This infodemiological study used a convergent mixed methods approach [29,30] to analyze publicly available Twitter content related to dental care affordability. This study was submitted to the Institutional Review Boards of New York University (IRB-FY2021-5634) and the University of Washington (STUDY00013725). In alignment with federal regulations regarding the use of publicly available data for research, both institutional reviews determined that the study did not meet the criteria for research involving human subjects and that no further review was required.

Data were obtained from Twitter, a free social media website created in 2006. Tweets can remain visible to the public or can be made visible only to approved followers, at the discretion of the user. Only publicly available tweets were used in this study, and usernames were removed for privacy protection.

Search terms were generated to identify tweets that discussed financial considerations associated with dental treatment. These search terms were tested and expanded through pilot queries and assessments until they were refined to the following word stems: “dental,” “dentist,” “tooth,” “teeth,” “root canal” AND “expensive,” “pay,” “afford,” and “money.”

Using a newly created account on Twitter, tweets were collected between September 1, 2017, and September 17, 2021, using Snscrape [31-33], an open-source web scraper written in Python (Python Software Foundation). The data collection script included a specific query that consisted of any combination of the search terms. Duplicate tweets, foreign language tweets, and retweets were then excluded. In addition to the tweets, metadata was collected as follows: “url,” “date,” “renderedContent,” “user,” “replyCount,” “retweetCount,” “likeCount,” “quoteCount,” “retweetedTweet,” “quotedTweet,” “mentionedUsers,” “coordinates,” “place,” “hashtags,” and “cashtags.”

Metadata related to the users’ location were used in order to limit the included tweets to those in the United States. Content was excluded from the data set of tweets for any of the following reasons: (1) content was unrelated to dental-treatment needs or experiences; (2) content was determined to be an advertisement; (3) content pertained to purely cosmetic or orthodontic dental procedures; (4) content pertained to veterinary dentistry; (5) content was classified as a joke or sarcasm; and (6) content was in reference to dental policies outside of the United States.

Data were analyzed qualitatively using both deductive and inductive approaches. Two members of the research team with clinical dental knowledge (SY and LB) co-coded all of the tweets. In total, 8% (440/5500) of all of the included tweets were coded in order to identify prominent themes and subthemes. A final codebook was developed through consensus among the members of the research team. The themes identified for the codebook included (1) general sentiments; (2) delaying or forgoing dental care; (3) payment strategies; (4) insurance; and (5) policy statements. Using the codebook, the entire sample of the included tweets was independently coded by the aforementioned researchers. In instances where the coders disagreed, they reached a consensus through discussion.

Quantitative data analysis included geographic distribution of tweets by state, volume analysis of tweets over time, and distribution of tweets by content theme. The tweets were mapped to an individual state within the United States, and the volume of tweets over time were depicted using Tableau (Tableau Software Inc), a data visualization software.

Using the search terms, we collected a total of 18,685 tweets from September 1, 2017, to September 30, 2021. Of these tweets, 2617 (14%) were removed as duplicates, and 10,754 (57.6%) were removed based on preestablished exclusion criteria. A final sample of 5314 (28.4%) tweets (from 4963 unique users) were included in the study (Figure 1). The top 5 pairs of search terms were as follows: dental_afford (520/5314, 9.8% tweets); teeth_pay (461/5314, 8.7% tweets); dentist_pay (436/5314, 8.2% tweets); dental_money (356/5314, 6.7%, tweets); and dental_expensive (353/5314, 6.6% tweets).

Thematic analysis identified the following prominent themes (Table 1): (1) general sentiments (1614/5314 tweets, 30.4%); (2) delaying or forgoing dental care (1190/5314 tweets, 22.4%); (3) payment strategies (1019/5314 tweets, 19.2%); (4) insurance (767/5314 tweets, 14.4%); and (5) policy statements (724/5314 tweets, 13.6%). The identified subthemes and illustrative quotes from the data are presented below and in Table 2.

Overall, 5314 tweets about dental affordability were collected (Table 1), with 277 (5.2%) in 2017, 1227 (23.1%) in 2018, 1341 (25.2%) in 2019, 1090 (20%) in 2020, and 1379 (26%) in 2021. The period of data collection was less than a year in 2017 (September 1 to December 31; 4 months) and 2021 (January 1 to September 30; 9 months).

The volume of tweets over the study period is depicted in Figure 2. There was a monthly average of 116 included tweets. The lowest number of tweets was observed in April 2020 (58 tweets, 1.1%), and the highest number of tweets was observed in September 2021 (673 tweets, 12.7%). Overall and across each year of collection, “general sentiments” was the most prevalent thematic category (Table 1). “Delaying or foregoing care” was the second most prevalent thematic category, except in 2019 and 2021, when “payment strategies” and “policy statements,” respectively, were the more frequently observed categories. “Policy statements” was consistently the least frequently observed thematic category, except for a significant increase in 2021.

The geographic location of tweets was determined and mapped in Figure 3. Tweets originated the most from the most populous states in the United States, which are California, Texas, Florida, and New York [34].

This study leveraged Twitter to examine public sentiments toward dental care affordability in the United States. Twitter users expressed dissatisfaction with the cost of dental treatment and the ability to access dental care. The overwhelmingly negative sentiments found in this study provide insight into how individuals are coping with financial barriers, including delaying or foregoing care and pursuing various payment strategies. Our study findings support the conclusions of the existing dental literature on dental care affordability [2-5,7]. In the 2019 National Health Interview Survey, it was found that 19.2% of women and 15.6% of men did not access their needed dental care because of cost in the prior 12 months [35]. Rates of forgoing dental care are particularly high for uninsured adults, where 1 in 2 had not seen a dentist because of costs [36].

Moreover, our study findings highlight the fact that financial barriers to accessing dental care are prevalent even among those with dental insurance. The expressed sentiments of discontent with dental insurance payment structures echoed those of the existing reports in the literature. In a study on US health care spending, among 154 conditions examined, oral disorders requiring dental care had the highest out-of-pocket costs [37]. Another study that supports the sentiments of insured tweeters in this study reported that among individuals who were insured all year, US adults were significantly more likely than adults in other developed countries to go without care because of costs, the possibility of facing high out-of-pocket spending, or the financial burden of medical bills [36].

In an effort to come up with the money required to be paid out of pocket, users reported a variety of payment strategies that included credit cards, loans, stimulus checks, tax refunds, and donations. Crowdfunding, the web-based solicitation of public donations, has become a major financer of health care–related costs [38-41]. This trend was reflected in our study findings and reinforces sentiments that the use of crowdfunding to cover direct health care expenses may be a sign of a failing system [42]. Future research may explore crowdfunding for dental procedures.

Two relevant societal events occurred during the study period, which were the COVID-19 pandemic and the election of President Biden. COVID-19 resulted in dental office closures across the country in the spring and summer of 2020, with the greatest decline in weekly visits compared with 2019 observed in the week of April 12, 2020 [43]. This trend corresponds with the quantitative findings of this study, in which the lowest number of tweets were observed in April 2020. Further, COVID-19 had an economic impact that appeared to exacerbate financial barriers to accessing dental care for some people. One user tweeted, “…I can't wait to get an eye exam and new glasses, not to mention long overdue dental care! Or just not because I can't afford it … because of #COVID19”. For others, economic stimulus checks distributed during the COVID-19 pandemic helped reduce financial barriers to dental care; for example, “@politico Got my stimulus money today. Thinking of getting much-needed dental work. Several teeth fell out when I had COVID in Dec. and Jan. Thank you President Joe!” Lastly, once dental offices reopened after COVID-19–related closures, there were tweets expressing displeasure with increased costs of care due to safety measures; for example, “…Had to pay an up charge for PPE for a dental appointment over the summer. Was told insurance was not likely to cover that aspect of the cleaning....” The election of President Biden in 2021 renewed conversations about health care reform, and the volume of tweets coded as “policy statements” increased in correspondence. In September 2021, there were 3.4 times the monthly average number of tweets. We hypothesize that this increase is a reaction to the announcement in August 2021 that President Joe Biden’s budget-reconciliation package included funding for a standard Medicare dental benefit [44-46]. Ultimately, the Medicare dental benefit was not included in the House of Representatives’ passed legislation.

The trends in the volume of tweets over time as well as the geographic distribution of the tweets support the generalizability and transferability of the findings of this study. However, this study is not without limitations. While all demographic categories have been shown to engage with social media to varying extents in the United States, Twitter users are not necessarily representative of the US population [47]. On average, Twitter users are younger, are more likely to identify as Democrats, are more highly educated, and have higher incomes than US adults overall [48]. Further, the sample of tweets collected may have some sampling bias, as the data set was not necessarily limited to a single tweet per unique user.

The findings illustrate the real-time sentiment of Twitter users toward the cost of dental treatment and suggest shortcomings in funding, which may be representative of a greater systemic failure in the provision of dental care. Thus, this study provides insights for policy makers and dental professionals who strive to increase access to dental care. Limitations of dental insurance payment models, both public and private, are one such area that may be explored.