A clinical data acquisition system for respiratory patients and a data correction method

Abstract

The present application provides a clinical data acquisition system and data correction method for respiratory patients. First, the historical clinical data set of all respiratory patients is acquired through acquisition instructions, and then the test indicators of different respiratory signs are extracted from the historical clinical data set; then all the test indicators are subjected to multi-level pathological feature association analysis to obtain the dependency relationship of the pathological structure between each test indicator; further, the pathological features of all respiratory signs are associated and clustered to obtain multiple pathological feature clusters, and the feature dependency tree of each pathological feature cluster is constructed through the dependency relationship of the pathological structure between each test indicator in each pathological feature cluster; further, the clinical data of the target respiratory patient is corrected and processed based on all the feature dependency trees to obtain the proofreading data of the clinical data of the target respiratory patient. The multivariate correlation proofreading of the clinical data of respiratory patients can be realized, thereby improving the reliability of the clinical data.

Description

Technical Field

[0001] The present application relates to the technical field of medical data processing, and more specifically, to a clinical data acquisition system and a data correction method for respiratory patients. Background Art

[0002] With the rapid development of information technology, medical data processing technology has become an important part of the medical industry. Medical data includes structured data, such as patients' basic information, diagnosis results and treatment plans, and unstructured data, such as doctors' notes, medical images and gene sequences. Medical data processing refers to the process of cleaning, integrating and analyzing medical data through big data technology and extracting medical information.

[0003] Clinical data collection for respiratory patients refers to the recording and analysis of the physiological parameters and treatment responses of patients during the treatment and monitoring of respiratory patients. Clinical data for respiratory patients include respiratory rate, blood oxygen saturation, vital capacity and airway resistance. With the development of medical Internet of Things technology, medical equipment such as monitors, ventilators, and blood gas analyzers can collect patients' vital signs and respiratory data in real time. Although the existing clinical data collection technology for respiratory patients has made certain progress, the clinical data collection process for respiratory patients is subject to measurement errors, data delays, noise interference and other problems due to the unstable breathing patterns of patients and interference from respiratory activities, making it difficult to ensure data quality and effectively identify the correlation between various physiological parameters, resulting in insufficient data integrity. Therefore, how to achieve multivariate correlation proofreading of clinical data for respiratory patients and thus improve the reliability of clinical data has become a difficult problem faced by the industry. Summary of the invention

[0004] The present application provides a respiratory patient clinical data acquisition system and a data correction method, which can realize multivariate correlation proofreading of respiratory patient clinical data, thereby improving the reliability of clinical data.

[0005] In a first aspect, the present application provides a method for correcting clinical data of respiratory patients, which is used for data correction in a process of collecting clinical data of respiratory patients, and comprises the following steps:

[0006] Receiving a collection instruction of clinical data of respiratory patients, obtaining historical clinical data sets of all respiratory patients through the collection instruction, and extracting test indicators of different respiratory signs from the historical clinical data sets;

[0007] Based on the knowledge graph of respiratory diseases, a multi-level pathological feature correlation analysis is performed on all test indicators to obtain the dependency relationship of the pathological structure between each test indicator;

[0008] Performing association clustering on the pathological features of all respiratory signs to obtain multiple pathological feature clusters, and constructing a feature dependency tree of the pathological structure in each pathological feature cluster through the dependency relationship of the pathological structure between each test index in each pathological feature cluster;

[0009] Acquire clinical data of a target respiratory patient according to the acquisition instruction, perform correlation matching on the clinical data according to all feature dependency trees, and thereby obtain correlation correction parameters of all respiratory signs of the target respiratory patient;

[0010] The clinical data of the target respiratory patient is proofread using the associated correction parameters to obtain proofread data of the clinical data of the target respiratory patient.

[0011] Preferably, a multi-level pathological feature association analysis is performed on all the test indicators based on the respiratory disease knowledge graph to obtain the pathological structure dependency relationship between the various test indicators, specifically including:

[0012] Mapping all test indicators into the respiratory disease knowledge graph;

[0013] All the test indicators are correlated from the symptom surface, structural middle and pathological deep layers, and then the characteristic correlations of different levels between the test indicators are extracted;

[0014] The pathological structure dependency between each test index is determined based on all feature correlations.

[0015] Preferably, the pathological features of all respiratory signs are clustered by association to obtain multiple pathological feature clusters, specifically including:

[0016] Extracting the pathological characteristics of each respiratory sign from the test index of each respiratory sign;

[0017] Based on the normal distribution assumption, correlation analysis was performed on all pathological features to obtain the mutual correlation between pathological features;

[0018] All pathological features are hierarchically clustered according to the mutual correlation to obtain multiple pathological feature clusters.

[0019] Preferably, constructing a feature dependency tree of the pathological structure in each pathological feature cluster through the dependency relationship of the pathological structure between each inspection index in each pathological feature cluster specifically includes:

[0020] Determine the associated attention for each pathological feature cluster;

[0021] The feature dependency tree of the pathological structure in each pathological feature cluster is determined by the dependency relationship of the pathological structure between each inspection index in each pathological feature cluster and the associated attention of each pathological feature cluster.

[0022] The method also includes generating an electronic case report form of the target respiratory patient from the collated data of the clinical data of the target respiratory patient obtained after collation.

[0023] Preferably, the respiratory signs are physiological symptom indicators that reflect the functional state of the respiratory system.

[0024] Preferably, the test index is an index reflecting changes in the physiological state of the respiratory patient.

[0025] In a second aspect, the present application provides a respiratory patient clinical data acquisition system, the respiratory patient clinical data acquisition system includes a data correction unit, the data correction unit includes:

[0026] An acquisition module, used for receiving an acquisition instruction of clinical data of respiratory patients, acquiring historical clinical data sets of all respiratory patients through the acquisition instruction, and extracting test indicators of different respiratory signs from the historical clinical data sets;

[0027] The processing module is used to perform multi-level pathological feature correlation analysis on all test indicators based on the knowledge graph of respiratory diseases to obtain the dependency relationship of the pathological structure between various test indicators;

[0028] The processing module is further used to perform association clustering on the pathological features of all respiratory signs to obtain multiple pathological feature clusters, and to construct a feature dependency tree of the pathological structure in each pathological feature cluster through the dependency relationship of the pathological structure between each inspection index in each pathological feature cluster;

[0029] The processing module is further used to obtain clinical data of the target respiratory patient according to the acquisition instruction, and to perform association matching on the clinical data according to all feature dependency trees, thereby obtaining association correction parameters of all respiratory signs of the target respiratory patient;

[0030] The execution module is used to perform proofreading processing on the clinical data of the target respiratory patient according to the associated correction parameters to obtain proofreading data of the clinical data of the target respiratory patient.

[0031] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores codes, and the processor is configured to obtain the codes and execute the above-mentioned method for correcting clinical data of respiratory patients.

[0032] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned method for correcting clinical data of respiratory patients is implemented.

[0033] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects:

[0034] In an embodiment of the present application, first, a collection instruction for clinical data of a respiratory patient is received, and a historical clinical data set of all respiratory patients is obtained through the collection instruction, and test indicators of different respiratory signs are extracted from the historical clinical data set; a multi-level pathological feature association analysis is performed on all test indicators based on a knowledge graph of respiratory diseases to obtain a dependency relationship of the pathological structure between each test indicator; the pathological features of all respiratory signs are clustered by association to obtain a plurality of pathological feature clusters, and a feature dependency tree of the pathological structure in each pathological feature cluster is constructed through the dependency relationship of the pathological structure between each test indicator in each pathological feature cluster; the clinical data of the target respiratory patient is obtained according to the collection instruction, and the clinical data is matched by association according to all feature dependency trees to obtain association correction parameters of all respiratory signs of the target respiratory patient; the clinical data of the target respiratory patient is proofread by the association correction parameters to obtain proofreading data of the clinical data of the target respiratory patient.

[0035] It can be seen that the present application proofreads the clinical data of the target respiratory patient by associating correction parameters, and then generates an electronic case report form of the target respiratory patient from the proofread clinical data; firstly, a feature dependency tree is constructed by using the dependency relationship of the pathological structure between each test index in the pathological feature cluster, which can refine the subordinate dependency relationship of the test index within the pathological feature cluster, and express the multi-level pathological variable correlation relationship in a tree structure, and analyze the superior-subordinate relationship and influence between different test indexes through the feature dependency tree, which is convenient for hierarchical correlation analysis of each pathological variable, and can improve the reliability of clinical data during the final proofreading; secondly, the feature dependency tree and real-time clinical data are correlated and matched to obtain the respiratory patient The associated correction parameters of all respiratory signs of the patient are obtained, wherein the associated correction parameters can effectively correct errors in real-time clinical data, thereby forming sign data that is more in line with the actual situation, so that the correction parameters are generated on the basis of multivariate association to ensure the consistency of the proofread clinical data; finally, the clinical data of the target respiratory patient is proofread and processed using the associated correction parameters, and an electronic case report is generated. The results of the multivariate association analysis can be applied to the patient's actual clinical data to improve the accuracy of the clinical data, which is conducive to improving the reliability of the generated electronic case; in summary, the present application scheme can realize multivariate correlation proofreading of the clinical data of respiratory patients, thereby improving the reliability of the clinical data. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 is an exemplary flow chart of a method for correcting clinical data of a respiratory patient according to some embodiments of the present application;

[0037] Figure 2is a schematic diagram of a process for determining a pathological feature cluster according to some embodiments of the present application;

[0038] Figure 3 is a schematic diagram of a process for determining an associated correction parameter according to some embodiments of the present application;

[0039] Figure 4 is a structural schematic diagram of a data correction unit according to some embodiments of the present application;

[0040] Figure 5 It is a structural schematic diagram of a computer device for implementing a method for correcting clinical data of respiratory patients according to some embodiments of the present application. DETAILED DESCRIPTION

[0041] In order to better understand the technical solution of the present application, the technical solution of the present application will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0042] refer to Figure 1 , which is an exemplary flow chart of a method for correcting clinical data of a respiratory patient according to some embodiments of the present application. The method 100 for correcting clinical data of a respiratory patient mainly includes the following steps:

[0043] In step 101, a collection instruction of clinical data of respiratory patients is received, historical clinical data sets of all respiratory patients are acquired through the collection instruction, and test indicators of different respiratory signs are extracted from the historical clinical data sets.

[0044] It should be noted that the respiratory signs in this application refer to physiological symptom indicators that reflect the functional state of the respiratory system. The respiratory signs usually include: dyspnea, wheezing, coughing and shortness of breath. Respiratory signs can be obtained through observation, auscultation, measurement, etc.; in addition, the test indicators in this application refer to indicators that reflect changes in the physiological state of respiratory patients. The test indicators usually include: respiratory rate, blood oxygen saturation, vital capacity and airway resistance.

[0045] In specific implementation, first, the data acquisition system receives a collection instruction through an interface. The collection instruction usually includes the patient's unique identification (such as ID number), the time period to be obtained, and specific test indicators. After receiving the collection instruction, the data acquisition system will initiate a query request to the historical data repository to extract the clinical data of all respiratory patients in the historical time period, and use all the extracted clinical data as the historical clinical data set; then, after receiving the collection instruction, the data acquisition system accesses the historical data storage area (i.e., the historical clinical data set) according to the patient's ID and other retrieval conditions (such as date range, test items), and the data that can be filtered and extracted through predefined feature fields include test indicators related to respiratory signs, such as respiratory rate, blood oxygen saturation, vital capacity and airway resistance. It should be noted that for massive data, distributed computing technology (such as Hadoop or Spark) can be combined to accelerate data extraction, which will not be elaborated here.

[0046] It should be noted that the historical time period in the present application refers to the time period corresponding to the current moment to the past 2 years. In other embodiments, the historical time period may also be other time periods, which are not limited here.

[0047] In step 102, a multi-level pathological feature association analysis is performed on all the test indicators based on the respiratory disease knowledge graph to obtain the dependency relationship of the pathological structure between the various test indicators.

[0048] It should be noted that the respiratory disease knowledge graph in this application is constructed by medical experts or based on literature data, and the different symptoms, signs, pathological characteristics and their relationships of respiratory diseases are structured. The storage of the respiratory disease knowledge graph can adopt a graph database (such as Neo4j), and the association analysis of data can be performed through SPARQL or Cypher query language.

[0049] In some embodiments, a multi-level pathological feature association analysis is performed on all test indicators based on the respiratory disease knowledge graph to obtain the dependency relationship of the pathological structure between the various test indicators. The following steps can be used to achieve this:

[0050] Mapping all test indicators into the respiratory disease knowledge graph;

[0051] All the test indicators are correlated from the symptom surface, structural middle and pathological deep layers, and then the characteristic correlations of different levels between the test indicators are extracted;

[0052] The pathological structure dependency between each test index is determined based on all feature correlations.

[0053] It should be noted that in this application, multi-level feature association analysis can comprehensively analyze the correlation between test indicators from different angles (symptoms, anatomical structure, pathological mechanism), which is helpful to analyze the relationship between pathologies and better understand the complex mechanism of the disease and the dependency between signs.

[0054] In the specific implementation, first, it is necessary to identify and match the corresponding nodes of each test indicator in the respiratory disease knowledge graph, and map all test indicators (such as CRP, PaO2, etc.) to the nodes representing the corresponding pathological characteristics in the respiratory disease knowledge graph. This mapping can be performed through standard medical terminology systems (such as ICD-10, SNOMED CT, etc.), which can achieve standardized matching of indicators to nodes. In addition to the direct mapping of test indicators, it is also necessary to determine the relationship between nodes. For example, CRP may be related to inflammatory response, while PaO2 is related to hypoxemia. These relationships can be associated through existing edges in the respiratory disease knowledge graph, and can also be further expanded through literature mining and expert rules; secondly, feature association analysis of test indicators is performed from the symptom surface, structure middle layer and pathology deep layer. When performing feature association analysis on the symptom surface, graph-based reasoning can be used. Reasoning, using the structure of the respiratory disease knowledge graph, find the direct relationship chain between the test indicators, calculate their association strength, and use the association strength as the correlation between the test indicators at the symptom surface level. When performing feature association analysis in the middle layer of the structure, medical image analysis and structural modeling (such as CT, MRI and other data) can be used to combine the test indicators with the structural nodes in the graph for reasoning. The structured graph reasoning algorithm can be used to extract the intermediate node (structural layer) association between different test indicators, and the intermediate node association can be used as the structural layer correlation between the test indicators. When performing feature association analysis in the deep pathology layer, gene expression analysis, protein interaction and other technologies can be used to associate the test indicators with the deep pathology nodes. The pathology nodes (such as genes, proteins, pathological reactions, etc.) in the respiratory disease knowledge graph are used to calculate the pathology level correlation, and the pathology level correlation is used as the pathology deep correlation between the test indicators. Furthermore, for each level of the test indicators, a machine learning model can be used to extract their correlation (i.e., the correlation of the symptom surface, the correlation of the structural layer and the correlation of the pathology deep layer), and then mutual information can be used. Information) method is used to quantify the correlation strength between each pair of test indicators. In some embodiments, for any two test indicators, the average value of the correlation strength between the three levels (i.e., the correlation degree of the symptom surface layer, the correlation degree of the structural layer, and the correlation degree of the pathological deep layer) can be used as the characteristic correlation degree of different levels between the two test indicators, and then the characteristic correlation degree of different levels between each test indicator is obtained; finally, for any two test indicators, the characteristic correlation degree of different levels between the two test indicators can be used as the dependency relationship of the pathological structure between the two test indicators.

[0055] It should be noted that the dependency of pathological structure in the present application refers to the interdependence between different test indicators in the pathological structure. The dependency of pathological structure can be used to describe how changes in one pathological feature affect another feature, especially in the analysis of respiratory signs, where different signs are often interrelated through complex physiological mechanisms.

[0056] In step 103, the pathological features of all respiratory signs are clustered to obtain multiple pathological feature clusters, and a feature dependency tree of the pathological structure in each pathological feature cluster is constructed through the dependency relationship of the pathological structure between each inspection index in each pathological feature cluster.

[0057] In some embodiments, reference Figure 2 As shown in FIG. 1 , this figure is a schematic diagram of a process for determining a pathological feature cluster in some embodiments of the present application. In this embodiment, the pathological features of all respiratory signs are associated and clustered to obtain multiple pathological feature clusters, which can be achieved by the following steps:

[0058] In step 1031, the pathological characteristics of each respiratory sign are extracted from the inspection index of each respiratory sign;

[0059] In step 1032, based on the normal distribution assumption, correlation analysis is performed on all pathological features to obtain the mutual correlation between the pathological features;

[0060] In step 1033, hierarchical clustering is performed on all pathological features according to the mutual correlation degree to obtain multiple pathological feature clusters.

[0061] It should be noted that the pathological characteristics in this application refer to indicators used in the medical field to describe diseases or pathological states, and the pathological characteristics can reflect the biological processes of the disease, abnormal changes in tissues or organs, and physiological dysfunctions associated with the disease; in addition, the corresponding pathological characteristics are defined in this application according to the clinical significance of each test indicator, for example: respiratory rate, frequency changes beyond the normal range (usually 12-20 times / minute) can be regarded as pathological characteristics, blood oxygen saturation: less than 95% is regarded as a hypoxic state as a pathological feature, vital capacity: compared with normal values, lower than the normal range can indicate lung disease (pathological characteristics), airway resistance: airway resistance higher than the normal range reflects airway dysfunction (pathological characteristics).

[0062] In specific implementation, the test indicators of each respiratory sign can be extracted from the historical clinical data set. For example, respiratory signs include dyspnea, wheezing, coughing and shortness of breath, and the test indicators include respiratory rate, blood oxygen saturation, vital capacity and airway resistance, and then feature extraction is performed on each test indicator. Specifically, representative pathological features can be extracted for different test indicators. For example, for respiratory rate, the respiratory rate with a frequency change beyond the normal range (usually 12-20 times / minute) can be used as the pathological feature of respiratory rate. For blood oxygen saturation, a value below 95% can be regarded as a hypoxic state as the pathological feature of blood oxygen saturation. For vital capacity, a vital capacity below the normal range can be used as the pathological feature of vital capacity. For airway resistance, airway dysfunction can be used as the pathological feature of airway resistance.

[0063] In addition, based on the normal distribution assumption, all pathological features are subjected to correlation analysis, and the mutual correlation between pathological features can be obtained in the following way, namely: first, all extracted pathological features can be subjected to a normality test (such as the Shapiro-Wilk test) to verify their distribution characteristics. If they do not conform to the normal distribution, appropriate transformation methods (such as logarithmic transformation) are applied to transform the pathological features to make them closer to the normal distribution. The processed pathological features are then substituted into the mutual information calculation formula, and the calculation results are used as the mutual information between pathological features. The mutual information calculation results can then be used to generate a mutual correlation matrix of pathological features. Each element in the matrix represents the correlation between two pathological features. The mutual correlation matrix can be used to describe the mutual correlation between pathological features. It should be noted that mutual information is a measure of the amount of information shared between two random variables. Through mutual information analysis, significant correlations between pathological features can be identified, revealing the potential mutual influence between respiratory signs, and providing a basis for subsequent clustering.

[0064] It should be noted that the correlation degree in the present application is an indicator to measure the degree of correlation between pathological features. Through correlation analysis in the present application, the mutual influence between various pathological features can be quantified, highly correlated features can be identified, and the complex relationship between respiratory signs and disease status can be further revealed.

[0065] In specific implementation, all pathological features are hierarchically clustered according to the mutual correlation to obtain multiple pathological feature clusters, which can be achieved in the following manner: first, each pathological feature is treated as an independent cluster during initialization, and then the Euclidean distance between each two clusters is calculated based on the mutual correlation, and the two closest clusters are further merged into a new cluster, and the merging is repeated until all pathological features are aggregated into a preset number of clusters, and each cluster is used as a pathological feature cluster. It should be noted that when determining the number of clusters, the optimal number of clusters can be determined by evaluating the clustering performance to ensure that the pathological features in each cluster are highly correlated.

[0066] In some embodiments, constructing a feature dependency tree of the pathological structure in each pathological feature cluster through the dependency relationship of the pathological structure between each inspection index in each pathological feature cluster can be implemented by the following steps:

[0067] Determine the associated attention for each pathological feature cluster;

[0068] The feature dependency tree of the pathological structure in each pathological feature cluster is determined by the dependency relationship of the pathological structure between each inspection index in each pathological feature cluster and the associated attention of each pathological feature cluster.

[0069] Preferably, in the above embodiment, determining the associated attention of each pathological feature cluster can be achieved by using the following steps:

[0070] Determine the distribution concentration of pathological features in each pathological feature cluster;

[0071] The associated attention of the corresponding pathological feature cluster is determined by the proportion of each distribution concentration.

[0072] In specific implementation, first, for each pathological feature cluster, the distribution concentration of the pathological features in the pathological feature cluster can be described by the information entropy in the prior art, that is, all the pathological features in the pathological feature cluster can be substituted into the information entropy calculation formula, and the calculation result is used as the distribution concentration of the pathological features in the pathological feature cluster, thereby obtaining the distribution concentration of the pathological features in all pathological feature clusters; then, the sum of all distribution concentrations is taken as the total distribution concentration, and the ratio of the distribution concentration corresponding to each pathological feature cluster to the total distribution concentration is taken as the associated attention of each pathological feature cluster; it should be noted that in this application, by linking information entropy with associated attention, weights can be assigned according to the concentration of features within the pathological feature cluster. A cluster with a low entropy value means that its internal features are highly correlated, so it is necessary to give a higher weight in subsequent analysis and diagnosis to ensure more accurate pathological feature modeling. The method of this embodiment can effectively reflect the internal structural differences of each pathological feature cluster, help prioritize feature clusters that are more clinically relevant, and thus improve the accuracy of diagnosis.

[0073] It should be noted that the associated attention in this application refers to a weight indicator assigned according to the importance or relevance of different feature clusters, which indicates the degree of attention that the feature cluster should receive in the overall analysis. For example, in the analysis of respiratory diseases, if certain pathological feature clusters (such as low blood oxygen and high respiratory rate) show strong correlation in most patients, the associated attention of this pathological feature cluster will be higher, and the model will assign more weight to this pathological feature cluster during analysis.

[0074] Preferably, in the above embodiment, determining the feature dependency tree of the pathological structure in each pathological feature cluster through the dependency relationship of the pathological structure between each inspection index in each pathological feature cluster and the associated attention of each pathological feature cluster can be implemented by the following steps:

[0075] For each pathological feature cluster, the pathological features in the pathological feature cluster are used as tree nodes;

[0076] The dependency relationship between the pathological structures of the various test indicators in the pathological feature cluster is adjusted through the associated attention of the pathological feature cluster, and then the edge weights between the various pathological features are determined according to the dependency relationship of all the pathological structures after the adjustment;

[0077] The feature dependency tree of the pathological feature clusters is determined through all tree nodes and all edge weights.

[0078] In the specific implementation, first, for each pathological feature cluster, each pathological feature in it is defined as a node of the tree, and a unique identifier is assigned to each node so that the identity of the node can be clearly identified when the tree structure is constructed later; then, the importance of each pathological feature in the whole is determined through the associated attention of the pathological feature cluster. The more important the feature is, the greater the weight in the dependency relationship, that is, the dependency relationship of the pathological structure between each node and other nodes is weighted according to the adjustment value of the associated attention. For example, the dependency weight between features with a high degree of mutual influence is set to a larger value. It should be noted that the adjusted dependency relationship of the pathological structure can accurately reflect the pathological characteristics. The mutual influence of the features can be improved to improve the scientificity and accuracy of the dependency tree. Further, the edge weights can be assigned between each pair of adjacent pathological features based on the dependency relationship of the adjusted pathological structure. The size of the edge weight is determined by the associated attention value and the feature dependency strength. The higher the edge weight, the stronger the correlation between the two pathological features, which means that the mutual dependence and influence are greater. Finally, a graph algorithm (such as the maximum spanning tree algorithm) can be used to integrate all tree nodes and edge weights into a dependency tree, and the dependency tree is used as the feature dependency tree. Among them, the maximum spanning tree algorithm will select the connection method with the largest edge weight to ensure that the constructed tree reflects the dependency relationship of the pathological features as completely as possible.

[0079] It should be noted that the feature dependency tree in this application refers to a data structure constructed based on the pathological knowledge graph of respiratory diseases, which is used to describe the hierarchical dependency relationship and association strength between different respiratory signs (such as respiratory rate, blood oxygen saturation, lung function indicators, etc.). The nodes of the feature dependency tree represent various pathological features, and the edge weights between the nodes represent the degree of association between these features.

[0080] In step 104, clinical data of the target respiratory patient is acquired according to the acquisition instruction, and the clinical data is associated and matched according to all feature dependency trees, thereby obtaining associated correction parameters of all respiratory signs of the target respiratory patient.

[0081] In specific implementation, obtaining the clinical data of the target respiratory patient according to the acquisition instruction can be achieved in the following manner, namely: after receiving the acquisition instruction, the data acquisition system initiates a query request to the data repository and obtains the clinical data of the target respiratory patient from the data repository.

[0082] In some embodiments, reference Figure 3 As shown, this figure is a schematic diagram of the process of determining the associated correction parameters in some embodiments of the present application. In this embodiment, the clinical data is associated and matched according to all feature dependency trees, and then the associated correction parameters of all respiratory signs of the target respiratory patient are obtained, which can be achieved by the following steps:

[0083] In step 1041, pathological features of different respiratory signs are extracted from the clinical data;

[0084] In step 1042, all pathological features are matched with each feature dependency tree to obtain the matching degree of each feature dependency tree, and then the optimal matching degree is obtained by screening;

[0085] In step 1043, the associated correction parameters of all respiratory signs of the target respiratory patient are determined according to the optimal matching degree.

[0086] In specific implementation, first, clinical data of target respiratory patients are collected and preprocessed, and feature extraction technology (such as principal component analysis) can be used to perform structured processing on clinical data of respiratory signs, and pathological features closely related to the respiratory system can be extracted, including but not limited to respiratory rate, blood oxygen saturation, vital capacity and airway resistance; secondly, matching algorithms (such as Euclidean distance, cosine similarity) can be used to quantify the adaptability of pathological features and feature dependency trees, and each feature dependency tree will receive an overall matching score, and the matching score will be used as the matching degree. Further, the matching scores of all feature dependency trees are sorted, and the feature dependency tree with the highest matching degree is selected as the optimal dependency tree, and the matching degree of the optimal dependency tree is used as the optimal matching degree; then, the dependency relationship and associated attention of the pathological structure of each node in the optimal dependency tree will be used to generate the associated correction parameters of each pathological feature, that is, the correction parameters of each node in the dependency tree can be combined into a comprehensive correction vector using the weighted average or weighted accumulation method, and numerical adjustments are made for each pathological feature.

[0087] It should be noted that the association correction parameters in this application refer to a set of parameters used to correct data deviations according to the association relationship and dependency structure between different pathological features during the clinical data correction process of respiratory patients.

[0088] In step 105, the clinical data of the target respiratory patient is proofread using the associated correction parameters to obtain proofread data of the clinical data of the target respiratory patient.

[0089] In some embodiments, the calibration process of the clinical data of the target respiratory patient using the associated calibration parameters can be implemented by the following steps:

[0090] constructing a data correction model according to the associated correction parameters;

[0091] The clinical data of the target respiratory patient is corrected by using the data correction model to obtain corrected clinical data.

[0092] In the specific implementation, first, a data correction model (such as a linear regression model, a neural network correction model, or a graph-based model) is selected, and the associated correction parameters are mapped to the constraint parameters of the data correction model. Further, the associations between different respiratory signs in the optimal dependency tree are mapped to the data correction model structure, so that the data correction model can capture the structural relationship between data features. It should be noted that in this application, the dependency between different levels or nodes can be set according to the weight of the associated correction parameters, and the associated correction parameters are used as coefficients of the correction model. If a linear model is used, the correction parameters are the linear weighted coefficients of the features. If a neural network model is used, the associated correction parameters are loaded into the network level to further refine the correction ability of the model; then, the clinical data of the target respiratory patient is used as input data and passed into the constructed data correction model. Further, the input data is corrected by the associated correction parameters through the data correction model. The data correction model will adjust each feature value accordingly according to the correction parameters to make it conform to the pathological feature dependency relationship and reduce errors or abnormal deviations. Finally, the proofread data is output through the data correction model. The proofread data can be directly used as the basic data for subsequent analysis after verification.

[0093] It should be noted that the proofread clinical data in this application can eliminate potential system errors and abnormal deviations while maintaining the original clinical information, providing more accurate basic data for the subsequent diagnosis and treatment of respiratory patients. It should also be noted that the data correction model ensures the validity and stability of the clinical data through structured proofreading processing.

[0094] It should also be noted that an electronic case report form for the target respiratory patient can be generated from the proofread data of the target respiratory patient's clinical data obtained after proofreading, that is, the proofread data of the clinical data obtained after proofreading can be organized and output into a standardized electronic case report. This report can be used to summarize and present the patient's condition, diagnostic information, and personalized treatment recommendations, etc., to help doctors and nursing teams more efficiently understand the patient's health condition and make scientific medical decisions.

[0095] On the other hand, in some embodiments, the present application provides a respiratory patient clinical data acquisition system, the respiratory patient clinical data acquisition system includes a data correction unit, reference Figure 4 , which is a schematic diagram of the structure of a data correction unit according to some embodiments of the present application, the data correction unit 400 includes: an acquisition module 401, a processing module 402 and an execution module 403, which are described as follows:

[0096] Acquisition module 401, in the present application, the acquisition module 401 is mainly used to receive a collection instruction of clinical data of respiratory patients, obtain the historical clinical data set of all respiratory patients through the collection instruction, and extract the test indicators of different respiratory signs from the historical clinical data set;

[0097] Processing module 402, in this application, processing module 402 is used to perform multi-level pathological feature association analysis on all test indicators based on the respiratory disease knowledge graph to obtain the dependency relationship of the pathological structure between each test indicator;

[0098] The processing module 402 in the present application is also used to perform association clustering on the pathological features of all respiratory signs to obtain multiple pathological feature clusters, and to construct a feature dependency tree of the pathological structure in each pathological feature cluster through the dependency relationship between the pathological structures of the various inspection indicators in each pathological feature cluster;

[0099] The processing module 402 in the present application is also used to obtain clinical data of the target respiratory patient according to the acquisition instruction, and to perform association matching on the clinical data according to all feature dependency trees, thereby obtaining association correction parameters for all respiratory signs of the target respiratory patient;

[0100] The execution module 403 in the present application is mainly used to calibrate the clinical data of the target respiratory patient using the associated correction parameters to obtain calibration data of the clinical data of the target respiratory patient.

[0101] In addition, the present application also provides a computer device, which includes a memory and a processor, wherein the memory stores codes, and the processor is configured to obtain the codes and execute the above-mentioned respiratory patient clinical data correction method.

[0102] In some embodiments, reference Figure 5 , which is a schematic diagram of the structure of a computer device for implementing a method for correcting clinical data of respiratory patients according to some embodiments of the present application. The method for correcting clinical data of respiratory patients in the above embodiments can be Figure 5 The computer device 500 shown in the figure is implemented, and the computer device 500 includes at least one processor 501, a communication bus 502, a memory 503 and at least one communication interface 504.

[0103] The processor 501 may be a general-purpose central processing unit (CPU) or an application-specific integrated circuit (ASIC).

[0104] The communication bus 502 may be used to transmit information between the above-mentioned components.

[0105] The memory 503 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, an optical disc storage (including a compressed optical disc, a laser disc, an optical disc, a digital versatile disc, a Blu-ray disc, etc.), a magnetic disk or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 503 may exist independently and be connected to the processor 501 via the communication bus 502. The memory 503 may also be integrated with the processor 501.

[0106] The memory 503 is used to store the program code for executing the solution of the present application, and the execution is controlled by the processor 501. The processor 501 is used to execute the program code stored in the memory 503. The program code may include one or more software modules. The clinical data correction method for respiratory patients in the above embodiment can be implemented by the processor 501 and one or more software modules in the program code in the memory 503.

[0107] The communication interface 504 uses any transceiver or other device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0108] In a specific implementation, as an embodiment, a computer device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (such as computer program instructions).

[0109] The above-mentioned computer device may be a general-purpose computer device or a special-purpose computer device. In a specific implementation, the computer device may be a desktop computer, a portable computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device or an embedded device. The embodiment of the present application does not limit the type of computer device.

[0110] In addition, the present application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned method for correcting clinical data of respiratory patients is implemented.

[0111] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0112] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for correcting clinical data of respiratory patients, used for data correction in the process of collecting clinical data of respiratory patients, characterized in that: The steps include: Receiving a collection instruction of clinical data of respiratory patients, obtaining historical clinical data sets of all respiratory patients through the collection instruction, and extracting test indicators of different respiratory signs from the historical clinical data sets; Based on the knowledge graph of respiratory diseases, a multi-level pathological feature correlation analysis is performed on all test indicators to obtain the dependency relationship of the pathological structure between each test indicator; Performing association clustering on the pathological features of all respiratory signs to obtain multiple pathological feature clusters, and constructing a feature dependency tree of the pathological structure in each pathological feature cluster through the dependency relationship of the pathological structure between each test index in each pathological feature cluster; Acquire clinical data of a target respiratory patient according to the acquisition instruction, perform correlation matching on the clinical data according to all feature dependency trees, and thereby obtain correlation correction parameters of all respiratory signs of the target respiratory patient; Proofreading the clinical data of the target respiratory patient using the associated correction parameters to obtain proofreading data of the clinical data of the target respiratory patient; Among them, based on the knowledge graph of respiratory diseases, a multi-level pathological feature correlation analysis was performed on all the test indicators, and the dependency relationship of the pathological structure between each test indicator was obtained, including: Mapping all test indicators into the respiratory disease knowledge graph; All the test indicators are correlated from the symptom surface, structural middle and pathological deep layers, and then the characteristic correlations of different levels between the test indicators are extracted; The pathological structure dependency between each test index is determined based on all feature correlations.

2. The method according to claim 1, characterized in that The pathological characteristics of all respiratory signs are clustered and multiple pathological characteristic clusters are obtained, including: Extracting the pathological characteristics of each respiratory sign from the test index of each respiratory sign; Based on the normal distribution assumption, correlation analysis was performed on all pathological features to obtain the mutual correlation between pathological features; All pathological features are hierarchically clustered according to the mutual correlation to obtain multiple pathological feature clusters.

3. The method according to claim 1, characterized in that Constructing a feature dependency tree of the pathological structure in each pathological feature cluster by using the dependency relationship of the pathological structure between each inspection index in each pathological feature cluster specifically includes: Determine the associated attention for each pathological feature cluster; The feature dependency tree of the pathological structure in each pathological feature cluster is determined by the dependency relationship of the pathological structure between each inspection index in each pathological feature cluster and the associated attention of each pathological feature cluster.

4. The method according to claim 1, characterized in that Also includes: An electronic case report form of the target respiratory patient is generated from the collated data of the clinical data of the target respiratory patient obtained after the collation.

5. The method according to claim 1, characterized in that The respiratory signs are physiological symptom indicators that reflect the functional state of the respiratory system.

6. The method according to claim 1, characterized in that The test index is an index reflecting changes in the physiological state of a respiratory patient.

7. A respiratory patient clinical data acquisition system, which uses the method according to any one of claims 1 to 6 to perform data correction, and the respiratory patient clinical data acquisition system includes a data correction unit, characterized in that: The data correction unit comprises: An acquisition module, used for receiving an acquisition instruction of clinical data of respiratory patients, acquiring historical clinical data sets of all respiratory patients through the acquisition instruction, and extracting test indicators of different respiratory signs from the historical clinical data sets; The processing module is used to perform multi-level pathological feature correlation analysis on all test indicators based on the knowledge graph of respiratory diseases to obtain the dependency relationship of the pathological structure between various test indicators; The processing module is further used to perform association clustering on the pathological features of all respiratory signs to obtain multiple pathological feature clusters, and to construct a feature dependency tree of the pathological structure in each pathological feature cluster through the dependency relationship of the pathological structure between each inspection index in each pathological feature cluster; The processing module is further used to obtain clinical data of the target respiratory patient according to the acquisition instruction, and to perform association matching on the clinical data according to all feature dependency trees, thereby obtaining association correction parameters of all respiratory signs of the target respiratory patient; The execution module is used to perform proofreading processing on the clinical data of the target respiratory patient according to the associated correction parameters to obtain proofreading data of the clinical data of the target respiratory patient.

8. A computer device, comprising a memory and a processor, wherein the memory stores a code, characterized in that: The processor is configured to obtain the code and execute the respiratory patient clinical data correction method according to any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for correcting clinical data of a respiratory patient according to any one of claims 1 to 6 is implemented.

Images