A network analysis of ICD-11 complex PTSD symptoms in the treatment-seeking population in Iran

Abstract

Introduction

The 11th revision of the World Health Organization’s International Classification of Diseases (ICD-11) introduces a new disorder called complex posttraumatic stress disorder. This disorder is heterogeneous, and identifying its core symptoms is important for understanding its different aspects. The network approach to psychopathology allows for examining the structure of Complex PTSD at a symptom level, which helps in analyzing direct interactions between symptoms. This study aimed to explore the symptom structure of complex PTSD and identify critical symptoms in the treatment-seeking population in Iran.

Methods

Participants consisted of 463 people referred to comprehensive health centers in Tehran from September to December 2023 with psychopathological syndromes who had a history of trauma at different developmental stages. Complex PTSD symptoms were assessed using the International Trauma Questionnaire (ITQ) and International Measurement of Exposure to Traumatic Event checklist. Network analysis was applied to identify the most central symptoms (nodes) and associations between symptoms (edges) by the graphical LASSO algorithm and the EBCglasso method for network estimation.

Results

The result showed that the network of estimated symptoms for Complex PTSD in Iranian culture was highly accurate and stable. The most central symptoms in this network were feelings of failure and worthlessness. Additionally, “long-term upset” was identified as the connection between PTSD symptoms and DSO.

Conclusions

The study determined that feelings of failure and worthlessness are the most central symptoms in the Complex PTSD network. It was suggested that these symptoms should be given priority in theoretical and treatment models of Complex PTSD.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12888-024-06444-1.

Introduction

The World Health Organization (WHO) has developed the 11th version of the International Classification of Diseases (ICD-11), which now includes two types of stress and trauma disorders: Posttraumatic Stress Disorder (PTSD) and Complex Posttraumatic Stress Disorder (Complex PTSD [1]. Complex PTSD is considered a “sibling” disorder to posttraumatic stress disorder (PTSD) and has its own distinct yet interconnected conceptual framework that organizes symptomatology. The presence of a stressor determines the diagnosis of Complex PTSD or PTSD, and regardless of the nature of the stressor, the diagnosis is based on the symptom profile. This simplifies the diagnostic process for clinicians, allowing them to focus on treatment targets, symptoms, and issues instead of solely on trauma history [2].

Complex PTSD includes six symptom clusters: three PTSD criteria involving avoidance of trauma reminders, disturbing experiences, and feelings of current threat, as well as three disturbances of self-organization (DSO) symptoms. The DSO symptoms are defined as affect dysregulation (such as heightened emotional reactivity to minor stressors or emotional numbing), negative self-concept (including feelings of failure, worthlessness, inadequacy, shame, and loss of self-awareness), and disturbances in relationships (such as difficulty in establishing and maintaining interpersonal relationships, avoiding relationships, isolation, or exclusion) [2–4].

In the ICD-11, it is defined that both PTSD and Complex PTSD may develop following exposure to an extremely threatening or horrific event or series of events [1, 5]. Previous studies have shown that Complex PTSD occurs after exposure to trauma, such as childhood trauma, domestic violence, and sexual assault, which is prolonged, repetitive, interpersonal, and irreversible. These are situations in which physical, psychological, maturity, environmental, or social limitations make escape impossible. However, non-interpersonal traumas such as natural disasters or serious accidents might be more related to PTSD than Complex PTSD risk [6–11].

The multidimensional nature of Complex PTSD symptoms makes it challenging to explain the variability of symptoms with a comprehensive model. This diversity in symptoms can have significant implications, as different symptoms may impact functioning levels differently. Additionally, specific symptoms of Complex PTSD may hold greater importance than others in treatment plans [12, 13]. Therefore, determining Complex PTSD severity by solely employing a model assigning equal importance to all symptoms could be misleading. This is because specific symptoms may hold greater significance in the context of Complex PTSD [14].

An innovative approach to evaluating the validity of mental disorders involves conceptualizing psychopathology as a network of causal relationships based on network theory. This differs from traditional approaches that view symptoms as indicators of an underlying cause [15]. In the network approach to studying Complex PTSD, the various symptoms are seen as interconnected and potentially reinforcing. Complex PTSD is viewed as a network of nodes that represent specific symptoms that have strong causal connections with each other. These nodes also have notable associations with other symptoms in the network and are considered connectors between two different symptoms. Certain connections between two Complex PTSD symptoms referred to as edges, may be notably influential and stronger than others. A strong edge indicates a robust relationship between two symptoms that may co-occur or not. In other words, a strong edge suggests that the presence or activation of one Complex PTSD symptom also indicates the activation or presence of another symptom (i.e., the pathological network), and the inactivation or absence of one symptom indicates the inactivation or absence of another symptom (i.e., the healthy network) [16, 17]. Recent observations suggest that symptoms occupying central positions within psychopathology networks need to be prioritized in conceptual models of the disorder and targeted in treatment strategies [18] as it appears that central symptoms hold more significant potential for the prognosis of mental disorders [19, 20].

Previous studies have investigated network analysis of Complex PTSD symptoms in various countries, including Western countries, Africa, Israel, and China [21–25]. A study was conducted in three African countries, including Nigeria, Ghana, and Kenya [24], and in Western countries such as Germany [23], England, the United States, Ireland [21], Austria, Lithuania [22], and Israel [21] identified “feelings of worthlessness” as the central symptom of Complex PTSD. Another study conducted in China identified “hypervigilance” as a central symptom of Complex PTSD [25]. Since cultural factors influence the symptoms and clinical manifestations of mental disorders, resulting in diverse response patterns to psychological disorders among different populations, it is imperative to replicate this research in other countries and cultures [26, 27]. Iran’s cultural dynamics deeply influence the manifestation and comprehension of trauma, shaping individuals’ experiences and their approaches to seeking help. The intersection of cultural norms, historical context, and social structures forms a unique framework that affects how trauma is understood and addressed. Therefore, this study intends to examine the structure of the Complex PTSD symptoms network within Iranian culture.

Materials and methods

Participants and procedures

The current descriptive cross-sectional study was conducted between September 2023 and December 2023 at comprehensive health centers (CHC) in Tehran, which are affiliated with the Iran University of Medical Sciences. Patients with psychiatric symptoms and a history of trauma during different developmental stages were referred to the researcher by physicians at the CHC. The inclusion criteria were: [1] age between 20 and 50 years [2], ability to read and write in Persian, and [3] the absence of psychotic symptoms, bipolar disorder, brain injury, and substance use disorder. Individuals who did not meet these criteria were excluded from the study. The researcher re-administered the screening questions to ensure that participants met the inclusion criteria. The study included 463 patients, and their data were analyzed using R Studio software version 4.2.1 with a significance level of 0.05.

Measures

International trauma questionnaire

The ITQ is an 18-item self-reported measure used to evaluate PTSD and Complex PTSD based on the ICD-11 criteria. The scale ranges from 0 (not at all) to 4 (extremely) and utilizes both categorical and dimensional approaches to diagnose and rate the severity of symptoms. For PTSD diagnosis, items 1 to 9 are evaluated, while for Complex PTSD, in addition to items 1 to 9, scores related to disturbed self-organization (items 10 to 18) should also be obtained. To rate the severity of symptoms on a scale of 0 to 24, scores from questions 1 to 6 are used to evaluate the severity of PTSD symptoms, whereas scores related to questions 10 to 15 are utilized to assess the severity of disturbed self-organization (DSO) [28]. The Cronbach’s alpha coefficients for all PTSD and DSO subscales were reported to be over 0.77 for all subscales, except for the avoidance dimension, with a value of 0.67, indicating the acceptable validity of almost all subscales (> 0.79) [29]. In Iran, Cronbach’s alpha coefficient for internal consistency for the overall scale, PTSD, and DSO subscales was reported to be 0.89, 0.83, and 0.88, respectively. Test-retest reliability over a one-month duration was reported as excellent for both total scores (ICC = 0.868) and subscale scores (ICC = 0.847 to 0.854) [30]. In this study, Cronbach’s alpha of the ITQ score was 0.94.

International trauma exposure measure

The International Measurement of Exposure to a Traumatic Event (ITEM) checklist consists of 21 life-threatening events that an individual may have experienced at different stages of development (childhood, adolescence, and adulthood). Sixteen of these events align with the DSM-5 definition of trauma exposure, while the remaining five events correspond to psychological traumas based on the ICD-11 trauma exposure criteria. The cumulative score of childhood, adolescence, and adulthood traumas can be computed separately by adding up all events experienced before 12, between the ages of 13–18, and beyond 18, respectively. Additionally, the overall score for traumatic events across an individual’s lifetime can be determined by summing the scores of all events in each developmental stage. Furthermore, this checklist can identify the most distressing traumatic event, known as the trauma index, along with its frequency and the timing of its occurrence [31, 32].

Data analysis

Initially, we used the basic R software package (base R) to give a descriptive overview of the data. This involved representing demographic variables like age, education, gender, and marital status.

Subsequently, we used the R-package qgraph to estimate the symptom network that included all symptoms of Complex PTSD (DSO and PTSD symptoms) [33]. We used regularized partial correlation models to show the independent and unique relationships between symptoms. In this context, symptoms are represented as nodes, and their connections are called edges. The network is weighted and undirected due to the study’s cross-sectional nature. We estimated a polychoric matrix because the questionnaire data were measured at an ordinal scale. We then used a Gaussian Graphical Model (GGM) to calculate the partial pairwise correlation parameters between all nodes [34, 35].

We chose a cautious, data-driven method to visualize the networks with no directed hypotheses. Utilizing the graphical least absolute shrinkage and selection operator (glasso) implemented in qgraph, we constructed sparse networks by partial correlations while accommodating the questionnaire ordinal scale. This approach can directly estimate the covariance matrix inverse and shrink small edges and numerous parameters to zero using a penalized maximum likelihood solution according to the Extended Bayesian Information Criterion (EBIC) [35, 36].

The resulting graph features nodes represented as circles connected by lines denoting edges, with thicker edges indicating stronger relationships between nodes. The direction of these relationships is illustrated by color, with positive relationships depicted in blue and negative relationships in red. Additionally, the size of the nodes corresponds to their frequency in the raw data, with nodes appearing more frequently depicted as more significant in size. The graph’s layout was determined using the Fruchterman-Reingold algorithm, which clusters interconnected nodes closer together toward the center of the network [37].

In the qgraph library, we used centrality measures to identify important nodes. Different centrality measures capture different relationships between a focal unit and other units. For this study, we calculated node strength and a new centrality measure called “expected influence,” which is especially relevant and strong in psychopathology networks compared to other indicators [34, 38]. Node strength represents the total sum of all edges attached to a specific node [39]. Expected influence (EI) indicates the number of links associated with a node, illustrating its significance within the network. This measure is relative because, even in networks with generally low edge weights, there will always be a node with considerable EI [33]. The expected influence is determined by the sum of edge weights (with negative signs preserved) directly linked to a node [40], representing the node’s active potential throughout the Complex PTSD symptom network [41].

The accuracy and stability of the network model were evaluated using three approaches. Firstly, the non-parametric bootstrap method was used to estimate edge weight accuracy by calculating confidence intervals (CIs) with 1000 bootstrapped samples. The original dataset was resampled to generate new datasets, and 95% of CIs were derived from the average bootstrapped edge weights. This allowed comparison with the edge weights of the entire sample to assess concordance [34]. Secondly, subset bootstraps were used to compute the coefficient of correlation stability (CS-C) to evaluate the stability of expected influence (EI). CS-C indicates the maximum proportion of cases that can be excluded to achieve a specified correlation value with a 95% probability. A CS-C value ideally exceeds 0.5 but should not fall below 0.25. Lastly, bootstrap difference tests were conducted to assess differences in network characteristics. These tests used 95% CIs to indicate whether there were significant differences in centrality indices between two edge weights or nodes. The bootnet package examined the bootstrap method and overall stability [42].

Results

In a study, 463 people were evaluated, with 283 (61.1%) of them being aged 20 to 25. Furthermore, 291 (62.9%) of the participants were female, and 172 (37.1%) were male. Two hundred seventy-nine participants (60.3%) were single, and 166 (35.9%) had a bachelor’s degree. Additionally, 16 (3.5%) reported experiencing one traumatic event, 22 (4.8%) reported two traumatic events, and 425 (91.7%) reported three or more traumatic events in their lifetime. Among the participants, 392 (84.7%) reported childhood traumatic events, 409 (88.3%) reported adolescent traumatic events, and 401 (86.6%) reported adult traumatic events. The mean scores were as follows: PTSD (M = 8.75, SD = 5.37), DSO (M = 11.30, SD = 6.58).

The LASSO graphical algorithm illustrated the network of Complex PTSD symptoms following traumatic experiences in different developmental stages (Fig. 1). The centrality indices indicated that “feelings of failure” and “feelings of worthlessness” were the strongest nodes, while “having disturbing dreams” and “feeling uncomfortable” were the weakest nodes in the Complex PTSD symptom network. Additionally, AD1 (long-term upset) was identified as a mediating node between PTSD and DSO (Fig. 2).

Fig. 1.

Gaussian graphical model. Note. Nodes (circles) are scores of Complex PTSD symptoms. Undirected edges, indicated by lines between nodes, represent partial correlations between variables after controlling for relationships with all other nodes. Blue lines between the circles indicate positive conditional associations

Fig. 2.

Standardized node centrality indices of Complex PTSD symptoms

The bootstrapping results of edge weight CIs show that the red line indicates the weight value of each edge, and the gray area indicates the 95% CI. The CI near the edges, particularly the edges with large weights, was low, indicating a high level of stability of network estimation (Fig. 1; Supplementary Material). In addition, the bootstrapping subset results showed that the curve decreases slowly, and the centrality values of the original data and the subsets are highly correlated even after removing many subjects. Therefore, the centrality estimation can be regarded as stable. The correlation accuracy coefficient was above 0.50 (Strength = 0.672), which means that strength was still related to the original data after removing 75% of the data (Fig. 2; Supplementary Material).

Furthermore, the edge weight difference test results show that Small black boxes indicate a significant difference between the corresponding edge weights. The power centrality index difference test findings suggest that black boxes indicate significant differences in the power centrality index between the two connected nodes. The symptoms with higher centrality showed higher values compared to other symptoms (Figs. 3 and 4; Supplementary Material).

Discussion

We investigated a study to assess the network structure of Complex PTSD symptoms in treatment-seeking populations in Iran. Our findings showed that the estimated symptom network was highly accurate and stable. In Iranian culture, the most central symptoms of Complex PTSD were “feelings of failure” and “feelings of worthlessness,” which were strongly associated with other symptoms. Additionally, “long-term upset” was identified as the bridge between PTSD symptoms and DSO.

Previous research has shown that negative self-concept clusters, such as “feelings of worthlessness” and “feelings of failure,” are central in the network structure of ICD-11 Complex PTSD [21–25, 43, 44]. This finding can be explained by considering the identity of individuals with Complex PTSD symptoms, which reflects a history of repeated and prolonged childhood maltreatment. As a result, the individual’s expectations that the self is valuable, others are benevolent, and the world is meaningful are severely questioned. Complex PTSD often involves self-representations derived from previous experiences [45].

Although the three DSO symptom clusters in Complex PTSD (notions of the self, emotional expression, and interpersonal relationships) are at the core of what has been studied in cultural psychology over the past few decades [46], an interesting finding is that symptoms related to the negative self-concept cluster are central across different countries with different cultures. This suggests that cultural differences in the conceptualization of the self may have an impact on psychological processes related to PTSD and Complex PTSD. Cultural differences in self-concept can be analyzed by comparing flexibility versus self-consistency. In Eastern cultures, people are more likely to adapt to various situations while maintaining a relatively stable self. On the other hand, in American culture, there is a belief that the social world is flexible and that individuals can shape it, reflecting the concepts of personal agency and control [47, 48]. Heine et al. described that self-esteem, or having positive feelings about oneself, is a crucial aspect of American culture. In contrast, in Asian culture, self-criticism motivates individuals to improve themselves and adapt to their social environment. According to American culture, having self-esteem or self-respect, which involves positive feelings about oneself, is sometimes interpreted as a sign that one is not striving hard enough to become a better person [49]. DSO symptoms tend to be more prevalent in Asian cultures because of a higher tendency to agree with negative self-statements [50]. However, negative self-concept is a crucial symptom in the network of symptoms for Complex PTSD across all cultures. One’s self-concept is influenced by cultural norms and intersubjective concepts, which shape how individuals perceive themselves within their cultural group. Intersubjective perceptions may differ from personal beliefs and values, as well as shared cultural beliefs and values. When an individual perceives a deviation from normative intersubjective meaning, this may lead to negative self-appraisal. Therefore, while negative self-appraisal is likely universal, its meaning is culturally influenced [51].

Our findings showed that “long-term upset,” which is one of the symptoms of DSO, is the central symptom between PTSD and DSO. A study that examined the network analysis of PTSD symptoms showed that " anhedonia” and “dysphoria” are the PTSD central symptoms, which indicated that emotional dysregulation could have a significant effect on PTSD symptom clusters​ [52]. Thus, “long-term upset” may be the bridge between DSO and PTSD through emotion dysregulation. This finding supports the potential value of psychological interventions based on emotion regulation in the treatment of Complex PTSD patients.

The findings of this research are important for developing effective therapeutic interventions. Previous studies have shown that treating Complex PTSD involves focusing on symptoms that are comorbid with PTSD. Considering that ICD-11 Complex PTSD is a new disorder, it will take time to gather sufficient evidence regarding its treatment. However, there is evidence of interventions that have partially addressed symptoms of Complex PTSD, including those of DSO [53]. Furthermore, Interventions that address these problems in ways relevant to their specific content and dynamic may improve treatment outcomes [54]. For example, it has been proposed that profoundly negative self-concepts may respond well to compassion-focused interventions [55]. Another practical approach for treating Complex PTSD involves a phased intervention. Initially, the focus is on addressing DSO symptoms and related challenges in daily life, such as enhancing safety, emotion regulations, and improving social skills. Subsequently, the treatment delves into exploring the underlying trauma. This sequence is supported because bolstering emotional, psychological, and social resources can enhance daily functioning and elevate trauma-focused therapy [56, 57].

In conclusion, further research is imperative to explore the significance of different symptom clusters in Complex PTSD, which could enhance sufficient treatment.

Strength and limitations

The strength of the current research can be considered to be using the network analysis approach as an innovative method to assess the network structure of Complex PTSD symptoms in Iranian culture. Such research can determine the focus of therapeutic interventions by identifying the core symptoms of Complex PTSD. This research introduces a novel approach to visually representing Complex PTSD, offering three levels of insight: firstly, robust connectivity of symptoms across the network; secondly, distinctive patterns distinguishing symptoms of PTSD from those of disturbances in Self-Organization (DSO); thirdly, aligning with the hierarchical framework of Complex PTSD and PTSD in the ICD-11, the entire network serves as a visual portrayal of Complex PTSD. This innovative, multifaceted methodology can provide clinicians and researchers with a clearer understanding of the symptoms associated with these two distinct disorders. Our findings affirm that the clusters of DSO symptoms and PTSD symptoms are individual entities, each with potent internal associations, suggesting that DSO symptoms represent enduring constructs with internal coherence. These findings prompt consideration of whether DSO symptoms can constitute an independent disorder entity, given the substantial interconnections observed among DSO symptoms, even in comparison to connections within PTSD symptoms. Considerable evidence [58] indicates that DSO symptoms are integral to Complex PTSD, thereby bolstering the validity of its fundamental concept of Complex PTSD and hierarchical structure.

Network analysis is a valuable approach to understanding psychopathology, but it has limitations that need to be addressed in future research. Most network studies in psychopathology, including our research, have employed cross-sectional designs, which limit the ability to make causal inferences. Longitudinal network studies may identify predictive relationships, but causal inferences cannot be drawn from network analyses without experimental designs [59, 60]. Network studies have mainly relied on self-reported symptoms indicated by a single item, leading to concerns about the reliability of symptom measurement within these networks. Furthermore, network analyses have been criticized for their generalizability (i.e., the convergence of results across similar samples) and stability (i.e., the replicability of results across randomly selected subsamples) [60]. These criticisms include issues related to overfitting networks to particular datasets, the accuracy of parameter estimation, and the degree to which network parameters are influenced by sampling variation [35, 61, 62]. In summary, network analysis provides valuable insights into the interaction of psychological variables. However, researchers must acknowledge the limitations associated with data interpretation, cross-sectional designs, and causal inferences.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

We all confirm that we have approved the order and contribution of the authors listed in the manuscript. ZM: Conceptualization, Project Administration, investigation, writing-original draft; MD: Writing- review & editing, Supervision, Conceptualization, visualization; FFL: Writing- review & editing, Validation, Visualization; HF: Writing- review & editing, formal analysis, methodology, data curation; AA: Writing- review & editing, resources; ZO: Writing- review & editing, investigation.

Funding

This research received no specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The raw data supporting the conclusions of this article will be made available by the authors without undue reservation.

Declarations

Ethics approval and consent to participate

The study was approved by the Ethics Committee of the Iran University of Medical Sciences (ethics code: IR.IUMS.REC.1402.454) and in accordance with the guidelines outlined in the Declaration of Helsinki. All participants provided written informed consent and were informed that their participation was voluntary. They were also assured of their right to withdraw from the study at any time. Additionally, they were informed about the confidentiality of their information.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.