A disease condition stage identification method and device based on disease condition evolution

Abstract

The present application relates to the field of intelligent medical auxiliary technology, and more particularly to a disease stage recognition method and device based on special disease condition evolution, which comprises the following steps: collecting historical clinical data of patients with acute gastrointestinal hemorrhage; based on the historical clinical data, constructing a set of historical clinical state nodes; based on the set of historical clinical state nodes, constructing a causal chain structure of the special disease of acute gastrointestinal hemorrhage; obtaining target clinical data that changes over time; based on the causal chain structure of the special disease of acute gastrointestinal hemorrhage and the target clinical data, constructing an individualized disease state graph; based on the individualized disease state graph, using the stage competition method to recognize the comprehensive disease stage, and by means of the constructed causal chain structure of the special disease of acute gastrointestinal hemorrhage, the target clinical data of the individual is judged to accurately recognize the disease stage of the patient, so as to support subsequent prediction, recognition, decision-making and feedback optimization.

Description

Technical Field

[0001] This invention relates to the field of intelligent medical assistance technology, and in particular to a method and device for identifying disease stages based on the evolution of a specific disease. Background Technology

[0002] Acute gastrointestinal bleeding (AGIB) is a common critical condition in emergency medicine and the management of digestive system diseases. It is characterized by rapid onset and deterioration, with patients potentially developing circulatory instability, hemodynamic abnormalities, coagulation disorders, and multiple organ complications within a short period. Clinically, these patients typically require comprehensive management including hemostasis, fluid resuscitation, blood transfusions, acid suppression, and interventional or endoscopic treatments. In actual clinical practice, some patients may further develop secondary infections due to persistent bleeding, repeated blood transfusions, increased invasive procedures, heightened stress, and complex underlying diseases. In severe cases, this can progress to sepsis or multiple organ dysfunction.

[0003] From a clinical perspective, the progression of acute gastrointestinal bleeding is not driven by a single indicator, but rather exhibits distinct staged evolutionary characteristics and event chains. For example, the patient initially presents with a clear or unclear source of bleeding, increased bleeding volume, decreased hemoglobin, and rebleeding. Subsequently, the patient may enter a coagulation imbalance stage, characterized by prolonged prothrombin time, elevated international normalized ratio, decreased platelet count, or the effects of anticoagulant drug exposure. Further, with the combined effects of blood transfusions, catheterization, mechanical ventilation, and intensive care, the risk of infection gradually increases, leading to abnormal body temperature, elevated white blood cell count, elevated inflammatory factors, and positive cultures. This process is essentially a continuous clinical evolution chain consisting of "bleeding events—coagulation changes—infection evolution."

[0004] However, most existing medical information systems primarily rely on discrete field storage and single-point-of-time display. Typically, bleeding events, laboratory indicators, treatment measures, and infection manifestations are recorded separately in different business modules. For example, endoscopic hemostasis information is recorded in the examination system; coagulation indicators are in the laboratory system; transfusion records are in the transfusion management system; body temperature and vital signs data are recorded in the nursing system; and infection diagnosis and medication information are scattered across progress notes and medical orders. This fragmented data distribution, inconsistent semantics, and weakened temporal relationships mean that while clinicians can rely on experience to comprehensively assess a patient's condition from multiple systems, human judgment is time-consuming and inaccurate.

[0005] Therefore, improving the accuracy of patient condition stage identification to support subsequent prediction, identification, decision-making, and feedback optimization is a pressing technical problem that needs to be solved. Summary of the Invention

[0006] In view of the above problems, the present invention provides a method and apparatus for identifying disease stages based on the evolution of a specific disease to overcome or at least partially solve the above problems.

[0007] In a first aspect, the present invention provides a method for identifying disease stages based on the evolution of a specific disease, including: Historical clinical data of patients with a history of acute gastrointestinal bleeding were collected. The historical clinical data included: basic information data, bleeding-related data, coagulation-related data, infection-related indicator data, and vital signs time series. Based on the historical clinical data, a set of historical clinical status nodes is constructed, which includes: node type, node status value, node activation strength, and node timestamp. Based on the aforementioned set of historical clinical status nodes, a causal chain structure for the acute gastrointestinal bleeding disease is constructed. Obtain target clinical data for individuals that change continuously over time; Based on the causal chain structure and target clinical data of acute gastrointestinal bleeding, an individualized disease status map is constructed. The individualized disease status map is used to represent the current disease status, stage transition relationship and potential evolution direction of an individual. Based on the individualized disease status map, the stage competition method is used to identify the comprehensive disease stage.

[0008] Preferably, based on the set of historical clinical state nodes, a causal chain structure for the specific disease of acute gastrointestinal bleeding is constructed, including: Based on the set of historical clinical state nodes, the predecessor and successor nodes in the causal relationship are determined to obtain the causal relationship set. Based on the predecessor node and the successor node, calculate the first probability of the successor node occurring under the condition that the predecessor node occurs, and the basic probability of the successor node occurring. Based on the first probability and the basic occurrence probability, calculate the causal contribution of the predecessor node to the data-driven successor node. Based on the causal contribution, the causal strength is calculated to obtain the edge weight set; Based on the set of historical clinical state nodes, the set of causal relationships, and the set of edge weights, a causal chain structure for acute gastrointestinal bleeding is constructed.

[0009] Preferably, based on the causal chain structure and target clinical data of acute gastrointestinal bleeding, an individualized disease status map is constructed, including: In the causal propagation of the causal chain structure of acute gastrointestinal bleeding, the activation status of each target node state value in the target clinical data is evaluated. Based on the activation state and the influence of the upstream nodes of each target node, the intermediate state of each target node is determined. Based on the intermediate and active states of each target node, the final state of each target node is obtained through time decay and confidence evaluation. Based on the final state of each target node, an individualized disease status map is constructed.

[0010] Preferably, based on the intermediate and active states of each target node, the final state of each target node is obtained through time decay and confidence evaluation, including: Based on the intermediate states of each target node, time decay is performed according to the following formula to obtain the decayed state of each target node: in, The attenuation coefficient is... The initial time when the state is generated. For the current time, The initial state is generated based on the states determined by the intermediate states of each target node. This represents the decayed state of each target node; Obtain the total number of relevant information and the number of valid information for each target node; Based on the total amount of information and the amount of valid information related to each target node, the confidence assessment result of each target node is obtained; Based on the decayed state of each target node, the confidence evaluation result of each target node, and the activation state, the final state of each target node is obtained.

[0011] Preferably, based on an individualized disease status map, a stage competition method is used to identify the comprehensive disease stages, including: Based on the individualized disease status map, each target node is divided into stages to obtain a set of target nodes for each stage; Calculate the state strength of each stage based on the set of target nodes for each stage. Calculate the competition probability at each stage based on the state strength at each stage; Based on the state strength of each stage, the dynamic threshold of each stage is calculated. Based on the competition probability, state intensity, and dynamic threshold of each stage, the comprehensive disease stage is identified.

[0012] Preferably, based on the node set of each stage, the state strength of each stage is calculated, specifically according to the following formula: in, For any stage, the set of target nodes This represents the decayed state of each target node in the target node set for this stage. This represents the activation state of each target node in the target node set for this stage. This represents the confidence assessment results for each target node in the target node set for this stage. The first weight coefficient corresponding to the final state. This is the second weighting coefficient corresponding to the active state. This is the third weighting coefficient corresponding to the confidence assessment result. This represents the state intensity at this stage.

[0013] Preferably, the competition probability of each stage is calculated based on the state strength of each stage, specifically according to the following formula: in, Let be the probability of competition at any stage. For the state strength of each stage in all stages, For circular index, The state strength of any stage in all stages.

[0014] Preferably, the dynamic threshold for each stage is calculated based on the state strength of each stage, including: Based on the state intensity at each stage, the changing trend at each stage is determined using the following formula: in, The state strength of the previous moment in this phase. The state strength at the current moment of this phase. This represents the trend of change during this stage; Based on the changing trends at each stage, the dynamic threshold for each stage is calculated using the following formula: in, Based on the threshold, This is the trend influence coefficient. , which is the dynamic threshold for any stage.

[0015] Preferably, based on the competition probability of each stage, the state intensity of each stage, and the dynamic threshold of each stage, the comprehensive disease stage is identified, including: Determine whether the state strength corresponding to the maximum value of the competition probability in each stage is greater than the dynamic threshold of the corresponding stage. If so, then the corresponding stage is determined by the overall stage of the illness; If not, then the patient is in a transitional stage based on the overall condition.

[0016] Secondly, the present invention also provides a disease stage identification device based on the evolution of a specific disease, comprising: The data acquisition module is used to collect historical clinical data of patients with a history of acute gastrointestinal bleeding. The historical clinical data includes: basic information data, bleeding-related data, coagulation-related data, infection-related indicator data, and vital signs time series. The first construction module is used to construct a set of historical clinical status nodes based on the historical clinical data. The set of historical clinical status nodes includes: node type, node status value, node activation intensity, and node timestamp. The second construction module is used to construct a causal chain structure for acute gastrointestinal bleeding based on the set of historical clinical status nodes. The acquisition module is used to acquire target clinical data of individuals that change continuously over time; The third construction module is used to construct an individualized disease status map based on the causal chain structure and target clinical data of acute gastrointestinal bleeding. The individualized disease status map is used to represent the current disease status, stage transition relationship and potential evolution direction of an individual. The identification module is used to identify the overall disease stage based on the individualized disease status map and using a stage competition method.

[0017] Thirdly, the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described in the first aspect.

[0018] Fourthly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0019] One or more technical solutions in the embodiments of the present invention have at least the following technical effects or advantages: This invention provides a method for identifying disease stages based on the evolution of a specific disease, comprising: collecting historical clinical data of patients with acute gastrointestinal bleeding; constructing a set of historical clinical state nodes based on the historical clinical data; constructing a causal chain structure for acute gastrointestinal bleeding based on the set of historical clinical state nodes; acquiring target clinical data that changes continuously over time; constructing an individualized disease state map based on the causal chain structure for acute gastrointestinal bleeding and the target clinical data; identifying the comprehensive disease stage based on the individualized disease state map using a stage competition method; and accurately identifying the patient's disease stage by evaluating the individual's target clinical data using the constructed causal chain structure for acute gastrointestinal bleeding, thereby supporting subsequent prediction, identification, decision-making, and feedback optimization. Attached Figure Description

[0020] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A flowchart illustrating the disease stage identification method based on the evolution of a specific disease is shown in an embodiment of the present invention. Figure 2 This invention illustrates a schematic diagram of a disease stage identification device based on the evolution of a specific disease in an embodiment of the present invention. Figure 3 A schematic diagram of the structure of a computer device for implementing a disease stage identification method based on the evolution of a specific disease is shown in an embodiment of the present invention. Detailed Implementation

[0021] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the invention and to fully convey the scope of the invention to those skilled in the art.

[0022] Example 1

[0023] Embodiments of the present invention provide a method for identifying disease stages based on the evolution of a specific disease, such as... Figure 1 As shown, it includes: S101, Collect historical clinical data of patients with a history of acute gastrointestinal bleeding. This historical clinical data includes: basic information data, bleeding-related data, coagulation-related data, infection-related indicator data, and vital signs time series. S102, Based on historical clinical data, construct a set of historical clinical status nodes, which includes: node type, node status, node activation intensity, and node timestamp; S103, based on the set of historical clinical status nodes, constructs a causal chain structure for the specific disease of acute gastrointestinal bleeding; S104, Obtain target clinical data for individuals that change continuously over time; S105. Based on the causal chain structure and target clinical data of acute gastrointestinal bleeding, an individualized disease status map is constructed. This individualized disease status map is used to represent the current disease status, stage transition relationship and potential evolution direction of an individual. S106, based on an individualized disease status map, uses a stage competition method to identify the comprehensive disease stage.

[0024] This invention is not limited to determining the stage of a patient's disease by a single indicator, but rather transforms the complex clinical process into a computable, traceable, and interpretable disease state expression model, thereby providing support for subsequent risk propagation analysis, threshold identification, intervention strategy generation, and feedback updates.

[0025] First, execute S101 to collect historical clinical data of patients with a history of acute gastrointestinal bleeding. This historical clinical data includes: basic information data, bleeding-related data, coagulation-related data, infection-related indicator data, and vital signs time series.

[0026] The diagnosis and treatment of patients with acute gastrointestinal bleeding generate a large amount of historical clinical data from various systems, such as laboratory indicators, vital signs, transfusion records, endoscopic procedure records, and medical records. These historical clinical data come from different sources and have significantly different formats; therefore, data cleaning and standardization are necessary.

[0027] These historical clinical data include: Basic information data: age, gender, medical history, and history of anticoagulant medication use; Bleeding-related data: source of bleeding, bleeding volume classification, rebleeding events, endoscopic hemostasis procedures; Coagulation-related data: INR (International Normalized Ratio), PT (Prothrombin Time), APTT (Activated Partial Thromboplastin Time), PLT (Platelet Count); Infection-related data: body temperature, white blood cell count, CRP (C-reactive protein), PCT (procalcitonin); Vital signs time series: heart rate, blood pressure, respiratory rate, and blood oxygen saturation.

[0028] After obtaining this data, the data is cleaned by deleting duplicate records, obviously erroneous data, and unreasonable physiological values.

[0029] Next, the units of data from different sources need to be standardized. For example, CRP uses the unit mg / L, and hemoglobin uses the unit g / L. The units of the two need to be standardized.

[0030] Time alignment can be achieved by organizing data in time windows, such as summarizing patient status data within the same time window (4 hours).

[0031] Next, execute S102 to construct a set of historical clinical status nodes based on historical clinical data. This set of historical clinical status nodes includes: node type, node status value, node activation strength, and node timestamp.

[0032] First, historical clinical data is divided into three categories of status nodes: bleeding status nodes, coagulation status nodes, and infection-related nodes.

[0033] For example, bleeding status nodes include: bleeding source (upper gastrointestinal tract / lower gastrointestinal tract / unknown), bleeding volume classification, rebleeding events, and endoscopic hemostasis-related events.

[0034] Coagulation status indicators: abnormal INR, prolonged PT, abnormal APTT, and decreased PLT.

[0035] Infection-related symptoms: fever, elevated WBC (white blood cell count), and elevated CRP.

[0036] Each type of state node contains the following attributes: node type (bleeding / clotting / infection), node state value, node activation strength, and node timestamp.

[0037] For each node type, the following annotation method exists: in, For any node type, the node state value ( A value of 0 indicates inactive. A score of 1 indicates a mild abnormality. A score of 2 indicates a significant abnormality. For any node, the corresponding index value. The threshold value for dividing node state values. The upper limit threshold for dividing node state values.

[0038] The state value of this node is corrected by introducing a time rate of change: in, The rate of change of the indicator value over time. In order to be in The index value at that moment, In order to be in The index value at that moment, For time window, The threshold for abnormal changes in indicator values. Increase the weight of trends. This is an indicator function (1 if true, 0 otherwise). This refers to the corrected node state value for any node type.

[0039] Since the same node comes from multiple data sources (validation, text, operation records), a fusion function is constructed: in, For the first The activation strength of each node. For the first Data sources This is a mapping function, which can be implemented using NLP recognition or rule-based mapping, thereby mapping to a computable single activation strength. Weighting based on data source The number of data sources.

[0040] For node types, another mapping function can be used to determine whether it is a bleeding, coagulation, or infection type.

[0041] Node timestamps can be marked based on the time of occurrence, resulting in... .

[0042] Therefore, any state node in the historical clinical state node set is constructed as follows: .in, This refers to the node type. The historical clinical status node set is... .

[0043] Next, execute S103 to construct a causal chain structure for acute gastrointestinal bleeding based on the set of historical clinical status nodes.

[0044] Specifically, based on the set of historical clinical state nodes, the predecessor and successor nodes in the causal relationship are determined to obtain the causal relationship set; Based on the predecessor node and the successor node, calculate the first probability of the successor node occurring and the basic probability of the successor node occurring under the condition that the predecessor node occurs. Based on the first probability and the basic occurrence probability, calculate the causal contribution of the predecessor node to the successor node in data-driven operations. Based on the causal contribution, the causal strength is calculated to obtain the set of edge weights; Based on the set of historical clinical state nodes, the set of causal relationships, and the set of edge weights, a causal chain structure for acute gastrointestinal bleeding is constructed.

[0045] When determining the precursor and successor nodes in a causal relationship, the specific determination is based on the definitions in the medical knowledge base. For example, long-term oral administration of nonsteroidal anti-inflammatory drugs (precursor node) leads to gastric mucosal erosion and damage (successor node); portal hypertension (precursor node) leads to esophageal and gastric variceal rupture and bleeding (successor node), and so on.

[0046] This set of causal relationships is represented by the following formula: For a set of causal relationships, As a predecessor node, It is the successor node.

[0047] Next, calculate the first probability of the successor node occurring given that the predecessor node has occurred: in, Predecessor node and successor nodes The number of times they co-occur. Predecessor node The number of times it appears, For the predecessor node Under the conditions of occurrence, successor nodes The first probability of occurrence.

[0048] The basic occurrence probability of this successor node Specifically, this successor node exists in the entire population. The natural probability of occurrence, which is the overall epidemiological background probability.

[0049] The causal contribution of the predecessor node to the successor node in data-driven processes is as follows: in, This causal contribution is specifically the degree of influence after removing background probabilities. Using causal contribution can quantify the true causal influence between nodes, thereby avoiding misjudgments caused by relying solely on co-occurrence.

[0050] To ensure that the causal relationship conforms to the chronological logic, false causal relationships are filtered out. For example, the time of occurrence of the successor node must be after the time of occurrence of the predecessor node, and the time difference between the occurrence of the successor node and the occurrence of the predecessor node event cannot exceed the maximum allowed time interval.

[0051] Next, based on this causal contribution, the causal strength is calculated to obtain the set of edge weights, as shown in the following formula: in, The weight of the causal edge corresponding to the causal strength. The edge weights are set based on medical rules. For causal contribution (data-driven). This refers to the fourth weight coefficient corresponding to the edge weights set based on medical rules. This is the fifth weighting coefficient corresponding to the causal contribution.

[0052] Based on the causal strength of all causal relationships, the set of edge weights is obtained.

[0053] Finally, based on the set of historical clinical state nodes, the set of causal relationships, and the set of edge weights, a causal chain structure for acute gastrointestinal bleeding is constructed. This yields a directed graph of the causal chain structure: in, Represents the set of historical clinical state nodes. Represents a set of causal relationships. This represents the set of edge weights.

[0054] Next, perform S104 to obtain the target clinical data for individuals that change continuously over time.

[0055] It is possible to collect targeted clinical data from individual patients over a period of time.

[0056] Then, S105 is executed to construct an individualized disease status map based on the causal chain structure of acute gastrointestinal bleeding and the target clinical data. The individualized disease status map is used to represent the individual's current disease status, stage transition relationships, and potential evolutionary direction.

[0057] The following is a detailed description of how to construct an individualized disease status map: In the causal propagation of the causal chain structure of acute gastrointestinal bleeding, the activation status of each target node state value in the target clinical data is evaluated. Based on the activation state and the influence of the upstream nodes of each target node, the intermediate state of each target node is determined. Based on the intermediate and active states of each target node, the final state of the target node is obtained through time decay and confidence evaluation. Based on the final state of each target node, an individualized disease status map is constructed.

[0058] The activation status of each target node in the target clinical data is evaluated. Specifically, it is determined whether the target node triggers the occurrence of a certain type of event, such as rebleeding, INR increase, or CPR increase. Specifically, when a rebleeding event occurs, the bleeding node is identified as activated; when INR increases, the coagulation abnormality node is identified as activated; and when CRP increases, the infection node is identified as activated.

[0059] This activation state The values ​​are represented by 0 and 1. A 1 indicates the target node is active, and a 0 indicates it is inactive. The evaluation process compares the actual state value of the target node with a corresponding threshold. If the value exceeds the threshold, the node is considered active; otherwise, it is considered inactive.

[0060] Next, considering the influence of upstream nodes, the intermediate states of each target node are calculated: This is an intermediate state. This refers to the target node's own state. The set of all upstream nodes that point to the target node. The weights of all causal edges pointing to the target node. This indicates the active state of the upstream node.

[0061] This approach can demonstrate the transmission of influence between target nodes, giving the disease state an evolutionary driving force, which is different from static rules.

[0062] Then, based on the intermediate and active states of each target node, the final state of each target node is obtained through time decay and confidence evaluation.

[0063] Specifically, this includes: based on the intermediate states of each target node, performing time decay according to the following calculation formula to obtain the decayed state of each target node:

[0064] in, The attenuation coefficient is... The initial time when the state is generated. For the current time, The initial state is generated based on the states determined by the intermediate states of each target node. This represents the decayed state of each target node; Obtain the total number of relevant information and the number of valid information for each target node; Based on the total amount of information and the amount of valid information related to each target node, the confidence assessment results of each target node are obtained; Based on the decayed state of each target node, the confidence evaluation result of each target node, and the activation state, the final state of each target node is obtained. For example, the decay coefficient is 0.035 by default.

[0065] The confidence assessment results for each target node are as follows: The confidence assessment results for each target node. The number of valid information related to each target node. This represents the total amount of information related to each target node.

[0066] Before assessing the confidence level, the information related to the target node is evaluated from multiple aspects, including the legality of the source, the timeliness and validity, the standardization of the format and content, and the legality of the business logic. This process yields the number of valid information items for each target node.

[0067] Therefore, based on the decayed state of each target node, the confidence evaluation result of each target node, and the activation state, the final state of each target node is obtained. .

[0068] Based on the final state of each target node, an individualized disease status map is obtained. .

[0069] Next, S106 is executed, and based on the individualized disease status map, the stage competition method is used to identify the comprehensive disease stage.

[0070] Specifically, based on the individualized disease status map, each target node is divided into stages to obtain a set of target nodes for each stage; Calculate the state strength of each stage based on the set of target nodes for each stage. Calculate the competition probability at each stage based on the state strength at each stage; Based on the state strength of each stage, the dynamic threshold of each stage is calculated. Based on the competition probability, state intensity, and dynamic threshold of each stage, the comprehensive disease stage is identified.

[0071] Each target node, according to its original classification, corresponds to a specific stage, resulting in a set of target nodes for three stages. ,in, For the set of bleeding nodes, This is a set of nodes in the coagulation imbalance stage. This is the set of nodes during the infection phase.

[0072] The state strength at each stage is calculated below: in, For any stage, the set of target nodes This represents the decayed state of each target node in the target node set for this stage. This represents the activation state of each target node in the target node set for this stage. This represents the confidence assessment results for each target node in the target node set for this stage. The first weight coefficient corresponding to the final state. This is the second weighting coefficient corresponding to the active state. This is the third weighting coefficient corresponding to the confidence assessment result. This represents the state intensity at this stage. The default value is 0.55. The default value is 0.25. The default value is 0.20.

[0073] Node-level information is aggregated into stage-level expressions, forming the basic input for stage competition.

[0074] The competition probability at each stage is calculated below, using the following formula: in, Let be the probability of competition at any stage. For the state strength of each stage in all stages, For circular index, The state strength of any stage in all stages.

[0075] By calculating the competition probability at each stage, the competitive relationship between stages is modeled, and the output is a continuous probability rather than a discrete label, thus ensuring the stability of the algorithm.

[0076] Next, based on the state strength of each stage, the dynamic threshold of each stage is calculated, specifically including: Based on the state intensity at each stage, the changing trend at each stage is determined using the following formula: in, The state strength of the previous moment in this phase. The state strength at the current moment of this phase. This represents the trend of change during this stage; Based on the changing trends at each stage, the dynamic threshold for each stage is calculated using the following formula: in, Based on the threshold, This is the trend influence coefficient. This is the dynamic threshold for any stage. The baseline threshold is a static benchmark threshold determined without overlaying disease trend changes. It serves as a baseline cutoff value for this specific disease population and is unaffected by real-time score changes. It can be obtained by collecting data from a large number of historical patients at different symptom levels. The trend influence coefficient defaults to 0.12.

[0077] Finally, based on the competition probability, state intensity, and dynamic threshold of each stage, the comprehensive disease stage is identified. The specific identification process is as follows: Determine whether the state strength corresponding to the maximum value of the competition probability in each stage is greater than the dynamic threshold of the corresponding stage. If so, then the corresponding stage is determined by the overall stage of the illness; If not, then the overall condition is in a transitional stage.

[0078] The judgment process is represented by the following formula: in, For the comprehensive stage of the disease, It represents the maximum value of the competition probability in any stage, obtained as it changes over time. To and correspond .

[0079] This effectively identifies the boundary states between stages. Simultaneously, it can output the contribution source for that corresponding stage.

[0080] Of course, to improve the accuracy of identification, the difference between the identified comprehensive disease stage and the actual disease stage can be calculated to obtain the error.

[0081] Next, based on this error, the weights of the causal edges in the individualized disease status graph are adjusted. The specific adjustment formula is as follows: in, This indicates the active state of the upstream node. For this error, For learning rate, The adjusted causal edge weights, The weights of the causal edges before adjustment are shown. This approach allows for the gradual optimization of causal relationships.

[0082] In addition, to ensure a more optimized causal chain structure, the weights of causal edges with low long-term contributions can be removed by setting a minimum weight threshold. Based on the following judgment: This approach can reduce the complexity of individualized disease status maps and improve computational efficiency.

[0083] Finally, this transforms the overall system from a static recognition scheme into a dynamic evolutionary scheme with continuous learning capabilities.

[0084] One or more technical solutions in the embodiments of the present invention have at least the following technical effects or advantages: This invention provides a method for identifying disease stages based on the evolution of a specific disease, comprising: collecting historical clinical data of patients with acute gastrointestinal bleeding; constructing a set of historical clinical state nodes based on the historical clinical data; constructing a causal chain structure for acute gastrointestinal bleeding based on the set of historical clinical state nodes; acquiring target clinical data that changes continuously over time; constructing an individualized disease state map based on the causal chain structure for acute gastrointestinal bleeding and the target clinical data; identifying the comprehensive disease stage based on the individualized disease state map using a stage competition method; and accurately identifying the patient's disease stage by evaluating the individual's target clinical data using the constructed causal chain structure for acute gastrointestinal bleeding, thereby supporting subsequent prediction, identification, decision-making, and feedback optimization.

[0085] Example 2 Based on the same inventive concept, embodiments of the present invention provide a disease stage identification device for the evolution of a specific disease, such as... Figure 2 As shown, it includes: The data acquisition module 201 is used to collect historical clinical data of patients with a history of acute gastrointestinal bleeding. The historical clinical data includes: basic information data, bleeding-related data, coagulation-related data, infection-related indicator data, and vital signs time series. The first construction module 202 is used to construct a set of historical clinical status nodes based on the historical clinical data. The set of historical clinical status nodes includes: node type, node status value, node activation intensity, and node timestamp. The second construction module 203 is used to construct a causal chain structure for acute gastrointestinal bleeding based on the set of historical clinical state nodes. Acquisition module 204 is used to acquire target clinical data of individuals that change continuously over time; The third construction module 205 is used to construct an individualized disease status map based on the causal chain structure of acute gastrointestinal bleeding and target clinical data. The individualized disease status map is used to represent the current disease status, stage transition relationship and potential evolution direction of an individual. The identification module 206 is used to identify the comprehensive disease stage based on the individualized disease status map and using the stage competition method.

[0086] In one alternative implementation, the second building module 203 is configured to: Based on the set of historical clinical state nodes, the predecessor and successor nodes in the causal relationship are determined to obtain the causal relationship set. Based on the predecessor node and the successor node, calculate the first probability of the successor node occurring under the condition that the predecessor node occurs, and the basic probability of the successor node occurring. Based on the first probability and the basic occurrence probability, calculate the causal contribution of the predecessor node to the data-driven successor node. Based on the causal contribution, the causal strength is calculated to obtain the edge weight set; Based on the set of historical clinical state nodes, the set of causal relationships, and the set of edge weights, a causal chain structure for acute gastrointestinal bleeding is constructed.

[0087] In one alternative implementation, the third building module 205 is configured to: In the causal propagation of the causal chain structure of acute gastrointestinal bleeding, the activation status of each target node state value in the target clinical data is evaluated. Based on the activation state and the influence of the upstream nodes of each target node, the intermediate state of each target node is determined. Based on the intermediate and active states of each target node, the final state of each target node is obtained through time decay and confidence evaluation. Based on the final state of each target node, an individualized disease status map is constructed.

[0088] In one alternative implementation, the third building module 205 is configured to: Based on the intermediate states of each target node, time decay is performed according to the following formula to obtain the decayed state of each target node: in, The attenuation coefficient is... The initial time when the state is generated. For the current time, The initial state is generated based on the states determined by the intermediate states of each target node. This represents the decayed state of each target node; Obtain the total number of relevant information and the number of valid information for each target node; Based on the total amount of information and the amount of valid information related to each target node, the confidence assessment result of each target node is obtained; Based on the decayed state of each target node, the confidence evaluation result of each target node, and the activation state, the final state of each target node is obtained.

[0089] In one alternative implementation, the identification module 206 is used for: Based on the individualized disease status map, each target node is divided into stages to obtain a set of target nodes for each stage; Calculate the state strength of each stage based on the set of target nodes for each stage. Calculate the competition probability at each stage based on the state strength at each stage; Based on the state strength of each stage, the dynamic threshold of each stage is calculated. Based on the competition probability, state intensity, and dynamic threshold of each stage, the comprehensive disease stage is identified.

[0090] In one optional implementation, the identification module 206 is specifically used for: The state strength is calculated using the following formula: in, For any stage, the set of target nodes This represents the decayed state of each target node in the target node set for this stage. This represents the activation state of each target node in the target node set for this stage. This represents the confidence assessment results for each target node in the target node set for this stage. The first weight coefficient corresponding to the final state. This is the second weighting coefficient corresponding to the active state. This is the third weighting coefficient corresponding to the confidence assessment result. This represents the state intensity at this stage.

[0091] In one optional implementation, the identification module 206 is specifically used for: The competition probability at each stage is calculated using the following formula: in, Let be the probability of competition at any stage. For the state strength of each stage in all stages, For circular index, The state strength of any stage in all stages.

[0092] In one optional implementation, the identification module 206 is specifically used for: Based on the state intensity at each stage, the changing trend at each stage is determined using the following formula: in, The state strength at the current moment of this phase. The state strength of the previous moment in this phase. This represents the trend of change during this stage; Based on the changing trends at each stage, the dynamic threshold for each stage is calculated using the following formula: in, Based on the threshold, This is the trend influence coefficient. , which is the dynamic threshold for any stage.

[0093] In one optional implementation, the identification module 206 is specifically used for: Determine whether the state strength corresponding to the maximum value of the competition probability in each stage is greater than the dynamic threshold of the corresponding stage. If so, then the corresponding stage is determined by the overall stage of the illness; If not, then the patient is in a transitional stage based on the overall condition.

[0094] Example 3: Based on the same inventive concept, embodiments of the present invention provide a computer device, such as... Figure 3 As shown, it includes a memory 304, a processor 302, and a computer program stored in the memory 304 and executable on the processor 302. When the processor 302 executes the program, it implements the steps of the above-mentioned disease stage identification method based on the evolution of specific disease conditions.

[0095] Among them, Figure 3 In this document, a bus architecture (represented by bus 300) is used. Bus 300 may include any number of interconnected buses and bridges, linking various circuits including one or more processors represented by processor 302 and memory represented by memory 304. Bus 300 may also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. Bus interface 306 provides an interface between bus 300 and receiver 301 and transmitter 303. Receiver 301 and transmitter 303 may be the same element, i.e., a transceiver, providing a unit for communicating with various other devices over a transmission medium. Processor 302 is responsible for managing bus 300 and general processing, while memory 304 can be used to store data used by processor 302 during operation.

[0096] Example 4: Based on the same inventive concept, embodiments of the present invention provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method for identifying disease stages based on the evolution of specific disease conditions.

[0097] The algorithms and displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings herein. The required structure for constructing such systems is apparent from the above description. Furthermore, this invention is not directed to any particular programming language. It should be understood that the contents of the invention described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing the best mode of implementation of the invention.

[0098] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0099] Similarly, it should be understood that, in order to simplify the invention and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof. However, this disclosure should not be construed as reflecting an intention that the claimed invention requires more features than are explicitly recited in each embodiment. Rather, as reflected in each embodiment, inventive aspects lie in fewer than all features of the single foregoing disclosed embodiment. Therefore, the claims, following the detailed description, are hereby expressly incorporated into this detailed description, wherein each claim itself is a separate embodiment of the invention.

[0100] Those skilled in the art will understand that modules in the device of the embodiments can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiments can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all processes or units of any method or device so disclosed. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0101] Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are meant to be within the scope of the invention and form different embodiments. For example, in the specific implementation, any of the claimed embodiments can be used in any combination.

[0102] The various component embodiments of the present invention can be implemented in hardware, or as software modules running on one or more processors, or a combination thereof. Those skilled in the art will understand that microprocessors or digital signal processors (DSPs) can be used in practice to implement some or all of the functions of some or all of the components in the disease stage identification device or computer device based on disease evolution according to embodiments of the present invention. The present invention can also be implemented as a device or apparatus program (e.g., a computer program and computer program product) for performing part or all of the methods described herein. Such programs implementing the present invention can be stored on a computer-readable medium or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0103] It should be noted that the above embodiments are illustrative of the invention and not restrictive, and that those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names.

Claims

1. A method for identifying disease stages based on the evolution of a specific disease, characterized in that, include: Historical clinical data of patients with a history of acute gastrointestinal bleeding were collected. The historical clinical data included: basic information data, bleeding-related data, coagulation-related data, infection-related indicator data, and vital signs time series. Based on the historical clinical data, a set of historical clinical status nodes is constructed, which includes: node type, node status value, node activation strength, and node timestamp. Based on the aforementioned set of historical clinical status nodes, a causal chain structure for the acute gastrointestinal bleeding disease is constructed. Obtain target clinical data for individuals that change continuously over time; Based on the causal chain structure and target clinical data of acute gastrointestinal bleeding, an individualized disease status map is constructed. The individualized disease status map is used to represent the current disease status, stage transition relationship and potential evolution direction of an individual. Based on the individualized disease status map, the stage competition method is used to identify the comprehensive disease stage.

2. The method as described in claim 1, characterized in that, Based on the aforementioned set of historical clinical status nodes, a causal chain structure for the specific disease of acute gastrointestinal bleeding is constructed, including: Based on the set of historical clinical state nodes, the predecessor and successor nodes in the causal relationship are determined to obtain the causal relationship set. Based on the predecessor node and the successor node, calculate the first probability of the successor node occurring under the condition that the predecessor node occurs, and the basic probability of the successor node occurring. Based on the first probability and the basic occurrence probability, calculate the causal contribution of the predecessor node to the data-driven successor node. Based on the causal contribution, the causal strength is calculated to obtain the edge weight set; Based on the set of historical clinical state nodes, the set of causal relationships, and the set of edge weights, a causal chain structure for acute gastrointestinal bleeding is constructed.

3. The method as described in claim 1, characterized in that, Based on the causal chain structure and target clinical data of acute gastrointestinal bleeding, an individualized disease status map is constructed, including: In the causal propagation of the causal chain structure of acute gastrointestinal bleeding, the activation status of each target node state value in the target clinical data is evaluated. Based on the activation state and the influence of the upstream nodes of each target node, the intermediate state of each target node is determined. Based on the intermediate and active states of each target node, the final state of each target node is obtained through time decay and confidence evaluation. Based on the final state of each target node, an individualized disease status map is constructed.

4. The method as described in claim 3, characterized in that, Based on the intermediate and active states of each target node, the final state of each target node is obtained through time decay and confidence evaluation, including: Based on the intermediate states of each target node, time decay is performed according to the following formula to obtain the decayed state of each target node: in, The attenuation coefficient is... The initial time at which the state is generated. For the current time, The initial state is generated based on the states determined by the intermediate states of each target node. This represents the decayed state of each target node; Obtain the total number of relevant information and the number of valid information for each target node; Based on the total amount of information and the amount of valid information related to each target node, the confidence assessment result of each target node is obtained; Based on the decayed state of each target node, the confidence evaluation result of each target node, and the activation state, the final state of each target node is obtained.

5. The method as described in claim 1, characterized in that, Based on an individualized disease status map, a stage competition method was used to identify the comprehensive disease stages, including: Based on the individualized disease status map, each target node is divided into stages to obtain a set of target nodes for each stage; Calculate the state strength of each stage based on the set of target nodes for each stage. Calculate the competition probability at each stage based on the state strength at each stage; Based on the state strength of each stage, the dynamic threshold of each stage is calculated. Based on the competition probability, state intensity, and dynamic threshold of each stage, the comprehensive disease stage is identified.

6. The method as described in claim 5, characterized in that, Based on the node set of each stage, the state strength of each stage is calculated according to the following formula: in, For any stage, the set of target nodes This represents the decayed state of each target node in the target node set for this stage. This represents the activation state of each target node in the target node set for this stage. This represents the confidence assessment results for each target node in the target node set for this stage. The first weight coefficient corresponding to the final state. The second weighting coefficient corresponds to the active state. This is the third weighting coefficient corresponding to the confidence assessment result. This represents the state intensity at this stage.

7. The method as described in claim 5, characterized in that, Based on the state strength at each stage, the competition probability at each stage is calculated using the following formula: in, Let be the probability of competition at any stage. For the state strength of each stage in all stages, For circular index, The state strength of any stage in all stages.

8. The method as described in claim 5, characterized in that, Based on the state strength at each stage, the dynamic threshold for each stage is calculated, including: Based on the state intensity at each stage, the changing trend at each stage is determined using the following formula: in, The state strength of the previous moment in this phase. The state strength at the current moment of this phase. This represents the trend of change during this stage; Based on the changing trends at each stage, the dynamic threshold for each stage is calculated using the following formula: in, Based on the threshold, The trend influence coefficient. , which is the dynamic threshold for any stage.

9. The method as described in claim 5, characterized in that, Based on the competition probability, state intensity, and dynamic threshold of each stage, the comprehensive disease stages are identified, including: Determine whether the state strength corresponding to the maximum value of the competition probability in each stage is greater than the dynamic threshold of the corresponding stage. If so, then the corresponding stage is determined by the overall stage of the illness; If not, then the patient is in a transitional stage based on the overall condition.

10. A disease stage identification device based on the evolution of a specific disease, characterized in that, include: The data acquisition module is used to collect historical clinical data of patients with a history of acute gastrointestinal bleeding. The historical clinical data includes: basic information data, bleeding-related data, coagulation-related data, infection-related indicator data, and vital signs time series. The first construction module is used to construct a set of historical clinical status nodes based on the historical clinical data. The set of historical clinical status nodes includes: node type, node status value, node activation intensity, and node timestamp. The second construction module is used to construct a causal chain structure for acute gastrointestinal bleeding based on the set of historical clinical status nodes. The acquisition module is used to acquire target clinical data of individuals that change continuously over time; The third construction module is used to construct an individualized disease status map based on the causal chain structure and target clinical data of acute gastrointestinal bleeding. The individualized disease status map is used to represent the current disease status, stage transition relationship and potential evolution direction of an individual. The identification module is used to identify the overall disease stage based on the individualized disease status map and using a stage competition method.

Images