Topical Fluoride Applications Related Posts Analysis on Twitter Using Natural Language Processing.

Basak Kiziltan Eliacik

Oral health & preventive dentistry 2021 Jan 07

filter terms:

Social media is today a comprehensive source of data that can serve as a guide to professionals in issues related to public health. The purpose of this paper is to investigate the content of topical fluoride-related Twitter posts made over a 3-year period in order to improve our understanding of Twitter users' perceptions and treatment experiences. A continuous cross-sectional sample of Tweets on the subject of 'approaches to the topical fluoride treatment of tooth decay' was collected from the Twitter social networking platform between 1 January 2017 and 1 January 2020 using a software application developed for this research that makes use of the Twitter advanced search API. The words and phrases used for the identification of related Tweets were determined through a screening of the topical fluoride keywords of previous studies, and a search was conducted in the English language. To better arrange the collected Tweets and to make the data more meaningful, firstly one of the natural language process techniques - Tokenization - was applied, after which the Tweets were converted into a set of meaningful words and regular expressions. The Tweets were then compared with each other, word-by-word, with the help of a word-based Levenshtein distance algorithm, after which two experts in the computational social science domain labelled each Tweet. A total of 132,358 Tweeter posts referencing topical fluoride applications were collected, of which 110,847 were eliminated through the use of a word-based Levenshtein distance algorithm, and the remaining corpus of 21,511 posts was analysed and evaluated for specific content. Within the garnered data, 48.5% (n = 10,428) of the Tweeter posts concerned topical fluoride treatments, and 7% (n = 1,507) reported experiences with topical fluoride treatment. Negative Tweeter posts about topical fluoride treatment (5,679, 26.4%) vastly outnumbered those that were positive (3,897, 18.1%). The current study achieved its main objectives of analysing topical fluoride application-related posts made on social media. From the garnered Twitter data, it can be understood that Twitter users regularly share their concerns and negative sentiments about the side effects of topical fluoride treatments on the platform. Future explorations of social media may aid public health and dental professionals in the development of strategies to educate the public and to raise awareness of the importance of topical fluoride applications.

Citation

Basak Kiziltan Eliacik. Topical Fluoride Applications Related Posts Analysis on Twitter Using Natural Language Processing. Oral health & preventive dentistry. 2021 Jan 07;19(1):457-464