Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Facebook as communication support for persons with potential mild acquired cognitive impairment: A content and social network analysis study

  • Aboozar Eghdam ,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Resources, Software, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliation Health Informatics Centre (HIC), Department of Learning, Informatics, Management and Ethics (LIME), Karolinska Institutet, Stockholm, Sweden

  • Ulrika Hamidi,

    Roles Data curation, Formal analysis, Software, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliation Health Informatics Centre (HIC), Department of Learning, Informatics, Management and Ethics (LIME), Karolinska Institutet, Stockholm, Sweden

  • Aniko Bartfai,

    Roles Conceptualization, Investigation, Methodology, Resources, Supervision, Validation, Writing – original draft, Writing – review & editing

    Affiliation Division of Rehabilitation Medicine, Department of Clinical Sciences Danderyd Hospital, Karolinska Institutet, Stockholm, Sweden

  • Sabine Koch

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Resources, Supervision, Validation, Writing – original draft, Writing – review & editing

    Affiliation Health Informatics Centre (HIC), Department of Learning, Informatics, Management and Ethics (LIME), Karolinska Institutet, Stockholm, Sweden



Social media has the potential to increase social participation and support for the well-being of individuals with chronic medical conditions. To date, Facebook is the most popular social medium for different types of communication. However, there is a lack of knowledge about the potential use of Facebook as a means of communication for persons with potential Mild Acquired Cognitive Impairment (MACI), a non-progressive mild cognitive impairment after an acquired brain injury. The aim of this study was to explore how persons with potential MACI, specifically persons with perceived brain fatigue after brain injury, communicate through Facebook, to classify the content of the communication and to visualize the frequency and types of interactions.

Methods and materials

A social network analysis of the interactions between members’ and a qualitative content analysis of a whole year’s communication of a public Facebook group for Swedish speaking persons (1310 members) with perceived brain fatigue after an illness or injury to the brain were performed.


The results showed how members use social media technology and Facebook as a means for communication and support for their condition. Individual group members showed very different patterns of communication and interactions. However, for the group as a whole, the most frequent topics in their communication were related to informational support and banter in posts, and socialization in comments. The findings also showed that the majority of members only communicated with few other members and had few direct communications. The most used communication feature of Facebook was likes in form of “thumbs-up”.


This study indicated that social media and in this case Facebook is used for communication and social support by persons with potential MACI, and revealed that their communication behavior is similar to the healthy population. Further studies relating specific cognitive problems of the participants to the use of social media would provide more reliable results for this specific group.


One of the leading causes of disability in many developed countries is brain injury [1]. Mild Acquired Cognitive Impairment (MACI) is a term used to describe persons with non-progressive Mild Cognitive Impairment (MCI) acquired from a brain injury. They can have multiple cognitive (e.g. memory impairment) and/or somatic (e.g. headaches, fatigue) in addition to mild physical disabilities and their access to rehabilitation services to improve their affected skills is limited [2]. One approach that can be considered for further developing treatment of MACI is information and communication technology (ICT) and based on previous findings from Sweden [3], the most used types of ICT services are communication services. In Sweden over 70,000 persons suffer from brain injuries every year and the most common causes are stroke and head injuries [4]. About 80% of these persons suffer from mild impairments[4].

A serious illness affects various aspects of an individual’s life such as communication, self-care, social relationships, employment and mobility [5]. A person with a chronic condition is in great need of different kinds of social support on a long-term basis [6]. Previous research has shown that social networks in general may improve disease management for persons with chronic conditions, suggesting that online social networks have a great potential to improve health care [710]. Therefore, one of the biggest benefits of online social media is social support.

Internet and social media

The use of the Internet and especially social media as a means for communication has increased over the past decades [11]. Social media has changed how health consumers and health care providers communicate [11,12]. McGowan and coworkers have shown that the use of social media in health care may be an efficient way to share up-to-date medical knowledge with others within a community and improve the quality of patient care [13]. Due to the increase of interest in online social support, many researchers have been investing in understanding the contents, motivations and the effects from their use and the interest in interactive consumer health technology applications has also increased [14,15].

Today, gaining knowledge and sharing experiences with others in similar situations are common phenomena in social media. Based on the results of a national survey that we have conducted recently, one of the most important types of support for persons with mild brain injury is communication services and especially “social interaction and relationships with families and friends” [3]. Sweden is one of the top countries among European individuals. According to Eurostat, in 2016 about 94% of Swedish households have access to the Internet and 91% of the population use the Internet at least once a week [16].

Facebook is the largest social networking site in the world with 1.65 billion monthly active users on average from which about 92% access it through mobile devices [17]. Generally, people use Facebook as a means for communication with relatives and friends and other people with similar interests [18]. For people with common interest, the group function is useful to share their opinions and communicate on the topic or purpose of the group. Users can join or create new public or private groups and post or share content on the group’s wall. The group’s privacy settings are controlled by ‘‘administrators” who have the ability to manage the content in addition to invite or approve the membership of new users. The content of public groups is visible to any user of the Internet even to any non-Facebook user.

Social media for supporting patients with mild acquired cognitive impairment

Communicating with people in similar situations is helpful to cope with daily life as it reduces feeling of loneliness and isolation when dealing with a rare condition [19]. Generally, sharing experience of having the same condition as informational support in addition to socializing seems to be beneficial for persons with chronic conditions [20]. There is a socializing need for persons with MACI but since performing physical socializing requires so much energy due to problems with concentration and loosing track in a conversation, social activities appear to be more difficult in real-life situations. For persons with MACI, physical socializing decreases after brain injury and people get lonelier but the need to socialize still remains [21]. Therefore, social media would be an ideal solution since its use can be regulated depending on the individual’s condition and at one’s own pace. The long distance between individuals with MACI may hinder face-to-face conversations but the online communication would ease general accessibility and increase inclusiveness [22].

Today, many disease-specific groups exist on Facebook, people share various health issues in this platform [23]. and it became popular as a support-seeking- providing tool for patients with breast cancer and HIV [24,25]. For persons who are dealing with their situation by themselves, in addition to socializing, social media could be used for gaining knowledge from others as well as sharing experiences [26]. These kinds of social interactions among persons with a similar condition could also be beneficial for the social well-being of persons with MACI. Based on our previous research, the amount of persons with MACI who use Facebook is rather remarkable despite the reported cognitive problems [3]. Given the higher risk for a significant decrease in their friendships and social support and due to the large number of persons with mild brain injury, providing face-to-face social support may be problematic as a result of limitations in mobility, access, or communication. Therefore, online support groups may be a particularly useful alternative for these individuals [6].

Despite the potential value of online support, very little is known about how people with different health issues use Facebook groups and there is a lack of research examining social networking sites in the context of MACI [27,28]. To our knowledge, no studies on the use of Facebook groups as a means of providing social support for persons with brain injuries have been published. This study aims to explore how persons with potential MACI, specifically persons with brain fatigue after brain injury communicate through Facebook, to classify the content of the communication and to visualize the frequency and types of interactions.

Methods and materials

The main methods for this study were content and social network analysis of communication within a Facebook group for Swedish speaking persons with brain fatigue after brain injury.

All content of the Facebook group’s timeline between September 1, 2014 until August 31, 2015 was collected. In October 2015, the group consisted of 1310 members. To reduce the ratio of potential missing data in the collected data set and to avoid seasonal effects on individuals’ behavior, the time window was set to a whole year [29]. The data was collected from users’ interactions in the group such as posts, written comments to posts or acknowledged “likes” to posts.

The researchers collecting the data presented themselves in the Facebook group as “researchers who would observe the group”. Individual informed consent was not acquired as the data was collected retrospectively. The approval of the study was obtained from the Regional Ethics Review Board (EPN) in Stockholm (2015/1284-31/5 October 01, 2015). The Regional Ethics Review Board (Regionala etikprövningsnämnden— has the task to vet cases within the field of medical science (medicine, pharmacology, odontology, the science of health care and clinical psychology). The secretariats of the regional boards are situated at the different universities in Sweden.

Data collection

In this study Netvizz v1.25 was chosen as the data collection software. Netvizz is a Facebook application (data access via the Facebook API) that allows researchers to collect anonymous data from open (Public) Facebook groups [30]. The data collected by Netvizz was obtained in comma separated value (CSV) file and graph data file (GDF) format. This data collection was performed without any interaction with the Facebook group’s members and no attributes from the participants´ Facebook profiles (e.g. name, usernames, email, gender, etc.) were collected. Regarding the anonymization and to have unique identifiers, the participants were given automated anonymous Ids by the data collection software. The CSV files contained the texts and symbols in the posts and comments. The GDF files had tables with relational data on the members´ involvement in communicative-interactions.

Data analysis

Content analysis.

To understand the members’ contributions to the Facebook group, contents of the posts and comments were analyzed using a direct content analysis approach in which the initial starting categories and definitions for the themes were retrieved from literature (Table 1) [31]. The content analysis began with repeated reading of all posts and comments by the first and second authors. Basically the content was categorized into social supports and others. For social supports the posts and comments were grouped based on the initial codes retrieved from typology by Cutrona and Suhr’s [25,32]. The five broad categories for social support according to the Social Support Behavior Code (SSBC) developed by Cutrona and Suhr are defined as “informational”, “emotional”, “esteem”, “network”, and “tangible support” [32]. Table 1 shows the Cutrona and Suhr definitions for social support categories.

Table 1. Definitions of social support categories according to [32].

For other categories, the initial codes were retrieved from the Gaysynsky et al. study about analyzing a Facebook group for young adults living with HIV [25]. The definitions for other categories are shown in Table 2.

During the content analysis process additional categories emerged that were classified by the authors of this paper. (Table 3)

Table 3. Definitions of additional categories as defined by the authors.

In addition to the posts and comments, the likes were counted. Furthermore, the context of a particular post or comment (i.e. content that precedes or follows the message of interest) was used to interpret the intention of the message, categorization and to see if the group member was seeking or providing support.

Microsoft Excel was used as a tool for organizing and coding the collected data. The first and second authors performed the analysis process and validated (to evaluate the correctness of the assigned categories) through discussion with the other authors.

Social network analysis.

One method for measuring relations, associations and flows between different information and knowledge entities that can reveal patterns of the activities within social media groups is social network analysis (SNA) [33,34]. Generally, a SNA map consists of nodes (persons) that are linked to show the relationships and flows (information) between them.

A social network analysis was performed to analyze and visualize the distribution of the group members’ involvement in information exchange. In the Facebook group, participants are conceptualized as actors (nodes) and their comments and likes to each other’s posts are incorporated as the relations between them. In a social network, the nodes with more relations to others, are less dependent and more influential. Therefore, they may have alternative ways to satisfy needs [35,36]. The number of nodes that are in relation with one node, is defined as degree centrality and the number of interactions (comments on, or likes to posts) with each node is defined as weighted-centrality. Also, the comments and likes around the posts between two unique nodes are defined as edges. In this study, the location/degree centrality of nodes is evaluated and visualized to provide an understanding of the network and its members.

In order to provide the drawing of social network graphs as well as calculations of social network concepts, Gephi 0.8.2 beta software was used. Gephi is an open-source software that provides the opportunity to interact with the graph drawing algorithms [37]. Gephi provides several force-directed algorithm alternatives and visual representations of the data. The decision on choosing the graph drawing algorithm is a subjective task but it is advised to try all of them to be able to present the most comprehensible graph [38]. For constructing and to be able to modify the density of the visualization setting of the graph, the ForceAtlas2 algorithm was used. This algorithm was continuous and it was stopped when relatively stable and clear patterns were perceived.


During the data collection the contents that appeared on the Facebook group’s timeline one-year period were collected. Fig 1 shows the numbers and frequency of posts and comments during this one-year time period. The maximum amount of posts and comments occurred during May 2015. Within the time window 1092 (83%) unique active members’ contributed to the Facebook group with 630 posts, 4323 comments and 10187 likes on posts. The amount of self-comments (comments on own posts) was 23%. 218 members did not have any interaction in the group.

Fig 1. Number of posts and comments during the study period.

Content of the communication

Content analysis was applied to all posts and comments. Table 4 shows the incidence of the main interaction categories for posts and comments.

Table 4. The incidence of posts and comments based on main social support categories.

The majority of posts (57%) and comments (62%) were categorized as different types of social support. From 362 posts in the social support categories, most of them (72%) and a large amount of comments (43%) were peer to peer exchange of “informational” social support. Only 17% of social support posts were seeking support and 40% were providing support. Regarding the other categories, “banter” in posts and “socializing” in both posts and comments were the most frequently applied codes (excluding self-comments) (Table 4).

The informational social support comments and posts were mostly about the members’ health related problems or sharing an experience of having a health condition and strategies about how to deal with it (Table 5). Despite clinical differences, expressing their experiences made them feel less alone since they had a common empathic feeling. However, in some cases, the group members lacked the required medical knowledge to address specific questions or issues (such as adequate medication). In general, the group platform had an open learning atmosphere. As shown in Table 4, advertisement and negative interaction were quite rare in this group. The members had very few posts and comments coded as negative interactions (e.g. benefits of narcotic drugs).

Table 5. Exemplifying quotations from posts and comments.

Frequency and types of interactions

The social network analysis provided a clear visualization of how members were connected in this group and the nodes were divided based on their level of connectivity (number of connected nodes) (Fig 2 and Table 6). In this study only direct connections between different nodes were considered and based on their interaction types, different groups of nodes were identified (Figs 3, 4, 5 and 6).

Fig 2. The whole network graph.

The size of each node represents the degree centrality and the darkness indicates weighted-centrality. The marked nodes A-R are well-connected and very-well-connected nodes in the network. The x and y nodes are two other active nodes who just made “comments” and “likes”.

Fig 3. The colored nodes shows the members who were involved in publishing posts (16% of the nodes).

The size of each node represents the degree centrality. The marked nodes A-R are well-connected and very-well-connected nodes in the network. The x and y nodes are two other active nodes who just made “comments” and “likes”.

Fig 4. The colored nodes shows the members who were involved in making comments (49% of the nodes).

The size of each node represents the degree centrality. The marked nodes A-R are well-connected and very-well-connected nodes in the network. The x and y nodes are two other active nodes who just made “comments” and “likes”.

Fig 5. The colored nodes shows the members who were only making comments (5% of the nodes).

The size of each node represents the degree centrality. The marked nodes A-R are well-connected and very-well-connected nodes in the network. The x and y nodes are two other active nodes who just made “comments” and “likes”.

Fig 6. The colored nodes shows the members who were only making likes (48% of the nodes).

The size of each node represents the degree centrality. The marked nodes A-R are well-connected and very-well-connected nodes in the network. The x and y nodes are two other active nodes who just made “comments” and “likes”.

Table 6 shows the division of nodes based on the level of connectivity and different interaction types. In this Facebook group only 18 nodes had more than 100 connections and most nodes were sparsely-connected to 3–10 other nodes followed by single and intermediate-connected nodes (Table 6). Also the nodes with higher degree centrality differed regarding their types of contributions (posts, comments and likes) compared to some of the single, bi and sparsely-connected nodes who had mainly a liker or a commenter personality.

As shown in the Fig 2, the graphs generated from the social network analysis were very nested and the edges between different nodes were difficult to distinguish. However, the density element was not high (0.011 in 0–1 range representing the complete connectivity between the nodes) due to incomplete connectivity within the whole group members. This also showed that only a few nodes were in contact with many others and there were many nodes who were only communicating with one or a few others. Fig 7 shows the degree centrality and Fig 8 weighted-centrality distributions in the group.

Fig 7. The degree centrality distribution for the whole network.

Fig 8. The weighted-centrality distribution for the whole network.

To understand the nature of active nodes’ contributions to the group, and different personalities, nodes that had the most contacts in the very well and well-connected categories were marked in the graphs as A-R. Also two other nodes who had an interesting behavior were marked as x and y (Fig 2). The node x was an intermediate-connected node who just made comments and likes towards others without having any posts. The node y’s contribution to the group was only likes. Table 7 shows the details of active nodes’ contribution to the group.

Table 7. The numbers of post, comment, likes, connected nodes and directed interactions for selected nodes.

The social network analysis showed that the active nodes had their own single-connected followers in addition to other multi-connected nodes (Fig 9). During identification of the most connected nodes, to understand the interaction between other nodes, the very well–and well connected nodes were removed to apprehend the network with less active nodes (Fig 10). As shown in Fig 10, the network between single, bi, sparsely and intermediate nodes is also pretty dense even without active nodes.

Fig 9.

The proportion of network connected to the nodes A, B and C (colored). The size of each node represents the degree centrality. The marked nodes A-R are well-connected and very-well-connected nodes in the network. The x and y nodes are two other active nodes who just made “comments” and “likes”.

Fig 10. The network nodes and their edges excluding active nodes.

The size of each node represents the degree centrality. The x and y nodes are two other active nodes who just made “comments” and “likes”.

In this group, three different types of nodes were identified. The nodes who mostly post in the group, nodes who get into conversation by commenting on the posts and others’ comment, in addition to nodes who just like others’ posts and comments. However, the majority of the nodes (43%) were identified as commenters and likers during the data collection period.

According to Table 7, A, B and C nodes had the highest number of posts in the group in addition to having the highest degree centrality and weighted-centrality in the group (very-well-connected nodes) and as shown in Fig 9, i.e. they have been in contact with a much larger part of the network than the others. Although A, B and C had published the highest amount of posts, they did make highest number of comments and likes per individual. However, they mostly responded to other nodes’ posts than their own published materials. The M, x, D and L nodes made highest amount of comments excluding self-comments. Despite having a small size in the graph, the node M was the most frequent commenter in the group. The x node was chosen to be among this sample due to having the high number of comments but no post or like contribution to the group. This node was only addressing other nodes by responding to already published content instead of posting to the whole group. The most frequent likers in the group were M and y. Regarding receiving comments and likes in posted content, the most frequent receivers of comments were, in order, nodes C, A and H. The node H, was the second most frequent receiver of comments by having very few posts in the group (Table 7).

Fig 11 shows that many of the most frequent producers and receivers of comments or “likes” were communicating with each other. The thickness of the lines between different nodes shows the edges between them. This figure shows that the information exchange between M and A was the most frequent in the whole network. Interestingly the node H was communicating with a group of single-connected nodes which were not in contact with other active nodes. Furthermore, the exclusion of the active nodes (A, B, C, D, H, M, L, x and y) also resulted in a highly nested network (density 0.007) demonstrating a good communication within the network.

Fig 11. The selected active nodes and their edges.

The size of each node represents the degree centrality and the darkness indicates weighted-centrality.

With the intention of getting deeper into the active nodes’ contributions, the top posters and commenters were selected to analyze the semantics of their posts (Table 8), and made and received comments without considering self-loops (comments on their own posts) (Table 9). Most of the post (except “Emotional support”) were providing support.

Table 8. The content analysis of active members’ contributions regarding seeking and providing support categories, identified in the social network analysis.

Table 9. The content analysis of active members’ contributions considering different social support categories, identified in the social network analysis.

Bearing in mind the nodes with the highest number of published posts (A, B and C), their contributions were the most liked in the network and they typically responded to posts from other members. Moreover, in addition to making the highest number of likes, these nodes received the highest number of likes in different support categories. Remarkably, A and B had fewer informational posts than other active nodes in addition to the whole group. Also the C node had a similar amount of posts in different category to the whole group. Regarding influential posts, the H node had a few seeking emotional support posts which received the highest amount of comments per post in the network. None of the active nodes were involved in negative interaction. Also the most frequent commenters’ (M and x) contribution to the group had a similar theme to the whole group overall comments’ categories. Moreover, the C node received the highest number of comments in the “Socializing related” category.


The results of this content and social network analysis concerning persons with brain fatigue after acquired brain injury showed their capacity of using Facebook as a means for communication and support for their condition. However, the lack of data about the participants’ cognitive and general health problems as well as the insecurity about the real identity of persons behind Facebook profiles, needs to be considered.

Indicating the same pattern as other studies for typical Facebook users, the members of this Facebook group acted similar as other Facebook users in general by using features such as likes on posts (as the most common communicative interaction) followed by comments and publishing posts [39,40] (Figs 3, 4, 5 and 6). Group members used posts to address the general public, and comments, to address specific group members. Obviously, the interaction patterns were dependent on the Facebook platform features, thus, the results would not show how this group will use other types of social media such as Twitter, YouTube etcetera [41].

Based on the results (Table 4) and specifically the number of likes per posts, providing any type of support in posts and comments was appreciated by the group members. Similar to other online forums for patients with other health conditions, informational support was a common contribution [33,42]. In addition to providing informational support, banter in posts, and socialization in comments were also quite popular (Table 4). Usually, the informational support would be beneficial for all members and non-members of a group (even the ones who are not a member of the same group) [20] and the provider, will be encouraged and supported by other members’ likes and comments. Regarding the access to information about health, the results replicated the findings by Bunner et al. showing social participation as common reason to use social media for individuals with traumatic brain injury [20]. In this groups, the amount of socializing, non-verbal cues, expression of gratitude and offering congratulation showed that members would appreciate the contact with people in similar situations. Furthermore, the publicness of the group was discussed between group members and some of the members indicated that it should be public to improve the understanding of their condition the among general public. This preference shows social media (in this case Facebook) could be a means to provide a better understanding of persons with specific clinical conditions for the general public in addition to all other potential benefits for its’ own members.

Based on the results (Figs 7 and 8), the degree centrality and weighted-centrality followed a long tail distribution, indicating that the majority of members only communicated with few others and had few direct communications. Bearing in mind the likes and comments per posts, the results also showed that the popularity of active members was dependent on the frequency of communications. Obviously, the members with a higher number of posts had higher degree centrality and weighted-centrality in the network. This finding is in line with other studies on Facebook groups over 1000 members [43]. Unfortunately, the Facebook platform does not show passive members’ activities within the group. In addition, the comment on comment feature on Facebook was not activated during the time of this study.

The results also showed different patterns in content and interaction of a few active individual members of this group. Considering the association of different personalities with self-disclosure and social orientation [44], the members showed different personalities in terms of liking, commenting and publishing posts. However, since we did not have access to their clinical data, no conclusion can be drawn regarding phenomenon being due to the brain injury or a personality trait. Generally, in a real-life support group, these active attention-seeking individuals could take up a lot of space, monopolize other people’s time and use the group for their own purposes [45]. However, in our study removing the active members resulted in a nested network of interactions indicating that if a member did not like one or more active members, they could easily interact with other members and ignore active members and their contribution.

Despite the association of the content analysis results in our study with previous research, none of them focused on different categories in posts and comments nor the number of likes per category. Given the popularity of Facebook for different persons with special health conditions, our methodology could provide a new way to get deeper into unexplored social media features.

Certainly, social media (in this case Facebook) would be a great means of communication and social support for helping persons with MACI. Moreover, most of them with basic knowledge of using ICT tools would be able to use this medium easily on different devices.

Implication/trustworthiness and patients’ advocacy

Since social contact decreases after cognitive disturbances [21], persons with brain fatigue after brain injury appreciate the information support in communication. This might have an association with the general use of social media which has to be followed in further research. However, it has been known that patients suffering from a mild brain injury show a lower occurrence of mental fatigue symptoms when perceiving social support [46].

In this study, the accuracy and benefits of the exchanged information have not been analyzed. However, the potential risks of inaccurate information and unhealthy behavior are substantial. Noticeably, providing the informational support by most frequent publishers, was highly appreciated and received a high number of likes. Based on previous research, people with cognitive problems may have a tendency to judge information based on being familiar with the source rather than the trustworthiness of it [47]. Therefore, the probability of trusting information provided by active members might be higher than by other information sources. As shown in Fig 9, the A, B and C nodes were in contact with a big portion of the members and nearly covered the whole network. Given the trustworthiness of these members, a suggestion for the healthcare system and patient organizations would be to identify the active users through a social network analysis and voluntarily provide them with basic training to moderate or monitor the group in a way that other members benefit from the knowledge of the effects of brain injury. This idea has to be further explored with clinicians to see the burdens regarding choosing a patient ambassador.


A major concern was to establish the representability of this Facebook group compared to the clinical population. Generally, recognizing mild cognitive impairment is problematic even in elderly populations. Therefore, the manifestation of cognitive impairment is relatively unknown for individuals who are suffering from it and also for healthcare professionals [48,49]. After discussing this issue with MACI rehabilitation professionals’ and medical experts, it seemed that for this type of studies, it is not feasible to achieve a higher level of evidence concerning pathology of the participants in this group. It would be applicable to find the target group in laboratory science but not in the real world. Therefore, this new research area and the methodology should be investigated more. In addition, further studies are needed to replicate this study’s results with persons with similar and different chronic conditions in other Facebook groups to show any disease specific behavior. One of the main limitations of this study was that the data collection platform Facebook does not provide any information about all potential members of the group and their activities (rather than likes, comment and posts), and that the directions of the information (social support) flows were not included. Moreover, the reliability of the results could be biased since this study did not provide any information about the cognitive failure levels of individuals behind the Facebook profiles and their health in general.

There are two possible sources of systematic error in the study: 1) the inclusion of members with cognitive problems due to other than somatic disease and 2) inclusion of non-serious members. However, based on the content analysis, the group administrators excluded supposedly non-serious individuals from the group based on other members questioning their behavior. It must be emphasized that this study was performed on a Facebook group for Swedish speaking persons with brain fatigue and cultural influences might have affected the interaction between the group members. Based on participants’ contribution to the Facebook group and support seeking categories, they had common symptoms with persons with MACI but clinical data is missing [3]. Further studies with more controlled variables would probably provide more reliable results for this specific group.

In this study the intention was to investigate different actors’ behavior in such a community and the majority of the participants had low degree centrality but due to practical and technical reasons, the focus was on participants with the highest degree centrality on individual levels and the behaviors of participants with low degree centrality was not examined. It would be interesting to explore and visualize all members’ activities in terms of themes and content of contributions to the group.


The aim of this study was to explore the behavior in social media for persons with potential MACI. The results of the content and social network analysis of a public Facebook group addressed the lack of knowledge about the potential use of social media as a means for communication for persons with acquired brain injury and potential MACI.

The results showed different patterns in themes and activities of different individual members compared to the whole group. The most used communication feature of Facebook was likes in form of “thumbs-up” and the most common themes of the members’ contribution in this group were characterized by informational support and socializing. The results of this study showed that the communication behaviors of persons with potential MACI could be similar to the healthy population. However, further studies considering also the cognitive failure levels of individuals would provide more reliable results for this specific group.


  1. 1. WHO. The top 10 causes of death [Internet]. World Health Organization, Geneva. World Health Organization; 2015. Available:
  2. 2. Lin F, Gleason CE, Heidrich SM. Illness representations in older adults with mild cognitive impairment. Res Gerontol Nurs. NIH Public Access; 2012;5: 195–206. pmid:22716654
  3. 3. Eghdam A, Bartfai A, Oldenburg C, Koch S. How Do Persons with Mild Acquired Cognitive Impairment Use Information and Communication Technology and E-Services? Results from a Swedish National Survey. PLoS One. 2016;11: e0159362. pmid:27427947
  4. 4. Hjarnkraft. Hjärnskador [Internet]. 2015 [cited 29 Sep 2015]. Available:
  5. 5. Braithwaite DO. From majority to minority: An analysis of cultural change from ablebodied to disabled. Int J Intercult Relations. Pergamon; 1990;14: 465–483.
  6. 6. Braithwaite DO, Waldron VR, Finn J. Communication of Social Support in Computer-Mediated Groups for People With Disabilities. Health Commun. Taylor & Francis Group; 1999;11: 123–151. pmid:16370973
  7. 7. Christakis NA, Fowler JH. The Collective Dynamics of Smoking in a Large Social Network. N Engl J Med. Massachusetts Medical Society; 2008;358: 2249–2258. pmid:18499567
  8. 8. Christakis NA, Allison PD. Mortality after the Hospitalization of a Spouse. N Engl J Med. Massachusetts Medical Society; 2006;354: 719–730. pmid:16481639
  9. 9. Sayers SL, Riegel B, Pawlowski S, Coyne JC, Samaha FF. Social Support and Self-Care of Patients with Heart Failure. Ann Behav Med. Springer-Verlag; 2008;35: 70–79. pmid:18347906
  10. 10. Thoren EM, Metze B, Bührer C, Garten L. Online support for parents of preterm infants: a qualitative and content analysis of Facebook “preemie” groups. Arch Dis Child Fetal Neonatal Ed. BMJ Publishing Group Ltd and Royal College of Paediatrics and Child Health; 2013;98: F534–8. pmid:23759520
  11. 11. boyd danah m., Ellison NB. Social Network Sites: Definition, History, and Scholarship. J Comput Commun. Blackwell Publishing Inc; 2007;13: 210–230.
  12. 12. Moorhead SA, Hazlett DE, Harrison L, Carroll JK, Irwin A, Hoving C. A new dimension of health care: systematic review of the uses, benefits, and limitations of social media for health communication. J Med Internet Res. JMIR Publications Inc.; 2013;15: e85. pmid:23615206
  13. 13. McGowan BS, Wasko M, Vartabedian BS, Miller RS, Freiherr DD, Abdolrasulnia M. Understanding the factors that influence the adoption and meaningful use of social media by physicians to share medical information. J Med Internet Res. 2012;14: e117. pmid:23006336
  14. 14. Walker KK. A Content Analysis of Cognitive and Affective Uses of Patient Support Groups for Rare and Uncommon Vascular Diseases: Comparisons of May Thurner, Thoracic Outlet, and Superior Mesenteric Artery Syndrome. Health Commun. Routledge; 2015;30: 859–871. pmid:24877701
  15. 15. Jimison HB, Pavel M, Parker A, Mainello K. The Role of Human Computer Interaction in Consumer Health Applications: Current State, Challenges and the Future. In: Patel VL, Kannampallil TG, Kaufman DR, editors. Cognitive Informatics for Biomedicine: Human Computer Interaction in Healthcare. Cham: Springer International Publishing; 2015. pp. 259–278.
  16. 16. Eurostat. Digital economy and society statistics—households and individuals. In: Eurostat [Internet]. 2016 [cited 17 Nov 2017]. Available:
  17. 17. Facebook. Company Info | Facebook Newsroom [Internet]. [cited 29 Sep 2015]. Available:
  18. 18. Loving M, Ochoa M. Facebook as a classroom management solution. New Libr World. Emerald Group Publishing Limited; 2011;112: 121–130.
  19. 19. Cotten SR, Anderson WA, McCullough BM. Impact of internet use on loneliness and contact with others among older adults: cross-sectional analysis. J Med Internet Res. JMIR Publications Inc.; 2013;15: e39. pmid:23448864
  20. 20. Brunner M, Hemsley B, Palmer S, Dann S, Togher L. Review of the literature on the use of social media by people with traumatic brain injury (TBI). Disabil Rehabil. 2015;37: 1511–21. pmid:25985717
  21. 21. Morton M V, Wehman P. Psychosocial and emotional sequelae of individuals with traumatic brain injury: a literature review and recommendations. Brain Inj. 1995;9: 81–92. pmid:7874099
  22. 22. Demir SÖ, Görgülü G, Köseoglu F. Comparison of rehabilitation outcome in patients with aphasic and non-aphasic traumatic brain injury. J Rehabil Med. 2006;38: 68–71. pmid:16548091
  23. 23. Greene JA, Choudhry NK, Kilabuk E, Shrank WH. Online Social Networking by Patients with Diabetes: A Qualitative Evaluation of Communication with Facebook. J Gen Intern Med. Springer-Verlag; 2011;26: 287–292. pmid:20945113
  24. 24. Bender JL, Jimenez-Marroquin M-C, Jadad AR. Seeking support on facebook: a content analysis of breast cancer groups. J Med Internet Res. 2011;13: e16. pmid:21371990
  25. 25. Gaysynsky A, Romansky-Poulin K, Arpadi S. "My YAP Family": Analysis of a Facebook Group for Young Adults Living with HIV. AIDS Behav. 2015;19: 947–62. pmid:25186783
  26. 26. Gattiker UE. Social media audits: achieving deep impact without sacrificing the bottom line.
  27. 27. Eghdam A, Scholl J, Bartfai A, Koch S. Information and communication technology to support self-management of patients with mild acquired cognitive impairments: systematic review. J Med Internet Res. 2012;14: e159. pmid:23165152
  28. 28. Eghdam A. Understanding how persons with mild acquired cognitive impairment use and experience information and communication technology: an exploratory study [dissertation]. Karolinska Institutet. 2016.
  29. 29. Statistics Sweden. Seasonal Adjustment [Internet]. 2012 [cited 27 Dec 2015]. Available:
  30. 30. Rieder B, Bernhard. Studying Facebook via data extraction. Proceedings of the 5th Annual ACM Web Science Conference on—WebSci ‘13. New York, New York, USA: ACM Press; 2013. pp. 346–355. 10.1145/2464464.2464475
  31. 31. Hsieh H-F, Shannon SE. Three approaches to qualitative content analysis. Qual Health Res. SAGE Publications; 2005;15: 1277–88. pmid:16204405
  32. 32. Cutrona CE, Shur JA. Controllability of Stressful Events and Satisfaction With Spouse Support Behaviors. Communic Res. SAGE Publications; 1992;19: 154–174.
  33. 33. Scott J. Social network analysis: a handbook. SAGE Publications; 2000.
  34. 34. Nygren E. Simulation of User Participation and Interaction in Online Discussion Groups. 2011 IEEE Third Int’l Conference on Privacy, Security, Risk and Trust and 2011 IEEE Third Int’l Conference on Social Computing. Springer Berlin Heidelberg; 2011. pp. 1390–1397. 10.1109/PASSAT/SocialCom.2011.83
  35. 35. Robert A. Hanneman MR. Introduction to Social Network Methods—Robert A. Hanneman, Mark Riddle—Google Books [Internet]. 1st ed. Riverside: University of California; 2005. Available:
  36. 36. Emile Durkheim GS and FT . Social Differentiation and Functionalist Sociology. In: Hess Andreas, editor. Concepts of Social Stratification. London: Palgrave Macmillan UK; 2001. pp. 36–49.
  37. 37. López-Ferrer M. Social Network Analysis Tools to Understand How Research Groups Interact: A Case Study. In: Juan AA, Daradoumis T, Roca M, Grasman SE FJ, editor. Collaborative and Distributed E-Research: Innovations in Technologies, Strategies and Applications: Innovations in Technologies, Strategies and Applications. 1st ed. Hershey, PA, USA: IGI Publishing; 2012. p. 209.
  38. 38. Combe D, Largeron C, Egyed-Zsigmond E, Géry M. A comparative study of social network analysis tools. Soc Networks. 2010;2: 1–12.
  39. 39. Gerolimos M. Academic Libraries on Facebook: An Analysis of Users’ Comments. D-Lib Mag. Corporation for National Research Initiatives; 2011;17: 4.
  40. 40. Thackeray R, Neiger BL, Smith AK, Van Wagenen SB, Safko J, Brake D, et al. Adoption and use of social media among public health departments. BMC Public Health. BioMed Central; 2012;12: 242. pmid:22449137
  41. 41. Bruns A. How Long Is a Tweet? Mapping Dynamic Conversation Networks on Twitter Using Gawk and Gephi. Information, Commun Soc. Routledge; 2012;15: 1323–1351.
  42. 42. Coursaris CK, Liu M. An analysis of social support exchanges in online HIV/AIDS self-help groups. Comput Human Behav. 2009;25: 911–918.
  43. 43. Sreenivas S. Social Network Analysis of the Facebook Group—SVUCE Engineer | SVUCE Engineer [Internet].
  44. 44. Lombardo JP, Fantasis SC. The relationship of self-disclosure to personality, adjustment and self-actualization. J Clin Psychol. 1976;32: 765–9. Available: pmid:977737
  45. 45. Montgomery JM, Espie CA. Behavioural Management of Hysterical Pseudoseizures. Behav Psychother. Cambridge University Press; 1986;14: 334–340.
  46. 46. Zeng EQ, Zeng BQ, Tian JL, Du B, Tian XB, Chen H. Perceived Social Support and Its Impact on Mental Fatigue in Patients with Mild Traumatic Brain Injury. Balkan Med J. 2016;33: 152–7. pmid:27403383
  47. 47. Kahneman D. Thinking, fast and slow. 1st ed. New York: Farrar, Straus and Giroux; 2013.
  48. 48. Morley JE, Morris JC, Berg-Weger M, Borson S, Carpenter BD, del Campo N, et al. Brain Health: The Importance of Recognizing Cognitive Impairment: An IAGG Consensus Conference. Journal of the American Medical Directors Association. 2015. pmid:26315321
  49. 49. Tierney MC, Naglie G, Upshur R, Jaakkimainen L, Moineddin R, Charles J, et al. Factors associated with primary care physicians’ recognition of cognitive impairment in their older patients. Alzheimer Dis Assoc Disord. 28: 320–5. pmid:24632991