885610

research-article2019
JDRXXX10.1177/0022034519885610Journal of Dental ResearchPublic Sense of Water Fluoridation on Twitter

Discovery!
Journal of Dental Research
2020, Vol. 99(1) 11­–17
Public Sense of Water Fluoridation © International & American Associations
for Dental Research 2019

as Reflected on Twitter 2009–2017 Article reuse guidelines:

sagepub.com/journals-permissions
DOI: 10.1177/0022034519885610
https://doi.org/10.1177/0022034519885610
journals.sagepub.com/home/jdr

H.J. Oh1,2, C.H. Kim2,3, and J.G. Jeon2,4

Abstract
Though controversial, water fluoridation has been hailed as one of the top-ten public-health achievements of the 20th century in the
United States of America. In this article, we aim to investigate the public sense of water fluoridation as reflected on Twitter, using data
from 2009 to 2017. To this end, tweets related to water fluoridation were collected using queries such as “fluoridated water or fluoride
water,” “water fluoridation or fluoridation of water,” and hashtags related to water fluoridation. The collected tweets (n = 218,748)
were examined through informetric, linguistic (word sentiment, word frequency, and word network analyses), and issue tweet analyses.
We found that Twitter users who tweeted about water fluoridation in English between 2009 and 2017 constituted about <0.01% of all
users including non-English users. In their tweets, words such as “poison” and “waste” were the strong negative sentiment words most
often used. Of the top 30 words most frequently used, words related to information sources on water fluoridation and the safety of
water fluoridation appeared more often than words related to its efficacy. Additionally, the words related to information sources on
water fluoridation and the safety of water fluoridation were found to be core terms in the sentences of tweet mentions. Our linguistic
analyses indicate that Twitter users responded sensitively to words that emphasize negative aspects of fluoridation. This is clearly shown
in our issue tweet analysis, where tweet mentions expressing negative opinions about water fluoridation accounted for at least 59.2%
of all mentions. By contrast, <15% of tweet mentions were found to be positive. These findings suggest that professionals need to
reevaluate the current state of online information about water fluoridation, and improve it in a way so that the public can easily access
reliable information sources.

Keywords: informetrics, linguistic analysis, fluoride, safety, efficacy, social networking services

Introduction
Water fluoridation has been hailed as one of the top-ten public-
health achievements of the 20th century in the United States of
America (Centers for Disease Control and Prevention, 1999).
Furthermore, many research articles have asserted its efficacy,
safety, and cost-effectiveness (Kaminsky et al. 1990; Iheozor-
Ejiofor et al. 2015; Ajiboye et al. 2018) to the effect that it may
have been a major contributor to the dramatic decline in dental
caries from the 1950s to the 1980s. In contrast to this wide-
spread acceptance of water fluoridation and its role in aca-
demic society, the general public’s attitudes toward it have
largely been, and still remain, unknown.
Monitoring public response to a social issue, such as water
fluoridation, is of considerable interest to health and dental
professionals, health policymakers, and government agencies.
Traditionally, institutions have had to rely on expensive and
time-consuming surveys administered by polling agencies in
order to measure public response. However, the recent emer-
gence of social networking services (SNSs) such as Twitter,
Facebook, and Instagram, offers new opportunities for tracking
public response. Data can be collected and analyzed from
SNSs in real time with the objective of surveying public opin-
ions toward a particular subject (Eysenbach 2011). Of the vari-
ous forms of SNSs, Twitter, which enables users to publish
short messages no more than 140 characters (whose length has
been increased from 140 to 280 characters since November
2017), has quickly grown in popularity amongst researchers
because it is easy to use and cheap, and it can further offer
contextual insights not captured by traditional survey question-
naires (Davis et al. 2017).
Indeed, in recent years, researchers have analyzed massive
amounts of data produced by SNSs such as Twitter, in order to
glean insights into topics of interest to public health. Their out-
puts are increasingly being considered valuable sources of
1
Department of Library and Information Science, Jeonbuk National
University, Jeonju, Republic of Korea
2
Institute of Medical Information Convergence Research in JBNU, Jeonju,
Republic of Korea
3
Department of English Literature, Jeonbuk National University, Jeonju,
Republic of Korea
4
Department of Preventive Dentistry, School of Dentistry, BK21 plus
program, Jeonbuk National University, Jeonju, Republic of Korea
A supplemental appendix to this article is available online.
Corresponding Author:
J.G. Jeon, Department of Preventive Dentistry, School of Dentistry,
BK21 Plus Program, Jeonbuk National University, 567 Baekje-daero,
Jeonju, 54896, Republic of Korea.
Email: dentjjk@jbnu.ac.kr
12
health information (Eysenbach 2009; Moorhead et al. 2013).
For example, Twitter data have been analyzed to study natural
disasters (Black et al. 2015), infectious disease outbreaks
(Chew and Eysenbach 2010), drug and alcohol use (Cavazos-
Rehg et al. 2015), and public responses to health policies (King
et al. 2013). Such online information, however, has been barely
exploited by dental researchers.
In this article, we use English-language Twitter data to probe
public response to a specific topic in dentistry – water fluorida-
tion. The findings we report may offer a new perspective in the
area of dentistry that would not have otherwise been made
available by the traditional way of measuring public opinion.
Materials and Methods
Data Source and Search Strategy
The experimental strategy employed in this study is shown in
Appendix Figure 1. Data collection from Twitter was per-
formed using an open Application Programming Interface
(API) on GitHub produced by Twitter. Using the API, one can
collect tweets matching certain query criteria and access meta-
information including user IDs, mentions, tweet generation
time, and the number of replies, retweets, and favorites
(Appendix Fig. 2). The queries used in this study are shown in
the Appendix.
Data Collection
The selected queries were submitted to Twitter through an API,
and our target tweet information between 2009 and 2017 was
retrieved. We collected the retrieved data using an automatic
web crawler (https://github.com/Jefferson-Henrique/GetOld
Tweets-python) and converted the collected tweets (n =
219,224) to XML format (Oh et al. 2017). Of the data col-
lected, non-English tweets (n = 476) were excluded. With the
removal of non-English tweets, the information from 218,748
tweets—mentions, tweet generation time, and the number of
replies, retweets, and favorites—was used for informetric, lin-
guistic, and issue tweet analyses (Appendix Fig. 1). To check
the validity of this method for identifying tweets related to
water fluoridation, a random sample of 1,000 tweets was
selected from our collected data and analyzed. Two authors
(H.J.O. and J.G.J) reviewed each tweet separately to determine
if it was indeed related to water fluoridation. Only 0.1% (10
tweets) of 1,000 tweets were found to be unrelated to water
fluoridation.
Informetric Analyses
Of the tweet information gathered, tweet generation time and
the number of replies, retweets, and favorites were informetri-
cally analyzed. To identify Twitter users’ interest in water fluo-
ridation and its (de)growth pattern over time, we calculated the
proportion of Twitter users who tweeted about water fluorida-
tion relative to all the Twitter users and the change in this
Journal of Dental Research 99(1)
proportion over time. In addition, we traced the changes over
time in the number of tweets, retweets, and favorites related to
water fluoridation.
Linguistic Analyses
Word Sentiment Analysis. We performed word sentiment anal-
ysis to reveal the negative and positive words used in tweet
mentions. A lexicon of words each with previously assigned
numeric measures of emotion is commonly used for word sen-
timent analysis. For our purposes, we used labMT 1.0, a lexi-
con developed by Dodds et al. (2011) (see labMT 1.0 in the
Appendix for additional information). After the collected tweet
mentions were processed to remove stopwords, the assigned
scores for words in a given text were analyzed. We analyzed
not just the entirety of tweets related to water fluoridation, but
also isolated and analyzed tweets that received more than 10
retweets or favorites. Normalized frequency was used to assess
how many times a word (positive or negative score) occurred
in tweets during a selected year via the following equation:
number of occurrences of a word in a selected year divided by
number of tweets in the same year. In addition, the frequency
of negative score words from 1 to 3 and positive score words
from 7 to 9 were expressed as word clouds (see Word Clouds
in the Appendix for further information).
Word Frequency Analysis. Word frequency analysis was con-
ducted to find out which topics related to water fluoridation
people were most interested in. For the analysis, functional
contentless words (such as articles and conjunctions) were
removed from the collected tweet mentions before the fre-
quency of word use during a selected year was investigated.
Subsequently, the words ranked from 1 to 30 were classified
according to subject of interest.
Word Co-occurrence Network Analysis. From the list of words
obtained by word frequency analysis, we selected the 100
words most frequently used during the period 2009–2017. We
performed word co-occurrence network analysis on these
selected words in order to understand how they are intertwined
in a body of text (for detailed information, see Word Co-
occurrence Network and Fig. 3 in the Appendix). The analyzed
network was visualized by the NetDraw program in UCINET
(Analytic Technologies), a public software package for net-
work visualization tools.
Issue Tweet Analyses
The contents and meaning that each tweet mention intends to
convey are barely captured by word sentiment, word frequency
or word network analyses, as these analytic methods are
designed to analyze the individual words used in tweet men-
tions. Thus, we performed issue tweet analysis to identify the
content of tweet mentions (the meaning of sentences) that
Twitter users are interested in (see Determination of Issue
Tweet in the Appendix for additional information). In our
Public Sense of Water Fluoridation on Twitter
study, tweet mentions with tweet importance ranking 1–3
between 2009 and 2017 were sorted and analyzed.
Results
Informetric Analyses
While the number of Twitter users increased over time from
2009 to 2017, the proportion of Twitter users who mentioned
water fluoridation gradually decreased and constituted about
0.005% to 0.010% of all Twitter users (Fig. 1A). A total of
218,748 tweets related to water fluoridation were generated
during that period (Fig. 1B). The number of retweets and
favorites related to water fluoridation continued to increase
from 2012 onward, even though the generation of tweets
remained stagnant from 2013. Interestingly, tweets related to
water fluoridation showed sharp spikes in number in January
2011, May 2013, and April 2015 (Fig. 1B).
Linguistic Analyses
Word Sentiment Analysis. In the mentions of tweets related to
water fluoridation, positive words were used more often than
negative words and the ratio of the 2 was 2.25:1 (Fig. 2A-1).
Amongst words expressing strong negative sentiment (scores
between 1 and 3), words such as “poison” and “waste” were
used most frequently (Fig. 2A-2). Words such as “health,”
“like,” and “good” were found to be the ones most frequently
used to express strong positive sentiment (scores between 7
and 9) (Fig. 2A-2). In the mentions of tweets with more than 10
retweets (n = 2,424) or favorites (n = 2,390) (Fig. 2B-1), the
ratio of positive and negative word usage was similar to that in
all water fluoridation related tweets (Fig. 2B-2-1, B-2-2). The
most frequently used negative (scores between 1 and 3) and
positive words (scores between 7 and 9) in tweet mentions with
more than 10 favorites or retweets were also “poison” and
“health,” respectively, in line with the general trends identified
in all tweets related to water fluoridation (Fig. 2B-3-1, B-3-2).
Word Frequency Analysis. Regardless of the year, the word
“drinking” was the most common in tweet mentions related to
water fluoridation (Table). The words ranked from 1 to 30
between 2009 and 2017 were generally related to information
sources on water fluoridation, water fluoridation targets or sup-
pliers, effect or effect areas of water fluoridation/fluoride, safety
of water fluoridation, and water fluoridation regions. Of these,
words related to the safety or health impact of water fluorida-
tion were the most common, accounting for at least a third of
the top 30 most frequently used words. By contrast, the propor-
tion of words related to effect or effect areas of water fluorida-
tion such as “decay,” “cavities,” and “tooth” was very small.
Word Co-occurrence Network Analysis. Word co-occurrence
network analysis of tweet mentions, visualized based on
betweenness centrality, is shown in Appendix Figure 4. As
shown in Appendix Figure 4, the words ranked from 1 to 5 in
the centrality were mainly composed of terms that reflect the
13
relationship between water fluoridation and health (such as
“water,” “fluoride,” “drinking,” and “health”). The words
ranked from 6 to 20 in the centrality were composed of those
related to information sources on water fluoridation and the
safety of water fluoridation (such as “article,” “poison,” “news,”
“level,” “kid,” and “brain”).
Issue Tweet Analyses
Tweet mentions that Twitter users were interested in (i.e., men-
tions with high importance) between 2009 and 2017 are shown
in the Appendix Table. The mentions roughly fall into 3 groups:
mentions providing basic objective information about water
fluoridation, mentions emphasizing a prudent approach to
water fluoridation, and mentions expressing negative opinions
about water fluoridation. Of these groups, the mentions
expressing negative opinions about water fluoridation
accounted for at least 59.2% of all mentions (16 mentions).
These negative mentions asserted their opinion mainly by
using unproven fluoride toxicity theories or conspiracy theo-
ries. The second group of mentions constituting 25.9% of all
mentions (7 mentions), a sizable portion, provided objective
information about water fluoridation. Even these informative
mentions, however, can be construed as being not entirely
objective but loaded with negative, though subtle, sense,
because they mainly focus on the decrease of fluoride concen-
tration in water. Only 14.8% (4 mentions) of the total mentions
suggested that people should approach the topic of water fluo-
ridation rationally.
Discussion
The level of support for, or opposition against, water fluorida-
tion varies from community to community depending on its
demographic and socioeconomic characteristics (Peckham and
Awofeso 2014). This means that water fluoridation can poten-
tially be a contentious social issue. For decades, surveys con-
ducted via phone, mail, or face-to-face have been used to
measure public sensitivities to water fluoridation. While these
traditional survey methods have their own merits, they have
shortcomings as well due to limitations related to sampling, the
respondent’s attitude, and question types. To compensate for
the shortcomings inherent in traditional methods, we used
Twitter and investigated all of the data related to water fluori-
dation generated on the SNS between 2009 and 2017. This new
method of study does not just give us access to entire popula-
tions rather than small samples, but it also allows us to analyze
opinions freely described by people regardless of time point.
Twitter users may not be representative of the general popu-
lation. As described in a previous study (Davis et al. 2017),
Twitter users are considered to be distinct—they tend to be
younger and more tech-savvy. This implies that there may be
limitations in using data from Twitter for routine public health
surveillance or research for the general population.
Additionally, the content of a tweet can be contaminated by
commercial intentions given the recent trend of SNS advertis-
ing (Alalwan 2018). The interest of commercial organizations
14 Journal of Dental Research 99(1)

the rejection of voters in Portland, Oregon,

to the addition of fluoride to drinking water.
In April 2015, the US Department of Health
and Human Services updated its recom-
mendation for the optimal fluoride level in
drinking water. These results suggest that
while water fluoridation is not a major
social issue, it can be a serious social issue
that can catch the public’s attention quickly.
Additionally, the fact that the number of
retweets and favorites related to water fluo-
ridation continued to increase sharply from
2012 onward suggests that people have
begun to not only describe their individual
thoughts but also actively exchange and
share their opinions.
In the individual words constituting
tweet mentions related to water fluorida-
tion, the ratio of positive and negative
words was 2.25:1 (Fig. 2A-1). The domi-
nance of positive words is not surprising,
as it was found to be the typical pattern
exhibited by all Twitter users regardless of
the topic they are involved in (Kloumann
et al. 2012). What is of interest to us are
negative words and which of them are used
in tweets because our main concern is to
understand the nature of the arguments of
the negative words. Of the negative words
in tweet mentions related to water fluorida-
tion, “poison” and “waste” were used most
frequently (Fig. 2A-2, B-3-1, B-3-2). The
high frequency of the words “poison” and
“waste” may reflect the strong concerns of
Figure 1. Changes in the number of Twitter users or tweets related to water fluoridation Twitter users regarding possible health
from 2009 to 2017. (A) Changes over time in the proportion of Twitter users who mention damage by water fluoridation. The Table
water fluoridation. (B) Changes over time in the number of tweets, retweets, and favorites
related to water fluoridation. *The United States federal government’s recommendation to
also reflects similar concerns of Twitter
change the amount of fluoride in drinking water for the first time in 50 years (January 2011), the users regarding water fluoridation. Words
rejection of voters in Portland, OR, to the addition of fluoride to drinking water (May 2013), or related to the safety or health impact of
the US Department of Health and Human Services updating its recommendation for the optimal water fluoridation accounted for at least a
fluoride level in drinking water (April 2015).
third of the top 30 most frequently used
water fluoridation-related words. Issue
in SNSs and their role in attracting customers and building a tweet analysis in the Appendix Table adds to the conclusion
profitable marketing relationship with customers have been that Twitter users are negatively concerned about water fluori-
well documented in many studies (Alalwan 2018). Yet, even dation, as shown by the increased interest in tweet mentions
admitting these potential limitations, the data from SNSs may describing conspiracy theories, which may be due to the clas-
provide a potential indication of the general interest of SNS sification of fluoride as a pollutant rather than as a nutrient or
users on a particular issue at any point in time. medicine (Lewis and Milgrom 2001). All in all, our results
On Twitter, water fluoridation was not a popular topic (Fig. clearly show that Twitter users were strongly influenced by,
1A). However, the number of tweets related to water fluorida- and interested in, the fears and concerns surrounding water
tion experienced sharp, though temporal, increases in January fluoridation though it has proven to be a highly safe, beneficial,
2011, May 2013, and April 2015 (Fig. 1B). The sharp increase cost-effective and nondiscriminatory means to prevent caries
in tweet number in January 2011 coincides with the event in in individuals of all ages (Harris and Garcia-Godoy 2004).
which the US federal government recommended changing the In the top 30 frequent words in the Table, “News” and
amount of fluoride in drinking water for the first time in 50 “YouTube” appeared often. This suggests that Twitter users
years. The May 2013 increase in tweet number was driven by may be using news services and video-sharing websites, either
Public Sense of Water Fluoridation on Twitter 15

Figure 2. Sentiment analysis on the words in tweet mentions related to water fluoridation: (A) all tweets and (B) tweets with more than 10 retweets
or favorites. (A-1) Frequency and (A-2) clouds of words with negative and positive sentiment in all tweets. (B-1) Relationship between the number of
tweets and favorites or retweets. (B-2-1, B-2-2) Frequencies and (B-3-1, B-3-2) clouds of words with negative and positive sentiment in tweets with
more than 10 retweets or favorites.
16 Journal of Dental Research 99(1)

Table. Top 30 Words in Frequency between 2009 and 2017 in Tweet Mentions Related to Water Fluoridation.

Ranking 2009 2010 2011 2012 2013 2014 2015 2016 2017 Total
1 drinking drinking drinking drinking drinking drinking drinking drinking drinking drinking
2 tinyurl tap health filter health health health health health bit
3 tap health tap health remove added bit bottled bottled tap
4 health filter supply remove supply toothpaste teeth remove people health
5 supply supply teeth city portland people levels teeth teeth remove
6 bottled tinyurl city portland drink city org people news news
7 put poison levels poison city teeth people filter remove filter
8 teeth teeth remove drink filter free public toxic toothpaste teeth
9 new children filter put toothpaste poison thyroid city public put
10 added youtube government teeth know public government toothpaste youtube added
11 know filters calgary youtube youtube filter linked poison toxic ow
12 toothpaste toxic poison added poison stop toothpaste public food city
13 good city stop berkey teeth pic time safe poison people
14 city put video crystal gmo bottled free government filter bottled
15 poison dental put toothpaste good toxic filter food government toothpaste
16 toxic video county quest public dental adhd mercola utm know
17 news free adding su free good cancer youtube brands poison
18 public toothpaste good bottled anti harvard bottled calgary city just
19 dangers good toxic tinyurl foods food city dental brain youtube
20 people public watch big toxic government youtube kids chemtrails get
21 want supplies dental toxic sodium children poison brands dental public
22 video gl toothpaste reverse food decay community avoid gmo too
23 filter bottled public stop dental safe depression cancer epa free
24 chlorine remove tinyurl food bottled uk toxic complete children stop
25 blog reverse system council chemical iq uk community iq toxic
26 think system council osmosis decay brain decay cavities chlorine good
27 dental study people black government disorders co utm decay filters
28 stop sodium children countertop council research dental reduce gay government
29 check dangers natural news dental children supplies children children chemicals status
30 bad chlorine free children articles cancer food council mercola adding

   , related to information sources on water fluoridation;    , related to water fluoridation targets or suppliers;    , related to effect or effect
areas of water fluoridation/fluoride;    , related to safety of water fluoridation;    , related to water fluoridation regions.

as information sources on water fluoridation or as support for
their claims. Recently, several studies have revealed the posi-
tive aspects brought about by the internet health information
boom (Baker et al. 2003; Karamitros et al. 2017). However,
numerous negative outcomes have also been highlighted; take,
for example, how anyone on the Internet can publish health
information, which could be incorrect, biased, prejudiced, mis-
leading or even potentially dangerous (Karamitros et al. 2017).
In general, our study indicates that professionals who are
involved in water fluoridation need to actively take advantage
of news services and YouTube in order to guide people to reli-
able information.
Word network analysis reiterates the above points that
Twitter users are concerned about the safety of water fluorida-
tion and their source of concern is from the Internet and media.
As shown in Appendix Figure 3, words related to information
sources and the safety of water fluoridation were not only fre-
quently used but also functioned as the main terms in sentences
of tweet mentions. This result confirms the trend indicated by
the results in the Table.
In conclusion, Twitter users were mainly interested in
tweets that emphasized the negative health aspects of water
fluoridation as opposed to its efficacy. This phenomenon is
apparently driven by the biased and inaccurate information
about water fluoridation or fluoride readily accessible on the
Internet. Given that the fluoridation of water as a preventive
dental health measure has been a contentious issue from its
birth, it needs to be ensured that people are provided with
unbiased and scientific information. Professionals, academic
institutions, and professional organizations should join forces
to improve the quality of information on water fluoridation or
fluoride by providing content with scientific and unbiased
Public Sense of Water Fluoridation on Twitter
explanations and by directing people to reliable information
sources.
Author Contributions
H.J. Oh, J.G. Jeon, contributed to conception, design, data acquisi-
tion, analysis, and interpretation, drafted and critically revised the
manuscript; C.H. Kim, contributed to design, data acquisition, and
analysis, drafted and critically revised the manuscript. All authors
gave final approval and agree to be accountable for all aspects of
the work.
Acknowledgments
This work was supported by the National Research Foundation of
Korea (NRF) grant funded by the Korea government (MSIP) (No.
2019R1A2C1085086). The authors declare no potential conflicts
of interest with respect to the authorship and/or publication of this
article.
References
Ajiboye AS, Dawson DR 3rd, Fox CH, AADR Science Information Committee.
2018. American Association for Dental Research policy statement on com-
munity water fluoridation. J Dent Res. 97(12):1293–1296.
Alalwan AA. 2018. Investigating the impact of social media advertising fea-
tures on customer purchase intention. Int J Inf Manage. 42:65–77.
Baker L, Wagner T, Singer S, Bundorf M. 2003. Use of the internet and E-mail
for health care information. JAMA. 289(18):2400–2406.
Black DR, Dietz JE, Stirratt AA, Coster DC. 2015. Do social media have a place
in public health emergency response? J Emerg Manag. 13(3):217–226.
Cavazos-Rehg PA, Krauss M, Fisher SL, Salyer P, Grucza RA, Bierut LJ. 2015.
Twitter chatter about marijuana. J Adolesc Health. 56(2):139–145.
Centers for Disease Control and Prevention, 1999. Ten great public health
achievements—United States, 1900–1999. Morb Mortal Wkly Rep.
48(12):241–243.
17
Chew C, Eysenbach G. 2010. Pandemics in the age of twitter: content analysis
of tweets during the 2009 H1N1 outbreak. PLoS One. 5(11):e14118.
Davis MA, Zheng K, Liu Y, Levy H. 2017. Public response to Obamacare on
Twitter. J Med Internet Res. 19(5):e167.
Dodds PS, Harris KD, Kloumann IM, Bliss CA, Danforth CM. 2011. Temporal
patterns of happiness and information in a global social network: hedono-
metrics and Twitter. PLoS One. 6(12):e26752.
Eysenbach G. 2009. Infodemiology and infoveillance: framework for an emerg-
ing set of public health informatics methods to analyze search, communica-
tion and publication behavior on the Internet. J Med Internet Res. 11(1):e11.
Eysenbach G. 2011. Infodemiology and infoveillance tracking online health
information and cyberbehavior for public health. Am J Prev Med. 40(5
Suppl 2):S154–S158.
Harris NO, Garcia-Godoy F. 2004. Primary preventive dentistry. In: Neenan
EM, Easley MW, Ruiz M, editors. Water fluoridation. 6th ed. Upper Saddle
River (NJ): Pearson Prentice Hall. p. 181–240.
Iheozor-Ejiofor Z, Worthington HV, Walsh T, O’Malley L, Clarkson JE,
Macey R, Alam R, Tugwell P, Welch V, Glenny AM. 2015. Water fluo-
ridation for the prevention of dental caries. Cochrane Database Syst Rev.
18:CD010856.
Kaminsky LS, Mahoney MC, Leach J, Melius J, Miller MJ. 1990. Fluoride:
benefits and risks of exposure. Crit Rev Oral Biol Med. 1(4):261–281.
Karamitros GA, Kitsos NA, Sapountzis S. 2017. Systematic review of quality
of patient information on phalloplasty in the internet. Aesthetic Plast Surg.
41(6):1426–1434.
King D, Ramirez-Cano D, Greaves F, Vlaev I, Beales S, Darzi A. 2013. Twitter
and the health reforms in the English National Health Service. Health
Policy. 110(2–3):291–297.
Kloumann IM, Danforth CM, Harris KD, Bliss CA, Dodds PS. 2012. Positivity
of the English language. PLoS One. 7(1):e29484.
Lewis CW, Milgrom P. 2001. Fluoride. Pediatr Rev. 24(10):327–336.
Moorhead SA, Hazlett DE, Harrison L, Carroll JK, Irwin A, Hoving C. 2013.
A new dimension of health care: systematic review of the uses, benefits,
and limitations of social media for health communication. J Med Internet
Res. 15(4):e85.
Oh HJ, Oh HW, Lee DW, Kim CH, Ahn JY, Kim Y, Shin HB, Kim CY, Park
SH, Jeon JG. 2017. Chronologic trends in studies on fluoride mechanisms
of action. J Dent Res. 96(12):1353–1360.
Peckham S, Awofeso N. 2014. Water fluoridation: a critical review of the phys-
iological effects of ingested fluoride as a public health intervention. Sci
World J. 2014:293019.