An information digital management system based on the critical care medicine department

Abstract

The present invention relates to the field of data management technology, and in particular to an information digital management system based on the Department of Critical Care Medicine. The system is used to perform the following steps: inputting patient information data of the Department of Critical Care Medicine using a terminal device; performing missing field value filling processing on the patient information data of the Department of Critical Care Medicine to obtain complete critically ill patient data; performing physiological indicator time risk analysis on the complete critically ill patient data to obtain physiological indicator abnormal risk data; performing risk intelligent warning based on physiological indicator abnormal risk data to generate intelligent disease risk warning data; performing risk intervention timing reasoning on the intelligent disease risk warning data to obtain risk intervention timing decision data; performing medical resource optimization and adjustment based on the risk intervention timing decision data to obtain an individualized diagnosis and treatment auxiliary decision strategy. The present invention realizes individualized intervention timing decision-making through intelligent disease risk warning in the system, which significantly improves the management efficiency and resource utilization of the Department of Critical Care Medicine.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and in particular to an information digital management system based on a critical care medicine department. Background Art

[0002] In today's medical and health field, the intensive care unit (ICU), as an important part of the hospital, bears the heavy responsibility of rescuing critically ill patients and maintaining life safety. Its information construction level is directly related to the success rate of patient treatment and the rational use of medical resources. Patients in the ICU are usually in critical condition, with frequent fluctuations in physiological indicators, and need to receive complex monitoring and treatment. This puts higher requirements on the rational allocation of medical resources, the optimization of medical workflows, and the efficiency of information management systems. An effective information digital management system is of great significance to improving the work efficiency of the ICU and the quality of patient care. However, the traditional management model of the ICU mainly relies on the experience and manual records of medical staff, and requires continuous monitoring and intervention by medical staff, which brings huge work pressure to medical staff, and is prone to human misjudgment and delays in treatment, which in turn causes unreasonable allocation of medical resources in the ICU. Summary of the invention

[0003] Based on this, the present invention provides an information digital management system based on the critical care medicine department to solve at least one of the above technical problems.

[0004] To achieve the above purpose, an information digital management system based on the critical care medicine department includes:

[0005] Control circuit;

[0006] A patient data acquisition module is electrically connected to the control circuit, and the patient data acquisition module includes: a mobile computing terminal, a multifunctional vital signs monitor and an encrypted storage component. The patient data acquisition module is used to input the patient information data of the critical care medicine department using the mobile computing terminal; perform differential privacy protection processing on the patient information data of the critical care medicine department to generate patient privacy protection data; perform distributed data storage according to the patient privacy protection data to obtain preliminary patient information storage data;

[0007] A physiological risk assessment module is electrically connected to the control circuit, and includes an intelligent data analysis component. The physiological risk assessment module is used to predict missing field values ​​for preliminary patient information storage data, and perform data filling processing to obtain complete critically ill patient data; perform a quantitative assessment of the patient's risk level on the complete critically ill patient data to generate patient's disease severity grading data; perform a physiological indicator time risk analysis based on the patient's disease severity grading data to obtain physiological indicator abnormal risk data;

[0008] A disease risk warning module is electrically connected to the control circuit, and includes a case retrieval database and an intelligent warning terminal, which is used to process case retrieval conditions according to abnormal risk data of physiological indicators, and to retrieve similar cases to obtain similar case data; to construct a patient group interaction relationship network according to similar case data; to perform risk intelligent warning according to the patient group interaction relationship network, and to generate intelligent disease risk warning data;

[0009] A risk intervention timing decision module is electrically connected to the control circuit, and the risk intervention timing decision module includes: an integrated interface of a Bayesian network analysis tool and a clinical information system feedback workstation, for constructing a Bayesian network model; using the Bayesian network model to perform risk intervention timing reasoning on the intelligent disease risk warning data to obtain risk intervention timing decision data; performing warning information feedback statistics on the risk intervention timing decision data and the intelligent disease risk warning data to generate warning information feedback data;

[0010] An individualized diagnosis and treatment decision-making assistance module is electrically connected to the control circuit. The individualized diagnosis and treatment decision-making assistance module includes: a scheduling instruction generator, a doctor's order execution terminal and an integrated interface of a medical resource management system, which is used to analyze the disease risk treatment plan based on the early warning information feedback data, and optimize and adjust the medical resources to obtain an optimized medical resource allocation plan; process the medical resource scheduling instructions according to the optimized medical resource allocation plan to obtain an individualized diagnosis and treatment decision-making assistance strategy; integrate the individualized diagnosis and treatment decision-making assistance strategy into the intensive care system to implement digital information management of the critical care department.

[0011] The present invention inputs the patient information data of the Department of Critical Care Medicine through the terminal device and performs differential privacy protection processing, which can effectively ensure the security and privacy of the patient's personal information and prevent the leakage of sensitive information. At the same time, the distributed data storage method is used to save the patient information after privacy protection processing, which not only improves the security of data storage, but also enhances the data access efficiency and system stability. By predicting and filling the missing field values ​​in the preliminary patient information storage data, the data integrity can be improved, and the deviation or wrong conclusion caused by incomplete data can be reduced, so that the decision made based on these data is more reliable. The complete critical patient data is quantitatively evaluated for the degree of danger, and the graded data is generated, which helps medical staff to quickly identify high-risk individuals who need priority attention and take targeted measures. And according to the patient's condition severity graded data, the physiological indicator time risk analysis can more accurately capture the key nodes of disease progression, discover potential life-threatening factors in advance, and win precious time for timely intervention. By formulating case retrieval conditions based on abnormal risk data of physiological indicators and performing similar case searches, we can find historical treatment cases under similar circumstances. These similar case data provide a reference for the diagnosis and treatment of current patients, improve the accuracy and effectiveness of diagnosis and treatment, and are of great significance for controlling nosocomial infections and optimizing ward layout. Using this network for intelligent risk warning can achieve the transition from grasping the overall trend at the macro level to accurately locating individual problem points at the micro level, greatly improving the sensitivity and accuracy of the early warning system. The constructed Bayesian network model provides scientific decision-making support for the analysis of intelligent disease risk warning data and enhances the cognition and understanding of potential risks. Through the reasoning of risk intervention timing, the precise timing of clinical intervention can be grasped, avoiding waste of resources and unnecessary intervention. According to the early warning information feedback data, the analysis of disease risk disposal plans can achieve accurate analysis of individual patient needs and improve the pertinence and effectiveness of medical intervention. The optimization and adjustment of medical resources ensures the rational allocation of resources in the critical care department and reduces resource waste. The formulation of individualized diagnosis and treatment decision-making assistance strategies and the integration of the intensive care system provide clinicians with intuitive and practical decision-making tools, support the precise treatment and monitoring of critically ill patients, and ultimately improve the level of information digital management in the critical care department and the treatment effect of patients. Therefore, the information digital management system based on the critical care department of the present invention inputs patient information through terminal equipment, improves the accuracy and completeness of data through missing field value prediction, uses quantitative assessment of the severity of the disease and real-time deviation analysis of physiological indicators to timely identify high-risk patients, warn in advance and take intervention measures, and recommends intervention timing based on the Bayesian network model, optimizes the medical resource allocation plan, and significantly improves the management efficiency and resource utilization of the critical care department, realizes scientific auxiliary decision-making of individualized intervention timing, and optimizes the allocation of medical resources.

[0012] Preferably, the patient data acquisition module is used to perform the following steps:

[0013] Step S11: inputting patient information data of the Department of Critical Care Medicine using a terminal device;

[0014] Step S12: performing data standardization processing on the patient information data of the Department of Critical Care Medicine to generate standard patient information data;

[0015] Step S13: desensitizing the standard patient information data and performing differential privacy protection processing to generate patient privacy protection data;

[0016] Step S14: data association is performed on the standard patient information data through the patient privacy protection data to obtain patient information index association data;

[0017] Step S15: Based on the patient information index association data, the standard patient information data is distributedly stored using a distributed storage system, thereby obtaining preliminary patient information storage data.

[0018] The present invention uses a terminal device to input patient information data of the Department of Critical Care Medicine, ensuring the real-time and accuracy of data collection. Through a centralized data entry method, it is ensured that patient information can enter the system in a timely and complete manner. Data standardization processing of patient information data of the Department of Critical Care Medicine helps to eliminate format differences and inconsistencies between data from different sources. Standardization processing makes the data more unified and standardized, facilitates cross-platform and cross-system data exchange and sharing, and improves the comparability and operability of the data. Through desensitization, the patient's sensitive information is protected, the risk of information leakage is reduced, and the application of differential privacy technology ensures that valuable statistical information can still be obtained without compromising the patient's privacy during data analysis. Data association is performed on standard patient information data through patient privacy protection data, which improves the traceability and utilization efficiency of the data. Effective index association can quickly locate and retrieve relevant patient information, reduce the time cost of information retrieval, and improve the efficiency of clinical decision-making. Standard patient information data is distributedly stored through a distributed storage system, ensuring high availability and reliability of the data. Distributed storage not only improves the access speed of data, but also enhances the fault tolerance of the system, avoids the risk of single point failure, and provides a guarantee for the long-term storage and management of data.

[0019] Preferably, the physiological risk assessment module is used to perform the following steps:

[0020] Step S21: Perform integrity check on the preliminary patient information storage data to obtain data integrity assessment data;

[0021] Step S22: Identify missing data fields based on the data integrity assessment data, and predict missing field values ​​to obtain missing field recommended value data;

[0022] Step S23: Feedback the missing field recommended value data to the terminal device to remind the medical staff to make additional inputs to obtain the patient information supplementary data;

[0023] Step S24: performing data filling processing on the preliminary patient information storage data based on the patient information supplementary data to obtain complete critically ill patient data;

[0024] Step S25: Quantitatively assess the patient's risk level on the complete critically ill patient data to generate patient's disease severity grading data; perform real-time physiological deviation analysis based on the patient's disease severity grading data to generate physiological parameter real-time deviation data;

[0025] Step S26: assigning time sensitivity weights to the real-time deviation data of physiological parameters based on the preset critical illness time sensitivity rules, and performing physiological indicator risk processing to obtain abnormal physiological indicator risk data.

[0026] The present invention can timely find missing or incomplete parts in the data by performing an integrity check on the preliminary patient information storage data. Through automated identification and recommendation of missing fields, the need for manual intervention is reduced, and the predicted recommended values ​​can be used as a reference to help medical staff to supplement missing information more efficiently, thereby improving the efficiency and accuracy of data supplementation. Through a timely feedback mechanism, it is ensured that medical staff can quickly obtain the need for supplementary information, enhance the dynamic management of patient information, help improve the personalized care level of critically ill patients, and ensure that patients obtain more accurate medical services. Based on the patient information supplementary data, the preliminary patient information storage data is filled with data to ensure the comprehensiveness and consistency of the data. The complete critically ill patient data is quantitatively evaluated for the patient's risk level, and the quantitative evaluation of the risk level provides a scientific basis for disease classification, helping medical staff to understand the severity of the patient's condition in a timely manner and guiding clinical decision-making. Real-time physiological deviation analysis can dynamically monitor the changes in the patient's physiological parameters, detect abnormal conditions in a timely manner, and provide important information for early intervention. Through the weight assignment of time sensitivity, the priority processing of important physiological indicators in real-time monitoring is ensured, and the accuracy of risk assessment is improved.

[0027] Preferably, step S22 includes the following steps:

[0028] Step S221: Identify missing data fields based on data integrity assessment data and generate missing field data;

[0029] Step S222: Identify the missing field type of the missing field data to obtain the missing field type data;

[0030] Step S223: query the historical filling records based on the missing field type data, and perform similarity matching of the complete historical records to generate complete historical filling record data;

[0031] Step S224: extracting context-related fields from missing field type data by filling in the complete historical record data, and generating context-related field data;

[0032] Step S225: Based on a preset conventional data filling strategy, preliminary data supplementation processing is performed on the missing field data through the context-related field data to generate preliminary supplementary suggestion data;

[0033] Step S226: Perform logical verification on the preliminary supplementary suggestion data, conduct validity evaluation, and generate missing field recommendation value data.

[0034] The present invention can systematically find incomplete parts in the data by identifying missing data fields for data integrity assessment data, and provides a clear goal for subsequent data supplementation and processing. By identifying the specific type of missing fields (such as numerical type, text type, etc.), a suitable filling method can be selected more specifically. This classification processing improves the refinement of data processing, ensures the effectiveness and rationality of the filling strategy, and improves the overall quality of the data. By querying historical records and performing similarity matching, a reliable data source related to the current missing field can be found. Extracting context-related fields helps to capture the inherent connection between data, provide more background information for data filling, make the filling more in line with the actual situation, reduce errors in the data filling process, and improve the authenticity and accuracy of the data. Using the preset filling strategy combined with context information for preliminary supplementation, reasonable data filling suggestions can be quickly generated, which not only improves the efficiency of data filling, but also ensures the consistency and coherence of the filled data with the existing data. The preliminary supplementary suggestion data is logically verified to ensure the scientificity and reliability of the supplementary data. The implementation of logical verification can effectively identify unreasonable or unrealistic filling suggestions and avoid incorrect data input.

[0035] Preferably, step S25 comprises the following steps:

[0036] Step S251: Acquire the patient's physiological sign data and electronic medical record data based on the complete critically ill patient data;

[0037] Step S252: extracting the medical condition text diagnosis information from the electronic medical record data to obtain the medical condition diagnosis information data;

[0038] Step S253: evaluating the severity of the patient's condition by using the disease diagnosis information data to evaluate the patient's physiological sign data, and obtaining the patient's severity classification data;

[0039] Step S254: marking high-risk patients according to the patient's disease severity classification data, and reading the high-risk patient's physiological data to generate high-risk patient physiological parameters;

[0040] Step S255: dynamically regularize the physiological parameters of high-risk patients in time series to generate physiological parameters of time-series aligned patients;

[0041] Step S256: constructing a physiological parameter dynamic baseline model using a preset time series model and a preset physiological parameter baseline reference rule;

[0042] Step S257: Perform real-time physiological deviation analysis on the physiological parameters of the time-series aligned patients through a physiological parameter dynamic baseline model to generate physiological parameter real-time deviation data.

[0043] The present invention obtains the patient's physiological sign data and electronic medical record data based on the complete critical patient data, and can fully grasp the patient's health status and historical medical information. By automatically extracting the key diagnostic information in the electronic medical record, the patient's condition description and diagnosis results can be quickly obtained, reducing the time and workload of manual sorting, and improving the efficiency and accuracy of data processing. A comprehensive evaluation is performed in combination with the diagnostic information and physiological sign data, and the severity of the patient's condition can be more accurately judged. High-risk patients are marked according to the patient's critical degree classification data, ensuring that high-risk patients are focused on. The generated high-risk patient physiological parameters provide a key basis for monitoring and intervention for medical staff, which can improve the management ability of critically ill patients and reduce potential risks. Through dynamic regularization of time series, the consistency of physiological data collected at different time points on the time axis is ensured. By constructing a dynamic baseline model, a standardized reference framework can be established for comparing and analyzing the patient's physiological parameters. This model helps to identify whether the patient's physiological parameters deviate from the normal range. Real-time deviation analysis can promptly detect the deviation between the patient's physiological parameters and the dynamic baseline, and provide early warning information. This analysis helps medical staff respond quickly to abnormal situations and take timely intervention measures to reduce risks to patients and improve the safety and effectiveness of treatment.

[0044] Preferably, the disease risk warning module is used to perform the following steps:

[0045] Step S31: Process case retrieval conditions according to the patient's disease severity classification data and physiological index abnormal risk data to obtain case retrieval condition data;

[0046] Step S32: Acquire a historical case database; use the case search condition data to perform a similar case search on the historical case database to obtain similar case data;

[0047] Step S33: constructing a patient group interaction relationship network based on similar case data; identifying key nodes of the patient group interaction relationship network to obtain key node case patient data; performing case development path analysis based on the key node case patient data to generate case development path data;

[0048] Step S34: performing multi-dimensional index correlation analysis of physiological indexes, test indexes, treatment plans and nursing measures on key node case patient data to obtain key node correlation index data;

[0049] Step S35: Perform intelligent risk warning on the case development path data through key node association indicator data to generate intelligent disease risk warning data, wherein the intelligent disease risk warning data includes warning level data, warning reason data and recommended measures data.

[0050] The present invention processes case retrieval conditions according to the patient's critical condition classification data and physiological index abnormal risk data, which can ensure the accuracy and pertinence of the retrieval conditions. Obtain a historical case database and use the case retrieval condition data to retrieve similar cases to ensure that historical cases that are highly similar to the current patient's condition can be identified. Constructing a patient group interaction relationship network based on similar case data can effectively demonstrate the correlation and mutual influence between different patients. The implementation of case development path analysis helps to identify the law of patient condition evolution, support the medical team to accurately predict the trend of disease changes, and optimize clinical management and intervention strategies. Multi-dimensional indicator correlation analysis of physiological indicators, test indicators, treatment plans and nursing measures for key node case patient data can fully understand the impact of different factors on the patient's condition. Through the key node correlation indicator data, the case development path data is intelligently warned of risks, which can promote rapid intervention in high-risk patients, reduce the probability of adverse events, and improve the overall safety and treatment effect of critical care medicine.

[0051] Preferably, step S33 includes the following steps:

[0052] Step S331: constructing a patient group feature vector for similar case data to obtain patient group feature vector data;

[0053] Step S332: using the fuzzy C-means clustering algorithm to perform patient group characteristic clustering analysis on the patient group characteristic vector data to obtain patient group characteristic clustering data;

[0054] Step S333: extracting relevant physiological indicators according to the patient group characteristic clustering data to obtain case-related physiological indicator data;

[0055] Step S334: analyzing the case development stage according to the case-related physiological indicator data to obtain the case development stage data;

[0056] Step S335: Process the key physiological indicator mutation nodes on the case development stage data, and perform case development path trend analysis to obtain case development path data.

[0057] The present invention constructs a patient group feature vector for similar case data, enhances the quantitative analysis capability of the patient group, and the construction of the feature vector can numerically represent information such as the patient's clinical characteristics, pathological state and treatment response. The fuzzy C-means clustering algorithm is used to perform cluster analysis on the patient group feature vector data, which improves the heterogeneity analysis capability within the group. The cluster analysis can classify patients with similar characteristics into the same category, revealing the potential characteristics and similarities of different patient groups. The extraction of relevant physiological indicators helps to more accurately monitor and evaluate the health status of patients. According to the case-related physiological indicator data, the case development stage analysis can be performed, and the patient's disease development can be divided into different stages, which helps medical staff to better understand the dynamic change process of the disease, adjust the treatment plan in time, and improve the continuity and systematicness of the treatment. The key physiological indicator mutation node processing of the case development stage data ensures sensitive monitoring of the patient's disease changes. The processing of the key indicator mutation node can timely identify the moment when the disease changes sharply, providing an important decision-making basis for the medical team.

[0058] Preferably, step S333 includes the following steps:

[0059] The nodes of the patient group interaction network are constructed according to the patient group characteristic clustering data, and the edges of the patient group interaction network are constructed according to the disease correlation and treatment scheme similarity between the patient groups, thereby constructing the patient group interaction network;

[0060] Conduct topological structure analysis on the patient group interaction network, identify key nodes in the interaction network, and obtain key node case patient data;

[0061] According to the patient data of key node cases, the severe type of the case is identified to obtain the severe type data of the case;

[0062] Relevant physiological indicators are extracted from the severe type data of the case through the abnormal risk data of physiological indicators to obtain the case-related physiological indicator data.

[0063] The present invention constructs a patient group interaction network based on patient group characteristic clustering data, which can reveal the mutual relationships and influences between different patient groups. The topological structure analysis of the patient group interaction network can reveal the important nodes and connection paths in the network, and identify the key nodes that have a significant impact on the overall network structure. These key nodes are usually important hubs for disease transmission or treatment effects, and focusing on monitoring and managing them helps to improve the utilization efficiency and treatment effects of medical resources. By identifying the types of severe illnesses of key node patients, the characteristics and commonalities of different types of severe illnesses can be more accurately understood. This classification helps medical staff to adopt more targeted treatment strategies for different types of severe illnesses. By extracting relevant physiological indicators from the case severe illness type data through the abnormal risk data of physiological indicators, indicators closely related to specific severe illness types can be screened out from a large number of physiological indicators, reducing data redundancy and improving the efficiency of data analysis.

[0064] Preferably, the risk intervention timing decision module is used to perform the following steps:

[0065] Step S41: constructing a Bayesian network model based on preset critical care medical knowledge data, wherein the Bayesian network model includes risk factor nodes, physiological indicator nodes, and risk event nodes;

[0066] Step S42: using the abnormal risk data of physiological indicators and the intelligent disease risk warning data to assign time node states to the Bayesian network model, and using the Bayesian network inference algorithm to calculate the event probability of the risk event node to obtain risk probability deduction data;

[0067] Step S43: constructing a fuzzy reasoning decision model through a Bayesian network model based on the risk probability deduction data and the preset risk intervention threshold, wherein the fuzzy reasoning decision model includes a risk probability fuzzy variable, a risk development speed fuzzy variable, and a patient individual characteristic fuzzy variable;

[0068] Step S44: performing risk intervention timing reasoning according to the fuzzy reasoning decision model to obtain risk intervention timing decision data;

[0069] Step S45: supplementing the recommended measures data in the intelligent disease risk warning data with the risk intervention opportunity decision data to generate supplementary disease risk warning data;

[0070] Step S46: Utilize the preset medical staff information receiving authority matrix to perform multi-level linkage push processing on the supplementary disease risk warning data, and perform warning information feedback statistics to generate warning information feedback data.

[0071] The present invention constructs a Bayesian network model based on preset critical care medical knowledge data to form risk factor nodes, physiological indicator nodes and risk event nodes, and optimizes the modeling ability of complex physiological and pathological relationships. The construction of the Bayesian network model provides a systematic framework, which makes the relationship between different risk factors and physiological indicators clearly visible, thereby supporting a comprehensive assessment of the patient's condition. The Bayesian network model is assigned a time node state using physiological indicator abnormal risk data and intelligent disease risk warning data, and the event probability of the risk event node is calculated in combination with the Bayesian network reasoning algorithm, which can reflect the dynamic characteristics of the patient's condition change in real time. Risk probability deduction provides a quantitative risk assessment basis for medical personnel, making the identification of potential risks more accurate. Based on the risk probability deduction data and the preset risk intervention threshold, a fuzzy reasoning decision model is constructed through the Bayesian network model, which enhances the comprehensive assessment ability of risks. The reasoning of the timing of risk intervention provides a clear action guide for the medical team, which can respond quickly at critical moments and reduce the risk of worsening of the patient's condition. Through the multi-level linkage push mechanism, it is ensured that the warning information can be delivered to the corresponding medical staff in a timely manner according to the authority and division of responsibilities, thereby improving the efficiency and accuracy of information transmission.

[0072] Preferably, the individualized diagnosis and treatment decision-making assistance module is used to perform the following steps:

[0073] Step S51: Analyze the disease risk treatment plan according to the early warning information feedback data to generate disease risk treatment plan data;

[0074] Step S52: performing a medical resource demand assessment on the disease risk treatment plan data, and generating a medical resource allocation plan to obtain a medical resource allocation plan;

[0075] Step S53: Calculate the patient waiting time according to the medical resource allocation plan to obtain patient waiting time data;

[0076] Step S54: Perform resource allocation feasibility analysis on the medical resource allocation plan based on the patient waiting time data, and make optimization adjustments to obtain an optimized medical resource allocation plan;

[0077] Step S55: Process the medical resource dispatching instructions according to the optimized medical resource allocation plan, and send the medical resource dispatching instructions to relevant medical staff and medical equipment, so as to obtain an individualized diagnosis and treatment auxiliary decision-making strategy.

[0078] The present invention analyzes the disease risk treatment plan according to the early warning information feedback data, can systematically evaluate the patient's disease risk, and formulate a corresponding treatment plan. By comprehensively analyzing the early warning information, targeted treatment suggestions and intervention measures can be provided to improve the scientificity and effectiveness of medical decision-making. By evaluating the medical resources (such as manpower, equipment, drugs, etc.) required for different treatment plans, the use of resources can be reasonably planned. The generated medical resource allocation plan helps to optimize resource allocation and ensure that the needs of patients are maximized under limited resources. By calculating the waiting time of patients, the service efficiency under different allocation plans can be predicted. This calculation method helps to evaluate the practical feasibility of different plans. Based on the feasibility analysis of waiting time data, bottlenecks and deficiencies in existing plans can be found, so as to make targeted optimization adjustments. The optimized allocation plan can make more reasonable use of medical resources, reduce patient waiting time, and improve the overall efficiency and response speed of medical services. Through specific scheduling instructions, it is ensured that medical resources can be in place in time to support individualized diagnosis and treatment decisions. This scheduling mechanism not only improves the flexibility and efficiency of resource use, but also enhances the personalization and accuracy of medical services, and improves the treatment effect and satisfaction of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 It is a module block diagram of an information digital management system based on the critical care medicine department;

[0080] Figure 2 Detailed implementation flow chart of step S3 of the present invention;

[0081] Figure 3 Detailed implementation flow chart of step S4 of the present invention;

[0082] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0083] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are 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 technicians in this field without creative work are within the scope of protection of the present invention.

[0084] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0085] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0086] To achieve this, please refer to Figures 1 to 3 The present invention provides an information digital management system based on the critical care medicine department, comprising:

[0087] Control circuit;

[0088] A patient data acquisition module is electrically connected to the control circuit, and the patient data acquisition module includes: a mobile computing terminal, a multifunctional vital signs monitor and an encrypted storage component. The patient data acquisition module is used to input the patient information data of the critical care medicine department using the mobile computing terminal; perform differential privacy protection processing on the patient information data of the critical care medicine department to generate patient privacy protection data; perform distributed data storage according to the patient privacy protection data to obtain preliminary patient information storage data;

[0089] A physiological risk assessment module is electrically connected to the control circuit, and includes an intelligent data analysis component. The physiological risk assessment module is used to predict missing field values ​​for preliminary patient information storage data, and perform data filling processing to obtain complete critically ill patient data; perform a quantitative assessment of the patient's risk level on the complete critically ill patient data to generate patient's disease severity grading data; perform a physiological indicator time risk analysis based on the patient's disease severity grading data to obtain physiological indicator abnormal risk data;

[0090] A disease risk warning module is electrically connected to the control circuit, and includes a case retrieval database and an intelligent warning terminal, which is used to process case retrieval conditions according to abnormal risk data of physiological indicators, and to retrieve similar cases to obtain similar case data; to construct a patient group interaction relationship network according to similar case data; to perform risk intelligent warning according to the patient group interaction relationship network, and to generate intelligent disease risk warning data;

[0091] A risk intervention timing decision module is electrically connected to the control circuit, and the risk intervention timing decision module includes: an integrated interface of a Bayesian network analysis tool and a clinical information system feedback workstation, for constructing a Bayesian network model; using the Bayesian network model to perform risk intervention timing reasoning on the intelligent disease risk warning data to obtain risk intervention timing decision data; performing warning information feedback statistics on the risk intervention timing decision data and the intelligent disease risk warning data to generate warning information feedback data;

[0092] An individualized diagnosis and treatment decision-making assistance module is electrically connected to the control circuit. The individualized diagnosis and treatment decision-making assistance module includes: a scheduling instruction generator, a doctor's order execution terminal and an integrated interface of a medical resource management system, which is used to analyze the disease risk treatment plan based on the early warning information feedback data, and optimize and adjust the medical resources to obtain an optimized medical resource allocation plan; process the medical resource scheduling instructions according to the optimized medical resource allocation plan to obtain an individualized diagnosis and treatment decision-making assistance strategy; integrate the individualized diagnosis and treatment decision-making assistance strategy into the intensive care system to implement digital information management of the critical care department.

[0093] In the embodiment of the present invention, reference Figure 1 The above is a schematic diagram of a module block diagram of an information digital management system based on a critical care medicine department of the present invention. In this embodiment, the information digital management system based on a critical care medicine department includes:

[0094] Control circuit;

[0095] A patient data acquisition module is electrically connected to the control circuit, and the patient data acquisition module includes: a mobile computing terminal, a multifunctional vital signs monitor and an encrypted storage component. The patient data acquisition module is used to input the patient information data of the critical care medicine department using the mobile computing terminal; perform differential privacy protection processing on the patient information data of the critical care medicine department to generate patient privacy protection data; perform distributed data storage according to the patient privacy protection data to obtain preliminary patient information storage data;

[0096] In an embodiment of the present invention, a tablet computer or mobile terminal equipped with dedicated data acquisition software is used to enter patient information data in the critical care medicine department. These data include basic patient information (such as name, age, gender, admission diagnosis, etc.), vital signs data (such as heart rate, blood pressure, blood oxygen saturation, respiratory rate, etc.), laboratory test data (such as blood routine, biochemical indicators, etc.), imaging examination data (such as X-ray, CT, MRI, etc.) and treatment information (such as medication, surgical records, etc.). Differential privacy technology is used to perform privacy protection on the entered patient information data. For example, the Laplace mechanism can be used to add random noise that meets a specific privacy budget to numerical data (such as blood pressure, heart rate), or the exponential mechanism can be used to perturb discrete data (such as gender, diagnosis), thereby retaining the overall statistical characteristics of the data set without leaking individual patient information. A distributed database (such as Apache Cassandra or MongoDB) is used to store patient privacy protection data, and the data is dispersed and stored on multiple nodes to avoid single point failures and improve data security and scalability. At the same time, the database is access controlled and encrypted to further ensure data security.

[0097] A physiological risk assessment module is electrically connected to the control circuit, and includes an intelligent data analysis component. The physiological risk assessment module is used to predict missing field values ​​for preliminary patient information storage data, and perform data filling processing to obtain complete critically ill patient data; perform a quantitative assessment of the patient's risk level on the complete critically ill patient data to generate patient's disease severity grading data; perform a physiological indicator time risk analysis based on the patient's disease severity grading data to obtain physiological indicator abnormal risk data;

[0098] In an embodiment of the present invention, for the missing field values ​​existing in the preliminary patient information storage data, a machine learning method is used to predict the missing field values. For example, a prediction based on a K-nearest neighbor algorithm or a historical fill field recommendation value can be used to predict the missing field value based on other complete field values, and data filling processing can be performed to finally obtain complete critically ill patient data. A quantitative assessment model for the patient's risk level is constructed using a machine learning algorithm, such as a support vector machine (SVM) or a random forest algorithm. The input features of the model may include various physiological indicators, test results, treatment information, etc. of the patient, and the output is a score or classification of the severity of the patient's condition, for example, mild, moderate, severe and critical. According to the patient's severity classification data, a physiological indicator time risk analysis is performed. For example, a time series analysis method can be used to analyze the change trend of physiological indicators (such as heart rate, blood pressure, and blood oxygen saturation) of patients with different severity levels over time, and identify abnormal physiological indicator patterns with significant risks. For example, it can be found that the blood pressure of severe patients continues to decline within 24 hours after admission, and the indicator fluctuations deviate from the normal range, and the risk level is assessed to generate physiological indicator abnormal risk data.

[0099] A disease risk warning module is electrically connected to the control circuit, and includes a case retrieval database and an intelligent warning terminal, which is used to process case retrieval conditions according to abnormal risk data of physiological indicators, and to retrieve similar cases to obtain similar case data; to construct a patient group interaction relationship network according to similar case data; to perform risk intelligent warning according to the patient group interaction relationship network, and to generate intelligent disease risk warning data;

[0100] In an embodiment of the present invention, case retrieval conditions are processed according to the abnormal physiological indicator risk data. For example, the abnormal physiological indicator pattern can be converted into a specific retrieval condition, such as "blood pressure continues to drop by more than 20% within 24 hours after admission". Then, similar case data with similar abnormal physiological indicator patterns as the current patient are retrieved in the patient database using a similarity measurement method (such as Euclidean distance, cosine similarity). A patient group interaction relationship network is constructed based on similar case data. For example, each patient can be taken as a node. If there is a similar abnormal physiological indicator pattern between two patients, a connecting edge is established between the two nodes, and the weight of the edge can represent the magnitude of the similarity. Risk intelligent warning is performed based on the patient group interaction relationship network. For example, a graph theory algorithm can be used to analyze the topological structure of the network to identify key nodes in the network (such as nodes with high connectivity), and the patients corresponding to these nodes have higher risks. At the same time, the path and speed of risk propagation in the network can be analyzed, the trend of risk diffusion can be predicted, and intelligent disease risk warning data can be generated, such as "patient A is at risk of shock, and the risk level is high".

[0101] A risk intervention timing decision module is electrically connected to the control circuit, and the risk intervention timing decision module includes: an integrated interface of a Bayesian network analysis tool and a clinical information system feedback workstation, for constructing a Bayesian network model; using the Bayesian network model to perform risk intervention timing reasoning on the intelligent disease risk warning data to obtain risk intervention timing decision data; performing warning information feedback statistics on the risk intervention timing decision data and the intelligent disease risk warning data to generate warning information feedback data;

[0102] In an embodiment of the present invention, a Bayesian network model is constructed. The patient's various physiological indicators, laboratory examination indicators, imaging examination indicators, treatment information, and intelligent disease risk warning data are used as nodes of the Bayesian network, and the causal relationship between the nodes is defined according to medical knowledge and clinical experience, such as "hypotension can lead to shock". The Bayesian network model is used to reason about the risk intervention timing of the intelligent disease risk warning data. For example, the patient's current physiological indicators and risk warning information can be input into the Bayesian network as evidence, and the success probability of different intervention measures (such as infusion, use of pressor drugs) at different time points can be calculated through the reasoning function of the Bayesian network, and the intervention time with the highest success probability is selected to generate risk intervention timing decision data, such as "it is recommended to start infusion treatment immediately when the blood pressure of patient A drops to 80 mmHg". The risk intervention timing decision data and the intelligent disease risk warning data are fed back to the medical staff through the intensive care system, and the feedback of the medical staff on the warning information is recorded, such as whether the warning suggestions are adopted, what intervention measures are taken, and how the intervention effect is, etc., to generate warning information feedback data.

[0103] An individualized diagnosis and treatment decision-making assistance module is electrically connected to the control circuit. The individualized diagnosis and treatment decision-making assistance module includes: a scheduling instruction generator, a doctor's order execution terminal and an integrated interface of a medical resource management system, which is used to analyze the disease risk treatment plan based on the early warning information feedback data, and optimize and adjust the medical resources to obtain an optimized medical resource allocation plan; process the medical resource scheduling instructions according to the optimized medical resource allocation plan to obtain an individualized diagnosis and treatment decision-making assistance strategy; integrate the individualized diagnosis and treatment decision-making assistance strategy into the intensive care system to implement digital information management of the critical care department.

[0104] In the embodiment of the present invention, the risk treatment plan of the condition is analyzed according to the feedback data of the early warning information. For example, different risk warning types, success rates of different intervention measures, complication rates and other indicators can be counted to evaluate the effectiveness and reliability of the early warning system, and the risk treatment plan of the condition can be optimized and adjusted according to the analysis results. For example, if it is found that the complication rate of a certain intervention measure is high, it can be considered to adjust the implementation conditions of the intervention measure or choose other alternatives. According to the optimized risk treatment plan of the condition and the current use of beds in the critical care medicine department, the configuration of medical staff, the use of medical equipment and other information, the medical resources are optimized and adjusted. For example, the operation research method can be used to establish a medical resource allocation model, and the patient's severity of the condition, treatment needs, expected hospitalization time and other factors are used as the input of the model to optimize the allocation of beds, medical staff scheduling, medical equipment use and other plans, and generate an optimized medical resource allocation plan, such as "arrange patient A to bed No. 3 in the ICU ward, and assign nurse B to be responsible for the nursing work of the patient". According to the optimized medical resource allocation plan, a medical resource scheduling instruction is generated, such as "transfer the ventilator from ward No. 5 to ward No. 3". Individualized diagnosis and treatment decision-making support strategies (including risk warning information, risk intervention timing recommendations, optimized disease risk management plans, etc.) are integrated into the intensive care system. For example, risk warning information and intervention recommendations are displayed in the electronic medical record system to assist medical staff in making clinical decisions. Medical resource scheduling instructions are executed through the intensive care system to achieve digital management of critical care information and improve the efficiency of medical resource utilization.

[0105] Preferably, the patient data acquisition module is used to perform the following steps:

[0106] Step S11: inputting patient information data of the Department of Critical Care Medicine using a terminal device;

[0107] Step S12: performing data standardization processing on the patient information data of the Department of Critical Care Medicine to generate standard patient information data;

[0108] Step S13: desensitizing the standard patient information data and performing differential privacy protection processing to generate patient privacy protection data;

[0109] Step S14: data association is performed on the standard patient information data through the patient privacy protection data to obtain patient information index association data;

[0110] Step S15: Based on the patient information index association data, the standard patient information data is distributedly stored using a distributed storage system, thereby obtaining preliminary patient information storage data.

[0111] In an embodiment of the present invention, each bed in the intensive care unit (ICU) is equipped with a dedicated terminal device, such as a tablet computer or a mobile workstation. Medical staff enter various information of the patient through the terminal device. The entered information should be as comprehensive as possible and entered in a predefined format. For example, the basic information of the patient can be entered using a structured form, including name (replaced by a predefined unique identifier), gender, age, admission time, chief complaint, etc. Vital sign data, such as heart rate, blood pressure, respiratory rate, blood oxygen saturation, etc., can be collected in real time by sensors connected to the terminal device and automatically uploaded to the system, or can be manually entered by medical staff. Test and examination data can be automatically obtained by docking with the hospital information system (HIS) and the picture archiving and communication system (PACS). Due to the diversity of data sources and differences in entry formats, the original data needs to be standardized. Data cleaning and conversion tools, such as OpenRefine or Talend, can be used to unify the data format, convert the code, unify the unit, etc. For example, ISO8601 format. The disease codes used in different systems are different and need to be uniformly converted to ICD-10 codes. The blood pressure data collected by different devices have different units and need to be uniformly converted to mmHg. Perform data desensitization on the standard patient information data to remove or replace the patient's direct identification information, such as replacing the patient's name with a unique patient ID, and use differential privacy technology to further protect the privacy of the desensitized data. For example, the Laplace mechanism can be used to add random noise that meets a specific privacy budget to numerical data (such as blood pressure and heart rate), or the exponential mechanism can be used to perturb discrete data (such as gender and diagnosis), so as to retain the overall statistical characteristics of the data set without leaking individual patient information. Use a hash function to encrypt the patient's privacy protection data and generate a unique hash value as the index of the patient. Store the indexes of all patients in a data table and associate the table with the standard patient information data. In this way, when performing data analysis, the patient index can be used to associate with the corresponding standard patient information data without directly accessing the patient's privacy protection data. The data can be sharded according to the patient index, and the data of different shards can be stored on different nodes. At the same time, access control and encryption are performed on the database to further ensure data security. Only authorized users can access and use the data. Through distributed storage, massive amounts of patient information data can be effectively managed and processed.

[0112] Preferably, the physiological risk assessment module is used to perform the following steps:

[0113] Step S21: Perform integrity check on the preliminary patient information storage data to obtain data integrity assessment data;

[0114] Step S22: Identify missing data fields based on the data integrity assessment data, and predict missing field values ​​to obtain missing field recommended value data;

[0115] Step S23: Feedback the missing field recommended value data to the terminal device to remind the medical staff to make additional inputs to obtain the patient information supplementary data;

[0116] Step S24: performing data filling processing on the preliminary patient information storage data based on the patient information supplementary data to obtain complete critically ill patient data;

[0117] Step S25: Quantitatively assess the patient's risk level on the complete critically ill patient data to generate patient's disease severity grading data; perform real-time physiological deviation analysis based on the patient's disease severity grading data to generate physiological parameter real-time deviation data;

[0118] Step S26: assigning time sensitivity weights to the real-time deviation data of physiological parameters based on the preset critical illness time sensitivity rules, and performing physiological indicator risk processing to obtain abnormal physiological indicator risk data.

[0119] In the embodiment of the present invention, the missing rate, null value rate and other indicators of each data field can be counted, and whether the data field is complete can be determined according to a preset threshold. It is also possible to check whether the value of the data field conforms to the predefined data type and value range, such as checking whether the value of the age field is a positive integer, checking whether the value of the blood pressure field is within a reasonable range, etc. The data integrity check can find problems such as missing values ​​and wrong values ​​in the data. For example, if the blood oxygen saturation data of a patient is missing, the K nearest neighbor algorithm can be used to predict the missing blood oxygen saturation value of the patient based on the patient's other vital signs data (such as heart rate, blood pressure) and the blood oxygen saturation data of other patients with similar characteristics. The missing field value prediction can use the existing data information to restore the missing data as much as possible and improve the integrity of the data. The recommended value data of the missing field is displayed on the terminal device in a striking manner to remind medical staff to confirm and supplement. For example, on the patient information interface, the missing field is marked in red, and the predicted value is displayed next to it, which is convenient for medical staff to quickly locate and process. Medical staff can judge whether the recommended value is reasonable based on their professional knowledge and clinical experience, and make necessary corrections. Update the missing field values ​​supplemented by the medical staff to the database, or directly fill the recommended values ​​into the missing fields (if the medical staff did not modify them). Based on the APACHEII scoring system, the patient's various physiological indicators, laboratory test indicators, imaging test indicators and treatment information are used as input features to predict the risk of worsening of the patient's condition and divide the patient into different levels of severity (such as mild, moderate, severe, and extremely severe). Compare the patient's current physiological indicator data with the physiological indicator baseline data of patients of the same level, and calculate the real-time deviation value of each physiological indicator, such as heart rate deviation, blood pressure deviation, blood oxygen saturation deviation, etc. According to the preset critical illness time sensitivity rules, the real-time deviation data of different physiological parameters are assigned time sensitivity weights. For example, for patients with critical illness, slight fluctuations in blood pressure are more important than those for patients with stable conditions, so a higher weight needs to be assigned, and the system will determine that the indicator has a higher risk. For example, based on clinical experience, risk thresholds and time windows for different physiological indicators can be set. For example, if the heart rate exceeds 120 beats / minute for more than 30 minutes, it is considered that there is a risk of tachycardia, and if the blood pressure is continuously lower than 90 / 60 mmHg for more than 1 hour, it is considered that there is a risk of hypotension. Then, based on the real-time deviation data of physiological parameters and preset rules, it is determined whether the patient has a risk of abnormal physiological indicators, and a risk score is given according to the severity and duration of the risk to generate abnormal physiological indicator risk data. For example, the risk score for tachycardia can be 3 points, the risk score for hypotension can be 5 points, etc. The time sensitivity weight assignment can reflect the importance of different physiological indicators in different time periods. For example, for critically ill patients, blood pressure fluctuations in a short period of time need more attention than long-term heart rate fluctuations.

[0120] Preferably, step S22 includes the following steps:

[0121] Step S221: Identify missing data fields based on data integrity assessment data and generate missing field data;

[0122] Step S222: Identify the missing field type of the missing field data to obtain the missing field type data;

[0123] Step S223: query the historical filling records based on the missing field type data, and perform similarity matching of the complete historical records to generate complete historical filling record data;

[0124] Step S224: extracting context-related fields from missing field type data by filling in the complete historical record data, and generating context-related field data;

[0125] Step S225: Based on a preset conventional data filling strategy, preliminary data supplementation processing is performed on the missing field data through the context-related field data to generate preliminary supplementary suggestion data;

[0126] Step S226: Perform logical verification on the preliminary supplementary suggestion data, conduct validity evaluation, and generate missing field recommendation value data.

[0127] In an embodiment of the present invention, the system analyzes the data integrity assessment data and identifies all missing fields in the preliminary patient information storage data. For example, the data integrity assessment data shows that patient A lacks the "allergy history" field, and patient B lacks the "most recent renal function test results" field, etc. The missing field type is identified for the missing field data to obtain the missing field type data. For example, the missing fields can be divided into different types according to the data type and value range of the field, such as numerical missing fields (such as blood pressure, heart rate), text missing fields (such as chief complaint, diagnosis), categorical variable missing fields (such as gender, blood type), etc. The missing fields can also be divided into different types according to the semantic information and clinical significance of the field, such as missing fields related to the severity of the disease, missing fields related to the treatment plan, and missing fields related to prognosis. According to the type of missing fields and other characteristic information of the patient (such as age, gender, diagnosis, etc.), the complete history filling records of patients with similar characteristics are retrieved from the database. Then, the similarity between the current patient and each historical record can be calculated using a similarity measurement method (such as cosine similarity, Euclidean distance, etc.), and several historical records with the highest similarity are selected as the complete history filling record data. For example, for the missing "Latest renal function test result" field, its context-related fields include "Renal function test time", "Blood creatinine", "Urea nitrogen", etc. The system records the information of these related fields and the type of relationship between them and the missing fields, such as "time association", "value association", etc. The system uses the context-related field data to preliminarily supplement the missing fields according to the preset general data filling strategy and generates preliminary supplementary suggestion data. Common strategies include: Default value filling: For some fields, default values ​​can be set according to medical common sense or experience, such as setting the default value of the "Is there a smoking history" field to "No". Related field inference: Infer the value of the missing field based on the value of the context-related field. For example, if the patient's "Renal function test time" is "three days ago", and the values ​​of "Blood creatinine" and "Urea nitrogen" are both within the normal range, it can be inferred that the patient's "Latest renal function test result" is "Normal". Borrowing similar patient information: Borrowing the information of similar patients to fill in missing fields. For example, if patient A lacks the "allergy history" field, while patient B, who is similar to him, has the "allergy history" field filled in completely, then patient B's "allergy history" information can be used as the preliminary supplementary suggestion for patient A. The system performs logical verification on the preliminary supplementary suggestion data, such as checking whether the data type matches, whether the data range is reasonable, and whether there are logical contradictions between the data. At the same time, the system will also evaluate the effectiveness of the preliminary supplementary suggestion data based on preset evaluation indicators, such as calculating the confidence and accuracy of the predicted value.For example, if the system infers that the patient's "most recent renal function test result" is "normal" based on the "associated field inference" strategy, but the patient's "serum creatinine" and "urea nitrogen" values ​​are higher than normal, the system will determine that there is a logical contradiction in the inference result and mark it as "unreliable". If the preliminary supplementary suggestion data does not conform to the logical rules or has low validity, it needs to be corrected or regenerated; if the preliminary supplementary suggestion data conforms to the logical rules and has high validity, it can be used as the missing field recommended value data for subsequent data filling processing. The system will use the preliminary supplementary suggestion data that has undergone logical verification and validity evaluation as the missing field recommended value data, and feedback it to medical staff for confirmation and supplementation.

[0128] Preferably, step S25 comprises the following steps:

[0129] Step S251: Acquire the patient's physiological sign data and electronic medical record data based on the complete critically ill patient data;

[0130] Step S252: extracting the medical condition text diagnosis information from the electronic medical record data to obtain the medical condition diagnosis information data;

[0131] Step S253: evaluating the severity of the patient's condition by using the disease diagnosis information data to evaluate the patient's physiological sign data, and obtaining the patient's severity classification data;

[0132] Step S254: marking high-risk patients according to the patient's disease severity classification data, and reading the high-risk patient's physiological data to generate high-risk patient physiological parameters;

[0133] Step S255: dynamically regularize the physiological parameters of high-risk patients in time series to generate physiological parameters of time-series aligned patients;

[0134] Step S256: constructing a physiological parameter dynamic baseline model using a preset time series model and a preset physiological parameter baseline reference rule;

[0135] Step S257: Perform real-time physiological deviation analysis on the physiological parameters of the time-series aligned patients through a physiological parameter dynamic baseline model to generate physiological parameter real-time deviation data.

[0136] In an embodiment of the present invention, the system extracts the patient's physiological sign data and electronic medical record data from the complete critically ill patient data based on the patient's unique identifier. Physiological sign data: including heart rate, blood pressure, respiratory rate, blood oxygen saturation, body temperature, etc. These data can be obtained from a real-time database connected to medical equipment, or extracted from stored historical records. Electronic medical record data: including admission records, diagnosis records, examination and inspection reports, operation records, medical order records, etc., which are usually stored in the electronic medical record system in text form. The system uses natural language processing (NLP) technology, such as named entity recognition and relationship extraction, to extract information related to the patient's condition diagnosis from the electronic medical record data and generate condition diagnosis information data. For example, the system can identify that the patient's main diagnosis is "acute respiratory distress syndrome", and comorbidities include "hypertension", "diabetes", etc., as well as related condition descriptions, such as "the patient has shortness of breath, sitting breathing, and cyanosis of the lips". Based on the disease diagnosis information data and the pre-set disease severity assessment rules, the system conducts a comprehensive analysis of the patient's physiological sign data, assesses the severity of the patient's disease, and divides it into different levels, such as mild, moderate, severe and critical. The assessment rules can refer to commonly used clinical scoring systems such as APACHEII score and SOFA score, or can be adjusted according to specific circumstances. For example, the rule can be set as: if the patient's main diagnosis is "acute respiratory distress syndrome" and the blood oxygen saturation is continuously lower than 90%, it will be assessed as "severe" critical. The system marks "severe" and "critical" patients as high-risk patients according to the preset severity threshold. For patients marked as high-risk patients, the system reads their detailed physiological sign data, such as heart rate per minute, blood pressure per hour, etc., to generate physiological parameters of high-risk patients. Since the physiological parameters of different patients are collected at different times, for example, the blood pressure of patient A is collected once an hour, while the blood pressure of patient B is collected once every half an hour, this will lead to difficulties in subsequent analysis. Therefore, it is necessary to dynamically regularize the time series of physiological parameters of high-risk patients, align data at different time intervals to a unified time axis, eliminate differences caused by different sampling frequencies, different time starting points, etc., so that the time series of physiological parameters of different patients can be compared and analyzed in the time dimension, and generate time-aligned patient physiological parameters. Use the preset time series model and the preset physiological parameter baseline reference rules to build a dynamic baseline model of physiological parameters. For example, time series models such as recurrent neural networks (RNN) and long short-term memory networks (LSTM) can be used to learn the dynamic change law of each physiological parameter based on historical data, and constrain the model according to the preset physiological parameter baseline reference rules (such as normal heart rate range, normal blood pressure range, etc.), and build a dynamic baseline model that can predict the change trend of each physiological parameter in the future.The system inputs the time-series aligned patient physiological parameters into the physiological parameter dynamic baseline model, predicts the expected value of the patient's physiological parameters at the current time point, and compares it with the actual observed value to calculate the real-time deviation. For example, if the model predicts that patient A's blood pressure should be 120 / 80mmHg at the current time point, and the actual observed value is 110 / 70mmHg, the patient's blood pressure real-time deviation is -10 / -10mmHg. The system continuously monitors the patient's physiological parameters and updates the deviation data in real time.

[0137] As an example of the present invention, refer to Figure 2 As shown, the disease risk warning module described in this example includes:

[0138] Step S31: Process case retrieval conditions according to the patient's disease severity classification data and physiological index abnormal risk data to obtain case retrieval condition data;

[0139] In an embodiment of the present invention, case retrieval condition processing is performed based on the patient's disease severity grading data and physiological indicator abnormal risk data to obtain case retrieval condition data. For example, the patient's disease severity level, physiological indicator abnormality type, abnormality level and other information can be converted into specific retrieval conditions, such as "the disease severity level is severe", "the heart rate continues to exceed 120 times / minute for more than 30 minutes", "the blood pressure continues to be lower than 90 / 60mmHg for more than 1 hour", etc. Some combined conditions can also be set based on clinical experience and medical knowledge, such as "the disease severity level is severe and there is a risk of respiratory failure", etc. Case retrieval condition processing can convert the patient's disease information into conditions that can be used to retrieve historical cases.

[0140] Step S32: Acquire a historical case database; use the case search condition data to perform a similar case search on the historical case database to obtain similar case data;

[0141] In an embodiment of the present invention, a historical case database is obtained, which contains a large number of case data of critically ill patients who have completed the diagnosis and treatment process, including the patient's basic information, condition information, diagnosis and treatment process, treatment plan, prognosis, etc. The historical case database is retrieved for similar cases using the case retrieval condition data to obtain similar case data. For example, a rule-based retrieval method can be used to filter out qualified cases from the historical case database according to the case retrieval condition data; or a similarity-based measurement method can be used to calculate the similarity between the current patient and each historical case, and select several cases with the highest similarity as similar case data. The similarity measurement method can consider multiple factors such as the severity of the patient's condition, abnormal physiological indicators, and disease diagnosis information. For example, the similarity between patients can be calculated using cosine similarity, Euclidean distance, etc. For example, if the current patient is a patient with a severe degree of severity and a risk of respiratory failure, the case data of other patients with a severe degree of severity and a risk of respiratory failure can be retrieved from the historical case database, and the similarity can be calculated, and the cases with the highest similarity can be selected as similar case data.

[0142] Step S33: constructing a patient group interaction relationship network based on similar case data; identifying key nodes of the patient group interaction relationship network to obtain key node case patient data; performing case development path analysis based on the key node case patient data to generate case development path data;

[0143] In an embodiment of the present invention, a patient group interaction network is constructed based on similar case data. For example, each patient can be taken as a node. If there is similarity between two patients (for example, similar severity of illness, similar abnormal physiological indicators, etc.), a connecting edge is established between the two nodes, and the weight of the edge can represent the magnitude of the similarity. A patient group interaction network is constructed based on similar case data. For example, each patient can be taken as a node. If there is similarity between two patients (for example, similar severity of illness, similar abnormal physiological indicators, etc.), a connecting edge is established between the two nodes, and the weight of the edge can represent the magnitude of the similarity. The system uses complex network analysis techniques, such as degree centrality, betweenness centrality and other indicators, to identify key nodes in the network, that is, cases with a high degree of association with other patients. These key node cases usually represent a typical disease development pattern. The system analyzes key node case patient data, such as the time sequence of disease development, the trend of changes in key physiological indicators, the implementation effect of the treatment plan, etc., to generate case development path data. These path data reflect the disease evolution process of similar patient groups at different stages, and provide a reference for predicting the disease development trend of the current patient.

[0144] Step S34: performing multi-dimensional index correlation analysis of physiological indexes, test indexes, treatment plans and nursing measures on key node case patient data to obtain key node correlation index data;

[0145] In an embodiment of the present invention, a multi-dimensional indicator correlation analysis of physiological indicators, test indicators, treatment plans and nursing measures is performed on the patient data of key node cases to obtain key node correlation indicator data. For example, statistical analysis methods (such as correlation analysis, regression analysis, etc.) or machine learning methods (such as association rule mining, decision trees, etc.) can be used to analyze whether there is a correlation between the physiological indicators, test indicators, treatment plans and nursing measures of key node case patients, and identify key indicators related to the progression of the disease, prognosis, etc. For example, the correlation between physiological indicators such as heart rate, blood pressure, and blood oxygen saturation and the deterioration of the disease can be analyzed, or the prognosis of patients after treatment with a certain drug can be analyzed.

[0146] Step S35: Perform intelligent risk warning on the case development path data through key node association indicator data to generate intelligent disease risk warning data, wherein the intelligent disease risk warning data includes warning level data, warning reason data and recommended measures data.

[0147] In an embodiment of the present invention, the system integrates the key node associated indicator data and the case development path data to predict the future disease risk of the current patient. For example, a risk warning model can be constructed based on the key node associated indicator data and the case development path data. For example, a model can be trained using machine learning methods (such as logistic regression, support vector machine, etc.) to predict the probability of the patient's disease deterioration or complications in the future, and the prediction results are converted into warning levels (such as low risk, medium risk, high risk), and intelligent disease risk warning data is generated, including: Warning reason data: Explain the reason for triggering the warning, such as "the patient's blood oxygen saturation continues to decline, and there is a risk of respiratory failure." Recommended measures data: Based on the treatment experience of similar cases, corresponding recommended measures are given, such as "it is recommended to perform a bedside chest X-ray examination" and "it is recommended to evaluate whether endotracheal intubation is needed." For example, if the system finds that the current patient has a high degree of similarity with a key node case, and the key node case has a sudden drop in blood pressure in the next 24 hours, the system will generate a high-risk warning to remind medical staff to pay close attention to the patient's blood pressure changes and make corresponding rescue preparations.

[0148] Preferably, step S33 includes the following steps:

[0149] Step S331: constructing a patient group feature vector for similar case data to obtain patient group feature vector data;

[0150] Step S332: using the fuzzy C-means clustering algorithm to perform patient group characteristic clustering analysis on the patient group characteristic vector data to obtain patient group characteristic clustering data;

[0151] Step S333: extracting relevant physiological indicators according to the patient group characteristic clustering data to obtain case-related physiological indicator data;

[0152] Step S334: analyzing the case development stage according to the case-related physiological indicator data to obtain the case development stage data;

[0153] Step S335: Process the key physiological indicator mutation nodes on the case development stage data, and perform case development path trend analysis to obtain case development path data.

[0154] In an embodiment of the present invention, a patient group feature vector is constructed for similar case data to obtain patient group feature vector data. For example, the age, gender, disease diagnosis information, main physiological indicators, treatment plan and other information of each patient can be extracted, and these information can be converted into numerical or categorical features, such as converting age into continuous numerical features, converting gender into binary categorical features, converting disease diagnosis information into multi-valued categorical features, etc. Then, the features of each patient are combined into a feature vector, for example, one-hot encoding can be used to convert categorical features into numerical features, and all features can be spliced ​​into a vector. The patient group feature vector data is clustered and analyzed by the fuzzy C-means clustering algorithm to obtain patient group feature clustering data. For example, the number of cluster clusters can be set, such as dividing the patient group into 5 clusters, and then using the fuzzy C-means clustering algorithm to assign each patient to one or more clusters, each cluster representing a class of patient groups with similar characteristics. According to the changing trend of physiological indicators, the development process of the case is divided into different stages, such as stable disease period, disease exacerbation period, disease recovery period, etc. For example, time series analysis methods (such as hidden Markov models, dynamic time warping, etc.) can be used to analyze the time series of physiological indicators and identify different development stages. Identify the time points when physiological indicators suddenly change, such as a sudden increase in heart rate, a sudden drop in blood pressure, etc., and use these time points as key nodes. Then, the change trend of physiological indicators before and after the key nodes can be analyzed, such as whether the increase in heart rate is accompanied by a drop in blood pressure, a drop in blood oxygen saturation, etc., and a case development path diagram can be constructed. For example, a directed acyclic graph can be used to represent the case development path, with nodes representing different development stages and edges representing the transfer relationship between stages. For example, the system identifies a key node case of the "respiratory failure" type, whose blood oxygen saturation continues to decrease within 24 hours after admission, and a sudden drop mutation node occurs at 48 hours. The system analyzes the development path trend of the case and believes that the patient's condition gradually worsens in the early stages and deteriorates sharply at 48 hours, and more active treatment measures need to be taken in time, such as endotracheal intubation and mechanical ventilation.

[0155] Preferably, step S333 includes the following steps:

[0156] The nodes of the patient group interaction network are constructed according to the patient group characteristic clustering data, and the edges of the patient group interaction network are constructed according to the disease correlation and treatment scheme similarity between the patient groups, thereby constructing the patient group interaction network;

[0157] Conduct topological structure analysis on the patient group interaction network, identify key nodes in the interaction network, and obtain key node case patient data;

[0158] According to the patient data of key node cases, the severe type of the case is identified to obtain the severe type data of the case;

[0159] Relevant physiological indicators are extracted from the severe type data of the case through the abnormal risk data of physiological indicators to obtain the case-related physiological indicator data.

[0160] In an embodiment of the present invention, nodes of a patient group interaction network are constructed based on clustering data of patient group characteristics, and each node represents a patient group. Then, edges of a patient group interaction network are constructed based on the correlation of disease conditions and similarity of treatment plans between patient groups. For example, if two patient groups suffer from the same disease or receive similar treatment plans, a connecting edge is established between the two nodes, and the weight of the edge can represent the degree of correlation of disease conditions or similarity of treatment plans. The system analyzes the topological structure of the patient group interaction network, such as network diameter, clustering coefficient, node centrality and other indicators, and identifies key nodes in the network. Key nodes usually refer to nodes with higher indicators such as degree centrality and betweenness centrality. They play an important role in the network, such as connecting different patient groups and transmitting important disease information. The system identifies the type of severe illness to which the case belongs, such as respiratory failure, heart failure, sepsis, etc., based on the characteristics of the patient data of the key node case, such as diagnostic information, physiological indicators, treatment plans, etc., and stores the identification results as case severe illness type data. According to the case severe illness type data, the system extracts physiological indicators related to the type of severe illness from the physiological indicator abnormal risk data. For example, for cases of the "respiratory failure" type, relevant physiological indicators include respiratory rate, blood oxygen saturation, PaCO2, PaO2, etc.

[0161] As an example of the present invention, refer to Figure 3 As shown, the risk intervention timing decision module described in this example includes:

[0162] Step S41: constructing a Bayesian network model based on preset critical care medical knowledge data, wherein the Bayesian network model includes risk factor nodes, physiological indicator nodes, and risk event nodes;

[0163] In the embodiment of the present invention, critical care medical knowledge data can be derived from clinical guidelines, expert experience, medical literature, etc. For example, diseases such as hypertension and diabetes can be used as risk factor nodes, physiological indicators such as heart rate, blood pressure, and blood oxygen saturation can be used as physiological indicator nodes, and complications such as heart failure, respiratory failure, and septic shock can be used as risk event nodes. Then, based on medical knowledge and clinical experience, the causal relationship between nodes is defined, for example, hypertension can lead to heart failure, and a fast heart rate can lead to myocardial ischemia, etc., and a conditional probability table is used to represent the probability distribution of each node when its parent node takes different values.

[0164] Step S42: using the abnormal risk data of physiological indicators and the intelligent disease risk warning data to assign time node states to the Bayesian network model, and using the Bayesian network inference algorithm to calculate the event probability of the risk event node to obtain risk probability deduction data;

[0165] In an embodiment of the present invention, the patient's current heart rate, blood pressure, blood oxygen saturation and other physiological index values ​​are assigned to the corresponding physiological index nodes, and the risk events predicted in the intelligent disease risk warning data are assigned to the corresponding risk event nodes. Then, the Bayesian network inference algorithm is used to calculate the event probability of the risk event node to obtain the risk probability deduction data. For example, the confidence propagation algorithm or the variable elimination algorithm can be used to calculate the probability of the patient experiencing risk events such as heart failure and respiratory failure in the future.

[0166] Step S43: constructing a fuzzy reasoning decision model through a Bayesian network model based on the risk probability deduction data and the preset risk intervention threshold, wherein the fuzzy reasoning decision model includes a risk probability fuzzy variable, a risk development speed fuzzy variable, and a patient individual characteristic fuzzy variable;

[0167] In an embodiment of the present invention, the system constructs a fuzzy reasoning decision model based on the risk probability deduction data and the preset risk intervention threshold. The model uses fuzzy logic to handle uncertainty and ambiguity, and includes the following fuzzy variables: Risk probability fuzzy variable: The risk probability is converted into a fuzzy language variable, such as "low", "medium", and "high". Risk development speed fuzzy variable: According to the changing trend of the risk probability, the risk development speed is converted into a fuzzy language variable, such as "slow", "stable", and "rapid". Patient individual characteristic fuzzy variable: The individual characteristics of the patient, such as age, comorbidities, etc., are converted into fuzzy language variables, such as "young", "old", "healthy", "weak", etc. The fuzzy reasoning decision model uses fuzzy rules to define the intervention measures to be taken in different situations. For example, a rule can be defined as: If the "risk probability" is "high" and the "risk development speed" is "rapid", then "immediate intervention is recommended".

[0168] Step S44: performing risk intervention timing reasoning according to the fuzzy reasoning decision model to obtain risk intervention timing decision data;

[0169] In the embodiment of the present invention, the system inputs the current risk probability, risk development speed and individual characteristics of the patient into the fuzzy reasoning decision model, performs reasoning based on fuzzy rules, determines the best intervention time, and generates risk intervention time decision data. For example, the model recommends "intervene immediately", "continue observation for 2 hours", "recheck tomorrow morning", etc.

[0170] Step S45: supplementing the recommended measures data in the intelligent disease risk warning data with the risk intervention opportunity decision data to generate supplementary disease risk warning data;

[0171] In the embodiment of the present invention, the system integrates the risk intervention timing decision data into the intelligent disease risk warning data, supplements the recommended measures, and generates more comprehensive supplementary disease risk warning data. For example, the system can supplement the recommended measure of "immediate tracheal intubation is recommended" to "tracheal intubation is recommended within 1 hour, and non-invasive ventilation can be tried for 2 hours if the patient's condition permits."

[0172] Step S46: Utilize the preset medical staff information receiving authority matrix to perform multi-level linkage push processing on the supplementary disease risk warning data, and perform warning information feedback statistics to generate warning information feedback data.

[0173] In an embodiment of the present invention, a preset medical staff information receiving authority matrix is ​​used to perform multi-level linkage push processing on the supplementary disease risk warning data. For example, high-risk warning information can be pushed to the attending physician and the nurse on duty, medium-risk warning information can be pushed to the doctor on duty, and low-risk warning information can be pushed to the responsible nurse, etc. The medical staff information receiving authority matrix can be set according to the hospital's management system and actual conditions. For example, different authorities can be set according to factors such as the doctor's title and the nurse's level. Then, the warning information feedback statistics are performed to generate warning information feedback data. For example, it can record whether the medical staff has viewed the warning information, whether the corresponding measures have been taken, and how the intervention effect is.

[0174] Preferably, the individualized diagnosis and treatment decision-making assistance module is used to perform the following steps:

[0175] Step S51: Analyze the disease risk treatment plan according to the early warning information feedback data to generate disease risk treatment plan data;

[0176] Step S52: performing a medical resource demand assessment on the disease risk treatment plan data, and generating a medical resource allocation plan to obtain a medical resource allocation plan;

[0177] Step S53: Calculate the patient waiting time according to the medical resource allocation plan to obtain patient waiting time data;

[0178] Step S54: Perform resource allocation feasibility analysis on the medical resource allocation plan based on the patient waiting time data, and make optimization adjustments to obtain an optimized medical resource allocation plan;

[0179] Step S55: Process the medical resource dispatching instructions according to the optimized medical resource allocation plan, and send the medical resource dispatching instructions to relevant medical staff and medical equipment, so as to obtain an individualized diagnosis and treatment auxiliary decision-making strategy.

[0180] In an embodiment of the present invention, the system analyzes the early warning information feedback data, such as the early warning level, the reason for the early warning, the recommended measures, the feedback from medical staff, etc., and combines the patient's real-time condition data and historical medical record data to generate personalized condition risk treatment plan data. For example, suppose a patient triggers a "high risk" early warning, and the reason for the early warning is "blood oxygen saturation continues to decline, and there is a risk of respiratory failure", and the recommended measure is "immediate endotracheal intubation". The system will generate different treatment plans based on factors such as the patient's age, comorbidities, and course of disease, as well as feedback from medical staff, such as "the patient's family refuses endotracheal intubation", etc., for example: Plan 1: Immediate endotracheal intubation and connect a ventilator to assist breathing. Plan 2: Try non-invasive ventilation for 2 hours. If the blood oxygen saturation cannot be maintained within a safe range, perform endotracheal intubation. Plan 3: According to the wishes of the patient's family, give up endotracheal intubation and perform comfort treatment. Each plan contains detailed operation steps, required resources, expected results, and risk assessment information. The system evaluates the required medical resources, such as doctors, nurses, ventilators, beds, etc., based on the data of the disease risk treatment plan, and generates a medical resource allocation plan based on the current resource availability. For example, operations research methods such as linear programming or integer programming can be used to build a medical resource allocation model, with the goal of minimizing the patient's waiting time or maximizing the utilization of medical resources. The constraint is to meet the treatment needs and medical resource needs of all patients. The system calculates the patient's expected waiting time based on the medical resource allocation plan. For example, if there is only one ventilator available and two other patients also need to use a ventilator, the system will determine the order of ventilator use based on the severity of the patient's condition and the waiting time, and calculate the expected waiting time for each patient. The system analyzes the patient's waiting time data and evaluates the feasibility of the medical resource allocation plan. If the patient's expected waiting time is too long, it will delay treatment and cause the condition to worsen. In this case, the system needs to optimize and adjust the medical resource allocation plan, for example: Resource scheduling: allocate resources from other departments or hospitals. Plan adjustment: select a treatment plan with less resource requirements. Priority sorting: based on the severity of the patient's condition, give priority to treating patients with more serious conditions. Medical resource dispatch instructions are processed according to the optimized medical resource allocation plan, and sent to relevant medical staff and medical equipment, so as to obtain individualized diagnosis and treatment auxiliary decision-making strategies. For example, specific medical resource dispatch instructions can be generated according to the medical resource allocation plan, such as "transfer patient A to bed No. 3", "notify doctor B to go to the emergency room immediately", etc., and these instructions are sent to relevant medical staff and medical equipment through the hospital's information system.For example, message queues or real-time communication technologies can be used to send medical resource scheduling instructions to relevant medical staff and medical equipment. For example, message queue systems such as Kafka and RabbitMQ can be used to send instructions to relevant applications, or real-time communication technologies such as WebSocket can be used to push instructions to medical staff's mobile devices.

[0181] The beneficial effect of this application is that the terminal device is used to collect the patient's physiological sign data in real time, and the diagnosis information of the disease is extracted and the severity of the patient's disease is assessed in combination with the electronic medical record data, so as to realize the real-time monitoring of the patient's disease. By constructing a patient group interaction relationship network, analyzing the development path of similar cases, and combining the preset critical medical knowledge data to build a Bayesian network model, the prediction of the patient's future disease development trend and the risk probability deduction are realized, and the potential disease risk is discovered in time, avoiding the limitation of relying on the experience of medical staff for judgment in the traditional mode, and improving the accuracy and timeliness of risk identification. Through technical means such as data standardization, data desensitization and differential privacy protection, the standardized storage and safe sharing of patient information are realized, avoiding the problems of incomplete, inaccurate and easy loss of information caused by manual records in the traditional mode. In addition, the missing field value prediction is also carried out, and the medical staff is reminded to make additional inputs through the terminal device, which further improves the integrity and accuracy of the data, reduces the occurrence of human errors and reduces the work pressure of medical staff. Based on the Bayesian network model, a fuzzy reasoning decision model is constructed, and the risk probability, risk development speed and individual characteristics of the patient are comprehensively considered to reason the timing of risk intervention, providing personalized intervention timing suggestions for medical staff. It avoids the "one-size-fits-all" intervention approach in the traditional model, realizes accurate intervention timing decision-making, and improves the effectiveness and safety of intervention. The system automatically evaluates the demand for medical resources based on early warning information and treatment plans, and generates optimized medical resource allocation plans based on real-time resource availability. Through intelligent scheduling instructions, it realizes the rational allocation and efficient use of medical resources.

[0182] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0183] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. An information digital management system based on the Department of Critical Care Medicine, characterized in that: include: Control circuit; A patient data acquisition module is electrically connected to the control circuit, and the patient data acquisition module includes: a mobile computing terminal, a multifunctional vital signs monitor and an encrypted storage component. The patient data acquisition module is used to input the patient information data of the critical care medicine department using the mobile computing terminal; perform differential privacy protection processing on the patient information data of the critical care medicine department to generate patient privacy protection data; perform distributed data storage according to the patient privacy protection data to obtain preliminary patient information storage data; A physiological risk assessment module is electrically connected to the control circuit, and includes an intelligent data analysis component. The physiological risk assessment module is used to predict missing field values ​​for preliminary patient information storage data, and perform data filling processing to obtain complete critically ill patient data; perform a quantitative assessment of the patient's risk level on the complete critically ill patient data to generate patient's disease severity grading data; perform a physiological indicator time risk analysis based on the patient's disease severity grading data to obtain physiological indicator abnormal risk data; The disease risk warning module is electrically connected to the control circuit, and includes a case retrieval database and an intelligent warning terminal, which is used to process case retrieval conditions according to the abnormal risk data of physiological indicators, and to retrieve similar cases to obtain similar case data; to build a patient group interaction relationship network according to the similar case data; to perform risk intelligent warning according to the patient group interaction relationship network, and to generate intelligent disease risk warning data; wherein, the disease risk warning module is specifically used to perform the following steps: Step S31: Process case retrieval conditions according to the patient's disease severity classification data and physiological index abnormal risk data to obtain case retrieval condition data; Step S32: Acquire a historical case database; use the case search condition data to perform a similar case search on the historical case database to obtain similar case data; Step S33: construct a patient group interaction network based on similar case data; identify key nodes of the patient group interaction network to obtain key node case patient data; perform case development path analysis based on key node case patient data to generate case development path data; wherein step S33 is specifically as follows: Step S331: constructing a patient group feature vector for similar case data to obtain patient group feature vector data; Step S332: using the fuzzy C-means clustering algorithm to perform patient group characteristic clustering analysis on the patient group characteristic vector data to obtain patient group characteristic clustering data; Step S333: extracting relevant physiological indicators according to the patient group characteristic clustering data to obtain case-related physiological indicator data; Step S334: analyzing the case development stage according to the case-related physiological indicator data to obtain the case development stage data; Step S335: Process the key physiological indicator mutation nodes on the case development stage data, and perform case development path trend analysis to obtain case development path data; Step S34: performing multi-dimensional index correlation analysis of physiological indexes, test indexes, treatment plans and nursing measures on key node case patient data to obtain key node correlation index data; Step S35: Perform intelligent risk warning on the case development path data through key node association indicator data to generate intelligent disease risk warning data, wherein the intelligent disease risk warning data includes warning level data, warning reason data and recommended measures data; A risk intervention timing decision module is electrically connected to the control circuit, and the risk intervention timing decision module includes: an integrated interface of a Bayesian network analysis tool and a clinical information system feedback workstation, for constructing a Bayesian network model; using the Bayesian network model to perform risk intervention timing reasoning on the intelligent disease risk warning data to obtain risk intervention timing decision data; performing warning information feedback statistics on the risk intervention timing decision data and the intelligent disease risk warning data to generate warning information feedback data; An individualized diagnosis and treatment decision-making assistance module is electrically connected to the control circuit. The individualized diagnosis and treatment decision-making assistance module includes: a scheduling instruction generator, a doctor's order execution terminal and an integrated interface of a medical resource management system, which is used to analyze the disease risk treatment plan based on the early warning information feedback data, and optimize and adjust the medical resources to obtain an optimized medical resource allocation plan; process the medical resource scheduling instructions according to the optimized medical resource allocation plan to obtain an individualized diagnosis and treatment decision-making assistance strategy; integrate the individualized diagnosis and treatment decision-making assistance strategy into the intensive care system to implement digital information management of the critical care department.

2. The information digital management system based on the critical care medicine department according to claim 1 is characterized in that: The patient data collection module is used to perform the following steps: Step S11: inputting patient information data of the Department of Critical Care Medicine using a terminal device; Step S12: performing data standardization processing on the patient information data of the Department of Critical Care Medicine to generate standard patient information data; Step S13: desensitizing the standard patient information data and performing differential privacy protection processing to generate patient privacy protection data; Step S14: data association is performed on the standard patient information data through the patient privacy protection data to obtain patient information index association data; Step S15: Based on the patient information index association data, the standard patient information data is distributedly stored using a distributed storage system, thereby obtaining preliminary patient information storage data.

3. The information digital management system based on the critical care medicine department according to claim 1 is characterized in that: The physiological risk assessment module is used to perform the following steps: Step S21: Perform integrity check on the preliminary patient information storage data to obtain data integrity assessment data; Step S22: Identify missing data fields based on the data integrity assessment data, and predict missing field values ​​to obtain missing field recommended value data; Step S23: Feedback the missing field recommended value data to the terminal device to remind the medical staff to make additional inputs to obtain the patient information supplementary data; Step S24: performing data filling processing on the preliminary patient information storage data based on the patient information supplementary data to obtain complete critically ill patient data; Step S25: Quantitatively assess the patient's risk level on the complete critically ill patient data to generate patient's disease severity grading data; perform real-time physiological deviation analysis based on the patient's disease severity grading data to generate physiological parameter real-time deviation data; Step S26: assigning time sensitivity weights to the real-time deviation data of physiological parameters based on the preset critical illness time sensitivity rules, and performing physiological indicator risk processing to obtain abnormal physiological indicator risk data.

4. The information digital management system based on the critical care medicine department according to claim 3 is characterized in that: Step S22 includes the following steps: Step S221: Identify missing data fields based on data integrity assessment data and generate missing field data; Step S222: Identify the missing field type of the missing field data to obtain the missing field type data; Step S223: query the historical filling records based on the missing field type data, and perform similarity matching of the complete historical records to generate complete historical filling record data; Step S224: extracting context-related fields from missing field type data by filling in the complete historical record data, and generating context-related field data; Step S225: Based on a preset conventional data filling strategy, preliminary data supplementation processing is performed on the missing field data through the context-related field data to generate preliminary supplementary suggestion data; Step S226: Perform logical verification on the preliminary supplementary suggestion data, conduct validity evaluation, and generate missing field recommendation value data.

5. The information digital management system based on the critical care medicine department according to claim 3 is characterized in that: Step S25 includes the following steps: Step S251: Acquire the patient's physiological sign data and electronic medical record data based on the complete critically ill patient data; Step S252: extracting the medical condition text diagnosis information from the electronic medical record data to obtain the medical condition diagnosis information data; Step S253: evaluating the severity of the patient's condition by using the disease diagnosis information data to evaluate the patient's physiological sign data, and obtaining the patient's severity classification data; Step S254: marking high-risk patients according to the patient's disease severity classification data, and reading the high-risk patient's physiological data to generate high-risk patient physiological parameters; Step S255: dynamically regularize the physiological parameters of high-risk patients in time series to generate physiological parameters of time-series aligned patients; Step S256: constructing a physiological parameter dynamic baseline model using a preset time series model and a preset physiological parameter baseline reference rule; Step S257: Perform real-time physiological deviation analysis on the physiological parameters of the time-series aligned patients through a physiological parameter dynamic baseline model to generate physiological parameter real-time deviation data.

6. The information digital management system based on the critical care medicine department according to claim 1 is characterized in that: Step S333 includes the following steps: The nodes of the patient group interaction network are constructed according to the patient group characteristic clustering data, and the edges of the patient group interaction network are constructed according to the disease correlation and treatment scheme similarity between the patient groups, thereby constructing the patient group interaction network; Conduct topological structure analysis on the patient group interaction network, identify key nodes in the interaction network, and obtain key node case patient data; According to the patient data of key node cases, the severe type of the case is identified to obtain the severe type data of the case; Relevant physiological indicators are extracted from the severe type data of the case through the abnormal risk data of physiological indicators to obtain the case-related physiological indicator data.

7. The information digital management system based on the critical care medicine department according to claim 1 is characterized in that: The risk intervention timing decision module is used to perform the following steps: Step S41: constructing a Bayesian network model based on preset critical care medical knowledge data, wherein the Bayesian network model includes risk factor nodes, physiological indicator nodes, and risk event nodes; Step S42: using the abnormal risk data of physiological indicators and the intelligent disease risk warning data to assign time node states to the Bayesian network model, and using the Bayesian network inference algorithm to calculate the event probability of the risk event node to obtain risk probability deduction data; Step S43: constructing a fuzzy reasoning decision model through a Bayesian network model based on the risk probability deduction data and the preset risk intervention threshold, wherein the fuzzy reasoning decision model includes a risk probability fuzzy variable, a risk development speed fuzzy variable, and a patient individual characteristic fuzzy variable; Step S44: performing risk intervention timing reasoning according to the fuzzy reasoning decision model to obtain risk intervention timing decision data; Step S45: supplementing the recommended measures data in the intelligent disease risk warning data with the risk intervention opportunity decision data to generate supplementary disease risk warning data; Step S46: Utilize the preset medical staff information receiving authority matrix to perform multi-level linkage push processing on the supplementary disease risk warning data, and perform warning information feedback statistics to generate warning information feedback data.

8. The information digital management system based on the critical care medicine department according to claim 1 is characterized in that: The individualized diagnosis and treatment decision-making support module is used to perform the following steps: Step S51: Analyze the disease risk treatment plan according to the early warning information feedback data to generate disease risk treatment plan data; Step S52: performing a medical resource demand assessment on the disease risk treatment plan data, and generating a medical resource allocation plan to obtain a medical resource allocation plan; Step S53: Calculate the patient waiting time according to the medical resource allocation plan to obtain patient waiting time data; Step S54: Perform resource allocation feasibility analysis on the medical resource allocation plan based on the patient waiting time data, and make optimization adjustments to obtain an optimized medical resource allocation plan; Step S55: Process the medical resource dispatching instructions according to the optimized medical resource allocation plan, and send the medical resource dispatching instructions to relevant medical staff and medical equipment, so as to obtain an individualized diagnosis and treatment auxiliary decision-making strategy.

Images