The different faces of online support: A thematic analysis of responses to help-seeking related to depression and suicidal ideation in public online forums

Introduction

Suicide continues to be a leading cause of mortality worldwide, with close to 800 000 deaths yearly (WHO, 2021). Although suicidal plans and ideation are linked to psychiatric disorders, such as major depressive disorder, most people dying from suicide are not in current contact with mental health services (Turecki et al., 2019). In the digital age, online environments make up a growing arena for help-seeking by individuals suffering from mental health issues, including depression and suicidal thoughts. But what help do online forums provide? From the help-seekers’ perspective, forums appear to have several compelling aspects compared to traditional offline help. Users state that online environments can offer a useful option when offline support is unavailable. Furthermore, they facilitate discussions of stigmatized health conditions and provide access to others with similar problems, others who may have an increased understanding of their problems and experiences (Wright, 2016). It is common that suicide-risk individuals utilize online resources for suicide-related purposes (Harris et al., 2009). Compared to suicide-risk individuals that does not use online resources, this group has been found to report lower levels of perceived social support, as well as more time spent online. They also seemed to differ through a greater use of social networking resources, which may be indicative of them searching for other people with similar experiences (Harris et al., 2009, 2014). It has been noted that these users tend to view the online communication around the topic as supportive, rather than negative (Eichenberg, 2008; Harris et al., 2009), but also that social networking sites and participation in discussion forums could be important sources for exposure to suicide stories and even increase suicidal ideation (Dunlop et al., 2011).

Online environments commonly suffer from several potential problems in relation to mental health issues. Often, professional expertise is lacking, potentially leading to unintentional unhelpful or even harmful advice, or evaluations of the situation. There is also a risk of intentional harmful responses, including trolling, bullying, and encouragement of problematic behavior, such as whether to seek help or not, and how to approach health-care services or treatment. In relation to depression and suicidal ideation, this is particularly relevant as several studies have noted that strategies employed by responding users online may include malicious or excluding intent (Gough, 2016; Kaufman & Whitehead, 2018; Lundström, 2018; Westerlund, 2013; Westerlund et al., 2015). An extreme example of the potentially detrimental effects of forum usage is a case where forum participation appears to have pushed one user towards suicide (Westerlund et al., 2015).

In this study, we aim to examine the types of responses that posts seeking help for depression and suicidal ideations elicit within online discussion forums. The insights garnered could clarify both the positive and negative impacts of seeking help in this particular online context. Additionally, it could point to critical elements that might influence the usefulness of support, thereby contributing to a more nuanced understanding of the support process in digital spaces. Following the advice of several previous studies that noted difficulties with classifying the content in online settings, we opted for a bottom-up approach for analyzing the forum content (Braithwaite et al., 1999; Coulson & Greenwood, 2012; Hwang et al., 2010). In doing so, we attempt to gain a deeper understanding of the online experiences of individuals grappling with these issues.

Online Health Forums

Over the last decades, the internet has transformed the landscape of healthcare information and led to a growing usage of different virtual health resources. Recent survey data collected in European countries suggests that 55% of the adult population has sought health-related information online, with numbers as high as over 70% in many countries (Eurostat, 2021). Similarly, in the U.S., a recent study indicated that 74% had sought health information online for their most recent search (Finney Rutten et al., 2019). One significant development in this trend is the emergence of online health forums. These digital platforms provide users with spaces to share experiences, seek advice, and discuss health-related topics (Eysenbach et al., 2004). Although it is difficult to assess how widespread their usage is, the vast proliferation of forums underscores their significance in contemporary society. Health forums can be focused on specific medical conditions, treatments, or other wellness topics. They can also be broader, more general forums that offer a space to discuss many different topics within the same resource, sometimes branching out into specific subsections. Today, these resources are typically provided through internet discussion forums (digital bulletin boards) or social media platforms. In essence, these resources provide users with a space to create threads, engage in discussions, and seek information from a community of peers. Different forums vary in design and scope. Typically features include that users register and create accounts to post their queries, stories, and insights, but they vary to the extent in which they manage anonymity, user profiles, and use of moderators to facilitate engagement and a supportive environment. 

Opportunities and Drawbacks of Internet Health Forums

From a user perspective, there are several practical opportunities afforded by forum technology. Important aspects include accessibility and convenience, affordability, and the possibility of being anonymous (Hanley et al., 2019). By interacting in the forum, users receive stand alone or complementary informal support, and are provided informational and emotional social support that can lead to a sense of empowerment, community, and hope. It can also provide emotional relief, reduce isolation, and encourage individuals to be more active in their own health management (Hanley et al., 2019; Wright, 2016).

However, the nature of these forums also possesses some potential drawbacks. The open nature of these platforms can lead to the dissemination of inaccurate or unverified information, potentially compromising users’ health decisions. Forums may also inadvertently perpetuate negative health behaviors or foster anxiety by showcasing severe cases or unrepresentative outcomes, thus having a negative influence for the user. Hence, relying solely on forum content could lead to misdiagnoses or inappropriate treatment decisions as professional expertise is often lacking (Jacobs et al., 2017).

The Content in Health Forums 

It is possible to categorize the exchanges on forums in several ways. An influential approach for content analysis of forums has been to view it in terms of social support. Defined as “actions taken by individuals who intend to alleviate the distress of others”, social support has been found to have positive influence across a range of health outcomes (Gleason & Iida, 2015), including functioning as a protective factor for depression (Gariépy et al., 2016). There has been a large research interest directed at examining how online forums can provide social support to users (Griffiths, Calear, Banfield, & Tam, 2009; Prescott et al., 2017; Schotanus-Dijkstra et al., 2014; Smedley et al., 2015).

Supportive actions have been categorized into at least five macro-categories using a coding framework originally developed to study spouse support behaviors (Cutrona & Suhr, 1992). First, attempts at helping can focus on action-prompting by providing information, e.g., support in the form of messages that provide facts, guidance, or advice, or attempts to change appraisal of the situation. Second, it can take the form of emotional support, e.g., through expressions of physical affection, messages of empathy, understanding or encouragement. Third, some actions have also been labeled as esteem support, i.e., support conveyed by compliments and expressions of agreement. Fourth, offerings of tangible support, such as offers of physical or monetary assistance. And fifth, help can be in the form of network support, e.g., attempts to help expand someone’s social connections by providing access to social networks, spending time with someone or offering companionship. Throughout much of the social-support literature however, the various forms of support are reduced to two overarching types: emotional and instrumental-practical support (Carver et al., 1989; Gleason & Iida, 2015). Rains et al. (2015) performed a review of computer mediated support and concluded that online support messages examined in different studies could essentially be classified into these types. Further, they also highlight that informational and emotional support types are identified in virtually all studies of online support and are the most common forms. In contrast, esteem support, network support, and tangible support were included less frequently. 

Within the social support literature, it has also been hypothesized that there is an important distinction between perceived support (support experienced by the recipient) and enacted support (any action meant to be helpful; Barrera, 1986). Perceived support is thought to alleviate stress while enacted support may have mixed effects, potentially due to interactions with contextual circumstances such as the type of stressor and individual factors such as personality, developmental history, and past experiences (Uchino, 2009). The match between the type of support individuals seeks and the type of support they receive has also been hypothesized as crucial for its beneficial effect (Cutrona & Suhr, 1992; Gleason & Iida, 2015). Hence, the matching of different forms of support has been proposed as an explanation of differences in the effects of various kinds of supportive interactions. Instrumental-practical support is similar to problem-focused coping and is consequently expected to be most helpful for controllable problems, as information, advice, and guidance can help the help-seeker deal with the cause of their difficulties. Conversely, emotional support is comparable to emotion focused coping, which is more relevant for uncontrollable problems, as it aids the help-seeker to endure the negative and stressful feelings of being in an uncontrollable situation. This overall pattern of support matching has also been found in the meta-analysis of content of support in forums (Rains et al., 2015). When stressor characteristics were considered nurturant forms of support were more common in relation to health conditions likely to threaten personal relationships as well as among content analyses focusing on health conditions with a greater potential for loss in the form of death. Action-facilitating types of support were instead more commonly found in content analyses examining chronic conditions.

What Are the Effects of Using Forums?

While social support is generally viewed as beneficial, evidence regarding the effects of online forum participation on health outcomes are ambiguous (Ali et al., 2015; Eysenbach et al., 2004), including for depression (Griffiths, Calear, & Banfield, 2009). Notably, one systematic review of online peer support for young people with mental health problems found that only two out of six studies showed significant positive outcomes (Ali et al., 2015). One randomized controlled trial of an online support forum for individuals with psychiatric disabilities even found that participants with a higher degree of forum participation reported higher levels of distress as compared to individuals who participated less (Kaplan et al., 2011).

Findings of varying effects of participating in internet forums mirrors one of the main controversies in the broader social support literature. While an individual’s perception of available support is generally associated with positive health outcomes, many studies examining actual received (enacted) support find that supportive interactions are correlated with negative mood and distress. These findings have been described as a paradox of social support (Gleason & Iida, 2015). There are several possible explanations to this effect; it has been proposed that the effectiveness of the support depends on the match between the type of support provided and the help-seekers needs (Barrera, 1986). There may also be potential negative effects of social support if it undermines an individual’s autonomy or promote negative behavior or dependency (Thoits, 2011). Rafaeli and Gleason (2009) have argued that support attempts can often be miscarried and performed in an unskilled manner and therefore may results in greater costs than benefits for the receiver of the support. Factors that can influence the effectiveness include such things as timing of the support, individual factors related to the involved parties and their relationship, and the actual content of the support (Gleason & Iida, 2015; Rafaeli & Gleason, 2009).

Given the lack of strong empirical evidence that forum usage benefits health outcomes directly it is important to explore possible reason for this lack of effectiveness. It might be that the quality of the support (e.g., the accuracy of information or how skillfully it is conveyed), or the composition (e.g., how different simultaneous attempts at support is interpreted and managed) matters. Beneficial effects of support could also be hampered by the effects from negative social interactions, be them intentional or unintentional, or by inducing unhelpful notions that a help-seeker might construe from their forum participation.

The Present Study

To better understand the different responses that help-seeking posts related to depression and suicidal ideations generates, this study aimed to examine discussion threads of posts from two Swedish online forums. In particular, we wanted to capture a broad impression of forum participation, including qualitative aspects of the communication (such as tone and intentions) that more general conceptualizations might miss, but which might hold importance for the quality of the responses. Our specific research question was: What themes can be found in responses to help-seeking for depression and suicidal thoughts in internet discussion forums, and what characterize these themes? Our assumption was that a closer examination of responses could further the understanding of the varying effects of forum participation in relation to health outcomes.

Methods

This study used reflexive thematic analysis (Braun & Clarke, 2021) of discussion threads posted on two large public internet forums with sub-forums about mental health. Threads that dealt with depression and suicidal thoughts were selected based on the titles and first posts from the threads. Ethical approval for the study was obtained from the Swedish Ethical Review Authority (Dnr 2020-02506). All analyzed materials were publicly accessible and available for anyone to read. All forum posts were written under a registered username or a pseudonym, but the forums did not collect any other personal information from the users. Due to the public availability of the data and the anonymity of the forum users, the forums were considered part of the public domain. Consequently, consent was not sought from participants before conducting the study (Roberts, 2015). However, all data has been stripped from usernames and other possibly identifying information, and all quotes presented have been translated into English and slightly rewritten for anonymity purposes.

Materials

Data were collected from two large Swedish internet forums (https://www.flashback.org/ and https://www.familjeliv.se/forum). The forums are similar in that they are very broad in scope (in practice spanning virtually any possible topic) and have a very large and active user base. Both forums are among the most visited Swedish websites and according to the Swedish internet Foundation, in 2021, 27% of the population had visited Forum 1 and 7 % had visited Forum 2. Both forums allow anyone with an email address to register and make posts (Forum 2 also allowed posting without logging in). Both forums have rules about keeping to the topic, keeping a civil tone and not discussing suicide techniques or concrete suicide plans. Both forums also have moderators that can be alerted if postings break these rules, but in practice the moderation can be described as minimal. All threads in the sub-forums about mental health active in June 2020 and that had been active sometime in the past 12 months were screened by the authors. Relevant posts about depression and suicidal thoughts were selected based on the titles and first post of the thread. The resulting dataset consisted of 1,475 posts from 72 different threads. When the study was conducted, most of the threads analyzed had been viewed a few thousand times. The included forum threads had between 4 and 200 replying posts. Forum 1 (35 threads) had a mean of 31,1 replying posts per thread, and Forum 2 (32 threads) had a mean of 10,5 replying posts per thread. As a result, our dataset consisted of 1,087 posts from Forum 1 and 388 posts from Forum 2. The included threads from Forum 1 contain a total of 479 unique usernames. Forum 2 did not require users to register to participate in discussions, which prevents us from having an exact account of the number of active users. Based on the user count for Forum 1, the total number of participants for both forums combined can be estimated to be approximately 650–750.

The first posts of each thread were not analyzed but were saved along with the responses for contextual purposes. The quote below is an example of what a typical first post could look like and is one of the shorter first posts in the dataset.

“I have an unbearable depression. Life is just painful. When everything is at its worst, I also get abandoned by the people in my life. Now I am completely alone and have lost faith in humanity. Everyone knows how vulnerable and lonely I am, but they still don’t care which is incomprehensible and disgusting. Not even my family cares about me. I just want to die but I’m not going to take my own life. However, I wish I could die in my sleep, just fall asleep and never wake up again.”

Data Analysis and Rigor

 The entirety of the included threads was copied, processed, and saved onto NVivo 1.3, where the data were coded and analyzed. In the coding process, each reply was coded in segments (i.e., a longer reply could generate several codes) based on what strategies and meaning they appeared to convey to the original question (i.e., “how” the poster could be helped), as well as the specific content of the help (“what” the help constituted).

The data were analyzed using reflexive thematic analysis following the six-step process suggested by Braun and Clarke (2021): i.e., first we familiarized and engaged critically with the data by reading through each tread multiple times (authors HG & EJ); then systematically reviewing and assigning codes to the posts based on the content, interpreted intentions and tones of the posts; in the next step, we generated initial themes from the coding; reviewed and developed these themes through discussions in the research group and finally; naming themes and writing them up. An inductive approach to the data was used, and coding was open and done without any preexisting coding framework. Instead, our reflective practice included that all authors, after an initial selection and coding of forum threads, in parallel read, coded, and, importantly, critically discussed how the coded data could be best summarized in themes (also acknowledging and discussing how our subjective and pre-analytical predispositions could influence the analytical process). Our approach could be described as that of epistemological realists, but our interpretations of the forum posts have also moved beyond the explicit and considered latent meanings (such as the perceived intentions and tones of the posts). Data collection and analysis were also guided by methodological concepts adopted from grounded theory, namely theoretical sampling (screening sub-forums for relevant threads), theoretical saturation (collecting data until no new codes were generated), and the practice of continually collecting data and analyzing the material (Corbin & Strauss, 2012).

The goal of our analysis was to create themes that represent as much of the data as possible. This has led to some potential overlaps between themes. For example, posts about healthcare and medicine can be included in most of the themes, while it is primarily the strategy applied by the responding users that distinguish one theme from another.

Results

Through the analysis, six themes were identified, which could be described as six different response types: Giving tips, Sharing One’s Own Experiences, Consoling, Educating, Coaching, and Bullying. These types, as well as subthemes and a summary of their respective primary response strategy and example quotes, can be found in Table 1. The first five types use strategies aimed at helping the original poster, while the last type did not appear to have this aim.

Table 1. Themes (Number of Times Coded in the Data), With Their Respective Subthemes (in Italicized Font),
Primary Strategies, and Example Quotes.

The number of sequences from all the posts (1,472) that comprise each theme are included in the table to illustrate the frequency of different types of strategies. Giving tips was the most frequent type (coded 781 times), Consoling (494 times), and Coaching types (487 times) were also common, Sharing One’s Own Experience (341 times), and the Educating (283 times) types were less frequent. Bullying type (70 times) were the most uncommon type of message. In the following sections, each type, including its subthemes, are presented, along with examples of anonymized quotes. Subthemes are italicized in the text.

Theme 1: Giving Tips

More or less, every thread in the dataset included numerous posts aimed at supplying the thread starter with practical tips on things to do. At the core of these posts were the strategy of quickly arriving at some kind of concrete simple action that could be of help for the help-seeker. This theme also differed from the remaining themes in that they were not explicit in what the suggestions were based on and not presented within a coherent framework or rationale. The advice varied widely, ranging from trying a specific diet (a Healthy Mind in a Healthy Body) to starting over on another continent (Time for a Change). Some tips were short and straightforward and could include prompts about seeking professional help (Get Help) or trying a specific medicine (Medication and Self-medication). A few examples of tips could be: “Take multivitamins, omega-3 and 5-htp and everything will be easier”, “go talk to someone!” or “try micro dosing LSD”. Other tips were longer and contained detailed descriptions of activities but still centered on fairly a simple solution to what was complex problems. For example, one user suggested that a person who described feelings of depression and low self-esteem should try doing squats:

“Squats are a good exercise that includes many muscles and raises testosterone levels in the body. You can start today; why not? It’s fine to just hold on to a door frame for balance or pull yourself up with your arms. I’ve seen many who can’t do a single rep without help so there’s no shame in that. Try to do five reps, rest for a few minutes and do this twice more. Today’s workout done! Try to make it a routine, and you’ll be on your way. Maybe I get a little too stuck on training, but the body’s mood and psyche are quite connected. Then, of course, there is no miracle cure that solves everything but at least it won’t hurt.”

It is important to note that different posts of this type could contradict one another completely. There was no consensus in the forums about the preferred lifestyle or which drugs or dietary supplements one should use, and even though a lot of users recommended professional health services, others could tell the thread starter to avoid the healthcare system altogether.

Theme 2: Sharing One’s Own Experiences

A frequently employed strategy among responding users was to base contributions on their personal experiences. The phrase “We Are in the Same Boat” reoccurred frequently, and comments often emphasized similarities between responder and thread starter, such as shared experiences of a particular diagnosis, medication, or situation. Many users emphasized shared experiences of struggling and suffering, often with the explicit purpose of reducing stigma or loneliness and creating a sense of community.

“Let me start by saying that I recognize myself in the original post. The similarities with my life are so striking it might as well have been me writing. If it’s any consolation you should know you are not alone.”

A subset of these comments was specifically concerned with sharing experiences of The Road to Recovery from depression and other mental health issues. These respondents were often explicitly looking to convey hope by telling stories of their journey from a dark period in their life to a brighter future. Their comments could include a wide array of advice or suggestions but were nonetheless grounded in the responders’ lived experiences. Additionally, they were often written in the tone of a mentor or a sponsor, someone who knew the thread starters’ pain and struggles but also knew a way out.

“I’m not exactly happy today, I don’t appreciate things like I should, and so on. But it’s far from the hell that I used to live in. And I’m so damn grateful for that. But you should know that the road is not straight. [...] I didn’t really notice myself getting better and better. Only when I look back at how things were before do I see the enormous difference. My life is far from perfect, but at least it’s not all bad.”

Posts like the one above used the responder’s personal experience both as a means of communicating hope and describing how the road to recovery can be long and complicated.

Theme 3: Consoling

Virtually every thread contained supportive or consoling posts. Comments within this theme were characterized by attempts to tell the thread starters that there were, in fact, people on the forum who cared about them. The message these responders tried to bring about was that things would get better and that there was still hope, even though it might be difficult to see. The consoling response strategy could generally be interpreted as some sort of emotional support and had little focus on explicit content or information. These messages of hope could resemble the comments in the Sharing one’s own experiences theme, but differed from these by referring less to the responder’s own lived experiences. Comments within the Consoler theme were often short but personal and aimed at making the thread starter feel more hopeful and less alone: “Shit, that sounds tough”; “I hope that you don’t give up, even if I know things might seem bad right now”; “it will get better, I promise.” In conclusion, the Consoling theme was present in most of the analyzed threads, and its comments distinguished themselves by their empathic and emotionally supportive character.

Theme 4: Educating

Many forum users sought to educate the thread starter and other users about medicine and healthcare (Healthcare Class) or the world in general (Life Lessons). The Educating theme was characterized by an expert position and a tendency to view one’s thoughts and opinions as true. These posts did not focus directly on addressing the original post; instead, at their core was the intent to present the respondents views on some topic in a lecturing manner. For example, comments within this theme could contain information about how the healthcare system works, theories about psychological phenomena, or opinions about politics or medicine. These comments were often written in a confident tone, and the responders seemed to view their knowledge of things as essential and correct. One comment within the theme could contain a more extended exposition about a phenomenon, while another could consist of a few words of wisdom. An example of a typical Educating comment could look like this user’s statement about anti-depressive medication:

“No one knows exactly how they work. It’s all about hit and miss and only naive idiots think that they solve problems without affecting anything else. Sure, maybe you’ll become less depressed but you’re not going to feel anything else either, neither joy nor depression, so say hello to your new apathetic self. How about some erectile dysfunction with that?”

Although this comment seemed to stem from a negative view of the healthcare system and medicine, peoples’ opinions about healthcare differed widely. Similarly, the opinions and theories that responders shared about the world could touch almost any subject and be aimed in any direction. Some Educating comments seemed to be based on established psychological or political theories. In contrast, others were more private opinions: “we’re expected to walk around like zombies and fill our lives with drugs, charter travel, furnishing, and sex games. Don’t mess with the gender roles, we are created the way we are for a reason.” While comments within the Educating theme often consisted of opinions or speculations, they could also be aimed at sharing information in objective ways, even though the information provided was not necessarily correct.

Theme 5: Coaching

Some of the comments in our dataset employed strategies with a higher degree of complexity, combining suggestions, information, and consolation into attempts at guiding or leading the thread starters in solving their problems. These posts were characterized by containing some sort of personalized model or explanation of the experiences relating to the original question, as well as some idea of how to manage or change the situation. In contrast to the Giving tips theme which were concrete, unsubstantiated suggestions, the Coaching theme was more concerned with using general psychological theory to understand or improve the thread starter’s life and then to describe the resulting process of working to change it. These posts also did not limit themselves to being purely supportive (like the Consoling theme) or to conveying general information (like the Educating theme), but instead used a complex combination of these approaches. Another unique feature of these posts was that they included questions for the original poster to answer.

A subset of the coaching responders was explicitly interested in Finding the Reason for the thread starter’s suffering, offering explanations in terms of psychiatric diagnoses or more general psychological constructs, such as insecure attachment, trauma, or low self-esteem. These comments were generally based on the idea that the thread starters needed a better understanding of themselves and their problems to facilitate change. The following quote is an example of an attempt at explaining depression as a result of childhood trauma: “You suffer from trauma. Think of your childhood: what were your worst experiences? What experiences affected you the most?”

Other coaching posts were more concerned with suggesting concrete strategies for dealing with psychological problems (Control Your Mind). Different users suggested different, and at times contradictory, strategies but generally, the suggested methods had much in common with practices in cognitive behavioral therapy. These practices included techniques such as ignoring negative thoughts or letting anxiety pass without interfering: “thoughts come and go by themselves! They disappear quicker if you don’t listen to them.”

Some users also employed the “tough love” strategy, urging the thread starters to Pull Themselves Together. These users were looking to make the thread starters take responsibility for their own lives and were often concerned with what they saw as unhealthy self-victimization or the tendency of some thread starters to view their situation as hopeless. These posts were often written in harsh and sometimes accusive language: “Your life is hell because you’re letting it fall apart. Stop smoking pot, stop sitting by the computer, get a hold of your life and do something. Get it together and stop feeling sorry for yourself.”

The Coaching theme can be conceptualized as quite complex and consists of a combination of elements from other themes; coaching comments can contain tips, support, and information, while it is distinguished by its tendency to use these different strategies to guide the thread starters through their situations.

Theme 6: Bullying

The comments from our first five themes could in large be interpreted as attempts to help or support the thread starter in one way or another. Our last theme, Bullying, consists of posts with what we interpreted as malicious or antagonistic intent. The tone in these comments was harsh or mocking and seemed to aim to ruin the climate of the forum conversations or hurt and insult other forum users. Bullying comments were rare in our dataset, but the theme still appeared relevant for inclusion. The theme never constituted the majority of a thread but was nevertheless present in many of them and was hard to ignore due to its distinct nature.

The general message of the Bullying theme was that other users were weak and useless and that the thread starters’ help-seeking only confirmed their weakness. A general strategy applied in posts within the Bullying theme was to tell other users that they were in no way special and that they should be ashamed of themselves for complaining about their ill-being. For example, in a thread where the thread starter disclosed recurring suicidal thoughts, a responder replied: “The fact that you’re screwing everything up because you’re weak and worthless is just pathetic, and then you come here to cry about it.” Comments like these illustrated how the Bullying responses would use other peoples’ suffering as proof of their failing as human beings. Another typical Bullying strategy was to tell the thread starter that they had no reason to complain and that they should not share their stories on the forums because no one cared:

“What a whiner you are! The slightest setback and you’re talking about suicide, because you’re having such a hard time huh? People lose arms and legs, get chronic diseases and disabilities, while you sit here crying because you can’t carry your own weight? You should be ashamed of yourself...”

Both of the analyzed forums had rules about the tone and content of posts, which might have affected the number of bullying comments in our dataset. In some cases, it was possible to see traces of posts that had been removed because they violated the rules of the forum. Hence, the true number of bullying posts were higher than found in this analysis.

Discussion

The present study aimed to investigate the responses that posts about depression and suicide generated on two Swedish internet forums. A thematic analysis of the underlying themes resulted in six types of responses: Giving tips, Educating, Coaching, Sharing one’s own experiences, Consoling, and Bullying. These themes capture the diverse ways forum users engage with individuals seeking help or support online, including different qualities of the responses, as they may be perceived by the help-seeker or passive users of these forums. 

At their core, all but one of these themes appear to represent genuine attempts at helping, mirroring findings from studies on the content of internet support forums for health issues, where a plethora of different forms of enacted support is readily offered (Prescott et al., 2017; Rains et al., 2015). A subset of responses however made up the Bullying theme and appears to have destructive intent. These responses are expressed as attacks on the help-seeker, a phenomenon that is common in many online environments. While this type of response has generally not been included in the literature examining social support content in internet discussion forums, similar findings have been reported in studies related to postings about suicide (Westerlund, 2013; Westerlund et al., 2012). Potential negative effects of internet use have also been a highlighted issue within research focusing on self-harm (Daine et al., 2013), and cyberbullying has been identified as a factor in several documented cases leading to suicide (Barlett et al., 2018; Westerlund et al., 2015). Compared to the study reported here, previous research on social support and internet forums have focused on relatively narrower online forums with specific focuses (i.e., support groups). While the literature acknowledges that there is indeed a breadth of different types of forums, not many have focused on the type of broad topic forums that were the focus of our study. It might be that the harsher tone and the malicious comments that characterize the Bullying theme are more frequent in this type of forum. There are several features of online communication that have been proposed as contributing to what has been termed as toxic disinhibition in participant interactions (Suler, 2004). The forums included in our study are characterized by anonymity, invisibility and lack of direct contact between users, factors that have been found to contribute to a disinhibition effect (Lapidot-Lefler & Barak, 2012). Although the Bullying theme is represented in only a relatively small number of posts, the effect of these comments may be especially detrimental to the forums due to their disproportional effect on the tone of the discussions (Fullwood et al., 2019). Future research should investigate the impact of this type of responses on the supportive qualities of forum usages as a whole and how they affect help-seekers.

The five help-oriented themes identified in the thematic analysis can be broadly matched to three of the categories of support that the social support literature outlines: comments that primarily provide instrumental-practical support (i.e., the Giving tips, Educating, and Coaching themes) and comments that providing emotional or esteem support (i.e., the Consoling, and Sharing One’s Own Experiences themes). However, when viewed as threads of responses given to a specific help-seeker, the present thematization capture more nuances of both the content of the support (such as the accuracy and ambiguity of information), as well as the delivery of the support (such as the tone and skillfulness conveyed by the responses). Only focusing on the frequency of these broad categories of support at a forum level would likely fail to consider such qualitative aspects of the content that is critical for the help-seeker of a specific thread. This aspect has not been thoroughly addressed in the literature to date and could potentially be important to understand the effectiveness of forum interactions. It is also important to note that in our study, help-seekers received numerous responses (typically between 10 and 20 replies), indicating that the factor of total number of responses (the thread’s attractiveness) may also be important to consider. The relative frequency of different types of responses (the balance within the thread) might influence how helpful the support appears. While more numerous responses increase the enacted support and increase the likelihood of relevant suggestions, it also increases the potential for diverging and even obfuscating messages.

The Giving Tips theme was the most common help-oriented response in this study. This theme could subsequently be divided into four sub-themes, based on different focuses on the suggestions offered; many responders were encouraging the help-seeker in addressing the connection between physical and mental health (A Healthy Mind in a Healthy Body sub-theme), encouraging the help-seeker to initiate life changes (Time for a Change sub-theme), to use medication or drugs to alleviate their mood (Medication and Self-medication sub-theme), or encouraging the thread starter to seek help or support outside of the forum (Get Help sub-theme). The Giving Tips theme was characterized by offering short and generally simplistic suggestions of concrete actions that the help-seeker could try. The theme was also characterized by considerable differences and contradictions in proposed solutions to the same problem. Although many potential suggestions and strategies were offered, the discussions typically failed to reach any consensus, leaving the help-seeker without definitive guidance. Similar findings have been reported in online support groups for childhood cancer (Coulson & Greenwood, 2012). The sheer amount of support of this type can easily appear overwhelming and misguided, potentially encumbering the help-seekers limited cognitive and motivational capabilities and thus increasing distress, as theorized in previous literature (Gleason & Iida, 2015; Rafaeli & Gleason, 2009).

Like Giving Tips, the Educating and Coaching themes also focused on instrumental-practical suggestions but were more complex in their approach. The educating theme could be characterized as respondents’ attempts to help by explaining what they perceived as important fundamental truths about the world (Life Lessons sub-theme) or about medicine and mental health issues (Healthcare Class sub-theme). In the Coaching theme were responses attempting to guide the help-seeker through their situation. These responses could be further characterized into three sub-themes. First, there were messages intended to develop an understanding of the help-seekers situation and address the core problem (Find the Cause sub-theme). Second, there were messages suggesting different concrete strategies or techniques to manage problems (Control Your Mind sub-theme). Third, there were messages encouraging the help-seeker to adopt a more proactive stance in their situation (Pull Yourself Together sub-theme). It is relevant to note that while all these themes could be characterized as predominantly attempts at instrumental-practical support, the actual content of the replies (the quality of the responses) again varied considerably. As did the tone of different messages within these themes (e.g., Educating responses could appear unrelated to the help-seeker’s question, and Coaching responses could be perceived as harsh). Many messages in the Educating and Coaching themes also adopted a clear expert position, emphasizing the responder’s own knowledge of the topic and establishing a hierarchy separating helper and help-seeker. Previous research has associated the expert position with a lack of communicated empathy and with resistance on the part of the help-seeker (e.g., Kaufman & Whitehead, 2018). Hence, the helpfulness of these different attempts is difficult to assess without knowing more about how the help-seeker interpret them. The sheer breadth of suggestions, solutions, and attempts at guidance in the instrumental-practical-oriented responses appear to be a central aspect of the support these forums offered. It might be that the broad scope of these forums leads to more heterogeneous user base, and thus a greater variation in the content of support offered. Further research could explore how this is experienced from the help-seekers perspective and if the specific information or suggestions that are given impact the utility of the support.

In this study, we also identified two themes that appear to constitute predominantly emotional or esteem support. The Consoling theme and the Sharing One’s Own Experiences theme. These were more uniform in their tone and content, focusing on conveying empathy and comforting the help-seeker. Consoling messages focused on conveying hope, optimism, and belonging. The messages were usually short, and the theme was common in virtually all threads in these forums. Irrespective of other forms of support, virtually all help-seekers were met, usually prompt, with such basic emotional support. Responses in the Sharing One’s Own experience theme used respondents’ personal experiences as a starting point for emotional support. In the sub-theme In the Same Boat, the messages focused on the similarities between the help-seeker and the responders’ situation or experience, and in the sub-theme The Road to Recovery, they focused on the responder’s experiences of having overcome similar difficulties as the help-seeker. These messages attested to a likeness or bond between the users on the forum, often with the promise that the help-seeker’s problems could be solved. Several previous studies have pointed out a continued cyclical process of personal sharing between users as central to emotional support in forums (Eichhorn, 2008; Smedley et al., 2015; Tichon & Shapiro, 2003; Winzelberg, 1997). One user’s personal story seems to encourage other users to be more open, sparking a loop of sharing where one sharing comment leads to another. In the present study we noted that messages in the Sharing One’s Own Experience theme, often contributed to a sense of community around psychological ill-being in certain threads. The personal stories that are being shared in the forums seem to work as a basis for self-identification and self-comparison. In general, the more active threads in our material are the ones that include personal stories and a sense of community around depression, suicidal thoughts, or general ill-being. The Sharing One’s Own Experience theme could be of particular importance in building a sense of community. Zhu and Stephens (2019) have suggested that two group communication mechanisms could function as mediating factors to explain perceived social support in online support groups: common identity (i.e., identification with the group); and common bond (i.e., having interpersonal bonds with others). It is also interesting to note that sharing of personal experiences can be promoted by the anonymity and invisibility in forums, in what has been termed a benign disinhibition effect (Lapidot-Lefler & Barak, 2015). These disinhibiting factors likely also contribute to the bullying responses, making such features a double-edged sword of forums, contributing to toxic as well as constructive effects.

When a sense of community around mental illness is established, the forum interactions share a resemblance with peer-support interventions that have gained increased popularity in the healthcare system of today (Repper & Carter, 2011; Watson, 2019). These interventions use personal experiences of illness and hardships as a source of knowledge and insight, and the sharing of experiences may reduce the stigma associated with mental health struggles and provide a sense of hope and solidarity. The use of lived experience provides credibility to the users sharing their support and establishes a common ground for the community. As with peer-support work, support offered by others that share personal experiences reduces the risk that they adopt an expert position, instead it is received with a sense of equality (Kaufman & Whitehead, 2018). However, it should also be noted that in some examples, comments from the Sharing One’s Own Experience could also appear as self-involved or even unrelated to the thread starters’ experiences, lending a sense of false identification. It would be interesting for future studies to examine more in detail whether and when help-seekers might perceive sharing of this kind as more or less helpful. 

As previous studies have paid certain attention to the frequency of different forms of support, this aspect was also assessed in the present study. Using the classification of two categories of broad social-support, approximately two-thirds of the coded material in this study would represent instrumental-practical support, and one-third represent emotional or esteem support. In their literature review, Rains et al. (2015) found no difference between the number of emotional and instrumental-practical support messages in different forums, across studies. However, it can be noted that the distribution varied significantly between studies. Out of 41 studies included in their analysis 16 reported roughly equal numbers of messages, while a further 16 showed a preponderance of instrumental-practical support messages, and a further 9 studies reported emotional support messages to be in majority. A number of the included studies also reported distributions comparable to what was found in our study (i.e., two thirds vs. one third; e.g., Coursaris & Liu, 2009; Frohlich & Zmyslinski-Seelig, 2012; Robinson et al., 2011). Optimal matching theory (Cutrona & Suhr, 1992) postulates that the type of social support provided is most helpful when it matches the needs of the help seeker. In that context, depression has been considered a controllable stressor (Rains et al., 2015), and thus an emphasis on instrumental-practical support should be particularly relevant. However, the three types of responses identified here that can be characterized as predominantly instrumental-practical support differ widely in their content and approach. It is also worth noting that in our material, there are almost no support being offered in the forums that can be understood as tangible or network support. Indeed, these forms of support is also less commonly found in the Rains et al. (2015) meta-analytic review, but it is still present in about half of the studies appraised. Again, the type of forum studied might explain this difference. As the studies included in their review focuses on more narrowly scoped discussion forums centered around specific diagnoses or problems, it is possible that this constitutes an important difference between smaller more specific forums (such as support groups), and broader forums as examined in our study. Future research should seek to explain the variations in content found in different types of forums, and its potential impact on the effects of forum participation.

Limitations and Future Research

This study utilizes a naturalistic dataset that does not include any demographic data of the respondents. Previous research has pointed to differences in the utilization of online communities, e.g., by men and women, patients and family members (Ginossar, 2008). Consequently, it is unknown to what extent user characteristics may have impacted the results, and the overall transferability of the findings to specific populations or type of internet forum. It should also be noted that the results are generated from Swedish internet forums and general cultural differences, as well as of the specific forums, may be reflected in the findings. Additionally, we have no data on forum users’ experiences of responses, or their intentions, which precludes an assessment of the reliability of our interpretations through informant feedback. Lack of user involvement in research is a well-known problem in online research (Roberts, 2015), and future research might address this issue by increasing contact with participants. For example, studies based on qualitative interviews might provide insight into users’ experiences of participation in online forums. More research is needed to examine how help-seekers perceive the responses they get and how they use them (especially, it would be relevant to know what instrumental-practical support, if any, is acted on). These questions should also be expanded to include passive users of the forums, as these users might be influenced differently than the original poster. Examining the effects of forum participation for such passive users might be an important understudied area, given that on many health forums most users only read threads (Nonnecke & Preece, 2003), possibly in lieu of seeking help themselves.

Conclusion

Posts about depression or suicidal ideation are met with a multitude of responses in online forums. Most responses appear to constitute genuine attempts to provide support. By looking at the responses, their underlying themes and their characteristics, our findings suggest that a distinction between enacted and perceived support must be considered. Since forum participation is spontaneous and unorganized, it constitutes a situation where every forum reader must interpret and choose between a multitude of viewpoints, tips, and attempts at support. Many times, the support may also be unskillfully or harshly expressed, or miscarried. This considerable complexity might offer an explanation to the variation in effects of forum participation found in previous studies. It is also interesting to note that when users are asked about the reasons for turning to online forums for help, a breadth of reasons are generally reported that encompass all the different help-oriented themes (excluding the bullying theme) found here (Hanley et al., 2019; Prescott et al., 2019). Hence, users appear to appreciate, or at least not be deterred, by this complexity. From a healthcare perspective, the study highlights the importance of other complimentary sources that can provide more personalized information and guidance to people suffering from depression and suicidal thoughts, such as evidence based online health services. Particularly, mental health professionals also need to have an understanding of the idiosyncratic nature of different ideas that individuals may encounter when seeking help online and that may influence how they then approach traditional health-care services.

Conflict of Interest

The authors have no conflicts of interest to declare.

Acknowledgement

We would like to thank Tove Lundberg at department of psychology, Lund University, who contributed to the conception and design for this paper.

Ethical approval for the study was obtained from Swedish Ethical Review Authority in 2020.