +0.01 ± 0.16; analysis of variance revealed significant differences between categories (P < 0.001). Web sites categorized as “legal” and “news/media” had a slightly negative mean sentiment, whereas those categorized as “medical,” “reference,” and “patient generated” had slightly positive mean sentiments. Tukey test showed differences between all category pairs except the “medical” versus “reference” in comparison with the largest mean difference (−0.13) seen in the “legal” versus “reference” comparison.
Web sites related to vaginal mesh have an overall mean neutral sentiment, and Web sites categorized as “medical,” “reference,” and “patient generated” have significantly higher sentiment scores than related Web sites in “legal” and “news/media” categories.
From the *Division of Female Pelvic Medicine and Reconstructive Surgery, Department of Obstetrics & Gynecology, University of Louisville School of Medicine, Louisville, KY; and
†Baylor Scott & White Medical Center, Irving, TX.
Correspondence: Deslyn T. G. Hobson, MD, Department of Obstetrics & Gynecology, University of Louisville School of Medicine, 550 S Jackson St, Louisville, KY 40202. E-mail: firstname.lastname@example.org.
No funding was received for this study.
This study was presented as oral poster at the annual scientific meeting of the American Urogynecologic Society in Providence, RI (October 4–7, 2017).
The authors have declared that there are no conflicts of interest.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.fpmrs.net).