key: cord-0786360-5woq1vgm
authors: Mavroeidi, A.; Innes, R.; Miyake, E.; Pennington, D.
title: Why were Twitter Users Obsessed with Vitamin D during the first year of the pandemic?
date: 2022-04-18
journal: nan
DOI: 10.1101/2022.04.13.22273830
sha: d076df3dccd49a00928343147eeb7c19f153ffb5
doc_id: 786360
cord_uid: 5woq1vgm

The aim of this study was to explore how the relationship between vitamin D and COVID-19 has been represented on the social media site Twitter. NCapture was used to collect textual Tweets on a weekly basis for three months during the pandemic. In total, 21,140 Tweets containing the keywords vitamin D and COVID were collected and imported to NVivo12. An inductive thematic analysis was carried out on the Tweets collected on the first (12/2/2021) and last week (21/5/2021) of the recording period to identify themes and subthemes. Quality control of the coding was conducted on a sample of the dataset (20%). Data were also compared to the ground truth to explore the accuracy of media outputs. The four main themes identified were a)association of vitamin D with COVID-19, b)politically informed views, c) vitamin D deficiency and d) vitamin D source. When compared to the ground truth, the majority of information relating to the key findings was incorrect for all of the findings. This study contributes to the area of research by highlighting the extent of the issue social media sites face with health-related misinformation. In the context of COVID-19, it is important that sites such as Twitter improve their existing misinformation policies, as misinformation can be detrimental in disease prevention.

The COVID-19 pandemic caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a subject of global concern (Misra et al., 2020) .

There has been an unprecedented response across several sectors to develop targeted therapeutics to slow the rate of infection (Carter et al., 2020) . One modifiable lifestyle intervention that has received particular interest is vitamin D (Lanham-New et al., 2020) .

The scientific debate basis for the potential relationship of vitamin D status with COVID-19 is based upon the association of low serum concentration of 25(OH)D with increased susceptibility to acute respiratory tract infections (Ilie et al., 2020) . Several correlational studies have reported a relationship between vitamin D deficiency and poor COVID-19 outcomes (Ilie et al., 2020; Sulli et al., 2021; Baktash et al., 2020) . However, a number of risk factors for poor COVID-19 outcomes are similar to those for vitamin D deficiency such as old age and darker skin (The World Health Organization, 2020; Dey & Sinha, 2020) . Consequently, there are several other risk factors that must be considered before the true extent of the cause-andeffect relationship between vitamin D and COVID-19 can be extrapolated.

The potential impacts of social media on the COVID-19 pandemic are currently unknown, as it is the first pandemic to develop in the era of mass social media use. Social media differs to traditional media as it allows users to both consume and contribute to news in real time (Das & Ahmed, 2020) . This participatory culture in news and social media could have both positive and negative consequences (Almgren & Olsson, 2016; Robinson & Wang, 2018) . Factually correct information can be circulatedby official government sources e.g., Scottish Government (https://twitter.com/scotgov) -on social media faster than is possible with traditional media sources. Sharing accurate information at the earliest time possible such as local guidelines can help reduce the rate of infection (Mejia et al., 2020) . However, social media is also synonymous with misinformation (Mancosu & Vegetti, 2020; Shu et al., 2020) . The WHO stated that there was not only a pandemic, but also an "infodemic" which was defined as an outbreak of All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 misinformation causing mass anxiety and uncertainty (Das & Ahmed, 2020) . In the pandemic response, misinformation is dangerous, as information is a key factor in disease prevention (Mejia et al., 2020) .

The aim of this study was to explore how the relationship between vitamin D and COVID-19 was presented on the social media site Twitter during the first year of the pandemic, and assess the level of misinformation presented in Twitter around this issue.

Ethical approval was obtained through the Strathclyde University Computer and Information Sciences Ethics Committee on 1/2/2021 (ID: 1303).

Textual data were collected from the social media site Twitter. NCapture (NVivo 12 plugin) was used to collect all Tweets in English containing the keywords "vitamin D" and "COVID" using the advanced search function on Twitter; no hashtags were used. NCapture does not have the ability to capture images, videos or emojis, so these were not included in the analysis. This was also the main reason why other platforms such as Instagram and TikTok were not included in this analysis, as they mainly contain image and video data. Facebook was also not included because its content is not considered publicly available and thus there are additional ethical considerations. Tweets were collected on a weekly basis for three months from 12/2/2021 until 21/5/2021-resulting in a total of 21,140 Tweets being collected. This All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 time period was chosen as vaccination programmes were rolled out in most countries in the developed world, including the UK. As a consequence, alternative preventative remedies were a trending topic on Twitter. Collecting Tweets on a weekly basis made it possible to track the prevalence of Tweets containing the keywords (Figure 1 ).

Inductive thematic analysis was used to investigate how the relationship between vitamin D and COVID-19 has been represented on Twitter. Thematic analysis is a qualitative technique that enables researchers to extract meanings and concepts from data sets that cannot be quantitively analysed, such as focus group transcripts (Javadi & Zarea, 2016) . The technique is commonly utilized in health settings to analyse patient experiences (Hewis, 2015; Attard & Coulson, 2016; Lane et al., 2016 ). An inductive approach was used, meaning that the analysis was data driven (Braun & Clarke, 2006) . Therefore, the themes identified are strongly linked to the data as the student researcher (RI) coded the data without trying to fit it into a pre-existing coding frame (Braun & Clarke, 2006) . The student researcher (RI) followed the step-by-step guide outlined in (Braun & Clarke, 2006) to analyse the data. The six steps are as follows: familiarising yourself with your data, generating initial codes, searching for themes, reviewing themes, defining and naming themes, and producing the report.

The inductive thematic analysis was carried out in NVivo 12 -a software package commonly used for qualitative analysis (Edhlund & McDougall, 2019) . Only Tweets collected on 12/2/2021 and 21/5/2021 were analysed due to time constraints All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 (Appendix A & B). Quality control of the coding was conducted by another postgraduate research student with previous experience of thematic analysis (anonymous 1), who coded a sample of the dataset (20%). The student who carried out the quality control (anonymous 1) agreed with the provisional themes and sub themes identified by the student researcher (anonymous 2).

The codes identified from analysis of the Twitter data set collected on 12/2/2021 were also compared to the "ground truth" in order to explore the accuracy of media outputs. The NICE report titled "COVID-19 rapid guideline: vitamin D" was used as the 'ground truth' (National Institute for Health and Care Excellence, 2020).

The four key findings identified from the NICE report were:

Do not offer a vitamin D supplement to people solely to treat COVID-19.

Do not offer a vitamin D supplement to people solely to prevent COVID-19.

Low vitamin D status was associated with more severe outcomes from COVID-19. However, it is not possible to confirm causality because many of the risk factors for severe COVID-19 outcomes are the same as the risk factors for low vitamin D status.

Adults, young people and children over 4 years should consider taking a daily supplement containing 10 micrograms of vitamin D between October and early March. Groups at high risk of vitamin D deficiency should also consider taking a daily supplement containing 10 micrograms of vitamin D year-round.

Data that did not relate to the key findings outlined above were excluded from this portion of the analysis. Appropriate data were compared to the key findings and coded in NVivo into either 'correct' or 'incorrect' nodes.

All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 

The number of original Tweets shared containing the keywords "vitamin D" and "COVID" peaked in mid-February before steadily declining until mid-March, at which point, the number of Tweets generally trended upwards until late April. The number of Tweets then rapidly declined over the month of May. (Figure 1 ). (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The belief that scientific literature had 'proven' the relationship between vitamin D and COVID-19 infection was also common.

"@redacted There is by now so much literature that vitamine D deficiency exacerbates the chance to get infected by COVID-19 or a worse outcome, All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. However, there were also users that highlighted vitamin D was not a proven treatment for COVID-19.

"@redacted Agreed that Vitamin D is good for you in appropriate amounts, but as far Covid treatment is concerned, it and Ivermectin seem more like false promise wrapped in a thin layer of anecdote and rationalisation." (Twitter User 10)

"@redacted There is no evidence that vitamin D has anything to do with COVID. It certainly doesn't seem to have any impact on treating it." (Twitter User 11)

All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022.

Lockdown measures and the restrictive effect that these might have had to people's exposure to sunlight, and thus diminishing vitamin D status, were heavily criticised.

"@redacted For morons out there that don't get my joke, lock downs are moronic, cause more harm than good. While it's all theories with COVID, will be interesting to see if low Vitamin D caused bad outcomes, because that would mean the lockdowns more adversly affected dark skinned people." (Twitter User 12)

There was also a widespread belief that face masks were ineffective in reducing the risk of COVID-19 transmission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ; https://doi.org/10.1101/2022.04.13.22273830 doi: medRxiv preprint COVID-19 vaccines were also often described in a negative light (e.g. unsafe and useless) while vitamin D was presented as the miracle safe cure for COVID. "@redacted There is an argument to be made that your skin color affects your outcome. It's well known that the darker your skin, the less Vitamin D you get from sunlight." (Twitter User 24) "Obesity is strongly correlated with Vitamin D deficiency. Vitamin D deficiency is a major factor in whether someone becomes seriously ill or dies from Covid.

All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The positive impact on general health was linked to improved outcomes in COVID-19 sufferers.

However, it was also pointed out that vitamin D status alone does not determine if an individual is healthy.

"@redacted Vitamin D important in context of overall Health. Overall Health significant in lessening severity of COVID. But overall health is more than Vitamin D, it is not a therapy for COVID or alternative to vaccines and will be unable to protect those who most need protection from COVID" (Twitter User 28)

All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 

Food fortification programs were linked to COVID-19 outcomes.

Foods's rich in vitamin D were often listed with claims that they could prevent mortality in COVID-19 sufferers.

All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 There was also particular interest in the homeless population due to their increased exposure to sunlight.

"@redacted The only thing the homeless have going for them is plenty of sunshine and vitamin D. Maybe some "scientist" should pick up on that as they are virtually immune to Covid." (Twitter User 37) All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 "@redacted Am I reading this right? I think they're saying why aren't homeless people dying -well, they've probably got very good vitamin D levels, much like those farm workers who all had covid and didn't know. The people who are dying are old people who don't get out enough" (Twitter User 38)

There was a widespread belief that the government recommended dose of 10 micrograms was too low for both general health and any perceived COVID-19 benefits.

"@redacted RDA is typically grossly underestimated (I wonder why that it is?). 5,000iu D3 to maintain your existing level. 10,000iu to increase your D3 level." Twitter users also commonly stated that they were personally taking a vitamin D supplement that was much higher than the 10 micrograms recommended. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ; https://doi.org/10.1101/2022.04.13.22273830 doi: medRxiv preprint

Some users highlighted that even if vitamin D is not associated with COVID-19, it still plays an important role in the maintenance of general health. Figure 2 shows the number of 'correct' and 'incorrect' codes identified when data from Twitter (12/2/2021 Tweets) were compared to the key findings from the NICE report titled "COVID-19 rapid guideline: vitamin D" (National Institute for Health and Care Excellence, 2020). The majority of the information relating to the key findings was 'incorrect' for all of the findings. All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ; https://doi.org/10.1101/2022.04.13.22273830 doi: medRxiv preprint 

The NICE report states that vitamin D supplementation should not be used as a COVID-19 treatment. 'Incorrect' Tweets were over five times more common than 'correct' Tweets.

"@redacted There is no evidence that vitamin D has anything to do with COVID. It certainly doesn't seem to have any impact on treating it." (Twitter User 11) Incorrect All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ; https://doi.org/10.1101/2022.04.13.22273830 doi: medRxiv preprint "I'm obviously thrilled with this study result, but I also feel pretty sad. It was pretty clear to me in August that vitamin D was a potent treatment for Covid, 1.5 million people have died since then." (Twitter User 45)

Vitamin D supplementation should not be offered to prevent COVID-19 infection.

COVID-19 infection was the most discussed theme on Twitter. This theme also had the highest proportion of 'incorrect' information.

Correct "@redacted They didn't. They've been telling people to take vitamins for the last 50 years. Telling people to specially take vitamin D for COVID like it'll make them immune when it won't even come close is however a murderously bad idea in anyone with a brain's book." (Twitter User 46) Incorrect "@redacted The HSE chiefs should have acted earlier and informed the public that vitamin D is a great preventative for covid.... This needs to be addressed by the nurses union" (Twitter User 47)

Low vitamin D status is associated with more severe COVID-19 outcomes; however, it is not possible to confirm causality. 'Incorrect' Tweets were almost four times more common than 'correct' Tweets.

Correct All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 

Adults, young people, and children over 4 years should consider taking a daily supplement containing 10 micrograms of vitamin D between October and early

March. Groups at high risk of deficiency should also consider year-round supplementation. Supplementation was the least discussed theme on Twitter.

However, this theme had the highest proportion of 'correct' Tweets when compared to the other key findings.

Correct "Surprised to discover that the NHS recommends *everyone* take Vitamin D from October to March, during Covid." (Twitter User 50) Incorrect "@redacted RDA is typically grossly underestimated (I wonder why that it is?). 5,000iu D3 to maintain your existing level. 10,000iu to increase your D3 level." (Twitter User 39) Discussion All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 This study identified how the relationship between vitamin D and has been represented on the social media site Twitter. The four main themes identified through inductive thematic analysis were association of "vitamin D with COVID-19", "politically informed views", "vitamin D deficiency" and "vitamin D sources" (Table 1) . When compared to the ground truth, the majority of information relating to the key findings was 'incorrect' for all of the findings (Figure 2 ). The ground truth analysis carried out in this study is similar to the fact checking processes traditional journalists engage in as part of their professional practices, especially on social media (Juneström, 2020) .

Considering the implications of these findings on public health, high levels of misinformation are dangerous in the context of COVID-19, as information is a key factor in disease prevention (Mejia et al., 2020) . Ofcom reported that in 2020, 45% of people in the UK used social media to follow breaking news stories (Ofcom, 2021) .

Due to the large quantity of misinformation identified (Figure 2) , it is likely that many in this portion of the population are exposed to similar misinformation on a regular basis. Consider one of the 45% reading the Tweet "@redacted The HSE chiefs should have acted earlier and informed the public that vitamin D is a great preventative for covid.... This needs to be addressed by the nurses union" (Twitter User 49). If they believe the Tweet contains correct information, they may stop following guidelines such as social distancing that aim to prevent COVID-19 infection, and instead begin supplementing with vitamin D. In this case, their risk of infection, and consequently the risk faced by others in the local community would increase. However, no data on behaviour change in the public was collected, so it is not possible to confidently draw any conclusions regarding the impact of the misinformation on public health. Though, similar previous research measuring the All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ;  impact of COVID-19 vaccine misinformation on vaccination intent highlighted that exposure to online misinformation reduced intent to accept a vaccine (Loomba et al., 2021) . Therefore, in this study, it is hypothesised that public health may have been negatively impacted by the high levels of misinformation shared on Twitter.

Comparing this study to previous research concerning social media response to infectious disease outbreaks, both the methodology and results are similar. (Ahmed et al., 2019) explored Twitter users views on the H1N1 virus by thematically analysing Tweets collected using keyword searches. Some of the 8 key themes and 38 subthemes identified by researchers are comparable to themes and subthemes identified in this study relating to vitamin D and COVID-19. Politics and food were two of the main themes, while prevention products, medication and media criticism were sub themes.

Another similar study, (Oyeyemi, Gabarron & Wynn, 2014) used keyword searches on Twitter to test the quality of Ebola related information shared by users in Guinea, Liberia, and Nigeria from 1st to 7th September 2014. It was reported that 55.5% of Tweets contained medical misinformation, while only 36% contained medically correct misinformation. Notably, the proportion of incorrect information relating to vitamin D and COVID-19 reported in this study is even more skewed (Figure 2 ). High levels of misinformation are synonymous with social media sites such as Twitter because users sharing 'news' do not need to adhere to the same verification processes as traditional journalists (Boididou et al., 2017) . Consequently, users can freely share misinformation with no repercussions. Current Twitter policy states sharing content that may mislead people about "the efficacy and/or safety of preventative measures, treatments, or other precautions to mitigate or treat the disease (COVID-19)" may not be shared on Twitter (COVID-19 misleading All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 information policy, 2021). However, it is evident that this policy is not being enforced effectively ( Figure 2 ). Therefore, Twitter should consider improving their current methods of combating misinformation in order to effectively enforce their policy.

The strengths and weaknesses of this study must be considered. Collecting

Tweets on a weekly basis provided a high volume of data and allowed researchers to track changes in the number of Tweets shared (Figure 1) , providing a good insight into Twitter users behaviours throughout this time period. However, NCapture can only collect Tweets published in the previous eight days, meaning that Twitter data from the early stages of the pandemic were unavailable. It is possible that insights from Twitter would be different if data from earlier periods of the pandemic were available. Furthermore, data was only obtained from Twitter. Different social media platforms contain different types of information (Cinelli et al., 2020) , so it is likely that if data were also analysed from other platforms that the results would differ.

Additionally, as only textual data was collected, other types of data such as images, videos and emojis were not included in the analysis. Therefore, the nuance of other data types was not captured. It is also important to note that the only keywords used were "vitamin D" and "COVID", meaning that Tweets using abbreviations such as "vit D" were not collected.

This study highlights the extent of the issue social media sites face with misinformation. Companies like Twitter have an obligation to fight misinformation, as it can be detrimental in disease prevention (Mejia et al., 2020) . Future research should consider insights from various social media platforms. Different social media platforms contain different types of information (Cinelli et al., 2020) , so it is possible that insights relating to COVID-19 may also differ. All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 18, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 

Novel insights into views towards H1N1 during the 2009 Pandemic: a thematic analysis of Twitter data

Commenting, Sharing and Tweeting News

A thematic analysis of patient communication in Parkinson's disease online support group discussion forums

Vitamin D status and outcomes for hospitalised older patients with COVID-19

Detection and visualization of misleading content on Twitter

Using thematic analysis in psychology

Considerations for Obesity, Vitamin D, and Physical Activity Amid the COVID-19 Pandemic

The COVID-19 social media infodemic

COVID-19 misleading information policy

Despite concerns, COVID-19 shows how social media has become an essential tool in the democratisation of knowledge (Blog)

Ethnicity and COVID-19 -A commentary on "World Health Oganization declares global emergency: A review of the 2019 novel coronavirus (COVID-19)

NVivo 12 essentials

Do MRI Patients Tweet? Thematic Analysis of Patient Tweets About Their MRI Experience

The role of vitamin D in the prevention of coronavirus disease 2019 infection and mortality

Understanding Thematic Analysis and its Pitfall

An emerging genre of contemporary fact-checking

Vitamin D and SARS-CoV-2 virus/COVID-19 disease

Factors that shape the patient's hospital experience and satisfaction with lower limb arthroplasty: an exploratory thematic analysis

Measuring the impact of COVID-19 vaccine misinformation on vaccination intent in the UK and USA

Is It the Message or the Messenger

Vitamin D supplementation to prevent acute respiratory tract infections: systematic review and meta-analysis of individual participant data

The Media and their Informative Role in the Face of the Coronavirus Disease 2019 (COVID-19): Validation of Fear Perception and Magnitude of the Issue (MED-COVID-19)

Rheumatologists' perspective on coronavirus disease 19 (COVID-19) and potential therapeutic targets

COVID-19 rapid guideline: vitamin D

Fewer people getting news from social media

Retrieved 13 th

Ebola, Twitter, and misinformation: a dangerous combination

Networked News Participation: Future Pathways

Disinformation, Misinformation, and Fake News in Social Media

Vitamin D and Lung Outcomes in Elderly COVID-19

Report of the WHO-China Joint Mission on Coronavirus Disease