Gender differences in prolonged grief: a network perspective on symptom interconnections and functional impairment

ABSTRACT

Background: While prior studies have reported gender differences in prolonged grief symptom severity, less is known about symptom interconnections and their associations with functional impairment. Clarifying these may inform more targeted bereavement care.

Methods: A total of 1,093 bereaved adults (274 males; mean age = 37.03) in Hong Kong completed Prolonged Grief Disorder-13-Revised. Network analysis was used to examine gender differences in edge strength, centrality (expected influence; EI), global strength, and network structure. Functional impairment was included to explore its associations with grief symptoms.

Results: Females reported higher overall symptom severity than males. Meaninglessness and Disengagement showed the strongest EI in both networks. Network structure differed significantly by gender (M = 0.35, p = .018), whereas global strength did not (S = 0.75, p = .114). Identity Disruption showed higher EI in males. In exploratory analyses including functional impairment, symptom-function associations also differed by gender.

Conclusions: Findings indicate gender-related differences in the structure and functional correlates of prolonged grief symptoms. Meaninglessness and Disengagement were central symptoms across genders, suggesting that meaning-making and re-engagement may be important intervention targets. Identity Disruption appeared more central in males, and differences in symptom-function associations may help inform gender-sensitive assessment and clinical support for bereaved individuals.

HIGHLIGHTS

• Prolonged Grief Disorder-13-Revised symptom networks were compared by gender in 1,093 bereaved adults.

• Females reported higher symptom severity, but global strength was similar across gender groups.

• Symptom network structure and functional impairment links differed by gender, and meaninglessness and disengagement were the most central symptoms.

1. Introduction

Prolonged grief disorder (PGD) has been formally recognized in contemporary diagnostic guidelines and is increasingly viewed as an important mental health condition with substantial clinical and public health relevance (American Psychiatric Association [APA], 2022; Killikelly et al., 2025). PGD is characterized by persistent separation distress and related emotional, cognitive, and behavioural symptoms, together with functional impairment beyond expected social, cultural, or religious norms (APA, 2022). As research on PGD has expanded, increasing attention has been given to its symptom manifestations, assessments, treatments, and relevant individual difference factors. Among these, gender has been identified as a relevant risk factor for adverse bereavement outcomes (Stroebe et al., 2006), and many studies report higher grief symptom severity in females than males (e.g. Buur et al., 2024; Heeke et al., 2019). However, gender differences in symptom network structure, the centrality of specific symptoms, and symptom-function relationships remain underexplored.

A network perspective may help address this question by conceptualizing grief symptoms as a system of mutually connected experiences rather than as indicators of a latent condition (e.g. Djelantik et al., 2020; Jiao et al., 2025; Maccallum et al., 2017). Within this framework, edges represent associations between symptoms, and centrality helps identify symptoms that are more strongly connected within the network and may therefore be more influential in relation to other symptoms. Such an approach may complement mean-level comparisons by clarifying whether gender differences are reflected not only in symptom severity but also in symptom connectivity and centrality. This information could be useful for identifying clinically salient symptoms and patterns of co-occurrence.

To date, only a small number of studies have investigated gender differences in grief networks. Maccallum et al. (2023), using Prolonged Grief Questionnaire-13, reported substantial similarities between male and female prolonged grief symptom networks. Similarly, Caycho-Rodríguez et al. (2025) used the five-item Grief Impairment Scale and found that the networks were invariant across gender. Together, these studies suggest broad similarity rather than marked gender differences in grief-related networks. However, the evidence base remains limited. Grief network studies more broadly have shown that central symptoms and strongest edges may vary across samples and contexts (Djelantik et al., 2020; Jiao et al., 2025; Maccallum et al., 2017; Malgaroli et al., 2018; Pan, 2020; Xiong et al., 2026). Therefore, it remains unclear whether the gender invariance reported by Maccallum et al. (2023) and Caycho-Rodríguez et al. (2025) generalizes across different symptom sets, assessment tools, and cultural contexts, particularly in non-Western samples, which have rarely been examined.

Functional impairment, despite being a core diagnostic feature of PGD, has often been omitted from grief network studies because it is commonly assessed using a binary item rather than the ordinal response format used for symptom items (Prigerson et al., 2021). However, it remains important to investigate how specific symptoms relate to functional impairment, as this may clarify which symptoms are most closely associated with difficulties in daily functioning and whether such patterns differ by gender. Similar approaches have been used in recent network studies outside the grief literature, where binary clinical indicators (e.g. diagnosis of depression) have been integrated into symptom networks to explore their associations with individual symptoms (e.g. Alcalde et al., 2024). In addition, prior gender-focused grief network studies have not examined symptom networks using PG-13-R items aligned with DSM-5-TR PGD criteria, and most available evidence has come from Western samples.

Therefore, the current study aimed to examine gender differences in prolonged grief symptoms assessed with PG-13-R in a bereaved Hong Kong Chinese sample. Using network analysis, we compared symptom connectivity, expected influence, global strength, and overall network structure between males and females. We also explored how grief symptoms were associated with functional impairment and whether these associations differed by gender.

2. Method

2.1. Participants

This study is part of a larger project focused on the validation of bereavement outcome and risk assessment tools in Hong Kong (Jockey Club Bereavement Care and Support Alliance, 2025). Participants were recruited through online advertisements and bereavement-related community and service channels, including workshops, activities, and collaborating health and social care agencies. Participants were eligible if they were adults (18+ years) who had experienced the loss of a significant person between three months and three years before enrollment. This timeframe was selected to exclude individuals in the acute phase of grief (under three months) and those potentially affected by retrospective bias (beyond three years), while capturing variation across bereavement stages. Data collection was between May 2024 and November 2024. The final sample size was 1093. Detailed demographic information can be found in Table 1.

Table 1.

Demographic characteristics of the bereaved and the deceased (N = 1093).

	Total	Female	Male
	(N = 1093)	(N = 819)	(N = 274)
	M ± SD/N(%)
The Bereaved
Age	37.03 ± 13.64	37.50 ± 13.68	35.61 ± 13.44
Marriage status
Single	585 (53.52)	440 (53.72)	145 (52.92)
Married/cohabiting	399 (36.51)	285 (34.80)	114 (41.61)
Divorced/separated	39 (3.57)	33 (4.03)	6 (2.19)
Widowed	70 (6.40)	61(7.45)	9 (3.28)
The Deceased
Age at death	74.35 ± 17.82	74.21 ± 18.03	74.77 ± 17.21
Gender
Female	516 (47.21)	381 (46.52)	135 (49.27)
Male	577 (52.79)	438 (53.48)	139 (50.73)
Cause of death
Acute Illness	385 (35.22)	301 (36.75)	84 (30.66)
Chronic Illness	558 (51.05)	403 (49.21)	155 (56.57)
External causes	104 (9.52)	80 (9.77)	24 (8.76)
Others	46 (4.21)	35 (4.27)	11 (4.01)
Relationship with the bereaved
Spouse/Partner	86 (7.87)	74 (9.04)	12 (4.38)
Parent	375 (34.31)	276 (33.70)	99 (36.13)
Grandparent	420 (38.43)	309 (37.73)	111 (40.51)
Others	212 (19.40)	160 (19.54)	52 (18.98)
Time since loss
3 months–year	415 (37.97)	320 (39.07)	95 (34.67)
1–2year	376 (34.40)	280 (34.19)	96 (35.04)
2–3 year	302 (27.63)	219 (26.74)	83 (30.29)

Notes: Gender was assessed by self-report using a binary female/male item (‘My gender’). Sex assigned at birth and gender identity were not separately assessed.

2.2. Measures

The Prolonged Grief Disorder-13-Revised (PG-13-R) was used to assess grief symptoms (Prigerson et al., 2021). The translated Chinese version was used. The questionnaire included binary (yes/no) questions related to the occurrence of loss, time since loss, and functional impairment, and ten symptom items aligned with DSM-5-TR criteria (e.g. ‘Do you feel yourself longing or yearning for the person who died?’), rated on a 5-point Likert scale (1 = not at all, 5 = overwhelmingly). All ten items and the functional impairment item were included in the network analysis. Accordingly, the present study focused on prolonged grief symptoms assessed with the PG-13-R across 3 months to 3 years post-loss rather than on PGD diagnosis. The Chinese version PG-13-R showed good reliability in the study (Cronbach’s Alpha = .95).

2.3. Data analysis plan

2.3.1. Network estimation and centrality analysis

Regularized partial correlation networks were estimated using the Graphical Least Absolute Shrinkage and Selection Operator (LASSO) with Extended Bayesian Information Criterion (EBIC) model selection (Epskamp et al., 2018; Foygel & Drton, 2010; Friedman et al., 2008). Edges represent partial correlations between symptoms, with stronger weights indicating stronger associations. To enhance interpretability and sparsity, a tuning parameter (γ = 0.5) was applied, and small edges were pruned (Epskamp et al., 2018).

Expected influence (EI) was selected as the centrality index, as it better captures a node’s actual influence when negative associations are present (Robinaugh et al., 2016). Higher EI indicates stronger overall connectedness within the network. Edge-weight accuracy was evaluated using non-parametric bootstrapping, and the stability of EI was assessed using case-dropping bootstrapping. Correlation stability (CS) coefficients above 0.25 were considered acceptable, and above 0.5 as ideal (Epskamp et al., 2018). Comparative significance testing for both edge weights and EI were performed using bootstrapped difference tests.

2.3.2. Network comparison

The NetworkComparisonTest (NCT) package (van Borkulo et al., 2023) was used to compare male and female networks in terms of global strength (the sum of absolute edge weights) and overall network structure. When significant differences in network structure were detected, Edge Invariance Tests were conducted to identify specific edges contributing to the differences. Centrality Invariance Tests were also conducted to examine whether EI differed significantly between male and female networks at the node level.

To examine associations with functional impairment, the Functional Impairment item of the PG-13-R was included as a binary node, and the male and female networks were re-estimated. In this step, the analyses focused on network comparison. For simplification purposes, only comparisons related to Functional Impairment were reported. We acknowledge that including a binary node in EBICglasso estimation introduces scale heterogeneity, which may affect the precision of edge estimation. This step is exploratory in nature, and results involving this node should be interpreted with caution. Similar approaches have been used in previous network studies incorporating binary clinical indicators (e.g. Alcalde et al., 2024).

All analyses of the networks were conducted with RStudio Version 2024.12.1 with R Packages bootnet (Epskamp et al., 2018), qgraph (Epskamp et al., 2012), and networktools (Jones, 2025).

3. Results

3.1. Preliminary analyses

Preliminary comparisons indicated that males and females did not differ significantly in key loss-related characteristics, including cause of death, relationship with the deceased, deceased age, deceased gender, and time since loss (all p values > .05), suggesting broad comparability between gender groups on bereavement-related background characteristics.

Bereaved females experienced a significantly higher level of grief symptoms than bereaved males (t = 2.45, p = .015). Item-level analyses revealed that bereaved females reported statistically significantly higher scores on several grief symptoms than males, including Yearning, Disbelief, Avoidance, Emotional Pain, and Loneliness (all p values < .05; see Table 2). However, males and females did not statistically differ in whether their symptoms were associated with functional impairment (χ²(1) = 0.06, p = .803).

Table 2.

Descriptives and gender comparisons on PG-13-R items and total score.

	Total sample M ± SD	Female M ± SD	Male M ± SD	p	Effect size
[Yearning] Longing or yearning for the person who died	2.95 ± 1.09	3.02 ± 1.10	2.76 ± 1.01	.001	.13
[Preoccupation] Have trouble doing the things you normally do because you are thinking so much about the person who died	2.00 ± 1.08	2.02 ± 1.10	1.95 ± 1.02	.527	.02
[Identity Disruption] Feel confused about your role in life or feel like you don’t know who you are anymore	1.91 ± 1.08	1.93 ± 1.10	1.87 ± 0.99	.788	.01
[Disbelief] Have trouble believing that the person who died is really gone	2.41 ± 1.22	2.47 ± 1.26	2.23 ± 1.07	.015	.10
[Avoidance] Avoid reminders that the person who died is really gone	2.36 ± 1.21	2.43 ± 1.24	2.16 ± 1.08	.003	.12
[Emotional Pain] Feel emotional pain (e.g. anger, bitterness, sorrow) related to the death	2.62 ± 1.16	2.68 ± 1.19	2.46 ± 1.07	.018	.09
[Disengagement] Feel that you have trouble re-engaging in life (e.g. problems engaging with friends, pursuing interests, planning for the future)	1.96 ± 1.11	1.99 ± 1.13	1.86 ± 1.04	.112	.06
[Numbness] Feel numb or emotionally detached from others	1.96 ± 1.10	1.96 ± 1.11	1.93 ± 1.07	.812	.01
[Meaninglessness] Feel that life is meaningless without the person who died	1.92 ± 1.17	1.92 ± 1.19	1.91 ± 1.10	.690	−.02
[Loneliness] Feel alone or lonely without the deceased	2.32 ± 1.24	2.38 ± 1.27	2.17 ± 1.13	.041	.08
Total score	22.44 ± 9.43	22.82 ± 9.68	21.30 ± 8.57	.015	.16

Notes: For PG-13-R items, p values indicate gender differences assessed using Mann-Whitney U tests for each item because of the ordinal nature of the data, and effect sizes are reported as rank-biserial correlations. For total score, the p value indicates gender differences using independent-samples t test, and the effect size is reported as Hedges’ g.

3.2. Network estimation

Network analysis for the entire sample can be found in Appendix (Figure A1–A4). Networks for gender subgroups are shown in Figure 1a and 1b. For the male network, 31 of 45 edges are present (68.89%), while for the female network 30 of 45 edges are present (66.67%). These proportions reflect the regularized estimation approach, in which very small edges are shrunk to zero.

Figure 1.

Estimated Symptom Networks of Grief Symptoms for the Gender Subgroups. Edge thickness indicates the strength of association between symptoms, with green edges representing positive associations and red edges representing negative associations. To ensure comparability, the same edge weight scaling was applied across the male and female networks by fixing the minimum and maximum edge widths to the global values observed across both networks. To facilitate visual comparison, both male and female networks were plotted using the same layout, derived from the male network. The size of the outer pie segments reflects each node’s Expected Influence, normalized to a 0–1 scale. Yearn = Yearning, Preoc = Preoccupation, Ident = Identity disruption, Disbe = Disbelief, Avoid = Avoidance, Pain = Emotional Pain, Disen = Disengagement, Numb = Numbness, Mean = Meaninglessness, Lone = Loneliness.

Bootstrap analyses of edge weights are presented in Appendix Figures A5–A8. In the male network, the strongest connections involved Yearning to Loneliness, Identity Disruption–Disengagement, Yearning–Emotional Pain. In the female network, the strongest connections involved Meaninglessness to Loneliness, Disengagement–Numbness, and Preoccupation–Disengagement.

The stability of EI was acceptable in both networks (CS-coefficients = 0.44 for males and 0.75 for females). Meaninglessness showed the highest EI in both networks (male: EI = 1.20; female: EI = 1.30). Detailed EI results are presented in Appendix Figures A9 and A10.

3.3. Network comparison

Network comparison showed no significant gender difference in global strength (female = 5.79, male = 5.04; test statistic S = 0.75, p = .114), but a significant difference in overall network structure (test statistic M = 0.35, p = .018), indicating that the specific pattern of connections between grief symptoms differed significantly between male and female networks.

The Edge Invariance Test identified nine edges with significant gender differences in their edge weights, as presented in Table 3. The male network demonstrated stronger edge weights in the Disengagement–Disbelief, Disengagement–Avoidance, Identity Disruption–Loneliness, and Avoidance–Yearning edges. Notably, the Disengagement–Yearning edge was negative in both networks but was substantially stronger in the male network. In contrast, the female network showed stronger edge weights for the Disengagement–Preoccupation and Emotional Pain–Avoidance connections.

Table 3.

Edge invariance test.

Symptom 1	Symptom 2	Male	Female	E	p
Female > Male
Disengagement	Preoccupation	0	0.35	0.35	.005
Emotional Pain	Avoidance	0	0.24	0.24	.010
Male > Female
Meaninglessness	Emotional Pain	0.20	0	0.20	.017
Meaninglessness	Preoccupation	0.21	0	0.21	.040
Disengagement	Disbelief	0.16	−0.08	0.23	.009
Disengagement	Yearning	−0.30	−0.07	0.23	.012
Disengagement	Avoidance	0.19	0	0.19	.012
Identity Disruption	Loneliness	0.20	0	0.20	.025
Avoidance	Yearning	0.16	0	0.16	.016

The Centrality Invariance Test identified a significant gender difference in the EI of Identity Disruption (male: EI = 1.04; female: EI = 0.78; C = −0.177, p = 0.027).

3.4. Network comparison including functional impairment as an external node

In the exploratory analysis including Functional Impairment as an external node, symptom-function associations differed by gender. Networks displaying all edges are presented in Appendix Figure A11. Figure 2 highlights only the edges connecting grief symptoms with Functional Impairment. In the male network, Meaninglessness and Disengagement were most strongly connected to Functional Impairment, whereas in the female network, Disengagement and Preoccupation showed the strongest associations.

Figure 2.

Estimated Symptom Networks of Grief Symptoms Including Functional Impairment for the Gender Subgroups. Edge thickness indicates the strength of association between symptoms, with green edges representing positive associations and red edges representing negative associations. To ensure comparability, the same edge weight scaling was applied across the male and female networks by fixing the minimum and maximum edge widths to the global values observed across both networks. To facilitate visual comparison, both male and female networks displaying only the edges connected to the Functional Impairment node were plotted using the same layout, derived from the male network. The size of the outer pie segments reflects each node’s Expected Influence, normalized to a 0–1 scale. Yearn = Yearning, Preoc = Preoccupation, Ident = Identity disruption, Disbe = Disbelief, Avoid = Avoidance, Pain = Emotional Pain, Disen = Disengagement, Numb = Numbness, Mean = Meaninglessness, Lone = Loneliness, Func = Functional Impairment.

Results of the Edge Invariance Test are shown in Appendix Table A1. The connection between Functional Impairment and Meaninglessness was significantly stronger in the male network (edge = 0.36) than in the female network (edge = 0), with an edge difference of 0.36, p = .009. Furthermore, the connection between Functional Impairment and Disbelief differed in direction across gender: it was negative in the male network (edge = –0.12) but positive in the female network (edge = 0.10), and the difference was statistically significant (E = 0.22, p = .035).

4. Discussion

The current study examined gender differences in prolonged grief symptoms assessed with the PG-13-R in a bereaved Hong Kong sample using network analysis. Overall, the findings suggested gender differences in symptom severity, symptom network structure, and symptom-function associations. First, Meaninglessness and Disengagement were the most central symptoms in both gender-specific networks. Second, the male and female symptom networks differed in structure despite no significant difference in global strength, and Identity Disruption was more central in males. Third, in the exploratory analysis including Functional Impairment, the symptoms most strongly associated with impairment differed by gender. In addition to these network findings, bereaved females reported higher overall grief symptom severity than males, particularly in Yearning, Disbelief, Avoidance, Emotional Pain, and Loneliness. However, the effect sizes for the gender differences was generally small. This pattern is consistent with prior literature reporting higher prolonged grief symptom levels among women (Buur et al., 2024; Heeke et al., 2019).

4.1. Central symptoms in prolonged grief networks

The central positions of Meaninglessness and Disengagement suggest that these symptoms were most strongly connected within the broader grief symptom network across genders. The centrality of Meaninglessness is consistent with prior network studies of bereaved individuals following spousal or parental loss (Maccallum et al., 2017, 2023; Malgaroli et al., 2018). In contrast, Disengagement has not been identified as a central symptom in these or other grief network studies (Jiao et al., 2025; Pan, 2020; Xiong et al., 2026). The current sample consisted predominantly of young and middle-aged adults (92.41% aged 18–60; M = 37.03), most of whom (72.74%) had experienced the loss of a parent or grandparent. These sample characteristics, together with the Hong Kong Chinese cultural context, may help account for why both symptoms were especially prominent.

For Meaninglessness, the loss of a parent or grandparent may carry particular weight in a Hong Kong Chinese context in which family continuity and intergenerational obligation are highly valued. In this context, the death of an elder may challenge not only emotional attachment, but also a sense of meaning partly embedded in family roles and responsibilities. This interpretation is broadly consistent with earlier work suggesting that meaning making is a salient aspect of Chinese bereavement experience (Chan et al., 2005), while also fitting meaning-reconstruction perspectives on grief more generally (Neimeyer, 2001, 2019). For Disengagement, the predominantly young and middle-aged composition of the sample may be especially relevant, as difficulty re-engaging in everyday life may be particularly distressing when bereaved individuals must continue to manage work, family, and other daily responsibilities. Consistent with this interpretation, bereaved adult children in Hong Kong have been found to report struggles with changed family roles and caregiving responsibilities following a parent’s death (Chan et al., 2025).

Taken together, these findings support the clinical relevance of attending to both existential distress and difficulties re-engaging in everyday life after bereavement. The centrality of Meaninglessness may call for approaches that explicitly support the reconstruction of meaning following loss (Neimeyer, 2001, 2019), whereas the centrality of Disengagement points to approaches that directly target re-engagement in daily activities, like behavioural and cognitive–behavioural approaches (Hao et al., 2024).

4.2. Gender differences in symptom network structure

Unlike Maccallum et al.’s (2023) study, which identified largely similar grief symptom structures across genders, the current findings indicated a significant difference in network structure. Prior work has shown that grief-related emotional display and symptom expression may vary across cultural settings (Xiu et al., 2016; Zhou et al., 2023), suggesting that gendered grief patterns may not generalize uniformly across samples.

In the male network, the stronger negative association between Disengagement and Yearning, together with the stronger connections between Disengagement, Disbelief and Avoidance, may suggest that yearning occupies a different position within the overall symptom structure. This interpretation may be broadly consistent with the concept of instrumental coping, whereby grief is managed more through activity-based strategies than through emotional expression (Doka & Martin, 2010), and with qualitative findings among suicide-bereaved Chinese men in Hong Kong highlighting hidden grief and emotional control, alongside continued management of day-to-day responsibilities following loss (Chan & Cheung, 2022). In contrast, females showed stronger connections between Disengagement and Preoccupation, and between Emotional Pain and Avoidance. These stronger positive associations may indicate tighter co-activation among selected cognitive-emotional symptoms in females and may be broadly consistent with more intuitive forms of grieving, which place greater emphasis on emotional experience and relational processing (Doka & Martin, 2010).

Identity Disruption showed higher expected influence for males than females, suggesting that identity-related disturbance may occupy a more central position in the male grief network. As mentioned above, the connection between Identity Disruption and Loneliness was also stronger in the male network. Although females in our sample reported higher levels of loneliness overall, these findings suggest that, for bereaved men, loneliness may be more closely tied to disruptions in identity than to social isolation alone. In Hong Kong’s cultural context, where traditional male roles emphasize responsibility and strength (Leung & Chan, 2014), loss of a significant other may challenge core identity anchors. This pattern may have implications for assessment and intervention, including attention to identity reconstruction and role redefinition in bereaved men.

4.3. Gender differences in symptom-function associations

In the exploratory analysis including Functional Impairment as an external node, the symptom-function pattern extended the gender-specific findings observed in the primary symptom network. Disengagement was closely related to Functional Impairment for both genders, which is understandable given the role of re-engagement difficulties in daily functioning after bereavement. Meanwhile, the additional symptom most strongly tied to impairment differed by gender: Meaninglessness in males and Preoccupation in females. This suggests that functional impairment may be linked to somewhat different grief processes across gender: in males, impairment may be more closely tied to meaning-related distress and difficulties re-engaging in life, whereas in females it may be more closely tied to sustained cognitive-emotional involvement with the loss. In practice, these findings suggest that, for males, improving functioning may require particular attention to meaning reconstruction and existential integration (Neimeyer, 2001, 2019), whereas for females, interventions that help shift attention away from ruminative involvement with the loss and toward present-moment awareness, such as mindfulness-based practices, may be particularly useful (Cacciatore et al., 2014; Eisma et al., 2023).

Interestingly, Disbelief was negatively associated with Functional Impairment in males but positively associated in females. A similar gender divergence was observed in the primary symptom network, where the Disengagement–Disbelief edge was positive in males but negative in females. Although Disengagement was closely related to Functional Impairment in the exploratory network, these differing patterns suggest that difficulty re-engaging in life and functional impairment, while related, are not interchangeable and may connect to disbelief in partly different ways across gender. Accordingly, although recognizing the reality of the loss is often considered an important part of adaptive grieving, as emphasized in Worden's Four Tasks of Mourning (Worden, 2009), disbelief may warrant somewhat different clinical attention for males and females with respect to re-engagement and functioning. However, these interpretations should remain cautious given the exploratory nature of this analysis and the inclusion of a binary impairment node.

4.4. Limitations and future directions

Despite the above important findings, several limitations should be noted. First, the sample was gender-imbalanced, with more female than male participants, which may have affected the stability of gender-specific network estimates. Although this pattern is common in grief research (e.g. Maccallum et al., 2023), future studies should aim for more balanced recruitment to further examine the gender differences. Second, the cross-sectional design limits interpretation of temporal and causal relationships between symptoms and functional impairment. Longitudinal designs are needed to examine how these relationships evolve over time. Third, all variables were assessed using self-report measures, which may be influenced by reporting bias and shared method variance. In addition, functional impairment was assessed with a dichotomous item, potentially reducing sensitivity. Future studies may benefit from incorporating multi-method assessments and more dimensional measures of functional outcomes. Finally, as the sample was drawn from Hong Kong, cultural factors may influence grief expression and network structure. Replication in other cultural contexts is warranted to assess generalizability.

5. Conclusion

This study examined gender differences in prolonged grief symptoms assessed with the PG-13-R in a bereaved Hong Kong sample using network analysis. Although females reported higher overall symptom severity than males, the difference was small in magnitude, and the global strength of the symptom networks was similar. Gender differences were evident in symptom network structure. Meaninglessness and Disengagement were the most central symptoms across networks, whereas Identity Disruption was more central in males. Exploratory analyses further suggested that the symptoms most closely linked to Functional Impairment differed by gender. These findings highlight the value of moving beyond symptom severity to examine symptom interconnections and may help inform gender-sensitive assessment and clinical support following bereavement.

Funding Statement

This work was supported by the Hong Kong Jockey Club Charities Trust under the auspices of the Jockey Club Bereavement Care and Support Alliance, The University of Hong Kong (project serial number: 2022-0156). The funder had no role in study design; in the collection, analysis, or interpretation of data; in the writing of the report; or in the decision to submit the article for publication.

Disclosure statement

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

Data availability statement

The participants of this study did not give written consent for their data to be shared publicly; therefore, supporting data are not available due to the sensitive nature of the research.

Declaration of generative AI and AI-assisted technologies in the manuscript preparation process

During the preparation of this work the first author used ChatGPT (versions 4.0–5.2) to assist with generating analysis code and refining the language of the manuscript. After using this tool, the first author reviewed and edited the content as needed and takes full responsibility for the content of the published article.

Ethical approval statement

The study was approved by the University of Hong Kong (HKU) Human Research Ethics Committee (HREC; approval no. EA230512). All participants provided written informed consent before participating in the study. The authors assert that all procedures contributing to this work comply with the ethical standards of the HREC of HKU on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.

Supplemental Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/20008066.2026.2711483