Severe coagulopathy auxiliary analysis system

Abstract

The application belongs to the technical field of coagulopathy auxiliary analysis, and particularly relates to a severe coagulopathy auxiliary analysis system, which comprises a patient portrait module, a four-step analysis reasoning engine and a treatment suggestion module.The patient portrait module is used to collect patient information data and construct a patient information file according to the data.The four-step analysis reasoning engine is used to assist doctors in giving cause analysis, typing analysis, mechanism analysis and function analysis results of patients according to the patient information data collected in the patient portrait module.The treatment suggestion module gives corresponding treatment directions according to the analysis results given by the doctors with the assistance of the four-step analysis reasoning engine.Through systematic analysis of four dimensions of cause, typing, mechanism and function, and in combination with a knowledge graph, viscoelasticity experiment parameters, a multi-layer reasoning engine and a quantitative scoring model, the application can assist doctors in completing the evaluation of complex coagulation states in a short time, reduce human experience differences, and improve diagnosis accuracy and decision-making efficiency.

Description

Technical Field

[0001] This invention belongs to the field of auxiliary analysis technology for coagulation disorders, specifically an auxiliary analysis system for severe coagulation disorders. Background Technology

[0002] Severe coagulopathy is a common complication in critically ill patients, with an incidence rate as high as 40% to 67.6%. It is often accompanied by bleeding, shock, multiple organ dysfunction syndrome, and even death, with a mortality rate that can increase by more than four times. Its pathophysiological process is complex, involving multiple pathogenic factors, coagulation pathway disorders, vascular endothelial damage, and fibrinolytic system imbalance.

[0003] Traditional coagulation assessments primarily rely on static indicators such as routine coagulation markers, platelet count, and D-dimer, which are insufficient to comprehensively reflect the body's overall coagulation status, especially during dynamic processes involving both hypercoagulability and hypocoagulability. In recent years, viscoelastic coagulation tests (such as thromboelastography and coagulation and platelet function analyzers) have gradually become important tools for monitoring coagulation function in critical care due to their ability to assess coagulation initiation, clot formation, clot strength, and fibrinolysis in real time.

[0004] Currently, there is no intelligent auxiliary system for analyzing severe coagulopathy. Clinicians need to manually integrate a large amount of heterogeneous data (medical history, laboratory tests, imaging, bedside monitoring), which can easily lead to missed diagnoses or misjudgments, especially in high-pressure environments such as emergency rooms and ICUs, resulting in low decision-making efficiency.

[0005] Therefore, the present invention provides an auxiliary analysis system for severe coagulopathy. Summary of the Invention

[0006] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0007] The technical solution adopted by the present invention to solve its technical problem is: the critical coagulopathy auxiliary analysis system of the present invention includes a patient profiling module for collecting patient information data and constructing patient information files based on the data; The four-step analysis and reasoning engine is used to assist doctors in providing results of etiological analysis, subtyping analysis, mechanism analysis, and functional analysis of patients based on the patient information data collected in the patient profile module. The four-step analysis and reasoning engine includes: The etiology analysis unit analyzes the causes of coagulation disorders by acquiring patient information data and provides the etiology analysis results. The subtyping analysis unit acquires patient information data, performs a two-dimensional subtyping analysis of the patient's coagulation disease, and provides the subtyping analysis results; The mechanism analysis unit analyzes the patient's coagulation mechanism by acquiring patient information data and provides the mechanism analysis results; The functional analysis unit analyzes the patient's coagulation function abnormalities by acquiring the patient's information data and provides the functional analysis results; The treatment suggestion module provides corresponding treatment directions based on the analysis results provided by the doctor through a four-step analysis and reasoning engine.

[0008] Preferably, the patient profiling module includes: Information storage unit, used to store the patient's past medical records; The test result analysis unit is used to connect to the hospital's laboratory information management system and image storage and transmission system to obtain the patient's laboratory data, image data and clinical data; The information entry unit provides a data entry environment for medical staff to collect patient complaints. The vital signs monitoring unit is used to connect to the hospital's monitors and obtain data from the monitors in real time, thereby acquiring the patient's real-time vital signs information.

[0009] Preferably, the steps for the etiology analysis unit to analyze the etiology of the patient's coagulopathy are as follows: S1. Construct a knowledge graph of coagulation disorders etiology, and input common etiological types of coagulation disorders; the graph nodes include: clinical manifestations, abnormal laboratory data, and pathogenic events; and set the weight of each graph node; S2. Obtain the patient's past medical records through the information storage unit, obtain the patient's chief complaint through the information entry unit, and obtain the patient's clinical and laboratory data through the test result analysis unit; S3. Based on the patient's chief complaint, clinical and laboratory data, and the patient's past medical records, match the atlas nodes of each coagulation disorder. Calculate the confidence level of each coagulation disorder type based on the number and weight of matched atlas nodes. ; S4. According to the confidence level of each coagulation disorder etiology The values ​​are sorted by size and output in a table. The top three coagulopathy types with the highest confidence levels are highlighted, which are the results of the etiological analysis.

[0010] Preferably, the confidence level The calculation method is as follows: in Indicates symptoms Actual match with the cause of the disease The number of nodes in the graph. Indicates the cause of the disease The total number of graph nodes in a knowledge graph. Indicates the first The weights of each graph node, Indicates symptoms The weights of the graph nodes, Indicates symptoms The normalized intensity coefficient, Indicates the cause of the disease The score reflects the chronological order of coagulation abnormalities. and These are the weighting coefficients; Among them, the The logic for obtaining it is as follows: in The patient's original calcitonin level was obtained through laboratory data.

[0011] Preferably, the The logic for determining the value is as follows: If the cause If it occurs after a coagulation disorder, then output The value is 0; If the cause If the temporal relationship with coagulation abnormalities is unclear, then output The value is 0.3; If the cause If it occurs almost simultaneously with a coagulation abnormality, i.e., the time interval between the two events is within 1 hour, then the output... The value is 0.7; If the cause If the coagulation abnormality occurs prior to the coagulation abnormality and the time interval between the occurrences is greater than 1 hour, then output... The value is 1.

[0012] Preferably, the assignment logic for the graph node weights is as follows: The graph nodes are divided into highly specific nodes, moderately supportive nodes, and weakly correlated nodes; highly specific nodes are assigned a weight of 3, moderately supportive nodes are assigned a weight of 2, and weakly correlated nodes are assigned a weight of 1. in: Strongly specific nodes indicate that they occur at a very high frequency in the current etiology and at a very low frequency in other etiologies, and have a direct pathophysiological causal relationship. Moderately supportive nodes indicate a reasonable association with the target etiology, but are not specific and may also appear in 2-3 other etiologies; Weakly correlated nodes indicate that they appear in multiple etiologies and are non-specific systemic responses to the disease; their existence alone has no analytical significance.

[0013] Preferably, the logic of the typing analysis unit in analyzing the patient's coagulation mechanism is as follows: Based on the patient's basic coagulation test, platelet count, D-dimer, thromboelastography parameters, coagulation and platelet function analyzer parameters, and blood smear erythrocyte ratio, a dual-dimensional classification of phenotype and subphenotype is achieved. The thromboelastography parameters include: R time, K time, α angle, maximum amplitude, and LY30. The coagulation and platelet function analyzer parameters include activated clotting time, coagulation rate, and platelet function. The phenotypic classification logic is as follows: The patient's clinical presentation should be used to determine whether the hemorrhagic or thrombotic type is present. If the patient has both bleeding and thrombosis, the most life-threatening symptom should be used to determine whether the patient has the hemorrhagic or thrombotic type. The subtype classification logic is as follows: If the viscoelasticity test shows strong and rapid clot formation ability; such as shortened thromboelastography time, shortened K time, increased α angle and / or increased maximum amplitude; or shortened activation clotting time, increased clotting rate and / or increased platelet function in the coagulation and platelet function analyzer, then the analysis result of hypercoagulability is given. If the viscoelasticity test shows weak clot formation ability or excessively rapid dissolution; or if the R time, K time, α angle reduction and / or maximum amplitude of the thromboelastography are prolonged, or the activation coagulation time, coagulation rate and / or platelet function are prolonged and / or platelet function is reduced in the coagulation and platelet function analyzer, then the analysis result for hypocoagulability is given.

[0014] Preferably, the mechanism analysis unit analyzes the patient's coagulation mechanism logic as follows: Step 1: Construct a mechanism type library and input a set of evidence for the pathophysiological mechanisms of common coagulation abnormalities; The second step is to construct a multi-layered inference engine module to analyze the patient's coagulation mechanism through multi-layered inference. The specific multi-layered inference is as follows: First layer: Evidence extraction layer; receives patient vital sign data through information input unit and outputs a list of abnormal evidence; The second layer: mechanism hypothesis generation layer; based on the abnormal evidence list set, the mechanism type library is matched, the pathophysiological mechanism evidence set that intersects with the abnormal evidence list set is filtered, and the initial suspected pathophysiological mechanism library is generated. The third layer: integration and priority ranking layer; based on the number of intersections between the abnormal evidence list set and the pathophysiological mechanism evidence set, the pathophysiological mechanisms in the suspected pathophysiological mechanism library are ranked, and the top ten pathophysiological mechanisms are output as possible mechanisms.

[0015] Preferably, the logic of the functional analysis unit for analyzing the patient's coagulation function is as follows: The first step is to construct a four-dimensional analysis model based on coagulation function indicators, platelet function, tissue perfusion, and endothelial indicators. The second step is to obtain data on the patient in various dimensions and give scores for each dimension; The third step is to add up the scores of the four dimensions to obtain a comprehensive coagulation function score, and output the coagulation function analysis results based on the total score: coagulation function may be normal or close to normal, mild coagulation dysfunction may be present, moderate coagulation dysfunction may be present, or severe coagulation dysfunction may be present.

[0016] Preferably, it also includes a trend early warning module, which is used to monitor the patient's comprehensive coagulation function score interval to help determine the patient's condition trend and promptly issue an alarm to medical staff when the patient's condition deteriorates. The specific monitoring and early warning logic is as follows: The coagulation function comprehensive score is automatically calculated every 4-6 hours. If the patient's coagulation function score drops by 2 points or more during this period, it indicates that the patient's indicators have deteriorated, and a yellow warning is issued to remind medical staff to pay close attention to the patient's condition. If the patient's coagulation function score drops to 3 or less during this period, it indicates a precipitous deterioration in the patient's condition, giving a red alert and reminding medical staff to provide timely rescue. If the patient's coagulation function score rises by 1 point or more during this period, it indicates that the patient's condition has improved, and a green warning is given.

[0017] The beneficial effects of this invention are as follows: 1. The critical coagulopathy auxiliary analysis system described in this invention integrates patient history medical records, laboratory data, imaging data, bedside monitoring data, and medical staff input information through a patient profiling module to construct a dynamically updated holographic patient profile, solving the problem of scattered and fragmented data in traditional diagnosis and treatment, and providing complete information support for rapid judgment.

[0018] 2. The critical coagulation disease auxiliary analysis system described in this invention, through systematic analysis of four dimensions—etiology, subtype, mechanism, and function—combined with knowledge graphs, viscoelastic experimental parameters, multi-layer inference engines, and quantitative scoring models, assists doctors in completing the assessment of complex coagulation states in a short time, reducing human experience differences and improving diagnostic accuracy and decision-making efficiency.

[0019] 3. The critical coagulation disease auxiliary analysis system described in this invention realizes the quantitative assessment of coagulation function through a comprehensive coagulation function scoring system, and combines it with a trend early warning module for continuous monitoring. Once the score shows abnormal fluctuations, it automatically triggers graded early warnings to help medical staff identify the trend of disease deterioration in advance and strive for a treatment window. Attached Figure Description

[0020] The invention will now be further described with reference to the accompanying drawings.

[0021] Figure 1 This is a system framework diagram of the present invention. Detailed Implementation

[0022] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0023] The critical coagulopathy auxiliary analysis system described in this embodiment of the invention includes a patient profiling module for collecting patient information data and constructing patient information files based on the data; The four-step analysis and reasoning engine is used to assist doctors in providing results of etiological analysis, subtyping analysis, mechanism analysis, and functional analysis of patients based on the patient information data collected in the patient profile module. The four-step analysis and reasoning engine includes: The etiology analysis unit analyzes the causes of coagulation disorders by acquiring patient information data and provides the etiology analysis results. The subtyping analysis unit acquires patient information data, performs a two-dimensional subtyping analysis of the patient's coagulation disease, and provides the subtyping analysis results; The mechanism analysis unit analyzes the patient's coagulation mechanism by acquiring patient information data and provides the mechanism analysis results; The functional analysis unit analyzes the patient's coagulation function abnormalities by acquiring the patient's information data and provides the functional analysis results; The treatment suggestion module provides corresponding treatment directions based on the analysis results provided by the doctor through a four-step analysis and reasoning engine.

[0024] To address the problem of low decision-making efficiency caused by clinicians having to manually integrate large amounts of heterogeneous data in existing technologies, this invention can be implemented by setting up a patient profiling module to build a unique user profile for each patient, collect patient information data, and lay a solid foundation for doctors to analyze the condition and provide emergency care. The four-step analysis and reasoning engine assists doctors in analyzing patients' conditions through four steps: etiology, subtyping, mechanism, and function. It helps doctors integrate a large amount of data, making it easy for them to review and providing systematic analysis results, including etiology analysis results, subtyping analysis results, mechanism analysis results, and function analysis results. Based on these analysis results, doctors can quickly make correct judgments about the condition, thereby improving the efficiency of patient treatment and preventing the patient's condition from worsening. Finally, the embodiments of the present invention can also provide corresponding treatment directions based on the analysis results given by the four-step analysis and reasoning engine through the treatment suggestion module, thereby assisting doctors to quickly make treatment plans and improve the efficiency of patient treatment. In summary, the embodiments of the present invention, by establishing a patient profile module and combining it with an original four-step analysis and reasoning engine, assist doctors in quickly making diagnoses and treatment plans based on the actual situation of each patient, thereby improving the treatment efficiency of patients with severe coagulopathy.

[0025] Furthermore, the patient profiling module includes: The information storage unit is used to store the patient's past medical records; these mainly include: the patient's medical history, medication history, and trauma records. The test result analysis unit is used to connect to the hospital's laboratory information management system and image storage and transmission system to obtain the patient's laboratory data, image data and clinical data; The information entry unit provides a data entry environment for medical staff to collect patient complaints. Understandably, patient complaints are the patient's verbal description of their symptoms. For example, a patient may describe being bitten by a venomous snake or being hit by a vehicle. Such verbal descriptions are helpful in analyzing the patient's coagulation disorders. The vital signs monitoring unit is used to connect to the hospital's monitors and obtain data from the monitors in real time, thereby acquiring the patient's real-time vital signs information.

[0026] The information storage unit can effectively extract the patient's past medical history and medication information, assisting doctors in judging the medical condition and treatment, avoiding other complications caused by the patient's past medical history during the treatment process, and improving the efficiency of patient care.

[0027] Furthermore, the steps for the etiology analysis unit to analyze the etiology of the patient's coagulopathy are as follows: S1. Construct a knowledge graph of coagulation disorders etiology, and input common etiological types of coagulation disorders; the graph nodes include: clinical manifestations, abnormal laboratory data, and pathogenic events; and set the weight of each graph node; S2. The information storage unit retrieves the patient's past medical records, mainly including past medical history, medication history, and trauma records. The information entry unit retrieves the patient's chief complaint. It is understood that the patient's chief complaint is also a key clue for etiological analysis. For example, if the patient reports being bitten by a venomous snake, being hit by a vehicle, or falling from a height, it can quickly help the doctor diagnose the cause. It should be noted that since the patient's chief complaint is entered by medical staff through the information entry unit, the medical staff should independently extract the key information from the patient's chief complaint and integrate it into the standard information in the atlas nodes to reduce the difficulty of system recognition. The test result analysis unit retrieves the patient's clinical and laboratory data, mainly including present medical history, body temperature, infection markers, and etiological results. S3. Based on the patient's chief complaint, clinical and laboratory data, and the patient's past medical records, match the atlas nodes of each coagulation disorder. Calculate the confidence level of each coagulation disorder type based on the number and weight of matched atlas nodes. ; S4. According to the confidence level of each coagulation disorder etiology The values ​​are sorted by size and output in a table. The top three coagulopathy types with the highest confidence levels are highlighted, which are the results of the etiological analysis.

[0028] By constructing a knowledge graph of etiology, the etiology is broken down into multiple indicators, and the confidence level of each coagulation disorder type is calculated. This provides an analysis of the cause of the disease, helps doctors integrate a large amount of patient data, assists doctors in making quick decisions, and improves treatment efficiency.

[0029] Furthermore, the confidence level The calculation method is as follows: in Indicates symptoms Actual match with the cause of the disease The number of nodes in the graph. Indicates the cause of the disease The total number of graph nodes in a knowledge graph. Indicates the first The weights of each graph node, Indicates symptoms The weights of the graph nodes, Indicates symptoms The normalized intensity coefficient, Indicates the cause of the disease The score reflects the chronological order of coagulation abnormalities. and These are the weighting coefficients.

[0030] It should be noted that, based on clinical experience and research, it is recommended to... The value is set to 0.8. The value was set to 0.2; based on the confidence level of the etiology. The calculation and sorting process helps doctors identify possible causes of illness. It should be noted that in practical use, doctors can also view the specific symptoms displayed by the patient in each possible cause to assist in making quick decisions and improve treatment efficiency.

[0031] The The logic for obtaining it is as follows: in The patient's procalcitonin (PCT) level was obtained through laboratory data.

[0032] Understandably, in the analysis of severe infections, PCT > 10 ng / mL usually indicates severe bacterial infection / sepsis, and when PCT > 20 ng / mL, its strength as evidence of infection has reached the "strongest". Higher values ​​do not significantly increase the analytical confidence, so a threshold of 20 is used.

[0033] Furthermore, the aforementioned The logic for determining the value is as follows: If the cause If it occurs after a coagulation disorder, then output The value is 0; If the cause If the temporal relationship with coagulation abnormalities is unclear, then output The value is 0.3; If the cause If it occurs almost simultaneously with a coagulation abnormality, i.e., the time interval between the two events is within 1 hour, then the output... The value is 0.7; If the cause If the coagulation abnormality occurs prior to the coagulation abnormality and the time interval between the occurrences is greater than 1 hour, then output... The value is 1.

[0034] Specifically, based on the principles of causal analysis in clinical epidemiology and the grading of temporal evidence in evidence-based medicine, the causal event preceding the coagulation abnormality in time is a necessary condition for establishing a causal relationship. This setting can effectively eliminate causes with contradictory temporal logic in the etiology knowledge graph, thereby achieving the effect of rapid decision-making.

[0035] Furthermore, the assignment logic for the graph node weights is as follows: The graph nodes are divided into highly specific nodes, moderately supportive nodes, and weakly correlated nodes; highly specific nodes are assigned a weight of 3, moderately supportive nodes are assigned a weight of 2, and weakly correlated nodes are assigned a weight of 1; it is understood that the assignment of the above nodes is not unique and can be adjusted according to the actual situation. in: A highly specific node is one that occurs very frequently in the current etiology but very rarely in other etiologies, and has a direct pathophysiological causal relationship. For example, in septic coagulopathy, a positive blood culture is a highly specific node; in traumatic coagulopathy, a clear history of severe trauma accompanied by massive hemorrhage or shock is a highly specific node. Moderately supporting nodes indicate that there is a reasonable association with the target etiology, but it is not specific and may also appear in 2-3 other etiologies. These nodes usually need to be combined with other evidence to form an analysis. For example, septic coagulopathy, significantly elevated white blood cell count, and significantly elevated C-reactive protein are moderately supporting nodes. Weakly correlated nodes are those that appear in multiple etiologies and are non-specific systemic responses to the disease; their existence alone has no analytical significance. For example, fever, fatigue, and mild anemia are weakly correlated nodes.

[0036] Furthermore, the logic by which the typing analysis unit analyzes the patient's coagulation mechanism is as follows: Based on the patient's basic coagulation test, platelet count, D-dimer, thromboelastography parameters, coagulation and platelet function analyzer parameters, and the proportion of cleaved erythrocytes in the blood smear, a dual-dimensional phenotypic and subphenotypic classification is achieved. The thromboelastography parameters include: R time (reaction time, i.e., the time required from the start of detection to the tracing amplitude reaching 2 mm), K time (Kinetics Time, i.e., the time required from the end of R time to the tracing amplitude reaching 20 mm), α angle (the angle between the tangent line drawn along the tracing from the end of R time and the horizontal line), maximum amplitude (MA, i.e., the maximum width on the tracing), and LY30 (i.e., the percentage of blood clot dissolution 30 minutes after the appearance of the MA value). The coagulation and platelet function analyzer parameters include activated clotting time (ACT), clotting rate (CR), and platelet function (PF). The phenotypic classification logic is as follows: The patient's clinical presentation should be used to determine whether the hemorrhagic or thrombotic type is present. If the patient has both bleeding and thrombosis, the most life-threatening symptom should be used to determine whether the patient is hemorrhagic or thrombotic. The logic for determining phenotypic classification based on patients' clinical manifestations is as follows: The main characteristic of hemorrhagic coagulopathy is: inability to stop bleeding; its main symptoms include: Spontaneous bleeding is mainly manifested as: skin ecchymosis, epistaxis, and gingival bleeding without obvious cause; Post-traumatic bleeding is difficult to stop, mainly manifested as: prolonged bleeding from small wounds and persistent oozing after surgery; Deep bleeding mainly manifests as: joint cavity hematoma, muscle hematoma, internal organ bleeding (such as melena, hematuria, hemoptysis, menorrhagia), and in severe cases, intracranial hemorrhage; The main characteristic of thrombotic coagulopathy is that blood that shouldn't clot clots; its main symptoms include: Arterial thrombosis is mainly manifested as: sudden chest pain (myocardial infarction), severe pain and coldness in the limbs with no pulse (arterial embolism), and stroke symptoms (hemiplegia, aphasia). Venous thrombosis: Venous thrombosis is divided into deep vein thrombosis and pulmonary embolism; Deep vein thrombosis is mainly characterized by: unilateral limb swelling, pain, increased skin temperature, and a positive Homans sign. Pulmonary embolism is mainly characterized by sudden chest pain, difficulty breathing, hemoptysis, and syncope.

[0037] Microvascular thrombosis is mainly manifested as: reticular cyanosis of the skin, ischemic blackening of the fingertips, and multiple organ failure (such as hemolytic uremic syndrome and thrombotic thrombocytopenic purpura). The subtype classification logic is as follows: If the viscoelasticity test shows strong and rapid clot formation ability; such as shortened thromboelastography time, shortened K time, increased α angle and / or increased MA; or shortened ACT, increased CR and / or increased PF in the coagulation and platelet function analyzer, then the analysis result of hypercoagulability is given. If the viscoelasticity test shows weak clot formation ability or excessively rapid dissolution; such as prolonged R time, prolonged K time, reduced α angle and / or decreased MA in thromboelastography, or prolonged ACT, decreased CR and / or decreased PF in coagulation and platelet function analyzer, then the analysis result for hypocoagulability is given.

[0038] Furthermore, the mechanism analysis unit analyzes the patient's coagulation mechanism logic as follows: Step 1: Construct a mechanism type library, and input the evidence set of pathophysiological mechanisms of common coagulation abnormalities. Possible mechanism examples are given below: coagulation factor deficiency or inhibition (e.g., vitamin K deficiency, insufficient liver synthesis); platelet dysfunction (decreased number or activity); hyperfibrinolysis (primary or secondary); vascular endothelial injury (assessed by plasma thrombomodulin (TM) levels); increased pathological anticoagulants (e.g., lupus anticoagulants); microcirculatory disturbances (elevated lactate, poor tissue perfusion), etc. It is understood that the above are only some of the mechanisms, and the coagulation mechanism type library should be improved as much as possible in practical applications. Common mechanistic evidence mainly includes: routine coagulation tests, platelet count and function indicators (thromboelastography-MA, coagulation and platelet function analyzer-PF), fibrinolytic markers (D-dimer, plasmin-α2 plasmin inhibitor complex, fibrin / fibrinogen degradation products), endothelial injury markers [thrombomodulin, von Willebrand factor (vWF)], microcirculation indicators (lactic acid, tissue perfusion parameters), and clinical bleeding / thrombosis manifestations, etc. The second step is to construct a multi-layered inference engine module to analyze the patient's coagulation mechanism through multi-layered inference. The specific multi-layered inference is as follows: First layer: Evidence extraction layer; receives patient vital sign data through information input unit and outputs a list of abnormal evidence; The second layer: mechanism hypothesis generation layer; based on the abnormal evidence list set, the mechanism type library is matched, the pathophysiological mechanism evidence set that intersects with the abnormal evidence list set is filtered, and the initial suspected pathophysiological mechanism library is generated. The third layer: integration and priority ranking layer; based on the number of intersections between the abnormal evidence list set and the pathophysiological mechanism evidence set, the pathophysiological mechanisms in the suspected pathophysiological mechanism library are ranked, and the top ten pathophysiological mechanisms are output as possible mechanisms.

[0039] Furthermore, the logic of the functional analysis unit in analyzing the patient's coagulation function is as follows: The first step is to construct a four-dimensional analysis model based on coagulation function indicators, platelet function, tissue perfusion, and endothelial indicators. The second step is to obtain data on the patient in various dimensions and give scores for each dimension; The third step is to multiply the scores of the four dimensions by their corresponding weights and then sum them up to obtain the Coagulation Function Score (CFS). Based on the total score, the coagulation function analysis results are output as follows: coagulation function may be normal or close to normal, mild coagulation dysfunction may be present, moderate coagulation dysfunction may be present, or severe coagulation dysfunction may be present.

[0040] Specifically, the coagulation function indicators are based on routine coagulation test data to assess the overall functional status of the coagulation system. The assessment parameters include: prothrombin time (PT), activated partial thromboplastin time (APTT), international normalized ratio (INR), and fibrinogen (FIB). The specific scoring logic is as follows: If all indicators are within the normal range, output 3 points; If any indicator exceeds the normal range but does not reach the clinical intervention threshold, 2 points will be output. If all indicators reach the clinical intervention threshold but do not endanger life, output 1 point; If any of the indicators are significantly abnormal, the risk of bleeding or thrombosis will be output as 0 points. The clinical intervention thresholds for each indicator are as follows: PT / APTT > 1.5 times the upper limit of normal; FIB < 1.5 g / L; FIB < 1.0 g / L; The platelet function dimension assesses platelet quantity and function by combining platelet count and the MA value in thromboelastography parameters. Assessment parameters include: platelet count and thromboelastography-MA (maximum amplitude). The specific scoring logic is as follows: If the platelet count is greater than or equal to 100 × 10 9 If / L and MA are normal, then output 3 points; If the platelet count is 50–100 × 10 9 If / L or MA decreases slightly (by 10-20%), output 2 points; If the platelet count is 20–50 × 10 9 If / L or MA is moderately reduced (decreased by 20-30%), output 1 point; If platelet count <20×10 9 If / L or MA decreases significantly (decreases >30%), output 0 points; Tissue perfusion dimension assesses microcirculatory status by combining hemodynamic markers, including lactate and central venous oxygen saturation (ScvO2).

[0041] The specific scoring logic is as follows: If the patient's lactate level is normal (≤ 2.0 mmol / L) and ScvO2 is normal, a score of 3 will be output. If the patient has a slightly elevated lactate level (2.1 – 3.9 mmol / L), a score of 2 is awarded. If the patient has a moderately elevated lactate level (4.0 – 8.0 mmol / L) accompanied by a decreased ScvO2, then 1 point will be awarded. If the patient has persistently elevated lactate levels (>8.0 mmol / L), significant tissue perfusion impairment, or mottled limbs, a score of 0 will be awarded. Endothelial markers assess endothelial function based on biomarkers of vascular endothelial injury; the parameters assessed include: von Willebrand factor (vWF) and thrombomodulin.

[0042] The specific scoring logic is as follows: If the patient's endothelial markers are normal, a score of 3 is output. If the patient's endothelial markers are mildly abnormal (elevated by 50%), 2 points will be awarded. If the patient's endothelial markers are significantly abnormal (elevated by 100%), 1 point will be awarded. If the patient has severe endothelial damage (elevation of more than 100%), a score of 0 will be output. In summary, the total score for CFS ranges from 0 to 12. The logic for functional grading based on the total score is as follows: If the CFS score is 10–12, the output result is likely that the coagulation function is normal or close to normal. If the CFS score is 7–9, the possible outcome is mild coagulation dysfunction. If the CFS score is 4–6, the possible outcome is moderate coagulation dysfunction. If the CFS score is 0–3, the possible outcome of severe coagulation dysfunction will be output. It should be noted that when conducting comprehensive coagulation function scoring, the principle of dimensional orthogonality must be followed. That is, the same evidence should only be scored in its most specific dimension, and each dimension should assess different pathophysiological aspects to avoid overlap. For example, when scoring the platelet function dimension, the influence of coagulation factor abnormalities should not be considered; similarly, when scoring the coagulation function index dimension, the influence of platelet count should not be considered or assessed. This avoids the situation where the same pathological manifestation is repeatedly calculated in different dimensions, which could lead to artificially inflated or distorted scores. Since different etiologies lead to inconsistent clinical manifestations, the weights of each dimension need to be adjusted according to the actual etiology. For example: If the patient has sepsis-related coagulopathy, then endothelial parameters and tissue perfusion are the primary core dimensions, while platelet function is a secondary dimension. In this case: The weighting factor for the endothelial dimension is 1.5 because sepsis directly damages the endothelium, releasing vWF / TM, which is the initiating factor for coagulation disorders. The weighting factor for the tissue perfusion dimension is 1.0 because sepsis is often accompanied by microcirculatory disturbances, but these disturbances are nonspecific, and elevated lactate reflects the severity of the condition. The weighting coefficient for the platelet function dimension is 0.8, because thrombocytopenia is mostly secondary consumption and not a specific manifestation of sepsis. The weighting coefficient for the coagulation function index dimension is 0.7, because changes in PT / APTT occur relatively late and are influenced by multiple factors; For example, if the patient has traumatic blood loss, the weighting coefficient for the endothelial dimension is 0.7; the weighting coefficient for the tissue perfusion dimension is 1.5; the weighting coefficient for the platelet function dimension is 1.2; and the weighting coefficient for the coagulation function index dimension is 0.6. It is understandable that different causes of disease will lead to different clinical manifestations, and therefore will correspond to different dimensional weights. The dimensional weights corresponding to each cause will not be elaborated here. These weights need to be manually entered into the information storage unit inside the system. Once the doctor determines the cause of the patient's disease, the system can automatically retrieve the dimensional weights corresponding to that cause to calculate the CFS score. By conducting a comprehensive assessment across four key dimensions—coagulation system, platelets, clinical manifestations, and vascular endothelium—with quantitative scoring standards for each dimension, the system assists physicians in synthesizing data from these four key dimensions. This helps physicians reduce subjective judgment differences, make accurate medical assessments and clinical decisions, and improve patient treatment efficiency. Furthermore, the scores can be repeatedly calculated over time, which can also be used to monitor disease progression and treatment response, enabling dynamic management of patient conditions.

[0043] Furthermore, it also includes a trend warning module, which is used to monitor the patient's CFS score interval, assist in judging the patient's condition trend, and promptly issue an alert to medical staff when the patient's condition deteriorates. The specific monitoring and warning logic is as follows: The CFS score is automatically calculated every 4-6 hours. If the patient's CFS score drops by 2 points or more during this period, it indicates that the patient's condition has deteriorated, and a yellow warning is issued to remind medical staff to pay close attention to the patient's condition. If the patient's CFS score drops to 3 or less during this period, it indicates a precipitous deterioration in the patient's condition, giving a red alert and reminding medical staff to provide timely rescue. If the patient's CFS score rises by 1 point or more during this period, it indicates that the patient's condition has improved, and a green alert is given.

[0044] By monitoring patients' CFS scores at intervals, the analysis is expanded from a "single time point" to a "continuous time series," which better reflects the rapidly changing nature of critical illnesses. Furthermore, when malignant fluctuations occur in a patient's CFS score, timely intervention can be detected, preventing the condition from deteriorating to a critical stage before it is discovered. This achieves a leap from "identifying problems" to "predicting risks," securing a valuable intervention window for clinicians. It should be noted that the interval for monitoring patients' CFS scores can be set based on the patient's initial CFS score. If the initial CFS score is 0–3 or 4–6, it is recommended to reassess the CFS score every 3 hours; if the patient's CFS score is 7–9, it is recommended to reassess the CFS score every 8 hours; and if the patient's CFS score is 10–12, it is recommended to reassess the CFS score every 12 hours.

[0045] example: The hospital emergency department received a 58-year-old male patient. His family reported that he had a high fever and had been experiencing confusion for two hours upon admission. He also had a history of diabetes.

[0046] The medical staff entered the chief complaint into the system: "High fever, confusion, history of diabetes"; The system then automatically connects to LIS / PACS via the test result analysis unit to obtain laboratory data: PCT 35 ng / mL, platelets 65 × 10⁻⁶. 9 / L, D-dimer 15 mg / L; the monitor showed blood pressure 85 / 50 mmHg and heart rate 130 beats / min.

[0047] The system integrates the patient's historical medical records (history of diabetes), real-time vital signs (hypotension, high fever), and abnormal laboratory data (PCT↑, platelet↓, D-dimer↑) to generate a dynamic patient profile.

[0048] The four-step analysis and inference engine then runs: Etiological analysis: The system matched the knowledge graph and calculated the confidence level. The results showed that "septic coagulopathy" had the highest confidence level. =0.92); the system indicated that "septic coagulopathy" was highly likely and listed key evidence points such as blood culture and elevated C-reactive protein to assist doctors in making a judgment; Subtype analysis: Thromboelastography parameters showed a shortened R time, a decreased MA, and an increased LY30. The system determined the subtype to be "possibly hypercoagulable thrombosis" to assist doctors in making a judgment. Mechanism analysis: The multi-layer inference engine extracts abnormal evidence (thrombocytopenia, elevated D-dimer, abnormal endothelial injury markers), matches it with the mechanism library, and outputs possible mechanisms: "fibrinolysis inhibition", "platelet consumption" and "vascular endothelial injury" to assist doctors in making judgments. Functional analysis: Four-dimensional scoring results: coagulation function index 2 points, platelet function 1 point, tissue perfusion 1 point, endothelial index dimension 1 point, CFS total score 5 points. The system prompts "moderate coagulation dysfunction may be present" to assist doctors in making a judgment. Based on the analysis results, the system then proposed the following treatment recommendations: aggressive anti-infection treatment; consider antifibrinolytic drugs based on the thromboelastography results; monitor platelet count and prepare for platelet transfusion; assist doctors in making treatment decisions. Once the patient is transferred to the ICU, a trend warning system is activated, and the CFS score is dynamically monitored. The system automatically calculates the CFS score every 4 hours. After 6 hours, the CFS score dropped from 5 to 2, triggering a red alert: "The patient's coagulation function has deteriorated precipitously. Please intervene immediately!" The medical team immediately started the rescue, including platelet transfusion, antifibrinolytic therapy and enhanced anti-infection treatment.

[0049] Ultimately, after treatment, the patient's CFS score gradually rose to 7 points, and the system switched to a green alert: "Coagulation function improved."

[0050] During the treatment process, this invention continuously provides information on the subtyping and mechanism evolution to assist in adjusting the treatment plan; it realizes integrated closed-loop management of data integration, intelligent analysis, dynamic early warning and treatment support, assisting doctors to complete the assessment of complex coagulation status in a short time while monitoring the patient's later disease evolution trend, significantly improving the treatment efficiency and safety of patients with severe coagulopathy.

[0051] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A system for auxiliary analysis of severe coagulopathy, characterized in that: This includes a patient profiling module, used to collect patient information data and build patient information profiles based on the data; The four-step analysis and reasoning engine is used to assist doctors in providing results of etiological analysis, subtyping analysis, mechanism analysis, and functional analysis of patients based on the patient information data collected in the patient profile module. The four-step analysis and reasoning engine includes: The etiology analysis unit analyzes the causes of coagulation disorders by acquiring patient information data and provides the etiology analysis results. The subtyping analysis unit acquires patient information data, performs a two-dimensional subtyping analysis of the patient's coagulation disease, and provides the subtyping analysis results; The mechanism analysis unit analyzes the patient's coagulation mechanism by acquiring patient information data and provides the mechanism analysis results; The functional analysis unit analyzes the patient's coagulation function abnormalities by acquiring the patient's information data and provides the functional analysis results; The treatment suggestion module provides corresponding treatment directions based on the analysis results provided by the doctor through a four-step analysis and reasoning engine.

2. The auxiliary analysis system for severe coagulopathy according to claim 1, characterized in that: The patient profiling module includes: Information storage unit, used to store the patient's past medical records; The test result analysis unit is used to connect to the hospital's laboratory information management system and image storage and transmission system to obtain the patient's laboratory data, image data and clinical data; The information entry unit provides a data entry environment for medical staff to collect patient complaints. The vital signs monitoring unit is used to connect to the hospital's monitors and obtain data from the monitors in real time, thereby acquiring the patient's real-time vital signs information.

3. The auxiliary analysis system for severe coagulopathy according to claim 2, characterized in that: The steps for the etiology analysis unit to analyze the etiology of the patient's coagulation disorder are as follows: S1. Construct a knowledge graph of coagulation disorders etiology, and input common etiological types of coagulation disorders; the graph nodes include: clinical manifestations, abnormal laboratory data, and pathogenic events; and set the weight of each graph node; S2. Obtain the patient's past medical records through the information storage unit, obtain the patient's chief complaint through the information entry unit, and obtain the patient's clinical and laboratory data through the test result analysis unit; S3. Based on the patient's chief complaint, clinical and laboratory data, and the patient's past medical records, match the atlas nodes of each coagulation disorder. Calculate the confidence level of each coagulation disorder type based on the number and weight of matched atlas nodes. ; S4. According to the confidence level of each coagulation disorder etiology The values ​​are sorted by size and output in a table. The top three coagulation disorders with the highest confidence levels are highlighted, which are the results of the etiological analysis.

4. The auxiliary analysis system for severe coagulopathy according to claim 3, characterized in that: The confidence level The calculation method is as follows: in Indicates symptoms Actual match with the cause of the disease The number of nodes in the graph. Indicates the cause of the disease The total number of graph nodes in a knowledge graph. Indicates the first The weights of each graph node, Indicates symptoms The weights of the graph nodes, Indicates symptoms The normalized intensity coefficient, Indicates the cause of the disease The score reflects the chronological order of coagulation abnormalities. and These are the weighting coefficients; Among them, the The logic for obtaining it is as follows: in The patient's original calcitonin level was obtained through laboratory data.

5. The auxiliary analysis system for severe coagulopathy according to claim 4, characterized in that: The The logic for determining the value is as follows: If the cause If it occurs after a coagulation disorder, then output The value is 0; If the cause If the temporal relationship with coagulation abnormalities is unclear, then output The value is 0.3; If the cause If it occurs almost simultaneously with a coagulation abnormality, i.e., the time interval between the two events is within 1 hour, then the output... The value is 0.7; If the cause If the coagulation abnormality occurs prior to the coagulation abnormality and the time interval between the occurrences is greater than 1 hour, then output... The value is 1.

6. The auxiliary analysis system for severe coagulopathy according to claim 5, characterized in that: The assignment logic for the weights of the graph nodes is as follows: The graph nodes are divided into highly specific nodes, moderately supportive nodes, and weakly correlated nodes; highly specific nodes are assigned a weight of 3, moderately supportive nodes are assigned a weight of 2, and weakly correlated nodes are assigned a weight of 1. in: Strongly specific nodes indicate that they occur at a very high frequency in the current etiology and at a very low frequency in other etiologies, and have a direct pathophysiological causal relationship. Moderately supportive nodes indicate a reasonable association with the target etiology, but are not specific and may also appear in 2-3 other etiologies; Weakly correlated nodes indicate that they appear in multiple etiologies and are non-specific systemic responses to the disease; their existence alone has no analytical significance.

7. The auxiliary analysis system for severe coagulopathy according to claim 6, characterized in that: The logic of the typing analysis unit in analyzing the patient's coagulation disorder typing is as follows: Based on the patient's basic coagulation test, platelet count, D-dimer, thromboelastography parameters, coagulation and platelet function analyzer parameters, and blood smear erythrocyte ratio, a dual-dimensional classification of phenotype and subphenotype is achieved. The thromboelastography parameters include: R time, K time, α angle, maximum amplitude, and LY30. The coagulation and platelet function analyzer parameters include activated clotting time, coagulation rate, and platelet function. The phenotypic classification logic is as follows: The patient's clinical presentation should be used to determine whether the hemorrhagic or thrombotic type is present. If the patient has both bleeding and thrombosis, the most life-threatening symptom should be used to determine whether the patient is hemorrhagic or thrombotic. The subtype classification logic is as follows: If the viscoelasticity test shows strong and rapid clot formation ability; such as shortened thromboelastography time, shortened K time, increased α angle and / or increased maximum amplitude; or shortened activation clotting time, increased clotting rate and / or increased platelet function in the coagulation and platelet function analyzer, then the analysis result of hypercoagulability is given. If the viscoelasticity test shows weak clot formation ability or excessively rapid dissolution; or if the R time, K time, α angle reduction and / or maximum amplitude of the thromboelastography are prolonged, or the activation coagulation time, coagulation rate and / or platelet function are prolonged and / or platelet function is reduced in the coagulation and platelet function analyzer, then the analysis result for hypocoagulability is given.

8. The auxiliary analysis system for severe coagulopathy according to claim 7, characterized in that: The mechanism analysis unit analyzes the patient's coagulation mechanism logic as follows: Step 1: Construct a mechanism type library and input a set of evidence for the pathophysiological mechanisms of common coagulation abnormalities; The second step is to construct a multi-layered inference engine module to analyze the patient's coagulation mechanism through multi-layered inference. The specific multi-layered inference is as follows: First layer: Evidence extraction layer; receives patient vital sign data through information input unit and outputs a list of abnormal evidence; The second layer: mechanism hypothesis generation layer; based on the abnormal evidence list set, the mechanism type library is matched, the pathophysiological mechanism evidence set that intersects with the abnormal evidence list set is filtered, and the initial suspected pathophysiological mechanism library is generated. The third layer: integration and priority ranking layer; based on the number of intersections between the abnormal evidence list set and the pathophysiological mechanism evidence set, the pathophysiological mechanisms in the suspected pathophysiological mechanism library are ranked, and the top ten pathophysiological mechanisms are output as possible mechanisms.

9. The auxiliary analysis system for severe coagulopathy according to claim 8, characterized in that: The logic of the functional analysis unit in analyzing the patient's coagulation function is as follows: The first step is to construct a four-dimensional analysis model based on coagulation function indicators, platelet function, tissue perfusion, and endothelial indicators. The second step is to obtain data on the patient in various dimensions and give scores for each dimension; The third step is to add up the scores of the four dimensions to obtain a comprehensive coagulation function score, and output the coagulation function analysis results based on the total score: coagulation function may be normal or close to normal, mild coagulation dysfunction may be present, moderate coagulation dysfunction may be present, or severe coagulation dysfunction may be present.

10. The auxiliary analysis system for severe coagulopathy according to claim 9, characterized in that: It also includes a trend warning module, which monitors the patient's comprehensive coagulation function score at intervals to help determine the patient's condition trend and promptly issues an alert to medical staff when the patient's condition deteriorates. The specific monitoring and warning logic is as follows: The coagulation function comprehensive score is automatically calculated every 4-6 hours. If the patient's coagulation function score drops by 2 points or more during this period, it indicates that the patient's indicators have deteriorated, and a yellow warning is issued to remind medical staff to pay close attention to the patient's condition. If the patient's coagulation function score drops to 3 or less during this period, it indicates a precipitous deterioration in the patient's condition, giving a red alert and reminding medical staff to provide timely rescue. If the patient's coagulation function score rises by 1 point or more during this period, it indicates that the patient's condition has improved, and a green warning is given.

Images