The social health domain of people bereaved by a drug-related death and associations with professional help: A cross-sectional study

Abstract

People bereaved by traumatic deaths are vulnerable to long-lasting impairments in social health, including the quality of social relationships and the capacity to manage their social lives. In this Norwegian study involving 255 participants bereaved by a drug-related death, we aimed to investigate their social health and associations with professional help. The results of a cross-sectional survey showed that participants on average rated their social health as poor, though with large variations within the group. Participants who reported high satisfaction with professional help reported significantly higher scores on most social health-related variables. More research is needed on professional help focusing on the social health of traumatically bereaved people.

Introduction

Meaningful relationships with others are crucial when grieving; still, bereavement may lead to long-lasting difficulties in social interactions between people (Dyregrov & Dyregrov, 2008; Sajan et al., 2022). People who have been bereaved through traumatic deaths seem to be especially vulnerable to adversity in their social connection with others (Dutta et al., 2019, Dyregrov et al., 2003; McDonnell et al., 2022; Pitman et al., 2014; Sajan et al., 2022). Furthermore, studies exploring the situation of people bereaved by drug-related deaths (DRDs) show that this group also struggles with severe social challenges (Lambert et al., 2022; Titlestad & Dyregrov, 2022; Titlestad, Lindeman, et al., 2021). However, no quantitative study has, to our knowledge, investigated the social health of DRD-bereaved people until now.

According to Huber et al. (2011), social health is the third health domain besides physical and mental health. The empirical data for this study is situated at the micro-level of social health, which refers to the individual’s “quality of social relationships, and the capacity to manage social life” (Cho et al., 2020, p. 3). The study investigates three dimensions on this level: adjustment to work- and social activities, perceived and obtained social support, and connecting with or withdrawing from other people.

The first dimension is how bereaved people adjust to work-related and social activities. Bereaved populations often suffer impairment in this dimension, which seems to correlate highly with complicated grief reactions (Mauro et al., 2017; Shear et al., 2016; Tal et al., 2017).

A second dimension is perceived and obtained social support. The current literature finds perceived social support important for mental health outcomes (Wang et al., 2018). Furthermore, studies show that bereaved people experience social network support as essential (Dyregrov & Dyregrov, 2008) and that emotional caring and support are most helpful (Cacciatore et al., 2021). Low perceived social support is consistently reported as a major risk factor for mental distress and complicated grief reactions after bereavement (Lobb et al., 2010; Mason et al., 2020; Scott et al., 2020). Concerning bereavement by DRD, O’Callaghan et al. (2022) reported dialogue and social support as one of three main themes in DRD-bereaved people’s pathways to posttraumatic growth.

Third, connecting with or withdrawing from other people are essential social health dimensions. Bereaved people who have withdrawn from others are more likely to report high psychosocial distress, andprolonged grief symptoms (Dyregrov et al., 2003; Titlestad, Schmid, et al., 2021), and a recent study found significantly less psychological distress following increased social connection among bereaved people (Smith et al., 2020).

Health-related quality of life, “a multidimensional construct covering physical, emotional, mental, social and behavioral components of wellbeing and functioning” (Andersen et al., 2017, p. 3421), includes social health dimensions. Health-related quality of life often decreases significantly in the first months after bereavement, and some studies find long-term negative impacts (Liu et al., 2019; Song et al., 2010).

DRD-bereavement and social health

Some factors suggest that the social health of those bereaved by DRD may be particularly challenged compared to other traumatically bereaved people. One aspect is a severe strain on family members before the death, often including an ambivalent relationship with the drug-using family member, complex family dynamics, and withdrawal from social relationships outside the family (Lindeman et al., 2022). Another aspect is stigma toward drug using persons, which is more pronounced than stigma toward people with mental illness (Yang et al., 2017) and suicidal persons (Kheibari et al., 2022). This stigma can spill over to close relationships (Dyregrov & Selseng, 2022) and prevail after death. Two recent qualitative studies support these notions. First, an Irish study highlights how DRD-bereaved people experience challenges in family dynamics and relations with surrounding community members due to stigma (Lambert et al., 2022). Second, a Norwegian study found that DRD-bereaved parents struggled with shame, guilt, stigma, self-stigma, and challenging communication with their social network members (Titlestad, Mellingen, et al., 2021).

Still, there are many similarities between DRD-bereavement and other kinds of traumatic bereavement. Studies on those bereaved by suicides (Sajan et al., 2022; Shields et al., 2017) and parents bereaved through a child’s chronic illness (Dutta et al., 2019) report that many experience problems in familial communication and social relationships. Furthermore, qualitative studies on bereavement by suicides find that stigma, shame, guilt, and blame are frequently experienced by those bereaved (Sajan et al., 2022; Shields et al., 2017), adding to a socially strenuous bereavement. These experiences seem to parallel those of DRD-bereaved people, although the pre-loss strain and stigma may disfavor DRD-bereaved people even more.

Professional help, satisfaction with help, and social health

Professional help services can influence bereaved people’s social health indirectly or directly. The indirect path entails individual help to a bereaved person. The interactions in the helping relationship can then help the bereaved relate with people in their social network (Baddeley & Singer, 2009). The direct path entails including different social network members in the same meeting, for example, in a family or social network meeting (Seikkula, 2012) or bereavement support groups.

Regardless of the pathway, satisfaction with the service (Duggan & Thompson, 2011) and the alliance between helper and help-seeker are crucial in therapeutic relationships (Flückiger et al., 2018). Several studies have documented associations between satisfaction with services and better mental health and quality of life (Bamm et al., 2013; Oetzel et al., 2015; Petkari & Pietschnig, 2015). DRD-bereaved people in Norway who received professional help, reported significant variations in satisfaction with the received help (Kalsås et al., 2022). Services that provided psychotherapeutic help, or early and flexible help, were more often rated as satisfactory (Kalsås et al., 2022).

We have described the knowledge base showing that dimensions of social health often are negatively impacted after bereavement and that there are differences depending on the circumstances and manner of death. DRD-bereaved people’s social health is probably vulnerable due to relational strain, stigmatization, and internalization of stigma. Professional help can open paths for connection between bereaved people and their social network members in direct or indirect ways. The alliance and service satisfaction may be one key factor in accomplishing this. Hence, we wanted to map DRD-bereaved people’s social health and the relationship between satisfaction with help and social health.

Material and methods

The study is part of the END project, a Norwegian research project investigating DRD-bereaved people’s experiences, psychosocial health, experiences with help and support, and professional services’ way of relating to them. This study has a cross-sectional design, and the reporting is guided by the STROBE Checklist (Von Elm et al., 2007).

Data collection and participants

A survey was administered in March–December 2018 to a heterogeneous convenience sample of 255 DRD-bereaved Norwegian family members, partners, and close friends. The recruitment of 200–300 participants was conceived as feasible and adequate for cautious generalizations of findings to the target population. All participants were over 18 years, with at least three months separating them from the loss. The timeframe of three months was chosen in line with The Regional Committees for Research Ethics policies, as including more recently bereaved participants was considered ethically problematic. The recruitment strategy entailed information letters to all Norwegian municipalities and cooperation with health- and welfare services, hospital services, treatment centers, and non-governmental organizations. In addition, recruitment was promoted through advertising in various media, information at conferences, and snowball recruitment.

Sample characteristics

The time since death ranged from three months to 35 years, with minor statistical differences between the different relationships to the deceased. Almost all participants reported that the deceased had used drugs for several years before the death. Most participants worked, studied, or were on sick leave, but with considerable differences between relations, ranging from 48% (extended family members) to 85% (siblings). A majority of 82–96% of the deceased relatives reported feeling close or very close to the deceased at the time of death, while children stood out from the others, with only 64% reporting the same. More than one-third of the participants had experienced devaluating comments concerning the deceased after the death. At the same time, within the groups of children, close friends, and partners, about half of them reported this experience (Table 1).

Table 1. Sample characteristics (N = 255) presented with (min–max), mean (SD) or n (%).

Variable (min–max)	Mean (SD)	n (%)
Sociodemographic characteristics
Age of bereaved at the time of survey (18–80)	48 (14)
Age of bereaved at the time of loss (5–76)	40 (15)
Female sex		208 (82)
Educational status
College/university		125 (49)
Senior high school		97 (38)
Primary school		32 (13)
Employment
Working (full- or part-time)		155 (61)
Retired		29 (11)
Student		14 (5.5)
Other		58 (23)
Household income (USD)
≤50,000		85 (34)
50,000–100,000		121 (48)
≥100,000		45 (18)
Relational characteristics
Years since a family member’s or friend’s death (0–35)	8.1 (7.4)
Relation to deceased
Parent		95 (37)
Sibling		79 (31)
Child		25 (9.8)
Other kin		28 (11)
Close non-kin relation: partner (n = 13) or friend (n = 15)		28 (11)
Perceived closeness to deceased. Close/very close		222 (88)
Experienced devaluating comments of deceased post-loss		90 (35)
Characteristics of deceased
Deceased age in years at the time of death (15–68)	31 (9.9)
Deceased years of drug use before death (0–42)	13 (8.6)

Questionnaires and variables

Health-related quality of life

The RAND-12 health survey is the 12-item version of the RAND-36/SF-36, consisting of four nominal variables, two three-point Likert items, and six five-point Likert items regarding the situation for the last four weeks. Examples of questions are “have you felt downhearted and blue,” “how much of the time have your physical health or emotional problems interfered with your social activities (like visiting friends, relatives, etc.)?.” The instrument has been validated in several countries (Farivar et al., 2007) and proven cross-culturally reliable in Norway (Gandek et al., 1998). The oblique scoring method and the mental health component score (MCS) were used for this study (Farivar et al., 2007). Seven participants had one or two missing values, which were imputed based on the value of the adjacent variable/variables measuring the same Health-related quality of life dimension. Seven participants with missing values were not included in the analysis.

Perceived and obtained social support

The Crisis Support Scale (CSS) contains seven items measured by seven-point Likert scales (Joseph et al., 1992) and has shown good reliability and validity (Elklit et al., 2001; Joseph et al., 1992). The scale consists of five items tapping positive support, one item tapping negative social experiences, and one item asking about overall satisfaction with social support. The scale gives “a consistent and meaningful picture of both perceived and obtained social support” (Elklit et al., 2001, p. 1300). The first six items were used for the frequency analyses, and the negative experience item was reversed when calculating the sum score. The five first items measuring positive support were used for the subsequent correlation and group comparison analyses. The α of the six-item scale was .706, and for the five-item positive support subscale, it was .779. Five imputations were made on the positive support subscale on participants with one missing item, based on the mean of the participant’s other scores.

Work, social adjustment and social connectedness

The Work and Social Adjustment Scale (WSAS) consists of five 0–8 scored items and has shown good reliability and validity (Mundt et al., 2002). A higher sum score interprets as higher levels of impairment due to the bereavement. All five items were used for the descriptive analysis, allowing comparison with other populations. For the ANOVAs, the summary of items three and five that tap social connectedness were used. Three participants had one missing value imputed with the sample’s estimated mean. The α of the five-item scale on this sample was .907; for the two-item scale, it was .798 (Pearson’s r = .663).

Help and social withdrawal

The Assistance Questionnaire (AQ-R) has 22 items, where response alternatives are either nominal or five-point Likert items. The questionnaire is previously used in other studies with traumatized bereaved populations (Dyregrov et al., 2003; Wilson & Clark, 2005) and addresses bereaved people’s need for help and received help. We used five items in this study; “Needs for help” (five-point), “Received help” (nominal), “Satisfaction with help” (five-point, collapsed to three-point for analysis), “I have withdrawn from others” (five-point), and “Others have withdrawn from me” (five-point). Further details on the professional help and services involved with the bereaved in this sample can be found in Kalsås et al. (2022).

Ethical considerations

The Norwegian Regional Committee for Medical and Health Research Ethics has approved the END research project (ref. nr. 2017/2486/REK vest). All participants were informed in writing about the project’s aim before participating and were made aware of the possibility of contacting the project manager if answering the survey prompted a need to talk to someone. Furthermore, it was explained that the data would be published non-identifiable and stored on the research server at the university.

Statistical analyses

Social health was mapped through frequency analyses of RAND-12 MCS (Health-related quality of life, mental component score), WSAS (Work- and social adjustment), CSS (social support), “Own social withdrawal,” and “Others’ social withdrawal.” For correlations between dimensions of social health and time since death, a bivariate correlation analysis was conducted. For the single five-point Likert scaled items, Spearman’s rho was measured.

Concerning differences in received help, satisfaction with help, and social health, the first analysis was conducted with T-tests and Mann–Whitney U-tests, examining group differences in social health dimensions between those who had received help and those who had not received help. The aim was to determine differences in social health dimensions of people who reported needing professional help after the death (n = 230). The sample was stratified into those who had received professional help (n = 124) and those who had not received professional help (n = 106). The second analysis examined group differences in social health dimensions among bereaved who had reported different levels of satisfaction with help. The group that had received help was stratified into three groups based on satisfaction with help: low (n = 23), medium (n = 43), and high (n = 55). The analyses were conducted with one-way ANOVAs and the Kruskal-Willis H test. All analyses were done using IBM SPSS Statistics Version 27.

Results

Frequency analyses

Mean RAND-12 scores were generally better among extended family (43.1) and siblings (42.4) than among partners/friends (35.6), parents (39.3), and children (38). Relative parallel patterns were also observed for CSS and WSAS, while children seem to have withdrawn from others to a higher degree (3.4), especially compared to extended family members (2.3). A total of 67% of the sample reported having withdrawn from others to some-high degree, and 46% reported that others had withdrawn from them to the same degree. There were only minor differences between groups in the scorings of “Others’ withdrawal” (Table 2).

Table 2. Frequency analyses, whole sample and different relations to deceased.

Measure	Tot. sample (n = 249)	Parent (n = 92)	Sibling (n = 79)	Child (n = 24)	Ext. fam (n = 27)	Part/friend (n = 27)
	M (SD) range	M (SD)	M (SD)	M (SD)	M (SD)	M (SD)
RAND-12 MCS	40.2 (12.9) 12–65	39.3 (12.9)	42.4 (12.5)	38 (12.1)	43.1 (13)	35.6 (13.3)
CSS (six items)	28 (6.9) 7–42	28.9 (6.8)	28 (6.5)	25 (8.5)	29.6 (7.2)	25.9 (5.9)
WSAS	12.7 (10.9) 0–40	14.4 (10.3)	10.8 (10.7)	12.8 (11.1)	10 (10.2)	15.1 (12.6)
Own withdrawal	2.9 (1.2) 1–5	2.8 (1.1)	3 (1.2)	3.4 (0.8)	2.3 (1.1)	3 (1)
Others’ withdrawal	2.4 (1) 1–5	2.4 (1.1)	2.3 (1)	2.5 (0.7)	2.3 (1.1)	2.2 (1)

Correlation analysis

“Time since death” showed a small significant correlation only with RAND-12 MCS (r = .205). “I have withdrawn” correlated most with the WSAS subscale measuring social connectedness (r = .533) and showed a medium-high negative correlation with the RAND-12 mental component score (MCS) (r = −.447). Furthermore, the full WSAS scale and WSAS subscale correlated highly negatively with RAND-12 MCS (r = −.722 and r = −.657). “Others have withdrawn” showed small-medium correlations between r = .260 and r = .334 to all other dimensions except “Time since death.” CSS, scored as a five-item positive support subscale, correlated on a small-medium level with the other scales, highest at r = .323 with RAND-12 MCS. WSAS sub = WSAS sum score of items three and five (Table 3).

Table 3. Correlation matrix, social health dimensions and time since death.

	I withdraw	Others withdraw	RAND-12 MCS	CSS pos	WSAS full	WSAS sub	Time^b
I withdraw^a	1
Others withdr.^a	.334**	1
RAND-12 MCS	–.447**	–.239**	1
CSS pos	–.267**	–.244**	.323**	1
WSAS full	.501**	.298**	–.722**	–.219**	1
WSAS sub	.533**	.260**	–.657**	–.219**	.918**	1
Time^b	.014	.042	.205**	.046	–.098	–.047	1

*p < .05; **p < .001 (two-tailed). ^aCoefficients on “I have withdrawn” and “Others have withdrawn” calculated as Spearman′s rho. ^bTime = Time since death.

Analyses comparing groups

We planned to use “Time since death” as a covariate to control for possible confounding in all following analyses. However, correlation and linear regression analyses showed no relationship between “Time since death” and dependent variables. Therefore, we decided not to include the variable.

Social health dimensions: help group versus no help group

When comparing positive social support (CSS 5-items), no significant differences in scores between the group that had received professional help (M = 24.8, SE = .569) versus the group that had not received professional help (M = 23.5, SE = .648) emerged, homogeneity of variances assumed (Levene’s test p = .517), t(227) = 1.53, p = .127. Impairment in social connectedness (WSAS subscale) also showed no significant differences in scores between the help group (M = 5.98, SE = .436) and the no help group (M = 4.89, SE = .442), homogeneity of variances assumed (Levene’s test p = .320), t(226) = 1.75, p = .082. Considering “I have withdrawn,” the distribution of scores for the help group (M = 2.89) and no help group (M = 2.84) were similar assessed by visual inspection, and there were no significant differences between the two groups’ scores, U = 6621.5, z = .344, p = .731. Finally, the test for differences concerning “Others’ withdrawal” showed that the distribution of scores for the help group (M = 2.44) and the no help group (M = 2.28) were similar, assessed by visual inspection, and there were no significant differences between the two groups’ scores, U = 7055, z = .477, p = .262.

Social health dimensions at low, medium, and high satisfaction with help

All analyses showed a linear relationship where social health dimensions increased with higher-rated satisfaction with help. However, not all relationships were statistically significant at p < .05. When comparing positive social support (CSS 5-items) with a one-way ANOVA, the group means were significantly different, F(2,117) = 7.9, p < .001, ω² = .103), homogeneity of variances assumed (Levene’s test mean: p = .947). Tukey HSD Post hoc analysis showed that the difference between the high satisfaction group (M = 27.2, SD = 6.1) and the medium satisfaction group (M = 23.5, SD = 5.9) was statistically significant (3.70, 95% CI [0.81–6.59], p = .008), as well as the difference between the high satisfaction group and low satisfaction group (M = 22, SD = 5.7) (5.19, 95% CI [1.66–8.71], p = .002).

Impairment in social connectedness (WSAS subscale) showed significant differences in group means, F(2,117) = 3.26, p = .042, ω² = .036), homogeneity of variances assumed (Levene’s test mean: p = .694). Tukey HSD Post hoc analysis showed that the difference between the high satisfaction group (M = 5.25, SD = 4.7) and the low satisfaction group (M = 8.22, SD = 5.2) was significantly different (−2.96, 95% CI [−5.76 to −0.17], p = .035).

For the item “I have withdrawn,” the Kruskal–Wallis H test showed a statistically significant difference between the high satisfaction group (mean rank 52.15) and the low satisfaction group (mean rank 73.76), H(2) = 7.558, p = .023. Finally, the Kruskal-Wallis H test for differences concerning others’ withdrawal showed no significant differences between groups, although close (mean rank high satisfaction group 54.46, low satisfaction group 72.52), H(2) = 4.873, p = .087. The distributions of scores between the different satisfaction groups on both Kruskal-Wallis tests could not be confirmed as similar for all groups based on visual inspections of a boxplot. Pairwise comparisons (Dunn, 1964) and a Bonferroni correction for multiple comparisons were made.

Discussion

The frequency analyses of the different social health dimensions showed a low mean score on the mental health-related quality of life component (RAND-12 MCS). For work- and social adjustment (WSAS), the average score suggests significant impairment (Mundt et al., 2002), and the mean score for social support (CSS) was relatively low compared to other bereaved populations (Arnberg et al., 2012). Two-thirds of respondents reported having withdrawn from other people to some-high degree, and almost half of the sample reported that other people had withdrawn from them in correspondingly degree. All scores were quite evenly distributed across different relations, although extended family scored somewhat better on all measured variables. No social health dimensions correlated positively or negatively with “Time since death,” indicating that those bereaved a long time ago did not have better social health than those newly bereaved. This result suggests that social health variables do not improve substantially with time, or that the participants who lost a long time ago, reported low scores for reasons not investigated in this study. The analyses examining the social health dimensions between the groups who rated help satisfaction differently showed a positive linear relationship: higher satisfaction was associated with better scores on all social health dimensions, except “Others have withdrawn from me.” These differences in social health dimensions for the groups who rated satisfaction differently might be related to the help-providing.

DRD-bereaved people’s results in social health dimensions

The WSAS average score of 12.7 (SD 10.9) indicates that many participants scored at the same level as those seeking treatment for complicated grief and that a large group also had low work- and social impairment. A study including a random sample of bereaved people showed average WSAS scores of 0.8 (SD 2.4) (Mauro et al., 2017), while studies on bereaved people seeking treatment for complicated grief, have shown average WSAS scores from 19.7 to 26.3 (SD 8.3–10.1) (Mauro et al., 2017; Shear et al., 2016; Tal et al., 2017). Our sample of DRD-bereaved people thus scored averagely better than bereaved populations seeking treatment for complicated grief and considerably worse than a random sample of bereaved people (see Mauro et al., 2017; Shear et al., 2016; Tal et al., 2017).

The mean scores on health-related quality of life measured with RAND-12 MCS and social support measured with CSS were poor. Scores on RAND-12 MCS were significantly below the Norwegian norm, 40.2 vs. 51.8 (Statistics Norway, 2012), and a Danish study found higher mean scores in different groups of parentally bereaved youth measured with the SF-36 (M = 45.4–49.6) (Appel et al., 2019). The SF-36 MCS is highly correlated with the RAND-12 MCS (Gandek et al., 1998; Lee et al., 2008). The mean score on CSS (six items) was 28 (SD 6.9). In comparison, a large sample of bereaved Swedish people reported a mean of 30.7 (SD 7.6) 14 months after the 2004 tsunami (Arnberg et al., 2012), which is significantly better. The level at which participants reported “I have withdrawn” was considerably higher in our sample compared to people bereaved by suicides, accidents, and sudden infant deaths in another Norwegian study, respectively 67%, 45%, 50%, and 57% (Dyregrov et al., 2003). Though less stated, almost half of participants also reported that others had withdrawn from them to some-large degree, supporting findings from other studies of DRD-bereavement (Feigelman et al., 2020). Traumatically bereaved people are vulnerable to impaired social health (Dutta et al., 2019; Sajan et al., 2022; Shields et al., 2017), and we have shown that DRD-bereaved people score even poorer on many social health dimensions than other bereaved populations. Possible explanations for these results will be discussed.

Strain, stigma, and shame before and after the death

Several studies show that problematic substance use may severely impact close family members, affecting the family structure and increasing the family members’ risk for different mental and physical illnesses (Di Sarno et al., 2021; Lindeman et al., 2022; Orford et al., 2010). In addition, the experience of problematic substance use is considered a “family matter” for many families, contributing to feelings of shame and guilt for being closely related to the drug-using person (Lindeman et al., 2022). As a result, many distances themselves from social relationships outside the family, and feelings of isolation and loneliness are recurring themes (Lindeman et al., 2022).

This distancing is probably partly due to processes of stigmatization that drug-using persons are subjected to (Titlestad, Mellingen, et al., 2021) and have a spillover effect on family members (Dyregrov & Selseng, 2022). Experiences of stigmatization are closely connected with feelings of shame (Luoma et al., 2013), which signal threats to social bonds (Scheff, 2006). Approach behaviors toward other people can mend this threat to social bonds and alleviate the feelings of shame (De Hooge et al., 2010). However, the difficulty of shame repair through approach behaviors may be exacerbated by stigma (Cibich et al., 2016) and cultural expectations concerning what one should keep within the family (Lindeman et al., 2022). Thus, withdrawing from others is one way people can cope with stigma and feelings of shame, and try to protect the self from further social harm (De Hooge et al., 2010).

These dynamics show how stigma at a macro level (e.g., norms, public discourse, and jurisdiction) can connect with individual and familiar strain at a micro-level (e.g., shame, withdrawal, avoidance from other people). These dynamics seem to start before the loss and may also prevail after the death (Dyregrov & Selseng, 2022), likely impacting the social health of DRD-bereaved people negatively.

The circular causality of social health

The finding that many DRD-bereaved people also have experienced that others have withdrawn from them may illustrate how social health, including social support, is an interactional phenomenon (Lakey & Orehek, 2011). Both the interpersonal actions of the individual and other people’s way of relating to the bereaved person are essential. Social network members may avoid contacting bereaved people due to their own insecurity, or because they interpret the withdrawal of the one who is bereaved as a wish to be left in peace (Dyregrov & Dyregrov, 2008). The social network members’ withdrawal may be interpreted as motivated by prejudice by the bereaved person, possibly leading to a circle of misunderstanding, avoidance, and withdrawal from both parties. Bereaved people have stated that “openness” is vital for alleviating the problem; to tell the social network members their story, inform them, and clarify their needs (Dyregrov & Dyregrov, 2008). Social network members recommended the same for themselves; that is, openness from both parties (Dyregrov, 2006). Depending on the bereaved person’s resources and the state and quality of the existing social relationships, some bereaved people will manage to take this responsibility themselves, thus alleviating the interaction with others. The findings from this study may suggest that many DRD-bereaved people with their social network members could need professional help to manage these problems.

Professional help and social health

There were significant differences in the social health dimensions between the group that rated the professional help as highly satisfactory versus those with low satisfaction. This finding supports a hypothesis that the level of satisfaction with professional help might positively affect most social health dimensions, although a causal or directional relationship cannot be inferred from the cross-sectional data. We have stated a hypothesis of strain, stigma, shame, and withdrawal as drivers of the reduction in social health. If a directional relationship exists, a possible explanation might be that professional help has mitigated the impairing potential of these dynamics and facilitated openness between the bereaved and their social network members. If so, this kind of help is not only immediately helpful but also aids bereaved persons in turning to other people for connection and support later.

Furthermore, given that a directional relationship is present, the finding suggests that monitoring the helping alliance and satisfaction should be adopted as an integrated part of the help provisions. This is common in psychotherapy and other therapeutic settings (Kidd et al., 2017). However, the help to DRD-bereaved people does not necessarily include a structured intervention based on a diagnosis. Good outcomes in public mental health settings may differ from a highly structured therapeutic arena (Moltu et al., 2017). For bereaved people, we argue that social health outcomes should be included.

Strengths and limitations

This study has several strengths. To our knowledge, the sample of DRD-bereaved family members and friends is the largest in a cross-sectional study in one country until now. The sampling process ensured a large variety of participants concerning relationship, age, and geography. Furthermore, the study has user involvement in the study design, data collection, and the interpretation of findings. This ensures the study’s relevance for the primary stakeholder group, DRD-bereaved people. There are also limitations. Causal associations cannot be inferred from the cross-sectional design. Second, quantitative measuring of social relationships through a cross-sectional survey gives limited information on relationships and social interaction. Third, social withdrawal was reported using single Likert-scale items, and impairment in social connectedness using two items from WSAS. More complex measures would provide more nuances, for example, the newly developed “Oxford Grief Social Disconnection Scale” (Smith et al., 2020). Fourth, the sample may be biased because of self-selection and relations between participants and is skewed concerning gender, geographical representation, and above-average education level. Thus, generalization of the results to the target population must be made with some caution. Finally, the large variations in the time since death increase the possibility of recall bias and confounding variables, which makes it more difficult to assume causality.

Implications for practice and future research

Professional services should emphasize ways of working with the bereaved to improve social health dimensions. We argue that interventions that use the direct pathway, including more than one individual at a time, are likely the most effective for improving social health. Such interventions may include social network meetings and family meetings with or without psychoeducational elements, for example, social network meetings based on the “Open Dialogue” approach (Olson et al., 2014), “Systematic Early Intervention for Bereaved” (Pereira et al., 2016), or bereavement and family support groups (see O’Callaghan et al., 2022).

Since few studies have investigated early helping interventions to improve social health for bereaved people (Andriessen et al., 2019; Wittouck et al., 2011), we argue for both quantitative and qualitative studies on this topic. A longitudinal experimental study can yield important knowledge concerning possible benefits of such therapeutic approaches. A naturalistic study design is probably most feasible, where social health outcomes of bereaved populations in different geographic areas could be compared. An action research design could be beneficial in developing therapeutic approaches adapted to local sociocultural contexts. These study designs could be used independently, or they could be combined.

Conclusion

DRD-bereaved people reported poorer social health than comparable bereaved populations, and social health dimensions did not correlate with time since death. Reduced social health might be mitigated through professional help, and DRD-bereaved who reported high satisfaction with help also reported better social health. Helping measures that include social network members can potentially mitigate mutual social withdrawal and increase social connectedness. There is a need for more research on these types of helping interventions.

Acknowledgments

The authors acknowledge the END project and its user representatives for valuable inputs and interpretations concerning the findings. A special thanks to Professor Sigurd Hystad for his help with statistical analyses.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The author(s) reported that this study was funded by the Western Norway University of Applied Sciences and The Research Council of Norway.