Potential risk factors for acute grief during the COVID‐19 pandemic: The mediating role of avoidance processes

Abstract

Bereavement during the COVID‐19 pandemic may have some unique characteristics that become potential risk factors (e.g., absence of grief rituals, no opportunity to say goodbye to the deceased and loneliness caused by social distancing) for acute grief. Avoidance processes could be significant mediators in the context of the pandemic. The current study aimed to investigate whether and how these COVID‐19‐related risk factors were related to acute grief severity. Bereaved adults (n = 319) who lost significant others during the COVID‐19 pandemic completed a self‐report questionnaire package measuring COVID‐19‐related factors, grief severity and depressive and anxious avoidance. Regression analyses suggested that among the three potential risk factors (loneliness, grief rituals and opportunity to say goodbye), loneliness was significantly associated with acute grief after controlling for basic demographic and loss‐related information. Structural equation models suggested that depressive avoidance and anxious avoidance partially mediated the associations of loneliness with acute grief severity. The findings indicate that dealing with loss during the COVID‐19 pandemic warrants further exploration concerning how potential environmental risk factors may impede adaptation to loss. Depressive and anxious avoidance processes may play important roles in grief interventions for isolated and lonely bereaved people.

Key Practitioner Message.

• Bereavement during COVID‐19 may have some common characteristics and become potential risk factors for acute grief.

• As a probable product of social distancing during COVID‐19, loneliness may become a potential risk factor for developing higher levels of grief.

• Avoidance processes could partly explain the relationship between potential risk factors for the bereaved in the context of COVID‐19 and acute grief severity. Using exposure or behavioural activation treatment procedures to reduce avoidant thoughts and behaviours may be beneficial to the bereaved.

1. INTRODUCTION

1.1. Bereavement and grief reactions during COVID‐19

The worldwide outbreak of COVID‐19 caused more than 5 million deaths by October 31, 2021 (World Health Organization, WHO, 2021). Moreover, it is still having substantial impact on people's lives. Many families have experienced significant loss during the pandemic. Notably, because of policies targeting the pandemic, such as social distancing, changes in social and cultural practices have occurred, interfering with the grief process (Eisma, et al., 2020). Although grief is a natural reaction to losing loved ones, particularly in the first 6 months, it has been estimated that approximately 10% of the bereaved experience prolonged grief reactions and that these symptoms might not be alleviated over time (Lundorff et al., 2017). Prolonged grief disorder (PGD) is now included as a mental disorder in the International Classification of Diseases 11th Revision (ICD‐11) and Diagnostic and Statistical Manual of Mental Disorders‐5‐Text Revised (DSM‐5‐TR) (American Psychiatric Association, APA, 2022; WHO, 2018). The DSM‐5‐TR described PGD as a persistent and pervasive grief response characterized by longing for the deceased or persistent preoccupation with thoughts of the deceased with emotional distress. To reach the criteria to diagnose PGD, grief responses need to persist for over 12 months and clearly exceed expected social and cultural norms for the individual's context. The ICD‐11 stipulated that to be classified as PGD, grief responses should be associated with significant functional impairment and persist for longer than 6 months. In the context of the COVID‐19 pandemic, researchers have raised concerns that a worldwide rise may occur in prevalence of PGD (e.g., Kokou‐Kpolou et al., 2020). According to the diagnostic criteria for PGD in the DSM‐5‐TR, PGD can only be diagnosed when acute grief remains intense beyond 12 months, and acute grief responses may predict later prolonged grief symptoms (Boelen & Lenferink, 2021). Therefore, exploring the potential risk factors associated with grief reactions in the early stages of bereavement may help us implement preventive interventions against the development of prolonged grief symptoms.

Eisma and Tamminga (2020) found that people bereaved during the pandemic had higher grief levels than people bereaved before the pandemic. This suggests that identifying COVID‐19 related factors that would become potential risk factors for prolonged grief is a practical and scientific priority. Based on the current situation, researchers have highlighted that the circumstances surrounding death during the COVID‐19 pandemic may have some unique characteristics: (a) The restrictions of social distancing made it difficult to interact with relatives and friends compared with prepandemic interactions, causing more feelings of loneliness (Brooks et al., 2020; Neimeyer & Lee, 2021); (b) according to quarantine policies, funeral ceremonies were strictly restricted, and some family members could not attend ceremonies (Gesi et al., 2020; Ingravallo, 2020); and (c) because of the viral contagiousness of COVID‐19, hospital visits were strictly prohibited, with no opportunity for relatives to say goodbye to the deceased (Diolaiuti et al., 2021; Wallace et al., 2020). Although researchers have formed some projections on the potential risk introduced by changes in circumstances during COVID‐19, there are relatively few quantitative studies on the health consequences of bereavement during the pandemic. To date, a few studies have begun to explore which COVID‐19‐related factors may hamper processes of adaptation to loss. In a recent cross‐sectional study, researchers surveyed 831 adults bereaved due to COVID‐19 regarding their grief levels and how they were distressed by complications associated with the pandemic. The results showed that circumstantial risk factors such as feelings of not having a proper funeral for the deceased and not being able to say goodbye to the deceased were strongly associated with dysfunctional pandemic grief (Neimeyer & Lee, 2021). Another cross‐sectional study in China explored the associations of demographic and loss‐related characteristics with bereavement outcomes among 422 Chinese adults bereaved due to COVID‐19 and found that higher grief levels were associated with factors such as losing someone close and feeling more traumatized by the loss (Tang & Xiang, 2021).

Previous studies mainly focused on groups of bereaved due to COVID‐19. However, as suggested by some clinical scholars, the grief of people who have lost loved ones from other causes during the pandemic should not be overlooked (Reynolds et al., 2007). Eisma et al. (2021) found that people bereaved due to COVID‐19 showed significantly higher grief levels than people bereaved due to natural causes. From the results of another large‐scale survey, people bereaved due to COVID‐19 experienced higher levels of grief than those bereaved due to dementia, and was less likely to meet the criteria of PGD compared to those bereaved due to unnatural causes (Gang et al., 2022). People bereaved due to different causes shared the same complex contextual factors of the pandemic. This study aimed to systematically explore the effect of COVID‐19‐related factors on grief reactions using empirical data as an extension of previous studies. Furthermore, we wished to shed some light on potential underlying mechanism of grief reactions following the loss during the pandemic.

1.2. COVID‐19‐related factors

1.2.1. Loneliness

Loneliness is a psychological feeling that occurs when individuals' interpersonal relationships are below their expectations, often accompanied by negative feelings such as emptiness, boredom, helplessness and bitterness (Weiss, 1973). As previous research has suggested, during the pandemic, reduced social and physical contact with others was significantly related to higher amounts of perceived isolation and loneliness (Brooks et al., 2020; Smith & Lim, 2020). Several studies have shown that social isolation during COVID‐19 increased feelings of loneliness (Killgore et al., 2020; Tull et al., 2020), which could be one of the common potential risk factors for the onset of psychiatric disorders, such as depression, anxiety, posttraumatic stress disorder and prolonged grief disorder (Murata et al., 2020).

Loneliness was listed as an additional symptom of prolonged grief disorder in the DSM‐5‐TR (APA, 2022). Additionally, loneliness was a primary symptom after loss, which might activate other grief‐ or depression‐related symptoms from a network perspective (Fried et al., 2015; Robinaugh et al., 2014). Some researchers insist that loneliness mainly stems from the loss of an attachment figure (Mushtaq et al., 2014), while some researchers propose that loneliness can also be a product of social distancing restrictions during COVID‐19 (Wu, 2020). With considerable changes in how we grieve during the pandemic, loneliness may be a potential risk factor associated with higher levels of grief severity. A recent systematic literature review of loneliness in bereavement proposed that the particular circumstances of the pandemic may exacerbate loneliness, in turn impacting grieving processes (Vedder et al., 2021). Consistent with this assumption, one recent study by Murata et al. (2020) demonstrated that loneliness was significantly related to prolonged grief reactions.

1.2.2. Grief rituals

Grief rituals are considered essential in some cultures of Eastern and Western countries. Burrell and Selman (2020) offered the insight that meaningful grief rituals depend on the ability of the bereaved to attend a funeral and say goodbye to the deceased. In China, people attach considerable importance to holding formal funerals following local customs, and grief rituals are assumed to have psychotherapeutic functions (Jia, 2005). In traditional Chinese funerals, bereaved people cry out loud to vent negative emotions, and their friends and relatives give empathetic companionship and support (Qiu & Yan, 2014). The absence of funeral rituals may induce feelings of guilt and indignation, increasing the risk of developing prolonged grief (Cardoso et al., 2020). In addition to lacking traditional funerals, a lack of opportunity to say goodbye to the deceased can also increase the risk of developing prolonged grief (Kentish‐Barnes et al., 2015). Considering quarantine policies, the absence of grief rituals and the opportunity to say goodbye could be common and potential risk factors for grief severity during the COVID‐19 pandemic.

1.2.3. Mediators: Avoidance processes

Due to the strong infectivity and fast transmission of the coronavirus, people were inclined to avoid physical contact with others. Researchers have suggested that people may have developed more avoidance and withdrawal behaviours when dealing with loss during the COVID‐19 pandemic (Dominguez‐Rodriguez et al., 2022; Sirrine et al., 2021). According to the cognitive‐behavioural conceptual model of complicated grief, avoidance processes can be one of the core processes mediating characteristics of the loss event or sequelae and disturbed grief (Boelen et al., 2006). There are two distinct but correlated avoidance processes, depressive avoidance and anxious avoidance. Depressive avoidance mainly refers to avoiding activities that would contribute to adaptation, while anxious avoidance mainly refers to avoiding reminders of loss (Boelen & van den Bout, 2010). These avoidance strategies are seen as maladaptive emotion regulation strategies positively associated with grief reactions (Eisma & Stroebe, 2021). As cross‐sectional studies have suggested, depressive avoidance and anxious avoidance are positively correlated with concurrent grief symptoms (Boelen & van den Bout, 2010; Fernández‐Alcántara et al., 2021). In addition, targeting anxious and depressive avoidance with exposure and behavioural activation procedures would be useful to decrease prolonged grief symptoms (Eisma et al., 2015).

The complexity of the environment around death and multiple stressors after loss drive individuals to use more avoidance strategies. Boelen et al. (2015) found that people confronted with violent loss may be more inclined to engage in anxious and depressive avoidance, and avoidance processes emerged as significant mediators between violent/non‐violent loss and grief reactions. According to the cognitive‐behavioural conceptualization of complicated grief, avoidance strategies that are used to deal with loss are influenced by characteristics of the loss sequelae (Boelen et al., 2006). The stressors that need to be dealt with may be more complex in the context of the pandemic. With the absence of companionship and support from friends and relatives, the bereaved may be more inclined to adopt withdrawal behaviours and not willing to interact with others. However, these avoidance strategies may not be conducive to adjustment to loss.

The goal of the current study was to better understand whether and how COVID‐19‐related factors function as vulnerability factors for acute grief. The first aim was to determine whether the proposed factors were significantly related to grief severity. We hypothesized that loneliness, grief rituals and opportunity to say goodbye to the deceased would be all significantly associated with acute grief. The second aim was to clarify the pathways along which those risk factors contribute to grief severity. We hypothesized that depressive and anxious avoidance would function as mediators between the relationship of COVID‐19 related factors and acute grief. The hypothetical mediation model is presented below (Figure 1).

FIGURE 1.

Graphical display of the proposed mediation models

2. METHODS

2.1. Participants and procedure

The data were collected as a part of a larger research programme, “Psychological assistance for bereaved people during the COVID‐19 pandemic in mainland China.” For the outline of the whole programme, please see Tang et al. (2021). The study was approved by the Ethics Committee of Beijing Normal University. Participants were recruited via the internet with information on grief (e.g., WeChat). The inclusion criteria were as follows: (1) age over 18 years old; (2) loss of their significant others from 1 January 2020 to 30 June 2020, when the COVID‐19 pandemic had the most significant impact on people's lives in mainland China. The data were collected using Qualtrics software from May 2020 to January 2021. A total of 328 people participated in this study. Informed consent was obtained from participants electronically before the study. After completing the survey, the participants were given 150 RMB and short feedback for their psychological evaluation or reliable psychological support resources (e.g., 6–8 sessions of free grief counselling).

We used data from 319 participants (cases missing more than 20% of the data were excluded, n = 9). The sample comprised 86 (27%) men and 233 (73%) women who had experienced the loss of a significant other during the COVID‐19 pandemic. Their average age was 34.48 ± 11.91 (SD) years, and their ages ranged from 18 to 79 years. Additionally, 43.4% of the participants (n = 138) had lost their parents. The time since the loss ranged from 0 to 12 months, with an average of 4.00 ± 2.45 months. More details are presented in Table 1.

TABLE 1.

Demographic and loss‐related characteristics of the sample (N = 319)

Background and loss characteristics (N = 319)	N (%)/M (SD)
Demographic variables
Gender
Male	86 (27.0%)
Female	233 (73.0%)
Age	34.48 (11.91)
Highest education
High school/less than high school	39 (12.2%)
College/undergraduate degree	207 (64.9%)
Postgraduate degree	73 (22.9%)
Religious affiliation (yes)	44 (13.8%)
Loss characteristics
Time since loss (months)	4.00 (2.45)
Deceased person
Parent	138 (43.3%)
Partner	45 (14.1%)
Child	23 (7.2%)
Sibling	31 (9.7%)
Other	82 (25.7%)
Cause of death
Illness shorter than 1 month	95 (29.8%)
Illness longer than 1 month	123 (38.6%)
Violent cause (accident, suicide)	61 (19.1%)
COVID‐19	39 (12.2%)
Missing	1 (0.3%)
Grief rituals
None	47 (14.7%)
Individual rituals	22 (6.9%)
Collective rituals (informal)	106 (33.3%)
Collective rituals (formal)	144 (45.1%)
Opportunity to say goodbye (yes)	149 (46.7%)

2.2. Measures

2.2.1. Sociodemographic factors

Sociodemographic information included age, gender (male and female), educational background (high school or less than high school, college or undergraduate degree and postgraduate degree) and religious affiliation (yes or no). Loss‐related characteristics included relationship with the deceased (parents, spouses, children, siblings, others), time since the loss and cause of death (illness longer than a month, illness shorter than a month, violent causes and COVID‐19).

2.2.2. Grief severity

Grief severity was assessed using the 19‐item Chinese version of the Inventory of Complicated Grief (ICG‐19) (Li & Prigerson, 2016; Prigerson et al., 1995). Participants rated the frequency of symptoms on 5‐point scales ranging from 0 (never) to 4 (always). The scores for the items (e.g., “I feel I cannot accept his or her death”) were summed to form an overall grief severity score. The internal consistency of the questionnaire in the present sample was 0.94.

2.2.3. COVID‐19 related factors

Under the circumstances of death during the COVID‐19 pandemic, two questions were asked about traditional grief rituals: “Were you able to say goodbye when he or she died?” and “What kind of grief rituals were held to mourn the deceased?” The answer to the second question included “no grief rituals,” “personal rituals,” “collective rituals (informal)” or “collective rituals (formal),” and it was recoded to 0 = no grief rituals or personal rituals or 1 = collective grief rituals.

Loneliness was measured using the UCLA Loneliness Scale version 3 (Russell, 1996). The scale was translated and adapted in Chinese by Wang et al. (1999), and it comprises 20 items that evaluate the individuals' perceived sense of loneliness in the past week (e.g., “How often do you feel close to people?” and “How often do you feel that there is no one you can turn to?”). Participants rated how often they felt the way described in each statement on a 4‐point scale, ranging from 1 (never) to 4 (always). Nine items were reverse scored, and higher sum scores indicated higher levels of loneliness. In our sample, the Cronbach's α was 0.91.

2.2.4. Depressive and anxious avoidance

The Depressive and Anxious Avoidance in Prolonged Grief Questionnaire (DAAPGQ; Boelen & van den Bout, 2010) was specifically developed to measure two distinguishable avoidance processes following loss. Depressive avoidance, containing five items, measures behavioural avoidance of activities (e.g., “Since [–] is dead, I do much less of the things that I used to enjoy”). Anxious avoidance, assessed with four items, measures cognitive avoidance of loss‐related thoughts and events (e.g., I avoid dwelling on painful thoughts and memories connected to his or her death). Participants rated their agreement with each item on 8‐point scales ranging from 1 (not at all true for me) to 8 (completely true for me). The research team translated the original version into Chinese. The translated items were thoroughly discussed within the research team for accuracy and consistency with the original version. The internal consistencies of the two subscales were excellent (α = .96) and acceptable (α = .78). The two‐factor model with two distinct but correlated factors had an acceptable fit estimate (CFI = 0.979, TLI = 0.964, RMSEA = 0.086).

2.3. Statistical analysis

The missing proportion of all variables was less than 5%, and missing data were imputed using the expectation maximization (EM) algorithm (Enders, 2001) with 25 iterations, consistent with recommendations (Graham, 2009). Descriptive, correlation and regression analyses were conducted using SPSS version 22. First, the demographic variable “religious affiliation” was recoded as a yes‐no answer. The categorical variables, highest education levels, relationship with the deceased and cause of death, were transformed to dummy variables. Up to high school was set as a reference group, and two dummy variables were college/undergraduate degree and up to high school (“1” for college/undergraduate degree and “0” otherwise), and postgraduate degree and up to high school (“1” for postgraduate degree and “0” otherwise). Losing others was set as a reference group, so four dummy variables were parent and other (“1” for parent and “0” otherwise), spouse and other (“1” for spouse and “0” otherwise), child and other (“1” for child and “0” otherwise) and sibling and other (“1” for sibling and “0” otherwise). Bereaved by illness longer than a month was set as a reference group, and three dummy variables were illness shorter than a month and illness longer than a month (“1” for illness shorter than a month and “0” otherwise), violent causes and illness longer than a month (“1” for violent causes and “0” otherwise) and COVID‐19 and illness longer than a month (“1” for COVID‐19 and “0” otherwise).

Before the regression analyses, scatterplots, normal probability plots and residual plots were checked. The assumptions of linearity and homoscedasticity were met. Next, hierarchical regression analyses were performed with grief severity as the dependent variable, including demographic variables in the first step of the model; loss‐related variables, opportunity to say goodbye and grief rituals in the second step of the model and loneliness in the third step of the model. Finally, structural equation models were performed in Mplus version 8.3 (Muthén & Muthén, 1998‐2017). All the continuous variables (loneliness, depressive and anxious avoidance, and acute grief) were set as observable variables presented by the total score of the scales. We used the chi‐square, comparative fit index (CFI), Tucker Lewis index (TLI) (both ≥.90 indicating acceptable fit) and root mean square error of approximation (RMSEA) (≤.08 indicating acceptable fit) values to evaluate the model fit. Mediation analyses were conducted to test the mediating role of those two avoidance processes. Demographic and loss‐related variables significantly associated with grief severity were included in the mediation model as covariates. According to the results of regression analyses, COVID‐19‐related risk factors that were significantly associated with grief severity were set as independent variables, with depressive and anxious avoidance entered simultaneously as proposed mediators and grief severity as the dependent variable. Indirect effects reflected the effect of COVID‐19‐related factors on acute grief via depressive avoidance and anxious avoidance (path a × path b). The total effects were the effects of each independent variable on the dependent variable and the direct effects were that remained when controlling for the effects of the mediators (path c′). The bootstrapping approach with 2,000 resamples of Preacher and Hayes (2008) was employed to test the significance of the direct effects and the indirect effects using a 95% confidential interval (CI). Confidence intervals that do not contain zero were considered significant. In addition, we compared the indirect effects of these significant variables on acute grief by model constraint approach. The product of mediation path coefficients was subtracted to obtain a new indicator, diff. The differences in specific mediating effects were compared using 95% CI for the value of the new indicator. The effect sizes of the total indirect effect and each specific indirect effect were calculated by the proportion mediated, ab/(ab + c′) (MacKinnon et al., 2007).

3. RESULTS

3.1. Preliminary analyses

Higher grief severity was reported by people with religious beliefs (t = −2.28; p = .02), those who lost a partner or a child (F[4, 314] = 12.15; p < .001), or those bereaved due to violent death or COVID‐19 (F[3, 315] = 9.61; p < .001). Post hoc tests indicated that people bereaved due to COVID‐19 experienced higher levels of grief severity than those bereaved through illness longer than a month (p = .001), but no significant difference was found compared with those bereaved due to illness shorter than a month (p = .16) and violent causes (p = .38). Only a marginally significant difference was found in acute grief among those with different education levels (F[2, 316] = 3.00; p = 0.05). No significant difference was found in acute grief severity between the male and female participants (t[317] = −1.22, p = .22). No significant correlation was observed between age and acute grief (r = 0.04; p = .51), or time since loss and acute grief (r = −0.04; p = .46). Hence, religious affiliation, relationship with the deceased and cause of death were set as control variables in the mediation models with acute grief severity as the dependent variable.

3.2. Regression analyses of the effects of demographic, loss‐related, and COVID‐19‐related factors on acute grief severity

The results of hierarchical multiple regression analyses are presented in Table 2. The model was significant with acute grief as the dependent variable (F[16, 301] = 11.21; p < .001). In the first step, demographic variables, such as gender, age, educational level and religious affiliation, were added as predictive variables, explaining 3.8% of the variance. Loss‐related variables, such as time since loss, the relationship to the deceased, cause of death, opportunity to say goodbye and grief rituals, were entered in the second step, explaining another 19.7% of the variance. Loneliness, added in the third step, explained an additional 13.9% of the variance. The full model indicated that only loneliness was significantly correlated with higher levels of acute grief severity (β = .39, p = .38), while opportunity to say goodbye (β = −.04, p = .71) and grief rituals (β = −.02, p < .001) were not significantly associated with acute grief.

TABLE 2.

Hierarchical regression of risk factors related to COVID‐19 on complicated grief

	Acute grief symptoms
	ΔF (df)	ΔR2	β (step 1)	β (step 2)	β (step 3)
Step 1	2.47 (312)	.04 *
Gender			.07	.05	.01
Age			−.01	−.21 **	−.18 **
Education 1			−.12	−.05	−.05
Education 2			−.20 *	−.10	−.05
Religious affiliation			.12 *	.10	.06
Step 2	7.76 (302)	.20 ***
Parent (vs. other)				.36 ***	.30 ***
Partner (vs. other)				.40 ***	.34 ***
Child (vs. other)				.27 ***	.25 ***
Sibling (vs. other)				.06	.01
Time since loss				−.07	−.04
Illness shorter than a month (vs. illness longer than a month)				.15 *	.14 **
Violent cause (vs. illness longer than a month)				.21 **	.18 **
COVID‐19 (vs. illness longer than a month)				.14 *	.13 *
Grief rituals				−.03	−.04
Opportunity to say goodbye				−.06	−.02
Step3	66.57 (301)	.17 ***
Loneliness					.39 ***

Note: Kinship is dummy coded: parent (1) versus other (0), partner (1) versus other (0), child (1) versus other (0), sibling (1) versus other (0). Education is dummy coded: college/university (1) versus high school or less (1), graduate/postgraduate (1) versus high school or less (0). Cause of death is dummy coded: illness shorter than a month (1) versus illness longer than a month (0), violent causes (1) versus illness longer than a month (0), COVID‐19 (1) versus illness longer than a month (0).

^* p < .05.

^** p < .01.

^*** p < .001.

3.3. Mediation models with avoidance processes mediating the relationship between COVID‐19‐related factors and acute grief severity

According to the results of regression analysis, loneliness was included for further mediation analysis. The mediation model was tested if depressive and anxious avoidance were significant mediators between the relationship of loneliness and acute grief. Age, religious affiliation, relationship with the deceased and cause of death were regressed on acute grief as control variables according to the results of regression analysis, and the relationship to the deceased was also regressed on avoidance processes. The two mediators were set to be correlated with each other for their similarities. The model fit of the mediation model was good: χ² = 19.60; df = 10; p = .033; CFI = .980; TLI = .935; RMSEA = .055 (Figure 2).

FIGURE 2.

Mediation model. Note: ^** p < .01, ^*** p < .001

As Table 3 shows, the direct effect of loneliness on acute grief was significant (c′ = .17, p < .001) after controlling for the indirect effects of the mediators. Depressive avoidance (95% CI [.147, .273]) and anxious avoidance (95% CI [.012, .064]) were both significant mediators of the relationship between loneliness and grief severity, because the confidence intervals of the mediation effects did not include 0. Loneliness was positively associated with depressive avoidance (a ₁ = .44, p < .001) and anxious avoidance (a ₂ = .17, p = .002), and the two avoidance processes were positively associated with grief severity (b ₁ = .46, p < .001; b ₂ = .19, p < .001). Diff, the new indicator indicating the difference between the two mediating processes, was calculated by the formula a₁ × b₁ − a₂ × b₂. The 95% CI for the value of diff was [.17, .39]. The mediating effect of depressive avoidance was stronger than that of anxious avoidance. The total mediation effect size was .58. The mediation effect size of depressive avoidance was .49 and that of anxious avoidance was .09.

TABLE 3.

Summary of the mediation analysis in which the two avoidance processes mediate the relationship between loneliness and complicated grief, preparedness for death and complicated grief

Mediation path	Estimates	SE	p	95% CI
Loneliness‐acute grief	.23	.03	<.001	[.175, .306]
Loneliness‐DA‐acute grief	.20	.03	<.001	[.147, .273]
Loneliness‐AA‐acute grief	.03	.01	.011	[.012, .064]

Abbreviations: AA, anxious avoidance; DA, depressive avoidance.

4. DISCUSSION

Confronting grief might be challenging during the pandemic. The present study sought to explore the role of COVID‐19‐related factors in grief reactions. Consistent with our expectation, loneliness was positively associated with grief severity. Contrary to our expectation, this study did not find a significant difference in grief reactions between groups of people who had the opportunity to say goodbye to the deceased and those who did not. Similarly, no significant difference was found for groups of people who attended a funeral and those who did not. As we hypothesized, depressive and anxious avoidance partially explained the association between loneliness and acute grief.

In line with previous studies (Eisma et al., 2021; Gang et al., 2022), during the pandemic, people bereaved due to COVID‐19 reported more severe grief reactions than people bereaved due to illness longer than a month, but they did not report more severe grief than people bereaved due to illness shorter than a month and violent causes. People bereaved through COVID‐19 may have been unprepared for the death, and lower levels of preparedness for death were significantly related to higher levels of grief severity (Eisma et al., 2021; Nielsen et al., 2016).

This result was consistent with the assumption that loneliness was significantly associated with various psychiatric symptoms, including prolonged grief (Gorenko et al., 2020; Murata et al., 2020). People were restricted from social gathering and home visits during the outbreak of COVID‐19. Hence, the need for social interaction was difficult to meet, and increased levels of loneliness were seen as pandemic‐related challenges for the early bereaved (Selman et al., 2022). As suggested in micro‐sociological theory of adaptation to loss, bereaved people may need additional social interactions to fill the void in their social spaces that the deceased used to occupy. If their social needs cannot be satisfied, grief reactions may be more distressing and severe (Maciejewski et al., 2021). The positive relationship between loneliness and acute grief suggests that bereaved people feeling isolated and lonely during the quarantine may need more attention.

Unexpectedly, the results failed to demonstrate a predictive role of grief rituals, indicated by whether people had the opportunity to say goodbye to the deceased and attend the funeral. This outcome is contrary to that of Eisma and Tamminga (2022), who found that bereaved people who were unable to say goodbye experienced higher levels of grief during the pandemic. Some scholars thought that absence at the moment of death was common during the pandemic, it could be a significant predictor for worse bereavement outcomes (Wallace et al., 2020). However, a previous study providing evidence to support our results showed that only meaningful communication (the dying patient's ability to say goodbye) but not presence at the time of death itself was associated with grief reactions (Otani et al., 2017). The question in our study, “Were you able to say goodbye,” was misleading for differentiating whether meaningful communication occurred. Some participants have interpreted that question as asking if they were present when loved ones died. Although some of the bereaved may have the opportunity to stay with the deceased in the last moments before death, they may not have the opportunity to complete the farewells. Contradictory results were found in the literature about the effects of grief rituals on grief reactions. Simsek Arslan and Buldukoglu (2021) found that most of the participants stated that the COVID‐19 pandemic affected the grieving processes, but the implementation of grief rituals was not associated with grief reactions. Similarly, in this study, the healing effect of collective funeral rituals was not verified. This result may be explained by the fact that cultural and psychological connotations are complex and the direct influence cannot be seen in the short term (Mitima‐Verloop et al., 2019). The effect of funerals may depend on the sense of control over rituals, because simply attending a funeral without sharing one's feelings may not be helpful (Simsek Arslan & Buldukoglu, 2021). In addition, the average time since loss in the sample was 4.00 ± 2.45 months, with a maximum of 10 months. The bereaved may have been experiencing intense and acute grief with an insufficient time to heal and recover from loss (Prigerson et al., 2021). These findings may be somewhat limited without longitudinal follow‐up studies.

The results corroborate the findings of various studies in a cognitive‐behavioural conceptual model of complicated grief (e.g., Boelen et al., 2006; Boelen & Klugkist, 2011). One of the core mediating processes, avoidance processes, is a significant mediator between loneliness and grief severity. Stroebe and Schut (2010) suggested that the concepts of depressive and anxious avoidance were compatible with the Dual Process Model of coping with bereavement. Depressive avoidance refers to the avoidance of engagement in restoration‐oriented activities, and anxious avoidance refers to avoidance of confrontation with the reality of bereavement. Both processes were correlated with postbereavement stressors. Feelings of loneliness were both related to the departure of the deceased and the social isolation policy during the pandemic. When bereaved people find that their connectedness is becoming weaker and that the qualities of their relationships are below their expectations during the pandemic, with fewer supportive people encouraging them to engage in rewarding activities, they are more likely to withdraw (Boelen et al., 2006). When feelings of loneliness come from the absence of the significant attachment figure, the bereaved tend to avoid the memories, thoughts or reminders of loss events in order not to feel the pain of the loss (Ryckebosch‐Dayez et al., 2016). These avoidance behaviours may function as maladaptive coping strategies and lead to higher levels of grief severity (Eisma, de Lang, & Boelen, 2020). In addition, this study confirmed the postulation that depressive avoidance is critical to postbereavement adjustment (Boelen et al., 2015; Eisma, de Lang, & Boelen, 2020). For bereaved people who are more likely to feel lonely and isolated, encouraging them to engage in activities is more helpful to their adaptation compared to dealing with the avoidance of reminders of loss.

There are some important clinical implications from these findings. First, bereavement due to COVID‐19, illness shorter than 1 month and violent causes led to similar levels of grief reactions. People bereaved directly due to COVID‐19 did not experience higher levels of grief. The finding is compatible with the idea that preparedness for death may explain differences in grief reactions when comparing COVID‐19 and other causes of death (Eisma et al., 2021; Eisma & Tamminga, 2022). The contextual factors of death during the pandemic may be of more importance. Second, our findings suggest that bereaved people during the pandemic may need more companionship and attention to reduce the feelings of isolation. Bereaved individuals who have more social support structures and companions, and thus have their expectations met might be more inclined to become involved in productive activities. Although the study indicated that the presence of a farewell may not be associated with grief reactions, the importance of funerals and saying goodbyes cannot be ignored. The psychological meanings of grief rituals should be taken into account. Last, the mediating role of avoidance processes suggested that using targeted treatment procedures on avoidance behaviours would be useful to attenuate the effect on grief reactions brought by risk factors. Such targeted treatment procedures include, exposure (aiming to reduce avoidance of loss‐related memories, objects, or situations) and behavioural activation (aiming to increase the number of rewarding and valued activities) (Glickman et al., 2017; Lechner‐Meichsner et al., 2022). The reduction in grief severity was significantly associated with a reduction in avoidance processes (Boelen et al., 2011; Bryant et al., 2017). Therefore, targeting anxious and depressive avoidance behaviours may help lonely isolated bereaved people face the reality of bereavement and feel encouraged to recover from loss.

Although this study extends empirical studies on the characteristics of grief reactions during COVID‐19, some limitations are notable. First, a convenience sample was used, and most of the participants were female. Future studies with a representative group are required to confirm the results of this study. Second, there was no control group in this study. Therefore, how the potential risk factors worked before the pandemic remains unknown. Third, the cross‐sectional design calls for caution when interpreting the relationship between loneliness and grief severity. Inferences about causal relations and temporal precedence cannot be drawn. The bidirectional relationship between loneliness, avoidance processes and grief severity could be tested longitudinally in the future. For example, increased use of depressive avoidance may also lead to more loneliness. Last, two of the three proposed risk factors were measured using a single‐item binary question. Thus, the accuracy of the measurement needs improvement. More detailed questions are warranted concerning the evaluation of grief rituals and usage of funerals (Mitima‐Verloop et al., 2019). Moreover, the Chinese version of DAAPGQ has not been formally validated. In future studies, researchers can consider developing tools to measure individuals' perceptions of the impacts of the pandemic, such as the extent to which they were bothered by each risk factor (Neimeyer & Lee, 2021).

In conclusion, the present study indicates the effect of loneliness on acute grief reactions. Additionally, avoidance processes, especially depression avoidance, may further hamper the recovery of lonely bereaved individuals from the loss. More empirical studies of the impact of COVID‐19 circumstances on bereavement may be needed in the near future to provide finer‐grained information to understand bereavement during the pandemic.

CONFLICT OF INTEREST

No potential conflict of interest was reported by the authors.

ETHICS STATEMENT

The study was approved by the Ethics Committee of Beijing Normal University.

Zou, X. , Xu, X. , Tang, R. , Jiao, K. , Qian, W. , Shen, X. , & Wang, J. (2022). Potential risk factors for acute grief during the COVID‐19 pandemic: The mediating role of avoidance processes. Clinical Psychology & Psychotherapy, 1–11. 10.1002/cpp.2794

DATA AVAILABILITY STATEMENT

All data used in the current study are available from the corresponding author upon reasonable request.