Patient life quality assessment analysis method and system based on patient report

Abstract

The invention provides a patient life quality assessment analysis method and system based on a patient report, and relates to the technical field of data assessment management, the patient life quality assessment analysis method based on the patient report comprises the following steps: screening out a patient report outcome symptom by using pre-collected symptom information of a peripheral artery disease patient, constructing a multi-dimensional life quality index scale; based on the multi-dimensional life quality index scale, comprehensive risks of the life quality of the patient under different symptoms are calculated through a dynamic symptom tree model in combination with a Markov method; and evaluating the importance of each symptom on the life quality, and selecting an adjustment measure according to an importance evaluation result in combination with a pre-configured adjustment measure. The comprehensive risk of the life quality of the patient under different symptoms can be dynamically calculated, a doctor can track the disease change of the patient in real time, a treatment scheme can be adjusted in time, and the risk of decline of the life quality of the patient is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of data evaluation management. Specifically, it particularly relates to a method and system for evaluating and analyzing the quality of life of patients based on patient reports. Background Art

[0002] Peripheral arterial disease (PAD) is a common manifestation of systemic atherosclerosis, referring to a vascular disease with stenosis or occlusion of arteries in the upper or lower extremities, mostly referring to lower extremity arterial diseases, which can be manifested as calf pain, intermittent claudication, rest pain, ulcers, and gangrene. PAD is common in lower extremity arteriosclerosis obliterans, arterial embolism, diabetic foot, thromboangiitis obliterans, etc. The latest guidelines indicate that an ankle-brachial index (ABI) ≤ 0.90 is sensitive and specific for arterial stenosis / occlusion and is used to diagnose lower extremity PAD.

[0003] Currently, the main treatment methods for PAD include: controlling risk factors (such as diabetes, hypertension, smoking, etc.), drug treatment, surgical revascularization, and endovascular intervention (angioplasty or endovascular stent implantation) to improve limb ischemia symptoms. However, except for correcting risk factors, it is not clear whether specific treatments can improve functional levels and quality of life. A review of the existing literature shows that although the graft vessels are patent after surgical revascularization or interventional treatment, and the survival rate without amputation and the limb salvage rate are reasonable, the quality of life of patients after limb salvage intervention is not always significantly improved. Further research is needed to use quality of life tools in a measurable and clinically relevant manner to guide optimal quality care and maximize patient-centered outcomes. "Three points of treatment, seven points of care", high-quality care before and after surgery is crucial for the daily life of patients, especially for improving the quality of life.

[0004] Patient-reported Outcomes (PROs) is a broad term, which is characterized by starting from the patient's own feelings and interests, and collecting information related to the patient's own health during the illness through methods such as patient oral history, doctor-patient discussion, patient interview, questionnaire or scale survey, and summarizing the patient's feelings and evaluations about diseases, treatments, care, health policies, etc.

[0005] Quality of life (QOL) is officially defined as "the impact of a disease or its treatment on an individual's usual or expected physical, emotional, and social well-being. It is also called health-related quality of life (HRQOL), which is a subjective evaluation by patients of the impact of a disease or its treatment on themselves. Current quality of life assessment techniques are mostly limited to a single dimension, such as only focusing on physical symptoms while ignoring the comprehensive impacts on the psychological, emotional, and social aspects. This single-dimensional assessment method cannot comprehensively reflect the multi-faceted impacts of a disease on a patient's quality of life, resulting in inaccurate and incomplete assessment results.

[0006] Regarding the problems in the related technologies, no effective solutions have been proposed yet. Summary of the Invention

[0007] In view of this, the present invention provides a method and system for assessing and analyzing a patient's quality of life based on patient reports to solve the above-mentioned problems.

[0008] To solve the above problems, the specific technical solutions adopted by the present invention are as follows: According to one aspect of the present invention, there is provided a method for assessing and analyzing a patient's quality of life based on patient reports. The method for assessing and analyzing a patient's quality of life based on patient reports includes the following steps: S1. Use the symptom information of peripheral artery disease patients collected in advance to screen out patient-reported outcome symptoms, and based on the patient-reported outcome symptoms, construct a multi-dimensional quality of life index scale through symptom chain analysis; S2. Based on the multi-dimensional quality of life index scale, calculate the comprehensive risk of the patient's quality of life under different symptoms through a dynamic symptom tree model and in combination with the Markov method; S3. Based on the comprehensive risk of the patient's quality of life under different symptoms, evaluate the importance of each symptom to the quality of life, and select adjustment measures according to the importance evaluation results and in combination with the pre-configured adjustment measures; Preferably, using the symptom information of peripheral artery disease patients collected in advance to screen out patient-reported outcome symptoms, and constructing a multi-dimensional quality of life index scale through symptom chain analysis includes the following steps: S11. Sort and classify the symptom information of peripheral artery disease patients collected, screen out patient-reported outcome symptoms, and use the Delphi entropy method to discriminate and extract the screened patient-reported outcome symptoms to obtain an initial symptom data set; S12. Analyze the co-occurrence pattern, time series relationship, and severity change of each symptom data in the initial symptom data set, and construct a symptom chain network model according to the analysis results; S13. Use the fuzzy c-means clustering method weighted by knowledge amount to perform clustering analysis on the symptom nodes in the symptom linkage network model, and establish a multi-dimensional quality of life index scale according to the clustering results.

[0009] Preferably, analyze the co-occurrence pattern, temporal relationship, and severity change of each symptom data in the initial symptom dataset, and construct a symptom linkage network model according to the analysis results, including the following steps: S121. Preprocess each symptom data in the initial symptom dataset, and the preprocessing includes data cleaning and data standardization processing; S122. According to the preprocessed symptom data, calculate the co-occurrence pattern index, temporal relationship index, and severity change index of each symptom data respectively; S123. Calculate the relevance index of each symptom data according to the co-occurrence pattern index, temporal relationship index, and severity change index of each symptom data; S124. Select the symptom data that meets the relevance threshold according to the relevance index of each symptom data, use the symptoms corresponding to the selected symptom data as symptom nodes, and construct a symptom linkage network model according to the relevance index of each symptom data.

[0010] Preferably, calculating the co-occurrence pattern index, temporal relationship index, and severity change index of each symptom data respectively according to the preprocessed symptom data includes the following steps: S1221. Construct a co-occurrence matrix according to the preprocessed symptom data, and calculate the collinear pattern index of each symptom data based on the co-occurrence matrix; S1222. Construct the time series data of the symptom data according to the preprocessed symptom data, and calculate the temporal relationship index of each symptom data through the dynamic time warping algorithm; S1223. Quantify the severity of each symptom data based on the pre-configured severity score, and calculate the severity change index of each symptom data through regression analysis and mutual information.

[0011] Preferably, using the fuzzy c-means clustering method weighted by knowledge amount to perform clustering analysis on the symptom nodes in the symptom linkage network model, and establishing a multi-dimensional quality of life index scale according to the clustering results includes the following steps: S131. Intuitionistically fuzzify the symptom data of the symptom nodes through the membership degree generation function to obtain an intuitionistic fuzzy dataset; S132. Use the improved intuitionistic fuzzy knowledge measure model to calculate the knowledge amount of each symptom node in the intuitionistic fuzzy dataset, and assign weights to the knowledge amount of each symptom according to the preset weight assignment function; S133. Use the knowledge quantity of each symptom to assign weights to the intuitionistic fuzzy data set to obtain a weighted data set; S134. Use the weighted data set to select the initial clustering centers based on kernel space density and kernel distance, and obtain the optimal fuzzy partition matrix and clustering centers through iterative optimization according to the inter-class sample distance and the principle of minimum knowledge quantity; S135. Construct a multi-dimensional quality of life index scale according to the optimal fuzzy partition matrix and clustering centers.

[0012] Preferably, the expression of the improved intuitionistic fuzzy knowledge measure model is: ; In the formula, G(A) represents the knowledge quantity of the symptom nodes in the intuitionistic fuzzy data set; n represents the number of symptom nodes in the intuitionistic fuzzy data set; u A (x i ) represents the membership degree of the i-th symptom node to the intuitionistic fuzzy data set; v A (x i ) represents the non-membership degree of the i-th symptom node to the intuitionistic fuzzy data set.

[0013] Preferably, the expression of the preset weight assignment function is: ; In the formula, w j represents the weight of the j-th symptom node in the intuitionistic fuzzy data set; G(A tj ) represents the knowledge quantity of the j-th symptom node in the t-th intuitionistic fuzzy data set; m represents the number of intuitionistic fuzzy data sets; n represents the number of symptom nodes in the intuitionistic fuzzy data set.

[0014] Preferably, based on the multi-dimensional quality of life index scale, calculating the comprehensive risk of the patient's quality of life under different symptoms through the dynamic symptom tree model and combining with the Markov method includes the following steps: S21. Based on the multi-dimensional quality of life index scale, configure the dynamic symptom tree parameters and establish a dynamic symptom tree model. The dynamic symptom tree parameters include basic events, intermediate events, sequence-related gates and top events; S22. Modularize the dynamic symptom tree model to obtain several modular dynamic symptom trees; and calculate the risk values of the modular dynamic symptom trees using the Markov method; S23. According to the risk values of each module's dynamic subtree, and in accordance with different combinations of symptom chains, summarize the risk values of each dynamic subtree to obtain the comprehensive risk of the patient's quality of life.

[0015] Preferably, modularize the dynamic symptom tree model to obtain several module dynamic symptom trees; and the steps for calculating the risk value of the module dynamic symptom tree using the Markov method include the following: S221. Divide the dynamic symptom tree model into several dynamic subtrees, and use the top event of the dynamic subtree as the basic bottom event to obtain several module dynamic symptom trees; S222. For the basic events and intermediate events in the module dynamic symptom tree, take each basic event and intermediate event as the state of the Markov chain, and configure the state transition probability using the memoryless property of the Markov chain; S223. According to the state transition probability, calculate the probability of each module dynamic symptom tree from the initial state to the top event state; S224. In each module dynamic symptom tree, evaluate the degree of influence of the top event on the patient's quality of life, and calculate the risk value of each module dynamic symptom tree based on the degree of influence and the probability of each module dynamic symptom tree from the initial state to the top event state.

[0016] Preferably, based on the comprehensive risk of the patient's quality of life under different symptoms, evaluate the importance of each symptom to the quality of life, and select adjustment measures according to the importance evaluation results and in combination with pre-configured adjustment measures, including the following steps: S31. Based on the comprehensive risk of the patient's quality of life under different symptoms, calculate the contribution of each symptom to the comprehensive risk; S32. In combination with the contribution of the comprehensive risk, calculate the importance of each symptom through the proportion of symptom risk contribution; S33. Sort the symptoms according to the importance results of the symptoms, and identify the core symptoms that affect the quality of life; S34. In combination with the pre-configured adjustment measures, select adjustment measures using the identified core symptoms and formulate a personalized intervention plan.

[0017] According to another aspect of the present invention, there is provided a patient quality of life assessment and analysis system based on patient reports, and the patient quality of life assessment and analysis system based on patient reports includes: an index scale construction module, a comprehensive risk calculation module, and an importance evaluation module, and the index scale construction module, the comprehensive risk calculation module, and the importance evaluation module are connected in sequence; The index scale construction module is used to screen out patient-reported outcome symptoms using the symptom information of peripheral artery disease patients collected in advance, and construct a multi-dimensional quality of life index scale through symptom chain analysis based on the patient-reported outcome symptoms; A comprehensive risk calculation module, which is used to calculate the comprehensive risk of the patient's quality of life under different symptoms based on a multi-dimensional quality of life index scale, through a dynamic symptom tree model and in combination with the Markov method; An importance evaluation module, which is used to evaluate the importance of each symptom to the quality of life based on the comprehensive risk of the patient's quality of life under different symptoms, and select adjustment measures according to the importance evaluation results and in combination with pre-configured adjustment measures.

[0018] The beneficial effects of the present invention are as follows: 1. By collecting the symptom information reported by the patient himself, the present invention can more accurately reflect the individual differences and subjective feelings of the patient. By constructing a multi-dimensional quality of life index scale through the symptom chain analysis method, it can comprehensively evaluate multiple aspects such as the patient's physical, psychological, and social functions, which helps doctors better understand the patient's health status, identify various factors affecting the quality of life, and thus formulate a more comprehensive treatment plan. By using a dynamic symptom tree model and in combination with the Markov method, it can dynamically calculate the comprehensive risk of the patient's quality of life under different symptoms, which helps doctors track the patient's condition changes in real time, adjust the treatment plan in time, and reduce the risk of the patient's quality of life decline.

[0019] 2. By calculating the co-occurrence pattern, time series relationship, and severity change index, the present invention deeply reveals the complex relationships between symptoms, improves the accuracy of analysis. By calculating the correlation index of symptom data, it constructs a symptom chain network model, intuitively displays the mutual influence and correlation relationships between symptoms, which helps doctors better understand the development process of the disease and formulate a more accurate treatment plan. By calculating the knowledge amount of each symptom node and performing weighted processing according to the preset weight distribution function, it ensures that the influence of important symptom nodes is fully reflected in the clustering analysis process, can provide scientific, accurate, and comprehensive decision-making support for doctors, and helps improve the quality of life of patients with peripheral artery disease.

[0020] 3. The present invention modularizes the dynamic symptom tree model to obtain several modular dynamic symptom trees, making the model more flexible and easy to manage. By using the memoryless property of the Markov chain to configure the state transition probability, it can accurately calculate the probability of each modular dynamic symptom tree from the initial state to the top event state. By calculating the importance of each symptom based on the proportion of symptom risk contribution and sorting the symptoms, it can identify the core symptoms that have the greatest impact on the quality of life, providing a key direction for the formulation of intervention measures. Description of the Drawings

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required in the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings. In the drawings: Figure 1 is a flowchart of a method for evaluating and analyzing the quality of life of patients based on patient reports according to an embodiment of the present invention; Figure 2 is a schematic block diagram of a system for evaluating and analyzing the quality of life of patients based on patient reports according to an embodiment of the present invention.

[0022] In the figure: 1. Scaling table construction module; 2. Comprehensive risk calculation module; 3. Importance evaluation module. Specific implementation manners

[0023] In order to enable those skilled in the art of this technology to better understand the technical solutions in this application, the following will clearly and completely describe the technical solutions in the embodiments of this application in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this application.

[0024] According to an embodiment of the present invention, a method and system for evaluating and analyzing the quality of life of patients based on patient reports are provided.

[0025] Now, the present invention will be further described in conjunction with the accompanying drawings and specific implementation manners. As Figure 1 shown, according to an embodiment of the present invention, a method for evaluating and analyzing the quality of life of patients based on patient reports is provided. The method for evaluating and analyzing the quality of life of patients based on patient reports includes the following steps: S1. Screen out the patient-reported outcome symptoms using the symptom information of patients with peripheral artery disease collected in advance, and construct a multi-dimensional quality of life index scale through symptom chain analysis according to the patient-reported outcome symptoms; As a preferred implementation manner, screening out the patient-reported outcome symptoms using the symptom information of patients with peripheral artery disease collected in advance, and constructing a multi-dimensional quality of life index scale through symptom chain analysis includes the following steps: S11. Sort and classify the symptom information of the patients with peripheral artery disease collected, screen out the patient-reported outcome symptoms, and use the Delphi entropy method to discriminate and extract the screened patient-reported outcome symptoms to obtain an initial symptom data set; It should be noted that the Delphi entropy method is a method for determining weights based on information entropy. It determines the weights of symptoms according to the degree of variation and the amount of information of the symptoms. The greater the degree of variation and the more information of a symptom, the greater its weight.

[0026] Specifically, the process of sorting and classifying the symptom information of peripheral artery disease patients collected, screening out the patient-reported outcome symptoms, and using the Delphi entropy method to discriminate and extract the screened symptoms to obtain the initial symptom dataset includes the following steps: a. Data collection Data source: Collect the symptom information of peripheral artery disease patients from multiple channels such as medical records, patient self-reports, questionnaires, and interviews.

[0027] Data scope: Ensure that the collected symptom information comprehensively covers multiple aspects such as the patient's physiology, psychology, and social function.

[0028] b. Data sorting and classification Duplicate removal and cleaning: Perform duplicate removal on the collected symptom information, delete duplicate or redundant data, and ensure the uniqueness of each piece of symptom information. At the same time, clean the data and correct incorrect or inconsistent information.

[0029] Classification coding: Classify and code the symptoms according to the nature and characteristics of the symptoms for subsequent data processing and analysis. For example, symptoms can be classified into pain type, numbness type, difficulty walking type, psychological symptom type, etc.

[0030] c. Screening of patient-reported outcome symptoms Define outcome symptoms: Clearly define the patient-reported outcome symptoms, which are usually the symptoms that patients are most concerned about and have the greatest impact on quality of life. For example, difficulty walking, unrelieved pain, severe psychological depression, etc.

[0031] Screening method: Screen out the symptoms that meet the definition of outcome symptoms by reading patient reports, medical records, etc. Methods such as keyword search and symptom scoring can be used to assist in screening.

[0032] d. Discrimination and extraction by Delphi entropy method Construct a symptom evaluation matrix: Use the screened patient-reported outcome symptoms as the evaluation objects to construct a symptom evaluation matrix. The rows of the matrix represent patients, and the columns represent symptoms.

[0033] Standardization processing: Perform standardization processing on the symptom evaluation matrix to eliminate the influence of different symptom dimensions and magnitudes.

[0034] Calculate the information entropy: According to the standardized symptom evaluation matrix, calculate the information entropy of each symptom. The smaller the information entropy, the greater the degree of variation of the symptom and the more information it provides.

[0035] Calculate the weights: Calculate the weights of each symptom based on the information entropy. The greater the weight, the higher the importance of the symptom in the whole.

[0036] e. Extract the initial symptom dataset: According to the calculated weights, sort the selected patient-reported outcome symptoms, extract the symptoms with larger weights, and form the initial symptom dataset. These symptoms will serve as the basis for subsequent symptom chain analysis and the construction of the quality of life index scale.

[0037] S12. Analyze the co-occurrence pattern, temporal relationship, and severity change of each symptom data in the initial symptom dataset, and construct a symptom chain network model based on the analysis results; As a preferred implementation, analyzing the co-occurrence pattern, temporal relationship, and severity change of each symptom data in the initial symptom dataset, and constructing a symptom chain network model based on the analysis results includes the following steps: S121. Preprocess each symptom data in the initial symptom dataset. The preprocessing includes data cleaning and data standardization processing; It should be noted that data cleaning is to clean each symptom data in the initial symptom dataset, including removing duplicate data, handling missing values, correcting incorrect data, etc., to ensure the accuracy and consistency of the data.

[0038] Data standardization processing is to standardize the cleaned data so that data of different symptoms can be compared under the same dimension. This can be achieved through methods such as normalization and standardization.

[0039] S122. Calculate the co-occurrence pattern index, temporal relationship index, and severity change index of each symptom data according to the preprocessed symptom data; As a preferred implementation, calculating the co-occurrence pattern index, temporal relationship index, and severity change index of each symptom data according to the preprocessed symptom data includes the following steps: S1221. Construct a co-occurrence matrix according to the preprocessed symptom data, and calculate the collinear pattern index of each symptom data based on the co-occurrence matrix; Specifically, according to the preprocessed symptom data, list all symptoms. For each pair of symptoms, count the number of times they appear simultaneously in the same record. This can be completed by traversing the data of all patients. Create an h×h matrix C, where Cij represents the number of times symptom Si and symptom Sj appear simultaneously.

[0040] The collinear mode index is used to quantify the co-occurrence intensity and relevance between symptoms. Commonly used indices include co-occurrence frequency, which converts the number of co-occurrences into co-occurrence frequency. Different methods can be used for normalization, such as simple frequency or conditional probability. Based on the co-occurrence frequency, the co-occurrence mode index of each symptom with other symptoms is calculated.

[0041] S1222. Construct time series data of symptom data based on the preprocessed symptom data, and calculate the temporal relationship index of each symptom data through the dynamic time warping algorithm. Specifically, time series data is a set of data points arranged in chronological order. In symptom analysis, each time point corresponds to the state of one or more symptoms (such as presence, absence, severity, etc.). Extract the timestamps and corresponding symptom states from the preprocessed symptom data. For each symptom, combine its timestamp and corresponding state into a time series.

[0042] The dynamic time warping algorithm (DTW) is used to measure the similarity between two time series, even if they are not perfectly aligned on the time axis. It calculates the distance between them by finding the optimal alignment path between the two sequences. The specific steps are as follows: For two time series Q=(q1,q2,…,qr) and P=(p1,p2,…,pe), construct an r×e distance matrix D, where Dij represents the distance between qi and pj (usually using the Euclidean distance).

[0043] Starting from the distance matrix D, gradually calculate the cumulative distance matrix C, where Cij represents the minimum cumulative distance from (1,1) to (i,j).

[0044] Find the optimal path from (1,1) to (r,e) in C, which minimizes the cumulative distance. The cumulative distance on the optimal path is the DTW distance, representing the similarity between the two time series. A smaller DTW distance indicates a stronger temporal correlation.

[0045] S1223. Quantify the severity of each symptom data based on the preconfigured severity score, and calculate the severity change index of each symptom data through regression analysis and mutual information.

[0046] Specifically, a severity score needs to be preconfigured before quantifying the severity of each symptom. Each symptom is assigned a severity score based on its impact on the patient's quality of life, which may be based on the judgment of clinicians, medical record data, or patient self-report.

[0047] Range of severity score: Generally, the severity score will adopt a numerical interval, such as 0-10 points, where 0 indicates no symptoms and 10 indicates extremely severe symptoms.

[0048] Suppose several typical symptoms of patients with peripheral artery disease are being evaluated: leg pain, difficulty walking, and fatigue.

[0049] Leg pain: 7 points (severely affecting walking and daily activities); Difficulty walking: 8 points (relatively high severity as the patient cannot move normally); Fatigue: 5 points (moderate impact, mainly worsening after the patient's activities); Regression analysis is a statistical method aimed at predicting the change of a dependent variable through one or more independent variables. Here, regression analysis is used to find out the trend of symptom severity change over time or some other factors (such as treatment, daily activities, etc.). Regression analysis can reveal how symptoms change over time, or the change pattern of symptoms when patients receive different treatment interventions. For example, whether pain decreases with the increase of a certain drug treatment, or whether fatigue intensifies with the increase of the patient's exercise volume.

[0050] For example, perform a regression analysis on the leg pain symptom to study how it changes after the patient receives drug treatment. For instance, before treatment, the leg pain score was 7.

[0051] In the next four weeks, data was collected, and the leg pain scores for each week were: 7, 6, 5, 3. Through regression analysis, it was found that the severity of pain is negatively correlated with the treatment time, and the pain gradually decreases as the treatment time increases.

[0052] Among them, mutual information is a statistical tool for measuring the dependence relationship between variables, and is often used to measure the degree of information sharing between two random variables. In the present invention, mutual information is used to evaluate the correlation between the severity changes of different symptoms and find out whether there is a potential association or mutual influence between certain symptoms.

[0053] Suppose we want to know whether the severity changes of difficulty walking and fatigue are related. If the mutual information value between the two is high, it means that their change trends have a strong dependence: Difficulty walking score: 8 (week 1), 7 (week 2), 6 (week 3), 5 (week 4); Fatigue score: 5 (week 1), 6 (week 2), 7 (week 3), 8 (week 4); By calculating the mutual information, it is found that the change trends of the severity of these two symptoms are correlated: as fatigue increases, difficulty walking also gradually worsens. This may prompt the doctor to pay attention to the intervention of difficulty walking while treating fatigue.

[0054] Through regression analysis and mutual information, a severity change index can be generated, which quantifies how the severity of each symptom changes and, in combination with the associations of other symptoms, identifies potential trends or patterns.

[0055] S123. Calculate the correlation index for each symptom data based on the co-occurrence pattern index, temporal relationship index, and severity change index of each symptom data; Specifically, after having the co-occurrence pattern, temporal relationship, and severity change indices, these information can be integrated into a correlation index for measuring the overall correlation between symptoms. Generally, these three indices can be combined through weighting or other combination methods: Correlation index = ω1 × Co-occurrence pattern index + ω2 × Temporal relationship index + ω3 × Severity change index; where ω1, ω2, and ω3 are weights set according to specific application scenarios for adjusting the influence degree of each index.

[0056] S124. Select the symptom data that meet the correlation threshold according to the correlation index of each symptom data, use the symptoms corresponding to the selected symptom data as symptom nodes, and construct a symptom chain network model based on the correlation index of each symptom data.

[0057] It should be noted that the calculated correlation index is used to screen the symptom data and construct a symptom chain network model to help understand the mutual relationships and mutual influences between symptoms. In this way, the key symptom chains can be extracted from the complex symptom data, facilitating subsequent risk assessment and the selection of intervention measures.

[0058] Among them, the correlation threshold is a preset value used to determine which symptoms have a strong enough correlation to be part of the symptom chain. The threshold can be set according to specific application scenarios or clinical experience. For example, the threshold can be 1.5. If the correlation index of a pair of symptoms is greater than this value, they will be regarded as a pair of symptoms with a strong correlation.

[0059] Suppose the correlation index of a pair of symptoms (such as pain and difficulty walking) is 2.1, and the set threshold is 1.5, then this pair of symptoms meets the threshold requirement.

[0060] If the correlation index of another pair of symptoms (such as fatigue and dizziness) is 1.2, then it is below the threshold and does not meet the requirement, so it will not be selected into the symptom chain.

[0061] For those symptom data screened by correlation, the corresponding symptoms need to be used as symptom nodes. Each node represents a specific symptom, and these symptom nodes are connected to each other through their correlation indices to form a symptom chain network.

[0062] Definition of symptom nodes: Each symptom node represents a specific symptom of a patient, such as pain, fatigue, difficulty walking, etc.

[0063] The connection weights between symptom nodes are represented by their correlation indices. The larger the weight, the stronger the correlation between the two symptoms.

[0064] By connecting all symptom nodes that meet the threshold requirements and their correlations, a symptom chain network model is constructed. This network model can visually display the mutual relationships between various symptoms, helping doctors or researchers better understand how symptoms affect each other.

[0065] Specifically, the characteristics of the symptom chain network model: Nodes: Each node represents a symptom.

[0066] Edges: Each edge represents the correlation between two symptoms, and the weight of the edge is the value of the symptom correlation index.

[0067] Network topology: According to the correlations between symptoms, the structure of the network can be tightly connected (indicating a high correlation between symptoms) or sparse (low correlation between symptoms).

[0068] S13. Use the knowledge quantity weighted fuzzy c-means clustering method to perform clustering analysis on the symptom nodes in the symptom chain network model, and establish a multi-dimensional quality of life index scale based on the clustering results.

[0069] As a preferred implementation, using the knowledge quantity weighted fuzzy c-means clustering method to perform clustering analysis on the symptom nodes in the symptom chain network model, and establishing a multi-dimensional quality of life index scale based on the clustering results includes the following steps: S131. Intuitionistically fuzzify the symptom data of the symptom nodes through the membership function to obtain an intuitionistic fuzzy data set; It should be noted that the intuitionistic fuzzification process is a process of converting the symptom data of the symptom nodes into an intuitionistic fuzzy set. Through the membership function, each symptom data point is assigned a membership degree (indicating the degree to which the symptom belongs to a certain fuzzy set) and a non-membership degree (indicating the degree to which the symptom does not belong to the fuzzy set); The membership function is a mathematical function used to calculate the relationship between symptom data and fuzzy sets. It converts the symptom data (such as severity, frequency, etc.) of each symptom node into a membership degree and a non-membership degree.

[0070] Membership degree u A (q i):Indicates the membership degree of the symptom node to a fuzzy set A, with a value between 0 and 1. The larger the value, the more likely the symptom q i is to belong to this fuzzy set.

[0071] Non-membership degree v A (q i ):Indicates the symptom node q i 's non-membership degree to the fuzzy set A, with a value also between 0 and 1, indicating the degree to which the symptom q i does not belong to this fuzzy set.

[0072] By fuzzifying the symptom data of all symptom nodes, an intuitionistic fuzzy data set can be constructed. In this data set, each symptom node corresponds to a set of membership degrees and non-membership degrees, used to represent its fuzziness in different fuzzy sets.

[0073] S132. Calculate the knowledge amount of each symptom node in the intuitionistic fuzzy data set using the improved intuitionistic fuzzy knowledge measure model, and assign weights to the knowledge amount of each symptom according to the preset weight assignment function; It should be noted that calculating the knowledge amount of each symptom node in the intuitionistic fuzzy data set is to measure the importance of this node to the entire data set. The calculation of the knowledge amount is based on the membership degree and non-membership degree of the symptom node, and reflects the information richness of this symptom in the network.

[0074] As a preferred implementation manner, the expression of the improved intuitionistic fuzzy knowledge measure model is: ; In the formula, G(A) represents the knowledge amount of the symptom node in the intuitionistic fuzzy data set; n represents the number of symptom nodes in the intuitionistic fuzzy data set; u A (q i ) represents the membership degree of the i-th symptom node to the intuitionistic fuzzy data set; v A (q i ) represents the non-membership degree of the i-th symptom node to the intuitionistic fuzzy data set.

[0075] As a preferred implementation manner, the expression of the preset weight assignment function is: ; In the formula, w j represents the weight of the j-th symptom node in the intuitionistic fuzzy data set; G(A tj ) represents the knowledge amount of the j-th symptom node in the t-th intuitionistic fuzzy data set; m represents the number of intuitionistic fuzzy data sets; n represents the number of symptom nodes in the intuitionistic fuzzy dataset.

[0076] S133. Perform weighted processing on the intuitionistic fuzzy dataset by allocating weights using the knowledge amount of each symptom to obtain a weighted dataset; It should be noted that adjusting the membership degree and non - membership degree in the intuitionistic fuzzy dataset according to the weight of each symptom node can be achieved by multiplying the membership degree and non - membership degree of each symptom node by its corresponding weight.

[0077] S134. Using the weighted dataset, select the initial clustering centers based on kernel space density and kernel distance, and through iterative optimization according to the inter - class sample distance and the principle of minimum knowledge amount, obtain the optimal fuzzy partition matrix and clustering centers; It should be noted that in clustering analysis, choosing a suitable initial clustering center is crucial. The kernel method helps to better select clustering centers by mapping data from a low - dimensional space to a high - dimensional space to handle non - linear relationships.

[0078] The kernel space density reflects the density value of a certain symptom node (data point) in the high - dimensional space. It indicates how many data points are dense near a certain point.

[0079] The kernel distance is a distance metric between symptom nodes in the high - dimensional space. By calculating the kernel distance between data points.

[0080] Among them, the steps of selecting the initial clustering centers include: Calculate the kernel space density: Perform kernel function transformation on the weighted dataset and calculate the space density value of each symptom node. A higher density value means that the data around this point is more dense.

[0081] Calculate the kernel distance: Calculate the kernel distance between data points to ensure that the selected clustering centers are far away from each other in space.

[0082] Select the clustering centers: First, select the data point with the highest kernel density as the first clustering center, and then select other points with the maximum distance from the existing clustering centers as additional clustering centers.

[0083] In addition, the inter - class sample distance is used to measure the relative distance of sample points between different clusters. The purpose is to make the distance between points in different categories as large as possible, so as to ensure good discrimination between categories. When the inter - class distance is too small, it may lead to confusion between different categories. Therefore, it is hoped to maximize the inter - class sample distance during the clustering optimization process.

[0084] The amount of knowledge is a measure of the importance of symptom nodes in the intuitionistic fuzzy dataset. The principle of minimum knowledge amount requires that in each iteration, the nodes included in each cluster of the clustering should have at least a certain amount of knowledge to ensure that each cluster contains valuable symptom nodes and there will be no invalid or noisy clusters. During the optimization process, the clusters containing nodes with a higher amount of knowledge are preferentially retained to avoid some categories in the clustering results being too sparse or insignificant.

[0085] Once the initial clustering centers are selected, the clustering results are then improved through an iterative optimization algorithm. Here, the fuzzy c-means clustering algorithm is used and optimized in combination with the amount of knowledge. The fuzzy partition matrix represents the membership degrees of each symptom node (data point) to each clustering center. Each element value in the matrix ranges from 0 to 1, indicating the degree to which the symptom node belongs to a certain clustering center.

[0086] The iterative optimization steps include: Calculate the distances from each data point to all clustering centers and update the fuzzy membership matrix according to the distances. The closer the symptom node is to its corresponding clustering center, the higher its membership degree.

[0087] Update the clustering centers: According to the current membership matrix, recalculate the positions of the clustering centers to ensure that the clustering centers continuously move towards the optimal positions.

[0088] Knowledge amount constraint: In each iterative update, utilize the principle of minimum knowledge amount to ensure that the knowledge amounts of the nodes in each cluster meet certain requirements and prevent the occurrence of empty clusters or clusters with a low amount of knowledge.

[0089] Check for convergence: Continue the iteration until the positions of the clustering centers and the membership matrix no longer change significantly, indicating that the clustering converges and the optimal fuzzy partition matrix and clustering centers are obtained.

[0090] S135. Construct a multi-dimensional quality of life index scale based on the optimal fuzzy partition matrix and clustering centers.

[0091] It should be noted that according to the different characteristics of the clustering centers, each clustering center can be classified into different dimensions related to the quality of life. Common dimensions of the quality of life include: Physical dimension: The symptoms may focus on physical fitness, pain, activity limitation, etc.

[0092] Psychological dimension: It includes symptoms related to mental state such as emotions, anxiety, depression, etc.

[0093] Social dimension: It reflects the symptoms in aspects such as the patient's social interaction ability, work ability, interaction with others, etc.

[0094] Functional dimension: Symptoms related to daily activities, work ability, or the ability to live independently.

[0095] Correlate the cluster centers with specific quality of life dimensions. This can be accomplished through further analysis of the clustering results, such as observing the characteristics of symptom nodes in each cluster and identifying the main dimensions. Symptom nodes in a cluster may exhibit similar patterns, which can help classify them into a certain dimension. For example: If a large number of symptoms related to pain, fatigue, muscle weakness, etc. are clustered around a certain cluster center, it can be classified into the physical dimension.

[0096] Another cluster center may contain a large number of symptoms related to depression, anxiety, mood swings, etc., and is classified into the psychological dimension.

[0097] Once the correspondence between the cluster centers and the quality of life dimensions is completed, the next step is to construct a multi-dimensional quality of life index scale based on the clustering results. Each dimension will have a set of symptoms as its indicators, and users can evaluate the patient's quality of life based on these indicators. Each indicator can be quantified through a scoring system, such as 1 to 5 points, representing the severity or impact degree. The scale structure design includes: Physical dimension: The indicators may include pain score, degree of activity limitation, daily living function, etc.

[0098] Psychological dimension: including anxiety score, depression score, mood swing situation, etc.

[0099] Social dimension: Evaluate social interaction ability, frequency of participating in social activities, etc.

[0100] Functional dimension: Working ability, daily self-care ability, etc.

[0101] S2. Based on the multi-dimensional quality of life index scale, calculate the comprehensive risk of the patient's quality of life under different symptoms through the dynamic symptom tree model and in combination with the Markov method; As a preferred implementation method, calculating the comprehensive risk of the patient's quality of life under different symptoms through the dynamic symptom tree model and in combination with the Markov method based on the multi-dimensional quality of life index scale includes the following steps: S21. Based on the multi-dimensional quality of life index scale, configure the dynamic symptom tree parameters and establish a dynamic symptom tree model. The dynamic symptom tree parameters include basic events, intermediate events, sequence-related gates, and top events; It should be noted that the dynamic symptom tree model is a hierarchical graphical model used to represent how different symptoms and their combinations affect the patient's quality of life. This model is similar to the Fault Tree Analysis (FTA), but is more dynamic and can reflect the impact of symptom changes on the final health status.

[0102] The basic event is the bottom - most node of the dynamic symptom tree, representing directly observed, single symptoms or risk factors. Each basic event corresponds to a symptom node in a quality - of - life indicator scale, such as pain, anxiety, insomnia, etc.

[0103] The intermediate event is an event composed of basic events, reflecting the combined impact of multiple symptoms or risk factors. Intermediate events usually describe more complex health problems caused by the combined action of multiple symptoms. These intermediate events can be obtained through the combination of basic events.

[0104] The sequence - related gate is a logical gate in the tree structure, used to represent the sequence relationship between events. That is, certain symptoms or events must occur in a specific order to cause the top event to occur. There may be a dependency relationship between some symptoms. For example, the occurrence of one symptom may trigger another symptom, and this sequence relationship can be represented by the sequence - related gate.

[0105] The top event is the final node of the dynamic symptom tree, usually representing a major change in quality of life or a health risk, such as a severe decline in quality of life or a sharp deterioration of the condition. The top event is the final result of the entire symptom chain, and the combination of all symptoms and intermediate events will lead to the occurrence of the top event.

[0106] The visual structure of the dynamic symptom tree model is an inverted tree, with basic events at the bottom of the tree and the top event at the top. Through logical gates (such as "AND gate" and "OR gate") and sequence - related gates, the basic events and intermediate events are connected to form a complete description of the relationship between symptoms and quality of life.

[0107] S22. Modularize the dynamic symptom tree model to obtain several modular dynamic symptom trees; and calculate the risk values of the modular dynamic symptom trees using the Markov method; As a preferred implementation, modularizing the dynamic symptom tree model to obtain several modular dynamic symptom trees; and calculating the risk values of the modular dynamic symptom trees using the Markov method includes the following steps: S221. Divide the dynamic symptom tree model into several dynamic sub - trees, and use the top events of the dynamic sub - trees as basic bottom events to obtain several modular dynamic symptom trees; Specifically, it can be divided according to different dimensions of quality of life (such as physical, psychological, social) or according to the similarity and dependency relationship of symptoms. Combine the relevant symptom and event nodes in the symptom tree into an independent sub - tree. Each sub - tree contains at least one top event, which serves as the final result of the sub - tree. After the division, each sub - tree will have a top event, indicating the final result caused by the combination of symptoms and events in the sub - tree. To construct a modular dynamic symptom tree, convert the top event of each dynamic sub - tree into a new basic bottom event to form multiple modular dynamic symptom trees.

[0108] S222. For the basic events and intermediate events in the module dynamic symptom tree, each basic event and intermediate event is regarded as a state of the Markov chain, and the state transition probability is configured by using the memoryless property of the Markov chain. Specifically, in the dynamic symptom tree, the basic event represents the directly observed symptom or risk factor, while the intermediate event is the combination and interaction effect of these basic events. For Markov chain analysis, these events can be regarded as different states of the system.

[0109] Basic event state: The basic event corresponds to a specific state in the Markov chain. For example, the fatigue symptom of the patient may be state 1, and the pain symptom is state 2.

[0110] Intermediate event state: The intermediate event can also be regarded as a state in the Markov chain, representing the result of the joint action of multiple basic events. For example, the combination of fatigue and pain may be an intermediate event of "decreased mobility", which can be defined as another state.

[0111] A core feature of the Markov chain is the memoryless property, that is, the future state depends only on the current state and is independent of the previous states. This feature simplifies the modeling of the dynamic relationship between symptoms, ensuring that the symptom state at a certain time point is only affected by the currently observed symptoms and not by the past symptom states. In the context of the dynamic symptom tree: The current symptoms (such as fatigue, anxiety) only affect the subsequent symptom development.

[0112] When in a certain state (such as "decreased mobility"), its subsequent evolution depends only on the current intermediate event, rather than what combination it was caused by before.

[0113] To make the Markov chain effective, it is necessary to set the state transition probability for each state (that is, each basic event and intermediate event), that is, the probability of transitioning from the current state to the next state. This transition probability reflects the dynamic evolution relationship between symptoms. It includes the following steps: Determine the possible transition paths between states: Define the transition paths from one basic event or intermediate event to other events. For example, the fatigue state (state 1) may transition to the pain state (state 2), or transition to the intermediate event state of "decreased mobility" (state 3).

[0114] Configure the transition probability: Assign a probability value to the transition between each state, reflecting the possibility of another state occurring in a certain symptom or event state.

[0115] If the fatigue state (state 1) has a possibility of leading to pain (state 2), and the possibility of "decreased mobility" (state 3) occurring is relatively low, then the corresponding transition probabilities can be set as follows: The probability from state 1 to state 2, P(1→2) = 0.6.

[0116] The probability from state 1 to state 3, P(1→3) = 0.3.

[0117] Calibrate the probabilities based on pathological knowledge or clinical data: The values of the state transition probabilities can be set and calibrated through clinical research data, the development patterns of patients' historical symptoms, or doctors' empirical knowledge. Reasonably configuring these probabilities is crucial for accurately modeling the symptom evolution process.

[0118] After configuring the transition probabilities between all states, the state transitions in the Markov chain can be represented as a state transition matrix. Each element of this matrix represents the probability of transitioning from one state to another. Suppose there are three states, state 1 is "fatigue", state 2 is "pain", and state 3 is "decreased mobility", the state transition matrix P is as follows: ; Among them, the sum of the probabilities of transitioning from the current state to other states in each row is 1.

[0119] S223. Calculate the probability of each module's dynamic symptom tree from the initial state to the top event state according to the state transition probability; It should be noted that when calculating the probability of each module's dynamic symptom tree from the initial state to the top event state, the state transition probability of the Markov chain needs to be utilized. The Markov chain is a mathematical system that describes the transition probabilities of a process between different states. In the dynamic symptom tree model, each state can represent a symptom or a combination of symptoms, and the state transition probability reflects the possibility of symptom development over time.

[0120] Specifically, when calculating the probability of each module's dynamic symptom tree from the initial state to the top event state, the state transition probability of the Markov chain is utilized. First, the states and the transition probabilities between them are defined. These states represent different symptoms or combinations of symptoms, and the transition probabilities reflect the possibility of symptom development over time. Then, a state transition probability matrix is constructed, and the probability distribution starting from the initial state is represented by initializing. Then, by iteratively calculating the product of the state transition probability matrix and the probability distribution, the long-term state probability distribution is obtained. Finally, the probability of the top event state is extracted from this distribution, which is the probability from the initial state to the top event state sought.

[0121] S224. In each module's dynamic symptom tree, evaluate the degree of impact of the top event on the patient's quality of life, and calculate the risk value of each module's dynamic symptom tree according to the degree of impact and the probability from the initial state to the top event state of each module's dynamic symptom tree.

[0122] It should be noted that the top event represents a state where the patient's quality of life has declined to a specific level, such as severe disability, significant reduction in quality of life, etc. Evaluating the degree of impact of the top event on the patient's quality of life requires comprehensive consideration of multiple aspects, such as the severity of symptoms, duration, treatment difficulty, restrictions on the patient's daily activities, etc. The evaluation method can be qualitative (such as expert scoring, questionnaire surveys) or quantitative (such as statistical analysis based on historical data); The risk value is the product of the degree of impact of the top event and the probability from the initial state to the top event state; the risk value quantifies the potential threat degree of the symptom combination represented by each module's dynamic symptom tree to the patient's quality of life. The higher the risk value, the greater the negative impact of the symptom combination on the patient's quality of life, and more attention and timely intervention are required.

[0123] S23. According to the risk values of each module's dynamic subtree, summarize the risk values of each dynamic subtree according to different symptom chain combinations to obtain the comprehensive risk of the patient's quality of life.

[0124] It should be noted that in multiple modules, different symptoms may be independent of each other or may interact with each other. This means that the risk values of multiple modules' dynamic subtrees need to be reasonably summarized in order to calculate the overall quality of life risk of the patient. The risk value of each module's dynamic subtree can be weighted and summarized according to its relative importance to the overall quality of life.

[0125] The situations of symptom chain combinations include: Independent symptom chains: If the symptom chains of each module are independent of each other, that is, the symptom development between modules does not affect each other, then the comprehensive risk can be directly calculated through the weighted average or linear combination of the risk values of each module.

[0126] Interacting symptom chains: If the symptom chains in multiple modules interact or are linked, then the relevance between these symptom chains needs to be considered, and the comprehensive risk can be calculated through corresponding combination methods (such as logical operations, joint probabilities).

[0127] S3. Based on the comprehensive risk of the patient's quality of life under different symptoms, evaluate the importance of each symptom to the quality of life, and select adjustment measures according to the evaluation results of the importance and in combination with pre-configured adjustment measures; As a preferred embodiment, based on the comprehensive risks of the patient's quality of life under different symptoms, evaluate the importance of each symptom to the quality of life. According to the evaluation results of the importance and combined with the pre-configured adjustment measures, the steps for selecting the adjustment measures are as follows: S31. Based on the comprehensive risks of the patient's quality of life under different symptoms, calculate the contribution of each symptom to the comprehensive risk; It should be noted that during the process of evaluating the comprehensive risks of the patient's quality of life under different symptoms, the impact can be quantified by calculating the contribution of each symptom to the comprehensive risk. First, establish a symptom risk model to associate the severity of each symptom with the comprehensive risk of the quality of life. Then, use the contribution coefficient and symptom severity of each symptom to calculate its specific contribution to the comprehensive risk, and determine the relative importance of each symptom through normalization.

[0128] S32. Combine the contributions of the comprehensive risk and calculate the importance of each symptom through the proportion of symptom risk contribution; It should be noted that after calculating the contribution of each symptom to the comprehensive risk, it is necessary to calculate the proportion of symptom risk contribution based on these contribution values, that is, the relative impact of each symptom on the overall quality of life risk. Through the proportion of symptom risk contribution, the importance of each symptom can be further calculated. The importance reflects the degree of influence of each symptom on the patient's quality of life. The importance of each symptom can be directly equivalent to its contribution proportion because it reflects the relative weight of the symptom in the overall risk.

[0129] Therefore, the importance Ii of the symptom can be expressed as: Ii = contribution proportion (symptom i); This formula indicates that the higher the value of the importance, the greater the impact of the symptom on the patient's comprehensive risk and the higher the importance.

[0130] S33. Sort the symptoms according to the importance results of the symptoms and identify the core symptoms that affect the quality of life; It should be noted that the core symptoms are those symptoms that have a significant impact on the quality of life and are ranked high in terms of importance. Usually, select the first few symptoms as the core symptoms, and the specific number can be determined according to the coverage of the overall risk or clinical practice. For example, select the symptoms whose total importance accumulates to 70% or 80% as the core symptoms.

[0131] S34. Combine the pre-configured adjustment measures, use the identified core symptoms to select the adjustment measures, and formulate a personalized intervention plan.

[0132] It should be noted that there is a pre-configured adjustment measure library, which lists the possible intervention measures for various symptoms. These measures can include drug treatment, physical therapy, psychotherapy, lifestyle intervention, etc. For example: Pain: painkillers, physical therapy, acupuncture; Fatigue: rest schedule adjustment, exercise therapy, nutritional supplementation; Anxiety: psychotherapy, cognitive behavioral therapy (CBT), relaxation training; Using the identified core symptoms, match appropriate intervention means from a pre-configured library of adjustment measures. Focus on the symptoms with the highest importance, and preferentially select effective adjustment measures for them.

[0133] For example, assume that pain and fatigue are core symptoms, and possible adjustment measures are: Pain: non-steroidal anti-inflammatory drugs (NSAIDs), physical therapy, acupuncture; Fatigue: interventions to improve sleep quality, reasonable exercise plans, nutritional supplementation; Once the adjustment measures for the core symptoms are selected, the next step is to develop a personalized intervention plan for the patient. This plan should be tailored to the specific situation of the patient, considering the following factors: Severity of symptoms: conduct more frequent or intensive interventions on the more severe parts of the symptoms.

[0134] Patient's living conditions and health background: such as past medical history, drug allergies, living habits, etc.

[0135] Priority and schedule of intervention measures: For example, first deal with the symptoms that most affect the quality of life (such as pain), and then gradually deal with secondary problems (such as anxiety or fatigue).

[0136] Long-term monitoring and adjustment: Regularly evaluate the intervention effect based on the patient's feedback and clinical effects, and make adjustments if necessary. Follow-up can be carried out through the quality of life data reported by the patient or symptom monitoring tools.

[0137] As Figure 2 shown, according to another embodiment of the present invention, a patient quality of life assessment and analysis system based on patient reports is provided. The patient quality of life assessment and analysis system based on patient reports includes: an index scale construction module 1, a comprehensive risk calculation module 2, and an importance assessment module 3, and the index scale construction module 1, the comprehensive risk calculation module 2, and the importance assessment module 3 are sequentially connected; The index scale construction module 1 is used to screen out patient-reported outcome symptoms by using the symptom information of peripheral artery disease patients collected in advance, and construct a multi-dimensional quality of life index scale through symptom chain analysis based on the patient-reported outcome symptoms; The comprehensive risk calculation module 2 is used to calculate the comprehensive risk of the patient's quality of life under different symptoms based on the multi-dimensional quality of life index scale through a dynamic symptom tree model and combined with the Markov method; The importance evaluation module 3 is used to evaluate the importance of each symptom to the quality of life based on the comprehensive risk of the patient's quality of life under different symptoms, and select adjustment measures according to the importance evaluation result and in combination with pre-configured adjustment measures.

[0138] In summary, by means of the above technical solutions of the present invention, the present invention can more accurately reflect the individual differences and subjective feelings of patients by collecting the symptom information reported by patients themselves. By constructing a multi-dimensional quality of life index scale through the symptom chain analysis method, it can comprehensively evaluate multiple aspects such as the physiological, psychological, and social functions of patients, which helps doctors understand the health status of patients more comprehensively, identify various factors affecting the quality of life, and thus formulate a more comprehensive treatment plan. By using the dynamic symptom tree model and combining with the Markov method, it can dynamically calculate the comprehensive risk of the patient's quality of life under different symptoms, which helps doctors track the changes in the patient's condition in real time, adjust the treatment plan in time, and reduce the risk of the patient's quality of life decline. The present invention deeply reveals the complex relationships between symptoms by calculating the co-occurrence pattern, temporal relationship, and severity change index, improving the accuracy of analysis. By calculating the correlation index of symptom data, a symptom chain network model is constructed, which intuitively shows the mutual influence and correlation relationship between symptoms, helping doctors better understand the development process of the disease and formulate a more accurate treatment plan. By calculating the knowledge amount of each symptom node and performing weighted processing according to the preset weight distribution function, it ensures that the influence of important symptom nodes in the clustering analysis process is fully reflected, and can provide scientific, accurate, and comprehensive decision-making support for doctors, which helps to improve the quality of life of patients with peripheral artery disease. The present invention modularizes the dynamic symptom tree model to obtain several modular dynamic symptom trees, making the model more flexible and easy to manage. By using the memoryless property of the Markov chain and configuring the state transition probability, it can accurately calculate the probability of each modular dynamic symptom tree from the initial state to the top event state. By calculating the proportion of symptom risk contribution to calculate the importance of each symptom and sorting the symptoms, it can identify the core symptoms that have the greatest impact on the quality of life, providing a key direction for the formulation of intervention measures.

[0139] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, optical storage, etc.) containing computer-usable program code.

[0140] The specific embodiments described above further elaborate on the objective, technical solution, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for evaluating and analyzing the quality of life of patients based on patient reports, characterized in that The patient-reported patient quality of life assessment and analysis method includes the following steps: S1. Screen patient-reported outcome symptoms using the symptom information of peripheral artery disease patients collected in advance, and construct a multi-dimensional quality of life index scale through symptom chain analysis based on the patient-reported outcome symptoms; S2. Based on the multi-dimensional quality of life index scale, calculate the comprehensive risk of the patient's quality of life under different symptoms through a dynamic symptom tree model combined with the Markov method; S3. Based on the comprehensive risk of the patient's quality of life under different symptoms, evaluate the importance of each symptom to the quality of life, and select adjustment measures according to the importance evaluation results combined with the pre-configured adjustment measures; The step of screening patient-reported outcome symptoms using the symptom information of peripheral artery disease patients collected in advance and constructing a multi-dimensional quality of life index scale through symptom chain analysis based on the patient-reported outcome symptoms includes the following steps: S11. Organize and classify the symptom information of peripheral artery disease patients collected, screen patient-reported outcome symptoms, and use the Delphi entropy method to discriminate and extract the screened patient-reported outcome symptoms to obtain an initial symptom dataset; S12. Analyze the co-occurrence pattern, time series relationship, and severity change of each symptom data in the initial symptom dataset, and construct a symptom chain network model according to the analysis results; S13. Use the knowledge quantity weighted fuzzy c-means clustering method to perform clustering analysis on the symptom nodes in the symptom chain network model, and establish a multi-dimensional quality of life index scale according to the clustering results.

2. The method for evaluating and analyzing the quality of life of a patient based on patient reports according to claim 1, wherein The step of analyzing the co-occurrence pattern, time series relationship, and severity change of each symptom data in the initial symptom dataset and constructing a symptom chain network model according to the analysis results includes the following steps: S121. Preprocess each symptom data in the initial symptom dataset, and the preprocessing includes data cleaning and data standardization processing; S122. Calculate the co-occurrence pattern index, time series relationship index, and severity change index of each symptom data according to the preprocessed symptom data; S123. Calculate the correlation index of each symptom data according to the co-occurrence pattern index, time series relationship index, and severity change index of each symptom data; S124. Select symptom data that meet the correlation threshold according to the correlation index of each symptom data, use the symptoms corresponding to the selected symptom data as symptom nodes, and construct a symptom chain network model according to the correlation index of each symptom data.

3. The patient quality of life assessment and analysis method based on patient reports according to claim 2, characterized in that The step of calculating the co-occurrence pattern index, time series relationship index, and severity change index of each symptom data according to the preprocessed symptom data includes the following steps: S1221. Construct a co-occurrence matrix according to the preprocessed symptom data, and calculate the collinear pattern index of each symptom data based on the co-occurrence matrix; S1222. Construct the time series data of the symptom data according to the preprocessed symptom data, and calculate the time series relationship index of each symptom data through the dynamic time warping algorithm; S1223. Quantify the severity of each symptom data based on a pre-configured severity score, and calculate the severity change index of each symptom data through regression analysis and mutual information.

4. A method for evaluating and analyzing the quality of life of a patient based on patient reports according to claim 2, characterized in that, The steps of using the knowledge volume weighted fuzzy c-means clustering method to perform clustering analysis on symptom nodes in the symptom linkage network model and establishing a multi-dimensional quality of life index scale according to the clustering results are as follows: S131. Perform intuitionistic fuzzification processing on the symptom data of symptom nodes through a membership degree generation function to obtain an intuitionistic fuzzy data set; S132. Calculate the knowledge volume of each symptom node in the intuitionistic fuzzy data set using an improved intuitionistic fuzzy knowledge measure model, and assign weights to the knowledge volume of each symptom according to a preset weight assignment function; S133. Perform weighted processing on the intuitionistic fuzzy data set using the weights assigned to the knowledge volume of each symptom to obtain a weighted data set; S134. Based on the weighted data set, select initial clustering centers based on kernel space density and kernel distance, and through iterative optimization, obtain the optimal fuzzy partition matrix and clustering centers according to the inter-class sample distance and the principle of minimum knowledge volume; S135. Construct a multi-dimensional quality of life index scale according to the optimal fuzzy partition matrix and clustering centers.

5. The method for evaluating and analyzing the quality of life of a patient based on patient reports according to claim 4, wherein The expression of the improved intuitionistic fuzzy knowledge measure model is: ； In the formula, G(A) represents the knowledge volume of symptom nodes in the intuitionistic fuzzy data set; n represents the number of symptom nodes in the intuitionistic fuzzy data set; u A (x i ) represents the membership degree of the i-th symptom node to the intuitionistic fuzzy data set; v A (x i ) represents the non-membership degree of the i-th symptom node for the intuitionistic fuzzy data set.

6. The patient quality of life assessment and analysis method based on patient reports according to claim 4, characterized in that The expression of the preset weight assignment function is: ； where, w j represents the weight of the j-th symptom node in the intuitionistic fuzzy dataset; G(A tj ) represents the knowledge quantity of the j-th symptom node in the t-th intuitionistic fuzzy dataset; m represents the number of intuitionistic fuzzy data sets; n represents the number of symptom nodes in the intuitionistic fuzzy data set.

7. A method for evaluating and analyzing the quality of life of a patient based on patient reports according to claim 1, characterized in that, The steps of calculating the comprehensive risk of the patient's quality of life under different symptoms through the dynamic symptom tree model and combining with the Markov method based on the multi-dimensional quality of life index scale are as follows: S21. Based on the multi-dimensional quality of life index scale, configure dynamic symptom tree parameters and establish a dynamic symptom tree model, where the dynamic symptom tree parameters include basic events, intermediate events, sequence-related gates, and top events; S22. Modularize the dynamic symptom tree model to obtain several modular dynamic symptom trees; and calculate the risk values of the modular dynamic symptom trees using the Markov method; S23. According to the risk values of each modular dynamic subtree, summarize the risk values of each dynamic subtree according to different symptom chain combinations to obtain the comprehensive risk of the patient's quality of life.

8. A method for evaluating and analyzing the quality of life of a patient based on patient reports according to claim 7, characterized in that, The step of modularizing the dynamic symptom tree model to obtain several modular dynamic symptom trees; And the steps of calculating the risk values of the modular dynamic symptom trees using the Markov method are as follows: S221. Divide the dynamic symptom tree model into several dynamic subtrees, and use the top events of the dynamic subtrees as basic bottom events to obtain several modular dynamic symptom trees; S222. For the basic events and intermediate events in the modular dynamic symptom trees, regard each basic event and intermediate event as the state of the Markov chain, and configure the state transition probability using the non-aftereffect characteristic of the Markov chain; S223. According to the state transition probability, calculate the probability of each modular dynamic symptom tree from the initial state to the top event state; S224. In each module's dynamic symptom tree, evaluate the degree of impact of the top event on the patient's quality of life, and calculate the risk value of each module's dynamic symptom tree based on the degree of impact and the probability from the initial state to the top event state of each module's dynamic symptom tree.

9. A method for evaluating and analyzing the quality of life of a patient based on patient reports according to claim 1, characterized in that, Based on the comprehensive risk of the patient's quality of life under different symptoms, evaluate the importance of each symptom to the quality of life, and select adjustment measures according to the importance evaluation results and in combination with pre-configured adjustment measures, including the following steps: S31. Based on the comprehensive risk of the patient's quality of life under different symptoms, calculate the contribution of each symptom to the comprehensive risk; S32. Combine the contribution of the comprehensive risk and calculate the importance of each symptom through the proportion of symptom risk contribution; S33. Sort the symptoms according to the importance results of the symptoms, and identify the core symptoms that affect the quality of life; S34. Combine the pre-configured adjustment measures, select adjustment measures using the identified core symptoms, and formulate a personalized intervention plan.

10. A patient-reported patient quality of life assessment and analysis system for implementing the patient-reported patient quality of life assessment and analysis method according to any one of claims 1-9, characterized in that, The patient quality of life assessment and analysis system based on patient reports includes: an index scale construction module, a comprehensive risk calculation module, and an importance evaluation module, and the index scale construction module, the comprehensive risk calculation module, and the importance evaluation module are connected in sequence; The index scale construction module is used to screen out patient-reported outcome symptoms using the symptom information of peripheral artery disease patients collected in advance, and construct a multi-dimensional quality of life index scale through symptom chain analysis based on the patient-reported outcome symptoms; The comprehensive risk calculation module is used to calculate the comprehensive risk of the patient's quality of life under different symptoms based on the multi-dimensional quality of life index scale, through a dynamic symptom tree model and in combination with the Markov method; The importance evaluation module is used to evaluate the importance of each symptom to the quality of life based on the comprehensive risk of the patient's quality of life under different symptoms, and select adjustment measures according to the importance evaluation results and in combination with pre-configured adjustment measures.

Images