Content analysis of social media regarding obesity as a chronic disease

Study setting and sample

This study evaluated a sample of Facebook posts and tweets of 11 major obesity-related health organizations and 10 commercial weight loss application accounts. The list of health organizations (Appendix A) was selected from a list of 26 organizations and events curated by the ACTION study, landmark research designed by a multi-disciplinary steering committee that comprised representatives from the Obesity Society, the Obesity Action Coalition and the Integrated Benefits Institute, and other obesity experts (Depue et al., 2015). As for commercial weight loss apps, we used Healthline lists for the best weight loss apps of 2021 (10 apps). The Healthline team manages an extensive medical network of over one hundred professionals covering more than fifty specialties, ensuring that published information is accurate and evidence-based (Wang et al., 2021). All posts and tweets were publicly accessible online, and no names were recorded.

Data collection

General information about the Facebook or Twitter account was collected. For the organizations, the following data were collected: the aim of the organization; whether it is local, international, or regional; the number of its members; the number of followers, and the total number of posts. As for weight loss applications, data collected included: focus of the app, whether the app was paid or free, the total number of posts, and total number of followers. Appendix A shows the characteristics of the organization or app’s social media accounts.

All posts (Facebook) and tweets (Twitter) between Aug 1, 2020, and Jan 31, 2021, were extracted manually to an excel document that is password-protected and shared among the research team. Duplicates of the same post or tweet within the same social media platform or across the two platforms were included. If the post was a retweet or share of another Facebook post, only the text posted by the concerned organization was included; any linked images, videos, or websites were excluded. Written posts other than English were excluded. In the excel sheet, each organization or app was given a code; therefore, the posts were deidentified. From each post or tweet, we recorded the date, the full post/tweet, and the code of the organization/weight-loss app.

Codebook development

The research team developed a codebook to interpret whether the meaning of the post/tweet text is related to the concept of obesity as a chronic disease. The initial codebook was derived from the elements of obesity as a chronic disease: chronicity, multifactorial causes, risk factors, symptoms, complications, prevention, and treatment. Using the initial codebook, a deductive approach was used to code 5–10% of the posts and tweets. Two investigators (DH and FH) coded each post or tweet independently, with the possibility to code a post with more than one code. The investigators used an inductive method to identify additional codes based on the data throughout this round. The study team collaboratively developed an updated final codebook applied to the data (Table 1).

Data analysis

Content analysis was used to interpret the meaning of the posts’ texts. The unit of analysis was the whole post or tweet. Two investigators independently identified the category of the posts/tweets based on the codebook. First, the post/tweet was identified as related to obesity as a chronic disease or irrelevant. If it was related to obesity as a chronic disease, the post’s category was identified using the codebook. Any discrimination was resolved by the third investigator (JA) after a discussion with the team. The excel sheet data was imported into IBM SPSS statistical package, version 26, for data analysis. Descriptive analysis was performed to quantify the number of codes and categories as percentages of the total posts. Sub-analysis was performed based on the health organizations and apps.

Comparison with prior work

Obesity discussions differ across social media platforms. Chou, Prestin & Kunath (2014) analyzed 2.2 million posts to study weight-related social media interactions. Condescending and derogatory comments with the content of weight stigmatization flooded Twitter and Facebook, while more insensitive comments flooded blogs and forums (Chou, Prestin & Kunath, 2014). In our study, we only analyzed Facebook and Twitter content. Still, we did not record any discriminatory or weight-stigmatizing comments as the social media accounts of both the obesity health organizations and weight loss applications are professional. Yet, in our study, weight bias was discussed in only 2.5% of the obesity health organization posts. It seems their social media accounts fail to promote awareness about obesity as a disease (as previously mentioned) and other important topics such as weight bias. In addition, Chou, Prestin & Kunath (2014) also reported the impact of language usage. Using the words obese and obesity rather than fat resulted in more discussions that focus on obesity as a disease (Chou, Prestin & Kunath, 2014). Our study findings show that even when the word obese is used, and obesity is addressed as a disease, obesity health organizations discuss it superficially without going into depth. Obesity health organizations need to move towards a more proactive role in changing the perception of obesity as a disease and decreasing weight bias.

In the past few decades, social media has proven its efficacy in raising awareness and creating change in public perceptions and attitudes on various social issues such as smoking (Naslund et al., 2017; Yoo, Yang & Cho, 2016), HIV testing, and adherence to medication (Bychkov & Young, 2018; Wang et al., 2019), and most recently COVID vaccinations (Puri et al., 2020). Social media has been used for public awareness regarding the appropriate use of antibiotics (Catalan-Matamoros, Pariente & Elias-Perez, 2019). Concerning obesity, news media frames show a significant impact on weight-related attitudes. Compared with people who read fat-positive frames (that highlight obesity as uncontrollable by individual factors), people who read fat-negative frames (that highlight obesity as a controllable issue) were more likely to negatively judge obese people and support weight-based job discrimination and prejudice. Yet, adding anti-stigma messages tempered the expression of prejudice (Frederick et al., 2020). Social media is a cost-effective tool that can hold obesity support groups (Dahl, Hales & Turner-McGrievy, 2016), fight weight stigma (Zavattaro, 2021), and potentially educate the public about obesity as a disease entity. Health organizations, healthcare providers, and governments should play a key role, through social media platforms, in providing the public with evidence-based information about obesity as a complex multifactorial disease.

Strengths and limitations

This study analyzes how social media portrays obesity as a chronic disease. It’s the result of a manual content analysis of many posts from two popular social medial platforms. It adds to our understanding of the information displayed to the public regarding the obesity pandemic. However, some limitations affect the generalization of the results. First, only Facebook and Twitter posts were analyzed. Even though these are still the most popular social media platforms, they are not representative of all social media platforms. Second, because only publicly accessible posts were mined, a substantial amount of private content may be left out of the analysis. Finally, this study does not look into social media conversations. It only analyzed initial posts as we were mainly interested in exploring the information organizations and commercial apps are disseminating.