Shame, depression, and complicated grief among suicide loss-survivors: the moderating role of self-disclosure

ABSTRACT

Background: Suicide-loss survivors (SLSs) are recognised as an at-risk population for several psychiatric complications, including complicated grief (CG) and depression (SI). However, whereas shame is known as one of the characteristics of this population, knowledge about possible psychological processes which may moderate the contribution of shame levels to CG and depression in the aftermath of suicide loss is sparse. This study examines the role of self-disclosure – the inclination to share personal information with others – as a possible moderator of the associations of shame with CG and shame with depression over time.

Method: Participants were 152 suicide-loss survivors, aged 18–70, who completed questionnaires tapping CG and depression at three time points (T1- index measurement, T2-two years after T1. and T3-four years after T1) and questionnaires tapping shame and SD at T3.

Results: Hierarchical regression analyses showed that shame significantly and positively contributed to CG at T3 and to depression at T3, beyond the CG/depression trajectories. Notably, two significant interactions were found: Self-disclosure moderated the contribution of shame to CG at T3 and to depression at T3. At lower self-disclosure levels, shame’s contribution to CG and depression was higher.

Conclusion: The study’s findings highlight shame as a significant facilitator of CG and depression in the aftermath of suicide loss. Moreover, the role of interpersonal interaction on SLSs’ distress levels and grieving process was underscored, as this interaction may serve as a buffer against the deleterious sequelae of the suicide of a loved one.

HIGHLIGHTS

• Shame significantly and positively contributed to CG and depression at trajectories.

• Self-disclosure moderated the contribution of shame to CG and depression at T3 beyond their natural trajectories.

• Interpersonal activities may serve as a buffer against the deleterious effects of suicide in the family.

Antecedentes: Se reconoce que los sobrevivientes a pérdida por suicidio (DPS) son una población de riesgo para diferentes complicaciones psiquiátricas, incluyendo al duelo complicado (DC) y a la depresión. Sin embargo, pese a que se conoce que la vergüenza caracteriza a esta población, el conocimiento de los posibles procesos psicológicos mediante los cuales los niveles de vergüenza ejercen un papel moderador entre estos, el DC y la depresión luego de un suicidio es escaso. Este estudio evalúa el papel de la divulgación voluntaria – la tendencia a compartir información personal con otros – como un posible moderador de las asociaciones entre la vergüenza y el DC, y entre la vergüenza y la depresión en el tiempo.

Métodos: Los participantes fueron 152 sobrevivientes a pérdida por suicidio entre 18 y 70 años quienes completaron cuestionarios que vinculaban el DC y la depresión en tres momentos (T1 – dato de base; T2 – dos años luego de T1; T3 – cuatro años luego de T1) y también cuestionarios que vinculaban la vergüenza con la depresión en T3.

Resultados: Los análisis de regresión jerárquica mostraron que la vergüenza contribuía de forma significativa y directamente proporcional al DC en T3 y a la depresión en T3, más allá de las trayectorias de duelo complicado / depresión. En particular, se encontraron dos interacciones significativas: La divulgación voluntaria tuvo un efecto moderador entre la vergüenza y el DC en T3, y entre la vergüenza y la depresión en T3. A menores niveles de divulgación voluntaria, la contribución de la vergüenza sobre el DC y la depresión era mayor.

Conclusión: Los hallazgos del estudio resaltan la vergüenza como un facilitador significativo para el DC y la depresión luego de un duelo por suicidio. Asimismo, resaltaron el papel de la interacción interpersonal sobre los niveles de angustia y el proceso de duelo de los sobrevivientes a pérdida por suicidio, pudiendo esta interacción servir para amortiguar las secuelas deletéreas del suicidio de un ser amado.

背景：自杀身亡幸存者 (SLS) 被认为是多种精神并发症的高危人群，包括复杂性哀伤 (CG) 和抑郁。 然而，虽然羞耻被认为是这一人群的特征之一，关于可能在自杀失败后可能调节羞耻水平对 CG 和抑郁作用的心理过程知之甚少。 本研究考查了自我披露—与他人分享个人信息的倾向，作为随时间推移羞耻与 CG 和羞耻与抑郁关联可能的调节者作用。

方法：参与者是 152 名18–70 岁的自杀失败幸存者，在三个时间点（T1-指数测量，T2-T1 后两年和 T3-T1 后四年）完成了探究 CG 和抑郁的问卷调查，在T3 时进行了探究耻辱和 SD的问卷调查。

结果：分层回归分析表明，在 CG/抑郁轨迹之外羞耻对 T3 的 CG 和 T3 的抑郁有显著和正向的作用。 值得注意的是，发现了两个重要的交互作用：自我披露调节了羞耻对 T3 的 CG 和 T3的抑郁的作用。 在较低的自我表露水平下，羞耻对 CG 和抑郁的作用更高。

结论：本研究结果强调了羞耻是自杀失败后 CG 和抑郁的重要促进因素。 此外，强调了人际互动对 SLS 的痛苦程度和哀伤处理的作用，因为这种互动可以缓冲亲人自杀的有害后遗症。

KEYWORDS:

• Complicated grief
• depression
• suicide-loss survivors
• shame
• self-disclosure

PALABRAS CLAVE:

• Duelo complicado
• depresión
• sobrevivientes a pérdida por suicidio
• vergüenza
• divulgación voluntaria

关键词:

• 复杂性哀伤
• 抑郁
• 自杀失败幸存者
• 羞耻
• 自我表露

1. Introduction

Suicide is one of the most disturbing public health problems. Annually, approximately 800,000 people worldwide die by suicide (WHO, 2019). Recent studies (Cerel et al., 2019; Feigelman et al., 2018a; Levi-Belz & Feigleman, 2021) noted that suicide impacts approximately 60 suicide-loss survivors (SLSs). These include nuclear family members, relatives, friends, and acquaintances who are profoundly and negatively affected by the suicide of their loved ones. Estimates are that more than 60 million people are added to the SLS population yearly. These numbers underscore the importance of understanding how we can help SLSs in their psychological journey of dealing with their suicide loss.

A large body of data has found SLSs at greater risk than other bereaved individuals and the general population for various severe psychological and health problems. Research has revealed that SLSs present higher levels of depression (Hamdan et al., 2019; Levi-Belz & Gilo, 2020; Maple et al., 2017) than other bereaved individuals. Moreover, SLSs’ grief experience can assume the pathological form of complicated grief (CG; Hamdan et al., 2019; Levi-Belz & Aisenberg, 2021), defined as prolonged and intense grief that leads to impairment in daily life functioning and significant health problems (Shear, 2015). CG is characterised by intense yearning, longing, or emotional pain, frequent preoccupying thoughts and memories of the deceased person, a feeling of disbelief or an inability to accept the loss, and difficulty imagining a meaningful future without the deceased person (Shear, 2015). Importantly, CG partially overlaps with the new ICD-11 and DSM-5-TR prolonged grief disorder (PGD) diagnosis, and studies have highlighted that they refer to the same clinical entity, reflecting a failure to successfully adapt after a loss (e.g. Golden & Dalgleish, 2010). However, as CG is a broader construct and encompasses individuals who would not be diagnosed as suffering from PGD (Maciejewski et al., 2016), this study focused on CG, which more accurately describes a broader symptom profile.

Alongside the negative psychiatric outcomes, SLSs encounter feelings of guilt, recognised as one of the most painful experiences in the aftermath of suicide loss (Cerel et al., 2008; Jordan, 2020). However, several studies have highlighted that the experience of shame – referring to an ‘intensely painful feeling or experience of believing we are flawed and therefore unworthy of acceptance and belonging’ (Brown, 2006, p. 45) – is also closely related to higher distress among SLSs (Jordan, 2020; Pitman et al., 2016). Shame arises when negative evaluation is applied to the whole self (e.g. ‘I am a bad person’; Tangney et al., 2007). Interestingly, several studies emphasized that those experiencing shame report more intense negative affect and greater feelings of inferiority and isolation following the traumatic event (Tangney & Dearing, 2002). Thus, Tangney et al. (2007) viewed shame as more emotionally painful than guilt because it threatens a person’s identity and is more difficult to resolve. An individual experiencing guilt can apologise or make amends, but someone experiencing shame must revise their entire sense of self (Tangney et al., 2007).

Among SLSs, intense feelings of shame often become an integral part of the suicide bereavement process. They can increase perceptions of negative self-stigmatization (Clark, 2012), which hinder help-seeking and lead to social isolation (Clark & Goldney, 2000). Their social isolation is due to the inclination of those who experience shame to be less likely to seek or receive social support from others (Groot et al., 2006; LeBlanc et al., 2020). Thus, it is not surprising that shame also demonstrates a consistent and robust relationship with psychopathology. Across a range of studies, researchers have shown that individuals with high shame report psychological distress that includes heightened levels of anxiety, depression, and grief reactions (Cândea & Szentagotai-Tătar, 2018; Kim et al., 2011). Taken together, SLSs characterised by higher shame levels in the aftermath of suicide loss are likely to present deleterious effects such as depression and CG. Surprisingly, there are only scant data to confirm this logical inference.

As shame is related to self-stigmatization and social isolation, one factor that may moderate its contribution to depression and CG may be the individual’s ability to self-disclosure. Self-disclosure refers to the process by which persons let themselves be known to others (Kahn et al., 2012). Several studies have demonstrated self-disclosure to be a prerequisite for healthy adjustment among SLSs (Levi-Belz, 2019; Levi-Belz & Lev-Ari, 2019; Smith et al., 2011). For example, SLSs who could talk about the death event showed significantly fewer grief difficulties and less mental health disturbance (Feigelman et al., 2018b). In another interesting study, Oexle et al. (2020) found that SLSs who felt the need to keep their suicide loss secret were more likely to exhibit increased grief difficulties and suicide ideation than those more open to sharing their loss with others.

Higher levels of self-disclosure were found to buffer against depression (Kahn & Garrison, 2009), partly through the mediation of loneliness (Wei et al., 2005). In some studies (e.g. Levi-Belz & Ben-Yaish, 2022; Shu-Liang et al., 2020), self-disclosure was found to be related to higher levels of perceived social support and belongingness, which have been recognised as factors that may contribute to SLSs’ recovery from their suicide loss (Levi-Belz & Aisenberg, 2021). Thus, self-disclosing emotional aspects of traumatic events to others may facilitate changes in loneliness and social support, as well as craft new ways of thinking about oneself and the suicide event (Levi-Belz, 2015), thus easing the harsh feelings of shame and its ramifications (Levi-Belz & Gilo, 2020). Thus, it can be suggested that self-disclosure may comprise a protective factor against depression and CG, especially when SLSs experience high shame in the aftermath of suicide loss. Importantly, however, self-disclosure’s role as a moderating factor in the relationship between shame and depression as well as between shame and CG among SLSs has yet to be examined.

1.1. The present study

This study aimed to examine the contribution of shame levels to CG and depression among SLSs and the moderating role of self-disclosure in these associations. To understand the contribution of shame and the moderative effect of self-disclosure on CG and depression beyond the natural trajectories of CG and depression over time, we used data from our longitudinal study (for details, see Levi-Belz, 2021). In the present study, we collected data at three measurement points: T1 (index measurement, baseline), T2 (24 months after baseline), and T3 (48 months after baseline) to assess the trajectories of CG and depression levels among SLSs. To the best of our knowledge, the current study is the first to examine self-disclosure as a buffer of the contribution of shame to CG and depression among SLSs over time. The study’s findings may facilitate designing and refining specific therapeutic interventions for SLSs that may diminish CG and depression levels for this population.

Two hypotheses lay at the core of this study:

(H1) Among SLSs, higher levels of shame will be associated with higher CG-T3 and depression-T3 beyond the CG/depression trajectory over time.

(H2) Among SLSs, self-disclosure levels will moderate the relationship between shame and CG and between shame and depression beyond the natural trajectories of CG and depression over time: As the self-disclosure levels increase, the relationships between shame and CG/depression will weaken.

2. Method

2.1. Participants

This study is part of the SLS longitudinal project, comprising four measurement points to date. In the current study, we used only three measurement points as we did not assess depression at one of the measurement points.

Participants comprised 189 suicide-loss survivors who participated in the study’s initial measurement point, conducted during 2015–2016. Of the initial participants, 152 (80.4%) SLSs participated in the last measurement points that provided the data for the current study (see Figure 1). Of the 37 dropouts from the first measurement, 23 could not be located, 13 did not respond to the invitation letter for the subsequent measurement points, and one died (due to cancer disease). These participants were excluded from the sample, leaving a total study sample of 152.

Figure 1. Flow chart of SLSs who participated in this study.

Participant recruitment for the longitudinal project transpired in two primary settings. Most participants were recruited through a nonprofit organisation, ‘The Path to Life,’ the national agency for suicide survivors in Israel. Other participants were recruited through the Israeli suicide survivors’ Facebook group and other social media groups of suicide survivors in Israel. No significant demographic or psychological differences were found between those who completed T3 and those who did not (dropouts).

The study’s sample of 152 suicide survivors (130 females) was aged 23–75. Inclusion criterion for SLSs was losing a family member or a close friend due to suicide (Feigelman et al., 2018a). Exclusion criteria were being under the age of 15 at the time of the suicide and lacking proficiency in reading and writing Hebrew.

2.2. Measures

2.2.1. Complicated Grief (CG)

CG levels were assessed using the Inventory of Complicated Grief-Revised (ICG-R), including only items reflecting the CG consensus criteria that obtained the correct classification of 93% and excellent interrater agreement (k = 1.0; Latham & Prigerson, 2004). These 17 ICG-R items tapped symptoms of separation distress and traumatic distress (e.g. ‘Do you ever feel yourself longing and yearning for ______?’; ‘To what extent do you feel like the future holds no meaning or purpose without _____?’). Anchors were presented on a 5-point Likert-type scale, ranging from 1 (never) to 5 (always). Sum scores were obtained, with higher scores indicating greater CG severity. CG levels were measured at all three measurement points. High internal consistency measures were obtained for the current sample at T1 (Cronbach’s α = .93), T2 (α = .92), and T3 (α = .89).

2.2.2. Depression

Depression was evaluated by the Patient Health Questionnaire-9 (PHQ-9; Kroenke et al., 2001). The PHQ-9 was used to assess for symptoms included in the diagnostic criteria for major depressive disorder (e.g. anhedonia, depressed mood, disturbances in sleep and appetite, self-disparagement, psychomotor agitation). Participants were asked, ‘Over the last two weeks, how often have you been bothered by any of the following problems?’ The frequency of each of the symptoms was rated on a 4-point Likert-type scale: 0 (not at all), 1 (several days), 2 (more than half the days), and 3 (nearly every day). Total scores ranging from 0 to 27 were calculated, with higher scores indicating higher depression levels. In this study, we measured depression at all three time points, yielding high Cronbach’s alpha coefficients: T1: α = .90; T2: α = .89; T3: α = .92.

2.2.3. Shame

Shame was measured by the trauma-related shame scale (Øktedalen et al., 2015), a short form of the Trauma-Related Shame Inventory (TRSI; Øktedalen et al., 2014), both developed by the same authors. The trauma-related shame scale includes five items tapped negative evaluations of the self (e.g. ‘As a result of my traumatic experience, I have lost respect for myself’) and representing the severity of shame related to the experienced trauma (Øktedalen et al., 2015). As Øktedalen et al. (2015) reported, four of the items drawn from the original 24-items version of the TRSI (Øktedalen et al., 2014) and one item drawn from the Posttraumatic Cognitions Inventory (PTCI, Ehlers et al., 1999). The items were selected by Øktedalen et al. (2015) based on content validity of shame and correlated highest with item-total score. Items of this measure were rated on an 11-point Likert-type scale ranging from 0 (does not match at all) to 10 (matches completely), yielding a sum score of 0–55. Øktedalen et al. (2015) reported the scale’s high reliability and content validity. Cronbach’s alpha for the present sample was 85.

2.2.4. Self-disclosure

The Distress Disclosure Index (DDI; Kahn et al., 2012) was used to measure the individual’s inclination to disclose personally distressing information (e.g. ‘I usually don’t share issues that bother me’ [reverse scored]; ‘I try to find people to talk to about my problems’). The 12-item DDI is presented on a 5-point Likert-type scale, ranging from 1 (strongly disagree) to 5 (strongly agree). The DDI’s confirmatory factor analysis suggested a single construct, yielding high reliability and validity coefficients (Kahn et al., 2012). Higher scores are reflective of higher disclosure levels. Cronbach’s alpha for the current sample was .94.

2.2.5. Demographic and personal characteristics

Concerning the suicide loss were collected for all participants, including the participant’s relatedness to the person who died by suicide (e.g. parent, spouse, child), the time since the suicide, and the survivor’s and the deceased’s ages at the time of suicide.

2.3. Procedure

The study was approved by the ethics committee at the Ruppin Academic Center. All potential participants were informed of the risks and compensation procedures. They were assured of anonymity, confidentiality, and their right to withdraw from the study at any time. Participants were required to affirm their willingness to participate by signing an informed consent form and then completing the questionnaire online (using Qualtrics online survey software). After completing the questionnaire, the participants were asked if they consented to be approached in the future for follow-up. Participants who agreed to be approached were invited again at the two subsequent measurement points (T2 and T3) and were requested to complete a short questionnaire incorporating the main study variables. After completing the online questionnaire, participants were compensated with gift vouchers for their participation at each measurement point (approximate value at each measurement: US $25).

2.4. Data analysis

Pearson correlation tests were calculated to examine the relationships among the study variables, followed by a hierarchical multiple regression, with CG or depression levels as the dependent variable. As Aiken et al. (1991) recommended, all continuous predictor variables were standardised, as were the cross-product interaction terms. To examine the nature of the interaction within a regression framework, moderation analysis was performed using the PROCESS macro [Model 1; Hayes, 2017]. The Statistical Package for the Social Sciences (SPSS, v26.0 for Windows) was used for all analyses. The level of statistical significance was set at p = .05.

3. Results

3.1. Demographics of the sample

At T3, the sample’s mean age was 45.9 (SD = 14.7). Of these participants, 25% were aged 55–73, and 18% were aged 21–28. Regarding participants’ family status, 72 (47%) were married, 61 (40%) were single, nine were divorced (6%), and 10 were widowed (7%). All participants identified as Jewish. Among these, 115 (75%) identified as secular, and 28 (18%) as religiously observant. Regarding socioeconomic status (SES), 36 (23%) participants reported a very low SES, 39 (25.6%) reported low SES, and 36 (23.6%) reported medium-to-high SES. Regarding schooling, almost all participants (n = 151, 99.3%) reported having completed at least 12 years, and almost 70% (n = 105) reported having college degrees.

3.2. Suicide-related demographic information

The participants reported various levels of relationship to the deceased: 29 (18.6%) were parents to the deceased, 26 (16.7%) were the deceased’s children, 43 (27.6%) were siblings, 16 (10.3%) were spouses, 13 (8.4%) were other family relatives, and 29 (18.6%) were best friends. All participants reported being devastated by the suicide: extremely devastated (48, 30.8%), highly devastated (85, 55.8%), and devastated (21, 13.4%). Time since the suicide varied among the participants (M_months= 132), with a range of 56–252 months: 27 participants (17.7%) had lost their significant other within 72 months, 38 (25%) within 72–102 months, 44 (29%) within 102–120 months, and the remainder (43, 28.2%) within ten years or more. At the time of the suicide, the participants’ mean age was 31.1 (SD = 15.3), ranging from 16 to 62.

3.3. Relationships between the study variables

Correlations among the study variables were calculated to test the hypothesis that shame and self-disclosure account for a high proportion of CG-T3 beyond the trajectory of CG and a high proportion of depression-T3 beyond the trajectory of depression. The intercorrelations for all variables are presented in Table 1. The matrix shows that shame levels were significantly and positively correlated with all CG and depression measures. Moreover, self-disclosure levels were significantly and negatively correlated with all CG and depression measures, except depression at T1. Time since suicide related negatively to CG-T1 and CG-T2 and was unrelated to depression levels at all measurement points.

Table 1. Means, standard deviations, and Pearson correlation coefficients between the study variables (N = 152).

	Measures	1	2	3	4	5	6	7	8	9
1	Time since suicide (months)	1
2	Shame (T3)	−.01	1
3	Self-disclosure (T3)	.09	−.09	1
4	Complicated grief (T1)	−.23**	.17*	−.18*	1
5	Complicated grief (T2)	−.17*	.33***	−.24**	.41***	1
6	Complicated grief (T3)	−.07	.51***	−.29***	.45***	.60***	1
7	Depression (T1)	−.11	.32***	−.01	.56***	.23***	.40***	1
8	Depression (T2)	−.05	.36***	−.20*	.14	.57***	.42***	.32***	1
9	Depression (T3)	−.00	.52***	−.18*	.32***	.41***	.62***	.42***	.46***	1
	M	132.72	7.63	43.88	34.64	37.24	36.26	6.97	7.61	6.41
	SD	22.06	5.42	8.32	7.51	8.80	7.29	4.04	4.23	3.61
	Range (min−max)	56–256	0–40	13–81	20–67	20–77	20–69	0–25	0–27	0–25

Note. *p < .05, **p < .01, ***p < .001.

3.4. Hierarchical analyses

To determine whether self-disclosure moderated the link between shame and CG and shame and depression beyond the CG/depression trajectories, two hierarchical regression analyses were applied, one with CG-T3 as a dependent variable and one with depression-T3 as a dependent variable.

3.4.1. CG-T3 as a dependent variable

To statistically control for the time since suicide, it was entered into the equation in Step 1. In Step 2, CG-T1 and CG-T2 were entered into the equation to control for the CG trajectory. In Step 3, the main effect of shame was entered to examine its contribution to CG-T3 beyond the CG trajectory. In step four, the main effect of self-disclosure was entered. In the final step, the interaction of shame and self-disclosure was entered to examine their combined contribution to CG-T3 beyond each variable’s discrete contribution.

Overall, the total set of variables explained 68% of the variance for CG-T3, F_(6,145) = 54.24, p < .001. As seen in Table 2, in Step 1, time since suicide did not significantly contribute to CG-T3. At Step 2, CG-T1 and CG-T2 contributed 41% to the variance, F_(6,145) = 54.24, p < .001. In Step 3, a model incorporating the main effect of shame accounted for an additional 11% of the variance and significantly predicted CG-T3 beyond the CG trajectory. Self-disclosure did not significantly contribute to CG-T3 at Step 4. At the final step, the main interaction effect of shame and SELF-DISCLOSURE predicted CG-T3, beyond the contribution of all other variables, accounting for an additional 16% of the total variance.

Table 2. Hierarchical regressions predicting complicated grief-T3 by the study variables (N = 152).

		Complicated Grief- T3
Step	Predictors	B	Std. Error	Beta	F change	ΔR^2	R^2
1	Time since suicide	.00	.00	−.07	.84	.00	.00
2	Complicated Grief (T1)	.02	.01	.26***	52.50	.41***	.41
	Complicated Grief (T2)	.03	.01	.50***
3	Shame (T3)	.02	.01	.34***	31.88	.11***	.52
4	Self-Disclosure (T3)	−.04	.03	−.07	1.57	.01	.52
5	Shame X Self-Disclosure	.05	.03	.68***	56.16	.16***	.68

Note. *p < .05, **p < .01, ***p < .001.

3.4.2. Depression-T3 as a dependent variable

To statistically control for the time since suicide, it was entered into the equation in Step 1. In Step 2, depression-T1 and depression-T2 were entered to control for the CG trajectory. In Step 3, the main effect of shame was entered to examine its contribution to CG-T3 beyond the depression trajectory. In step four, the main effect of self-disclosure was entered. In the final step, the interaction of shame and self-disclosure was entered to examine their combined contribution to depression-T3 beyond each variable’s discrete contribution.

Overall, the total set of variables explained 65% of the variance for depression-T3, F_(6,145) = 21.19, p < .000. As seen in Table 3, in Step 1, time since suicide did not significantly contribute to depression-T3. In Step 2, CG-T1 and CG-T2 contributed 30% to the variance. In Step 3, a model incorporating the main effect of shame levels accounted for an additional 10% of the variance and significantly predicted depression-T3 beyond the depression trajectory. Self-disclosure did not significantly contribute to depression-T3 in Step 4. At the final step, the main interaction effect of shame and SELF-DISCLOSURE predicted depression-T3 beyond the contribution of all other variables, accounting for an additional 7% of the total variance.

Table 3. Hierarchical regressions predicting depression-T3 by the study variables (N = 152).

		Depression- T3
Step	Predictors	B	Std. Error	Beta	F change	ΔR^2	R^2
1	Time since suicide	.00	.03	−.01	.84	.00	.00
2	Depression (T1)	2.26	.51	.33**	31.67	.30***	.30
	Depression (T2)	2.45	.48	.36**
3	Shame (T3)	.20	.04	.35***	24.91	.10***	.40
4	Self-Disclosure (T3)	−.11	.29	−.03	.14	.00	.40
5	Shame X Self-Disclosure	−.11	.03	.65***	17.79	.07***	.47

Note. * p < .05. **p < .01. ***p < .001.

3.4.3. Moderation analyses

Following the hierarchical regression results, we performed moderation analyses of significant interactions using the PROCESS macro (Hayes, 2017, Model 1).

Two moderation analyses were conducted with CG/depression at T3 as the dependent variables and shame as the independent variable. The moderator was self-disclosure. The trajectories of CG (CG-T1 and T2) and depression (depression-T2 and T3) were entered as covariates in the related analyses.

As seen in Figure 2, a significant interaction was found between shame and SELF-DISCLOSURE in predicting CG-T3, b = −0.02, SE = 0.01, 95% CI [−.03, −.01], t(146) = −2.88, p < .01. Probing the interaction revealed that for SLSs with low and moderate self-disclosure, shame contributed strongly and positively to CG-T3: for low- self-disclosure: b = 0.52, SE = 0.08, 95% CI [0.34, 0.68], t(146) = 6.52, p < .001; for moderate self-disclosure: b = 0.19, SE = 0.04, 95% CI [0.22, 0.49], t(146) = 5.12, p < .001. However, at high self-disclosure, shame did not significantly predict CG-T3.

Figure 2. The association between shame and complicated grief at T3 as moderated by levels of self-disclosure (N = 152).

As seen in Figure 3, we found a significant interaction between shame and SELF-DISCLOSURE in predicting depression-T3, b = −0.01, SE = 0.01, 95% CI [−.02, −.00], t(146) = −2.62, p < .001. Probing the interaction revealed that for SLSs with low and moderate SELF-DISCLOSURE, shame contributed strongly and positively depression-T3: at low self-disclosure levels: b = 0.27, SE = 0.04, 95% CI [0.18, 0.37], t(146) = 5.67, p < .001; for moderate self-disclosure: b = 0.36, SE = 0.07, 95% CI [0.11, 0.27], t(146) = 4.54, p < .001. However, at high self-disclosure levels, shame levels did not significantly predict depression-T3.

Figure 3. The association between shame and depression at T3 as moderated by levels of self-disclosure (N = 152).

4. Discussion

This study aimed to examine the role of shame in increasing the risk for depression and CG over time among SLSs, and, more specifically, to explore whether SELF-DISCLOSURE moderates the impact of shame on these outcomes. Our study revealed the critical effect of shame on both depression and CG, as it contributed to both depression-T3 and CG-T3 beyond the depression and CG trajectories over time. These findings align with several studies that highlighted the importance of feelings such as shame and guilt to the deleterious effects of a loved one’s suicide (Jordan, 2020; Maple et al., 2017).

Furthermore, we found that self-disclosure significantly moderated the impact of shame on depression and complicated grief beyond their trajectories over time. While other studies have emphasized that self-disclosure contributes significantly to healing processes (e.g. Levi-Belz, 2021; Levi-Belz & Aisenberg, 2021) in the aftermath of suicide loss, to our knowledge, this is one of the very few studies demonstrating the specific impact of shame on the negative psychological sequelae of SLSs and, more importantly, the specific role of self-disclosure in these effects.

Several studies have shown that individuals with high shame are likely to report a wide range of psychiatric symptoms, including heightened anxiety, depression, and suicidal ideation (Cândea & Szentagotai-Tătar, 2018; Tangney et al., 2007). Among SLSs, shame may be followed by specific experiences such as self-stigma and increased social withdrawal. The shame about the suicide event may enhance SLSs’ sense that something is wrong with them, leading to self-stigmatization and impeding their inclination to request support and assistance to cope with their grief (Pitman et al., 2016). Moreover, SLSs have reported higher levels of rejection which induces them to hide the circumstances of the suicide death. These dynamics make it much more challenging to share intimate feelings and emotions and process death (Sveen & Walby, 2008; Wertheimer, 2013). Thus, the vicious cycle of shame continues, as the lack of self-disclosure may exacerbate the loneliness and shame experiences.

4.1. The role of self-disclosure among SLSs

Numerous studies have shown self-disclosure to be a prerequisite for healthy adjustment among SLSs (Levi-Belz, 2019; Levi-Belz & Lev-Ari, 2019; Levi-Belz, 2016). In this study, we have shown that the inclination to disclose intimate feelings moderates the association between a sense of shame and depression and CG: Among those reporting high self-disclosure, the sense of shame was less likely to contribute to depressive and CG symptoms at T3, beyond the natural trajectories of depression and CG.

What can explain this moderation? Disclosing intimate feelings and thoughts with a significant other may facilitate emotional processing (Pennebaker, 1997). These emotional processes may inhibit the emergence of adverse sequelae of loss and even contribute to personal adaptation and growth among SLSs (Feigelman et al., 2018b; Levi-Belz, 2016, 2017). Another explanation for self-disclosure’s positive impact is the increased perception of family and friends’ support, which, in turn, helps the mourner deal with bereavement by alleviating the symptoms of grief and facilitating an adaptive adjustment in the aftermath of loss (Bottomley et al., 2017; Bottomley et al., 2019). Thus, it is plausible to suggest that SLSs with low levels of self-disclosure will be inclined to eschew social activities and engagement (Kreiner & Levi-Belz, 2019). Avoiding social interaction, in turn, may lead to solitude and loneliness, increasing self-stigmatization and, as a result, higher levels of depression and CG (Levi-Belz & Lev-Ari, 2019). However, by sharing personal and emotional information about the suicide loss with others, SLSs increase the likelihood of experiencing higher levels of social support and even a sense of belongingness, an essential component of mental health generally. Moreover, self-disclosure can carry the benefit of decreasing SLSs’ psychological difficulties, such as depression and suicide ideation (Van Orden et al., 2010). In recent studies, ii was found that experiences such as belongingness serve as a factor that diminishes CG levels (Levi-Belz & Aisenberg, 2021) and depression (Levi-Belz & Ben-Yaish, 2022) in the aftermath of suicide loss. Thus, grief and depression levels may also be related to the nature of SLSs’ interpersonal engagement.

Moreover, disclosing intimate feelings about suicide loss to significant others, such as the inherent guilt and shame following the suicide death, may allow SLSs to gain new perspectives about their situation. They can then adopt new narratives through reframing that may contribute to reducing depression, anxiety, and posttraumatic symptoms (Villacieros et al., 2014). Indeed, self-disclosure and social interaction may have a meaningful role in organising the intrapsychic response to bereavement and distress (Levi-Belz & Lev-Ari, 2019) and are more likely to yield a constructive mental perspective regarding the loss of loved ones. As Lepore et al. (1996) suggested, social sharing of grief may be a strategy to facilitate emotional desensitisation: ‘By talking with supportive and empathic others, trauma survivors may be able to contemplate and tolerate aversive trauma-related thoughts for a longer period of time than they would on their own’ (Lepore et al., 1996, p. 271). It can thus be suggested that self-disclosure may enable SLSs to better understand themselves in relation to the suicide loss and their life in its wake, resulting in lower levels of CG and depression, partly by lowering the shame levels that often accompany such a loss.

The current findings should be interpreted with caution due to several significant limitations. Firstly, whereas depression and CG were evaluated using well-validated questionnaires, the lack of direct interviews to assess these symptoms may have resulted in an incomplete or inaccurate appraisal. Second, applying a snowball sampling technique to increase the sample size (resulting in a disproportionality of female and highly educated participants) may limit the study’s generalizability. Moreover, the study sample’s specific cultural scripts relating to grief and sharing experiences may further constrain the findings’ generalizability to other cultures.

4.2. Conclusions and implications

The findings emphasize the role of shame in increasing both depression and CG among SLSs. Moreover, they highlight the importance of self-disclosure as a buffer against the harsh consequences of shame in the aftermath of suicide loss. Whereas shame may be inherent among SLSs, the study results emphasized the critical role of interpersonal abilities, such as self-disclosure, in facilitating inner psychological processes and gaining social support, which, together, may help SLSs cope better with their tragic loss, even years after the suicide event. Thus, we can cautiously surmise that the healing process of SLSs derives, at least partly, from their interpersonal experience, which may change the perceptions SLSs hold when dealing with the suicide of their loved one.

These findings indicate several practical implications. Our findings clearly point to the need to address the roles of self-disclosure as risk factors (if low) and protective factors (if high) when SLSs experience shame. For example, primary psychoeducational interventions designed explicitly for suicide survivors (Wittouck et al., 2014), such as facilitating self-disclosure behaviour, may help survivors acquire support and gain new perspectives on their situation. Psychotherapy protocols aimed at helping individuals understand their actual relationships with others may particularly benefit those with low self-disclosure. Interpersonal psychotherapy (IPT) is one such approach in its focus on social communication as an essential feature of treatment beyond the underlying psychiatric illness (Markowitz & Weissman, 2004). Another important protocol is The Expert Companionship approach (Tedeschi & Blevins, 2015), which integrates empathic support and psychological knowledge. This approach may help change the SLSs’ perspective by facilitating self-disclosure skills, thus helping decrease depression and CG symptoms among SLSs (Khazem et al., 2015). Indeed, considerably more research is needed to help determine the tailored interventions that may ease CG and depression among SLSs, highlighting the role of interpersonal factors such as self-disclosure. Questions relating to how much disclosure is desirable, who would be the best recipient for sharing, and what would be the optimum medium (e.g. Consedine et al., 2007; Towner et al., 2022) remain open. However, we hope this study facilitates more rigorous longitudinal explorations that include objective measures (e.g. interviews) on this issue.

Author statement contributors

YLB designed the study and analysed the data, and YLB and SH wrote the final manuscript. Both authors have approved the final article.

Acknowledgments

This study is dedicated to my dearest friend, Dafni Assaf, who was one of the greatest leaders of the suicide prevention programme in Israel.

Disclosure statement

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

Data availability statement

The data that support the findings of this study are available from the corresponding author, YLB, upon reasonable request.

Additional information

Funding

This work was supported by the American Foundation for Suicide Prevention [grant number PRG-2-083-14].