Critical patient past medical history tracing system and method

Abstract

The invention relates to the technical field of electric digital data processing, in particular to a critical patient past medical history tracing system and method, and the system comprises a data acquisition module, a data association module, a dynamic weight distribution module, a clustering analysis module, a time axis sorting module and a generation module. According to the invention, through data association and a dynamic weight distribution mechanism, deep mining and real-time updating of critical disease history data are realized; and meanwhile, the generation module is combined to generate a simple abstract, so that the information transmission efficiency is improved, and the practicability and the popularization value of the system are remarkably enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of digital data processing, and in particular to a critical patient past medical history tracing system and a tracing method. BACKGROUND

[0002] With the continuous development of medical informatization, critical patient medical history tracing plays a crucial role in improving diagnosis and treatment efficiency and accuracy. In complex medical scenarios, comprehensive and accurate tracing of patients' past medical history, especially multi-dimensional data integration and analysis, is a key link to optimize diagnosis and treatment decisions. However, the existing medical history tracing system still has significant shortcomings in data integration capability, analysis depth, and operation convenience, which is difficult to meet the special needs of critical patient management.

[0003] The existing tumor susceptible microenvironment detection device and kit with publication number CN106885910B and publication date December 21, 2018. The scheme comprehensively analyzes physiological indicators, living habits, and tumor history of three generations of direct relatives in the family through detection unit and analysis unit, mainly focusing on the detection of tumor susceptible microenvironment. However, this scheme has limited applicability in critical patient medical history tracing and does not cover multi-dimensional medical history data such as past diagnosis and treatment records, medication history, and multi-generation family history. In addition, this technology lacks support for dynamic updating and real-time analysis of data, making it difficult to meet the needs of rapid changes in critical patient conditions and unable to meet the requirements of efficient management and accurate diagnosis and treatment in complex medical scenarios. SUMMARY

[0004] The present application relates to the technical field of digital data processing, and in particular to a critical patient past medical history tracing system and a tracing method.

[0005] The present application relates to the technical field of digital data processing, and in particular to a critical patient past medical history tracing system and a tracing method.

[0006] According to an aspect of the present application, a critical patient past medical history tracing system is provided, comprising a processor, and further comprising a data acquisition module, a data association module, a dynamic weight distribution module, a clustering analysis module, a time axis sorting module and a generation module connected with the processor; wherein the data acquisition module is connected with a hospital information database through a standardized interface; the data association module is used for calculating the correlation between data nodes; the dynamic weight distribution module is used for adjusting the dynamic weight according to the importance and real-time changes of the data; the clustering analysis module is used for clustering the data nodes to generate a medical history tracing map; the time axis sorting module is used for sorting the key nodes in the medical history tracing map according to time axis; and the generation module is used for performing semantic analysis on the medical history data corresponding to each time node and generating an abstract.

[0007] In some embodiments, the data acquisition module is at least configured with a data parser and a data synchronization unit; the data parser uniformly converts medical data in different formats into structured data and stores them in a local cache, and the data synchronization unit monitors the changes of an external database in real time and synchronizes the latest data to the local cache.

[0008] In some embodiments, the data association module is at least configured with an association degree calculation model, and the association degree model expression is: , wherein, represents the association degree between data nodes and . represents the similarity of two nodes in the th dimension; is the dynamic weight of the th dimension; is the total number of dimensions.

[0009] In some embodiments, the dynamic weight distribution module is at least configured with a weight adjustment unit, which is connected with the data association module, and is used for adjusting the dynamic weight according to the importance and real-time changes of the data.

[0010] In some embodiments, the clustering analysis module is configured with a hierarchical clustering model, which is used to determine the adjacent node set of each data node through the association degree matrix, and to classify nodes with high correlation into the same cluster.

[0011] In some embodiments, the time axis sorting module is configured with a sorting model, which reevaluates the weight of each time node through a dynamic weight adjustment mechanism.

[0012] In some embodiments, the generation module is configured with at least a natural language processing unit, a topic modeling unit, a keyword extraction unit, and an abstract generation unit; wherein the natural language processing unit preprocesses the original medical history data through word segmentation, part-of-speech tagging, and entity recognition; the topic modeling unit identifies core topics in the data through a latent Dirichlet allocation algorithm; the keyword extraction unit extracts key information through a TF-IDF algorithm; and the abstract generation unit generates a concise abstract text based on the results of the topic modeling unit and the keyword extraction unit.

[0013] According to another aspect of the present application, a method for tracing past medical history of a critical patient is provided, which is executed by a processor and includes: obtaining multi-dimensional medical history data of the critical patient; based on the multi-dimensional medical history data, performing random screening of key data nodes at different time points and marking them as ; based on the key data nodes , performing calculation of correlation degrees between the data nodes; based on the correlation degree results, performing cluster analysis on the data nodes to generate a medical history tracing graph; performing time axis sorting on the key nodes in the medical history tracing graph; and performing semantic analysis on the medical history data corresponding to each time node, extracting core information and generating an abstract.

[0014] In some embodiments, the calculation of the correlation degrees between the data nodes adopts a formula as follows: , wherein, represents the correlation degree between the data nodes and ; and represents the similarity of the two nodes in the kth dimension. is the dynamic weight of the kth dimension; and m is the total number of dimensions.

[0015] According to the one or more technical solutions adopted by the present application, the following beneficial effects can be achieved: the present application realizes deep mining and real-time updating of medical history data of a critical patient through a data correlation and dynamic weight distribution mechanism; meanwhile, a concise abstract is generated by the generation module, which not only improves the efficiency of information transmission, but also significantly enhances the practicality and promotional value of the system. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor on the basis of these drawings.

[0017] Figure 1is a structural diagram of a previous medical history tracing system for critical patients of the present application; Figure 2 is a flow chart of a previous medical history tracing method for critical patients of the present application. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present application. Figures 1-2 The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present application.

[0019] Embodiment One: Figure 1 A structural diagram of a previous medical history tracing system for critical patients is provided in the embodiments of the present application, and the system comprises a processor. The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor, or the processor can also be any conventional processor, etc.

[0020] The system further comprises a data acquisition module, a data association module, a dynamic weight allocation module, a clustering analysis module, a time axis sorting module and a generation module which are in data connection with the processor, and the foregoing modules cooperate with each other to complete the whole process from data acquisition to medical history tracing. The data acquisition module extracts multi-dimensional medical history data of critical patients from multiple sources, including previous diagnosis and treatment records, medication history, family medical history and living habits, etc., and the foregoing data are at least from a hospital information system (HIS), an electronic medical record system (EMR), a laboratory information system (LIS) and other medical related databases. The data acquisition module is connected with the foregoing hospital information databases through a standardized interface, and ensures the integrity and consistency of the data. In actual operation, the data acquisition module will automatically screen out the required data set according to the preset rules, and the data set is: randomly screening out key data nodes, and corresponding marking as , wherein the subscripts 0, 1, …, n correspond to Key data nodes at each time point. The data association module connects to the data acquisition module to calculate the correlation between data nodes. The dynamic weight allocation module connects to the data association module to adjust dynamic weights based on data importance and real-time changes. The clustering analysis module connects to the data association module to cluster data nodes and generate a medical history tracing map. The timeline sorting module connects to the clustering analysis module to sort key nodes in the medical history tracing map over time. The generation module connects to the timeline sorting module to perform semantic analysis on the medical history data corresponding to each time point and generate a summary.

[0021] Furthermore, the data acquisition module is equipped with at least a data parser and a data synchronization unit. The data parser converts medical data of different formats into structured data and stores it in a local cache. The local cache is connected to the processor through the data synchronization unit. The data synchronization unit monitors changes in the external database in real time and synchronizes the latest data to the local cache to ensure data integrity and consistency.

[0022] Furthermore, the data association algorithm module calculates the correlation degree of the aforementioned key data nodes. This process is one of the core components of the entire system, aiming to reveal potential connections between different time points. The data association module is configured with at least a correlation degree calculation model, the expression of which is: ,in, Represents data nodes and The degree of correlation between them; This represents the similarity between two nodes in the k-th dimension; Let be the dynamic weight of the k-th dimension; m be the total number of dimensions. In practical applications, similarity... The dynamic weight allocation module can be calculated using various methods, such as cosine similarity or Euclidean distance, with the specific choice depending on the data characteristics and application scenario. Furthermore, the module includes at least a weight adjustment unit connected to the data association module. This unit adjusts the dynamic weights based on the importance of the data and real-time changes. Specifically, it generates dynamic weight values ​​by calculating the access frequency, update frequency, and clinical need priority of key data nodes within a specific time period and feeds these values ​​back to the data association algorithm module. The dynamic weights... A dynamic weighting adjustment mechanism is then used to adjust the weighting based on the importance of the data and real-time changes to ensure that the correlation calculation reflects the latest clinical needs. For example, for critically ill patients, recent medication history and vital sign data may have higher weighting, while earlier records of lifestyle habits may have relatively lower weighting.

[0023] Further, the clustering analysis module is configured with a hierarchical clustering model, which is used to determine the neighboring node set of each data node through the correlation matrix, and nodes with high correlation are classified into the same cluster. It can be understood that, based on the aforementioned correlation calculation results, the clustering analysis module performs clustering processing on the data nodes, thereby generating a medical history tracing map. The medical history tracing map contains multiple time nodes and their corresponding key nodes, and the association relationship between nodes is represented by a connection line. Among them, the specific steps of clustering analysis include: first, determining the neighboring node set of each data node according to the correlation matrix; second, using a hierarchical clustering algorithm or other suitable clustering method to classify nodes with high correlation into the same cluster; finally, integrating all clusters into a complete medical history tracing map. This process not only reveals the internal structure of the medical history data, but also helps medical staff quickly locate important time nodes.

[0024] Further, the time axis sorting module is configured with a sorting model, which reevaluates the weight of each time node through a dynamic weight adjustment mechanism. It can be understood that the sorting model of the time axis sorting module sorts the key nodes in the medical history tracing map according to the time axis. The purpose of time axis sorting is to arrange the data according to the chronological order of the time nodes, so as to facilitate subsequent analysis and display. In this process, the dynamic weight adjustment further plays a role, by reevaluating the weight of each time node to determine its importance in the medical history tracing. For example, certain time nodes may be assigned higher weights because they involve key treatment decisions or turning points in the patient's condition, and the aforementioned dynamic adjustment mechanism enables the system to flexibly respond to the needs of critical patients with rapidly changing conditions.

[0025] Further, the generation module is at least configured with a natural language processing unit, a topic modeling unit, a keyword extraction unit, and an abstract generation unit. The natural language processing unit preprocesses the original medical history data through word segmentation, part-of-speech tagging, and entity recognition. The topic modeling unit identifies the core topics in the data through the latent Dirichlet allocation algorithm. The keyword extraction unit extracts key information through the TF-IDF algorithm. The abstract generation unit generates a concise abstract text based on the results of the topic modeling unit and the keyword extraction unit. It can be understood that after completing the time axis sorting, the generation module performs semantic analysis on the medical history data corresponding to each time node, extracts core information, and generates an abstract. The specific steps of semantic analysis include: first, using natural language processing technology to perform word segmentation, part-of-speech tagging, and entity recognition on the original medical history data; second, identifying the core topics and key information in the data through topic modeling or keyword extraction methods; and finally, generating a concise and concise abstract text based on the extraction results. In practical applications, the generation module can customize the output format according to user needs, such as presenting key indicators in table form or summarizing main findings in paragraph form. This way not only improves the efficiency of information transmission, but also reduces the workload of medical staff.

[0026] To verify the effectiveness of the system, the present application selects a certain tertiary hospital intensive care unit (ICU) as a pilot unit for a 5-month practical application test. During the test, the system processed 50 cases of critical patient medical history data, covering cardiovascular diseases, respiratory failure, multiple organ dysfunction syndrome and other complex cases. The test results show that the system has a significant advantage in data integration and analysis. For example, in the medical history tracing of a patient with acute myocardial infarction, the system successfully identified the patient's multiple chest pain episodes in the past six months, and found that these records had a high degree of correlation with the final diagnosis through correlation calculation. In addition, the system generated a medical history tracing map that clearly showed the key time nodes of the patient's condition development, providing an important reference for doctors to develop treatment plans. At the same time, in the case of a patient with multiple organ dysfunction syndrome, the system prioritized recent laboratory test results and medication records through a dynamic weight adjustment mechanism, and generated a detailed condition abstract using semantic analysis technology. This abstract not only summarizes the patient's main symptoms and treatment measures, but also points out potential risk factors and precautions, greatly improving the work efficiency of medical staff. The test results show that the system has a significant effect on improving the comprehensiveness and accuracy of medical history tracing, and its response speed and operational convenience have also been highly evaluated by users.

[0027] In some preferred embodiments, the data association module further includes a dynamic similarity correction system, which includes a similarity correction unit and a data drift detection unit. The similarity correction unit corrects the similarity value by calculating the drift coefficient of the data node in different time periods. Specifically, it predicts the changing trend of the data node through a linear regression model and adjusts the similarity value in combination with the actual data deviation. The data drift detection unit monitors the change amplitude of the data node through a sliding window algorithm. When the change amplitude exceeds a preset threshold, it triggers the correction operation of the similarity correction unit.

[0028] In some preferred embodiments, the clustering analysis module further includes a multi-scale clustering optimization system, which includes a scale selection unit and a clustering result verification unit. The scale selection unit selects the optimal clustering scale by calculating the density distribution of data nodes. Specifically, it calculates the density function of data nodes using a kernel density estimation algorithm and selects the clustering scale based on the density peak. The clustering result verification unit evaluates the stability of the clustering results using a cross-validation algorithm. When the stability of the clustering results is lower than a preset threshold, the scale selection unit is triggered to reselect the clustering scale.

[0029] Example 2: Based on the same inventive concept as the critical patient medical history tracing system in Example 1, such as... Figure 2 As shown, the present invention also provides a method for tracing the past medical history of critically ill patients. This method is executed by a processor, which can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0030] Specifically, this includes: acquiring multi-dimensional medical history data of critically ill patients; and based on this multi-dimensional medical history data, randomly selecting key data nodes from different time points and marking them accordingly. Based on key data nodes The process involves calculating the correlation between data nodes; performing cluster analysis on the data nodes based on the correlation results to generate a medical history tracing map; sorting the key nodes in the medical history tracing map along a timeline; and performing semantic analysis on the medical history data corresponding to each time node to extract core information and generate a summary. for Key data nodes at any given moment.

[0031] Furthermore, the formula for calculating the correlation between data nodes is as follows: ,in, Represents data nodes and The degree of correlation between them; This represents the similarity between two nodes in the k-th dimension; For the first Dynamic weights for each dimension; This represents the total number of dimensions.

[0032] The specific example of the critical patient past medical history tracing system in the aforementioned embodiment 1 is also applicable to the critical patient past medical history tracing method in this embodiment. Through the aforementioned detailed description of the critical patient past medical history tracing system, those skilled in the art can clearly understand the critical patient past medical history tracing method in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0033] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered illustrative and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the scope of the invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0034] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A critical illness patient history tracking system comprising a processor, wherein, Also comprising a data acquisition module, a data association module, a dynamic weight allocation module, a clustering analysis module, a timeline sorting module and a generation module in data connection with the processor; The data acquisition module is connected with the hospital information database through a standardized interface; The data association module is used to calculate the correlation between data nodes; The dynamic weight allocation module is used to adjust the dynamic weight according to the importance and real-time changes of the data; The clustering analysis module is used to cluster the data nodes and generate a medical history tracking map; The timeline sorting module is used to sort the key nodes in the medical history tracking map by time axis; The generation module is used to perform semantic analysis on the medical history data corresponding to each time node and generate an abstract.

2. The system of claim 1, wherein, The data acquisition module is at least configured with a data parser and a data synchronization unit; the data parser uniformly converts medical data of different formats into structured data and stores them in the local cache, and the data synchronization unit monitors the changes of the external database in real time and synchronizes the latest data to the local cache.

3. The system of claim 1, wherein, The data association module is configured at least in relation to a correlation degree calculation model, and the correlation degree model expression is: , wherein, represents the association degree between data nodes and ; represents the similarity of two nodes in the kth dimension; is the dynamic weight of the kth dimension; m is the total number of dimensions.

4. The system of claim 3, wherein, The dynamic weight allocation module is at least configured with a weight adjustment unit, which is in data connection with the data association module; the weight adjustment unit is used to adjust the dynamic weight according to the importance and real-time changes of the data.

5. The system of claim 1, wherein, The clustering analysis module is configured with a hierarchical clustering model, which is used to determine the adjacent node set of each data node through the correlation matrix and classify nodes with high correlation into the same cluster.

6. The system of claim 5, wherein, The timeline sorting module is configured with a sorting model, which reevaluates the weight of each time node through a dynamic weight adjustment mechanism.

7. The system of claim 1, wherein, The generation module is at least configured with a natural language processing unit, a topic modeling unit, a keyword extraction unit and an abstract generation unit; wherein the natural language processing unit preprocesses the original medical history data through word segmentation, part-of-speech tagging and entity recognition; the topic modeling unit identifies the core topics in the data through the latent Dirichlet allocation algorithm; the keyword extraction unit extracts key information through the TF-IDF algorithm; and the abstract generation unit generates a concise abstract text based on the results of the topic modeling unit and the keyword extraction unit.

8. A method of critical illness history tracking, the method performed by a processor, the method comprising: It includes: Obtaining multidimensional medical history data of critical patients; Based on multi-dimensional medical history data, randomly screen out key data nodes at different times, and mark them as D n ; Based on key data nodes D n , the correlation between each data node is calculated Based on the correlation results, the data nodes are clustered and analyzed to generate a medical history tracking map; Performing time axis sorting on key nodes in the medical history tracking map; Performing semantic analysis on the medical history data corresponding to each time node, extracting core information and generating an abstract.

9. The method of claim 8, wherein, The execution calculates the correlation degree between each data node by using the formula: , wherein, represents the association degree between data nodes and ; represents the similarity of two nodes in the kth dimension; is the dynamic weight of the kth dimension; m is the total number of dimensions.

Images