Examining grieving problem correlates of anticipation of the death vs. shock among overdose death and suicide bereaved adults

Abstract

We conducted an online survey with individuals bereaved by drug deaths (n = 115), suicide (n = 185) and sudden natural deaths (n 103), including questions about anticipating the death vs. shock, other behaviors, established measures of mental health and grieving difficulties. Results showed more drug death bereaved anticipating their loved one’s death while suicide bereaved were more shocked. Death shocked respondents showed greater PTSD, other mental health, and grieving problems; those anticipating the death had less PTSD, grieving problems, and engaged in more meaning-making. These contrasting patterns should call for diverging clinical strategies in offering aid to these traumatically bereaved mourners.

Deaths of despair, defined by Case and Deaton (2020) as deaths occurring among middle- age white males due to alcohol-related causes, drug overdoses, and suicides, have shown sharp increases in recent years. While much of this analysis was conducted prior to the onset of the COVID-19 pandemic, most recent statistics show a 25% rise of U.S. alcohol-related fatalities from 2019 to 2020 (White et al., 2022). Between April 2020 and April 2021, during the first year of the deadly, isolation-inducing coronavirus pandemic, the Centers for Disease Control and Prevention estimated that the drug overdose death rate reached an all-time high of over 100,000 deaths (CDC, 2022). Suicide rates have also increased, although less dramatically compared to the numbers of drug deaths. The U.S. provisional count for 2020 was 45,855 (Curtin et al., 2021). Clinicians treating suicide and drug death bereaved clients may wonder how similar these two populations may be and how their needs for mental health care may differ. Existing research suggests much overlap between these two groups of mourners. Evidence suggests that both populations of decedents and their bereaved are likely to be highly stigmatized (Feigelman et al., 2012; Valentine, 2018). Commonly, decedents and their families are socially rebuked for their drug use, criminality, and/or perceptions of enablement (Feigelman et al., 2012). Suicide casualties and their collateral kinfolk report being admonished for harboring mentally ill members, encouraging self-stigmatization and deep shame (Carpiniello & Pinna, 2017). Indeed, virtually all qualitative studies of drug death bereavement have described their mourners as socially ostracized in much the same way as suicide survivors have been described (Cvinar, 2005; Titlestad et al., 2021; Valentine et al., 2016). The perceived and actual stigmatization makes it necessary for many to retreat, lest they be subjected to disenfranchized grief from social intimates.

The only two available comparative surveys of these two populations (Bottomley et al., 2022; Feigelman et al., 2011), also found much similarity between suicide and overdose loss survivors in their mental health functioning. Bottomley and colleagues found non-significant differences in the rates of clinically elevated symptoms of depression (51–54%), anxiety (31–42%), and PTSD (43–52%) among suicide and overdose bereaved, both of whom were significantly higher from rates experienced by sudden natural loss bereaved (e.g., fatal cardiac arrest). Results indicated much convergence in experiencing severe grieving, captured by prolonged grief disorder symptoms, compared to sudden-natural death bereaved. Very importantly, Bottomley et al. (2022) found suicide bereaved respondents, with a nuanced and intricate measure of suicidal thinking significantly higher in suicidal ideation, suggesting greater suicidal risks for these mourners.

Prior qualitative studies involving the drug death bereaved have shown that many anticipate the deaths of their deceased loved ones. The first reported observation of anticipation of the death was mentioned by da Silva and associates (2007). In this study of six Brazilian drug death bereaved, Da Silva and colleagues found three informants exhibited a “veiled preparation” for the death, anticipating the deaths of their loved ones, and concluded that such knowledge left these mourners with ambivalence, a mixture of pain and relief, ultimately placing them in better positions to adapt to their loved one’s deaths. For the remaining informants that were shocked at the deaths, a mixture of anger, guilt, helplessness, indignation, and shame was indicated. Since the da Silva study, other qualitative studies of drug death bereavement have found anticipation of the death among their overdose death bereaved informants (Feigelman et al., 2018; Templeton et al., 2016; Titlestad et al., 2021).

More recently, Valentine (2018) claimed that anticipation of the death acts as a stressor for the bereaved, exacerbating grief-related difficulties above and beyond other events, such as the deceased’s previous overdoses or suicide attempts, the stigma they are/ were exposed to, and the mourner’s own substance use. This hypothesis seems contrary to the notions advanced by da Silva and associates (2007), as well as with research evidence accumulated on surprise and shock in suicide bereavement. Maple et al. (2007) completed a qualitative analysis of 22 Australian suicide bereaved parents, some of whom had advance preparation for the death based upon their children’s previous suicide attempts or prior mental health treatment, reducing their grief difficulties after their children’s deaths. This was contrasted against other suicide bereaved parents whose children’s deaths were sudden and unexpected, leaving them in a constant state of grief, impinging adversely on familial and social functioning. Feigelman et al. (2012) investigated the Maple hypothesis and obtained ambiguous results, finding shock and surprise positively and significantly related to grief difficulties among all 575 bereaved parents, but these same variables did not relate significantly when tested exclusively among the 462 suicide bereaved parents. Thus, these results reflect ambiguity on overall patterns of correlations between anticipation and shock at the death and mental health complications associated with these traumatic losses.

Beyond this specialized realm of traumatic deaths (e.g., fatal drug overdose and suicide), the phenomenon of anticipatory grief has long been speculated as having an effect on the expression and trajectory of bereavement in the grief literature. The first use of this term was developed by Lindemann (1994) applying this term to soldier’s relatives adapting to their loved one’s military service absences, by finding it problematic to re-integrate them again into civilian life. This concept generated much discussion and criticism among thanatologists for its lack of conceptual and methodological precision (Gilliland & Fleming, 1998). However, despite such debates, much empirical research on anticipatory grief and bereavement has been conducted with few clear conclusions. By 2006, in a summary article (Reynolds & Botha, 2006), evaluating all studies found much conflicting results owing to diverging definitions of key terms, and lack of agreement about study time frames, among other methodological differences. The net result yielded inconsistent findings with some studies showing positive results, others negative, and still others demonstrating mixed results. A more recent investigation of anticipatory grief and posttraumatic growth among 120 adult respondents who anticipated deaths of loved ones from terminal illnesses found “anticipatory grievers who employ proactive coping techniques when facing the expected death of a loved one may exhibit more evidence of growth than those who remain inactive in their coping processes … [perceiving it as] a psychological threat” (Rogalla, 2020, p. 126). This finding suggests anticipation of a loved one’s death may not only be connected to posttraumatic growth, but could also be associated with meaning making, a related construct theorized to be an adaptive response to loss (Lancaster & Carlson, 2015; Neimeyer, 2006).

The degree of shock or surprise in the death, has been reported in the bereavement literature as having deleterious effects on the survivors’ well-being and adjustment course (Andriessen et al., 2019; Reed, 1998; Sveen & Walby, 2008). Research on the broad experience of sudden loss indicates that the degree of shock is associated with adverse outcomes, particularly PTSD (Hargrave et al., 2012).

To date, very little attention has been placed on investigating the role of anticipating the death versus the degree of shock at the death and the diverging mental health outcomes associated with each behavior pattern among the traumatically bereaved. We also need to know what factors might lead a traumatically bereaved person to anticipate their loved one’s death or remain shocked by it. We hypothesize that as significant others, specifically suicide and overdose loss survivors, acquire knowledge of their loved one’s challenges of mental health and drug use they will become more inclined to anticipate their deaths. We predict that as a significant other becomes more aware of a deceased’s suicide attempts, suicide warning signs, their prior drug overdoses or prior treatments, they will be more inclined to anticipate the deaths.

Another important study objective will be to clarify which bereaved, whether suicide or drug death, will be more likely to anticipate the death than the other. The prior research record suggests that drug death bereaved will be more likely to anticipate the death than the suicide bereaved. We suspect that the prior drug uses of the individual, their prior drug rehabilitation experiences, and contacts with law enforcement will make drug use more readily apparent to the drug user’s family members, encouraging anticipation. Contrastingly, in many cases of eventual suicides, the sometimes subtle evidence of harboring a mental illness, often eludes family members’ detection, leading to surprise at the death.

A final central aim of the present investigation will be to better illuminate the associated mental health and grief differences correlated with anticipation of the death or being shocked by it. From the evidence thus far, more evidence suggests shock and surprise at the death will be accompanied by more adverse mental health symptoms and greater grieving difficulties. By contrast, we also suspect that anticipation of the death will be associated with higher levels of meaning making and less PTSD. Meaning-making is claimed to be a fundamental part of the grieving process, enabling the bereaved to come to terms with a significant other’s death (Neimeyer, 2014; Stroebe & Schut, 2001). Previous research has documented that those scoring higher on the Integration of Stressful life Experiences Scale (the ISLES short form), measuring meaning making, will show evidence of better mental health and less prolonged grief (Holland et al., 2014). Exploring these hypotheses should advance bereavement knowledge and be useful to clinicians helping mourners cope with their losses and to navigate in society often encountering limited social supports.

Method

Participants and procedures

This research was based on a convenience sample of bereaved adults affiliated with support groups serving the suicide bereaved and the drug death bereaved, explicitly from the American Association of Suicidology list-serv, the American Foundation for Suicide Prevention and Grief Recovery After Substance Use Passing (GRASP) groups, and word of mouth or relevant social media exposure where they could be encouraged to participate in a bereavement survey research. Participants were offered a modest monetary incentive ($5 Amazon eGift Card or a donation made to an organization/charity of their choice for survey completion). Recruitment of individuals bereaved from the sudden natural death of a friend or loved one were solicited from online social media posts targeting general loss support organizations. In addition, undergraduate students from a large university located in the Mid-South region of the United States who had experienced the sudden natural death of a friend or loved one within the past five years were recruited through an online subject pool exchanging course credit for research participation.

Data collection occurred via online surveys using Qualtrics. Upon IRB and ethics review approval from The University of Memphis (PRO-FY2018–685), participants over 18 years of age, experiencing the death of a friend or loved one due to these diverging causes, within the previous five years, were recruited.

The current sample, as shown in Table 1, included 403 adults, aged 18–79 (M = 42.31 years, SD = 17.17) who lost a friend or loved one due to opioid overdose (n = 115), suicide (n = 185), and sudden natural causes (e.g., acute myocardial infarction; n = 103). The majority of participants lost a child (n = 154; 38.2%), sibling (n = 63; 15.6%), parent (n = 38; 9.4%), or grandparent (n = 38; 9.4%), with remaining participants experiencing death of a distant family member (e.g., cousin, aunt/ uncle; n = 32; 7.9%), spouse/partner (n = 36; 8.9%), or friend (n = 28; 6.9%). A small proportion of respondents consisted of other relationship losses (e.g., coworkers; n = 14; 3.5%). The vast majority identified as female (n = 338; 83.9%) and white (n = 301; 74.9%). Greater than a quarter represented racial and ethnic minorities, with 49 identifying as African American/ Black (12.2%), 25 as Hispanic/Latino (6.2%), 22 as Asian (5.5%), and 5 as Native American (1.2%). Regarding socioeconomic indicators, nearly half had completed a college degree (n = 195; 48.4%), and a sub stantial proportion completed some college (n = 142; 35.2%), while all others obtained a high school degree (n = 68; 16.4%). Yearly household income varied, with 10.9% reporting a household income below $10,000 (n = 43), 28.8% between $10,000 and $50,000 (n = 114), 34.8% between $50,000 and $100,000 (n = 138), and 25.5% above $100,000 (n = 101).

Table 1.

Demographic variables, characteristics of the death, and unadjusted and adjusted mean levels of anticipation and shock by loss type.

			Loss type
Features	Overdose loss (n = 115)		Suicide loss (n = 185)		Sudden-natural loss (n = 103)		Analysis
	n	%	n	%	n	%	χ2	p
Female	165	89.2	98	85.2	75	72.8	13.32	.001
Ethnicity							51.67	<.001
White	148	80.0	99	86.1	54	52.4
Black	13	7.0	5	4.3	31	30.1
Latinx	13	7.0	4	3.5	8	7.8
Person who died							120.66	<.001
Child	73	39.5	74	64.3	7	6.8
Sibling	37	20.0	19	16.5	7	6.8
Parent	17	9.2	2	1.7	19	18.4
Spouse/partner	24	13.0	8	7.0	4	3.9
Relative	15	8.1	3	2.6	53	51.4
Friend	15	8.1	6	5.2	7	6.8
	M	SD	M	SD	M	SD	F	p
Age	50.26	13.87	45.51	15.87	27.67	13.66	71.54	<.001
TSL	21.83	15.29	25.99	19.54	25.64	16.23	2.17	.116

Note. TSL: time since the death (in months).

Measures

Selected demographic variables were obtained, including sex, age, ethnicity, kinship category of the decedent, time since loss, and the respondent’s interpretation of the cause of death. Time since the loss varied as shown in Table 1, with a range between 2 months and 5 years (M = 25.12 months, SD = 18.09 months. We also employed a battery of questions assessing the respondent’s knowledge of the deceased’s mental health history and treatments prior to the death and known drug using behaviors. Pre-death levels of closeness with the decedent were assessed using the closeness subscale of the Quality of Relationships Inventory—Bereavement Version (QRI-B; Bottomley et al., 2019) measuring closeness to the deceased prior to the death. Questions included such items as, “To what extent could you count on this person to help you if a family member very close to you died?”). Items were rated on a 4-point Likert scale ranging from 1 (not at all) to 4 (very much). The Closeness subscale demonstrated an internal consistency reliability coefficient (α) of 0.88.

Consistent with findings from three qualitative studies of Brazilian (Da Silva et al., 2007), British (Valentine, 2018) and American drug death bereaved (Feigelman et al., 2018) who expressed widespread anticipation of the deaths of their deceased loved ones, we developed our own survey measure of this and its obverse, being shocked at the death. Anticipation and/or shock at the death was measured by single items: “In the time leading up to your loved one’s death, how likely did you consider their death to be?” Responses were measured on a five-point scale with the following choices: (1) Extremely unlikely (I was completely blindsided), (2) Unlikely, (3) Neutral, (4) Likely, (5) Extremely Likely (I anticipated it). Anticipation of the death yielded a mean of 2.22 (SD = 1.27). “Was your loved one’s death particularly shocking?” with the following responses measured on a five-point scale: (1) Not at all shocking, (2) Slightly shocking, (3) Somewhat shocking, (4) Moderately shocking, (5) Extremely shocking. Shock at the death yielded a mean of 4.40 (SD = .96), indicating a relatively high overall shock level.

An abbreviated version of the Grief Experience Questionnaire (GEQ) was administered to respondents including 37 items from the original 59 item-scale focusing on themes such as guilt, shame, stigma and responsibility (Bailley et al., 2000; Barrett & Scott, 1989). Scale items were highly inter-correlated generating an α =.94. Each scale item was graded on a five-point scale of from 1 (never) to 4 (almost always) of experiencing this event. Representative questions included: “feel like a social outcast,” “feel like others may have blamed you for the death.” This abbreviated GEQ scale generated a mean of 92.49 (SD = 27.08) with scores ranging from a low of 37 to a high of 165.

The Inventory of Complicated Grief (ICG; Prigerson et al., 1995) consists of 19 items that measure severity of PGD symptoms, such as yearning or longing for the deceased, numbness, meaninglessness, mistrust, difficulty with acceptance, and identity confusion. Items of the ICG are rated on a five-point Likert type scale that primarily assesses frequency (0 = never to 4 = always). The ICG yielded an internal consistency score of a 0.91, with a mean of 32.14 (SD = 14.52) and a range of 0 to 62.

We also administered the Integration of Stressful Life Experiences Scale short-form (ISLES-SF), encapsulating how well or poorly one may have succeeded at meaning-making after their loss (Holland et al., 2014). The ISLES-SF consists of six items: “I have difficulty integrating this loss into my understanding about the world”; “I am perplexed by this loss”; “since the loss I don’t know where to go in my life”; “I don’t understand myself anymore since this loss”; “this loss has made me less purposeful.” Respondents were asked to agree or disagree whether each of these six statements accurately reflected their feelings. Scale items were highly inter-correlated yielding α .90. The mean was 17.47 (SD = 6.67) and scores ranged from a low of 6 to a high of 30.

The PTSD Checklist for DSM-5 (Blevins et al., 2015) is a 20-item self-report measure of past month PTSD symptom criteria. All 20 items correspond with the DSM 5 symptoms and their corresponding clusters. Sample items include, “In the past month, how much were you bothered by: ‘repeated disturbing dreams of the stressful experience’ and ‘feeling jumpy or easily startled’”. Respondents asked to respond to items with the sudden death of a close other as the index event. Items were summed and rated on a scale from 0 (Not at all) to 4 (Extremely), with a range of scores from 0 to 80. In the current sample, the PTSD symptoms checklist yielded an α 0.93.

The Patient Health Questionnaire 8 (PHQ-8; Kroenke et al., 2009) is an 8-item self- report measure of depressive symptoms: (e.g., “Over the past 2 weeks, how often have you been bothered by: ‘little interest or pleasure in doing things’ and ‘Poor appetite or overeating’”). Items rated on the scale from 0 (Not at all) to 3 (Nearly every day), with a range from 0 to 24. The PHQ-8 showed an internal consistency of α 0.88.

To assess anxiety symptomatology, the GAD 7 item scale (GAD-7; Spitzer et al., 2006) was included. The GAD-7 assesses if participants experienced anxiety symptoms over the previous two weeks (e.g., “How often have you been bothered by: ‘feeling nervous, anxious, or on edge’ or ‘trouble relaxing’”). Items are rated on a scale from 0 (Not at all) to 3 (Nearly every day), with a range from 0 to 21 and summed to produce a total severity score. The GAD-7 yielded an alpha coefficient of α 0.90.

Suicide risk was assessed using the four-item Suicidal Behaviors Questionnaire Revised (SBQ R; Osman et al., 2001), a brief self-report instrument that measures previous suicide attempts, frequency of suicidal ideation, suicidal communication, and the subjective likelihood of a future suicide attempt, yielding an internal consistency of α 0.84.

Analytic plan

All analyses were conducted using the statistical package for the social sciences (SPSS) version 25 (SPSS Inc.). Upon inspection, a total of three respondents had missing values on a select number of outcome measures. Missing values were replaced with the respondent’s mean value for other items in the outcome scale as fewer than 25% of the values were missing for these cases (e.g., Unterhitzenberger et al., 2020). Correlation matrixes were employed to evaluate the relationship between anticipating the death, shock of the death, and pre-loss factors (e.g., decedent psychiatric diagnoses, previous suicide attempts, previous drug use) among the suicide and overdose bereaved participants. Next, we examined univariate associations between anticipation, shock, mental health, grieving difficulties, and meaning-making measures among all survivors (i.e., suicide, overdose, sudden-natural loss).

To evaluate the hypothesized differences in anticipating the death and the degree of shock in the death between suicide, drug death, and sudden-natural bereaved subgroups, controlling for the effects of a respondent’s age, pre-death closeness and other pre-death factors, we conducted two analysis of covariance (ANCOVA) models.

Results

Associations between anticipation, shock, pre-death factors, and outcomes

The correlation analysis shown in Table 2 found, as expected, anticipation and shock were significantly and inversely related to one another (r = −.52, p < .001). Previous decedent psychiatric diagnoses (r = .28, p < .001) and treatment (r = .28, p < .001), as well as previous suicide attempts (r = .13, p < .001) were significantly associated with greater anticipation of the death. Similarly, when the decedent had prior mental health diagnoses (r = −.13, p = .03) or a history of drug use (r = −.16, p =.04)survivors tended to experience less shock in the death.

Table 2.

Correlations between anticipation, shock, demographic and loss-related variables, and mental and grief-related outcomes {N = 403).

		1	2	3	4	5	6	7	8	9	10	11	12	13
1	Anticipation of the death	-
2	Shock of death	−.52**	-
3	Respondent age	.04	.09	-
4	Time since the death	.01	.03	.05	-
5	Pre-death closeness	−.11*	.21**	.04	.02	-
6	Respondent previous psychiatric Dx	.02	−.10*	.01	.07	−.08	-
7	Grief complications (GEQ)	−.11*	.28**	.24**	−.06	.08	−.21**	-
8	Prolonged grief symptoms	−.13**	.28**	.29**	−.20**	.28**	−.15**	.62**	-
9	PTSD symptoms	−.10*	.25**	.14**	−.20**	.18**	−.17**	.71**	.64**	-
10	Depressive symptoms	−.06	.14**	.01	−.24**	.11*	−.23**	.53**	.52**	.66**	-
11	Generalized anxiety symptoms	−.08	.13**	−.23**	−.17**	.08	−.21 **	.44**	.43**	.57**	.70**	-
12	Suicide risk	−.08	.13**	.01	−.04	.12*	−.27**	.33**	.26**	.36**	.42**	.34**	-
13	Meaning making	.16**	−.22**	−.27**	.14**	−.14**	.05	−.61**	−.56**	−.58**	−.49**	−.34**	−.29**	-

Note.

^*p < .05;

^**p < .01.

Next, we investigated with t-test comparisons, whether there were any significant differences between the subgroups of relationship categories and anticipating the death or being shocked by it. We found no significant differences between the parents (n = 154) and all other relationship categories (n = 249) in anticipating the death or being shocked by it.

The correlation analysis also found: anticipation of the death was inversely related to pre-death closeness (r = −.11, p = .03), grief complications (r = −.11, p = .03), PGD symptoms (r = −.13, p = .01), PTSD symptoms (r = −.10, p = .04), and positively associated with meaning making (r = .16, p < .01). With regard to the degree of shock, the closer the pre-death relationship, the more likely shock was experienced (r = .21, p < .001). Shock appeared to be significantly associated with all outcome measures, with greater levels of shock being significantly associated with grief complications (r = .28, p < .001), PGD symptoms (r = .28, p < .001), PTSD symptoms (r = .25, p < .001), depressive (r = .14, p < .001) and anxiety (r = .13, p < .001) symptoms, and suicide risk (r = .13, p < .001). Shock was inversely related to meaning making (r = −.22, p < .001), such that more shock was associated with lower levels of meaning made of the death.

Differences in anticipation and shock of death between loss groups

Table 3‘s ANCOVA results showed of the covariates entered into the model, only the age of the respondent was significantly related to the level of anticipation of the death, F(1, 395) = 4.06, p = .04. Importantly, there a was significant effect of cause of death on levels of anticipation of the death after controlling for the covariates, F(2, 395) = 20.648, p < .001. Estimated marginal means were significantly different across groups, with sudden-natural loss survivors reporting the greatest level of anticipation of the death (M_adj = 2.94, SE = .14), followed by overdose bereaved (M_adj = 2.24, SE = .12), survivors of suicide loss (M_adj = 1.80, SE = .09). A second ANCOVA model was run with the same variables examining the degree of shock between cause of death groups. Pre-death closeness between the respondent and the decedent was significantly related to the degree of shock in the death, F(1, 395) = 12.79, p < .001. Cause of death was significantly related to the degree of shock after controlling for the effects of respondent age and pre-death closeness to the decedent, F(2, 395) = 10.23, p < .001. Like anticipation of the death, estimated marginal means were significantly distinct across loss groups, with suicide loss survivors reporting the greatest level of shock in the death (M_adj = 4.62, SE = .08), followed by overdose loss (M_adj = 4.40, SE = .09), and those bereaved by a sudden-natural death (M_adj = 3.99, SE = .11).

Table 3.

Means (unadjusted and adjusted) of anticipation and shock between loss groups and ANCOVA results.

					Loss group
	Overdose loss (n = 115)				Suicide loss (n = 185)				Sudden-natural loss (n = 103)				Analysis
	M	SD	M adj	SE	M	SD	M adj	SE	M	SD	M adj	SE	F	p	${\eta }_{\text{p}}^2$
Anticipation	2.37	1.25	2.24	.12	1.93	1.21	1.80	.09	2.21	1.27	2.94	.14	20.65	<.001	.10
Shock	4.37	0.99	4.40	.09	4.63	0.75	4.62	.07	4.00	1.43	3.99	.11	10.29	<.001	.05

Note. Means adjusted after accounting for respondent age, pre-death closeness, decedent prior drug use, decedent mental health history and treatment, and suicide warning signs.

Discussion

A central focus of this research has been to investigate two potentially oppositional bereavement adaptations: the anticipation of the death vs. being shocked and surprised by it. We expected that those bereaved who experienced one of these sets of feelings would not experience the other. This survey data supported this with a −.52 correlation coefficient between these two variables of anticipation versus shock.

We further speculated that as a bereaved had greater knowledge of the potentially problematic pre-morbid behaviors of the deceased they would be more likely to anticipate their death. This basic hypothesis was supported with those bereaved that had greater knowledge of the deceased’s prior drug uses, their previous treatment experiences and those showing greater numbers of suicide warning signs—all these variables were significantly associated with being more inclined to anticipate the death. Thus, a bereaved person’s greater knowledge of the pre-morbid behaviors of the deceased sets the stage for anticipating their death.

Our comparative data also supported another expectation borne out from the annals of qualitative research findings showing drug death bereaved more inclined to anticipate the death (Feigelman et al., 2018; Templeton et al., 2016; Titlestad et al., 2021) than the suicide bereaved. Research among the suicide bereaved has more often shown shock and surprise at the death (Feigelman et al., 2012; Sveen & Walby, 2008). We found the drug death bereaved more inclined to anticipate their deceased loved one’s death, compared to the suicide bereaved. This was demonstrated in our ANCOVA results, which controlled for potential confounding factors such as age, closeness, prior drug use, prior treatment exposure and suicide warning signs. And equally importantly, we found corresponding empirical support showing the suicide bereaved significantly more likely to experience greater surprise and shock at the death compared to the other bereavement groups, again controlling for these same confounders. This confirms for the first time, with systematic survey data, what has been observed in various qualitative and quantitative studies of both groups. A larger question now looms ahead on what may be the significance of these differences in impacting the care and helping of these bereaved groups.

Although one analyst (Valentine, 2018) posited that anticipation of the death would act as a stressor, exacerbating the grief problems of the bereaved, most other researchers thought shock and surprise would be a stronger correlate of grief problems and mental health difficulties following a loss, than anticipation (da Silva et al., 2007; Feigelman et al., 2012; Maple et al., 2007). This data showed statistically significant correlations of shock with grief problems and poorer mental health and practically no evidence of similar associations with anticipation of the death. Those experiencing shock and surprise at the death were significantly more likely to have greater grief difficulties; more PGD symptoms, PTSD, depression, and anxiety; and to be at higher suicidal risk on our scales. Correspondingly, those anticipating the death were significantly less likely to experience grief problems, PGD and PTSD symptoms. Thus, our data showed that anticipation of the death was associated with helping to insulate the bereaved from grief complications and trauma.

The correlational analysis showed further adaptive response benefits from significant positive associations between anticipation of the death and meaning-making, an experience helping the bereaved to ultimately accept their loss and move forward with their lives. While these associations are unmistakable in our data, the correlations are modest in size. We suspect that among the newly bereaved the benefit of anticipation vs. shock will have its greatest impact. Among longer-term bereaved we do not imagine that the benefits of anticipation vs. shock will still be so impactful. Yet, it will be task for future research to investigate whether these hypotheses about the potential value and benefits of anticipation vs. shock will be further confirmed and will apply over the long-term trajectory of a mourner’s grief.

This is probably the first comparative research with overdose death and suicide bereaved respondents that has identified a potentially important difference that might suggest diverging strategies to best help these two diverging groups. Up until now, the limited research evidence could not suggest much if any different practices in attempting to allay their grieving difficulties. If anticipation of the death is more likely to occur among the drug death bereaved this may suggest that this population will be less subject to PTSD and other mental health deficits, compared to their suicide bereaved counterparts. It should also suggest that querying the bereaved on whether they anticipated the death of their loved one, or were shocked from it, will be helpful to know in providing support and substantive aid to these mourners.

During the past decade there has been a profound upsurge of drug death bereavement research, vastly outpacing the research output devoted to this topic from previous decades (see Titlestad et al., 2021). Yet, most all of this research has yielded discreet and non-comparative studies of drug death mourners. If we are to better understand the distinctive characteristics of these mourners, their unique loss experiences and how they may converge and diverge from those of other mourners we will need more comparative studies. Addressing this question in the past, has only produced two such studies, besides the present one (Bottomley et al., 2022; Feigelman et al., 2011). Clearly, future research should attempt to complete more comparative studies, utilizing similar study protocols for all bereaved subgroups.

We must be mindful that in any cross-sectional study such as the present one that assertions of causation or temporal priority can be hazardous to offer. It is entirely possible that those respondents endorsing more bereavement-related distress continue to perceive their losses as more shocking, partially as a function of their own ongoing distress.

Lest these findings be misinterpreted or misunderstood, it is important to keep in mind that the grief difficulties, complicated grief and poorer mental health does not differ all that substantially between overdose drug death bereaved and the suicide bereaved. The qualitative studies done by the British and Norwegian researchers has shown that even where the close relatives of the deceased anticipate the deaths of their loved ones, they still are likely to experience strong yearnings for their lost loved one (Templeton et al., 2016; Titlestad et al., 2021). In addition, two diverging quantitative studies have found great convergence between drug death and suicide bereaved in experiencing complicated grief and having poorer mental health (Bottomley et al., 2022; Feigelman et al., 2011). What then, explains this apparent anomaly? What brings down the mental health and ignites the grieving problems of these bereaved is their stigmatized positions in society. Never fully accepted after having lost a family member to drugs or a self-killing, these mourners are often denied social support and marginalized. Stigmatization has been repeatedly shown to be a high correlate of poorer mental health and enduring grief problems, long after the occurrence of a loss (e.g., Oexle et al., 2020; Scocco et al., 2019).

Limitations

Like many other surveys of the traumatically bereaved this study was based on a convenience sample of the subject populations and may not be fully representative as a more costly household-based-study would have yielded. We overrepresented women, whites, the more highly educated and those more inclined to seek treatments.

There is an additional potential bias problem from the sudden natural death bereaved group, which included many collegiate volunteers. They differed from the other two subgroups, in being younger, more racially diverse, and with fewer first-degree relatives of the deceased, compared to the overdose and suicide bereaved respondents. Yet, since the primary focus of this research was on the contrasts between the overdose death bereaved and the suicide mourners, this lack of complete matching had a negligible impact on our findings.

An additional limitation to the current study is its reliance on single-item measures of anticipation vs. shock at the death. Ideally a more elaborate measure of these behaviors would have been preferred. Yet, that these two question items correlated so highly as predicted, suggests general validity to the respondents’ understanding of these questions.

Future studies should employ large, representative samples whenever possible in order to obtain results that are more inclusive of all overdose loss survivors. Another inherent limitation in the current study relates to the utilization of self-report instruments. Though valid assessment tools that can differentiate those who may meet diagnostic criteria for psychiatric conditions, these measures are not sufficient to provide diagnostic determinations. As such, research employing structured clinical interviews with representative samples of traumatically bereaved loss survivors should provide the most accurate depiction of the rates of psychiatric diagnoses within this population.