Artigos

1. Introduction
The emergence of social media has created an environment that is a mixture of true and false information, leading to an unprecedented amount of data on various topics, including health1. Policymakers and the social media industry are faced with the challenge of controlling fake news, misinformation, and hate speech. In the same vein, the medical field also grapples with the spread of false, inaccurate, or incomplete health information2.
Although the manipulation of news is not a new phenomenon, false or highly misleading political "news" stories on social media came to the forefront during the 2016 US presidential elections and the UK Brexit Referendum3. In 2016, the Oxford Dictionary Word of the Year was "post-truth," which denotes that the public is more influenced by emotional appeals rather than objective facts4.
Defining misinformation can be a complex task, as the boundary between truth and falsehood is not always clear-cut and can shift over time with the emergence of new evidence and advancements in research methods and technologies5. In this sense, health misinformation can be defined as false health-related statements that lack scientific evidence6. It is prevalent on social media, and studies have shown a correlation between sharing misinformation and health-related knowledge, attitudes, and beliefs1 Misinformation can take various forms, including satire, parody, false connection, misleading content, false context, and imposter, manipulated or fabricated content7. For instance, there is evidence of health misinformation on social media posts related to communicable diseases8, general vaccination9, genetically modified organisms10, MMR vaccine11, zika virus12, cancer13, inflammatory bowel disease14, fluoride15, and health myths16. The consumption of incorrect oral health information can lead to beliefs that affect individual's health practices, which can have negative consequences on oral health outcomes17.
In Dentistry, misinformation about dental caries is prevalent on websites, and positive feelings are related to the spread of misinformation18. The misinformation includes a false relationship between antibiotics and dental caries, denial of the role of fluoride in preventing caries, and the disregard of sugar as a caries risk factor15, 19. This type of misinformation can lead to patients adopting harmful behaviors based on empirical evidence, which can damage the patient-professional relationship and oral health outcomes17. As such, dental professionals must pay attention to their patients' needs and produce high-quality digital materials while providing informative advice during clinical consultations20.
In light of the lack of evidence available on social media, this study aimed to perform the content analysis of Facebook posts related to dental caries in Brazilian Portuguese. The focus was on identifying and characterizing misinformation, as well as predicting the factors that influence users' interaction. Facebook has the second-highest rate of disinformation, trailing only behind Twitter21. Besides, it is recognized as a social media platform where misinformation spreads more rapidly compared to others22. In this sense, our hypothesis (H1) was that social media users are more likely to engage with posts that contain both misinformation and evoke positive emotions, compared to those that present only accurate information. This assumption is based on previous studies that have shown how emotional appeals can be more influential than factual information in shaping individuals' attitudes and behaviors3, 4.

2. Materials and Methods
2.1 Study Design
This digital study identified and characterized dental caries-related information from 500 posts published in Brazilian Portuguese on Facebook between August 2016 and August 2021. Two independent investigators (M.R. and O.S.J.) analyzed those posts qualitatively to define their author's profile, sentiment, aim of content, motivation, and facticity. Statistical analysis was performed regarding interaction metrics as mentioned below.

2.2 Ethics
This study did not require institutional review board approval from the Council of Ethics in Human Research of Bauru School of Dentistry, University of São Paulo because federal regulations do not apply to research using publicly available data that does not involve human subjects23.
The raw data of this paper have been anonymized and disclosed in an open data repository24.

2.3 Search strategy, data collection, and preprocessing dataset
To collect data, we used the Meta-owned web scrapping tool CrowdTangleTM. It tracks public interaction on content from Facebook pages and groups, Instagram accounts, and subreddits. We employed a specific search strategy and ranking criterion to collect Facebook posts and their interaction related to dental caries. The search strategy ("cárie” OR “carie" OR “cáries” OR "lesões cariosas" OR "dente cariado") was developed through exploratory analyses of hashtags and terms to ensure comprehensive coverage of dental caries-related content on Facebook. A dataset of 88,134 posts was downloaded in a .csv file format on August 26, 2021, covering a specific language (Portuguese) and timeframe (August 2016 to August 2021), with posts ranked by users' total interaction. The .csv files contain the respective link to the post, which allows the complete extraction and analysis of the content posted on Facebook by the research team, and also contain information on the date of publication and interaction metrics of each post, such as total interaction and overperforming score.
Two independent investigators (M.R. and O.S.J.) pre-processed the raw dataset before analyzing 783 posts in full to obtain a sample of 500 posts containing dental caries-related content. We excluded posts that were unrelated to dental caries (n=36), duplicated posts (n=225), and posts with unavailable links (n=22) (Figure 1). The selected posts were then printed, and anonymized by blacking out names, profiles, and people's eyes in images. To ensure standardization and prevent inconsistencies, the posts were numbered and saved in sequence in Google Slides (Google, Mountain View, CA, USA), which was later converted to a .pdf file. This systematic process allowed for ethical analysis of messages by different investigators at different times.
Total interaction represents the sum of all reactions, shares, and comments on a post on Facebook, while the overperforming score indicates the diffusion performance of a post relative to the interaction of the last 100 posts on the same account at the same time. The platform's algorithm disregards the top and bottom 25% of posts and determines the average number of interactions for the remaining middle 50% of posts across various time intervals (such as 15 minutes old, 60 minutes old, 5 hours old, etc.). Later, when the account under consideration uploads a new post, the platform contrasts its metrics with the calculated average and applies the corresponding weights from each dashboard to the difference obtained15, 25.

2.4 Data analysis
2.4.1 Qualitative analysis
Dental caries-related posts were characterized by passive qualitative analysis26, examining information patterns and interaction metrics. Two trained and calibrated investigators (M.R. and O.S.J.) classified the posts independently, according to the following criteria: author's profile (regular users, business, dental office, or news agency), sentiment (negative, neutral or positive), the aim of content (prevention or treatment), motivation (commercial or noncommercial), and facticity (information or misinformation). For the training process, several publications about dental caries were reviewed along with a third investigator (T.C.) for discussion and to learn the criteria for evaluation. Afterward, 10% of the sample (50 posts) was evaluated until the desired level of intra-examiner agreement was obtained (greater than 0.8). The posts that investigators divergently qualified were re-accessed until consensus.
Facebook profiles and pages were categorized based on their descriptions into regular users (including digital influencers), business pages, dental offices, or news agencies. The aim of content was determined by its perceived intention to control risk factors of the disease (prevention) or treat its clinical consequences (treatment). Posts were classified as commercial if they were published by profiles of stores and companies with an explicit sales intention. The sentiment of posts was categorized as positive if they contained signs such as smiles, motivational ideas, and/or happy emojis. Rational posts, such as journalistic news and scientific results, were classified as neutral, while negative sentiment posts were those that presented sad people, texts with words loaded with negative connotations, and conspiracy theories.
In terms of identifying the intentionality of a message, this study considered misinformation to be a broad term encompassing false or misleading content, regardless of whether or not there was intent to deceive or cause harm. This umbrella term includes two types of information disorders: misinformation per se and disinformation, as outlined by several sources7, 27, 28. To assess the accuracy of the content, the study relied on current scientific evidence from guidelines, consensuses, and systematic reviews on the management of dental caries29-31. A post was considered to contain misinformation only if it presented obvious false or misleading information that could potentially harm Facebook users.

2.4.2 Statistical analysis
The Statistical Package for Social Sciences (v. 28.0; SPSS; Chicago, IL, USA) was used for the statistical analysis. The variables were initially dichotomized based on specific criteria. These included time of publication (?1479 days or >1479 days), author's profile (regular users or business/dental office/news agency), sentiment (negative/neutral or positive), aim of content (prevention or treatment), motivation (commercial or noncommercial), facticity (information or misinformation), total interaction (?1659 or >1659), and overperforming score (?8.61 or >8.61). The continuous variables were dichotomized based on their median values. In author’s profile, dental offices, news agencies, and business profiles were grouped together due to their financial background. The choice to split sentiment into two categories was warranted based on prior studies indicating a link between positive emotions and higher social media user engagement rates32. This was done to examine if comparable results would be found for false messages concerning dental caries.
The inter-examiner agreement was determined by the Intraclass Correlation Coefficient (ICC), with values varying from 0.80 to 0.98.
The Kolmogorov-Smirnov and Levene tests were conducted to assess the normality and homogeneity of the data, respectively. Since the data were non-normally distributed, the Mann-Whitney U test was used to compare the total interaction and overperforming scores of the dichotomized variable groups.
Finally, the distribution of dichotomized variables based on the aim of content and facticity was evaluated using the Chi-Square Pearson test.
Multiple logistic regression models were created to investigate the possible association of distinct factors with misinformation, total interaction, and overperforming scores. Factors with Wald statistics with P<0.20 in the simple models were included in the multiple models. Statistical significance was considered when P values <0.05 for all analyses.

3. Results
As expected, most posts came from Brazil (n=451; 90.2%), although posts from 12 other countries were also identified. The majority were from business profiles (n=471; 94.2%) with noncommercial motivation (n=444; 88.8%, e.g. “End of the little motor? It is already possible to remove cavities with laser”). Additionally, most posts were about prevention (n=336; 67.2%, e.g. “On this world cavity day, advise your patients to increase the age at which sugar is introduced into their children's lives, as this is one of the most effective ways of prevention”) and expressed positive sentiment (n=327; 65.4%, e.g. “Home remedy to eliminate tartar forever”). A significant proportion of posts (n=302; 60.4%) contained accurate information about dental caries (e.g. “Dentists recommend that parents brush their children's teeth until they are 9 years old”). The overperforming scores were significantly higher in posts from business profiles, posts expressing positive sentiment, and posts with commercial motivation (Table 1).
Tables 2 and 3 provide an overview of the distribution of dichotomized variable groups according to the aim of content and facticity. The findings indicate that posts from regular users were more prevalent in treatment-related posts as opposed to preventive content. Conversely, preventive content was found to be more frequently associated with commercial motivation than its opposite. Moreover, posts containing misinformation (e.g. "Aspirin may regenerate teeth after decay, scientists say.") tended to have positive sentiment and commercial motivation more frequently than posts with only accurate information.
Table 4 displays the positive associations of misinformation with positive sentiments (OR = 1.778; P=0.005) and commercial motivation (OR = 1.846; P=0.038) (model 1). Similarly, it summarizes the significant positive associations of the overperforming score (>8.61) with positive sentiment (OR = 1.992; P<0.001) and a business profile (OR = 3.020; P=0.014) (model 2). Total interaction was not associated with any factor significantly.

4. Discussion
To our knowledge, this is the first study to examine Facebook posts about dental caries in Brazilian Portuguese, specifically focusing on misinformation and interaction metrics. The majority of the posts were originated from Brazil and were linked to business profiles that featured preventive content with noncommercial motivation. In addition, a high percentage of posts presented misinformation, which were closely linked to positive sentiment and commercial motivation. Concomitantly, business profile and positive sentiment were predictors of higher engagement of Facebook users. These findings support our hypothesis that social media users are more likely to engage with dental caries-related posts that express positive sentiment and contain misinformation than those that are based solely on accurate information.
A previous systematic review33 showed that social media platforms contain an average of 36.5% of health misinformation, which is similar to our findings (39.6%). According to Grinberg et al.4, misinformation spreads 70% more than true news. This means that people are more likely to believe in misinformation than accurate information, particularly when the false news is consistent with their political opinions34. This phenomenon aligns with the theory of innovations, which describes how new behaviors, practices, opinions, conventions, or technologies spread from person to person through social relationships35. Other factors that contribute to the spread of misinformation include individuals' lack of reflexivity regarding the accuracy of the information3, and overconfidence that prevents people from slowing down and engaging in reflective reasoning36. Additionally, belief in fake news is also associated with delusionality, dogmatism, religious fundamentalism38, bullshit receptivity, and overclaiming38.
One important factor to consider is the confirmation bias that people experience, which is the tendency to favor information that confirms their pre-existing beliefs and to reject information that contradicts them39. As a result, individuals may be more inclined to believe poorly supported claims that align with their strongly held beliefs40. Moreover, the availability heuristic is also relevant, as it involves the likelihood of believing information based on previous exposure to it41. In fact, a single exposure to a headline containing misinformation can increase people's later belief in the headline42.
Based on our study, it can be inferred that people tend to believe information from profiles they consider trustworthy, which is consistent with previous research indicating that users perceive business pages on Facebook as highly reliable43. Furthermore, our findings suggest that positive sentiment is a significant predictor of higher diffusion of posts and that misinformation tends to be emotionally evocative, eliciting strong reactions, as reported by Kozyreva44. This is in line with previous studies that have shown that individuals who experience more emotions, both positive and negative, are more likely to believe in false news45. The prevalence of positive emotions observed in our study may be attributed to the diverse range of content, including prevention and seemingly miraculous treatments, which is reminiscent of previous research on cancer-related misinformation16.
The association between commercial motivation and misinformation is worth considering, as social media engagement metrics such as likes and shares have been found to enhance the credibility of news content, particularly when it comes to misinformation46. This means that popular commercial pages with high engagement rates can leverage their social media popularity to sell questionable products (such as fluoride-free dental products, purportedly to prevent cavities), which may still attract a high rate of consumers who believe in the veracity of the information presented47.
Some limitations to this study should be considered. Firstly, we limited our analysis to posts written in Brazilian Portuguese. As a result, it is possible that the cultural nuances and characteristics unique to this language and region could have impacted the factors linked to the dissemination of information on dental caries. Secondly, although videos are known to have high engagement rates similar to photos48, we did not include them in this investigation due to the challenges associated with obtaining and analyzing the full-length content available on this media accurately. In addition, the sample size of our study was restricted due to the difficulties that arise when conducting content analysis through human evaluation. We had to take into account the feasibility of manually classifying datasets, which is in line with the methodology adopted in previous investigations15, 49.
In conclusion, these results indicate a significant proportion of dental caries-related posts containing misinformation, particularly those associated with positive emotions and commercial motivation. Moreover, positive emotions and business profiles are significant predictors of higher post dissemination. Thus, it is crucial to implement specific policies aimed at ensuring quality information on social media. This can be achieved through various measures, such as developing appropriate content, promoting critical thinking when consuming health-related information, and filtering information using technology-based solutions.

5. Acknowledgemets
This study was supported by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) (Grant #001).The authors thank Meta for granting the use of CrowdTangleTM platform.

Scielo Data Repository: https://doi.org/10.48331/scielodata.XTCTSF

6. References
1. Chou WYS, Oh A, Klein WMP. Addressing health-related misinformation on social media. JAMA. 2018; 320(23):2417-18.
2. Southwell B, Thorson E, Sheble L. The persistence and peril of misinformation. AmSci. 2017; 105(6):372.
3. Pennycook G, Rand DG. The psychology of fake news. Trends in Cogn Sci. 2021; 25(5):388–402.
4. Grinberg, N, Joseph K, Friedland L, Swire-Thompson B, Lazer D. Fake news on Twitter during the 2016 U.S. presidential election. Science. 2019; 363(6425), 374–8.
5. Swire-Thompson B, Lazer D. Public health and online misinformation: Challenges and recommendations. Annu Rev Public Health. 2020; 41:433-51.
6. Chou WYS, Gaysynsky A, Capella JN. Where we go from here: Health misinformation on social media. Am J Public Health Association. 2020; 110(S3):S273-5.
7. UNESCO. “Fake news” & disinformation: handbook for journalism education and training. 2018. https://unesdoc.unesco.org/ark:/48223/pf0000265552. Accessed 24 Apr 2023.
8. Adebimpe WO, Adeyemi DH, Faremi A, Ojo JO, Efuntoye AE. The relevance of the social networking media in Ebola virus disease prevention and control in Southwestern Nigeria. Pan Afr Med J. 2015; 22(Suppl 1):7.
9. Schmidt AL, Zollo F, Scala A, Betsch C, Quattrociocchi W. Polarization of the vaccination debate on Facebook. Vaccine. 2018; 36(25):3606–12.
10. Bode L, Vraga EK. See something, say something: correction of global health misinformation on social media. Health Commun. 2017; 33(9):1131–40.
11. Aquino F, Donzelli G, De Franco E, Privitera G, Lopalco PL, Carducci A. The web and public confidence in MMR vaccination in Italy. Vaccine. 2017; 35(35 Pt B):4494– 8.
12. Sommariva S, Vamos C, Mantzarlis A, ?ào LUL, Tyson DM. Spreading the (fake) news: exploring health messages on social media and the implications for health professionals using a case study. Am J Health Educ. 2018; 49:246–55.
13. Okuhara T, Ishikawa H, Okada M, Kato M, Kiuchi T. Assertions of Japanese websites for and against cancer screening: a text mining analysis. Asian Pac J Cancer Prev. 2017; 18:1069–75.
14. Groshek J, Basil M, Guo L, Ward SP, Farraye FA, Reich J. Media consumption and creation in attitudes toward and knowledge of inflammatory bowel disease: web-based survey. J Med Internet Res. 2017; 19(12):e403.
15. Lotto M, Menezes TS, Hussain IZ, Tsao S-F, Butt ZA, Morita PP, Cruvinel T. Characterization of false or misleading fluoride content on Instagram: infodemiology study. J Med Internet Res. 2022; 24(5):e37519.
16. Waszak PM, Kasprzycka-Waszak W, Kubanek A. The spread of medical fake news in social media – The pilot quantitative study. Health Policy Technol. 2018; 7(2):115–8.
17. Strieder AP, Aguirre PEA, Lotto M, Cruvinel AFP, Cruvinel T. Digital behavior surveillance for monitoring the interests of Google users in amber necklace in different countries. Int J Paediatr Dent. 2019; 29(5):603-14.
18. Vosoughi S, Roy D, Aral S. The spread of true and false news online. Science 2018;359:1146–51.
19. Blizniuk A, Furukawa S, Ueno M, Kawaguchi Y. Evaluation of English websites on dental caries by using consumer evaluation tools. Oral Health Prev Dent. 2016; 14:363–9.
20. Aguirre PEA, Aníbal I, Strieder AP, Lotto M., Rizatto VL, Cruvinel AFP, et al. Online quality and readability assessment of early childhood caries information available on websites from distinct countries: A cross-sectional study. Eur J Paediatr Dent. 2022; 23(1):15-20.
21. Trust Lab. Code of practice on disinformation. Brussels: European Union; 2023.
22. Guess AM, Nyhan B, Reifler J. Exposure to untrustworthy websites in the 2016 US election. Nat Hum Behav. 2020; 4(5):472-80.
23. Lotto M, Hanjahanja-Phiri T, Padalko H, Oetomo A, Butt ZA, Boger J, Millar J, Cruvinel T, Morita PP. Ethical principles for infodemiology and infoveillance studies concerning infodemic management on social media. Front Public Health. 2023; 11:1130079.
24. Remiro M, Santana Jorge O, Lotto M, Oliveira TM, Machado MAAM, da Silva TC. Raw data of the manuscript: Facebook users' engagement with dental caries misinformation in Brazilian Portuguese [Internet]. figshare; 2023 [cited 2023Nov9]. https://figshare.com/articles/dataset/Untitled_ItemRaw_data_of_the_manuscript_Facebook_users_engagement_with_dental_caries_misinformation_in_Brazilian_Portuguese/24534790/2
25. Caso P, Trevisan M, Vassio L. Disentangling the information flood on OSNs: Finding notable posts and topics. IEEE International Conference on Data Mining Workshops (ICDMW), Orlando, FL, USA. 2022; 1168-75.
26. Franz D, Marsh HE, Chen JI, Teo AR. Using Facebook for qualitative research: a brief primer. J Med Internet Res. 2019; 21(8):e13544
27. Molina MD, Sundar SS, Le T, Lee D. “Fake News” is not simply false information: a concept explication and taxonomy of online content. Am Behav Sci. 2019; 65(2):180-212
28. Wardle C, Derakhshan H. Information disorder: Toward an interdisciplinary framework for research and policy making. Strasbourg: Council of Europe; 2017.
29. Machiulskiene V, Campus G, Carvalho JC, Dige I, Ekstrand KR, Jablonski-Momeni A, et al. Terminology of dental caries and dental caries management: Consensus report of a workshop organized by ORCA and Cariology Research Group of IADR. Caries Res. 2020; 54(1):7-14.
30. Tinanoff N, Bazed RJ, Guillory CD, Donly KJ, Feldens CA, McGrath C, et al. Early childhood caries epidemiology, aetiology, risk assessment, societal burden, management, education, and policy: Globalperspective. Int J Paeditr Dent. 2019; 29:238-48.
31. Pitts NB, Zero DT, Marsh PD, Ekstrand K, Weintraub JA, Ramos-Gomez F, et al. Dental caries. Nat Rev Dis Primers. 2017; 3(1):17030.
32. Klassen K, Borleis E, Brennan L, Reid M, McCaffrey T, Lim M. What people “like”: analysis of social media strategies used by food industry brands, lifestyle brands, and health promotion organizations on Facebook and Instagram. J Med Internet Res. 2018; 20(6):e10227.
33. Suarez-Lledo V, Alvarez-Galvez J. Prevalence of health misinformation on social media: Systematic review. J Med Internet Res. 2021; 23(1):e17187.
34. Kahan DM. Ideology, motivated reasoning, and cognitive reflection. Judgm Decis Mak. 2013; 8(4).
35. Rogers EM. Diffusion of innovations. 5th ed. Florence: Free Press; 2003.
36. Salovich NA, Donovan AM, Hinze SR, Rapp DN. Can confidence help account for and redress the effects of reading inaccurate information? Mem Cognit, 2020; 49(2):293–310.
37. Bronstein MV, Pennycook G, Bear A, Rand DG, Cannon TD. Belief in fake news is associated with delusionality, dogmatism, religious fundamentalism, and reduced analytic thinking. J Appl Res Mem Cogn. 2019; 8(1):108–17.
38. Pennycook G, Rand DG. Who falls for fake news? The roles of bullshit receptivity, overclaiming, familiarity, and analytic thinking. J Pers. 2019; 88(2):185-200.
39. Nickerson RS. Confirmation bias: A ubiquitous phenomenon in many guises. Rev Gen Psychol. 1998; 2(2):175-220.
40. Grace PJ. Nurses spreading misinformation. Am J Nurs. 2021; 121(12):49-53.
41. Pennycook G, Cannon TD, Rand DG. Prior exposure increases perceived accuracy of fake news. J Exp Psychol Gen. 2018; 147(12):1865–80.
42. Smelter TJ, Calvillo DP. Pictures and repeated exposure increase perceived accuracy of news headlines. Appl Cogn Psychol. 2020; 34(5):1061-71.
43. Pornpitakpan C. The persuasiveness of source credibility: a critical review of five decades’ evidence. J Appl Soc Psychol. 2004; 34(2):243–81.
44. Kozyreva A, Lewandowsky S, Hertwig R. Citizens versus the internet: confronting digital challenges with cognitive tools. Psychol Sci Public Interest. 2020; 21(3):103–56.
45. Martel C, Pennycook G, Rand DG. Reliance on emotion promotes belief in fake news. Cogn Res Princ Implic. 2020; 5(1):1–20.
46. Avram M, Micallef N, Patil S, Menczer F. Exposure to social engagement metrics increases vulnerability to misinformation. 2020. https://misinforeview.hks.harvard.edu/article/exposure-to-social-engagement-metrics-increases-vulnerability-to-misinformation/. Accessed 24 Apr 2023.
47. Khobzi H, Lau RYK, Cheung TCH. The outcome of online social interactions on Facebook pages. Internet Res, 2019; 29(1):2–23.
48. Kite J, Foley BC, Grunseit AC, Freeman B. Please like me: Facebook and public health communication. PLoS One. 2016; 11(9):e016276.
49. Heaivilin N, Gerbert B, Page JE, Gibbs JL. Public health surveillance of dental pain via Twitter. J Dent Res. 2011; 90(9):1047–51.