The Impact of the COVID-19 "Infodemic" on Well-Being: A Cross-Sectional Study.

INTRODUCTION: The COVID-19 pandemic created a crisis in the world of information and digital literacy. The amount of misinformation surrounding COVID-19 that has circulated through social media (SM) since January 2020 is notably significant and has been linked to rising levels of anxiety and fear amongst SM users. AIM: This study aimed to assess SM practices during COVID-19 and investigated their impact on users' well-being.
METHODS: An online survey was distributed between June 10 and July 31 2020 via different SM platforms in the United Arab Emirates and other Arabic-speaking countries. Adults above 18 years of age who spoke Arabic or English were invited to complete the survey which covered multiple domains, use and practices related to social media platforms and mental health questions, including the WHO-5 Well-Being Index.
RESULTS: Out of 993 participants, 73% were females, 76% were non-Emirati, 91% were university graduates, and 50% were employed in various occupations, of which 20% were health care professionals. Participants indicated that they acquired COVID-19 related information primarily from social media and messaging applications of which WhatsApp was the most used. Most participants reported sharing information after verification. The mean well-being score was 12.6 ±5.6, with 49% of participants reporting poor well-being (WHO-5 score <12.5). Adjusted linear regression showed that Facebook usage was negatively associated with well-being scores. Additionally, high time use was associated with poorer well-being. When adjusting for other factors, including low confidence in information around COVID-19 and poor knowledge overall, SM usage was significantly associated with poorer well-being.
CONCLUSION: The study sheds light on the use of SM during the pandemic and its impact on well-being throughout the novel coronavirus pandemic. Social media practices during emergencies and disasters may impact public well-being. Authorities are advised to step in to minimize the spread of misinformation and more frequent use of social media as it may influence well-being. Public health specialists, information technology and communication experts should collaborate to limit the infodemic effect on communities.

Article Summary

To the best of our knowledge, this research is one of the earliest to investigate social media practices during the pandemic and their possible impact on well-being using the WHO-5 Well-Being Index (WHO-5) score. Results from an online study distributed in the United Arab Emirates and other Arabic speaking countries through SurveyMonkey has revealed a potentially negative relationship between social media use and general well-being, with most participants scoring “poor” well-being. The study findings indicate that people are using social media to access and share information related to COVID-19, claiming that they verify the information before sharing. This study highlights the possible effect of social media on well-being during this pandemic and other future crises. Therefore, global and local health authorities should be better equipped with strategies and interventions to tackle these issues in the future. However, further studies, including control group will be needed to investigate how social media practices directly impact well-being.

Introduction

An outbreak of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) appeared in Wuhan, China, leading to the emergence of a coronavirus disease 2019 (COVID-19) pandemic in 2020 worldwide.1 Regional and local efforts and policies have been influential in curbing the spread of this pandemic in community and health care settings.2,3 The uncertainty associated with COVID-19 has led to excessive global circulation and the spread of messages, information, disinformation and misinformation about the pandemic, its prevention, causes, complications and effects through the internet and social media.4,5 As of the date of the preparation of this manuscript, misinformation and disinformation are still spreading, especially with the appearance of variants. False information typically revolve around the origins of COVID-19, its long-term effects and its influence on different groups such as the elderly and children, as well as prevention measures and available vaccines.5

Social media platforms are communication tools hosted on and accessed via the internet that allow users to selectively communicate with others and acquire information from user-generated content.6 It has become an excellent means for sharing information, including medical and health information.7 Reliance on social media to collect information and news has increased significantly over the last two decades and reached its peak with the emergence of the COVID-19 pandemic.8

It is claimed that COVID-19 is the first true social media infodemic due to the accelerated rate of information and misinformation transmission around the globe, fuelling panic and fear within the global population.5 Consequently, the World Health Organization (WHO) issued a warning statement regarding the infodemic’s impact on people’s digital literacy, its effect on health literacy, and reactions to COVID-19.5 The WHO defined “infodemic” as ”an overabundance of information—some accurate and some not—that makes it hard for people to find trustworthy sources and reliable guidance when they need it, and which involves outbreak response and an influx of information about the disease, its spread and prevention.5 This influx of information also dilutes the public’s trust in official and verifiable sources and generates fear and panic due to unverified rumours and exaggerated claims. It has also promoted xenophobia and racism toward certain ethnic groups such as those of Chinese and East-Asian heritage. As well as having a tremendous psychological impact on general and vulnerable populations such as adolescents and youth overall.5,9–13

The internet, and specifically popular social media platforms such as Twitter, Instagram and Facebook, have been extensively used in the last two decades to analyse, detect, and forecast infections and epidemics, and to predict patterns of human behavior regarding health.4,14–16 Internet use is very much linked to behavior related to health information; users write posts concerning their health on different social media platforms, often seeking information and advice. Information epidemiology, known as Infodemiology, is a domain in health informatics designated as “the science of distribution and determinants of information in an electronic medium with the ultimate aim to inform public health and public policy”.17–19 However, SM’s role in disseminating information has proven beneficial for both society and authorities in managing the response to the pandemic. Social media use, frequency, and eHealth literacy were also associated with people adopting more preventive behaviors.4,20 It was also shown to improve information spread, health literacy levels, and changing perceptions and attitudes.21 Globally, people’s reliance on their traditional support groups has seen a reduction due to them substituting it with digital communication to share and connect with others. However, this was shown to increase anxiety and mental health issues following sudden restrictions in how we socialize with others.12,14,17,22,23 Many studies have pointed to the impact of social media use on mental health.24–27 This link was augmented during the pandemic due to the widespread physical, social distancing and isolation, fear, panic, and socio-political and economic factors induced by the pandemic.28

According to the media dependency theory by Ball-Rokeach, during a critical public disruption, there is a fundamental necessity for information and sense-making by people; in the modern world, mass media is typically best-suited to meet this need.29 During a crisis such as this pandemic, people choose social media platforms to seek information and share their opinions.13,30–33 Moreover, during the pandemic, misuse and political agendas were being pushed through social media.34 People tend to search for information to reduce the anxiety produced by the resulting ambiguity in a crisis event.32,33 Moreover, vulnerable groups are affected mentally mostly by the pandemic and tend to search social media platforms for information, especially Facebook.35,36 Schønning et al found that studies that have looked at positive and negative outcomes of social media use on mental health and wellbeing among adolescents are rare. In another study, the overall use of social media was not reported to have an impact on adolescents’ well-being, however, it was reported that it made them feel better. Social media navigators use the platforms to express their sentiments, opinions, and feelings, which can be a relevant source of data for researching mental health.37 In a recent study, students were found to seek emotional support and improve their social belongingness through social media use.38 Fake news distribution and sharing information reached their peak during the Covid-19 pandemic.39,40 People’s tendency to share “fake news” related to COVID-19 may depend on several factors and may be impacted by social media dependency and information-seeking needs. Understanding why people share fake news became the interest of many researchers worldwide during the pandemic.39 Evidence suggests that symptoms of anxiety, depression and self-reported stress are common psychological reactions to the COVID-19 pandemic and may affect sleep patterns and the general well-being of individuals.41,42

This study aims to explore social media practices and use during the COVID-19 pandemic and investigate the effect of social media on an individual’s well-being. Moreover, this study explores the use of social media to seek knowledge about COVID-19, treatment, and prevention during the pandemic.

Methods

Reporting Guidelines

We followed the CHERRIES checklist for reporting results of web-based surveys45 ().

Design and Study Participants

A cross-sectional study design was used to conduct this research during the COVID-19 outbreak between June 10, 2020 and July 31, 2020. Study participants were adults aged 18 years and above, of both genders who spoke Arabic or English. Participants were invited using a distribution through the researchers’ networks, social media and using snowball sampling techniques to complete the online survey hosted on the SurveyMonkey platform.46 The survey link was shared with participants via email, WhatsApp, Instagram, and Facebook in the UAE and other Arabic-speaking countries. We also invited Arabic speaking subjects who lived in western countries like (Australia, Canada, the United States and Europe).

Additionally, participants who completed the survey were asked to share the survey link with their contacts. The first page of the survey described the research objectives and assured participants of their anonymity and confidentiality. Only those who agreed to participate in the survey could initiate it by clicking “yes I agree to participate”. Ethical approval was obtained from the Abu Dhabi University Institutional Review Board on May 16, 2020 (CHS-20-05-00014).

Data Collection

A questionnaire comprising 25 questions was developed to achieve the study’s objectives. The questions were divided into four main sections: demographics, social media use and practices during the COVID-19 pandemic, impact of COVID-19 on health-related knowledge, and the overall well-being and impact of COVID-19 on mental health. Well-being was measured using the WHO-5 Well-Being Index44 and questions adapted from “Assessing mental health and psychosocial needs and resources toolkit for humanitarian settings” (section B).47 This paper presents data related to the WHO-5 Well-Being Index score and social media use.

WHO-5 Well-Being Index Score

The WHO-5 Well-Being Index (WHO-5) () is a short global rating scale that measures subjective well-being. The WHO-5 questions () include five items: i) “I have felt cheerful and in good spirits”, ii) “I have felt calm and relaxed”, iii) “I have felt active and vigorous”, iv) “I woke up feeling fresh and rested” and v) “My daily life has been filled with things that interest me”. Each item is scored from 0 to 5. The total raw score ranges from 0–25. The score is multiplied by four to achieve a score in percentage (%), with 0 representing the worst imaginable well-being and 100 representing the best imaginable well-being.48 A cut-off score of 12.5 multiplied by four was used to differentiate between good and poor well-being in this study. The WHO Well-Being index is scored based on a cut-off score of 50 with a total raw score ranging from 0–25 that is multiplied by four to give the final score, with 0 representing the worst imaginable well-being and 100 representing the best imaginable well-being.43 We used the cut-off score of 12.5 multiplied by 4 out of 25 to be the score used in this study as used in other literature.49

Social Media Platforms and Messaging Applications

The most commonly used social media platforms and messaging applications globally and in the region were included in this study. These included WhatsApp, Facebook, Snapchat, Twitter, Instagram, LinkedIn, YouTube and Reddit. Whilst WhatsApp, Instagram, and Facebook are all grouped under Facebook, Inc., WhatsApp is a messaging platform. It differs in format from Instagram and Facebook, whereas the latter two function via user posts on an algorithmically served timeline. However, all of these platforms are generally used to share information. Regardless of the social media platform used, information on these tools might be accepted and shared without validation and adequate supervision from authorities on content.20

Previous studies have classified the WhatsApp and Snapchat messaging applications under social media platforms.36,50 Therefore, we decided to include them in our study as also being social media platforms yet refer to them as messaging applications. The questionnaire was developed in English, translated to Arabic, and – back translated to ensure translation accuracy and validity. The questionnaire was piloted among 25 students and health care professionals to ensure face and content validity. The survey was also sent to five public health experts to assess content validity and to ensure that survey questions assessed the constructs being measured. No modifications were made based on the pilot testing. The underlying constructs, “types of social media”, “practices of looking for information”, and “attitudes and beliefs in abilities and knowledge of social media”, showed moderate to high levels of internal consistency with Cronbach’s alpha of 0.744, 0.688 and 0.791 respectively.

The survey was distributed mainly in the UAE and through social media and to other Arabs in other countries. Only one survey entry was allowed from the same IP address to avoid multiple participant entries.

Sampling

The minimum required sample size was calculated as being 384 adults considering a 95% confidence level, a margin error of 5%, and a power of 80%. However, to account for low response rate attrition and sensitivity analysis, we increased our sample size by 60%, making the minimum required sample 615 participants.

Statistical Analysis

Data were extracted in Microsoft Excel and analyzed in Stata version 16.0.48 Qualitative variables were tabulated as frequencies with corresponding percentages (%) and quantitative variables as mean with corresponding standard deviations (±SD). The overall well-being score was computed by summing up WHO-5 questions and further categorized into poor (<12.5) and good well-being (≥12.5). Unpaired t-tests and one-way analysis of variance (ANOVA) tests were applied to compare the mean well-being score with qualitative variables as appropriate. Chi-square or Fisher Exact tests were applied to compare well-being status with qualitative variables as appropriate. The p-value <0.05 was chosen to determine statistical significance. Simple and multiple linear regression was applied to assess the association of well-being score with social media use. In simple linear regression, crude estimates (β) of well-being score with social media use were reported with corresponding 95% confidence intervals (CI) and p-values. In multiple regression, the well-being score with each question related to social media use was adjusted for age, sex, marital status, education, nationality, and job. Upon initial analysis, we found no significant difference in well-being score among participants from different nationalities and countries and no differences in terms of social media practices and use. To ensure comparisons were meaningful and statistically appropriate, the responses “disagree“ and ”strongly disagree” and ‘agree’ and ‘strongly agree’ were merged.

Results

Demographics

A total of 999 responses were collected. Only six surveys were returned incomplete and were excluded from the analysis (Table 1). The majority of study participants were females (73%) with a mean age of 34 ±12.2 years. Most participants had either undergraduate or post-graduate education. Almost three-quarters of the participants were non-Emiratis (76%), with (64%) residing in the UAE. Half of the participants (50%) reported that they were working. Of those employed, the majority were health care professionals, followed by professionals in the education sector and academia.

Table 1

Demographic Characteristics and Comparison with Well-Being Scores

Variable	All	Well-Being Score	P-value*	Good Well-Being (≥12.5)	Poor Well-Being (<12.5)	P-value*
Overall	993	12.6 ±5.6	–	511 (51.5)	481 (48.5)	–
Age, Mean ±SD	34.2 ±12.2	−0.02!	0.581	34.1 ±12.1	34.3 ±12.4	0.813
Age groups, n (%)
<20 years	88 (8.9)	12.8 ±4.8	0.344	42 (47.7)	46 (52.3)	0.602
20–29 years	320 (32.3)	12.6 ±5.3		160 (50.0)	160 (50.0)
30–39 years	252 (25.5)	13.2 ±5.9		111 (44.1)	141 (55.9)
40–49 years	194 (19.6)	12.1 ±5.6		98 (50.5)	96 (49.5)
50+ years	136 (13.7)	12.5 ±6.1		68 (50.0)	68 (50.0)
Sex, n (%)
Female	726 (73.2)	12.5 ±5.5	0.119	361 (49.7)	365 (50.3)	0.063
Male	266 (26.8)	13.1 ±5.7		150 (56.4)	116 (43.6)
Education, n (%)
Primary	7 (0.7)	13.1 ±9.2	0.060	4 (57.1)	3 (42.9)	0.074
Secondary	80 (8.1)	12.3 ±5.7		41 (51.3)	39 (48.7)
Undergraduate	469 (47.3)	12.2 ±5.4		222 (47.3)	247 (52.7)
Postgraduate	436 (43.9)	13.2 ±5.7		244 (56.0)	192 (44.0)
Marital status, n (%)
Single	448 (45.2)	12.5 ±5.2	0.429	226 (50.5)	222 (49.5)	0.636
Married	495 (49.9)	12.8 ±5.8		263 (53.1)	232 (46.9)
Divorced	39 (3.9)	11.9 ±6.3		17 (43.6)	22 (56.4)
Widow(er)	10 (1.0)	10.5 ±7.0		5 (50.0)	5 (50.0)
Nationality, n (%)
Emirati	180 (23.1)	14.1 ±5.2	<0.001	295 (49.3)	304 (50.7)	<0.001
Non-Emirati	599 (76.9)	12.4 ±5.6		115 (63.9)	65 (36.1)
Country of Residence, n (%)
UAE	644 (64.9)	12.6 ±5.6	0.519	337 (52.3)	307 (47.7)	0.189
Saudi Arabia	112 (11.3)	13.4 ±5.4		64 (57.1)	48 (42.9)
Lebanon	88 (8.9)	11.6 ±6.2		33 (37.5)	55 (62.5)
Egypt	27 (2.7)	12.2 ±5.9		14 (51.9)	13 (48.1)
Other Arab countries	43 (4.3)	13.0 ±5.1		22 (51.2)	21 (48.8)
Australia	33 (3.3)	12.0 ±4.5		15 (45.5)	18 (54.5)
Canada/USA	16 (1.6)	14.7 ±4.6		10 (62.5)	6 (37.5)
Europe	18 (1.8)	12.8 ±5.9		8 (44.4)	10 (55.6)
Other	11 (1.1)	13.5 ±6.5		8 (72.7)	3 (27.3)
Job, n (%)
Employed for wages	451 (45.5)	13.1 ±5.6	0.122	241 (53.4)	210 (46.6)	0.718
Self-employed	45 (4.5)	12.9 ±6.3		25 (55.6)	20 (44.4)
Unemployed	119 (12.0)	11.9 ±5.8		56 (47.1)	63 (52.9)
Retired/Cannot work	19 (1.92)	10.4 ±6.8		9 (47.4)	10 (52.6)
Home-maker	92 (9.3)	11.8 ±5.6		43 (46.7)	49 (53.3)
Student	266 (26.8)	12.6 ±5.2		137 (51.5)	129 (48.5)
Field of work, n (%)
Medical/Healthcare	203 (20.5)	12.9 ±5.5	0.030	111 (54.7)	92 (45.3)	0.065
Education sector	150 (15.1)	12.2 ±5.7		68 (45.3)	82 (54.7)
Academic/Research	142 (14.3)	13.7 ±6.0		86 (60.6)	56 (39.4)
White-collar	57 (5.7)	11.5 ±5.5		26 (45.6)	31 (54.4)
Police/Military	21 (2.1)	14.4 ±6.8		14 (66.7)	7 (33.3)
Blue-collar	8 (0.8)	10.0 ±6.4		3 (37.5)	5 (62.5)

Notes: *Unpaired t-tests or one-way analysis of variance (ANOVA) were applied to compare the mean well-being score with qualitative variables as appropriate. Chi-square and Fisher Exact tests were applied to compare binary variable of well-being with qualitative variables as appropriate. !Pearson correlation coefficient.  The bold value represents the number of participants.

Sources of Information and Frequency of Use

The majority of participants (72.3%) reported using social media platforms as a main source of information for COVID-19. Television was the third most commonly used tool for information on COVID-19, while newspapers, radio and scientific publications were less commonly used as sources of information (Figures 1). WhatsApp was reported as the most commonly used social media platform and was reportedly used for at least 1–3 hours a day (39.2%) (Figures 2). While most participants reported using Snapchat, Twitter, Instagram, and YouTube between 1–5 hours a day on average (Table 2). The least used platforms by participants were Reddit (91.2% do not use) and LinkedIn (71.2% do not use). Almost half of the respondents reported using the internet to access information about COVID-19.

Figure 1

Medium of information of COVID-19 (N:993).

Figure 2

Use of social media (hours per day) (N: 993).

Table 2

General Media Sources Use and Well-Being Score by Media Use

Variable	All	Well-Being Score	P-value*	Good Well-Being (≥12.5)	Poor Well-Being (<12.5)	P-value*
Medium of information on COVID-19
Social media
No	275 (27.7)	13.1 ±5.5	0.134	157 (57.1)	118 (42.9)	0.029
Yes	717 (72.3)	12.5 ±5.6		354 (49.4)	363 (50.6)
Internet
No	488 (49.2)	12.6 ±5.6	0.909	253 (51.8)	235 (48.2)	0.837
Yes	504 (50.8)	12.7 ±5.6		258 (51.2)	246 (48.8)
TV
No	570 (57.5)	12.5 ±5.6	0.593	298 (52.3)	272 (47.7)	0.573
Yes	422 (42.54)	12.7 ±5.5		213 (50.5)	209 (49.5)
Newspaper
No	861 (86.8)	12.7 ±5.6	0.553	448 (52.0)	413 (48.0)	0.400
Yes	131 (13.2)	12.3 ±5.7		63 (48.1)	68 (51.9)
Radio
No	940 (94.8)	12.7 ±5.6	0.419	485 (51.6)	455 (48.4)	0.823
Yes	52 (5.2)	12.0 ±5.1		26 (50.0)	26 (50.0)
Health agencies
No	958 (96.6)	12.6 ±5.6	0.280	492 (51.4)	466 (48.6)	0.604
Yes	34 (3.4)	13.6 ±5.5		19 (55.9)	15 (44.1)
Scientific literature
No	975 (98.3)	12.6 ±5.6	0.413	504 (51.7)	471 (48.3)	0.467
Yes	17 (1.7)	11.5 ±6.8		7 (41.2)	10 (58.8)
Word of mouth
No	979 (98.7)	12.6 ±5.6	0.373	503 (51.4)	476 (48.6)	0.581
Yes	13 (1.3)	14.0 ±5.0		8 (61.5)	5 (38.5)
Hospital/healthcare professional
No	984 (99.2)	12.6 ±5.6	0.851	507 (51.5)	477 (48.5)	1.000
Yes	8 (0.8)	13.0 ±6.7		4 (50.0)	4 (50.0)

Notes: *Unpaired tests were applied to compare well-being scores with social media. Chi-square or Fisher Exact tests were applied to compare binary variable of well-being with social media.

Social Media Practices and Frequency of Use

When asked about sharing information about COVID-19, 86% of participants claimed they validated information before sharing. In contrast, 62% reported they felt responsible for sharing knowledge. Moreover, 78% stated that they felt confident in their abilities to differentiate between false and correct information about COVID-19 and 65% agreed that they would send back comments or advice to the sender if they felt that information was dubious. Only 4.1% felt that their health-related knowledge and awareness had not increased due to the pandemic, while 17% had no opinion.

Well-Being Score and Comparison with Other Variables

Overall, almost half (51.5%) of participants in this study scored above average in the WHO well-being score (). Further analysis was performed to compare between sociodemographic factors and well-being scores, social media use and well-being scores. It was found that being a non-Emirati and a blue-collar worker was negatively associated with well-being score (p≤0.001 and P=0.03) respectively (Table 1). At the same time, non-Facebook users had significantly higher well-being scores than Facebook users (p≤0.001) (Table 3). Additionally, participants who felt confident with their ability to differentiate between false and correct information and those who were likely to inform the sender of dubious information had significantly higher well-being scores (p≤0.001 and p=0.027) (Table 4). Furthermore, participants who indicated an increase in knowledge and awareness of health-related information due to the COVID-19 pandemic reported higher well-being scores (p=0.005) (Table 3).

Table 3

Frequency of Social Media Use and Comparison with a Well-Being Score

Variable	All	Well-Being Score	P-value*	Good Well-Being (≥12.5)	Poor Well-Being (<12.5)	P-value*
Social media
No	275 (27.7)	13.1 ±5.5	0.134	157 (57.1)	118 (42.9)	0.029
Yes	717 (72.3)	12.5 ±5.6		354 (49.4)	363 (50.6)
Facebook
Do not use	406 (43.7)	13.4 ±5.4	<0.001	243 (59.9)	163 (40.1)	<0.001
<1 hour	251 (27.0)	12.6 ±5.7		127 (50.6)	124 (49.4)
1–3 hours	188 (20.2)	11.3 ±5.5		71 (37.8)	117 (62.2)
3–5 hours	59 (6.3)	12.5 ±5.5		27 (45.8)	32 (54.2)
>5 hours	26 (2.8)	9.5 ±6.3		9 (34.6)	17 (65.4)
Twitter
Do not use	498 (55.3)	12.4 ±5.7	0.224	246 (49.4)	252 (50.6)	0.193
<1 hour	223 (24.8)	13.3 ±5.2		130 (58.3)	93 (41.7)
1–3 hours	104 (11.6)	12.4 ±5.8		49 (47.1)	55 (52.9)
3–5 hours	40 (4.4)	13.7 ±5.1		22 (55.0)	18 (45.0)
>5 hours	35 (3.9)	12.6 ±5.7		19 (54.3)	16 (45.7)
Instagram
Do not use	211 (22.6)	12.4 ±5.9	0.130	104 (49.3)	107 (50.7)	0.352
<1 hour	242 (25.9)	12.9 ±5.7		129 (53.3)	113 (46.7)
1–3 hours	281 (30.1)	12.9 ±5.6		147 (52.3)	134 (47.7)
3–5 hours	136 (14.5)	13.0 ±4.9		76 (55.9)	60 (44.1)
>5 hours	65 (6.9)	11.1 ±5.4		27 (41.5)	38 (58.5)
YouTube
Do not use	140 (14.9)	12.3 ±5.7	0.221	72 (51.4)	68 (48.6)	0.716
<1 hour	316 (33.6)	13.1 ±5.7		168 (53.2)	148 (46.8)
1–3 hours	288 (30.6)	12.8 ±5.3		149 (51.7)	139 (48.3)
3–5 hours	115 (12.2)	12.9 ±5.3		63 (54.8)	52 (45.2)
>5 hours	82 (8.7)	11.6 ±5.7		37 (45.1)	45 (54.9)
LinkedIn
Do not use	638 (71.2)	12.6 ±5.5	0.345	320 (50.2)	318 (49.8)	0.472
<1 hour	201 (22.4)	13.0 ±5.5		111 (55.2)	90 (44.8)
1–3 hours	42 (4.7)	12.4 ±5.9		24 (57.1)	18 (42.9)
>3 hours	15 (1.7)	14.7 ±5.5		9 (60.0)	6 (40.0)
Reddit
Do not use	798 (91.2)	12.8 ±5.6	0.102	419 (52.5)	379 (47.5)	0.287
<1 hour	51 (5.8)	11.1 ±5.1		21 (41.2)	30 (58.8)
≥1 hours	26 (3.0)	13.0 ±4.7		13 (50.0)	13 (50.0)

Notes: *One-way ANOVA tests were applied to compare well-being scores with the frequency of social media use. Chi-square or Fisher Exact tests were applied to compare binary variable of well-being with the frequency of social media use.

Table 4

Social Media Practices and Comparison with a Well-Being Score

Variable	All	Well-Being Score	P-value*	Good Well-Being (≥12.5)	Poor Well-Being (<12.5)	P-value*
How often do you look up information and updates about COVID-19?
Not at all	39 (3.9)	13.1 ±6.0	0.146	22 (56.4)	17 (43.6)	0.342
Rarely	113 (11.4)	12.1 ±5.4		60 (53.1)	53 (46.9)
Occasionally	268 (27.0)	12.9 ±5.4		141 (52.6)	127 (47.4)
Often	371 (37.4)	12.6 ±5.9		176 (47.4)	195 (52.6)
Very often	201 (20.3)	13.8 ±5.2		112 (55.7)	89 (44.3)
Do you follow influencers on social media for COVID-19 information and updates?
Yes	208 (21.0)	12.9 ±5.9	0.354	110 (52.9)	98 (47.1)	0.622
No	654 (65.9)	12.4 ±5.4		330 (50.5)	324 (49.5)
I do not know any influencer	130 (13.1)	13.1 ±5.8		71 (54.6)	59 (45.4)
I only share COVID-19 related information from official sources or after validating from the official sources?
Strongly Disagree/ Disagree	39 (3.9)	13.6 ±7.3	0.021	24 (61.5)	15 (38.5)	0.069
Neither disagree nor agree	100 (10.1)	11.2 ±5.6		42 (42.0)	58 (58.0)
Strongly Agree/ Agree	853 (86.0)	12.8 ±5.5		445 (52.2)	408 (47.8)
I share COVID-19 related information because: via social media
It gives me some sense of satisfaction
No	819 (82.6)	12.5 ±5.5	0.093	410 (50.1)	409 (49.9)	0.047
Yes	173 (17.5)	13.3 ±5.9		101 (58.4)	72 (41.6)
I want to stay socially interactive
No	863 (87.0)	12.6 ±5.6	0.579	444 (51.5)	419 (48.5)	0.917
Yes	129 (13.0)	12.9 ±5.3		67 (51.9)	62 (48.1)
I have no specific reason
No	859 (86.6)	12.8 ±5.5	0.018	454 (52.9)	405 (47.1)	0.032
Yes	133 (13.4)	11.6 ±6.0		57 (42.9)	76 (57.1)
I was asked to forward it
No	956 (69.4)	12.7 ±5.6	0.436	495 (51.8)	461 (48.2)	0.387
Yes	36 (3.6)	11.9 ±4.9		16 (44.4)	20 (55.6)
I feel responsible to share knowledge
No	377 (38.0)	12.3 ±6.1	0.201	192 (50.9)	185 (49.1)	0.773
Yes	615 (62.0)	12.8 ±5.3		319 (51.9)	296 (48.1)
I wish to stay informed
No	948 (95.6)	12.6 ±5.6	0.607	491 (51.8)	457 (48.2)	0.411
Yes	44 (4.4)	12.2 ±5.6		20 (45.5)	24 (54.5)
I am confident in my ability to differentiate between true and false information related to COVID-19?
Strongly disagree/Disagree	38 (3.8)	14.1 ±5.5	≤0.001	24 (63.2)	14 (36.8)	≤0.001
Neither disagree nor agree	180 (18.1)	11.1 ±5.3		64 (35.6)	116 (64.4)
Strongly agree/ Agree	774 (78.0)	13.0 ±5.6		423 (54.7)	351 (45.3)
When I receive dubious or false COVID-19 related information, I inform the sender or try to correct them
Strongly disagree/ Disagree	23 (2.3)	9.7 ±4.8	0.011	7 (30.4)	16 (69.6)	0.027
	73 (7.4)	12.4 ±5.6		32 (43.8)	41 (56.2)
Neither disagree nor agree	251 (25.3)	12.1 ±5.5		120 (47.8)	131 (52.2)
Strongly agree /Agree	349 (35.2)	12.7 ±5.3		183 (52.4)	166 (47.6)
	296 (29.8)	13.3 ±5.9		169 (57.1)	127 (42.9)
My knowledge and awareness of health-related issues has increased as a result of the COVID-19 pandemic
Strongly disagree/Disagree	41 (4.1)	13.0 ±5.9	0.046	23 (56.1)	18 (43.9)	0.003
Neither disagree nor agree	169 (17.0)	11.7 ±5.7		67 (39.6)	102 (60.4)
Strongly agree/ agree	782 (78.8.)	12.8 ±5.5		421 (53.8)	361 (46.2)

Notes: *Unpaired t-tests or one-way ANOVA tests were applied to compare well-being score with the frequency of social media use. Chi-square or Fisher Exact tests were applied to compare binary variable of well-being with the frequency of social media use.

Regression Analysis of Social Media Use with Well-Being Scores

Simple and multiple linear regression analysis of well-being scores with social media use showed that Facebook use was a significant predictor of well-being score even after adjusting for covariates (Table 5). More use significantly predicted lower well-being scores, with 1–3 hours use compared to no use (β = −1.53, 95% CI= −2.79 – −0.27, p=0.017) and over 5 hours use (β =−0.32, 95% CI= −5.78 – −0.55, p=0.018) than no use. Similarly, the use of Reddit for less than one hour predicted lower well-being scores than no use (β=−2.24, 95% CI= −3.98 –-0.51, p=0.011). Additionally, often looking up COVID-19 information and updates predicted lower well-being scores (β=−2.99, 95% CI −5.03- −0.95, p=0.004). An increase in knowledge and awareness of health-related issues due to the COVID-19 pandemic did not remain a significant predictor of well-being scores after adjusting for covariates (β=0.69, 95% CI= −0.47 – –1.86, p=0.244).

Table 5

Crude and Adjusted Linear Regression Analysis of Well-Being Score with Social Media Use and Practices

Variable	Unadjusted Estimate (95% CI)	P-value*	Adjusted Estimate (95% CI)	P-value*
Facebook
<1 hour/	−0.78 (−1.65–0.09)	0.078	−0.19 (−1.30–0.92)	0.734
1–3 hours/	−2.09 (−3.04 – −1.13)	<0.001	−1.53 (−2.79 – −0.27)	0.017
3–5 hours/ Do not use	−0.95 (−2.45–0.56)	0.216	−0.04 (−2.00–1.91)	0.965
>5 hours/ Do not use	−3.87 (−6.05 – −1.68)	0.001	−0.317 (−5.78 – −0.55)	0.018
R2	0.028		0.054
Twitter
<1 hour/ Do not use	0.92 (0.05–1.80)	0.039	0.85 (−0.17–1.87)	0.101
1–3 hours/ Do not use	0.01 (−1.17–1.18)	0.998	−0.38 (−1.67–0.92)	0.569
3–5 hours/ Do not use	1.23 (−0.56–3.02)	0.179	0.74 (−1.23–2.71)	0.461
>5 hours/ Do not use	2.07 (−1.70–2.11)	0.831	−0.52 (−2.67–1.63)	0.633
R2	0.006		0.046
Instagram
<1 hour/ Do not use	0.46 (−0.57–1.49)	0.381	1.08 (−0.10–2.26)	0.072
1–3 hours/ Do not use	0.48 (−0.51–1.49)	0.339	0.54 (−0.67–1.76)	0.379
3–5 hours/ Do not use	0.62 (−0.58–1.82)	0.313	0.13 (−1.37–1.64)	0.860
>5 hours/ Do not use	−1.33 (−2.88–0.21)	0.091	−1.09 (−2.92–0.73)	0.241
R2	0.008		0.048
You \Tube
<1 hour/ Do not use	0.76 (−0.35–1.87)	0.178	0.27 (−1.03–1.58)	0.683
1–3 hours/ Do not use	0.45 (−0.67–1.57)	0.433	−0.22 (−1.57–1.13)	0.748
3–5 hours/ Do not use	0.58 (−0.80–1.95)	0.411	−0.27 (−1.89–1.34)	0.738
>5 hours/ Do not use	−0.75 (−2.27–0.76)	0.329	−0.98 (−2.78–0.82)	0.285
R2	0.006		0.040
LinkedIn
<1 hour/ Do not use	0.48 (−0.40–1.36)	0.285	0.99 (−0.11–2.09)	0.078
1–3 hours/ Do not use	−0.18 (−1.91–1.55)	0.838	0.55 (−1.58–2.69)	0.610
>3 hours/ Do not use	2.17 (−0.67–5.01)	0.133	2.40 (−1.08–5.88)	0.177
R2	0.004		0.047
Reddit
<1 hour/ Do not use	−1.70 (−3.26 – −0.13)	0.034	−2.24 (−3.98 – −0.51)	0.011
≥1 hours/ Do not use	0.19 (−1.98–2.35)	0.865	1.42 (−1.50–4.34)	0.341
R2	0.005		0.050
How often do you look up information and updates about COVID-19?
Rarely/ Not at all	−1.29 (−3.22–0.74)	0.214	−2.13 (−4.36–1.00)	0.060
Occasionally/ Not at all	−0.98 (−2.85–0.90)	0.307	−2.08 (4.16 – −0.01)	0.050
Often/ Not at all	−1.74 (−3.58–0.11)	0.065	−2.99 (−5.03 – −0.95)	0.004
Very often/ Not at all	−0.78 (−2.69–1.14)	0.427	−1.91 (−4.04–0.23)	0.080
R2	0.007		0.055
Do you follow influencers on social media for COVID-19 information and updates?
Yes/ No	0.47 (−0.40–1.34)	0.291	0.21 (−0.77 – −1.19)	0.680
I do not know any influencer/ No	0.63 (−0.42–1.68)	0.241	0.20 (−1.01–1.40)	0.749
R2	0.002		0.041
I only share COVID-19 related information from official sources or after validating from the official sources?
Disagree/ Neither disagree nor agree	2.24 (−0.29–4.77)	0.083	1.34 (−1.54–4.22)	0.362
Strongly disagree/ Neither disagree nor agree	2.51 (−0.43–5.45)	0.094	3.78 (0.36–7.20)	0.030
Agree/ Neither disagree nor agree	0.99 (−0.31–2.30)	0.137	0.67 (−0.85–2.19)	0.389
Strongly agree/ Neither disagree nor agree	1.70 (0.52–2.88)	0.005	1.11 (−0.29–2.50)	0.121
R2	0.011		0.048
I share COVID-19 related information because:
It gives me some sense of satisfaction (Yes/No)	0.79 (−0.13–1.70)	0.093	0.83 (−0.20–1.85)	0.115
R2	0.003		0.044
I want to stay socially interactive (Yes/No)	0.29 (−0.74–1.33)	0.579	−0.21 (−1.38–0.96)	0.720
R2	0.001		0.041
I have no specific reason (Yes/No)	−1.23 (−2.24 – −0.21)	0.018	−0.89 (−2.05–0.27)	0.131
R2	0.006		0.044
I was asked to forward it (Yes/No)	−0.74 (−2.60–1.12)	0.436	−2.05 (−4.19–0.09)	0.060
R2	0.001		0.045
I feel responsible to share knowledge (Yes/No)	0.47 (−0.25–1.18)	0.202	0.19 (−0.62–1.01)	0.646
R2	0.002		0.041
I wish to stay informed (Yes/No)	−0.44 (−2.13–1.25)	0.606	−0.82 (−2.58–0.95)	0.364
R2	0.001		0.042
I am confident in my ability to differentiate between true and false information related to COVID-19?
Disagree or Strongly disagree/ Neither disagree nor agree	3.02 (1.08–4.96)	<0.001	2.81 (0.49–5.13)	0.018
Agree/ Neither disagree nor agree	1.52 (0.56–2.48)	0.002	1.61 (0.52–2.70)	0.004
Strongly agree/ Neither disagree nor agree	2.31 (1.30–3.31)	0.002	1.70 (0.56–2.83)	0.003
R2	0.023		0.056
When I receive dubious or false COVID-19 related information, I inform the sender or try to correct them
Disagree/ Neither disagree nor agree	0.32 (−1.13–1.77)	0.662	−0.92 (−2.55–0.71)	0.267
Strongly disagree/ Neither disagree nor agree	−2.43 (−4.81 – −0.06)	0.045	−2.50 (−5.13–0.13)	0.063
Agree/ Neither disagree nor agree	0.63 (0.27–1.53)	0.172	0.35 (−0.69–1.38)	0.509
Strongly agree/ Neither disagree nor agree	1.18 (0.25–2.12)	0.013	0.57 (−0.51–1.64)	0.301
R2	0.013		0.051
My knowledge and awareness of health-related issues has increased as a result of the COVID-19 pandemic
Disagree or Strongly disagree/ Neither disagree nor agree	1.33 (−0.57–3.23)	0.170	0.98 (−1.29–3.25)	0.396
Agree/ Neither disagree nor agree	0.91 (−0.07–1.90)	0.068	0.49 (−0.61–1.60)	0.380
Strongly agree/ Neither disagree nor agree	1.50 (0.45–2.53)	0.005	0.69 (−0.47–1.86)	0.244
R2	0.008		0.043

Note: The multiple linear regression model was adjusted for age, sex, education, marital status, nationality, and job.

Discussion

This study explored social media use and its impact on well-being during the COVID-19 pandemic in the UAE, and other Arab-speaking countries. It also explored the use of social media as a mean for seeking knowledge about COVID-19. Over half of participants in our study reported an above-average well-being score, with lower scores found among participants who were non-Emirati and blue-collar workers. Participants who reported using any social media had poorer well-being scores. Being a health care worker and an Emirati showed poorer well-being. Specifically, using Facebook for, extended periods and to to seek COVID-19 information and updates predicted lower well-being scores. In this study, most participants believed in their abilities to validate shared information and shared it as a social responsibility. However, further research should be undertaken to investigate the accuracy of the practice of validating information and the sources used and accessed for validation. According to Figueiras et al, trust in authorities and information sources impacts behaviors and preventive measures during the pandemic.51

Although this study cannot indicate a direct and causal relationship between social media use and poorer well-being, these effects should not be ignored. In fact, other studies have investigated anxiety, stress, and depression related to the COVID-19 pandemic in the community and among subgroups. For example, Zhong et al suggested that social media may mediate impact on mental well-being. Also, during the pandemic, social media was described as a sources of information about COVID-19 and a means to share information.27

Despite social media being a valuable means of communication, evidence from previous outbreaks suggests a negative impact on people’s knowledge, practices, beliefs, and mental health status. Earlier studies have identified social media’s role in intensifying anxiety, stress and depression during the Ebola, SARS, ZIKA and MERS outbreaks. A study in South Korea (2020) found that social media use was related to anger and fear. It significantly increased preventive behaviors via risk perceptions and substantially influenced public perceptions of risk issues. Other studies that focused on H1N1 flu,52 Avian flu,53 and bovine spongiform encephalopathy also suggested similar findings to the South Korean study.54 Similarly, a recent study published in China showed that mental health problems, specifically depression and anxiety, were prevalent and positively associated with social media use during the COVID-19 outbreak, highlighting the significant role of governments in combatting the “infodemic”.55

In this study, being Emirati indicated poorer well-being when associated with social media use. This might be explained by the fact that our sample was mainly Emiratis, or due to the high amount of misinformation that spread at the early stages of the UAE pandemic compared to other neighbouring countries.56

Social media platforms are becoming more commonly used to acquire news and information and behavioral change.57 This study has shown high use of social media among participants compared to other traditional media platforms such as newspapers, television, and radio programs. However, in this study overuse of certain social media platforms was negatively associated with well-being; participants who used social media platforms for longer periods of time had lower well-being scores. A similar study showed that social media influenced people’s mental health and psychological well-being in Iraq, Kurdistan, and its usage was correlated with fear spread and panic.10 In the US, the pandemic was linked to a longitudinal decrease in well-being and poorer mood as presented by social media.58

In general, over half of the participants scored above average on well-being, highlighting that there might be an equitable effect for social media in accessing information. Our participants indicated strong confidence in their knowledge, practices, and the ability to control the spread of information through WhatsApp and other social media platforms.27 This confidence might be because most participants were educated (health care professionals, academics, and students), or that large amounts of information were transmitted before and during the time of data collection for this study and the onset of the pandemic. Zhong et al discussed similar findings that social media affects mental health because younger people rely more on social media to receive information about the pandemic.27 It is worth noting that our sample had a considerable number of respondents stating that they are health care workers, which may introduce bias to our results. However, when we controlled our sample by occupation, the differences were not significant. Furthermore, based on recent research, it may be argued that health care professionals have shown higher rates of depression and anxiety during the COVID-19 pandemic, affecting their well-being.59–62 Two different studies conducted in the UAE during the COVID-19 pandemic found that health care workers in the UAE suffer from anxiety, psychological distress and burnout.62,63 However, our study mainly correlated the well-being of participants with social media practices, not with the type of work.59–61

In particular, WhatsApp messaging has been used to help families cope with social isolation and support social inclusion.33 WhatsApp was the most common and often used social media platform among participants in our study. Although we did not investigate the reasons for the use of WhatsApp in our research, the everyday use of this platform as a tool for social networking is consistent with other findings.33 In the internet-driven information age, there is significant pressure on governmental entities, health officials, and public health specialists to come up with ways to combat the considerable damage of social media on trust in scientific evidence. These effects will likely be felt for more time to come, even after the COVID-19 pandemic wanes. It is becoming increasingly necessary for health and government officials to promote digital literacy as a method to combat the spread of false, misleading, or unreliable information. Several studies reported the role of social media in promoting health protection and highlighted the significance of public health messaging. For example, past research conducted by Collinson et al. showed the value of controlling the dissemination of influenza and decreasing the infection outcomes on a population.14 Social media campaigns regarding epidemics or pandemics are in communicating information to the general public, thereby producing positive attitudes and behaviors that may reduce the spread of the disease by means such as hand-washing and social distancing.4,17,19,57 It is worth pointing out that answering “very often” and “often” to the question of “how often do you look up information” may be an indicator for anxiety and may reflect negatively on well-being. However, in our results, this was not significant and has not shown any impact on well-being.

Various studies have shown a high prevalence of COVID-19 awareness and its impact on practices and attitudes being propagated by the rapid spread of information across the globe through many platforms, including social media. It is worth noting that during the COVID-19 pandemic, both the virus and the information/misinformation around the outbreak spread very fast.3,5,18 This awareness could provide valuable context to the levels of confidence that participants of this study expressed they had in their knowledge and their indication that they only shared correct information and confirmed the truthfulness of information before disseminating it further. It can be argued that factors that may have led to amplified use of social media during this pandemic can be linked to the pandemic, its subsequent social isolation, and stringent containment measures that led to marked restrictions in typical social communication, unlike any of the previous historical epidemics and pandemics.64

Throughout pandemics, the primary research on social media was during the 2009 H1N1 outbreak, tracing the prevalence of misinformation (reported as 4.5%), terminology use (H1N1 or swine flu), public attitudes and fear, and relationships between case incidence and public concern.65 Furthermore, some individuals generate assumptions regarding disease transmission, often based on incorrect risk perceptions and pre-existing social prejudices, to ease the anxiety felt by the accusers or stigmatize those who are blamed. This was reported in April 2009 when a new strain of human influenza H1N1 emerged in Mexico and spread quickly worldwide.66 This resulted in an economic crisis that affected Mexican goods as they were considered disease vectors.67 To decrease the adverse outcomes of stigmatization during pandemics, public health officials must attempt to identify the dynamics that underlie this process, with particular attention given to protecting disadvantaged people.10,21,57,68,69

Moreover, a past study investigating media dependency amongst Chinese nationals during the SARS epidemic of 2003 found that the internet was a unique empowerment tool that allowed people to avoid official authority and challenge official claims during the crisis.31 In 2015, Misra et al. developed a model to explore the impact of awareness realized by social media campaigns on infectious disease prevalence. The study findings revealed that social media health campaigns lead to behavioral changes among individuals, causing them to isolate themselves and protect themselves from infection, thereby decreasing the number of infected people.32

In the earlier stages of the pandemic (during the survey distribution period), misinformation and disinformation spread was at its peak. Interventions by leading authorities around the globe and international organizations started to appear as the infodemic effect started, especially with vaccine uptake and testing and preventive measures. The UAE’s efforts to contain the pandemic and spread awareness were especially impressive, placing it at the top of the list of countries with effective mass-testing and spread awareness was especially remarkable, with effective mass-testing and vaccine coverage.70 Globally, governments and corporations have used various means to try and stem the tide of misinformation. Recently, YouTube and Twitter tried to introduce terms and policies regarding content deemed to be COVID-19 and vaccination misinformation. Programs and calls for e-literacy and digital health education are already in place to combat the social media-caused negative impact on knowledge and awareness, which we hope will decrease the negative effect on well-being.69

We acknowledge that there are limitations to this study, including the cross-sectional design and online nature of the survey, which may lead to bias. Moreover, other confounding factors may affect well-being scores such as self-esteem, family relations, physical health, and other factors not considered in this study. Additionally, not having a control group of non-social media users for comparison of baseline well-being scores is a limitation to this study and future studies should consider this. As the survey was online, there is potential for more educated, younger, and affluent people with internet access to respond. However, given the nature of the pandemic and the limitations associated with conducting research using more robust methods, the results are helpful for providing baseline data to inform further research investigating the impact of social media on well-being in the future.

Conclusion

This study confirms the importance of authority intervention and guidelines that may be needed to control facets of social media use and reduce the negative impacts on health and well-being. Awareness campaigns and collaborative efforts between public health authorities and specialists and information technology and communication specialists, are needed. Further studies on the impact of social media during pandemics are needed too to ensure responsible and constructive use of social media. Providing factual, evidence-based information and making them accessible to consumers is highly important to reduce negative impact on the well-being. Public health communities need to learn how to engage the public to make healthy choices while raising public awareness.