Social networks applied to Dengue,H1N1,and Zika epidemics: An integrative literature review

Abstract

BACKGROUND:

Health crises occur both regionally and globally. Online social networks are widely used technical resources that allow users to share large amounts of information with increasing reach and velocity. Thus, the capacity of spreading information about epidemics through social media allows members of a population and health professionals or agencies to collaborate.

METHOD:

This study presents results obtained in an integrative review, including examples of how social media enabled collaboration in health surveillance to treat the epidemies of Dengue, Zika, and H1N1. The literature review covers studies published between 2009 and 2017.

RESULTS:

The studies reviewed indicate that social media interactions are tools for the rapid dissemination of information. These networks operate at low cost and allow information to reach audiences in need of information and who otherwise would not receive it. Social media allowed researchers to monitor evolving epidemics and obtain epidemiological data useful for decision-making in health surveillance.

CONCLUSIONS:

Despite the widespread use of social networks, there are opportunities for improvement, especially in technology for treatment.

Keywords

Social media collaboration in health Dengue Zika H1N1

1 Introduction

In recent decades, the world has suffered a variety of public health epidemics, including H1N1 in 2009, Ebola in 2013 and 2016, Zika in 2015, and Dengue in 2016. These epidemics have often affected large populations; in some cases, these crises led to large-scale mortality. Ebola killed about 11,000 people in recent years, while the H1N1 flu epidemic led to the death of approximately 8,000 to 18,000 people. In some cases, these epidemics are ongoing - Dengue, for example, infects as many as 400 million people annually. During ongoing epidemics, the affected populations lack information on the diseases, the geographical location(s) affected, the risks to those impacted, and methods of prevention [1]. Epidemics such as these do not occur as frequently as floods and hurricanes; thus, information on the response is partial and conflicting, and the population obtains information via social media. The consequence for society is that people use this often ambiguous infor-mation to decide about travel, whether to seek treatment and what to do for prevention [2].

Social media represents a change in how information is accessed and shared, allowing the rapid sharing of text-based content and videos. Users can sort news and information, exchange opinions, and provide recommendations. Thus, social networks allow people with similar characteristics or common goals to connect and collaborate on common interests. In the age of social networking, studies about the use of these networks for the sharing of public health information and local events increasingly appear in literature Social media information can be used to understand how local populations view health problems or even to predict, detect, and track disease outbreaks [3].

Online communication through social networks like Facebook, Instagram, and Twitter has created new ways for people to share information about eme-rgencies, especially those of a public health nature [4]. Traditionally, public health agencies have transmitted information to the public primarily through print media, television, and radio. These forms of media still play a significant role both in intensifying and attenuating risk within a population. However, with regards to the public health H1N1 emergency in 2009, Twitter was cited as the most frequently-used source of information on the pandemic by respondents to a population-based survey cited [5].

Public health professionals can also use health data obtained from social media to detect potential risks of epidemics. It has been shown that by performing qualitative and quantitative analysis of data from social media, it is possible to draw a picture of the be-havior and the opinion dynamics of a population [5]. With the information available, health officials can establish a cycle of monitoring and public response online, identifying emergencies, and using these net-works’ communication strategies for adaptation of educational campaigns. This transition has gone be-yond public health emergency settings. Public he-alth officials recognize the value of social media. Research conducted in the US on the extent of social media use in public health more generally has shown that 60% of local health departments used social media to disseminate information. The most common platform for that dissemination was Twitter [6], a social network and server for microblogging texts of up to 140 characters.

In this article, we explore how social media platforms are being used to promote the collaborative participation of its members, researchers, and health professionals to generate relevant information and expand communication on epidemiological health emergencies. We aim to establish how data on social networks are used to help identify, track, and generate useful information on disease outbreaks, with a focus on recent H1N1, Dengue, and Zika virus epidemics. We chose these diseases due to their high dissemination capacity and the potential for significant global health damage. This review will offer information about the use of social networks in collaboration and monitoring to prevent Dengue, H1N1, and Zika epidemics.

2 Background

In modern society, the web has become an ess-ential tool for connecting people, spread clear he-alth information allowing new forms of interaction between users that sharing knowledge and interests. These relationships create networks that disseminate information of common interest [7]. Among this information, we highlight themes associated with public health, especially in crises such as epidemics. Sabeeh et al. [8] point out that techniques and methods that allow collaboration are necessary to meet population needs in healthcare, including by strengthening the use of technological solutions to share peer knowledge.

In 1963, Alexander Langmuir conceptualized he-alth surveillance as the continuous observation of the distribution and trends of the incidence of diseases through the systematic collection, consolidation, and evaluation of reports of morbidity and mortality, as well as other relevant data, and to the regular dissemination of such information to all who need to know it. Public health professionals and researchers need access to health information, educational status, and training tools, especially to respond to emergencies [9]. Social networks provide access to millions of records, an unprecedented opportunity for large-scale inference and information dissemination among the general population. Thus, new monitoring and reporting strategies for health applications can be developed, especially in health surveillance [10].

The research studies reviewed herein focus on the epidemiological contexts of recent outbreaks of Influenza A (H1N1), Dengue, and Zika virus. Human influenza is an infectious viral disease that is highly communicable, of global reach, and which goes through seasonal epidemic patterns. Small frequent changes in common antigens (drifts) make individuals infected before only partially protected from re-infection, which facilitates the circulation of the virus annually. Influenza type A (H1N1) had not been identified as a cause of infections in people before the pandemic of 2009. Genetic analyses of this virus have shown that it originated from animal influenza viruses and is unrelated to the human seasonal viruses that have been in circulation since 1977. Unlike typical seasonal flu patterns, however, H1N1 caused high levels of summer infections in the Northern Hemisphere, and activity increased even further during the colder months.

Influenza H1N1 was the first influenza pandemic to happen in the globalized modern world. Rapid and intense traffic of people and goods has a significant influence on the spread of the virus among countries. However, also due to globalization, information about this epidemic was shared and transmitted in real-time. The prevalence of web-based social communication introduced new tools for follow-up and monitoring through collaboration between researchers, public health officials, and the general population [11]. Research on this last outbreak of the H1N1 virus revealed that Twitter was the social network that collaborated the most. As a result of the tremendous overall use of Twitter, researchers could measure the real-time perceptions of the population on the situation and their interactions with official agencies and health professionals.

Dengue fever is a debilitating mosquito-borne in-fection caused by any one of four closely-related Dengue viruses. If untreated, Dengue can lead to potentially lethal complications, including Dengue hemorrhagic fever. In fact, despite massive media propaganda and costly household control measures, governments have been unable to reduce the prevalence of mosquito-based epidemics via information transmission alone. According to the World Health Organization (WHO), Dengue is currently the tropical disease that spreads most quickly around the world. WHO considers the infection to be already endemic in more than 125 countries, and it has the potential to become a worldwide epidemic.

Zika virus, the third disease covered in this review, is a flavivirus transmitted through the bite of the Aedes Aegypti mosquito. This disease causes moderate fever and skin rashes during a period of 2 to 7 days. However, it is estimated that 80% of infected people have no symptoms [26]. Although symptoms have a short duration, this disease has a high potential for dissemination and may cause significant health complications, including Guillain-Barre syndrome, an autoimmune disease targeting the peripheral nervous system causing weakness, numbness and even paralysis. Zika can infect people of any age and has no identified cure, but can be particularly devastating for pregnant women. This infection can cause fetal brain malformation and microcephaly –when babies are born with a smaller than average head circumference, which is usually larger than 32 cm, causing brain damage, rare and irreversible birth defects, or even fetal death [12].

3 Methodology

We assembled relevant studies using integrative review, a broad method of review, allowing the inclusion of theoretical and empirical literature and also studies with a quantitative and qualitative methodological approach. Using the methodology of Cooper [13], phases of each study were assessed and de-lineated, including the definition of the research problem, data collection methods, and verification, analysis, and interpretation of the data. We proceeded through the following steps:

Theme identification, selection of research hypothesis or research question, and keywords selection;

Inclusion and exclusion criteria definition, database selection, and literature search;

Categorization of studies via information extraction, organization and summarization to form a research database;

Evaluation of studies included in the review and critical analysis of selected studies;

Interpretation of results and discussion of results for future research;

Knowledge synthesis and detailed description of the review.

Ganong recommends the use of criteria for the inclusion and exclusion of work in integrative reviews [14]. Thus, we only included papers that had been published and made available in full, which encompassed scientific reports (such as research articles, technical reports, and project analyses), working documents, government documents, and evaluations supported by Grey Literature. Moreover, we only considered reports published from 2009 onwards, in either Portuguese or English, relevant to the themes of health or informatics/information technology, and focused on the application of social networks and collaborative cases for health surveillance. Studies that dealt with collaboration in health surveillance but did not address H1N1, Dengue, or Zika or that did not evaluate the utilization of social media were excluded.

For the search, we selected the databases Science Direct (www-sciencedirect-com.web.bisu.edu.cn); Springer (www.springerlink.com); Scielo (www.scielo.org); and IEEE Xplore (www.ieeexplore.ieee.org). We se-arched for articles based on Grey Literature in re-search institutions and specialized health agencies (WHO, Fiocruz, and others) with studies related to occupational diseases.

The terms we searched for in the databases were: “Social Media”, “Collaboration in Health”, “Health Surveillance”, “Dengue”, “Zika”, “H1N1”, “Twitter”, “Facebook”, and “Instagram”. There were no controlled descriptors for a broader search. After the first step, terms were crossed using conjunctions and varied by alternating diseases, for instance (H1N1) AND EPIDEMICS AND (SOCIAL MEDIA OR SOCIAL NETWORK AND DATA) AND HEA-LTH AND EMERGENCY; (DENGUE) AND EPIDEMICS AND (SOCIAL MEDIA OR SOCIAL NETWORK AND DATA) AND HEALTH AND EMERGENCY; (ZIKA) AND EPIDEMICS AND (SOCIAL MEDIA OR SOCIAL NETWORK AND DATA) AND HEALTH AND EMERGENCY.

4 Results

Database searches were initially carried out separately by disease. In this initial phase, we found 3,291 articles, out of which 1,145 articles were on Dengue disease (34.80%), 2,131 on H1N1 (64.75%), and 15 on Zika (0.45%).

In the second phase, we applied the selection criteria. We only included studies in Portuguese or English, related to the theme, and published since 2009. We selected 449 studies (13.64% of the total 3,291), 174 of which were about Dengue, 260 about H1N1, and 15 about Zika. Then, we proceeded to read the studies’ titles and abstracts. Due to their relative scarcity, articles related to Zika were all read in their entirety.

Out of the 499 studies, we selected 21 for complete analysis via integrative review. Only nine were relevant for the discussion on the theme. Figure 1 presents the flowchart used in the selection process of the articles included in the integrative review, and Table 1 shows a summary of these nine studies.

Fig. 1

Flowchart related to the process of the selection of articles.

Table 1

Summary of the work

Research base	Title	Abstract	Kind of study	Type of evidence
Scielo	Visual archives concerning	Visual files posted on the Instagram network, concerning the	Case study	4
	Zika virus: images on Instagram	Zika virus epidemic were studied by Antunes et.al (2016),
	as part of the constitution	helped to build a visual memory of the epidemic and concluded
	of an epidemic memory [3].	that the feeling of Internet users can provide solutions for care and prevention.
PubMed	Pandemics in the age of twitter:	Via Twitter, understand public perceptions in relation	Case study	4
	content analysis of tweets during	to the pandemic of influenza H1N1. The authors concluded that
	the 2009 H1N1 outbreak [9].	Twitter is potentially suitable for text mining and longitudinal analysis.
PubMed	Dengue prediction by the web:	This work showed that the association of tweets with dengue cases	Quantitative study	4
	Tweets are a useful tool for estimating	can help the traditional surveillance in real time and at low cost,
	and forecasting Dengue at	for being able to successfully transmit a dengue outbreak and predict
	country and city level. [29].	the disease in up to 8 weeks in the future in comparison with other
	available web-based data, Google Trends, and Wikipedia access logs.
PubMed	Health Tweets: An Exploration	The study showed the Twitter was used as a communication modality	Qualitative study	4
	of Health Promotion on	to update the public about waiting times at influenza vaccination sites
	Twitter [13].	and for distribution of government alerts, during the H1N1 pandemic in Canada.
IEEE Xplore	Dengue surveillance based	This paper showed the potential of Twitter data for dengue surveillance,	Experiment	4
	on a computational model	utilizing methodology based on four dimensions: volume, location,
	of spatio-temporal locality	time and content. The extent of public perception was explored by
	of Twitter [19].	performing sentiment analysis that allowed filtering
		content not relevant for dengue surveillance.
		After verifying, the high correlation between
		the number of cases reported by the official statistics
		and the number of tweets posted during the same period.
Fiocruz	Manual of Social Media	Manual of the institution’s official communication	Grey literature
	Fundação Oswaldo Cruz [15].	standards for social networks.
Fiocruz	Fiocruz Dengue Network [14].	Dynamics of disease control through an inter sectoral	Grey literature
		and multidisciplinary action of surveillance,
		prevention, and health promotion in the areas
		of environmental control, communication and
		information, management, social mobilization,
		laboratory services, reference attention, education
		and research, available through a website.
Science Direct	VazaDengue: An information system for	This paper discusses the limitations of existing alternatives,	Qualitative study	4
	preventing and combating mosquito-borne	as well as describes a technological solution (VazaDengue)
	diseases with social networks [39].	for notification and visualization in real-time incidence
		for notification and visualization in real-time incidence of
		Dengue outbreaks in geographic regions. Through of
		data mining of social networks, the system shows
		the situation of foci and other information correlated to the mosquito.
Science Direct	Zika virus Pandemic-analysis	Study about information source for the Zika virus	Case study	4
	of Facebook as a social media	pandemic in the United States; found that the misleading
	health information platform [37].	posts were far more popular than the real ones,
		dispersing accurate and relevant public health.

The studies selected in the second phase are distributed by research base as follows: 48 articles were obtained from PubMed, of which three were selected according to the criterion and 45 excluded; IEEE Xplore yielded four articles in total, of which none were excluded; ScienceDirect revealed 72 articles in total, of which six were selected and 66 excluded; Springer returned 323 articles in total, of which five were selected and 318 excluded; and the SciELO database had two relevant articles, from which one was selected and one excluded. Using the Grey Literature, two works were found and selected.

4.1 Works related to H1N1 influenza

According to the literature, nearly all humans have had contact with the H1N1 virus since the Spanish flu epidemic of 1918. The contact has been either by third or fourth generation infections. By swapping and changing its genes, H1N1 has managed to continue circulating for more than a century [15, 16]. However, H1N1 prevalence saw its apex in 2009 with a global outbreak that, according to the World Health Organization, caused the death of more than 18,000 people.

The Influenza A virus global pandemic (H1N1) initially emerged in Mexico in early April 2009. It soon spread rapidly around the world, putting the global public health community on alert. In the second half of April of that year, the World Health Organization (WHO) issued its first alert about the new epidemic.

At that moment, social networking tools helped to predict the behavior of the population and signaled appropriate actions for health surveillance. Donelle and Booth demonstrated the efficacy of Twitter during the 2009 H1N1 pandemic in their study “Health Tweets: An Exploration of Health Promotion on Twitter” [17]. Twitter was used to update the public on waiting times for influenza vaccination and the distribution of government alerts. Their study demonstrated that determinants of overall health are widely discussed by users of the technology of social networks like Twitter, showing that such platforms can play a fundamental role in the dissemination of information.

Eysenbach and Chew presented another example of the use of social media in health in their study “Pandemics in the age of Twitter: the content analysis of tweets during the 2009 H1N1 outbreak” [5]. The authors describe how they were able to assess public perceptions of the H1N1 pandemic via Twitter. There was a sharp increase of messages on Twitter with the term H1N1; the authors related these mentions to the volume of news on the subject. They concluded that Twitter is potentially suitable for text mining and longitudinal analysis.

Researchers have shown that for public health educators, Twitter and other social networks may serve as a good indicator of seasonal H1N1 epidemics –in particular, factors that influence the emergence and spread of the disease. These platforms play a critical role in disseminating information only if communication is done strategically, through reliable processes, and via high-yield avenues (through a dense network of consistent and real-time followers) [18].

4.2 Works related to the Zika virus

The Zika virus was first detected in Brazil in May 2015. In November of that year, the Brazilian Ministry of Health confirmed the probable relationship between Zika and microcephaly. By January 2016, the Ministry of Health had investigated 3,448 cases of microcephaly in the country and its possible relationship to prenatal Zika virus infection. The WHO then called the Emergency Committee to Regulate International Health to analyze the Zika virus and its dissemination and to determine the consequences for public health in the global context. As a result, strategies were defined, calling for research to clarify aspects of the association between the Zika virus and microcephaly. Data from the Pan-American Health Organization (PAHO), considered a regional structure of the WHO, 3–4 million cases of Zika were confirmed in 2016 on the American continent alone [19].

Health surveillance was known to play a vital role in improving the understanding of the disease in real-time, including its potential epidemiological and clinical effects, and the circulation dynamics of the virus. In this context, social networks were vital. Al-though these networks could replace in-person health education, social networks could be used to identify and track health populations [20].

Social network analysis showed that there was a collaboration between community and organization/health professionals in response to the Zika epi-demic. This collaboration was observed on Fiocruz’s social network and in the article “Visual Archives concerning Zika virus: images on Instagram as part of the constitution of an epidemic memory” [21] (sel-ected works found in the systematic review).

Social media are important tools for information dissemination. Facebook has been reported as the largest social network in constant use, serving as a strong platform for providing specific intervention to the use of information in a contemporary way [26]. Fiocruz’s Facebook page was widely used for monitoring and disseminating information about the Zika virus, especially during the epidemic, to promote educational media campaigns. Many visual cues and standardization of records were employed to ensure that the public identified the information as coming from a secure source. Figure 2 illustrates this form of communication on the network.

Fig. 2

Zika virus campaign in the media (Facebook). Source: Highlights Fiocruz Network 2016 [24].

Furthermore, the institution monitored the comments of its members in response to these official announcements. Through this monitoring, Fiocruz was able to identify the evolution of cases and themes about the disease with higher indices of doubt in the general population. Then, based on the feedback, new information was posted to help settle questions, promote guidelines to the public about customer service networks relevant to Zika, advertise jobs for workers to perform clinical exams, and indicate locations that searched for complications like microcephaly.

This network demonstrated that, in addition to the communication professionals of the Fiocruz, resea-rchers and health professionals were vital collaborators. The goal was to ensure the alignment and cor-rectness of content produced in order to educate the population about available services adequately. In the same vein, health professionals and researchers have appropriated the reports and facts brought forward by the public, generating recommendations and communicating new strategies and procedures to hospitals and laboratories.

Another study related to the Zika virus is “Visual archives concerning Zika virus: images on Instagram as part of the constitution of an epidemic memory” [21]. The authors used ImageCloud –software developed by the Laboratory of Studies on Image and Cyberculture (Labic) at the Federal Universi-ty of Espírito Santo, Brazil. The software selected the most-liked posts on Instagram with the hashtag #Zikavirus in November and December 2015 to form an image cloud –an image mosaic.

4.3 Works related to Dengue

Since 2003, Fiocruz instituted the Dengue Network in order to promote Dengue control actions in Brazil. The strategy of disease control was an intersectional and multidisciplinary combination of surveillance, prevention, and health promotion activities in the areas needing environmental control. These activities were coupled with communication and management of information, social mobilization, laboratory services, referral, education, and search strategies.

The Dengue Network integrates three lines of action: local socio-environmental diagnosis, continuing formation of local social agents, and ongoing actions to reduce Dengue. In 2015, Fiocruz restructured the network and started using social networks (Twitter, Facebook, YouTube, and Instagram) as spa-ces of interaction between the scientific community and people during the epidemics of the disease. As a side effect, major urban problems became evident: high demographic concentration, lack of basic sanitation, inadequate housing, and poor-quality education –factors that contribute significantly to the spread of Dengue. Citizens on Facebook and Twitter helped to map Aedes Aegypti’s focus sites so that authorities could carry out actions in response [22].

In their study, Marques et al. analyzed “the potential of Twitter data for estimating and forecasting De-ngue cases.” They showed that tweets related to De-ngue cases could successfully nowcast (estimate Dengue in the present week) and forecast (predict Dengue up to 8 weeks in the future) at country and city levels with high estimation accuracy [23]. The authors quantitatively evaluated the usefulness of the data acquired from Twitter for the early detection and weekly monitoring of Dengue epidemics. They also compared the use of tweets with other available web-based data, Google Trends, and Wikipedia access logs.

Also, Chen et al. showed that models built on the fraction of Google search volume for Dengue-related queries were able to adequately estimate true activity according to official data reported by national ministries of health or the WHO during the time analyzed [1].

The populations of many different countries have been affected by Dengue. Because of this, commun-ities can play an essential role in combating and pre-venting mosquito-borne diseases. The article “VazaDengue: An information system for preventing and combating mosquito-borne diseases with social networks” presents a collaborative computerized system called VazaDengue. VazaDengue offers users web and mobile platform on which they can help prevent and combat mosquito-borne diseases. Citizens can report mosquito breeding and Dengue cases to each other, and the reports are made available to the community and health agencies. The system also was designed to proactively monitor social media networks like Twitter to enrich the information provided [24].

5 Discussion

An integrative review of relevant literature demonstrated that social media platforms are essential tools for the rapid dissemination of information at low cost. These platforms reach audiences in need of accurate information, especially in times of public health emergencies. Besides, they have proved useful to professionals and health departments in monitoring and generating information on epidemics. In many cases, social networks can become platforms for promoting health education campaigns to the general public, both for general health and in response to specific needs or disease outbreaks.

The usefulness of social media in promoting public health arises, among other reasons, from the ability of users to have many close contacts, social influence, and access to resources that bind people together. With this, these media become important in disease prevention, detection, and treatment.

There are limitations to these studies. Although it is clear that people use social networks to communicate health problems, there are relatively few high-quality studies on how to best use social networks to promote public health. Additionally, the studies presented here raise issues of data reliability, objectivity, and relevance, despite their de-monstrated importance in health. For instance, social networks can amplify rumors on the gravity of a situation, potentially inducing people to make wrong decisions or even causing panic. The misunderstanding is mostly due to the user’s inability to verify the source of information obtained from person-to-person [25] –a social network issue that still needs improvements.

Completeness and clarity of information is an-other issue to improve. When information is spread over multiple posts, the overall message can become ambiguous. Thus, the more precise the subject in the message(s), the higher the chances of adequate understanding and clarity of the data by the population.

The dynamic aspect of networks is also a challenge for epidemiological scenarios. Constant changes oc-cur in public health settings. Therefore, postings need frequent monitoring, and responses should be quick enough to generate the desired mass effect and be used for real-time health surveillance.

Several of the papers we analyzed dealt with the theme of collaboration in social networks for health goals in the cases of H1N1, Zika, and Dengue epidemics. The main aspects of these studies are summarized in Table 2.

Table 2
Summary of the works analyzed

Title Main purpose Social media Results Disease Challenge

Visual archives concerning Zika virus: images on Instagram as part of the constitution of an epidemic memory [3]. Pandemics in the age of twitter: content analysis of tweets during the 2009 H1N1 outbreak [9]. The use of ImageCloud, a tool developed by Laboratory for studies of Image and Cyberculture, from Federal University of Spirit Santo-Brazil to visualize a lot of images in databases like, for example, on social networks. Via Twitter, understand public perceptions in relation to the pandemic of influenza H1N1. Instagram To understand the images of the epidemic and its debate on the Internet that had a greater number of ‘tanned’, which represent quantitatively the number of hits. Images collected on Instagram from the #zikavirus hashtag formed an “image cloud” (or “cloud of images”) according to their importance and redundancy. Zika The difficulty of obtaining a single conclusion from the discourse analyzed research period. The result could and may be different, if another period were chosen, another vehicle, other textualities and other experiences.

Manual of Social Media Fundação Oswaldo Cruz [15]. Provide information and health surveillance of Zika using social networks through the collaboration of users of the network and the institution. Facebook, Instagram, Twitter The institution can monitor the evolution and new cases of the disease, as well as health professionals and researchers, have appropriated the reports and facts brought in by the public generating reports to communicate new strategies and procedures to hospitals and laboratories. Zika Raise awareness of users of the correct use of the communication channel, avoiding rumors and fake news.

Fiocruz Dengue Network [14]. Combat dengue disease caused by the Aedes Aegypti mosquito. Facebook, Instagram, Twitter Dengue transmission reduction; Information on advances in research and new discoveries in relation to the disease, as well as new modalities of combat to the mosquito. Dengue To promote actions of health surveillance and collaboration between health professionals and the population.

Dengue prediction by the web: Tweets are a useful tool for estimating and forecasting Dengue at country and city level [29]. Quantitatively assess the usefulness of data acquired by Twitter for the early detection and monitoring of Dengue Epidemics. Twitter (Wikipedia, Google Trends and sociodemographic data for comparison) Tweets with Dengue content were used to estimate weekly Dengue cases occurrence at the country level. Dengue Improving Dengue surveillance in a cost-effective way.

Health Tweets: An Exploration of Health Promotion on Twitter [13]. A 24-hour cross-section of tweets containing the word or hashtag ‘health’ for analysis, revealing predominant themes of health services, personal health practices, and education. TwitterA cross-section of health-related Twitter® messages captured over a 24-hour period, using a DOH framework (Public Health Agency of Canada, 2003). H1N1, but health services were the focus This work demonstrates the fluid and dynamic linkages between mass media, microblogging, and the concept of health requires additional research inquiry.

Dengue surveillance based on a computational model of spatiotemporal locality of Twitter [19]. Analysis of how the Dengue epidemic is reflected on Twitter and the extent to which this information can be used for health surveillance purposes in Brazil. Twitter (Wikipedia, Google Trends and sociodemographic data) The construction of a simple model based on tweets that were able to ‘nowcast’, i.e. estimate disease numbers in the same week, but also ‘forecast’ disease in future weeks. At the country level, tweets are strongly associated Dengue cases, and can estimate present and future Dengue cases until 8 weeks in advance. At the city level, tweets are also useful for estimating Dengue activity. Dengue The accuracy of data can be found to be strongly influenced by socio-economic factors, particularly because it relies on internet searches. Social factors, however, may limit the value of using tweets to examine epidemics within a city.

VazaDengue: An information system for preventing and combating mosquitoborne diseases with social networks [39]. Strengthen the entomology-surveillance of the mosquito that transmits Dengue, Zika, and Chikungunya by providing geolocated reports, represented through dynamic maps. Twitter and Instagram A color scheme to mark the mosquito focus, as well as people sick or with suspected disease. Dengue Extend the classification of Instagram content, including the classification of pictures associated with potentially relevant posts.

Zika virus Pandemic-analysis of Facebook as a social media health information platform [37]. To verify the effective use of the social media site Facebook as an information source for the Zika virus pandemic. Facebook Most popular posts with relevant information on Zika virus pandemic. Zika To develop an effective strategy to deal with the spread of misinformation during times of pandemic.

Title	Main purpose	Social media	Results	Disease	Challenge
Visual archives concerning Zika virus: images on Instagram as part of the constitution of an epidemic memory [3]. Pandemics in the age of twitter: content analysis of tweets during the 2009 H1N1 outbreak [9].	The use of ImageCloud, a tool developed by Laboratory for studies of Image and Cyberculture, from Federal University of Spirit Santo-Brazil to visualize a lot of images in databases like, for example, on social networks. Via Twitter, understand public perceptions in relation to the pandemic of influenza H1N1.	Instagram	To understand the images of the epidemic and its debate on the Internet that had a greater number of ‘tanned’, which represent quantitatively the number of hits. Images collected on Instagram from the #zikavirus hashtag formed an “image cloud” (or “cloud of images”) according to their importance and redundancy.	Zika	The difficulty of obtaining a single conclusion from the discourse analyzed research period. The result could and may be different, if another period were chosen, another vehicle, other textualities and other experiences.
Manual of Social Media Fundação Oswaldo Cruz [15].	Provide information and health surveillance of Zika using social networks through the collaboration of users of the network and the institution.	Facebook, Instagram, Twitter	The institution can monitor the evolution and new cases of the disease, as well as health professionals and researchers, have appropriated the reports and facts brought in by the public generating reports to communicate new strategies and procedures to hospitals and laboratories.	Zika	Raise awareness of users of the correct use of the communication channel, avoiding rumors and fake news.
Fiocruz Dengue Network [14].	Combat dengue disease caused by the Aedes Aegypti mosquito.	Facebook, Instagram, Twitter	Dengue transmission reduction; Information on advances in research and new discoveries in relation to the disease, as well as new modalities of combat to the mosquito.	Dengue	To promote actions of health surveillance and collaboration between health professionals and the population.
Dengue prediction by the web: Tweets are a useful tool for estimating and forecasting Dengue at country and city level [29].	Quantitatively assess the usefulness of data acquired by Twitter for the early detection and monitoring of Dengue Epidemics.	Twitter (Wikipedia, Google Trends and sociodemographic data for comparison)	Tweets with Dengue content were used to estimate weekly Dengue cases occurrence at the country level.	Dengue	Improving Dengue surveillance in a cost-effective way.
Health Tweets: An Exploration of Health Promotion on Twitter [13].	A 24-hour cross-section of tweets containing the word or hashtag ‘health’ for analysis, revealing predominant themes of health services, personal health practices, and education.	TwitterA cross-section of health-related Twitter® messages captured over a 24-hour period, using a DOH framework (Public Health Agency of Canada, 2003).	H1N1, but health services were the focus	This work demonstrates the fluid and dynamic linkages between mass media, microblogging, and the concept of health requires additional research inquiry.
Dengue surveillance based on a computational model of spatiotemporal locality of Twitter [19].	Analysis of how the Dengue epidemic is reflected on Twitter and the extent to which this information can be used for health surveillance purposes in Brazil.	Twitter (Wikipedia, Google Trends and sociodemographic data)	The construction of a simple model based on tweets that were able to ‘nowcast’, i.e. estimate disease numbers in the same week, but also ‘forecast’ disease in future weeks. At the country level, tweets are strongly associated Dengue cases, and can estimate present and future Dengue cases until 8 weeks in advance. At the city level, tweets are also useful for estimating Dengue activity.	Dengue	The accuracy of data can be found to be strongly influenced by socio-economic factors, particularly because it relies on internet searches. Social factors, however, may limit the value of using tweets to examine epidemics within a city.
VazaDengue: An information system for preventing and combating mosquitoborne diseases with social networks [39].	Strengthen the entomology-surveillance of the mosquito that transmits Dengue, Zika, and Chikungunya by providing geolocated reports, represented through dynamic maps.	Twitter and Instagram	A color scheme to mark the mosquito focus, as well as people sick or with suspected disease.	Dengue	Extend the classification of Instagram content, including the classification of pictures associated with potentially relevant posts.
Zika virus Pandemic-analysis of Facebook as a social media health information platform [37].	To verify the effective use of the social media site Facebook as an information source for the Zika virus pandemic.	Facebook	Most popular posts with relevant information on Zika virus pandemic.	Zika	To develop an effective strategy to deal with the spread of misinformation during times of pandemic.

In analyzing the works presented above, we identified Twitter as the leading social network of interest in current literature. Twitter is unique among social media platforms in that it allows many technological approaches to treat data. It also has a broad reach and popularity.

Among the diseases analyzed, Dengue had the highest number of studies, possibly because this epidemic has been known for a long time and has a global reach.

Among the papers we selected, we highlight an article about an “infoveillance” system developed to continuously gather and mine textual information from Twitter via its Application Programming Interface [5]. The system gathered new publicly-available tweets containing keywords of interest and stored them in an internal relational database, including metadata such as username and time. This database served as the primary dataset for many statistical analyses presented elsewhere.

VazaDengue [24] is another example of such a system in that it proactively monitors social media networks like Twitter to enrich the information provided by users. It processes the natural language text from the network to classify tweets according to a set of predefined categories. After classification, relevant tweets are provided to users as geolocated reports. After analyzing the reports, health agents have tended to agree with the relevance of the classified tweets and confirm that the concentration of tweets is likely to be useful for monitoring mosquito manifestation in big cities. The relationship between health agents and society through this social network shows the potential for collaboration between the parties.

Another study [23] presented a system developed by the Observatories of Dengue Lab: a web-service that monitors tweeting activity associated with De-ngue. The method consists of collecting all tweets containing one or several keywords (“Dengue,” “Aedes,” and “Aegypti”). A machine-learning algorithm then identifies and selects only those suggestive of a personal experience with Dengue (i.e., being infected, knowing someone with Dengue, and similar). The automated classifier works based on pre-vious classifications performed by human specialists. Finally, this system uses a model applied to indicate whether tweets are a positive predictor for Dengue cases.

We particularly favor the strategy for information dissemination adopted by Fiocruz during the severe Zika virus epidemic in Brazil in 2012. The strategy to analyze posts on Facebook to inform decisions made by health teams and researchers in the monitoring of Zika was highly effective. Communication protocols were created on Fiocruz’s Facebook page to guarantee the authenticity of the messages, and the collaborative aspect of this network is evident in the information collected to promote the health research advances.

Finally, we highlight an article on Dengue surveillance based on a computational model of the spatio-temporal locality of Twitter [27]. This model created a surveillance methodology based on four dimensions: volume, location, time, and public perception. This analysis enabled the researchers to verify the high correlation between the number of cases reported by social statistics and the number of tweets posted during the same period. This corre-lation suggested a new approach for Dengue surve-illance: a weekly overview of relevant social media activity in each city compared with the weeks before. The most significant benefit of the methodology is the possibility of using it to predict other diseases.

The works highlighted above show the potential of social network data with or without the use of technological support tools to inform monitoring efforts and prediction of outbreaks for Dengue, Zika, and H1N1. The collaborative aspect was crucial to extract the analyses performed in all of them, either through technology or methodologies.

6 Conclusions

New forms of media, especially social media platforms, function as a means to provide timely updates and allowing rapid global public health collaboration. As such, these media fill information gaps often found in traditional disease surveillance.

Collaboration is essential in the application of social networks to health education and surveillance. This two-way communication allows a population-health professionals-population communication dyn-amic. To a certain extent, the selected papers analyzed in the integrative review portrayed the potential of social media collaboration between individual users to help in the prevention or the dissemination of in-formation about epidemic diseases. There are still opportunities for improvement, especially in the tech-nology itself, regarding its ability to inform frameworks, mechanisms, and algorithms of treatment; methods of searching for data relevant to a context; new ways of monitoring information; and compu-tational methods of evaluation of relationships bet-ween members of the network. Efforts are also needed to address problems related to objectivity, dyn-amism, rumor, and other aspects of the quality of the information that is disseminated. These are advancement opportunities for future work. In the case of VazaDengue, for example, a suggested further re-search step is to extend the classification of Instagram content, including the classification of pictures associated with potentially relevant posts, in order to include more posts in the analysis.

Social media-specific limitations need attention. One example is how not everyone who tweets about a disease is ill –they might be just interested or curious about it. Therefore, the detection of valid tweets for surveillance analysis will depend on a sufficient volume of interest to generate signals and compensate for other “noise”.

Overall, the analyzed studies reveal that there is still much to evolve in terms of technology and met-hodology. However, the potential for collaboration on social networks to contribute to health surveillance in cases such as epidemics of Dengue, Zika, and H1N1 is indisputable.

Conflict of interest

None to report.

References

Gui

, Kou

, Pine

, et al. Managing Uncertainty: Using Social Media for Risk Assessment during a Public Health Crisis. In: Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems - CHI ’17. Denver, Colorado, USA: ACM Press. 2017;4520-33. doi:10.1145/3025453.3025891

Villa

. Zika, or the burden of uncertainty. Clin Ter. 2016;7-9. doi: 10.7417/CT.2016.1907

Abdullah

, Wu

. An Epidemic Model for News Spreading on Twitter. In: 2011 IEEE 23rd International Conference on Tools with Artificial Intelligence. Boca Raton, FL, USA: IEEE 2011;163-9. doi:10.1109/ICTAI.2011.33

Chauhan

. Online media use and adoption by hurricane sandy affected fire and police departments. 2014.

Chew

, Eysenbach

. Pandemics in the Age of Twitter: Content Analysis of Tweets during the 2009 H1N1 Outbreak. PLOS ONE 2010;5:e14118. doi: 10.1371/journal.pone.0014118

Hempel

. The use of social media in environmental health research and communication: an evidence review.Vanc Environ Public Health. 2014.

7Roberts

, Callahan

, O’Leary

. Social media: A path to health literacy. Information Services & Use. 37(2), pp. 177-187. doi: 10.3233/ISU-170836.

Sabeeh

, Syed Mustapha

, Mohamad

. Healthcare knowledge sharing among a community of specialized physicians. Cogn Technol Work. 2018;20:105–24. doi: 10.1007/s10111-017-0453-z

Medronho

R de A

. Geoprocessamento e saúde: uma nova abordagem do espaço no processo saúde-doença. In: Geoprocessamento e saúde: uma nova abordagem do espaço no processo saúde-doença. 1995.

10.

Centola

. Social Media and the Science of Health Behavior. Circulation. 2013;127:2135–44. doi: 10.1161/CIRCU-LATIONAHA.112.101816

11.

Al-garadi

, Khan

, Varathan

, et al. Using online social networks to track a pandemic: A systematic review. J Biomed Inform. 2016;62:1–11. doi: 10.1016/j.jbi.2016.05.005

12.

Morens

, Fauci

. Zika Virus Infection and Zika-Associated With Fetal. Microcephaly Gen Update To 2016.

13.

Cooper

. The integrative research review: A systematic approach. Sage Beverly Hills, CA. 1984.

14.

Ganong

. Integrative reviews of nursing research. Res Nurs Health. 1987;10:1–11. doi: 10.1002/nur.4770100103

15.

Morens

, Taubenberger

, Fauci

. The H1N1 Pandemic Influenza Virus: What Next? MBio. 2010;1:e00211–10. doi: 10.1128/mBio.00211-10

16.

Nelson

, Viboud

, Simonsen

, et al. Multiple Reassortment Events in the Evolutionary History of H1N1 Influenza A Virus Since 1918. PLoS Pathog. 2008;4:e1000012. doi: 10.1371/journal.ppat.1000012

17.

Donelle

, Booth

. Health tweets: an exploration of health promotion on twitter. Online J Issues Nurs. 2012;17.

18.

Winston

, Dawn Medlin

, Adriana Romaniello

. An e-patient’s End-user community (EUCY): The value added of social network applications. Comput Hum Behav. 2012;28:951–7. doi: 10.1016/j.chb.2011.12.016

19.

World Health Organization. WHO | World Health Organization - Zika Virus Disease. WHO. http://www.who.int/emergencies/diseases/zika/en/(accessed 20 Dec 2019).

20.

Giustini

, Ali

, Fraser

, et al. Effective uses of social media in public health and medicine: a systematic review of systematic reviews. Online J Public Health Inform 2018;10. doi: 10.5210/ojphi.v10i2.8270

21.

Antunes

, Alves

, Goveia

, et al. Arquivos visuais relacionados ao vírus Zika: imagens no Instagram como parte da constituição de uma memória da epidemia. 2016.

22.

Fiocruz. Dengue. Fiocruz. https://portal.fiocruz.br/dengue (accessed 20 Dec 2019).

23.

Marques-Toledo

C de A

, Degener

, Vinhal

, et al. Dengue prediction by the web: Tweets are a useful tool for estimating and forecasting Dengue at country and city level. PLoS Negl Trop Dis. 2017;11:e0005729. doi: 10.1371/journal.pntd.0005729

24.

Sousa

, de Mello

, Cedrim

, et al. VazaDengue: An information system for preventing and combating mosquito-borne diseases with social networks. Inf Syst. 2018;75:26–42. doi: 10.1016/j.is.2018.02.003

25.

Halse

, Tapia

, Squicciarini

, et al. Tweet Factors Influencing Trust and Usefulness During Both Man-Made and Natural Disasters. In: ISCRAM. 2016.

26.

Inge

, Graham

, McLaughlin

, et al. Evaluating the effectiveness of Facebook to impactthe knowledge of evidence-based employment practices by individuals with traumatic brain injury: A knowledge translation random control study. Work. 2017;58(1):73–81. doi: 10.3233/WOR-172595.

27.

Gomide

, Veloso

, Meira

, et al. Dengue surveillance based on a computational model of spatio-temporal locality of Twitter. In: Proceedings of the 3rd International Web Science Conference on - WebSci ’11. Koblenz, Germany: ACM Press 2011. 1-8. doi: 10.1145/2527031.2527049