Discussion on Identified Research Gaps and

Possible Objectives
Nilesh Torwane
Ph.D. Student
School of Dentistry, UQ

CRICOS code 00025B

[Entity Name]

SOCIO-DIGITAL MEDIA ECOSYSTEMS:

BRIDGING THE INEQUALITY GAPS
• Social media can have a direct positive impact on inequality. We can see that it gives many people
many new opportunities.

• Social media gives disadvantaged people a cure against inequality because they can profit from access
to greater resources.

• They claim the internet is a space of egalitarianism, free speech, and democracy.

• Social media has created a public platform for the disadvantaged and underprivileged to raise their
voices.
• It gives individuals as well as communities the possibility to become active in the development
process.
• Social media platforms provide safe spaces for discussion.

Working Remotely | 2020 CRICOS code 00025B 2

[Entity Name]

• On March 23, 2012, a man broke into a hospital in China, fatally stabbing one doctor and injuring three others.
This tragic event was followed by a “public carnival” on the Internet, which included an online poll asking people
how they felt about this news, and 4,018 out of 6,161 (65%) people chose “happy” (People’s Daily Online, 2012).
Why did the public react positively to such tragic news? It makes sense that some netizens chose the happy
response to express their negative perceptions of the doctor-patient relationship in China. Furthermore, the
social media environment, in which official and nonofficial discourses fuse and news valence is sharpened,
intensifies such a mood.

Working Remotely | 2020 CRICOS code 00025B 3

[Entity Name]

Online social platforms like Facebook and Twitter provide millions of individuals with nearly unlimited access to
information and connectivity [8–10].
The content produced on such platforms has proved to impact society at large: from social and political discussions
[11–16], to emergency and disaster response [17–19], social media conversation affects the offline, physical world in
tangible ways.
The central issue that inspires this work is how the content produced and consumed on social media affects
individuals’ emotional states and behaviors.

Data from a 20-years longitudinal study suggest that emotions can be passed via social networks, and have long-
term effects [21].

Working Remotely | 2020 CRICOS code 00025B 4

[Entity Name]

Health and Social Media Relationship

• Over the last two decades, internet users have been increasingly using social media to seek and share health
information [1].
• These social platforms have gained wider participation among health information consumers from all social groups
regardless of gender or age [2].
• Health professionals and organizations are also using this medium to disseminate health-related knowledge on
healthy habits and medical information for disease prevention, as it represents an unprecedented opportunity to
increase health literacy, self-efficacy, and treatment adherence among populations [3-9].
• However, these public tools have also opened the door to unprecedented social and health risks [10,11].

• Although these platforms have demonstrated usefulness for health promotion [7,12], recent studies have suggested
that false or misleading health information may spread more easily than scientific knowledge through social media
[13,14].

• Therefore, it is necessary to understand how health misinformation spreads and how it can affect decision-making
and health behaviors [15].
Working Remotely | 2020 CRICOS code 00025B 5
[Entity Name]

In fact, at present, the propagation of health misinformation through social media has become a major public health
concern [17].

Working Remotely | 2020 CRICOS code 00025B 6

[Entity Name]

Current Knowledge Gaps…

BASED ON THE LITERATURE REVIEW: four knowledge gaps have been identified.
- First, we have to identify the dominant health misinformation trends towards CWF and specifically assess their
prevalence on different social platforms.
- Second, we need to understand the interactive mechanisms and factors that make it possible to progressively
spread health misinformation through social media.

- Factors, such as the sources of misinformation, structure, and dynamics of online communities, idiosyncrasies of
social media channels, content and framing of health messages, and the context in which misinformation is
shared, and health debates on social media are also affected by social bots are critical to understanding the
dynamics of CWF related misinformation through these platforms.

Working Remotely | 2020 CRICOS code 00025B 7

[Entity Name]

• Third, impact of these tools in the dissemination of health misinformation but also their impact on the
development and reproduction of unhealthy or dangerous behaviors.
• Finally, regarding health interventions, we need to know which strategies are the best in fighting and reducing the
negative impact of health misinformation without reducing the inherent communicative potential to propagate
health information with these same tools. In line with the abovementioned gaps, a recent report represents one of
the first steps forward in the comparative study of health misinformation on social media [16].
• Through a systematic review of the literature, this study offers a general characterization of the main topics, areas
of research, methods, and techniques used for the study of health misinformation. However, despite the
commendable effort made to compose a comprehensible image of this highly complex phenomenon, the lack of
objective indicators that make it possible to measure to the problem of health misinformation is still evident today.
Considering this wide set of considerations, this systematic review aimed to specifically address the knowledge gap.
In order to guide future studies in this field of knowledge, our objective was to identify and compare the prevalence of
health misinformation topics on social media platforms, with specific attention paid to the methodological quality of
the studies and the diverse analytical techniques that are being implemented to address this public health concern
(Prevalence of Health Misinformation on social media: Systematic Review)

[Presentation Title] | [Date] CRICOS code 00025B 8

[Entity Name]

Due to global digitization and inexpensive Internet services, various web portals and websites became encyclopedias for healthcare-
related information. More than 4 billion unique users joined the realm of the internet and the numbers are increasing day by day (Barua
et al., 2020). It has been observed that more than 70% of adults utilized various web sources to search healthcare-related information
(Barua et al., 2020). Among various Internet services, social media platforms are used as the main source of information by people
whereby they interact with each other to create, share, and exchange information among their virtual communities (Niknam et al., 2020).
Facebook, Twitter, and Instagram are the social media giants. An unrestrained usage and access of social media platforms made them a
discussion portal and healthcare are an emerging topic among several other issues (Ghenai & Mejova, 2018). Before consulting a
physician, 44% of people often search about the cure of an illness on the internet and change their decision about the treatment (Ghenai,
2017). However, the quality of information and content posting intentions were always been questioned and must be treated as a
potential threat to society (Ghenai & Mejova, 2018). Moreover, the influence of propagated healthcare misinformation on people often
leads them to harmful health decisions (Ghenai, 2017).
Therefore, usage of social media and people's reliance on available content originates a problem of health misinformation dissemination.
Misinformation is the piece of content that fulfills the deceiving intention by spreading misleading information. Misinformation can be
categorized in many forms such as hoaxes, rumors, fake news, fake reviews, and false facts (Vyas & El-Gayar, 2020). These categories keep
on increasing. For example, in the current pandemic situation, “Infodemic” emerged as a new category that comprises the usage of
unreliable information, especially for healthcare. Hence, healthcare professionals are often facing the problem to eliminate non-credible
information from all official and non-official sources (Datta et al., 2020). Furthermore, in recent years, the proliferation of health
misinformation gauged the attention of researchers. Abul-Fottouh
Working Remotely | 2020 CRICOS code 00025B 9
[Entity Name]

Machine Learning
• Machine learning methods for emotion classification originated from Natural Language Processing research in
Computer Science.25
• Among them, the so-called deep learning models or neural networks have the advantage of being able to
consider not only word frequencies but also information such as word order and other features of the context.
• The usual approach for emotion classification in machine learning relies on supervised methods, which require
datasets of annotated texts with emotion labels for model training. These text labels are referred to as “ground
truth”, and try to capture how humans would most likely interpret or express emotions in text.
• To train machine learning models with such a dataset, the texts contained in it need to be transformed to a
numerical representation. This can be done through word embeddings or be constructed from unweighted or
weighted frequencies for single words or short sequences of words (n-grams), or from index positions of words in
vocabulary lists.

Working Remotely | 2020 CRICOS code 00025B 10

[Entity Name]

people’s emotions, or reflect on emotions they experienced recently. Yet, for research questions about collective
emotional states, or the emotions of populations, talking about the emotions of others may actually contribute
valuable information about users not active on the specific social media platform.
Social media emotion measures correlate with self-reported emotions and life satisfaction at the population level.
These studies provide some evidence that certain social media measures can be valid indicators for emotional
trends and well-being in societies at large.

Working Remotely | 2020 CRICOS code 00025B 11

[Entity Name]

RESEARCH RELATED TO SENTIMENTAL ANALYSIS IN A

HEALTH SECTOR
Machine learning techniques offer the opportunity to discover hidden patterns and relationships in the data (Christozov & Mitreva, 2020). Therefore, as a
potential problem, researchers have adopted the classification of user-generated content as health misinformation or genuine content.
Niknam et al. (2020) performed a thematic analysis on Instagram posts by extracting user-search tags (i.e., upload date, media type) to characterize the
public health information as misinformation.
Ghenai & Mejova (2018) classified the misinformation related to fake cancer cures on Twitter by utilizing the logistic regression technique. They have
extracted user-centric, tweet-centric, readability-centric, and medical-centric features. (See table 1).
Hou et al. (2019) adopted viewer-centric and linguistic features to automatically detect prostate cancer-related misinformation on YouTube by utilizing the
support vector machine classification technique.
Silva et al. (2020) have performed a classification task by utilizing machine learning techniques to investigate the relationship between users and
misinformation through tweet-related, engagement-related, and user-related features.
Sear et al. (2020) performed the topic modeling to extract the frequent health misinformation-related themes in Facebook posts by Latent Dirichlet
allocation (LDA) techniques through calculating coherence and subjectivity scores.
J. Li et al. (2020) utilized the ML-based exploratory analysis to classify the health-related false facts on Snopes.com and Poynter.com using fact-checking
reviews, source content, social engagement, and user reply features.
Cui et al. (2020) adopted the linguistic features to detect health fake news through a graph attention neural network.
Further, Cui et al. (2020) used a gated recurrent unit (GRU) to detect health misinformation on social media platforms by using user engagement features.

Working Remotely | 2020 CRICOS code 00025B 12

[Entity Name]

Working Remotely | 2020 CRICOS code 00025B 13

[Entity Name]

SNSs…
• recent years, researchers have analyzed massive amounts of data produced by SNSs such as Twitter to
glean insights into topics of interest to public health.

• Their outputs are increasingly being considered valuable sources of health information.

• For example, Twitter data have been analysed to study natural disasters (Black et al. 2015), infectious
disease outbreaks drug and alcohol, and public responses to health policies. Such online information,
however, has been barely exploited by dental researchers.

• I intend to use English-language Twitter data to probe public response to a specific topic in dentistry – water
fluoridation.

• 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 qualitatively.

Working Remotely | 2020 CRICOS code 00025B 14

[Entity Name]

Research Gaps in a brief-

SOCIAL INEQUALITY AND CWF-
1. Worldwide and in Australia; the phenomenon to establish fluoridation as a marker to reduce social
inequalities pertaining to oral health hasn’t been explored sufficiently to draw some conclusive
inferences.
2. In spite of having decades of experience in the implementation of CWF for public health benefits; the
Qualitative aspects related CWF are hardly been focused on from a research point of view.
Moreover, the use of social platforms for qualitative assessment has been completely neglected by
scientific societies which could provide extremely important data by looking at the strong influence of
social media ecosystems on public lives when it comes to decision making.
3. There’s been less scientific work has been performed and almost all of that work is outdated with
regards to scientific evaluation of the quality of online information available in relation to CWF.

Working Remotely | 2020 CRICOS code 00025B 15

[Entity Name]

Research Gaps in Sentimental Analysis of CWF

Authors Time period Social Keywords Feature Geospatial Topic Sentiment Locality Country Network
Media Analysis Modeling Analysis Analysis Models
H.J. Oh From 2009 to Twitter     
et al. 20017
Aaron At a point of Twitter/   
M. et al. time FB/
Website
Michael At a point of Twitter  
Mackert time

Propos From 2017 to         

ed 2022/23 (for
study twitter +
Instagram0

Cross Sec
Analysis- FB,
Youtube,
Websites

Working Remotely | 2020 CRICOS code 00025B 16

[Entity Name]

Possible Objectives Could Be -

1. To identify possible social inequalities pertaining to oral health by considering CWF as a marker among
the school children of  Queensland State, Australia.

2. To perform a comprehensive thematic analysis and visualize public sentiments and the spread of
information towards CWF on various socio-digital media ecosystems (Twitter, FB, Instagram, and
YouTube) worldwide and in Australia by using advanced qualitative machine learning
algorithms/techniques. 

3. To explore the network group analysis to track anti-fluoride activities performed by specific groups (if
any).

4. To evaluate the quality of online content on CWF available on various search engines available
worldwide and in Australia by applying data analytic techniques.

Working Remotely | 2020 CRICOS code 00025B 17

[Entity Name]

Expected results (but not limited to)- 

• Draw inferences to establish the relationship between Water Fluoridation and Social inequalities (If
there’s Any)
• would be able to compare the sentiments of water fluoridation blocked countries of the world with the
fluoridated ones) and get an overall picture of sentiments across the globe and in an Australian
perspective.
• Sentiment comparison between Fluoridated (favoring), non fluoridated,(neutral) and fluoridation
opposing and fluoridation ceased (negative) perceptions worldwide
(Country-wise + Comparison across Australian states)
• Quality Analysis of (websites, blogs, newspaper, etc.) CWF-related information – evidence-based vs Fake
material (world overview and Australian overview)

Working Remotely | 2020 CRICOS code 00025B 18

[Entity Name]

“THE IMPACT OF COMMUNITY WATER FLUORIDATION AS A MARKER FOR SOCIAL INEQUALITIES ALONG
WITH QUALITATIVE ESTIMATION OF PUBLIC SENTIMENTS AND STANDARD OF INFORMATION AVAILABLE
IN RELATION TO FLUORIDATION ON SOCIO-DIGITAL ECOSYSTEMS WORLDWIDE AND IN AUSTRALIA BY
USING MACHINE LEARNING ALGORITHMS”

Working Remotely | 2020 CRICOS code 00025B 19

[Entity Name]

Working Remotely | 2020 CRICOS code 00025B 20

[Entity Name]

Working Remotely | 2020 CRICOS code 00025B 21

[Entity Name]

Working Remotely | 2020 CRICOS code 00025B 22

[Entity Name]

Working Remotely | 2020 CRICOS code 00025B 23

[Entity Name]

Working Remotely | 2020 CRICOS code 00025B 24

[Entity Name]

Working Remotely | 2020 CRICOS code 00025B 25

[Entity Name]

THANK YOU

Working Remotely | 2020 CRICOS code 00025B 26