Critical disease risk assessment model construction system

Abstract

The invention belongs to the technical field of medical service management, and particularly relates to a critical disease risk assessment model construction system which comprises a model construction operation end and an operation monitoring management end. The model construction operation end comprises a critical symptom data integration processing unit, a critical symptom feature selection extraction unit, a risk assessment model construction optimization unit, a patient information input transmission unit and a clinical decision support output unit; according to the invention, the risk assessment model of critical diseases is constructed through the risk assessment model construction optimization unit, the risk assessment model analyzes patient information input by medical personnel, and the clinical decision support output unit sends a critical disease risk assessment report of a corresponding patient to the corresponding medical terminal for display. According to the invention, doctors can be helped to formulate more scientific and reasonable treatment schemes, and operation monitoring and abnormity judgment early warning are carried out on the model construction operation end through the operation monitoring management end, so that efficient, smooth and stable operation of medical work can be ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of medical service management, and in particular to a system for constructing a risk assessment model for critical illness. Background Art

[0002] Critical illness refers to an emergency condition that is so severe that it threatens the patient's life. This condition is usually accompanied by unstable vital signs and damage to multiple organ functions. The patient may develop serious complications such as heart failure, respiratory failure, liver and kidney dysfunction, and may even be in danger of death. Risk assessment of critically ill patients is a crucial link in medical practice; Traditional risk assessment methods mainly rely on the clinical experience of doctors, which are highly subjective and inaccurate. With the rapid development of medical big data and artificial intelligence technology, critical illness risk assessment models have been widely used in clinical practice. However, the construction of existing critical illness risk assessment models is relatively complex, and it is difficult to effectively protect and monitor the subsequent operation of the model. Administrators are unable to make reasonable optimization and improvement measures in a timely manner, which is not conducive to ensuring the efficient, smooth and stable progress of medical work. In view of the above technical defects, a solution is now proposed. Summary of the invention

[0003] The purpose of the present invention is to provide a critical illness risk assessment model construction system, which solves the problem that the construction of existing critical illness risk assessment models is relatively complex, and it is difficult to effectively protect and monitor the subsequent operation of the model, which is not conducive to ensuring efficient, smooth and stable medical work.

[0004] To achieve the above object, the present invention provides the following technical solutions: A critical illness risk assessment model construction system includes a model construction operation end and an operation monitoring and management end. The operation monitoring and management end monitors and manages the operation of the model construction operation end, and the model construction operation end includes a critical illness data integration processing unit, a critical illness feature selection and extraction unit, a risk assessment model construction optimization unit, a patient information input and transmission unit, and a clinical decision support output unit; the critical illness data integration processing unit collects multi-dimensional data of critical illness patients from multiple medical information systems, pre-processes the data using data cleaning, deduplication and normalization technologies, and sends the pre-processed data to the critical illness feature selection and extraction unit; The critical illness feature selection and extraction unit uses feature selection algorithms and feature extraction techniques to extract key features that have an important impact on critical illness risk assessment from the preprocessed data, and sends the feature selection and extraction results to the risk assessment model construction and optimization unit; The risk assessment model construction optimization unit uses the ensemble learning method to construct the risk assessment model based on the extracted features and stores the constructed risk assessment model. It trains the model and optimizes its parameters through cross-validation and grid search technology, and supports online updating and iterative optimization of the risk assessment model. The patient information input and transmission unit is used for medical personnel to input patient information. The risk assessment model construction and optimization unit analyzes patient information based on the constructed risk assessment model, and sends the analysis results to the clinical decision support output unit; the clinical decision support output unit clinically interprets the output results of the risk assessment model, generates an intuitive critical illness risk assessment report, and sends the critical illness risk assessment report of the corresponding patient to the corresponding medical terminal for display.

[0005] Furthermore, the medical information system includes an electronic medical record system, a laboratory information system, and a medical imaging system, and the multi-dimensional data collected include medical history records, laboratory test results, imaging examination reports, and physiological monitoring data.

[0006] Furthermore, the critical illness risk assessment report includes the patient's risk level information, key risk characteristic information, potential complication prediction information and personalized treatment recommendation information.

[0007] Furthermore, the operation monitoring and management end includes a security protection unit, a hidden danger analysis unit and an immediate alarm unit. The security protection unit is used to set up a secure local area network. The model building operation end uses the set secure local area network for operation. The security protection unit is also used to register medical terminals and medical personnel, mark successfully registered medical terminals and medical personnel as safe terminals and safe personnel, and store the information of the safe terminals and safe personnel, and only allow safe personnel to log in through the safe terminals; The hidden danger analysis unit analyzes the operational hidden dangers of the model construction operation end, and determines whether to generate a hidden danger warning signal through analysis. When the hidden danger warning signal is generated, it is sent to the immediate alarm unit. When the immediate alarm unit receives the hidden danger warning signal, it sends an alarm message to the corresponding administrator.

[0008] Furthermore, the specific analysis process of the hidden danger analysis unit includes: The frequency of disconnection of the secure LAN and the frequency of external attacks during the detection period are collected and marked as a LAN disconnection value and a LAN attack measurement value, and the LAN disconnection value and the LAN attack measurement value are numerically compared with a preset LAN disconnection threshold and a preset LAN attack measurement threshold, respectively. If the LAN disconnection value or the LAN attack measurement value exceeds the corresponding preset threshold, a hidden danger warning signal is generated; If the LAN disconnection value and the LAN attack detection value do not exceed the corresponding preset threshold value, the total duration of LAN disconnection and the total duration of LAN being attacked by external forces during the detection period are collected and marked as the LAN opening time value and the LAN attack time value respectively; and the real-time flow value of the LAN is collected, and the real-time flow value is compared with the preset real-time flow value range. If the real-time flow value is not within the preset real-time flow value range, it is determined that the corresponding moment is in an abnormal flow state; The total duration that the LAN is in an abnormal traffic state during the detection period is obtained and marked as the traffic characteristic value. The hidden danger characteristic coefficient is obtained by numerically calculating the LAN opening time value, the LAN attack time value and the traffic characteristic value. The hidden danger characteristic coefficient is numerically compared with the preset hidden danger characteristic coefficient threshold. If the hidden danger characteristic coefficient exceeds the preset hidden danger characteristic coefficient threshold, a hidden danger warning signal is generated.

[0009] Furthermore, the hidden danger analysis unit is communicatively connected to the operation efficiency analysis unit. If the hidden danger analysis unit does not generate a hidden danger warning signal, the operation effect of the risk assessment model during the detection period is analyzed by the operation efficiency analysis unit to determine whether to generate an operation efficiency warning signal, and the operation efficiency warning signal is sent to the immediate alarm unit. When the immediate alarm unit receives the operation efficiency warning signal, it sends an alarm message to the corresponding administrator.

[0010] Furthermore, the specific analysis process of the operation efficiency analysis unit is as follows: The moment when the risk assessment model receives the input patient information is obtained and marked as the information arrival moment, and the moment when the clinical decision support output unit outputs the critical risk assessment report is obtained and marked as the report generation moment, and the time difference between the report generation moment and the information arrival moment is calculated to obtain the assessment duration; the assessment duration is numerically compared with the preset assessment duration threshold, and if the assessment duration exceeds the preset assessment duration threshold, the corresponding assessment duration is marked as the characteristic duration; The number of characteristic durations in the detection period is obtained and the ratio thereof is calculated with the number of evaluation durations to obtain the characteristic duration table value, and the characteristic duration table value is numerically compared with the preset characteristic duration table threshold. If the characteristic duration table value exceeds the preset characteristic duration table threshold, an operation efficiency warning signal is generated; If the characteristic time table value does not exceed the preset characteristic time table threshold, the average of all evaluation durations in the detection period is calculated to obtain the evaluation time measurement value, and the operation efficiency correlation value is obtained by weighted summing the characteristic time table value and the evaluation time measurement value. The operation efficiency correlation value is numerically compared with the preset operation efficiency correlation threshold. If the operation efficiency correlation value exceeds the preset operation efficiency correlation threshold, an operation efficiency warning signal is generated.

[0011] Furthermore, the operation efficiency analysis unit is connected to the pressure analysis unit. If no operation efficiency warning signal is generated during the detection period, the operation pressure of the risk assessment model during the detection period is analyzed by the pressure analysis unit. The analysis is used to determine whether a pressure warning signal is generated, and the pressure warning signal is sent to the immediate alarm unit. When the immediate alarm unit receives the pressure warning signal, it sends an alarm message to the corresponding administrator.

[0012] Furthermore, the specific analysis process of the pressure analysis unit is as follows: The number of real-time tasks waiting for evaluation by the risk assessment model is obtained and marked as a queuing coefficient, and the queuing coefficient is numerically compared with a preset queuing coefficient threshold. If the queuing coefficient exceeds the preset queuing coefficient threshold, it is determined to be in a high-pressure operation state; Several detection time periods are set within the detection period, the total duration of the operation in the high-pressure state in the corresponding detection time period is collected and marked as a high-pressure characteristic value, the number of occurrences of the single duration of the operation in the high-pressure state in the corresponding detection time period exceeding the preset single duration threshold is collected and marked as a high-pressure risk value, and the maximum value of the single duration of the operation in the high-pressure state in the corresponding detection time period is marked as a high-pressure amplitude value; The high-pressure decision value is obtained by numerically calculating the high-pressure characteristic value, the high-pressure risk value and the high-pressure amplitude value, and the high-pressure decision value is numerically compared with the preset high-pressure decision threshold. If the high-pressure decision value exceeds the preset high-pressure decision threshold, the corresponding detection period is marked as a high-pressure period; if the high-pressure decision value does not exceed the preset high-pressure decision threshold, the corresponding detection period is marked as a low-pressure period; The number of high-pressure periods in the detection period is obtained and marked as high-pressure detection values, and the high-pressure detection values ​​are numerically compared with the preset high-pressure detection threshold. If the high-pressure detection value exceeds the preset high-pressure detection threshold, a pressure warning signal is generated.

[0013] Furthermore, if the high-pressure detection value does not exceed the preset high-pressure detection threshold, the number of high-pressure periods between two adjacent groups of low-pressure periods is marked as a high-pressure duration value, and the high-pressure duration value is numerically compared with the preset high-pressure duration threshold. If the high-pressure duration value exceeds the preset high-pressure duration threshold, the corresponding high-pressure duration value is marked as a high-pressure maintenance inspection value, and the number of high-pressure maintenance inspection values ​​in the detection period is obtained and marked as a high-pressure maintenance risk value; The high-pressure decision values ​​of all detection periods are averaged to obtain the high-pressure performance value, the high-pressure analysis coefficient is obtained by numerically calculating the high-pressure detection value, the high-pressure risk value and the high-pressure performance value, and the high-pressure analysis coefficient is numerically compared with the preset high-pressure analysis coefficient threshold. If the high-pressure analysis coefficient exceeds the preset high-pressure analysis coefficient threshold, a pressure warning signal is generated.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. In the present invention, the critical disease data integration processing unit collects data and performs preprocessing, the critical disease feature selection and extraction unit extracts key features from the preprocessed data, the risk assessment model construction optimization unit constructs a risk assessment model based on the extracted features, and the patient information entered by the medical staff is analyzed based on the risk assessment model and a critical disease risk assessment report is output, so as to achieve a comprehensive and accurate assessment of the risks of critically ill patients, which is conducive to helping doctors formulate more scientific and reasonable treatment plans and improve the diagnostic efficiency and accuracy of medical staff; 2. In the present invention, the security protection unit is used to perform security protection on the model construction and operation end, the hidden danger analysis unit analyzes the operation hidden dangers of the model construction and operation end, and when no hidden danger warning signal is generated, the operation effect of the risk assessment model during the detection period is analyzed by the operation efficiency analysis unit. When no operation efficiency warning signal is generated, the operation pressure of the risk assessment model during the detection period is analyzed by the pressure analysis unit, thereby realizing effective monitoring of the model construction and operation end and abnormal judgment and warning, which is conducive to the administrator to make reasonable optimization and improvement measures in time to ensure the efficient, smooth and stable progress of medical work. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings; Figure 1 This is a system block diagram of Embodiment 1 of the present invention; Figure 2 This is a system block diagram of Embodiment 2 and Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] Embodiment 1: Figure 1 As shown, a critical illness risk assessment model construction system proposed by the present invention includes a model construction operation end and an operation monitoring and management end, the operation monitoring and management end monitors and manages the operation of the model construction operation end, and the model construction operation end includes a critical illness data integration processing unit, a critical illness feature selection and extraction unit, a risk assessment model construction optimization unit, a patient information input and transmission unit and a clinical decision support output unit; The critical care data integration processing unit collects multi-dimensional data of critical care patients (including medical history records, laboratory test results, imaging examination reports and physiological monitoring data, etc.) from multiple medical information systems (including electronic medical record systems, laboratory information systems and medical imaging systems, etc.), and pre-processes the data using data cleaning, deduplication and normalization techniques to ensure data quality and consistency, and sends the pre-processed data to the critical care feature selection and extraction unit.

[0018] The critical illness feature selection and extraction unit uses feature selection algorithms (such as recursive feature elimination, model-based feature selection, etc.) and feature extraction techniques (such as principal component analysis, deep learning feature extraction, etc.) to extract key features that have an important impact on critical illness risk assessment from the preprocessed data. These features include both static physiological indicators and medical history information, as well as dynamic time series data and imaging features, and the feature selection and extraction results are sent to the risk assessment model construction and optimization unit.

[0019] The risk assessment model construction optimization unit constructs a risk assessment model based on the extracted features using ensemble learning methods (such as random forest, gradient boosting tree, XGBoost, etc.) and stores the constructed risk assessment model. It trains the model and optimizes its parameters through cross-validation and grid search techniques to improve the accuracy and generalization ability of the model. It also supports online updating and iterative optimization of the risk assessment model to adapt to the ever-changing medical environment and patient characteristics.

[0020] The patient information input and transmission unit is used by medical personnel to input patient information. The risk assessment model construction and optimization unit analyzes the patient information based on the constructed risk assessment model and sends the analysis results to the clinical decision support output unit. The clinical decision support output unit clinically interprets the output results of the risk assessment model and generates an intuitive critical illness risk assessment report, wherein the critical illness risk assessment report includes the patient's risk level information, key risk feature information, potential complication prediction information and personalized treatment recommendation information. The critical illness risk assessment report of the corresponding patient will be sent to the corresponding medical terminal for display, helping doctors to formulate more scientific and reasonable treatment plans, achieving a comprehensive and accurate assessment of the risks of critically ill patients, providing strong support for clinical decision-making, and improving the diagnostic efficiency and accuracy of medical staff.

[0021] Furthermore, the operation monitoring and management end includes a security protection unit, a hidden danger analysis unit and an immediate alarm unit. The security protection unit is used to set up a secure local area network. The model building and operation end uses the set secure local area network for operation, which is conducive to ensuring network security. The security protection unit is also used to register medical terminals and medical personnel, mark the successfully registered medical terminals and medical personnel as security terminals and security personnel, and store the information of security terminals and security personnel, allowing only security personnel to log in through security terminals. Unrelated personnel or other terminals cannot log in, so as to ensure safety of use and patient privacy.

[0022] The hidden danger analysis unit analyzes the operation hidden dangers of the model construction operation end, and determines whether to generate a hidden danger warning signal through analysis. When the hidden danger warning signal is generated, it is sent to the immediate alarm unit. When the immediate alarm unit receives the hidden danger warning signal, it sends an alarm message to the corresponding administrator to remind the administrator to take reasonable improvement measures, strengthen the subsequent supervision of the model construction operation end, and ensure the subsequent operation safety and stability. The specific analysis process of the hidden danger analysis unit is as follows: The frequency of disconnection of the secure LAN and the frequency of external attacks during the detection period are collected and marked as a LAN disconnection value and a LAN attack measurement value, and the LAN disconnection value and the LAN attack measurement value are numerically compared with a preset LAN disconnection threshold and a preset LAN attack measurement threshold, respectively. If the LAN disconnection value or the LAN attack measurement value exceeds the corresponding preset threshold, indicating that the operation safety hazard during the detection period is high, a hazard warning signal is generated; If the LAN disconnection value and the LAN attack detection value do not exceed the corresponding preset thresholds, the total duration of LAN disconnection and the total duration of LAN being attacked by external forces during the detection period are collected and marked as the LAN opening time value and the LAN attack time value respectively; The real-time traffic value of the local area network is collected, and the real-time traffic value is compared with the preset real-time traffic value range. If the real-time traffic value is not within the preset real-time traffic value range, it is determined that the corresponding moment is in a traffic abnormality state, and the total duration of the local area network in the traffic abnormality state during the detection period is obtained and marked as a traffic characteristic value; The hidden danger characteristic coefficient GW is obtained by numerically calculating the LAN opening time value QY, the LAN attack time value NS and the flow characteristic value LX through the formula GW=(wq×QY+hy×NS+ty×LX) / 3; wherein wq, hy and ty are preset weight coefficients with values ​​greater than zero, and the larger the value of the hidden danger characteristic coefficient GW is, the higher the overall operation safety hidden danger in the detection period is; The hidden danger characteristic coefficient GW is numerically compared with the preset hidden danger characteristic coefficient threshold. If the hidden danger characteristic coefficient GW exceeds the preset hidden danger characteristic coefficient threshold, it indicates that the operation safety hidden dangers during the detection period are generally high, and a hidden danger warning signal is generated.

[0023] Embodiment 2: Figure 2As shown, the difference between this embodiment and the first embodiment is that the hidden danger analysis unit is communicatively connected to the operation efficiency analysis unit. If the hidden danger analysis unit does not generate a hidden danger warning signal, the operation effect of the risk assessment model during the detection period is analyzed by the operation efficiency analysis unit to determine whether to generate an operation efficiency warning signal through analysis, and the operation efficiency warning signal is sent to the immediate alarm unit. When the immediate alarm unit receives the operation efficiency warning signal, it sends an alarm message to the corresponding administrator to remind the administrator to conduct a cause investigation and analysis and take corresponding optimization measures in time to ensure the efficiency of subsequent model analysis and the efficient and smooth progress of medical work. The specific analysis process of the operation efficiency analysis unit is as follows: The moment when the risk assessment model receives the input patient information is obtained and marked as the information arrival moment, and the moment when the clinical decision support output unit outputs the critical risk assessment report is obtained and marked as the report generation moment, and the time difference between the report generation moment and the information arrival moment is calculated to obtain the assessment duration; the assessment duration is numerically compared with the preset assessment duration threshold, and if the assessment duration exceeds the preset assessment duration threshold, the corresponding assessment duration is marked as the characteristic duration; The number of characteristic durations in the detection period is obtained and the ratio of the ratio to the number of evaluation durations is calculated to obtain the characteristic duration table value, and the characteristic duration table value is numerically compared with the preset characteristic duration table threshold. If the characteristic duration table value exceeds the preset characteristic duration table threshold, it indicates that the model operation efficiency in the detection period is poor, and an operation efficiency warning signal is generated; If the characteristic time table value does not exceed the preset characteristic time table threshold, the evaluation time value is obtained by averaging all evaluation durations in the detection period, and the characteristic time table value TF and the evaluation time value YP are weighted and summed to obtain the operation efficiency correlation value HX by the formula HX=qt×TF+sw×YP, where qt and sw are preset weight coefficients with values ​​greater than zero; Moreover, the larger the value of the operation efficiency correlation value HX is, the worse the overall performance of the model operation efficiency during the detection period is; the operation efficiency correlation value HX is numerically compared with the preset operation efficiency correlation threshold. If the operation efficiency correlation value HX exceeds the preset operation efficiency correlation threshold, it indicates that the overall performance of the model operation efficiency during the detection period is poor, and an operation efficiency warning signal is generated.

[0024] Embodiment 3: Figure 2As shown, the difference between this embodiment and the first and second embodiments is that the operation efficiency analysis unit is connected to the pressure analysis unit. If no operation efficiency warning signal is generated during the detection period, the operation pressure of the risk assessment model during the detection period is analyzed by the pressure analysis unit, and the pressure warning signal is determined by analysis to determine whether a pressure warning signal is generated, and the pressure warning signal is sent to the immediate alarm unit. When the immediate alarm unit receives the pressure warning signal, it sends an alarm message to the corresponding administrator to remind the administrator to improve the analysis ability of the model in time, so as to further ensure the efficient and smooth progress of subsequent medical work; the specific analysis process of the pressure analysis unit is as follows: The number of real-time tasks waiting for evaluation by the risk assessment model is obtained and marked as a queuing coefficient, and the queuing coefficient is numerically compared with a preset queuing coefficient threshold. If the queuing coefficient exceeds the preset queuing coefficient threshold, it is determined to be in a high-pressure operation state; Several detection time periods are set within the detection period, the total duration of the operation in the high-pressure state in the corresponding detection time period is collected and marked as a high-pressure characteristic value, the number of occurrences of the single duration of the operation in the high-pressure state in the corresponding detection time period exceeding the preset single duration threshold is collected and marked as a high-pressure risk value, and the maximum value of the single duration of the operation in the high-pressure state in the corresponding detection time period is marked as a high-pressure amplitude value; The high-pressure characteristic value WF, the high-pressure risk value GP and the high-pressure amplitude value LF are numerically calculated by the formula XM=rg×WF+up×GP+ew×LF to obtain the high-pressure decision value XM; wherein rg, up, and ew are preset weight coefficients with values ​​greater than zero, and up>ew>rg; and the larger the value of the high-pressure decision value XM is, the greater the overall model operation pressure in the corresponding detection period is; The high-pressure decision value XM is numerically compared with the preset high-pressure decision threshold. If the high-pressure decision value XM exceeds the preset high-pressure decision threshold, it indicates that the model operation pressure of the corresponding detection period is generally large, and the corresponding detection period is marked as a high-pressure period; if the high-pressure decision value XM does not exceed the preset high-pressure decision threshold, it indicates that the model operation pressure of the corresponding detection period is generally small, and the corresponding detection period is marked as a low-pressure period; The number of high-pressure time periods in the detection period is obtained and marked as high-pressure detection values, and the high-pressure detection values ​​are numerically compared with the preset high-pressure detection threshold. If the high-pressure detection value exceeds the preset high-pressure detection threshold, it indicates that the model analysis capability needs to be increased in time to reduce the subsequent operating pressure and ensure the efficient, smooth and stable progress of medical work, and a pressure warning signal is generated.

[0025] If the high-pressure detection value does not exceed the preset high-pressure detection threshold, the number of high-pressure periods between two adjacent low-pressure periods is marked as the high-pressure duration value, and the high-pressure duration value is numerically compared with the preset high-pressure duration threshold. If the high-pressure duration value exceeds the preset high-pressure duration threshold, the corresponding high-pressure duration value is marked as the high-pressure maintenance inspection value, and the number of high-pressure maintenance inspection values ​​in the detection period is obtained and marked as the high-pressure maintenance risk value; The high-voltage decision values ​​of all detection periods are averaged to obtain the high-voltage performance value, and the high-voltage detection value FX, the high-voltage holding value GN and the high-voltage performance value PF are numerically calculated by the formula TW=kp×FX+ng×GN+sq×PF to obtain the high-voltage analysis coefficient TW; wherein kp, ng, sq are preset weight coefficients with values ​​greater than zero, ng>kp>sq; Moreover, the larger the value of the high-pressure analysis coefficient TW is, the more necessary it is to increase the model analysis capability in a timely manner; the high-pressure analysis coefficient TW is numerically compared with the preset high-pressure analysis coefficient threshold. If the high-pressure analysis coefficient TW exceeds the preset high-pressure analysis coefficient threshold, it indicates that the model analysis capability needs to be increased in a timely manner to reduce the subsequent model operation pressure and ensure the efficient, smooth and stable progress of medical work, and a pressure warning signal is generated.

[0026] The working principle of the present invention is as follows: when in use, the multi-dimensional data of critically ill patients are collected from multiple medical information systems through the critical disease data integration processing unit and preprocessed; the critical disease feature selection and extraction unit uses feature selection algorithm and feature extraction technology to extract key features that have an important impact on critical disease risk assessment from the preprocessed data; the risk assessment model construction optimization unit constructs and stores a risk assessment model based on the extracted features; the risk assessment model analyzes the patient information entered by the medical staff; the clinical decision support output unit clinically interprets the output results of the risk assessment model, and sends the critical disease risk assessment report of the corresponding patient to the corresponding medical terminal for display, thereby achieving a comprehensive and accurate assessment of the risks of critically ill patients, which is conducive to helping doctors formulate more scientific and reasonable treatment plans, and improving the diagnostic efficiency and accuracy of medical staff; and the operation monitoring management end performs operation monitoring and abnormal judgment warning on the model construction operation end, which is conducive to ensuring efficient, smooth and stable medical work.

[0027] The above formulas are all dimensionless and numerical calculations. The formula is a formula obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formula are set by technicians in this field according to actual conditions. The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the contents of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that technicians in the relevant technical field can understand and use the present invention well. The present invention is only limited by the claims and their full scope and equivalents.

Claims

1. A critical illness risk assessment model construction system, characterized in that: It includes a model building and operation end and an operation monitoring and management end. The operation monitoring and management end monitors and manages the operation of the model building and operation end, and the model building and operation end includes a critical disease data integration processing unit, a critical disease feature selection and extraction unit, a risk assessment model construction and optimization unit, a patient information entry and transmission unit, and a clinical decision support output unit; the critical disease data integration processing unit collects multi-dimensional data of critically ill patients from multiple medical information systems, pre-processes the data using data cleaning, deduplication and normalization technologies, and sends the pre-processed data to the critical disease feature selection and extraction unit; The critical illness feature selection and extraction unit uses feature selection algorithms and feature extraction techniques to extract key features that have an important impact on critical illness risk assessment from the preprocessed data, and sends the feature selection and extraction results to the risk assessment model construction and optimization unit; The risk assessment model construction optimization unit constructs a risk assessment model based on the extracted features and stores the constructed risk assessment model, as well as trains the model and optimizes its parameters, and supports online updating and iterative optimization of the risk assessment model; The patient information input and transmission unit is used for medical personnel to input patient information, and the risk assessment model construction and optimization unit analyzes the patient information based on the constructed risk assessment model and sends the analysis results to the clinical decision support output unit; The clinical decision support output unit clinically interprets the output results of the risk assessment model, generates an intuitive critical illness risk assessment report, and sends the critical illness risk assessment report of the corresponding patient to the corresponding medical terminal for display.

2. A critical illness risk assessment model construction system according to claim 1, characterized in that: The medical information system includes the electronic medical record system, laboratory information system and medical imaging system. The multi-dimensional data collected include medical history records, laboratory test results, imaging examination reports and physiological monitoring data.

3. A critical illness risk assessment model construction system according to claim 1, characterized in that: The critical illness risk assessment report includes the patient's risk level information, key risk characteristic information, potential complication prediction information and personalized treatment recommendation information.

4. A critical illness risk assessment model construction system according to claim 1, characterized in that: The operation monitoring and management end includes a security protection unit, a hidden danger analysis unit and an immediate alarm unit. The security protection unit is used to set up a secure local area network. The model building operation end uses the set secure local area network for operation. The security protection unit is also used to register medical terminals and medical personnel, mark the successfully registered medical terminals and medical personnel as safe terminals and safe personnel, and store the information of the safe terminals and safe personnel, and only allow safe personnel to log in through the safe terminals; The hidden danger analysis unit analyzes the operational hidden dangers of the model construction operation end, and determines whether to generate a hidden danger warning signal through analysis. When the hidden danger warning signal is generated, it will be sent to the immediate alarm unit.

5. A critical illness risk assessment model construction system according to claim 4, characterized in that: The specific analysis process of the hidden danger analysis unit includes: The frequency of disconnection of the secure LAN and the frequency of external attacks during the detection period are collected and marked as LAN disconnection value and LAN attack detection value. If the LAN disconnection value or the LAN attack detection value exceeds the corresponding preset threshold, a hidden danger warning signal is generated; If the LAN disconnection value and the LAN attack detection value do not exceed the corresponding preset threshold value, the real-time traffic value of the LAN is collected. If the real-time traffic value is not within the preset real-time traffic value range, it is judged that the corresponding time is in an abnormal traffic state; the total duration of the LAN in the abnormal traffic state during the detection period is obtained and marked as the traffic characteristic value, and the hidden danger characteristic coefficient is obtained by numerically calculating the LAN opening time value, the LAN attack time value and the traffic characteristic value. If the hidden danger characteristic coefficient exceeds the preset hidden danger characteristic coefficient threshold, a hidden danger warning signal is generated.

6. A critical illness risk assessment model construction system according to claim 5, characterized in that: The hidden danger analysis unit is communicatively connected to the operation efficiency analysis unit. If the hidden danger analysis unit does not generate a hidden danger warning signal, the operation effect of the risk assessment model during the detection period is analyzed by the operation efficiency analysis unit to determine whether to generate an operation efficiency warning signal, and the operation efficiency warning signal is sent to the immediate alarm unit.

7. A critical illness risk assessment model construction system according to claim 6, characterized in that: The specific analysis process of the operation efficiency analysis unit is as follows: The time when the risk assessment model receives the input patient information is obtained and marked as the information arrival time, and the time when the clinical decision support output unit outputs the critical risk assessment report is obtained and marked as the report generation time, and the time difference between the report generation time and the information arrival time is calculated to obtain the assessment duration; if the assessment duration exceeds the preset assessment duration threshold, the corresponding assessment duration is marked as the characteristic duration; The number of characteristic durations in the detection period is obtained and the ratio thereof is calculated with the number of evaluation durations to obtain the characteristic duration value. If the characteristic duration value exceeds the preset characteristic duration threshold, an operation efficiency warning signal is generated; If the characteristic time table value does not exceed the preset characteristic time table threshold, the average of all evaluation durations in the detection period is calculated to obtain the evaluation time measurement value, and the operation efficiency correlation value is obtained by weighted summing the characteristic time table value and the evaluation time measurement value. If the operation efficiency correlation value exceeds the preset operation efficiency correlation threshold, an operation efficiency warning signal is generated.

8. A critical illness risk assessment model construction system according to claim 7, characterized in that: The operation efficiency analysis unit is connected to the pressure analysis unit. If no operation efficiency warning signal is generated during the detection period, the pressure analysis unit will analyze the operating pressure of the risk assessment model during the detection period to determine whether a pressure warning signal is generated through the analysis, and send the pressure warning signal to the immediate alarm unit.

9. A critical illness risk assessment model construction system according to claim 8, characterized in that: The specific analysis process of the pressure analysis unit is as follows: The number of real-time tasks waiting for evaluation by the risk assessment model is obtained and marked as a queuing coefficient. If the queuing coefficient exceeds the preset queuing coefficient threshold, it is judged to be in a high-pressure operating state. Several detection periods are set within the detection period, and a high-pressure decision value is obtained by numerically calculating the high-pressure characteristic value, the high-pressure risk value and the high-pressure amplitude value. If the high-pressure decision value exceeds the preset high-pressure decision threshold, the corresponding detection period is marked as a high-pressure period. Otherwise, the corresponding detection period is marked as a low-voltage period; The number of high-pressure time periods in the detection period is obtained and marked as a high-pressure detection value. If the high-pressure detection value exceeds a preset high-pressure detection threshold, a pressure warning signal is generated.

10. A critical illness risk assessment model construction system according to claim 9, characterized in that: If the high-pressure detection value does not exceed the preset high-pressure detection threshold, the high-pressure analysis coefficient is obtained by numerically calculating the high-pressure detection value, the high-pressure risk value and the high-pressure performance value. If the high-pressure analysis coefficient exceeds the preset high-pressure analysis coefficient threshold, a pressure warning signal is generated.

Images