Network analysis and risk identification method for emotional trouble of gynecological cancer patient

Abstract

The invention provides a network analysis and risk identification method for emotional troubles of gynecological cancer patients. The method comprises the following steps: S1, acquiring emotional symptom data and psychological social variable data from a plurality of gynecological cancer patients; s2, constructing an extended Gaussian graph model network comprising emotional symptom nodes and psychological social variable nodes, the emotional symptom nodes being selected based on a simple mood state scale simple version, and psychological social variables at least comprising neuroplasm and the like; s3, calculating and analyzing centrality indexes, bridge indexes and the like of all nodes in the network, through a symptom network analysis method, the limitation of traditional total score evaluation is broken through, and dynamic association between emotional symptoms and action paths of psychological and social factors can be revealed finely; identified core symptoms and key bridge nodes can provide specific targets for clinical intervention, and individualized and layered psychological and social support is realized; the method is suitable for psychological risk assessment and intervention planning of gynecological cancer patients.

Description

Technical Field

[0001] This invention relates to the field of mental health assessment technology, and in particular to a network analysis and risk identification method for emotional distress in gynecological cancer patients. Background Technology

[0002] Gynecological cancers pose a significant threat to women's health worldwide, and their treatment is often accompanied by significant emotional distress, such as anxiety, depression, and fatigue. These emotional symptoms not only affect patients' quality of life but may also reduce treatment adherence and worsen clinical outcomes. Currently, standardized questionnaire total scores or subscale scores are commonly used in clinical practice to assess emotional distress. While this method is simple, it fails to reveal the dynamic relationships between symptoms and makes it difficult to identify which symptoms play a key role in maintaining the emotional distress network. A systematic and granular analysis is still lacking on how psychosocial factors such as neuroticism, perceived burden of self, and economic toxicity interact with emotional symptoms. Summary of the Invention

[0003] In view of this, to address the problems existing in the technical background, this invention proposes a network analysis and risk identification method for emotional distress in gynecological cancer patients. Specifically, it includes the following: A network analysis and risk identification method for emotional distress in gynecological cancer patients includes the following steps: Step S1: Collect emotional symptom data and psychosocial variable data from multiple gynecological cancer patients; Step S2: Construct an extended Gaussian graph model network containing emotional symptom nodes and psychosocial variable nodes. The emotional symptom nodes are selected based on a simplified version of the Brief Mood State Scale, and the psychosocial variables include at least neuroticism, self-perceived burden, disease cognition, economic toxicity, and psychological capital. Step S3: Calculate and analyze the centrality index and bridging index of each node in the network to identify the core emotional symptoms and key psychosocial entry points in the network. Step S4: Based on the core emotional symptoms and key psychosocial entry points, generate an identification report for individualized psychosocial risk assessment and intervention strategy development. In one embodiment of the present invention, the emotional symptom data includes specific item scores under five dimensions: tension-anxiety, depression-frustration, anger-hostility, fatigue-burnout, and confusion-bewilderment. The psychosocial variable data are obtained by obtaining a total score through the corresponding standardized scale. In one embodiment of the present invention, when constructing the extended Gaussian graph model network, the EBICglasso regularization method is used to estimate the partial correlation coefficients between nodes to form network edges, and the nonparametric bootstrapping method is used to evaluate the accuracy of network edges and the stability of the node centrality index. In one embodiment of the present invention, the process of identifying core emotional symptoms includes calculating the strength centrality and expected influence centrality of nodes, and identifying the emotional symptom node with the highest centrality index as the core node for maintaining the emotional distress network. In one embodiment of the present invention, the process of identifying key psychosocial entry points includes calculating the bridge strength and expected bridge impact of nodes, and identifying the psychosocial variable node with the highest bridge centrality index as the key bridge node connecting external stressors and the internal emotional symptom network. In one embodiment of the present invention, before constructing the network model, the collected continuous variable data is subjected to a rank-based nonparametric normal transformation, and the emotional symptom data is subjected to regression analysis with respect to demographic and clinical covariates to obtain residuals. Based on the residuals, a residual symptom network after controlling for covariates is constructed. In one embodiment of the present invention, a network comparison step is further included: patients are divided into high-risk and low-risk groups based on the median of at least one psychosocial variable, their symptom networks are constructed respectively, and the differences between the two groups in terms of overall network strength and network structure are analyzed using a network comparison test method. In one embodiment of the present invention, the generation of the identification report includes: listing the identified core emotional symptoms and their connection patterns in the network, indicating key psychosocial entry points and their correlation strength with specific emotional symptoms, and recommending targeted stratified psychological interventions or care strategies accordingly. The above technical solution has the following beneficial effects: This invention, through symptom network analysis, overcomes the limitations of traditional total score assessment, enabling precise revelation of the dynamic relationships between emotional symptoms and the pathways of psychosocial factors. The identified core symptoms and key bridging nodes can provide specific targets for clinical intervention, achieving individualized and stratified psychosocial support. This invention has good operability and scalability, and is applicable to the psychological risk assessment and intervention planning of gynecological cancer patients. Attached Figure Description Figure 1 This is a schematic diagram of the extended Gaussian graph model network that integrates emotional symptoms and psychosocial variables according to the present invention. Figure 2 This is a schematic diagram of the centrality index of all nodes in the extended network of the present invention. Figure 3 This is a schematic diagram illustrating the expected bridge impact of nodes in the extended network of this invention. Figure 4 This is a schematic diagram of the residual emotional symptom network after controlling for covariates in this invention. Figure 5 This is a schematic diagram of the centrality index of nodes in the residual network of the present invention. Figure 6This is a schematic diagram illustrating the expected bridge impact of nodes in the residual network of this invention. Detailed Implementation The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. See Figures 1-6 The method for network analysis and risk identification of emotional distress in gynecological cancer patients, as shown, includes the following steps: Step S1: Collect emotional symptom data and psychosocial variable data from multiple gynecological cancer patients; Step S2: Construct an extended Gaussian graph model network containing emotional symptom nodes and psychosocial variable nodes. The emotional symptom nodes are selected based on a simplified version of the Brief Mood State Scale, and the psychosocial variables include at least neuroticism, self-perceived burden, disease cognition, economic toxicity, and psychological capital. Step S3: Calculate and analyze the centrality index and bridging index of each node in the network to identify the core emotional symptoms and key psychosocial entry points in the network. Step S4: Based on the core emotional symptoms and key psychosocial entry points, generate an identification report for individualized psychosocial risk assessment and intervention strategy development. Example 2, based on Example 1, in one embodiment of the present invention, the emotional symptom data includes specific item scores under five dimensions: tension-anxiety, depression-frustration, anger-hostility, fatigue-burnout, and confusion-bewilderment. The psychosocial variable data are obtained by obtaining a total score through the corresponding standardized scale. When constructing the extended Gaussian graph model network, the EBICglasso regularization method is used to estimate the partial correlation coefficients between nodes to form network edges, and the nonparametric bootstrapping method is used to evaluate the accuracy of network edges and the stability of node centrality indices. The process of identifying core emotional symptoms includes calculating the strength centrality and expected influence centrality of nodes, and identifying the emotional symptom node with the highest centrality index as the core node maintaining the emotional distress network. The process of identifying key psychosocial entry points includes calculating the bridge strength and expected influence of nodes, and identifying the psychosocial variable node with the highest bridge centrality index as the key bridge node connecting external stressors and the internal emotional symptom network. Before constructing the network model, the collected continuous variable data is subjected to a rank-based nonparametric normal transformation, and regression analysis is performed on the emotional symptom data with demographic and clinical covariates to obtain residuals. Based on these residuals, a residual symptom network after controlling for covariates is constructed. Example 3, based on Example 1, further includes a network comparison step: patients are divided into high-risk and low-risk groups based on the median of at least one psychosocial variable, symptom networks are constructed for each group, and network comparison tests are used to analyze the differences between the two groups in overall network strength and network structure. The generated identification report includes: listing the identified core emotional symptoms and their connection patterns in the network, indicating key psychosocial entry points and their association strength with specific emotional symptoms, and recommending targeted stratified psychological interventions or nursing strategies accordingly. Data Collection and Research Subjects: During the implementation of this invention, patients with histologically confirmed gynecological malignancies, including cervical cancer, ovarian cancer, endometrial cancer, vulvar cancer, and vaginal cancer, were recruited consecutively from the gynecologic oncology departments of three tertiary hospitals in Wuxi, China. Patients had to be at least 18 years old, have intact cognitive function, possess Chinese reading and writing abilities, and have completed at least one major treatment. This study excluded patients with a history of severe mental disorders, major comorbidities that might interfere with emotional assessment, or those unable to complete the assessment due to cognitive and language impairments. All participants were required to sign a written informed consent form. The study period was set from October 1, 2024 to May 31, 2025, employing a rolling recruitment strategy to ensure sample representativeness. The final sample size included in the analysis was 415 cases. This sample size was determined based on the rule of thumb that each node in network analysis requires 10 to 15 participants, providing sufficient statistical power for the extended network analysis, which included 16 emotional symptom nodes and 5 psychosocial variables. Assessment tools and variable measurement: The Short Form of the Mood State Scale (Simplified Chinese version) was used to measure emotional symptoms. This scale assesses five dimensions: tension-anxiety, depression-frustration, anger-hostility, fatigue-burnout, and confusion-lackness. Sixteen high-frequency negative emotion items were selected as symptom nodes for network analysis. Each item was scored on a Likert scale from 0 to 4, and the scale showed good internal consistency. The following standardized tools were used to measure psychosocial variables: The Big Five Personality Inventory - Neuroproton Scale is used to assess neurotic personality traits; a higher score indicates greater emotional instability. The Self-Perceived Burden Scale is used to assess patients' perceptions of physical, emotional, and economic burdens. The Brief Disease Perception Questionnaire is used to assess people’s knowledge of disease and perceived threat. The Financial Toxicity Assessment Scale is used to assess the financial stress caused by illness. The Positive Psychological Capital Questionnaire is used to assess hope, optimism, resilience, and self-efficacy. All scales have been validated in the Chinese cancer patient population and have shown acceptable reliability and validity. Additionally, patient demographic information and clinical characteristics should be collected as covariates. Data Processing and Network Construction Preparation: After data collection, preprocessing is performed first. For cases containing missing values, if the missing value ratio is less than 10%, a single imputation using the chain equation method can be used. To ensure that the variables conform to the assumption of multivariate normal distribution, all continuous variables undergo rank-based nonparametric normal transformation. To isolate the influence of demographic and clinical covariates on the association between symptoms, regression analysis is performed on each emotional symptom against covariates such as age, education level, marital status, cancer type, stage, and treatment method. Residuals are extracted, and the residuals are again subjected to nonparametric normal transformation to form a dataset used to construct the residual symptom network. Construction and Analysis of Extended Gaussian Graph Network: Network analysis was performed using R language and related software packages. First, an extended Gaussian graph network was constructed, containing 21 nodes, including 16 nodes representing emotional symptoms and 5 nodes representing psychosocial variables. The EBICglasso regularization method was used to estimate the partial correlation coefficients between nodes, thus forming network connection edges. The network visualization was performed using the Fruchterman-Reingold algorithm for layout. like Figure 1 As shown, this extended network visually illustrates the complex connections between emotional symptom nodes and psychosocial variable nodes. The diagram reveals that symptom nodes such as depression-frustration, tension-anxiety, and fatigue-burnout are closely linked, forming a high-density core cluster. Psychosocial variable nodes such as neuroticism and perceived burden of self are significantly associated with these core symptom clusters. To assess the reliability and stability of the network, a nonparametric bootstrapping method was used for 1000 resampling cycles to calculate the correlation stability coefficient between the 95% confidence interval of the edge weights and the centrality index. While the accuracy of the strongest connections is acceptable, it should be noted that the confidence intervals of the weaker connections are wider, and the stability coefficient of the centrality index does not meet high standards. Therefore, the interpretation of the centrality index should be approached with caution, focusing primarily on the most robust connections. Identification of core symptoms and key bridge nodes In network analysis, centrality indices are used to identify critical nodes in a network. A node's strength centrality and expected influence centrality reflect its connectivity and impact within the network. like Figure 2As shown, the centrality indices of each node in the extended network are visualized. The analysis results show that items in the fatigue-burnout dimension, as well as some depression-depression and tension-anxiety items, have the highest centrality values. This indicates that "fatigue" occupies a central hub position in the emotional distress network, not only highlighting its own symptoms but also extensively connecting other emotional and physical experiences, making it a key factor in maintaining the stability of the entire symptom network. Bridge centrality analysis is used to identify key variables connecting different node communities. Nodes are divided into emotional symptom communities and psychosocial variable communities, and bridge strength and expected bridge impact are calculated. like Figure 3 As shown, the bridge-expected impact analysis clearly reveals that self-perceived burden and neuroticism have the highest bridge centrality indices. This means that these two psychosocial variables are the most important bridges connecting external stressors and the internal emotional symptom network. Specific path analysis indicates that self-perceived burden is strongly associated with emotional symptoms such as depression and despair, while neuroticism is closely related to symptoms such as anxiety and tension. This reveals two key psychosocial pathways that exacerbate emotional distress: one is guilt and depression triggered by the perception of oneself as a burden on the family; the other is anxiety and confusion triggered by emotional instability and overestimation of threats due to high neuroticism. To explore the pure associations between symptoms after excluding the influence of demographic and clinical factors, a residual symptom network was constructed. like Figure 4 As shown, after controlling for covariates, the core topology of the network remains largely consistent with the extended network, and the tight cluster of depression-anxiety-fatigue remains stable. This indicates that the intrinsic connections between emotional symptoms are robust and not entirely driven by external clinical features. like Figure 5 and Figure 6 As shown, although the centrality index and the bridge expected influence index in the residual network have decreased in value, symptoms such as fatigue and tension-anxiety still maintain a relatively high influence, and some nodes still play a bridging role between symptom dimensions. This suggests that even excluding background factors, some symptoms themselves have the potential to drive the dynamic evolution of emotional distress. To investigate whether there are structural differences in the emotional distress networks of patients at different risk levels, patients were divided into high-risk and low-risk groups based on the median of each psychosocial variable. For example, patients were divided into high-neuroticism and low-neuroticism groups based on the median neuroticism score. Symptom networks were constructed for each subgroup, and network comparison tests were used to examine differences between groups in overall network connectivity and network structure. The analysis revealed no statistically significant differences in overall network strength and global structure between the high-risk and low-risk groups across all psychosocial variables. This suggests that, despite differences in psychosocial risk levels, the underlying patterns of association among patients' distress symptoms may be similar. However, the use of median segmentation in this analysis reduced statistical power, and the descriptive analysis showed that network connections were numerically denser in the high neuroticism group, thus the possibility of subtle structural differences cannot be completely ruled out. Future research could employ continuous variable moderating analysis or potential profile analysis to more accurately capture the impact of risk gradients. Based on the network analysis results above, the system can automatically generate a personalized psychosocial risk assessment and intervention strategy identification report. The report content specifically includes: Core symptom identification: List the most central emotional symptoms in the patient's network, such as "fatigue," "depression," or "anxiety," and describe the specific connection patterns of these symptoms in the network. Key risk factor analysis: Identify the psychosocial entry points that have the greatest impact on the patient's emotional network, such as "high level of self-perceived burden" or "high neuroticism," and clarify which specific emotional symptoms they are most strongly associated with. Network characteristics summary: Summarize the overall density of the patient's emotional network, the composition of the core symptom clusters, and the presence of prominent bridging symptoms. Stratified intervention recommendations: For patients whose core symptom is "fatigue", behavioral activation, activity rhythm regulation and energy management interventions are recommended. For patients with a prominent "self-perceived burden" bridging effect, family systems therapy, meaning-centered therapy, or dignity therapy are recommended to alleviate feelings of guilt and relational stress. For patients with a significant “neurotic” bridging effect, interventions focusing on emotion regulation and stress management, such as mindfulness training, cognitive restructuring, and acceptance and commitment therapy, are recommended. Clinical monitoring focus: It is recommended that clinicians focus on monitoring changes in the core symptoms and bridging symptoms identified in the report during follow-up, as these are key indicators for assessing the effectiveness of interventions and disease fluctuations. Through the above systematic steps, this invention provides a complete methodology from data collection, network modeling, key node identification to personalized report generation, enabling refined assessment and risk identification of emotional distress in gynecological cancer patients, and providing a scientific basis for implementing precise psychosocial interventions. The basic principles and main features of the present invention have been described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are only illustrative of the principles of the present invention. Various changes and modifications can be made to the present invention without departing from the spirit and scope of the present invention. All such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the invention is defined by the appended claims and their equivalents.

Claims

1. A network analysis and risk identification method for emotional distress in gynecological cancer patients, characterized in that, Includes the following steps: Step S1: Collect emotional symptom data and psychosocial variable data from multiple gynecological cancer patients; Step S2: Construct an extended Gaussian graph model network containing emotional symptom nodes and psychosocial variable nodes. The emotional symptom nodes are selected based on a simplified version of the Brief Mood State Scale, and the psychosocial variables include at least neuroticism, self-perceived burden, disease cognition, economic toxicity, and psychological capital. Step S3: Calculate and analyze the centrality index and bridging index of each node in the network to identify the core emotional symptoms and key psychosocial entry points in the network. Step S4: Based on the core emotional symptoms and key psychosocial entry points, generate an identification report for individualized psychosocial risk assessment and intervention strategy development.

2. The method for network analysis and risk identification of emotional distress in gynecological cancer patients according to claim 1, characterized in that, The emotional symptom data includes specific item scores under five dimensions: tension-anxiety, depression-frustration, anger-hostility, fatigue-burnout, and confusion-lackness. The psychosocial variable data are obtained by obtaining the total score through the corresponding standardized scale.

3. The method for network analysis and risk identification of emotional distress in gynecological cancer patients according to claim 1, characterized in that, When constructing the extended Gaussian graph model network, the EBICglasso regularization method is used to estimate the partial correlation coefficients between nodes to form network edges, and the nonparametric bootstrapping method is used to evaluate the accuracy of network edges and the stability of the node centrality index.

4. The method for network analysis and risk identification of emotional distress in gynecological cancer patients according to claim 1, characterized in that, The process of identifying core emotional symptoms involves calculating the intensity centrality and expected influence centrality of nodes, and identifying the emotional symptom node with the highest centrality index as the core node that maintains the emotional distress network.

5. The method for network analysis and risk identification of emotional distress in gynecological cancer patients according to claim 1, characterized in that, The process of identifying key psychosocial entry points includes calculating the bridge strength and expected impact of nodes, and identifying the psychosocial variable nodes with the highest bridge centrality index as key bridge nodes connecting external stressors and internal emotional symptom networks.

6. The method for network analysis and risk identification of emotional distress in gynecological cancer patients according to claim 1, characterized in that, Before constructing the network model, the collected continuous variable data were subjected to a rank-based nonparametric normal transformation, and the emotional symptom data were subjected to regression analysis with respect to demographic and clinical covariates to obtain residuals. Based on these residuals, a residual symptom network was constructed after controlling for covariates.

7. The method for network analysis and risk identification of emotional distress in gynecological cancer patients according to claim 1, characterized in that, It also includes a network comparison step: patients are divided into high-risk and low-risk groups based on the median of at least one psychosocial variable, their symptom networks are constructed separately, and the network comparison test is used to analyze the differences between the two groups in terms of overall network strength and network structure.

8. The method for network analysis and risk identification of emotional distress in gynecological cancer patients according to claim 1, characterized in that, The generated identification report includes: listing the identified core emotional symptoms and their connection patterns in the network, pointing out key psychosocial entry points and their correlation strength with specific emotional symptoms, and recommending targeted stratified psychological interventions or care strategies accordingly.

Images