This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
Diabetes patient associations and diabetes-specific patient groups around the world are present on social media. Although active participation and engagement in these diabetes social media groups has been mostly linked to positive effects, very little is known about the content that is shared on these channels or the post features that engage their users the most.
The objective of this study was to analyze (1) the content and features of posts shared over a 3-year period on 3 diabetes social media channels (Facebook, Twitter, and Instagram) of a diabetes association, and (2) users’ engagement with these posts (likes, comments, and shares).
All social media posts published from the Norwegian Diabetes Association between January 1, 2017, and December 31, 2019, were extracted. Two independent reviewers classified the posts into 7 categories based on their content. The interrater reliability was calculated using Cohen kappa. Regression analyses were carried out to analyze the effects of content topic, social media channel, and post features on users’ engagement (likes, comments, and shares).
A total of 1449 messages were posted. Posts of interviews and personal stories received 111% more likes, 106% more comments, and 112% more shares than miscellaneous posts (all
Diabetes social media users seem to be least engaged in posts with content topics that a priori could be linked to greater empowerment: research and innovation on diabetes, and health education. Diabetes social media groups, public health authorities, and other stakeholders interested in sharing research and innovation content and promoting health education on social media should consider including videos and emoji in their posts, and publish on popular and visual-based social media channels, such as Facebook and Instagram, to increase user engagement.
RR2-10.1186/s12913-018-3178-7
Patient associations and patient groups from around the world are increasingly more present on social media. Being both ubiquitous and freely accessible, social media channels allow patient associations to share content and connect with individuals interested in their health condition. Representing one of the most prevalent chronic diseases worldwide, diabetes associations and diabetes patient groups can also be found on social media [
Belonging to health-related groups on social media has been linked to several benefits for users, including a reduction in feelings of isolation [
However, although active participation and engagement in diabetes-specific social media groups is mostly linked to positive effects, very little is known about the content that is shared on these channels or what features of posts engage their users the most. In a previous study [
One way of assessing if posted messages engage users of diabetes-related social media is by measuring the posts’ received feedback in the form of likes, shares, and comments. Likes and shares are a form of communication that allows social media users to provide feedback to other users with a simple click [
The objective of this study was to analyze the content topic and features of posts that were shared over 3 years in the 3 diabetes-related social media channels (Facebook, Twitter, and Instagram) of the Norwegian Diabetes Association, as well as the users’ engagement with these posts. This study is part of a participatory research project on the use of social media for health promotion in diabetes [
All social media posts from the Norwegian Diabetes Association (on Facebook, Twitter, and Instagram) published between January 1, 2017, and December 31, 2019, were extracted and included in the study (no posts were excluded or removed from the analysis). The social media posts were extracted using the manager tool for Facebook, manually for Instagram, and using the standard application programming interface (API) for Twitter. Using the appropriate functions, the standard API allows us to gather all tweets from a specific timeframe. In this way, a PHP script was programmed to query Twitter for all the tweets made by the Norwegian Diabetes Association during the study’s timeframe. The Twitter data, including the text and the tweet metadata (ie, date of publishing, likes, retweets, etc), were then exported into a Microsoft Excel file document. The following information was extracted from each post: text message. post features (ie, use of emoji, picture, and/or video), and number of likes, comments, and shares. For Facebook, we collected the total number of likes, including the reactions, for each post. For Instagram, we only extracted likes and comments because shares were not an available option at the time of the study.
We classified the content topics into 7 categories. These categories were based on findings from our previous studies [
Two independent reviewers classified the text message of each post according to its main topic. When a post was considered to fall into more than one category, reviewers were trained and instructed to choose the main topic among the 7 possible options. Discrepancies in the posts’ classification were discussed with a third reviewer until reaching consensus. The inter-rater reliability was calculated using Cohen kappa analysis.
All descriptive and regression analyses were performed using SPSS software (version 25; IBM Corp). The dependent variables in the regression analyses were the number of likes, comments, and shares, which were count data and non-normally distributed. Negative binomial regression models emerged as most appropriate based on the overdispersion parameters and the goodness-of-fit indices. For each of the dependent variables, we performed multilevel negative binomial regression, with the predictors being the independent nominal variables. The largest category in each group was used as the reference: content topic (reference group: miscellaneous), social media channel (reference group: Twitter), and post features (reference group: text only). We determined the interaction between the content topic, social media channel, and post features. The level of significance was set at
The study protocol was exempted from requiring ethical approval by the Norwegian Regional Ethics Committee (2017/764/REK Sør-ØstC), as it falls outside the scope of the Norwegian Health Research Act. The treatment of personal information was approved by the Data Protection Officer at the University Hospital of North Norway (ref 0720).
During the 3-year period of the study, the Norwegian Diabetes Association posted a total of 1449 messages on their social media channels: 718 (49.55%) were posted on Twitter, 530 (36.58%) were posted on Facebook, and 201 (13.87%) were posted on Instagram. The number of posts on each social media channel according to post features (text only, inclusion of picture, video, and/or emoji) is summarized in
Number of posts on 3 social media channels according to post features.
|
|
Without emoji | With emoji | Total |
|
410 | 120 | 530 | |
|
Text post | 247 | 73 | 320 |
|
Text + picture | 105 | 30 | 135 |
|
Text + video | 58 | 17 | 75 |
|
698 | 20 | 718 | |
|
Text post | 455 | 8 | 463 |
|
Text + picture | 225 | 11 | 236 |
|
Text + video | 18 | 1 | 19 |
|
15 | 186 | 201 | |
|
Text post | N/Aa | N/A | N/A |
|
Text + picture | 12 | 169 | 181 |
|
Text + video | 3 | 17 | 20 |
Total | 1123 | 326 | 1449 |
aN/A: not applicable
The interrater agreement of the posts’ main topic was found to be substantial (κ=0.695, κ=0.780, and κ=0.789, for Twitter, Facebook, and Instagram posts, respectively) [
Number of diabetes-related posts on 3 social media channels according to content (January 1, 2017, to December 31, 2019).
The effect of content topic, social media channel, and post features on the measures of users’ engagement was analyzed using negative binomial regression.
The regression analysis showed that posts of interviews and personal stories received 111% more likes, 106% more comments, and 112% more shares than miscellaneous posts (
With regard to social media channel, Facebook posts were 35 times more likely than Twitter posts to receive likes, 60 times more likely to receive comments, and 13 times more likely to be shared (
In terms of post features, posts that included videos were 3 times more likely to receive likes, almost 4 times more likely to receive comments, and 2.5 times more likely to be shared than text-only posts (all
Effect of content topic, social media channel, and post features on users’ engagement (likes, comments, and shares).
Independent variables | n | Likes | Comments | Shares | |||||||
ORa (95% CI) | OR (95% CI) | OR (95% CI) | |||||||||
|
|
|
|||||||||
|
Interviews and personal stories | 124 | 2.11 (1.66-2.68) | 2.06 (1.37-3.08) | 2.12 (1.58-2.96) | ||||||
Awareness days and other celebrations | 158 | 1.41 (1.12-1.76) | .002 | 1.31 (0.91-1.89) | .15 | 1.23 (0.88-1.70) | .22 | ||||
Recipes and food-related information | 99 | 0.53 (0.40-0.70) | 0.41 (0.25-0.65) | 0.57 (0.35-0.92) | .02 | ||||||
Diabetes-related technology | 84 | 0.78 (0.59-1.04) | .09 | 1.33 (0.83-2.14) | .23 | 0.73 (0.49-1.06) | .10 | ||||
Research and innovation on diabetes | 207 | 0.69 (0.57-0.85) | 0.65 (0.45-0.93) | .02 | 0.75 (0.58-0.97) | .03 | |||||
Health education | 260 | 0.87 (0.72-1.04) | .13 | 0.62 (0.45-0.85) | .003 | 1.39 (1.09-1.78) | .01 | ||||
Miscellaneous | 517 | 1b | 1b | 1b | |||||||
|
|||||||||||
530 | 35.41 |
60.37 |
12.79 |
||||||||
201 | 34.99 |
9.68 |
N/Ac | ||||||||
718 | 1b | 1b | 1b | ||||||||
|
|||||||||||
Emoji (no picture, no video) | 81 | 1.71 (1.29-2.26) | 1.09 (0.69-1.70) | .70 | 2.44 (1.69-3.51) | ||||||
Emoji and picture | 210 | 1.30 (0.96-1.77) | .09 | 0.65 (0.37-1.12) | .12 | 1.33 (0.80-2.19) | .27 | ||||
Emoji and video | 35 | 6.83 (4.21-11.09) | 1.29 (0.66-2.56) | .45 | 1.82 (0.91-3.65) | .09 | |||||
Picture (no emoji) | 342 | 1.86 (1.58-2.19) | 1.12 (0.83-1.49) | .46 | 2.24 (1.81-2.77) | ||||||
Video (no emoji) | 79 | 3.29 (2.26-4.39) | 3.87 (2.37-6.31) | 2.50 (1.73-3.61) | |||||||
Text only (no emoji, no picture, no video) | 702 | 1b | 1b | 1b |
aOR: odds ratio.
bReference group in the corresponding independent variable.
cN/A: not applicable.
Between January 1, 2017, and December 31, 2019, the Norwegian Diabetes Association posted a total of 1449 messages on their social media channels. The posts that were most engaging to users were those that featured interviews and personal stories. In fact, those posts were twice as likely to receive likes, comments, and shares. On the other hand, posts containing recipes and food-related information, and posts on research and innovation, were the least engaging to social media users. Both types of content were less likely to receive likes, comments, and shares. Regarding the social media channel, Facebook and Instagram posts were both 35 times more likely to receive likes, and 60 times and almost 10 times more likely to receive comments, respectively, than posts on Twitter. Video and emoji were the most engaging post features. Posts with video had 3 times increased chance of receiving likes, almost 4 times increased chance of receiving comments, and 2.5 times increased chance of being shared. Including both video and emoji increased the chances of receiving likes by almost 7 times. The addition of an emoji to a post increased its chances of receiving likes and being shared by 71% and 144%, respectively.
We found that diabetes social media posts that engaged the most users were the ones that featured interviews and personal stories, and those that mentioned awareness days and other celebrations. Our results indicate that posting content based on interviews and personal stories on diabetes social media channels offers the highest chances of receiving likes, comments, and shares. This finding contrasts with the results of our previous survey [
Posts containing content about awareness days and other celebrations also showed a higher number of likes. Similarly, in a previous study [
Diabetes social media posts dealing with content that promotes empowerment have been previously linked with higher engagement [
The use of videos predicted higher chances of receiving likes, comments, and shares. The inclusion of pictures and emoji also predicted an increased number of likes and shares. Our results are in concordance with previous publications [
Our results on the effect of pictures on engagement are also in line with the findings of a previous study [
Disease-specific social media groups, such as the one we analyzed in this study, are recognized as trusted sources of information. Patient associations on social media reach and engage a considerable number of people, which can potentially benefit their users at many levels, including with respect to health outcomes [
Analyzing users’ engagement with social media posts is a way for patient associations, healthcare authorities, and other stakeholders to understand the voice of the patient [
This study refers to diabetes social media groups led by a national patient association. Although the channels are open, only social media administrators within the patient association are able to post. Users are only able to respond to these posts. The type of content posted by this organization might differ from that posted in other diabetes social media groups.
Diabetes social media users seem to be least engaged in post content that a priori could be linked to greater empowerment: research and innovation on diabetes, and health education. Diabetes social media groups, public health authorities, and other stakeholders interested in sharing research and innovation content and in promoting health education that engages social media users should consider including videos and emoji in their posts, and preferably publish on popular and visual-based channels, such as Facebook and Instagram, to increase user engagement.
application programming interface
We sincerely thank the Norwegian Diabetes Association (Diabetesforbundet) for their engagement in this study. This study was funded by Northern Norway Regional Health Authority (Helse Nord RHF) grant number HNF1370-17. ED receives funding and is supported by the V Plan Propio de Investigación de la Universidad de Sevilla, Spain.
EG, RW, and EÅ designed the study. EG and ED extracted the data. EG, DL, and PEH classified the posts. EG and DL analyzed the data. EG took the lead in writing the manuscript and all authors provided critical feedback. All coauthors have revised the manuscript and approved its final version.
None declared.