The Usage of Digital Resources by Swedish Suicide Bereaved in Their Grief Work: A Survey Study.

This study examined Swedish suicide bereaved individuals' use of different resources in their grief work and how they value these resources. The material consisted of a web-based survey, which was analyzed with quantitative methods. The results showed that the psychosocial ill-health was severe among the suicide bereaved participants and that a majority used digital resources in their grief work. The propensity to engage in online support groups or memorial websites was not predicted by the severity of psychosocial consequences following the suicide. However, multiple regressions showed that higher online support group activity predicted more satisfaction with current psychosocial health, while memorial websites seemed to have the opposite effect. This study not only indicates that some digital resources, for example, online support groups, may be an effective way of coping with grief related to suicide loss, but also suggests that memorial websites may increase rumination and in this way cause emotional distress.

Introduction

Digital communication on the Internet has radically changed the possibilities for people to manage and communicate about different types of grief and mourning. It has been estimated that more than half of all bereaved make use of digital resources in their grief work (Van der Houwen, Stroebe, Schut, Stroebe, & Van den Bout, 2010) and that these resources could have a positive impact on bereaved people’s quality of life (Vanderwerker & Prigerson, 2004). Digital resources for suicide bereaved are multifaceted and diversified, and include passive and relatively static information websites on how to manage grief and loss, active memorial websites, interactive online support groups, and online therapy and counseling (Krysinska & Andriessen, 2013). Among these, online support groups, memorial websites, and social media appear to be the most frequently used digital resources by the bereaved (Krysinska & Andriessen, 2017). Also, digital grief support seems to be more often searched for and used by bereaved women than men (e.g., Chapple & Ziebland, 2011; Feigelman, Gorman, Beal, & Jordan, 2008; Oliveri, 2003).

Regarding memorial websites, they are usually set up by the family or friends of the deceased, and could be in the form of individual freestanding webpages, webrings (where the webpages of the deceased for example share the same cause of death), or more formalized online memorial websites, often established and run by commercial or noncommercial organizations (Krysinska & Andriessen, 2013). Memorial websites seem to be used by bereaved people as a complement to traditional rituals of grief, mourning, and memorializing, and it has been pointed out that they serve to both enhance continuing bonds with the deceased and to communicate and share the grief with others (e.g., Roberts, 2004). Chapple and Ziebland (2011, p. 181) write that “A key feature of on-line memorials is that they are modifiable projects unbounded by the local space that physical memorials inhabit.” When the bereaved visit their memorial website, they can “talk to the dead” and continue to share things that happen in the world of the living, and in the age of digital connectivity, people can do this practically anywhere and at any time. Bereaved people who have lost a relative or close friend in suicide more often write about the specific cause of death on their memorial website than other bereaved (Roberts & Vidal, 2000), and they also use longer sentences and more complex writings about anger and sadness, which in turn indicates a high level of emotional distress (Lester, 2012). This can probably be explained by the shock and the difficulty to understand the causes and the meanings of these kinds of deaths.

The ability to communicate virtually around the clock, to find practical information, to receive and provide emphatic support, to inform about the death, to share experiences, and discuss taboo and stigmatized issues, are key factors that have been highlighted as reasons for why bereaved people join online grief and support groups (Chapple & Ziebland, 2011; Feigelman et al., 2008; Hollander, 2001; Oliveri, 2003). During the past years, these groups are also often constructed as open or closed Facebook groups. Previous studies have also shown that participants in online support groups find it easier to discuss grief issues with their group than with family members or friends, as the group members show better understanding and are easier to reach for support and help (Feigelman et al., 2008; Oliveri, 2003; Pendry & Salvatore, 2015). Participation in different kinds of digital grief forums can bring bereaved people together in their grief work to form communities and also increase the possibility to maintain and enhance their relationships and bonds to the deceased (Bailey, Bell, & Kennedy, 2015; Moss, 2004; Roberts, 2004; Sofka, Cupit, & Gilbert, 2012), and a majority of the users seem to find the support groups beneficial and also show some improvement regarding well-being and depression (Kramer et al., 2015). However, these new patterns of mourning and memorizing have also raised concerns that different digital resources may increase and amplify negative ruminations and a reluctance to “let go” of the deceased (e.g., Stroebe, Van der Houven, & Schut, 2008).

Suicide is one of the major causes of death in the world, leading to nearly 800,000 deaths per year (World Health Organization, 2015). In 2020, the number of deaths caused by suicide is expected to reach 1.5 million (Bertolote & Fleischmann, 2009). In Sweden, with a population of 10 million people, approximately 1,500 individuals commit suicide every year (Public Health Agency of Sweden, 2017). Beyond the tragic of loss of individual lives, every death to suicide has a ripple effect and is estimated to deeply affect 6 to 14 relatives and close friends (Clark & Goldney, 2000; Jones & Meier, 2011; Jordan & McIntosh, 2011b). Also, from a long-term perspective, what seems to influence the bereaved most in their grief is not primarily a kinship relation, but the perception of closeness to the deceased (Cerel, Maple, Aldrich, & Van de Venne, 2013).

Grief reactions and bereavement should not primarily be regarded as a pathological condition (Wagner, Knaevelsrud, & Maercker, 2005), but rather as a psychosocial and existential limit situation (Jaspers, 1932/1970) triggered by an overwhelming and disruptive loss in life. These are situations where people are utterly thrown into a new life situation, torn away from what they thought was safe and secure in life—an experience which today is often digitized and highly entangled with digital media and online practices (Lagerkvist, 2016). Besides the death of a child, homicide, and clearly unpredicted deaths, suicide is often stated as a disastrous event, where relatives and friends often search for answers and explanations as to why the suicide occurred (Neimeyer & Sands, 2011). What further complicates the grief reaction after a suicide is that a suicidal person in some sense intentionally “chose” to die, which can produce existential doubts and strong feelings of anger and resentment among those still alive (Bell, Stanley, Mallon, & Manthorpe, 2012; Lester, 2012).

Some aspects of grief then appear to be especially poignant after a suicide, such as guilt, shame, shock, anger, and a prolonged and intense period of grief (Bell et al., 2012; Grad, 2005; Jordan & McIntosh, 2011a; Ratnarajah & Maple, 2011). This can in some cases lead to what is referred to as complicated or traumatic grief (Jacobs & Prigerson, 2000; Wagner et al., 2005). To this can be added a social and cultural stigma that is closely associated with suicide acts, which for the bereaved can cause strong emotions of feeling rejected, isolated, and misunderstood. The stigma may lead to the grief not being accepted and socially sanctioned, and this can result in the bereaved individuals’ withdrawal from their sociocultural context (Bell et al., 2012). Doka (1989) has termed this social reaction to the suicide stigma a disenfranchised grief. However, a systematic review of the literature on suicide bereavement among adolescents showed that grief and mourning after a suicide does not always have to be a more problematic issue than other forms of bereavement, at least not among younger people (Andriessen, Dudley, Draper, & Mitchell, 2017). This does not take away the severity of suicide bereavement, but, again, points to that the grief and mourning that follows a suicide do not always have to be considered as pathological.

In a Swedish context, the suicide survivors’ organization SPES (National Association for Suicide Prevention and Survivors' Support, 2017; http://spes.se) is the leading force for organizing suicide bereaved people and offer different kinds of resources and support to cope with this specific form of grief and mourning. Survivor support groups and organizations have historically played an important role in helping other suicide bereaved people in their grief work and have also been active in suicide prevention and the work against stigma and shame (McIntosh, Bolton, Andriessen, & Campbell, 2017), and studies have shown that survivors often find these support groups helpful in their grief work (McIntosh, 2017). SPES was formed by suicide bereaved people in 1987, and today, they have around 1,500 members. This kind of suicide survivors’ organizations do exist in many other countries, and SPES is a part of the International Network of Survivors of Suicide. SPES online resources can be seen as an extension and expansion of the organizations’ information, meetings, support, and counseling activities that are carried out offline (Westerlund, 2010). Overall, a majority of the websites and digital resources that are available for suicide bereaved people are created and maintained by the bereaved themselves (Krysinska & Andriessen, 2010). The digital resources are built around a main website (http://spes.se), which includes information about grief and bereavement, upcoming support meetings, regional networks, telephone helpline, digital resources as well as a digital guest book, on which visitors can write messages and get support anonymously. SPES are also very active on Facebook and operate three open and two closed Facebook groups (https://facebook.com). The two closed groups have more than 2,000 members and can be said to function as digital complements to the face-to-face support groups that SPES organize in different regions in Sweden.

The main objective of the present study is to examine Swedish suicide bereaved individual’s use of different resources in their grief work, and how they value the strengths and weaknesses of these resources. The specific research questions concern whether there are any differences between users and nonusers of digital resources depending on age, gender, education, relation to the deceased, time elapsed after the suicide, and perceived psychosocial health. The study also aims to investigate if the usage of two of the most common types of digital resources (i.e., online support groups and memorial websites; Krysinska & Andriessen, 2017) are differently associated with psychosocial health for the suicide bereaved.

Method

Participants and Procedure

The present survey study is part of a larger research project, which include the issue of how Swedish suicide bereaved handle loss, grief, and mourning in the era of digital communication and exchange. Ethical approval for the study was granted by the Regional Ethical Review Board in Stockholm (2016-02-11) and a collaboration was established with the aforementioned suicide survivors’ organization SPES, in order to gather participants for the survey study. The questionnaires and implementation of the survey was also discussed and approved by representatives of the organization.

In March 2016, a web-based survey addressed to individuals (over 18 years) who had lost a relative, friend, or a significant other to suicide was announced and made available on SPES official website (http://spes.se) and Facebook groups, with permission from the administrator of the website and Facebook groups. The announcement stated that participation was voluntary, and that the purpose with the survey was to investigate which resources suicide bereaved people use in their grief work and how they value these resources, and also how the loss had affected their life situation and psychosocial health. In the announcement, it was also stated that participation in the survey could be terminated at any time during the procedure, in which case the participant’s data would be automatically deleted. The participants could also choose to skip questions in the survey, for example, if they found them painful to answer or if they did not apply to them. There was also detailed information about where and how the participants could get help and support if they felt anxious, depressed, or had suicidal ideations.

The survey was accessible for respondents on the website and Facebook groups between March 18 and September 4, 2016, and all responses were collected during the same period. The responses were anonymously received and saved digitally on a secure server, with no individual identifiable data. The online survey system was set so that only one questionnaire response per Internet Protocol address could be submitted. The study’s design was cross-sectional, and no follow up was possible due to the anonymous, nonidentifiable responses. In total, 327 suicide bereaved responded to the survey.

Variables and Scales

The total survey contained 21 questions with fixed response alternatives and one open-ended question. In the present study, 13 of the fixed questions were analyzed. Eight of those questions had the form of Likert scales, where the respondents were asked to rate different statements from 1 to 5. The questionnaire items and variables were originally constructed by the author on the basis of the study’s objective, and in line with previous studies in the field (e.g., Feigelman et al., 2008; McMenamy, Jordan, & Mitchell, 2008). Furthermore, the survey questions were constructed and formulated with the specific target group in mind. For example, as a relatively large portion of suicide bereaved have a sceptical, sometimes negative, attitude toward psychiatry, mental health care, and the pathologization of grief (cf. Feigelman et al., 2008; Schotanus-Dijkstra et al., 2014; Westerlund, 2010), commonly used and well-known measurements, like clinical depression scales, were not used. In this case, the participants were not asked to self-assess “depression” but “nedstämdhet” (i.e., “depressed mood” or “feeling low”), which, at least in a Swedish context, is a less pathological definition of the consequences of grief and mourning.

Data were collected regarding three different areas: (a) demographic data including age, gender, education, relation to the deceased, and time elapsed since the suicide; (b) perceived psychosocial consequences that followed the loss of a relative or significant other in suicide as well as satisfaction with current psychosocial health; and (c) evaluation of the different resources (off- and online) that were used by the participants in their grief work.

Beside the variables age and gender, the level of education was also measured in the study. The three respondent alternatives were “University education,” “High-school education,” or “Primary school education.” Regarding the relationship to the deceased, the participants could choose from eight different alternatives (“Partner,” “Child,” “Friend,” etc.; see Figure 1). The time elapsed after a relative’s or a significant other’s suicide was measured with a scale from less than 1 year to more than 10 years (see Figure 3).

Figure 1.

Relation to the deceased.

Figure 3.

Time elapsed after the relative’s suicide; the proportion of participants who lost a relative within a given timeframe.

The bereaved respondents’ overall use of different resources (both off- and online) in their grief work, and the perceived helpfulness of these resources, was measured with a scale consisting of 15 statements about different sources of support (see Table 2). Each source of support was rated from 1 to 5, where 1 was Not helpful at all and 5 was Very much helpful. The internal consistency among these 15 items was alpha = .69.

Table 2.

Rated Helpfulness of Various Sources of Support.

Source of support	Rated helpfulness		Association with support group activity		Association with memorial websites activity
	M	SD	r	p	r	p
Family	3.19	1.18	.11	.060	−.05	.414
Books, articles & support material	3.10	1.08	.16	.004	−.05	.417
Others with similar experience	3.03	1.35	.22	<.001	.00	.983
Online support group	2.97	1.24	.54	<.001	.08	.169
Individual therapy	2.97	1.38	.08	.161	−.11	.055
Offline support group	2.73	1.60	.01	.896	−.05	.439
Priest/Deacon	2.35	1.48	.02	.741	.04	.556
Coworkers	2.29	1.23	.08	.202	−.03	.667
Primary care	2.02	1.18	.09	.123	.05	.379
My Facebook page	1.88	1.08	.17	.009	.13	.047
Psychiatric care	1.79	1.11	.06	.353	−.04	.531
Facebook page of the deceased	1.78	1.11	−.02	.768	.28	<.001
Occupational health service	1.62	1.12	.07	.333	−.09	.202
Memorial website	1.60	.95	−.05	.494	.35	<.001
Online blog	1.50	1.04	.08	.322	.00	.987
*All scores range between 1 and 5, where higher scores indicate more helpfulness.

Note. Sources are ranked from least to most helpful according to their score in the total sample. Correlation with online support group and memorial website use are also presented.

Participants reasons for engaging in online support groups (if they did) was measured with a scale consisting of nine items (“Seeking and sharing information,” “Help to cope with commemoration days,” etc.; see Table 3). Each item was rated on a Likert scale from 1 to 5, where 1 was Not important and 5 was Very important. The internal consistency among these nine items was alpha = .91.

Table 3.

Correlates of Online Support Group Activity (Raw Scores and Pearson Correlation).

Motive for use	Mean	SD	Pearson r	p
Seeking and sharing information	3.64	1.18	.29	<.001
Help to cope with commemoration days	3.80	1.33	.29	<.001
Opportunity to communicate around the clock	3.62	1.36	.28	<.001
Receiving support and comfort	3.77	1.28	.25	<.001
Sharing experiences and helping others	3.91	1.23	.25	<.001
Easy accessibility	3.59	1.38	.22	.001
To openly discuss grief-related issues	4.03	1.14	.21	.001
Meeting others with similar experiences	3.84	1.19	.20	.002
Contributing to suicide preventive work	3.95	1.18	.06	.308

Note. All scores range between 1 and 5, where higher scores indicate more importance. Ranked from least to most important, based on correlation size.

The participants’ visitation rate on online support groups and memorial websites was measured with a set of five items each (see Figure 2 and Table 1). The first item asked how often the participant visited the support group or memorial website (if they were users) and was rated on a Likert scale from 1 to 5 (1 = Never; 2 = Very seldom; 3 = About once a month; 4 = Several times a week; 5 = Every day). The remaining four items asked how often the participant “Reads posts,” “Writes posts,” “Posts multimedia content,” and “Posts links” on the support group or memorial page (see Table 1). These statements were also rated on a Likert scale from 1 to 5, where 1 was Never and 5 was Very often. Participants who did not use digital resources were asked to skip the aforementioned questions. Subjects with missing values on these items were assumed to have not engaged in said activity, and a score of 0 was imputed in those cases (thus, participants who did not use digital resources at all were not excluded from analyses, but had an activity level of 0). The five items of each scale were then averaged to calculate an index of support group activity and memorial page activity, respectively. The internal consistency of the five items measuring support group activity was alpha = .89 and the internal consistency of the five items measuring memorial website activity was alpha = .93.

Figure 2.

Shows the rate at which the respondents visit online support groups and memorial websites (N = 327).

Table 1.

Rated Activity on Online Support Groups and Memorial Websites.

Type of online page	Activity	Mean	SD
Support groups	Visitation rate	3.09	1.78
	Reads posts	2.76	1.74
	Writes posts	1.54	1.04
	Posts multimedia content	1.20	1.00
	Posts links	1.14	0.97
	Total activity rate	1.95	1.12
Memorial websites	Visitation rate	1.17	1.68
	Reads posts	1.02	1.59
	Writes posts	0.68	1.10
	Posts multimedia content	0.61	1.00
	Posts links	0.45	0.74
	Total activity rate	0.78	1.13
Support groups and Memorial websites	Overall activity rate	1.37	0.88

Note. All scores range between 0 and 5, where higher scores indicate higher activity (0 = Not a user).

A scale consisting of nine statements was used to ask the bereaved about the feelings and perceived psychosocial consequences that followed in close proximity to the suicide. These both psychological and social consequences were Depressed mood, Meaninglessness, Anxiety, Sleep problems, Guilt, Difficulty managing daily routines, Loneliness and isolation, Anger, and Shame (see Table 4). The selection of the different consequences was partly based on commonly used terms in previous studies on the subject (e.g., Kramer et al., 2015; McMenamy et al., 2008). Each feeling or consequence was rated from 1 to 5, where 1 was Nothing and 5 was Very much. The nine items were then combined to create an index of the severity of psychosocial consequences that followed the suicide, which in the analyses was correlated with other variables (see Statistical Analyses section). The internal consistency among the items was alpha = .84. However, a Principal Component Analysis (PCA) was also conducted to explore the factor structure of this scale (varimax rotation was used for this analysis but direct oblimin rotation produced similar results). The PCA suggested a two-factor solution (see Results section), and these factors were analyzed in addition to the entire scale. Satisfaction with present psychosocial health was measured with one item that asked how satisfied the participants were with their current psychosocial health (at the time the participants completed the survey). It was rated on a Likert scale from 1 to 5, where 1 was Not satisfied at all and 5 was Very satisfied.

Table 4.

Previous Perceived Psychosocial Consequences Following the Suicide Event, Ranked From the Least to Most Common Consequence.

Perceived psychosocial consequences	Mean	SD	Rotated factor loadings
			Factor 1	Factor 2
(a) Depressed mood	4.45	.85	.80	(.11)
(b) Meaninglessness	3.84	1.27	.67	(.29)
(c) Anxiety	3.81	1.27	.77	(.26)
(d) Sleep problems	3.68	1.30	.76	(.16)
(e) Guilt	3.63	1.35	(.39)	.66
(f) Difficulty managing daily routines	3.26	1.23	.77	(.12)
(g) Loneliness and isolation	3.19	1.32	.60	(.34)
(h) Anger	3.10	1.43	(.05)	.70
(i) Shame	2.30	1.43	(.23)	.78
(a–i) Total psychosocial consequences	3.48	.85
(j) Satisfaction with current psychosocial health (1 item)	2.37	1.18

Note. Participants also rated satisfaction with their current psychosocial health, which is shown in the bottom field. The two columns to the right show the rotated factor loadings in the PCA analysis. All scores range between 1 and 5. On items (a–i) higher scores indicate more severe psychosocial problems. On item (j) (Participant’s satisfaction with their current psychosocial health), higher scores indicate more satisfaction with their psychosocial health. Rotated factor loadings below .4 are shown in parenthesis.

Statistical Analyses

Several analyses were performed to investigate potential predictors of online support group and memorial website usage. Usage was considered both in terms of dichotomous membership (i.e., being a user of a support group or memorial website vs. not being a user) and in continuous terms (i.e., being more or less active on support groups or memorial websites).

With regard to demographic background variables, independent samples t tests were calculated to see if there was an age difference between users and nonusers of digital resources, and χ2 tests were calculated to investigate the difference in gender and level of education between users and nonusers.

To investigate what impact time had on participant’s behaviors and psychosocial health, independent samples t tests were calculated to see if users and nonusers differed in how much time had elapsed since the suicide, and Pearson correlations were calculated to investigate the association between time and the level of activity on support groups and memorial websites. The time elapsed after the suicide was also correlated with participant’s satisfaction with their current psychosocial health, to see if newly exposed participants had more psychosocial problems compared with those that were exposed a longer time ago (i.e., to see if there could be a positive effect of time alone). Satisfaction with current psychosocial health was also correlated with the rated helpfulness of different sources of support (the average score of all items), to see if participants who perceived more support were more satisfied with their current psychosocial health.

Furthermore, perceived psychosocial health, both at the time of the suicide and at present, was correlated with online support group and memorial website activity, to see if participants with more psychosocial problems used these resources more extensively. Activity on these digital resources were also correlated with the rated helpfulness of various sources of support (including offline resources), to see if the perceived helpfulness, or the lack of it, in some areas would predict their use. Participants also ranked different motives for using online support groups, and these were correlated with their activity on support groups, to see if their motives were associated with their actual behaviors.

Tests were also performed to see if participants’ relationship to the diseased (“Partner,” “Child,” “Friend,” etc.; see Figure 1) influenced their use of digital resources. As a large proportion (40.1%) of the participants had lost a child to suicide, this group was contrasted to all other participants in these analyses. χ2 tests were used to investigate if these two groups differed in regard of being a user or nonuser, and t tests were calculated to test the difference in activity level. A t test was also calculated to determine if those who had lost a child to suicide had experienced more psychosocial problems than other participants.

Moreover, two analyses were conducted to investigate how the use of digital grief resources impacts the psychosocial recovery from the suicidal event. First, a multiple regression model was calculated to predict satisfaction with current psychosocial health. The predictors in this model were the level of support group activity, the level of memorial website activity, the time elapsed since the suicide, and the psychosocial health at the time of the suicide (i.e., baseline levels of psychosocial health). The two latter predictors were included because they were hypothesized to influence both how much participants use digital grief resources, and also their satisfaction with current psychosocial health.

Finally, an analysis of covariance model was calculated to more precisely estimate how satisfaction with current psychosocial health is predicted by the interaction between time and use of digital resources, while controlling for the baseline level of psychosocial problems. The independent variables (fixed factors) in this model were as follows: time elapsed since the suicide (trichotomized at three levels: 0 to 3 years, 3 to 6 years, and 6 + years) and use of any digital resource (online support groups or memorial websites or both: Yes/No). Psychosocial health at the time of the suicide was treated as a covariate. In other words, this model tested if users were more satisfied with their current psychosocial health than their nonusing peers, given that they have had the same recovery prognosis to begin with, and the same time to recover. All statistical analyses were performed in IBM SPSS statistics v23 with an alpha level of .05.

Results

Demographic Variations of Digital Resources Usage

The final sample consisted of a total of 327 respondents, aged between 18 and 79 years (mean age = 47 years, SD = 13 years), of which 90% were women. About 56% of respondents had a university education, 36% a high school education, and 8% a primary school education. About 40% of the respondents had lost a child to suicide, which was the most common experience in terms of a relative's suicide (Figure 1).

Figure 2 shows the rate at which the respondents visit online support groups and memorial websites, respectively. In the total sample, 260 respondents (80%) reported that they were members of an online support group, but 3 (1%) reported that they never visited it. Thus, there were 257 (79%) active support group users in the sample (Figure 2, left side). Further, 126 (39%) respondents reported that they had utilized memorial websites, but 20 (6%) reported that they never visited it. Thus, there were 106 (32%) active memorial website users in the sample (Figure 2, right side). Taken together, there were 271 (83%) respondents who used either support groups or memorial websites (digital resources) to some degree and 56 (73%) respondents who never used such resources or did not use them actively.

The mean activity rate was higher for online support groups (M = 1.95; SD = 1.12) than for memorial websites (M = .78; SD = 1.13), and the mean activity rate for both support groups and memorial websites was all together 1.37 (SD = .88; Table 1).

There was no age difference between users and nonusers of online support groups, t(323) = −.085; p = .932, or memorial websites, t(323) = .116; p = .908. As for gender differences, women were more likely than men to be members of online support groups (81% vs. 61%; χ21 = 6.794; p = .009; Phi = −.145), and they also show higher support group activity, t(322) = 2.752; p = .006; Cohen’s d = .51. As for memorial websites, there was no such gender difference regarding being a member (38% vs. 39%; χ21 .003; p = .958), or in activity rate, t(322) = 1.046; p = .296. Users and nonusers of online support groups did not differ regarding their level of education, χ22 = 3.253; p = .197, but as for memorial websites, users were less likely to have a university education (45% vs. 62%) and more likely to have a high school education (44% vs. 31%) or elementary school education (11% vs.7%) compared with nonusers (χ22 = 9.023; p = .011; Cramer’s V = .167).

Figure 3 shows how much time had elapsed after the relative’s suicide. The figure shows, for example, that it was most common for participants to have lost a relative less than a year ago (approximately 21% of the sample). The median time since the suicide took place was 2 to 3 years, and the respondents’ use of digital resources was partially dependent on the time that had elapsed. A more recent loss was associated with a greater likelihood of being a member of an online support group, t(325) = 2.230; p = .026; Cohen’s d = .29, as well as higher support group activity (r = −.14; p = .010). Conversely, the time elapsed after suicide did not differ between users and nonusers of memorial websites, t(325) = .562; p = .575, and did not correlate with memorial website activity (r = −.06; p = .246).

The association between the use of digital resources and the personal relation to the deceased was also tested. As approximately 40% of the respondents had lost a child to suicide, this specific group was contrasted to all other respondents. Respondents who had lost a child were not more likely than other respondents to be members of online support groups (χ22 = 1.153; p = .283), but nonetheless they had higher support group activity, t(325) = −3.181; p = .002; Cohen’s d = .36. Similarly, respondents who had lost a child were not more likely to be members of memorial websites (χ22 = 3.043; p = .081) but had higher memorial website activity, t(325) = −2.932; p = .004; Cohen’s d = .33.

Help-Seeking and the Use of Different Resources

The study participants were asked to rate the helpfulness of various sources of support that they had used in their grief process. The mean scores are presented in Table 2, where sources of help are ranked from least to most helpful. The correlations with online support group activity and memorial websites activity are also presented in Table 2. As can be seen, Family was ranked as the most helpful source of support by the bereaved, but had no association with how intensely respondents used online support groups and memorial websites. However, higher support group activity was positively predicted by the rated helpfulness of support from Online support group, Others with similar experience, My Facebook page, and Books, articles and support material (in that order). Moreover, higher memorial website activity was positively predicted by the rated helpfulness of Memorial website, Facebook page of the deceased, and My Facebook page (in that order). The other rated sources of support did not correlate with the use of digital resources.

The participants who were members (including inactive members) of online support groups also rated the importance of various reasons for engaging in them. Table 3 shows how different motives were rated, and ranks them from least to most important based on how well they correlate with the respondent’s support group activity. The table shows that the most important reason for engaging in online support groups is Seeking and sharing information and Help to cope with commemoration days. All listed motives were rated relatively high and significantly correlated with support group activity, except for the motive Contributing to suicide preventive work.

Psychosocial Health and Digital Resources

Respondents were also asked to rate the extent to which they had experienced psychosocial health problems due to the suicidal event, and to what extent they were satisfied with their psychosocial health at present (Table 4). Adverse psychosocial reactions were common, and the highest rated consequences following the suicidal event were Depressed mood, Meaninglessness, and Anxiety, and the least common consequence was Shame. Respondents who had lost a child to suicide did not differ from other participants regarding severity of psychosocial problems, t(324) = −1.012; p = .312.

A PCA was conducted to examine the structure of the scale. Bartlett’s test of sphericity was significant (p < .001), and the Kaiser–Meyer–Olkin Measure of Sampling Adequacy was .866, indicating that PCA is appropriate. The analysis suggested a two-factor solution, where six items loaded onto one factor (Factor 1: Depressed mood, Meaninglessness, Anxiety, Sleep problems, Difficulty managing daily routine, and Loneliness and isolation) and three of the items loaded onto another factor (Factor 2: Guilt, Anger, and Shame). The factor loadings (after Varimax rotation) are shown in Table 4. After rotation, the factors explained 37.8% and 20.5% of the variance, respectively. The internal consistency among these two factors was α = .85 and α = .62, respectively, and α = .84 for the entire nine-item scale.

Moreover, respondents’ lack of satisfaction with their current psychosocial health was relatively serious. About 29% of respondents were “not content at all” and 29% were only “partially content” with their current psychosocial health (not shown). Psychosocial consequences experienced due to the suicidal event correlated strongly with satisfaction with current psychosocial health (r = −.40; p < .001). However, the correlation was higher for Factor 1 (r = −.42; p < .001) than for Factor 2 (r = −.23; p < .001).

Participants’ psychosocial health was partially associated with their use of Internet-based resources. Support group activity was not significantly associated with psychosocial problems following the suicidal event (r = .10; p = .068) or satisfaction with current psychosocial health (r = .04; p = .521). Neither was memorial website activity associated with psychosocial problems following the suicidal event (r = .10; p = .062), but it was negatively correlated with satisfaction with current psychosocial health (r = −.12; p = .026). Moreover, satisfaction with current psychosocial health was correlated with the amount of elapsed time after the suicide (r = .18; p = .001), as well as the rated total helpfulness of support sources (see Table 2; r = .18; p = .001). Thus, less satisfaction with current psychosocial health was observed among participants who had lost someone to suicide relatively recently, perceived different sources of support as less helpful, and were more active on memorial websites.

To investigate these relationships in more detail, a multiple regression model was calculated to test if the use of online support groups and memorial websites predicts satisfaction with current psychosocial health, while accounting for the time elapsed since the suicide and psychosocial problems experienced due to the suicide (i.e., baseline levels of psychosocial health). The model was significant, F(4, 318) = 21.767; p < .001; r2adj = .21, as was each individual predictor, except for memorial website activity (Table 5). As expected, previous psychosocial problems were negatively associated with satisfaction with current psychosocial health (std. beta = −.41), while time elapsed after suicide was associated with more satisfaction with current psychosocial health (std. beta = .19). Further, online support group activity was a positive predictor of satisfaction with current psychosocial health (std. beta = .13). Memorial websites activity was an almost significant predictor of satisfaction with current psychosocial health, but in the opposite direction (std. beta = −.10; p = .052).

Table 5.

Multiple Regression Predicting Satisfaction With Current Psychosocial Health.

	Std. Beta	95% CI for Std. Beta		t	p
		Lower	Upper
(Constant)				.017	.986
Previous psychosocial health	−.41	−.50	−.31	−8.082	<.001
Time elapsed after suicide	.19	.10	.29	3.863	<.001
Online support group activity	.13	.03	.23	2.555	.011
Memorial website activity	−.10	−.20	.00	−1.946	.052

Accounting for age, gender and relationship (child vs. other) did not significantly alter these results, F change(3, 312) = .376; p = .771; not shown in table.

As noted earlier, satisfaction with current psychosocial health was correlated with previous psychosocial problems, but Factor 1 was more correlated than Factor 2. Therefore, the regression analysis was rerun with these two factors separated instead of combined. The results showed that Factor 1 was significantly related to the outcome variable (std. beta = −.41; 95% CI: −.52 – −.29; p < .001) but not Factor 2 (p = .583). Moreover, when omitting Factor 2 from the analysis, the effect of memorial website activity became statistically significant (std. beta = −.10; 95% CI: −.20 – .00; p = .050). All other results were virtually unchanged and are therefore not presented in more detail.

A final analysis was performed to test the interaction effect between time elapsed after suicide and the use of digital resources. A two-way analysis of covariance was calculated with satisfaction with current psychosocial health as a dependent variable, previous psychosocial health problems as a covariate, and two independent factors: time elapsed since the suicide (<3 years, 3–6 years, and > 6 years) and the use of any Internet-based resource (Yes/No). The model as a whole was significant, F(6, 316) = 13.551; p < .001; η2partial = .21, and there was a significant main effect of time elapsed after suicide, F(2, 316) = 4.157; p = .017; η2partial = .03, and previous psychosocial health problems, F(1,316) = 57.273; p < .001; η2partial = .15, but no main effect of Internet-based resources, F(1, 316) = 2.436; p = .120. There was a tendency toward an interaction effect (illustrated in Figure 4), such that users of Internet-based resources improved faster regarding their satisfaction with current psychosocial health than nonusers during the first 3 years following the suicide, and then declined slightly during later years, while nonusers improved slower but steadier. However, this interaction effect did not reach statistical significance, F(2, 316) = .171; p = .843. A power calculation showed low statistical probability to detect any interaction effect (Power = .08). The results did not substantially change when previous psychosocial problems were substituted with only Factor 1 or Factor 2 of the scale and are therefore not presented in more detail.

Figure 4.

Participants’ satisfaction with their current psychosocial health as a function of time (accounting for the effect of previous psychosocial health problems), presented separately for participants who use any kind of Internet-based resources and those who do not.

Discussion

What Are the Differences in the Use of Digital Resources?

The major respondent group in the study consisted of middle-aged women who lost a child to suicide (Figure 1). Women were also more likely than men to be members of online support groups, and they were also more active in their use (cf. Chapple & Ziebland, 2011; Feigelman et al., 2008; Oliveri, 2003). It has been argued that different grieving patterns exist for men and women, for example, that women are in general more emotionally expressive and that they to a higher degree socialize with others in their grief compared with men (e.g., Doka & Martin, 2010). Based on the present results, it appears as if these culturally and historically developed gender-specific grieving patterns have been transferred to and partly repeated on digital platforms. Regarding memorial websites, there were no significant gender differences concerning being a member (cf. Roberts & Vidal, 2000) or regarding activity level. This can probably be seen in light of that socializing in grieving and mourning seem to be in conflict with the (western) traditional male role (Grad, Treven, & Krysinska, 2017); visiting and spending time on a memorial website does not require the same level of involvement and communication with others (except with the deceased) as being engaged in an online support group. Visiting a memorial website is likely considered a more solitary activity.

The analysis did not show any significant age differences between users and nonusers of online support groups or memorial websites, and neither were there any differences regarding the level of education between users and nonusers of online support groups. However, users of memorial websites seem to have, on average, a lower level of education than nonusers.

The present study shows that a majority of the respondents use digital resources (i.e., online support groups and/or memorial websites) in their grief work (Figure 2) and that reading posts is the most common activity (Table 1). The visitation rate among the users was higher in online support groups than on memorial websites, which indicates that the participants find it more important, or urgent, to recurrently visit support groups than memorial websites. The median time elapsed after a relative’s suicide was 2 to 3 years, and the analyses show that a more recent loss affects the probability of joining an online suicide support group. Moreover, the participants who had experienced a more recent loss had also a higher level of support group activity, which suggests that meeting others with the same sad experiences for support and sense-making is mostly wanted in the beginning of the grief period. Regarding memorial websites, there was no such significant correlation between time elapsed after suicide and the participants use and level of activity, which further points to the difference between these two digital resources.

Even if the majority of the respondents in the survey had lost a child to suicide, this group was no more likely to join an online support group or a memorial website than respondents that had lost another relative. However, those who lost a child showed a higher activity rate in their use of digital resources.

Which Resources Are Perceived as Most Beneficial?

Family was stated as the most important source of help in the grief work by the respondents, closely followed by Books, articles and support material and Others with similar experience (Table 2). Also, Online support group and Individual therapy were rated relatively high.

In contrast to the aforesaid, Psychiatric care, Primary healthcare, and Occupational health service (Table 2) were rated relatively low. This is in line with previous research where bereaved are very concerned about the scarcity of community resources and the lack of response, empathy, and competence that they have felt when reaching out for help and guidance in their grief and despair (e.g., McMenamy et al., 2008). Also, many bereaved by suicide have had a negative experience with mental health practitioners, having experienced how they failed to help and treat their loved ones while they still were alive (cf. Feigelman et al., 2008; Schotanus-Dijkstra et al., 2014). Moreover, digital resources such as Memorial website, Online blog, or Individual Facebook page were not considered to be especially helpful by the bereaved.

The highest ranked source of support Family was not correlated with the use of online support groups or memorial websites, which was also the case regarding most other sources of support (Table 2). The present analyses suggest that most of the respondents are using digital resources independently of how they value other forms of help, that is, that a respondent devalues psychiatric or primary care does not mean that she or he “automatically” seeks out digital resources as an alternative or complement. This noncorrelation also applies in the other direction; a respondent that, for example, values individual therapy can also use and value an online support group as very helpful. In summation, it seems that the bereaved respondents experience that they cannot get “too much” help in their grief work, or alternatively stated: the bereaved experience that there is insufficient help to ease the pain and heal the wounds after a loved one’s suicide.

The results showed that all listed reasons to engage in online support groups were rated as important (mean scores higher than 3; Table 3). This is consistent with previous research that has shown that online forums developed for and by suicide bereaved are valued digital “places” where the participants can gain recognition and social support (Kramer et al., 2015; Schotanus-Dijkstra et al., 2014), and that the only ones that “truly can understand” what losing someone to suicide means are other people with the same experience (cf. Feigelman et al., 2008).

The analysis also showed that all but one listed reasons and motives were significantly correlated with the bereaved online support activity (Table 3). For example, the high valuation for the alternative Seeking and sharing information means that it correlates with high activity in the support group. However, although Contributing to suicide preventive work was rated high (mean score = 3.95 of 5), the association with actual support group activity was nonsignificant. This indicates that even if the respondents rated this alternative as important, it is not proportional to the actual degree of activity in the support group. Interestingly, this may suggest that contributing to suicide preventive work is a socially accepted motive that the bereaved feel that they are “obligated” to rank high, even if they actually value other more pragmatic and individualistic reasons as more important in their grief work and presence in online support groups.

Do Digital Resources Have any Impact on Suicide Bereaved People’s Psychosocial Health?

The fact that the respondents lost a relative or a close friend to suicide makes them a very vulnerable group (e.g., Hollander, 2001; Jacobs & Prigerson, 2000; Wagner et al., 2005). In line with this, the results showed high values for almost all perceived negative consequences at the time of the relatives’ suicide, that is, Depressed mood, Meaninglessness, Anxiety, and so forth (Table 4; cf. Dyregrov, 2002; McMenamy et al., 2008). Moreover, the majority of the respondents were not satisfied with their current psychosocial health, which also correlated strongly with the perceived negative consequences due to the suicide event. Many bereaved by suicide are at high risk of complicated or traumatic grief, which among other reactions can include the inability to accept the loss and strong feelings of purposelessness and hopelessness about their lives and the future (Neimeyer, Burke, Mackay, & Van Dyke Stringer, 2009; Stroebe, Schut, & Stroebe, 2007). Yet, the analyses indicated that the bereaved persons’ satisfaction with their current psychosocial health correlated positively with longer time elapsed after the loss. Also, more satisfaction with current psychological health was observed among participant who perceived more support from different sources, among the digital resources. Notably, the respondents who lost a child to suicide did not show more severe psychosocial distress than other respondents, suggesting that it is the loss to suicide in itself that is the primarily devastating experience and profound limit situation for those that are closely related to the deceased (cf. Cerel et al., 2013).

Several previous studies have pointed to shame as a clear characteristic in the grief process after a suicide (Bell et al., 2012; Grad, 2005; Jordan & McIntosh 2011a; McMenamy et al., 2008; Ratnarajah & Maple, 2011), but this was not the case for the respondents in the present study. On the contrary, shame was the only negative consequence that was rated relatively low by the bereaved (Table 4). Even if this is just the result from one single study, and that the explanation for this is probably complex and intertwined in more overarching processes, it may partly be understood in terms of the increased communication and information about suicide and its underlying causes on various Internet platforms (Westerlund, Hadlaczky, & Wasserman, 2012). This increased communication may have had a “normalizing” effect on the subject and thereby contributed to a gradual reduction of the shame, taboo, and stigma that historically, culturally, and religiously have surrounded the subject, at least for the Swedish bereaved in the present study, the majority of whom are users of digital resources and living in a highly secular culture and society (World Values Survey, 2015). This reduction of shame may in turn have an impact on disenfranchised grief (Doka, 1989), in so far as grief after suicide may be more easily accepted and socially sanctioned.

Another important finding in the present study is that when accounting for previous psychosocial health and time elapsed after suicide, online support groups seem to have a positive effect on the bereaved participants’ satisfaction with their current psychosocial health. The results also showed that more intensive use of memorial websites had an opposite negative effect. This result was obtained in the bivariate analysis, as well as in the multiple regression, but only when shame, anger, and guilt (i.e., Factor 2 which was non-significant) were omitted from analysis. This suggests that online support groups—due to their opportunities to meet others with similar experiences and discuss issues related to suicide, loss, and grief—provide some support and help to the bereaved, at least temporarily. Memorial websites may also be used in grief work; for communicating with other bereaved people online and sharing experiences (e.g., Roberts, 2004). But they also appear to be used for other reasons, for example, to memorialize, honor, and enhance the relationship with the deceased (Bailey et al., 2015). However, the present study indicates that this online activity does not seem to have any apparent positive outcomes regarding the bereaved person’s satisfaction with their present psychosocial well-being. Memorial websites might initiate a process of negative rumination and a reluctance to let go of the deceased (cf. Stroebe et al., 2008), which Mitchell, Stephenson, Cadell, and Macdonald (2012) describe as a tendency to “extend mourning indefinitely.” This can be seen as an example of the digital resources’ potential for continuing bonds to the deceased, and, in the present case, with a more adverse outcome.

Finally, in a concluding analysis, it was tested if users of digital resources experienced more satisfaction with their current health than nonusers, given that they have had the same recovery prognosis to begin with and the same time to recover (time elapsed after suicide). Although the interaction effect in the model was not significant, there was a tendency in respondents who use digital resources in their grief work to improve faster regarding satisfaction with their current psychosocial health than nonusers during the first period after the suicide event. This suggests a kind of “plateau”: that Internet-based resources might facilitate recovery up to a certain point—between 3 and 6 years after the suicide—after which using these resources might become less helpful or even maladaptive for the suicide bereaved, while satisfaction with current psychosocial health of bereaved nonusers continues to improve in a more linear fashion (Figure 4) (cf. Feigelman et al., 2008; Kramer et al., 2015). However, the cross-sectional design of this study precludes such a longitudinal assumption.

Limitations

A main limitation of the present study is the cross-sectional design which constitute a weakness regarding temporal changes, as no follow up of the anonymized bereaved participants was possible. Also, retrospective self-assessments include a risk for memory biased reporting by the participants. A further limitation is the self-selection bias: The survey was only made available on the Swedish suicide bereaved organization SPES’ website and Facebook groups, which filtered out other suicide bereaved individuals that do not use or visit these websites. This procedure makes it problematic to generalize the results to other Swedish suicide bereaved people. Finally, as been stated in the article, the gender distribution in the sample is very uneven, due to that women to a higher degree than men use online groups in their grief work (which also applies to SPES online grief groups). This means that it is primarily women's usage of different resources in their grief work that is measured in the study. Preferably, future studies may use a longitudinal design and more expanded recruitment and sampling procedures.

Conclusion

The main results of the present study indicate that the psychosocial ill-health is severe among Swedish suicide bereaved people and that they use a relatively large set of both offline and online resources in their grief work, and the perceived helpfulness of these were valued very differently. Most of these resources also seem to be used independently of each other. Regarding digital resources, more participants used online support groups than memorial websites, and support groups seem to be valued higher than the latter. Moreover, online support groups appear to have a positive impact on the users’ satisfaction with their current psychosocial health, while memorial websites seem to have the opposite effect. Thus, digital resources may be seen as a double-edged sword that can bring advantages as well as disadvantages into suicide bereaved peoples’ grief work, depending on which resources the bereaved choose to use and have access to.