Rheumatoid Arthritis Complications Prediction System Based on Asymmetric Causal Association

Abstract

The present invention discloses a prediction system for rheumatoid arthritis complications based on asymmetric causal association, which acquires the clinical time series data of patients and preprocesses it, counts the number of occurrences of complication disease labels in the clinical data, constructs a disease correlation matrix, calculates the in-degree and out-degree of each pair of complication disease labels, and constructs a complication causal association enhancement matrix; hierarchically clusters the complication diseases according to the correlation and causal association degree to obtain disease subgroups of complication disease combinations, and dynamically predicts the complication risk. The present invention can more accurately reflect the coexistence of multiple diseases and their dynamic evolution in patients with rheumatoid arthritis (RA), and provide a more accurate risk assessment and prediction tool for clinical decision support.

Description

Technical Field

[0001] The present invention relates to the field of medical and health information technology, and particularly to a prediction system for rheumatoid arthritis complications based on asymmetric causal associations. Background Art

[0002] Rheumatoid Arthritis (RA) is a chronic, systemic autoimmune disease, mainly manifested as chronic inflammation and destruction of joints. However, RA is not limited to joints and may also affect multiple systems and organs, leading to various complications. These complications are characterized by multiplicity and organ heterogeneity, severely affecting the quality of life of patients. Common complications include cardiovascular diseases, interstitial lung diseases, pulmonary hypertension, etc. These complications not only increase the pain of patients but also significantly increase the mortality rate. Therefore, it is crucial to develop effective prediction tools for early identification and intervention of these complications.

[0003] Asymmetric causal association refers to the fact that in the causal relationship between diseases, the influence of disease A on disease B is not symmetrically bidirectional, that is, disease A may cause the occurrence of disease B, but the reverse relationship does not hold, or the intensity of the influence between the two is different. For example, in RA patients, pulmonary hypertension and interstitial lung disease often coexist, but pulmonary hypertension is usually caused by interstitial lung disease, rather than the other way around. Therefore, the causal relationship between diseases is asymmetric, reflecting the directionality and intensity differences between diseases. This asymmetry reveals the complex interactions between diseases, and the study of asymmetric causal associations helps to identify the dominant factors in specific disease combinations, thereby optimizing early warning and intervention strategies.

[0004] Disease correlation refers to the statistically significant non-random co-occurrence between two or more health conditions or diseases within a specific population. This means that the frequency of these phenomena occurring together in the same group is higher than the expected level determined by chance alone. Specifically, disease correlation describes the frequency and pattern of the occurrence of these diseases within the same group or individual, indicating whether they occur together or there is a certain degree of association between their occurrences. Disease correlation can suggest the existence of potential causal relationships, common risk factors. This concept is used to identify potential connections between diseases, thereby revealing common pathological pathways, risk factors, or shared etiological mechanisms. In the field of medical and health information, understanding disease associations helps to improve diagnosis, predict comorbidities, and design effective treatment and prevention strategies.

[0005] Disease feature similarity refers to the assessment of the similarity degree between different diseases by comparing their clinical features, symptoms, signs, pathological features, etc. Specifically, disease feature similarity is to find and measure the similarity relationship or pattern between different diseases by analyzing the multi-dimensional features of diseases (such as symptoms, clinical data). By analyzing and comparing these features, researchers can discover the commonalities of diseases, which helps to understand the occurrence and development mechanisms of diseases, identify potential therapeutic targets, and develop new diagnostic tools.

[0006] Causal directionality refers to clearly determining the direction of causal effect in a causal relationship, that is, identifying which variable is the cause and which is the result. It describes which variable's change triggers the change of another variable among a pair of related variables. Causal directionality not only cares about whether there is an association (correlation) between variables, but more importantly, it is necessary to clarify the causal direction of this association, that is, which variable is in the position of triggering change in the causal chain and which is the affected object. In the field of medical informatics, disease causal directionality refers to clearly distinguishing the causal relationship between different diseases, that is, which disease is the cause of another disease and which is the result. Specifically, it focuses on whether the change of a certain disease or health status directly causes the occurrence or progression of another disease. The study of causal directionality between diseases helps to reveal the mutual relationship between diseases, understand the occurrence mechanisms of disease transmission and secondary diseases, and provide guidance for clinical decision-making.

[0007] Dynamic prediction refers to using time series data and dynamic models to predict the development trend of diseases, focusing on the temporality and variability in the disease process, rather than just predicting based on data at a single point in time. Different from static cross-sectional prediction, dynamic prediction takes into account the longitudinal data of individuals changing over time, including symptom fluctuations, laboratory test results, treatment responses, etc. In patients with RA, as the disease progresses, the health status of patients may change, and new complications (such as cardiovascular diseases or pulmonary hypertension) may occur. By modeling time series data, dynamic prediction can reveal the complex changes in the disease process. By capturing these dynamic changes, dynamic prediction can more accurately reflect the individualized disease trajectory, thus achieving the goal of personalized medicine, especially showing significant advantages in dealing with chronic diseases and complex comorbidities.

[0008] Although some progress has been made in the prediction of RA complications in recent years, the existing methods still have some limitations:

[0009] 1. Ignoring the complex correlations and synergistic effects among complications

[0010] Existing prediction models usually rely on static cross-sectional data, ignoring the complex interdependencies and dynamic associations between complications, and failing to comprehensively capture the causal relationships between diseases and their evolution processes. This limitation makes it difficult for the model to reveal the interactions and triggering mechanisms between complications and unable to effectively predict the combined risks of multiple complications. At the same time, existing methods often focus on the prediction of a single complication, failing to fully consider the heterogeneity and synergistic effects between complications, resulting in insufficient ability to identify multiple comorbidity patterns and predict high-risk complication combinations. This restricts the accurate prediction of the future development of patients' conditions and weakens the implementation effect of personalized health management and precise intervention.

[0011] 2. Lack of real-time monitoring of dynamic disease progression

[0012] Many existing technologies focus on static data analysis, ignoring dynamic health trends, disease trajectories, and treatment outcomes. They fail to effectively utilize dynamic time series data and cannot track the disease trajectories and comorbidity evolution of rheumatoid arthritis patients in real time, resulting in insufficient timeliness and accuracy of treatment decisions. Summary of the Invention

[0013] The purpose of the present invention is to propose a prediction system for rheumatoid arthritis complications based on asymmetric causal associations in view of the deficiencies of the prior art. It aims to solve the challenges in capturing the complex correlations between complications, insufficient ability to predict comorbidity patterns, and dynamic disease progression in current prediction methods. By introducing asymmetric causal association analysis, this system can deeply reveal the triggering and dependence relationships between complications, especially in scenarios where there is a one-way causal impact between some comorbidities, such as the relationship between pulmonary hypertension and interstitial lung disease, thus making up for the deficiencies of existing methods in capturing complex disease correlations. By modeling time-series dynamic data, it accurately tracks the dynamic development of rheumatoid arthritis and solves the deficiency of dynamic prediction. By constructing a multi-dimensional cosine similarity matrix and hierarchical clustering analysis, the system can identify high-risk complication combinations and accurately predict comorbidity patterns, further enhancing the ability to predict comorbidity patterns. Combining these innovative technologies, this system significantly improves the accuracy and robustness of predicting rheumatoid arthritis complications and provides strong support for clinical health management and personalized treatment.

[0014] The purpose of the present invention is achieved through the following technical solutions: A prediction system for rheumatoid arthritis complications based on asymmetric causal associations, the system includes:

[0015] A clinical data preprocessing module for obtaining and preprocessing patients' clinical time series data;

[0016] A complication correlation modeling module for counting the number of occurrences of complication disease labels in clinical data and constructing a disease correlation matrix;

[0017] A complication causal association enhancement module, which is used to calculate the in-degree and out-degree of each pair of complication disease labels based on a disease correlation matrix, and construct a complication causal association enhancement matrix;

[0018] A comorbidity pattern recognition module, which is used to hierarchically cluster complication diseases based on the complication causal association enhancement matrix according to the degree of correlation and causal association, and obtain disease subgroups of complication disease combinations;

[0019] A complication prediction module, which is used to dynamically predict the complication risk based on the disease subgroups of complication disease combinations.

[0020] Furthermore, in the clinical data preprocessing module, the time series data is processed and the missing values in the time series data are filled by fusing time stamps, missing masks, and time intervals.

[0021] Furthermore, in the complication correlation modeling module, the conditional probability of two complication disease labels appearing simultaneously is calculated to construct a disease correlation matrix.

[0022] Furthermore, in the complication causal association enhancement module, for a pair of complication disease labels, intermediate labels are enumerated, the contributions of the intermediate labels to the two complication diseases are accumulated, the in-degree matrix is used to measure whether a disease is a prerequisite for the occurrence of another disease, and the out-degree matrix is used to measure the possibility of a disease triggering another disease.

[0023] Furthermore, the system further includes a causal analysis auxiliary compensation module, which is used to calculate the similarity between two complication disease labels, and combine it with the complication causal association enhancement matrix to obtain a comprehensive association matrix.

[0024] Furthermore, the comprehensive association matrix fuses the complication causal relationship and feature similarity information, provides an embedding representation for each complication disease label, and captures the complex associations between complication diseases.

[0025] Furthermore, in the comorbidity pattern recognition module, the average value of the causal path weights of the in-degree and out-degree of two complication disease labels is taken as the degree of causal association between the two complication diseases.

[0026] Furthermore, in the comorbidity pattern recognition module, according to a comprehensive index that combines the correlation and causal association degree of two complication diseases, the similarity degree of two complication disease clusters is judged, and a hierarchical structure is generated by gradually merging the most similar clusters.

[0027] Furthermore, based on the results of the comorbidity pattern recognition module, a hierarchical neural network is designed to predict the development trend of a specific disease combination.

[0028] Furthermore, in the complication prediction module, the dynamic prediction result is the confidence level of each complication disease, reflecting the likelihood of a rheumatoid arthritis patient having each complication at different time points.

[0029] Advantages of the present invention:

[0030] 1. Deep understanding of the associative synergy among complications: In the clinical practice of RA, patients often suffer from multiple complications, such as cardiovascular diseases, pulmonary diseases, and osteoporosis. There are complex triggering, dependency, and causal relationships among these complications, and existing technologies often struggle to comprehensively capture these associations. To more accurately predict the disease development trend and complication combinations, the present invention proposes a method for deeply capturing the interactions between complication diseases. By introducing asymmetric causal association analysis, it can reveal the complex interactions and potential disease combination patterns among different complications, comprehensively improving the accuracy and effectiveness of disease prediction. Through this method, doctors can better understand the evolution patterns of RA and its complications, thereby formulating more precise treatment plans.

[0031] 2. Enhancement of comorbidity pattern recognition and personalized treatment: The condition of RA patients often presents multiple complications simultaneously, and the prediction of a single complication cannot meet the clinical needs. Traditional technologies ignore the heterogeneity and synergy among complications, resulting in insufficient recognition of high-risk complication combinations. The present invention proposes a method for enhancing comorbidity pattern recognition and personalized treatment plans for patients. Through multi-dimensional causal analysis and hierarchical clustering techniques, it can effectively identify and predict the synergy and potential high-risk combinations among complications. This not only helps to early warn of possible health problems but also enables the customization of personalized health management plans for each patient, ensuring the effectiveness and safety of treatment.

[0032] 3. Real-time monitoring of disease progression and optimization of treatment decisions: RA is a chronic inflammatory disease, and its condition changes over time. Therefore, real-time monitoring of the patient's disease state is crucial. Traditional static data evaluation methods are difficult to reflect the true disease dynamics, affecting the timeliness and accuracy of treatment decisions. The present invention proposes a perspective of dynamic prediction, combining dynamic time series data and asymmetric causal association analysis, which can real-time monitor the evolution process of the disease and provide accurate predictions and interventions based on the changes in the patient's health status. Based on dynamic prediction, doctors can detect disease changes earlier and adjust treatment strategies in a timely manner, ensuring that each patient can obtain the most suitable treatment plan for the current situation.

[0033] 4. The present invention significantly improves the prediction accuracy, robustness, and timeliness of rheumatoid arthritis complications, providing strong support for clinical health management and personalized treatment. This complication prediction system has significant advantages in dealing with the prediction of rheumatoid arthritis complications. Compared with existing technical solutions, it can not only effectively process irregular time series data, but also capture the complex associations between complications, reveal comorbidity patterns, and improve the prediction accuracy of rare complications. In addition, this complication prediction system also has good interpretability and can provide valuable decision-making support for clinicians. Brief Description of the Drawings

[0034] To more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following described drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0035] Figure 1 It is a schematic diagram of a rheumatoid arthritis complication prediction system based on asymmetric causal association provided by the present invention.

[0036] Figure 2 It is a schematic diagram of enhanced causal association.

[0037] Figure 3 It is a schematic diagram of the comorbidity identification module. Detailed Embodiments

[0038] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0039] As Figure 1 shown, the present invention proposes a rheumatoid arthritis complication prediction system based on asymmetric causal association, introducing dynamic time series data and non-asymmetric causal relationship modeling between diseases to effectively capture the complexity of rheumatoid arthritis (RA) comorbidity and its dynamic change characteristics. And feature similarity analysis and cosine similarity matrix are used to make up for the information loss in causal relationship modeling, significantly improving the accuracy and robustness of the comorbidity prediction model.

[0040] This system specifically includes a clinical data preprocessing module, a complication correlation modeling module, a complication causal association enhancement module, a causal analysis auxiliary compensation module, a comorbidity pattern recognition module, and a complication prediction module. The specific functions of each module are as follows.

[0041] Clinical data preprocessing module. This module comprehensively processes and fills missing values in time-series data by fusing timestamps, missing masks, and time intervals. The integrated time-series data is more complete and stable, providing reliable support for clinical decision-making and subsequent data analysis.

[0042] Complication correlation modeling module. When constructing a rheumatoid arthritis complication prediction system, the associations between diseases are crucial. As the core pathology, synovitis affects multiple systems throughout the body through inflammatory mediators, such as damage to the cardio-pulmonary system. In addition, treatment drugs (such as immunosuppressants) can control the disease but may cause side effects in systems such as the cardiovascular, liver, and kidneys. The complex interactions between these systems make the correlations between complications particularly prominent. Therefore, modeling the correlations between complication diseases can accurately reflect clinical reality and thus improve the system's performance.

[0043] The present invention constructs a disease correlation matrix D by counting the occurrence times of disease labels in the data;

[0044]

[0045] where j and k respectively represent the labels of different complications. Count the occurrence times of complication disease label j, the co-occurrence times of complication disease label j and complication disease label k, and calculate the conditional probability of complication disease label k occurring when complication disease label j appears .

[0046] Modeling the correlations between complication diseases can more realistically reflect the complex pathology and treatment mechanisms in clinical practice by capturing the mutual influences and dependencies between different diseases, thus significantly improving the system's prediction performance.

[0047] Complication causal association enhancement module. Since there are often complex causal associations between different complications, the disease relationships exhibit directional characteristics. For example, there may be a frequent co-occurrence relationship between "pulmonary hypertension" and "interstitial lung disease", but this relationship is often one-way and asymmetric. "Pulmonary hypertension" may often accompany "interstitial lung disease", but the reverse is not necessarily true. As Figure 2 shown, in order to accurately capture this asymmetric disease label relationship, the present invention constructs a complication causal association enhancement matrix based on the proposed disease correlation matrix, and uses the matrix representations of in-degree and out-degree to reveal the mutual influences and potential comorbidity patterns between complication diseases, deepening the understanding of specific complication disease combinations.

[0048]

[0049]

[0050] For each pair of complication disease labels Enumerate the intermediate label u and calculate and respectively normalize the denominators to and For all labels in the v - index label set, accumulate the contributions of u to diseases j and k. The in - degree matrix measures whether disease k is a prerequisite for the occurrence of disease j, and the out - degree matrix measures the possibility of disease j causing disease k.

[0051] Causal analysis auxiliary compensation module. Relying solely on the in - degree and out - degree matrices may be difficult to comprehensively capture the combination of non - direct causal associations of complications, especially the complication associations that show strong statistical correlations but lack significant causal relationships. Therefore, a cosine similarity matrix is introduced as a similarity measurement method independent of data statistical characteristics. Cosine similarity can identify the similarity of complication diseases in a multi - dimensional feature space, and can recognize those complication relationships with similar features but may not be obvious in statistical frequencies. It provides auxiliary information when the causal relationship is insufficient to explain, thus effectively making up for the blind spots in the directional analysis.

[0052]

[0053] Among them, represent the feature vector representations of complication disease label j and complication disease label k respectively.

[0054] The final comprehensive association matrix M that fuses the complication causal relationship and feature similarity information is:

[0055]

[0056] is the weight matrix for linear transformation of the input features, which determines how to adjust the importance of the input features and how to project the input data into a new space or dimension. is a bias term used to adjust the result after linear transformation, helping the model capture more complexity and non - linear relationships.

[0057] The comprehensive association matrix M provides a final embedding representation for each label.

[0058] By fusing the causal relationship and feature similarity information, the comprehensive association matrix M comprehensively captures the complex associations between RA complications, significantly improves the performance of the prediction system, and supports more accurate clinical decisions and personalized treatment planning.

[0059] Comorbidity Pattern Recognition Module: To further support the complication prediction system in realizing complication prediction and deeply exploring disease association patterns, the present invention designs an auxiliary task module with a hierarchical disease relationship to identify comorbidity patterns and reveal high-risk complication combinations.

[0060] As Figure 3 shown, the comorbidity pattern recognition module first clusters highly correlated and highly causally associated complications into the same group by quantifying the correlation and causal relationship between diseases. This step can not only reveal the comorbidity characteristics among complications, but also further conduct subgroup analysis on patients to help identify the comorbidity patterns and their potential mechanisms within the subgroups, and reveal high-risk complication combinations.

[0061] The present invention defines a comprehensive index to describe the degree of correlation and causal association between diseases j and k:

[0062]

[0063] where and respectively represent the causal path weights from the in-degree and out-degree of the disease, and is a parameter between 0 and 1 used to adjust the relative importance between correlation and causal relationship.

[0064] In hierarchical clustering, the two most similar clusters of elements are merged:

[0065]

[0066] where respectively represent the clusters of complication disease labels j and k. A hierarchical structure is generated by gradually merging the most similar clusters, revealing different subgroups of complication combination characteristics and potential high-risk complication combinations.

[0067] Based on the clustering results of the comorbidity pattern recognition module, the present invention further designs a hierarchical neural network focusing on predicting the development trend of specific disease groups. This architecture not only helps to understand the interaction rules between diseases, but also provides a new perspective for the precise grouping of patients, making it possible to identify high-risk complication combinations at an early stage.

[0068] Complication Prediction Module: Based on the above module, for the common complications of RA patients, including 7 complications: interstitial lung disease, mental disease, pulmonary hypertension, infection, cardiovascular disease, skin ulcer, and myelosuppression, multi-label prediction is performed.

[0069] The present invention has achieved good prediction results on the data of the direct reporting project for rheumatoid arthritis in 274 rheumatology centers from 31 provinces in China. The accuracy rates for predicting 7 complications, namely interstitial lung disease, mental illness, pulmonary hypertension, infection, cardiovascular disease, skin ulcer, and myelosuppression, are 0.941, 0.976, 0.996, 0.998, 0.944, 0.992, and 0.997 respectively, and the recall rates are 0.763, 0.746, 0.728, 0.668, 0.805, 0.747, and 0.659 respectively.

[0070] The system of the present invention can dynamically predict the complication risks of RA patients and display the prediction results in real time. The displayed results are the confidence levels for each complication, which can intuitively reflect the likelihood of RA patients having various complications at different time nodes. Through this prediction system, clinicians can make more accurate personalized treatment decisions based on the real-time updated risk assessment, thereby optimizing patient management and improving the effectiveness of clinical intervention.

[0071] The above embodiments are used to explain the present invention rather than limit the present invention. Any modifications and changes made within the spirit and scope of the claims of the present invention fall within the protection scope of the present invention.

Claims

1. A rheumatoid arthritis complication prediction system based on asymmetric causal association, characterized in that, The system includes: A clinical data preprocessing module, which is used to obtain the patient's clinical time series data and perform preprocessing; A complication correlation modeling module, which is used to count the number of occurrences of complication disease labels in clinical data, calculate the conditional probability of two complication disease labels occurring simultaneously, and construct a disease correlation matrix; A complication causal association enhancement module, which is used to calculate the in-degree and out-degree of each pair of complication disease labels based on the disease correlation matrix, and construct a complication causal association enhancement matrix; for a pair of complication disease labels, enumerate intermediate labels, accumulate the contributions of the intermediate labels to the two complication diseases, use the in-degree matrix to measure whether one disease is a prerequisite for the occurrence of another disease, and use the out-degree matrix to measure the possibility of one disease causing another disease; A causal analysis auxiliary compensation module, which is used to calculate the similarity between two complication disease labels, and combine it with the complication causal association enhancement matrix to obtain a comprehensive association matrix; the comprehensive association matrix integrates the complication causal relationship and feature similarity information, provides an embedded representation for each complication disease label, and captures the complex associations between complication diseases; A comorbidity pattern recognition module, which is used to judge the similarity degree of two complication disease clusters based on the complication causal association enhancement matrix according to a comprehensive index that combines the correlation and causal association degree of two complication diseases, generate a hierarchy by gradually merging the most similar clusters, perform hierarchical clustering on the complication diseases, and obtain disease subgroups of complication disease combinations; A complication prediction module, which is used to dynamically predict the complication risk based on the disease subgroups of complication disease combinations.

2. The rheumatoid arthritis complication prediction system based on asymmetric causal association according to claim 1, wherein In the clinical data preprocessing module, the missing values in the time series data are processed and filled by fusing time stamps, missing masks, and time intervals.

3. The rheumatoid arthritis complication prediction system based on asymmetric causal association according to claim 1, characterized in that In the comorbidity pattern recognition module, the average value of the causal path weights of the disease in-degree and out-degree of two complication disease labels is taken as the causal association degree of the two complication diseases.

4. The rheumatoid arthritis complication prediction system based on asymmetric causal association according to claim 1, wherein Based on the results of the comorbidity pattern recognition module, a hierarchical neural network is designed to predict the development trend of specific disease combinations.

5. The rheumatoid arthritis complication prediction system based on asymmetric causal association according to claim 1, wherein In the complication prediction module, the dynamic prediction result is the confidence level of each complication disease, which reflects the possibility of a rheumatoid arthritis patient having each complication at different time nodes.

Images