Children and adolescents who experience the loss of a close person in traffic accidents (TAs), which account for approximately 1.19 million deaths annually worldwide (Foreman et al., 2018; World Health Organization [WHO], 2023), are confronted with a sudden and violent form of bereavement. Such traumatic losses may pose risks for children’s psychological adjustment; particularly when deaths are sudden and unexpected, childhood bereavement has been linked to long-lasting psychological consequences (Berg et al., 2016; Melhem & Brent, 2016; Melhem et al., 2008). One possible outcome is prolonged grief (PG), which involves separation distress and symptoms such as loneliness and numbness. PG disorder is recognized in the Diagnostic and Statistical Manual of Mental Disorders Fifth Edition, Text Revision (DSM-5-TR; American Psychiatric Association [APA], 2022), with symptoms causing significant impairment at least six months (12 months for adults) after the loss. Traumatic loss has also been associated with heightened risks of symptoms of depression and posttraumatic stress (PTS) in children (Boelen et al., 2023; Melhem et al., 2011).

From a cognitive-behavioral perspective, post-loss psychopathology following traumatic loss may be partly explained by a sense of unrealness about the loss, negative cognitions, catastrophic misinterpretations of grief reactions, and avoidance behaviors (Boelen et al., 2006). In bereaved adults, such processes have been shown to mediate associations between violent loss and symptoms of PG, depression, and PTS (Boelen et al., 2015). Despite the high incidence of fatal TAs and the possible associated psychological risks for bereaved children, research on their impact on children remains limited (Huang et al., 2016).

Parenting and Bereavement

Grieving often occurs within an interpersonal context of friends, culture, and family (Alvis et al., 2023; Kaplow et al., 2013). Within the family context, positive parenting, characterized by warmth, consistent discipline, and emotional support, are associated with better psychological outcomes for bereaved children (Haine et al., 2008; Kwok et al., 2005; Morris et al., 2016; Shapiro et al., 2014). Interventions boosting positive parenting have also been shown to reduce children’s psychological symptoms (Sandler et al., 2018).

Notably, unhelpful parenting may play a larger role than positive parenting in contexts where children are coping with potentially traumatic events (PTEs) (Afzal et al., 2023). Parental expressed emotion (EE) may be particularly important. EE in parenting reflects the emotional climate of parent-child interactions, with higher EE indicating more criticism, hostility, and emotional over-involvement towards the child (Fahrer et al., 2022; Peris & Miklowitz, 2015). Higher EE has been consistently linked to a range of mental health problems in children, including mood and anxiety disorders (Fahrer et al., 2022; Peris & Miklowitz, 2015). This may partly be because higher EE can act as a stressor and interact with children’s vulnerability and stress reactivity (Fahrer et al., 2022). In the context of traumatic loss, this may be particularly relevant, as children may already experience elevated stress and vulnerability due to a sudden and violent loss, while compounded trauma and loss reminders may further strain parents’ emotional responses (Layne et al., 2017). Research on the link between parental EE and children’s loss-related outcomes remains limited. Examining this link is important for both theoretical and clinical reasons, as parenting represents a potentially modifiable target for therapeutic intervention (Goodman et al., 2020).

Loss-Specific Parental Responses

Parental EE encompasses a broad pattern of parental attitudes and behaviors in parent-child communication in general. In contrast, loss-specific parental responses may include both supportive and unhelpful responses that are specific to parent-child communication about the loss. Research shows that warm, engaged, and sensitive loss-specific parental responses are associated with lower symptom-levels of PG and depression in children whose fathers recently died (Shapiro et al., 2014). In contrast, parental avoidance behaviors specific to the loss, such as the reluctance to mention the deceased, or experiencing discomfort when discussing the loss, have been linked to elevated PG, PTS, and depression in children (Alvis et al., 2020). However, less is known about whether parents’ appraisals of the loss and their active encouragement of approach or avoidance coping during loss-related conversations are associated with children’s psychological functioning.

In the child anxiety and PTS disorder literature, these processes are described as cognitive-behavioral parental responses and have been implicated in the development and maintenance of children’s psychological symptoms (e.g., Afzal et al., TS2023; Hiller et al., 2018; Percy et al., 2016). In a study by Hiller et al. (2018), parents discussed with their children an incident in which the child was reported to the emergency room after an accident. These authors found that parents’ descriptions of the emergency as highly threatening were linked to increased PTS in children, as was parental promotion of avoidance (e.g., saying “Yes, let’s stay away from trees” after the child expressed fear of climbing trees after being hospitalized due to a fall from a tree). Such avoidance behaviors and negative appraisals may hinder children from confronting trauma-related stimuli and addressing their negative appraisals, potentially transferring these parental response patterns to children (Afzal et al., 2023; Hiller et al., 2018; Williamson et al., 2017). These findings are consistent with cognitive-behavioral models of PTS disorder, which emphasize the role of threat appraisals and avoidance behaviors in the development and maintenance of symptoms (Ehlers & Clark, 2000). Given the conceptual overlap between the cognitive behavioral framework of PG disorder and PTS disorder models (Boelen et al., 2006), parental appraisals and promotion of coping behaviors as expressed during the interaction with the child may also account for part of the variability in children’s psychological functioning after loss. However, whether this is the case, remains understudied.

Researchers have increasingly studied parental responses in real-time using observational methods. Unlike self-reports, these methods help capture nuanced parent-child interactions that might otherwise go unnoticed (Marsac & Kassam-Adams, 2016). Observational approaches are also more reliable, because they reduce the impact of subjective and memory biases often present when reporting on parent-child interactions, as these interactions are typically brief and unconscious (Herbers et al., 2017; Morsbach & Prinz, 2006). The value of observational methods is especially relevant given the variability in findings on the link between parenting behaviors and loss- and trauma-related outcomes; these often differ depending on whether the data come from parent or child reports (Afzal et al., 2023; Boelen & Spuij, 2024). Observational tasks in which the parent and child discuss the loss have been used to study parental use of positive emotion words and effective communication quality (Shapiro et al., 2014; Wardecker et al., 2017). However, researchers have not yet used such tasks to examine loss-specific parental cognitive behavioral responses, such as how parents interpret the loss or promote coping strategies.

The Current Study

The current study aimed to address the gaps in the literature by examining, for the first time and in an exploratory manner, how parental EE and loss-specific parental cognitive behavioral responses relate to children’s psychological outcomes following traumatic loss within a dyadic mixed-methods framework. First, we focused on dimensions of parent-reported parental EE and their associations with children’s loss-related outcomes (i.e., PG, and related outcomes: PTS and depression, and health-related quality of life; HRQoL). Second, we aimed to adapt an existing coding scheme (Hiller et al., 2018) to categorize loss-specific parental responses based on previous research. These include parental cognitive-behavioral responses to the loss: the promotion of coping strategies (i.e., avoidance or approach) and expressions of appraisals (i.e., positive or negative). We then explored how these loss-specific parental responses relate to children’s loss-related outcomes.

We expected that high levels of dimensions of EE (i.e., lack of emotional support, more intrusiveness, more irritation, and more criticism), as well as parental promotion of avoidance and parental negative appraisals would be positively associated with children’s PG, depression, and PTS, and negatively associated with children’s HRQoL. Further, parental promotion of approach and positive appraisals were hypothesized to be negatively related to children’s PG, depression, and PTS in children, and positively to their HRQoL. Given the limited research in this area, we did not have specific expectations about the relative contributions of dimensions of EE and loss-specific responses in explaining differences in children’s outcomes.

Methods

Participants and Procedures

This study was conducted in the Netherlands and part of the Dutch ‘TrafVic Kids’-project focusing on the psychological consequences of, and care after, losing a close person in a traffic accident. Ethical approval was obtained from the Ethics Review Board of the Faculty of Social & Behavioural Sciences of Utrecht University (ID number: 21-0140). Participants were children aged 8 through 18 years, and their parents, who experienced the loss of a close person in a traffic accident at least six months prior. Recruitment methods encompassed various strategies, such as contacting participants from a similar project in bereaved adults, outreaching to social media, mental health organizations, clinicians, (peer) support organizations, and schools, snowball sampling, and an online questionnaire accessible via a website about grief.

One parent participated in the study alongside his or her child. Parents could participate in the study multiple times if more than one child participated. Active informed consent was obtained from children (of 12 years and older) and their parent(s) or caregiver(s) (when their child was 15 years or younger). Children participated in an online interview via Microsoft Teams, lasting up to 90 min, which included the measures described below. Parents completed online measures, also described below. In the last part of the interview, the parent-child dyads participated in a joint narrative task.

Measures

Expressed Emotion Assessed in Parents Using a Self-Report Measure

The parent version of the Level of Expressed Emotion questionnaire (LEE; Cole and Kazarian (1988) was used to assess parental EE. The LEE consists of 38 items. Parents were asked to rate how accurately each statement described their behavior toward their child over the past three months on a four-point scale from 1 = untrue to 4 = true. If parents participated with multiple children, they were asked whether their responses to the statements differed between children. If not, the same responses for the dimensions of EE were applied to all children. Certain items are recoded when scoring the questionnaire. Higher scores reflect a greater perceived level of the dimensions of EE: lack of emotional support (19 items, e.g., “I am understanding when my child makes a mistake”, after recoding), intrusiveness (7 items, e.g., “I need to know everything of my child), irritation (7 items, e.g., “I get irritated when things don’t go well), and criticism (5 items e.g., “I try to change my child”). We considered all four dimensions of EE in our analyses. Prior research has confirmed that the LEE demonstrates robust psychometric properties (Hale et al., 2011). In this study, Cronbach’s alpha for the four dimensions of the LEE was: 0.68 for lack of emotional support, 0.90 for intrusiveness, 0.80 for irritation, and 0.63 for criticism.

Loss-Specific Parental Responses Assessed Using a Joint Narrative Task

Parents and children engaged in a conversation about the death of a close person following a traffic accident. The same procedure was used as in Hiller et al. (2018), with minor adjustments to align with the context of losing a close person in a traffic accident. Dyads were instructed to begin the conversation by discussing the moments just before the event happened and to include any subsequent details they considered important. They were then provided with eleven prompts (e.g. “What were you thinking at the moment when you heard that ____ died?”). There was no time limit. On average the task lasted around 20 min. The researcher was not present during the task, but available through the chat function to send the prompts and answer questions. Prior to the task, the dyad was asked to each assess their current emotional state on two visual analogue scales of 0–10 (i.e., not distressed – very distressed, and relaxed – agitated/stressed). After the task, the researcher returned to the virtual meeting and asked both the parent and child to reassess their emotional state. This information was only used to monitor for any heightened emotional reactions and to provide support if needed, and not used for analyses. See https://osf.io/bszdk for the full procedure.

Prolonged Grief in Children

PG in children was assessed using the TGI-K-CA (Van Dijk et al., 2023). This 16-item interview screens for PG disorder according to both DSM-5-TR and ICD-11 criteria. The eleven items corresponding to the DSM-5-TR symptoms (without the D-criterion) were used in this study. For the two items measuring emotional pain, the highest score from either item was used to assess the symptom “Intense emotional pain”. Children rated the frequency of each symptom in the past month on 5-point Likert scales (1 = never, 5 = always). An example item is “In the past month, did you feel very alone because of the death of ____?”. The scores range from 10 to 50 (excluding the functional impairment item). Probable PG disorder classification can be indicated with a score of ≥ 3 (consistent with earlier research, e.g., Heeke et al., 2023; Lenferink et al., 2017) on at least one item representing the B-criterion, at least three of eight items representing the eight symptoms of the C-criterion, and the item representing the D-criterion. Cronbach’s alpha for the DSM-5-TR items was 0.92.

Depression in Children

The Patient Health Questionnaire-9 Modified for Teens (PHQ-9 M; Aggarwal et al., 2017) was used to assess depression in children. The measure includes nine items that each represent a DSM-IV symptom for major depressive disorder (MDD). The PHQ-9 M was translated from English to Dutch by forward-backward translation by two independent translators (see: https://osf.io/5xuaf). Children rated the extent to which they experienced a symptom in the past week (for example “Did you have little interest or pleasure in doing things?”). Responses are made on a 4-point Likert scale, ranging from 0 = not at all, to 3 = almost every day. Scores range from 0 to 27. A cut-off score of 11 can be used to indicate probable MDD in youths (Richardson et al., 2010). The PHQ-9 M has adequate psychometric qualities (Nandakumar et al., 2019). In the current study, Cronbach’s alpha was 0.82.

Posttraumatic Stress in Children

Children’s PTS was measured using the Child PTSD Symptom Scale for DSM-5 – Self Report (CPSS-5-SR; Foa et al., 2018). This 27-item questionnaire includes 20 items assessing PTS disorder symptoms according to DSM-5, and seven items assessing functional impairment. In this study, only the PTS symptom items were utilized. The measure was translated from English to Dutch by forward-backward translation by two independent translators (see: https://osf.io/6p4sq). Children rated the frequency of a symptom over the past month on a 4-point Likert scale (0 = not at all, 4 = six or more times per week/severe). An example is “Having bad dreams or nightmares”. Scores range from 0 to 80. A score of 31 can be used as a cut-off score for probable PTSD classification (Foa et al., 2018). Research shows that the CPSS-5-SR has good psychometric qualities (Foa et al., 2018). In the current study, Cronbach’s alpha was 0.93.

Health-Related Quality of Life (HRQoL) in Children

HRQoL of children was assessed by the KIDSCREEN-27 (Ravens-Sieberer et al., 2007). This 27-item questionnaire assesses five domains of HRQoL: physical well-being, mental well-being, autonomy and relationship with parent(s), social support and friends, and school. Each item measures either the frequency of behaviors and feelings or the intensity of attitudes, with a recall period of the past week. Response categories ranged from 1 = poor to 5 = excellent, 1 = not at all to 5 = extremely, or from 1 = never to 5 = always, depending on the item. Scores range from 27 to 135. Higher scores indicate higher HRQoL, after four negatively formulated items are recoded. Research shows that the KIDSCREEN-27 is valid for children and adolescents of 8–18 years and that it has good test-retest reliability and internal consistency (Ravens-Sieberer et al., 2007). Cronbach’s alpha for the KIDSCREEN-27 in the current study was 0.92.

Demographic and Loss-Related Variables

Demographic variables (age, gender, educational level) were assessed, with parents reporting their own information and children reporting theirs. Loss-related variables (e.g., time since the loss, nature of TA) were reported by the parent.

Data-Analysis

Coding for Loss-Related Parental Responses

The conversations in the joint narrative task were videotaped, transcribed, and parental utterances were divided into codable units, referred to as ‘chunks’, following the coding manual of Hiller et al. (2018). Each chunk represented a clause or single thought. We also drew on the coding scheme developed by Hiller et al. (2018), which is based on well-established work in the child anxiety field. It aims to categorize parental responses to children’s frightening experiences, such as those resulting in emergency room visits.

In the coding scheme, loss-related parental responses fall into four overarching categories: (1) Avoidance, which includes statements encouraging cognitive or behavioral avoidance, such as “We don’t have to talk about it”; (2) Approach, which includes statements promoting cognitive or behavioral approach, such as “Talking about it will help”; (3) Negative appraisals, which include statements making negative evaluations of the loss or its consequences, such as blaming the child, emphasizing external threats, or perceiving permanent change, as in “I don’t think our family will ever be the same again.”; and (4) Positive appraisals, which encompass statements minimizing threat, avoiding blame or criticism, promoting autonomy, or endorsing temporary, positive, or no change, such as “You coped really well on your own”. We made several adjustments to the coding scheme to suit bereavement following a traffic accident. This was done by combining a deductive method (applying the predefined coding scheme) with an inductive method (modifying or developing codes based on the data) (Braun & Clarke, 2006). See https://osf.io/ruqcn for the coding manual and see the Supplementary Materials A for more detailed information on the development of the coding scheme.

Coding was conducted in ATLAS.ti 24 (ATLAS.ti Scientific Software Development GmbH, 2024) by the first author, with regular discussions with the last author to refine the coding scheme. Inter-rater reliability was calculated based on 25% of the transcripts, coded by the first author and the (trained) third author. Krippendorff’s α was used as the index for inter-rater reliability because, unlike indices such as Cohen’s κ, it considers the likelihood of raters agreeing by chance based on overall patterns in the data, rather than on individual rater preferences (Krippendorff, 2004). Values of ≥ 0.50 considered as acceptable, consistent with trauma memory research (Kozina et al., 2019). Discrepancies were resolved through discussion. The total number of parental chunks (excluding irrelevant chunks such as “uh”) was calculated for each transcript, which was used in the calculation of relative frequency of each of the coding categories.

Descriptive and Correlational Analyses

Descriptive statistics were calculated for all variables, including dimensions of EE, loss-specific parental responses, and children’s outcomes (symptoms of PTS, PG, and depression, and HRQoL). We performed Bayesian Kendall’s τ correlations using default priors in JASP 0.18.3 (JASP Team, 2024) to examine associations of children’s loss-related outcomes with dimensions of EE, as well as with the relative frequency of loss-specific parental responses. Bayesian analyses were conducted because they may be more accurate than frequentist statistics when there is a relatively small sample size (Natesan, 2015; Van De Schoot et al., 2017). Kendall’s τ is a non-parametric correlation method that does not require a normal distribution of data (El-Hashash & Hassan, 2022). Values of τ ≤ 0.20 were considered as weak, τ > 0.20 ≤ 0.34 as moderate, and > 0.34 as strong (Gilpin, 1993). The Bayes Factor (BF) was used to quantify the support for the alternative hypothesis (i.e., that there is a relationship between the two variables under consideration). For example, if BF = 2, then there is twice as much support for the alternative hypothesis than for the null hypothesis (i.e., that there is no relationship between these two variables).

Results

Sample Characteristics

Table 1 presents the sample characteristics. The sample included 20 parent-child dyads, with parents (87.5% mothers) having a median age of 46.00 years (SD = 5.57) and children (50% girls) a median age of 13.50 years (SD = 3.44). Three parents (15.0%) participated multiple times in the interaction task, each time with a different child. The majority of losses involved the death of a parent (55.0%) or sibling (25.0%) for children, and a partner (41.2%) or child (29.4%) for parents. The lost persons were most frequently traveling by car or bicycle. The median time since the loss was 42.56 months (SD = 52.29). About half of children had received psychological support related to the loss. None of the parents who participated with multiple children reported any differences in the responses to the LEE for each child. Three children (15.0%) met probable PG disorder classification, four (20.0%) probable PTS disorder, and seven (35.0%) probable MDD.

Associations Between Parental Expressed Emotion and Children’s Loss-Related Outcomes

Bayesian Kendall’s τ correlations, presented in Table 2, Sect. 1, indicated that parent-reported parental criticism showed strong associations with children’s depression (τ = 0.37, BF = 3.07) and PTS (τ = 0.37, BF = 2.92). This suggests that these associations were approximately 3 times more likely than no association. For all other dimensions of EE, including lack of emotional support, intrusiveness, and irritation, BF’s indicated weak or negligible evidence for associations with children’s outcomes, with BF’s below 1.3.

Associations Between Loss-Specific Parental Responses and Children’s Loss-Related Outcomes

An overview of the loss-specific parental responses, including definitions, examples, interrater reliabilities, and relative frequencies in the sample, is presented in Table 3. Across interactions, the number of chunks ranged from 133 to 1214 (Md = 322.5; per-dyad overview in Supplementary Materials B). Parents most often expressed positive appraisals associated with the loss/lost person during the entire interaction with their child, followed by negative appraisals, then promotion of avoidance, and then promotion of approach. Although the frequencies of avoidance and approach behaviors were low, these responses were not entirely absent, and therefore, correlational analyses were still conducted to examine potential relationships with children’s loss-related outcomes.

Bayesian Kendall’s τ correlations, shown in Table 2, Sect. 2, indicated that parents’ expression of relatively more positive appraisals associated with the loss/lost person was strongly associated with children reporting lower depression (τ = −0.36, BF = 2.79) and higher HRQoL (τ = 0.47, BF = 13.74). This means that these associations were around 3 and 14 times more likely than no association. There was no evidence for a relationship between the relative frequency of positive appraisals and children’s PTS (BF = 0.66) or PG (BF = 0.58). The relative frequencies of negative appraisals, avoidance, and approach showed no evidence of associations with any of the children’s outcomes, with all BFs being below 1.0.

Discussion

This study examined the role of parenting in children’s psychological outcomes following traumatic loss due to a TA. Specifically, it focused on (1) dimensions of parent-reported parental EE, and (2) loss-specific parental responses expressed during the interaction with the child (i.e., the promotion of coping strategies and appraisals surrounding the loss). Unlike prior research relying on self-report, this study used a combination of self-report and observational methods with the aim of capturing parenting more accurately and reliably. The role of parenting was examined in relation to children’s PG, depression, PTS, and HRQoL.

Parental Criticism and Its Association with Children’s Depression and PTS

Partly in line with the expectations, the findings show that solely parent-reported parental criticism, as a dimension of EE, was positively associated with both depression and PTS in children. This aligns with previous studies, which, although not focused on bereavement, have shown a strong association between EE (especially parental criticism) and children’s depression (Braet et al., 2013; Nelemans et al., 2014; Tompson et al., 2010). Criticism may inflate negative self-evaluations and self-blame that characterizes depression (Garnefski & Kraaij, 2018), likely through the internalization of negative self-related cognitions and impaired emotion regulation (Berla et al., 2021).

The overlap between depression and PTS (e.g., cognitive distortion and negative self-evaluations; Wang et al., 2024) may explain why parental criticism is related to both depression and PTS in children. Furthermore, speculatively, parental criticism contributes to PTS-related hyperarousal, given the link between parental criticism and neural regions of emotional arousal (Chaplin et al., 2021). Even though EE (particularly parental criticism) is connected to different types of child psychopathology (Peris & Miklowitz, 2015), this study suggests that PG and related HRQoL are less sensitive to parental criticism. Possibly, although children can form new attachment bonds with other people who fulfill the same attachment functions as the close person (Mısırlı & Karakuş, 2024), children with PG may remain so preoccupied with the lost person that they focus less on other relationships – including those with their parents. These are tentative explanations, which need to be tested in future research. A methodological explanation may be that the associations between parental criticism and children’s depression and PTS were stronger than for PG because symptom-levels for depression and PTS were slightly higher compared with PG in our sample. Regarding the link between parental criticism and children’s HRQoL, research suggests that factors directly related to the child – such as a chronic physical disease and poor social skills – may be more important than family and social network factors in predicting children’s HRQoL (Bastiaansen et al., 2020).

Absence of Associations between Parental Negative Appraisals, Avoidance, and Approach, and Children’s Loss-Related Outcomes

Contrary to the expectations, parents’ expression of negative appraisals during the joint narrative task was not associated with children’s loss-related outcomes. This suggests that these type of appraisals may not directly drive these outcomes. However, more deeply ingrained critical negative cognitions may be more impactful. For example, a surface-level appraisal such as “I cannot do that very well yet” may reflect a deeper, more critical cognition such as “People who cannot cope quickly after a loss are weak”. These underlying cognitions may not have been easily expressed in the joint narrative task.

Also contrary to the expectations, parental promotion of avoidance and approach was unrelated to children’s loss-related outcomes. This may be explained by the low frequency of these parental responses in the interactions, which likely limited the power of the correlational analyses. Consequently, the results should be interpreted cautiously. One possible explanation for the low frequencies of these responses is that the coding manual, which was adapted from a PTS- and anxiety-focused framework, may not adequately capture avoidance and approach behaviors relevant to PG. In PG, avoidance involves the avoidance of the reality of the loss, which can paradoxically be expressed through approach behaviors, such as repeatedly talking about the loss or the lost person (Eisma & Lenferink, 2023). These proximity-seeking behaviors are more common in PG than in PTS (Smith & Ehlers, 2021). Nevertheless, it can be challenging to measure such paradoxical approach behaviors using observational methods. Approach behaviors that aim to retain a connection with the deceased may also serve an adaptive function by helping people construct meaning following bereavement (Hewson et al., 2024). This makes it difficult to observe when these behaviors are maladaptive. However, since avoidance is also difficult to measure through self-report (Gamez et al., 2010), it is still valuable to refine the interaction task and coding manual through further qualitative research to better capture different (mal)adaptive coping behaviors and their differential impact on children’s loss-related outcomes.

Parental Positive Appraisals and Its Association with Children’s Depression and HRQoL

Partly in line with the expectations, parental expression of positive appraisals during the joint narrative task was negatively associated with depressive symptoms and positively with HRQoL. This aligns with prior research showing that parental use of positive emotion words during reminiscing about the deceased is related to lower depression, anxiety, and avoidant coping (Wardecker et al., 2017). Our findings suggest that parental positivity surrounding the loss may play a protective role in outcomes related to mood and well-being (i.e., depression and HRQoL), more so than in stress-related outcomes (i.e., PGD and PTSD). Specifically, since particularly depression is characterized by a dysregulation of positive affect (Werner-Seidler et al., 2013), parents may help children to upregulate their positive emotions by focusing on positive thoughts associated with the lost person (Fredrickson, 2004). Engaging in positive reappraisal of an event may lead to more positive affect and less depressive symptoms (Garnefski & Kraaij, 2018; Pavani et al., 2016).

Particularly in people with more PG, positive appraisals related to grief may evoke “bittersweet” feelings, where happy loss-related memories may also cause pain (Smith et al., 2020). Speculatively, this may explain why no link was found between positive appraisals and children’s PG. It also underscores the need for research to further explore how seemingly positive appraisals may have potentially both supportive and negative effects on children with PG.

Clinical and Practical Implications

Although the findings should be interpreted cautiously, they may have implications for professionals supporting bereaved children and their families. In particular, bereavement support may help parents reflect on harsh or critical communication and provide them strategies to express more positive loss-related appraisals. Such support could build on existing psychoeducational interventions targeting trauma-specific parent-child communication and parental coping strategies (Cobham et al., 2012; Haag et al., 2020), recognizing that parents may need support themselves in navigating conversations about the loss with their child. Educators and other supportive adults may also benefit from awareness of the family context of bereavement, as they may be well-positioned to notice when children or families are struggling and to facilitate referral to appropriate support.

Limitations and Future Directions

This study has several limitations. First, the correlational design precludes conclusions about causality. It is especially important to consider bidirectional relationships, as parental criticism is not only found to predict adolescent psychopathology, but also vice versa (Nelemans et al., 2014). Second, the study was conducted in a community sample and children’s symptom-levels were relatively low. Clinical populations may have more pronounced negative appraisals and coping strategies during the joint narrative task, because they often experience more family-based vulnerabilities (e.g., high conflict; Weymouth et al., 2016). Third, the small sample limits the generalizability of results. Moreover, the study was conducted in the Netherlands; findings may not generalize to parent-child interactions in other cultural contexts, where family hierarchies and norms surrounding parent-child communication about loss may differ. The BFs provided reasonable but not overwhelming evidence, and along with a low inter-rater reliability for parental approach promotion – likely due to the low frequency of these responses –this further highlights the limitations of the smaller sample size. In addition, two EE subscales (i.e., lack of emotional support and criticism) showed relatively low internal consistency. This may have likely reduced the associations of these scales with the other considered variables. Fourth, because the current study was exploratory in nature and the prior research on potential associations between these constructs is limited, we did not apply corrections for multiple testing or statistically account for potential non-independence in the data caused by family-level clustering when parents participated with more than one child. Such clustering applied to a small proportion of the sample (15%); in future studies with larger samples and more substantial nested data structures, multilevel analyses may be appropriate. Overall, findings should be considered with caution and regarded as pending replication in larger and more diverse samples.

Future research could also include factors such as age and gender differences, and time since loss, as these may influence the interaction (e.g., parents may express blame or criticism more often towards older children; Canale et al., 2022; Sloover et al., 2024). Parenting may also differ qualitatively for parents who lost their partner (i.e., the child’s other parent), as they often must continue parenting alone. This shift in responsibility may lead to distinct parenting behaviors compared to those of parents who lost another person. Moreover, the small size of the study did not allow for analysis of fine-grained aspects of the joint narrative interaction in relation to children’s PG and other loss-related outcomes. Future studies could explore these associations in more detail. For instance, speculatively, threat appraisals may be more strongly associated with PTS, and vulnerability appraisals (e.g., helplessness) more with PG, given the key characteristics of each (i.e., PTS is characterized by anxiety, and PG by separation distress; Jordan & Litz, 2014, Maercker & Znoj, 2010). Furthermore, future research can examine other aspects of the joint narrative interaction. For instance, parental practices regarding emotional socialization, such as parental egocentrism (i.e., focusing the conversation on parents’ own emotions and experiences) during parent-child communication about disasters, may negatively influence children’s well-being (Hendrickson et al., 2020). Finally, given the importance of observational methods that we emphasized in this study, using an observational measure for EE such as the Five Minute Speech Sample (FMSS) (Magaña et al., 1986) could minimize reporter bias. This is important, as prior work on parenting and child PTS shows that studies often rely on single-informant and self-report measures, resulting in a high risk of bias, while observational methods to assess parent-child interactions are rarely used (Afzal et al., 2023). The FMSS asks caregivers to discuss their thoughts and feelings about the child for five uninterrupted minutes. However, this measure is not without limitations, as one dimension of the measure yields inconclusive findings regarding validity (Rea et al., 2020). Relatedly, because measures were conducted online via videoconferencing, the setting may have influenced how parents and children interacted during the joint narrative task and may have made it more difficult to capture subtle interactional dynamics and nonverbal behavior (Khan & MacEachen, 2022). Nevertheless, this influence may have been limited, as the task required parents and children to engage primarily with one another rather than with the screen, and only verbal content was analyzed. Future observational studies could therefore examine whether similar patterns emerge in person, where facial expressions and other nonverbal aspects of parent-child interaction may also be more fully captured.

Conclusion

Despite being small in scale, this exploratory study is the first to provide preliminary insights into both parental EE and loss-specific parental responses in the context of traumatic loss. A key strength of the study is its use of an observational task combined with self-report surveys using a dyadic approach. The coding scheme for loss-specific parental responses offers a structured approach to measure parental appraisals and promotion of coping strategies. Without overrating findings, they may highlight aspects of parental support that can be linked to children’s loss-related outcomes. Specifically, parental criticism may act as a harmful factor for depression and PTS. Moreover, loss-specific parental responses might also play an important role. Particularly parental positive appraisals may be protective against depressive symptoms and improve HRQoL. In conclusion, this study gives valuable directions to future research on specific parenting dimensions and children’s loss-related outcomes.