A method and system for performing a vte system with a standardized scale

Abstract

The application relates to the technical field of biomedical engineering, and discloses a VTE system execution method and system combined with a standardized scale, wherein the method comprises the following steps: performing standardized processing on original data to obtain standardized data; constructing a risk intensity flow; performing integral operation to obtain a cumulative integral value of a risk factor; calculating a dynamic risk trajectory; fusing a baseline score value of the standardized scale with the dynamic risk trajectory to obtain a total risk trajectory; calculating an instantaneous risk probability, and judging whether a VTE risk early warning is triggered according to an adaptive threshold value; the four-dimensional risk intensity flow of a body position, anesthesia, hemostatic drugs and circulation parameters is constructed, continuous quantitative monitoring of VTE risk factors is realized, early identification and timely early warning of VTE risks are realized through a dual mechanism of continuous triggering and acute triggering, and the accuracy, real-time performance and individualization level of risk assessment are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of biomedical engineering technology, and more specifically, to a method and system for implementing a VTE system incorporating standardized scales. Background Technology

[0002] Venous thromboembolism (VTE), including deep vein thrombosis (DVT) and pulmonary embolism (PE), is a common and serious perioperative complication, especially during general anesthesia and the early postoperative period, when it easily progresses to a hypercoagulable and stasis-prone state. Failure to identify and intervene promptly can lead to adverse outcomes. Standardized scales such as the Caprini and Padua, widely used in clinical practice, provide hospitalized patients with structured and quantifiable risk assessment tools, helping to improve the consistency and operability of assessments.

[0003] Chinese patent application CN120089367A discloses a VTE risk assessment method based on the Caprini and Padua scales. The method involves: S1, inputting patient information; S2, quantifying and assessing the VTE risk of the patient information to obtain a corresponding score; S3, summarizing the various assessment results and scores to obtain the patient's total VTE risk score; and S4, providing surgical clinical assessment decisions based on the total VTE risk score. The advantages of this invention compared to existing technologies are: more detailed assessment standards, improving the accuracy and reliability of assessment results; flexible assessment modules that can be adjusted according to the patient's specific situation to ensure the accuracy of assessment results; and intelligent decision support that can automatically provide strong support for clinical decision-making based on assessment results, reducing the workload of surgical clinicians and improving the practicality and reliability of assessment results.

[0004] However, existing technologies employ a static, discrete-time-point scoring model, typically scoring at fixed points during admission assessment, preoperative assessment, or postoperative period, and providing preventative recommendations accordingly. While this model has some applicability in routine ward settings, it suffers from significant timeliness and sensitivity issues during the critical window of drastic physiological fluctuations, particularly during surgery and the early postoperative period. It lacks a mechanism for sensing critical events and for automatic re-scoring, resulting in risk assessment results lagging behind rapid changes in the patient's physiological state, thus impacting the timeliness and accuracy of clinical decision-making. Summary of the Invention

[0005] The purpose of this invention is to provide a VTE system execution method and system that incorporates standardized scales in order to solve the above-mentioned problems.

[0006] This invention provides a method for implementing a VTE system using standardized scales, comprising the following steps:

[0007] Collect the patient's raw data and standardize the raw data to obtain standardized data;

[0008] Based on the standardized data, a risk intensity stream for multiple risk factors is constructed;

[0009] The risk intensity flow of the risk factor is integrated within a preset time window to obtain the cumulative integral value of the risk factor;

[0010] Calculate the dynamic risk trajectory based on the cumulative integral value of the aforementioned risk factors;

[0011] The baseline scores of the standardized scales in the standardized data are fused with the dynamic risk trajectory to obtain the total risk trajectory;

[0012] The instantaneous risk probability is calculated based on the total risk trajectory, and a VTE risk warning is triggered based on an adaptive threshold.

[0013] Furthermore, risk factors include positional risk, anesthesia risk, hemostatic agent risk, and circulatory parameter risk; risk intensity flows include positional risk intensity flow, anesthesia risk intensity flow, hemostatic agent risk intensity flow, and circulatory parameter risk intensity flow; among which:

[0014] The postural risk intensity flow is the product of the postural risk kernel function and the postural indicator function; where the postural indicator function outputs a value of 1 when the current postural state is a specified postural state, and outputs a value of 0 otherwise. The specified postural state includes supine, semi-recumbent and sitting positions. The postural risk kernel function sets a preset risk coefficient for each specified postural state.

[0015] The calculation of the anesthesia risk intensity flow is based on the anesthesia risk kernel function, which is the ratio of the current dose to the reference dose multiplied by the preset drug type weight, and then multiplied by the exponential function with the product of the negative preset drug elimination constant and the drug administration time as the exponent. The drug administration time is the difference between the drug administration time and time t.

[0016] The risk intensity flow of hemostatic drugs is calculated through the hemostatic drug gain function. The hemostatic drug gain function accumulates and sums all medication events. The contribution of each medication is the preset medication gain amplitude multiplied by an exponential function with the natural constant e as the base and the negative time difference to the preset duration of action as the exponent, and then multiplied by a step function.

[0017] The cyclic parameter risk intensity flow is the product of the preset blood pressure weighting coefficient and the blood pressure term, plus the product of the preset heart rate weighting coefficient and the heart rate term. The blood pressure term is the absolute value of the difference between the current blood pressure and the preset baseline blood pressure divided by the preset baseline blood pressure. The heart rate term is the absolute value of the difference between the current heart rate and the preset baseline heart rate divided by the preset baseline heart rate.

[0018] Furthermore, the cumulative integral values ​​of risk factors include postural risk integral, anesthesia risk integral, hemostatic drug risk integral, and circulatory parameter risk integral, among which:

[0019] The method for calculating the postural risk integral is as follows: within the time interval from the current time t minus the preset time window length to the current time t, the postural risk intensity flow is continuously integrated.

[0020] The method for calculating the anesthesia risk score is as follows: within the time interval from the current time t minus the preset time window length to the current time t, the anesthesia risk kernel function value is continuously integrated;

[0021] The method for calculating the risk integral of hemostatic drugs is as follows: within the time interval from the current time t minus the preset time window length to the current time t, the gain function value of the hemostatic drug is continuously integrated;

[0022] The method for calculating the integral of the cyclic parameter is as follows: within the time interval from the current time t minus the preset time window length to the current time t, the cyclic parameter intensity flow value is continuously integrated.

[0023] Furthermore, the specific calculation methods for dynamic risk trajectories include:

[0024] For each risk factor, the cumulative integral value, the preset feature weight vector, and the time decay kernel function value corresponding to the risk factor are multiplied together to obtain an integral feature term. The integral feature term is then integrated within a time interval from 0 to the length of a preset time window.

[0025] The time decay kernel function is an exponential function with the natural constant as the base and the product of a negative preset decay parameter and the time window length as the exponent.

[0026] Furthermore, the total risk trajectory is the adaptive fusion factor multiplied by the baseline score, plus a dynamic risk term, where the dynamic risk term is the difference between 1 and the adaptive fusion factor multiplied by the dynamic risk trajectory. The adaptive fusion factor is calculated by multiplying the preset initial fusion weight by an exponential function with a negative preset adjustment parameter (base of the natural constant) and the data completeness index as the exponent, and then by a preset scenario adjustment factor. The data completeness index is calculated by summing the products of the preset weights of multiple risk factors and the availability indicator function, and then dividing by the sum of the preset weights of the multiple risk factors. The availability indicator function has a value of 1 when the risk intensity flow corresponding to the risk factor is not 0, and a value of 0 when it is 0.

[0027] Furthermore, methods for calculating instantaneous risk probability include:

[0028] The linear combination result is input into the sigmoid activation function; the linear combination result is a preset constant term plus a first risk coefficient term plus a second risk coefficient term; where the first risk coefficient term is the product of the preset first risk coefficient and the total risk trajectory value, and the second risk coefficient term is the product of the preset second risk coefficient and the time derivative of the total risk trajectory.

[0029] Furthermore, the adaptive threshold is the product of the preset adaptive adjustment factor and the sliding standard deviation of the total risk trajectory, plus the preset basic threshold. The sliding standard deviation of the total risk trajectory is calculated as follows: the square of the difference between the total risk trajectory value at each time point within the preset sliding window and the average value of the total risk trajectory within the sliding window is calculated, then the average value is calculated, and finally the square root is taken.

[0030] Furthermore, determining whether a VTE risk warning has been triggered includes:

[0031] Whether to trigger a VTE risk warning is determined by triggering conditions. Triggering conditions include continuous triggering and acute triggering. Continuous triggering is when the total risk trajectory value is greater than the adaptive threshold for a duration exceeding a preset duration threshold. Acute triggering is when the total risk trajectory change rate is greater than a preset risk change rate threshold and the total risk trajectory value approaches the adaptive threshold. The total risk trajectory change rate is the difference between the total risk trajectory value at the current time and the total risk trajectory value at the previous time divided by a preset time step. The total risk trajectory value approaches the adaptive threshold when the difference between the current total risk trajectory value and the adaptive threshold is less than a preset approach tolerance.

[0032] Furthermore, the raw data included body position angles, anesthetic drug dosages, drug type identification, blood transfusion and infusion rates, hemostatic drug dosage intensity, blood pressure, heart rate, blood oxygen saturation, and baseline scores of standardized scales.

[0033] This invention provides a VTE system execution system incorporating a standardized scale, used to store computer-readable instructions, which, when read, execute a VTE system execution method incorporating a standardized scale; the VTE system execution system incorporating a standardized scale includes:

[0034] The data processing module collects the patient's raw data and performs standardization processing on the raw data to obtain standardized data;

[0035] The risk intensity module constructs risk intensity streams for multiple risk factors based on the standardized data.

[0036] The cumulative integration module performs integration calculations on the risk intensity flow of the risk factors within a preset time window to obtain the cumulative integration value of the risk factors.

[0037] The risk trajectory module calculates the dynamic risk trajectory based on the cumulative integral value of the risk factors.

[0038] The risk fusion module fuses the baseline scores of the standardized scales in the standardized data with the dynamic risk trajectory to obtain the total risk trajectory;

[0039] The risk assessment module calculates the instantaneous risk probability based on the total risk trajectory and determines whether to trigger a VTE risk warning based on an adaptive threshold.

[0040] The beneficial effects of this invention are as follows: By constructing a four-dimensional risk intensity flow of body position, anesthesia, hemostatic drugs, and circulatory parameters, this invention achieves continuous quantitative monitoring of VTE risk factors, breaking through the limitations of traditional static assessment; by using event capsule technology to convert discrete monitoring data into continuous risk signals, it improves the accuracy and timeliness of risk identification; by innovatively introducing a time decay mechanism, it assigns higher weight to recent risk events, effectively balancing real-time performance with the retention of historical information; by using an exponential decay function to gradually reduce the impact of historical events, it ensures that the risk assessment results are both real-time and continuous, achieving intelligent fusion of standardized scale baseline scores and dynamic monitoring data, and dynamically adjusting the fusion weights according to data completeness and clinical scenarios, significantly improving the accuracy and adaptability of risk assessment without replacing existing Caprini or Padua scores.

[0041] Through a dual mechanism of continuous and acute triggering, early identification and timely warning of VTE risk are achieved. The adaptive threshold algorithm dynamically adjusts the warning threshold according to risk volatility, effectively reducing false alarm and false negative rates. Based on the instantaneous risk probability, a three-level risk stratification is performed to formulate differentiated intervention strategies. Specific intervention measures are generated in combination with patient characteristics, including precise matching of mechanical prevention, drug prevention, and monitoring programs. A complete closed-loop control system is constructed from risk identification, early warning push, intervention execution to effect evaluation. The traceability of the entire process is ensured through an audit chain, thereby improving the level of medical quality management.

[0042] The edge computing architecture enables low-latency early warning pushes, ensuring timely response at critical moments. High-frequency monitoring with a preset time step of 1 minute meets the needs of real-time monitoring during surgery. Through a capacity-aware scheduling algorithm, resource allocation is optimized based on minimizing the cost function, balancing resource costs and latency penalties to improve the efficiency of hospital resource utilization. A unified time base and data format ensure the comparability of various risk factors, and the trapezoidal rule numerical integration guarantees calculation accuracy and system reliability.

[0043] Through dynamic real-time monitoring and precise intervention, the early identification capability of VTE risk is significantly improved. Individualized intervention strategies effectively reduce the risk of VTE during and early postoperative periods. Multi-source data fusion and intelligent analysis reduce human judgment errors. Standardized risk assessment processes improve the scientificity and consistency of medical decisions. Automated risk monitoring reduces the workload of medical staff. Intelligent resource scheduling improves the efficiency of medical services and patient satisfaction.

[0044] This invention achieves an intelligent upgrade of VTE risk management through technological innovation. While maintaining compatibility with existing clinical practices, it significantly improves the accuracy, real-time nature, and individualization of risk assessment, and has important clinical application value and prospects for promotion. Attached Figure Description

[0045] Figure 1 This is a flowchart illustrating an example of a VTE system execution method combining a standardized scale according to the present invention.

[0046] Figure 2 This is an example diagram illustrating the risk intensity flow of multiple risk factors in a VTE system implementation method that combines standardized scales according to the present invention.

[0047] Figure 3 This is an example diagram illustrating the cumulative integral values ​​of risk factors obtained from a VTE system implementation method combining a standardized scale according to the present invention.

[0048] Figure 4 This is a module example diagram of a VTE system execution system that incorporates a standardized scale according to the present invention. Detailed Implementation

[0049] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0050] A method and system for implementing a VTE system that incorporates standardized scales, comprising the following embodiments:

[0051] Example 1:

[0052] A method for implementing a VTE system that incorporates standardized scales, such as Figure 1 As shown, it includes the following steps:

[0053] Step 100: Collect the patient's raw data and standardize the raw data to obtain standardized data.

[0054] Continuous data acquisition from various medical devices within the operating room was initiated to obtain raw data, including patient positioning, anesthetic drug dosage, drug type identification, minimum alveolar concentration, transfusion / infusion rate, hemostatic agent strength, blood pressure, heart rate, blood oxygen saturation, and baseline scores on standardized scales. Specifically, this includes:

[0055] The surgical positioning recording system was used to collect the patient's position angle and duration at time t. The anesthesia machine was used to collect parameters such as the dosage of anesthetic drugs, drug type identification, and minimum alveolar concentration at time t. The infusion pump was used to collect the blood transfusion rate and the hemostatic drug intensity at time t. The monitor was used to collect circulatory parameters such as blood pressure, heart rate, and blood oxygen saturation. The laboratory LIS system was used to collect discrete data such as D-dimer and coagulation function. The EMR system was used to obtain the patient's underlying diseases and baseline scores of standardized scales.

[0056] Because a single data source cannot fully reflect a patient's true risk status, traditional VTE risk assessment mainly relies on static standardized scale scores, lacking the ability to monitor dynamic changes during surgery in real time, resulting in insufficient timeliness and accuracy of risk assessment. This invention integrates multiple data sources, including surgical positioning recording systems, anesthesia machines, infusion pumps, monitors, laboratory LIS systems, and EMR systems, to achieve comprehensive coverage of VTE risk-related factors. The advantage of this multi-source data acquisition strategy is its ability to capture the combined impact of different physiological systems and medical interventions on VTE risk, providing a complete data foundation for subsequent dynamic risk assessment. The necessity of collecting positional angle data lies in the fact that changes in position directly affect lower extremity venous return, a fundamental risk factor for VTE formation. The necessity of collecting anesthetic drug-related parameters lies in the fact that anesthetic drugs significantly alter the patient's thrombosis risk by affecting vascular tone and coagulation function. The necessity of collecting hemostatic drug intensity data stems from the enhancing effect of hemostatic drugs on the coagulation system, which, while helpful for intraoperative hemostasis, also increases the risk of thrombosis. The collection of circulatory parameters is necessary because hemodynamic status directly reflects the physiological environment for thrombosis and is an important indicator for risk assessment. The collection of laboratory data is necessary because coagulation function indicators can objectively reflect changes in a patient's coagulation status. The collection of baseline scores from standardized scales is necessary to provide a reference benchmark for static risk assessment during dynamic monitoring, ensuring the clinical comparability of assessment results. Standardized scales include the Caprini score or Padua score.

[0057] The collected raw data undergoes standardization processing to obtain standardized data. Standardization includes steps such as timestamp calibration, unit conversion, semantic standardization, and conflict arbitration. Specifically, timestamp calibration unifies the timestamps of all data sources to the system's standard time; unit conversion standardizes the data units from different devices to the standard unit; semantic standardization converts data of different formats into a unified data model; and conflict arbitration outputs confidence intervals and conservative alternative values ​​for conflicting data values.

[0058] Because different medical devices and systems use different data formats, time bases, and units of measurement, direct fusion can lead to data inconsistencies and calculation errors. The advantage of standardized processing methods lies in eliminating format differences and time deviations between different data sources through a unified data preprocessing framework, ensuring that subsequent algorithms can perform accurate calculations based on a consistent data foundation. The necessity of timestamp calibration lies in ensuring the accurate correspondence of all risk events on the timeline, avoiding risk assessment errors caused by time deviations. The necessity of unit conversion lies in eliminating differences in measurement units between different devices, ensuring the accuracy of numerical calculations. The necessity of semantic standardization stems from the fact that different systems may use different representations of the same concept; a unified data model ensures the algorithm's correct understanding of the data's meaning. The necessity of conflict arbitration lies in handling the uncertainty when multiple data sources provide different values ​​for the same parameter; through the output of confidence intervals and conservative alternative values, the reliability and security of risk assessment are ensured.

[0059] This step outputs standardized data, providing a unified data foundation for subsequent processing.

[0060] Step 200: Based on the standardized data, construct a risk intensity stream for multiple risk factors, specifically as follows: Figure 2 As shown.

[0061] Based on the standardized data obtained in step 100, an event capsule is constructed and a multi-dimensional risk intensity stream is generated. The event capsule is a structured encapsulation of various intraoperative risk events, converting discrete monitoring data into continuous intensity stream signals to achieve real-time quantitative assessment of VTE risk factors. The entire intensity stream generation process follows a logical chain from physiological mechanisms to mathematical modeling, ensuring the interrelationship and clinical rationality among various risk factors. These risk factors include positional risk, anesthesia risk, hemostatic agent risk, and circulatory parameter risk; the risk intensity streams include positional risk intensity stream, anesthesia risk intensity stream, hemostatic agent risk intensity stream, and circulatory parameter risk intensity stream.

[0062] First, a postural risk intensity flow is constructed based on hemodynamic principles. Postural changes directly affect lower extremity venous return and are a fundamental risk factor for VTE formation. The postural risk intensity flow is calculated as the product of a postural risk kernel function and a postural indicator function. The postural indicator function outputs a value of 1 when the current postural position is a specified position, and 0 otherwise, ensuring the accuracy of risk assessment. A specified position refers to a predefined type of position with VTE risk significance, including three basic postural states: supine, semi-recumbent, and sitting. This classification is determined based on clinical physiological studies of the range of postural angles and their impact on lower extremity venous return. The postural risk kernel function sets differentiated preset risk coefficients according to different ranges of postural angles. When the postural angle is between 0 and 15 degrees, corresponding to the supine position, the preset risk coefficient is 0.2, a value determined based on clinical observations showing relatively low blood stasis in the supine position. When the body position angle is between 15 and 45 degrees, corresponding to a semi-recumbent position, the preset risk coefficient increases to 0.8, reflecting the physiological mechanism of increased lower limb blood flow resistance in the semi-recumbent position. When the body position angle is greater than 45 degrees, corresponding to a sitting position, the preset risk coefficient reaches the highest value of 1.0, reflecting the clinical evidence that lower limb blood flow stasis is most severe in the sitting position.

[0063] Building upon positional risk assessment, an anesthesia risk intensity flow is further constructed, which synergistically affects positional risk. Anesthetic drugs, by influencing vascular tone and coagulation function, work in conjunction with positional factors to increase VTE risk. The anesthesia risk intensity flow is calculated based on the anesthesia risk kernel function, which takes the drug dose at time t and the drug type identifier as input parameters, reflecting the core elements of the pharmacokinetic model. The anesthesia risk kernel function is calculated by multiplying a preset drug type weight by the ratio of the current dose to the reference dose, and then by an exponential function with a base of the natural constant e, a negative preset drug elimination constant, and the dosing time as the exponent. The dosing time is the difference between the dosing time and time t. This design fully considers the time decay characteristics of drugs. The preset drug type weights are differentiated according to the thrombosis risk of different drugs: propofol has a preset weight of 0.3, sevoflurane has a preset weight of 0.5, and muscle relaxants have a preset weight of 0.8. This weight gradient is determined based on clinical research data on the impact of various drugs on coagulation function. Reference doses are obtained from drug instructions and clinical guidelines to ensure standardized calculation benchmarks. Preset elimination constants reflect the rate of drug elimination in the body; the preset elimination constant for propofol is 0.3 h / h, for sevoflurane it is 0.8 h / h, and for muscle relaxants it is 0.2 h / h. These values ​​are set based on pharmacokinetic studies to ensure the timeliness of risk assessment.

[0064] Corresponding to the anesthesia risk, a hemostatic drug risk intensity flow was constructed, reflecting the enhancing effect of hemostatic drugs on the coagulation system. While the use of hemostatic drugs helps with intraoperative hemostasis, it also increases the risk of thrombosis, creating a compound effect with the aforementioned risk factors. The hemostatic drug risk intensity flow was calculated using a hemostatic drug gain function, which cumulatively sums all medication events. The contribution of each medication is calculated by multiplying a preset medication gain amplitude by an exponential function with a base of the natural constant e and a negative time difference to a preset duration of action as the exponent, and then by a step function. The preset medication gain amplitude reflects the difference in the intensity of different hemostatic drugs; the preset gain amplitude for tranexamic acid is 1.5, and for thrombin it is 2.0. These values ​​are determined based on clinical trial data on the enhancing effect of hemostatic drugs on coagulation function. The preset duration of action reflects the duration of the drug effect; the preset duration of action for tranexamic acid is 4 hours, and for thrombin it is 2 hours. These times are set based on the results of pharmacodynamic studies. The step function ensures that the effect only occurs after the medication is administered, thus guaranteeing the correctness of the time logic.

[0065] Based on the aforementioned risk factors, a circulatory parameter risk intensity flow is constructed, which reflects the impact of hemodynamic status on VTE risk. Abnormal changes in circulatory parameters are often the result of the combined effects of multiple risk factors and are also important indicators for VTE risk assessment. The calculation method for the circulatory parameter risk intensity flow is as follows: the product of the preset blood pressure weighting coefficient and (the absolute value of the difference between the current blood pressure and the preset baseline blood pressure) is divided by the preset baseline blood pressure, plus the product of the preset heart rate weighting coefficient and (the absolute value of the difference between the current heart rate and the preset baseline heart rate) is divided by the preset baseline heart rate. The preset baseline blood pressure refers to the standard blood pressure reference value of the patient in a stable state before surgery, including two components: systolic blood pressure and diastolic blood pressure. This baseline value is determined by averaging multiple blood pressure measurements taken by the patient within 24 hours before surgery. The default preset baseline value for systolic blood pressure is 120 mmHg, and the default preset baseline value for diastolic blood pressure is 80 mmHg. These values ​​are determined based on clinical medical standards for the normal blood pressure range in adults. The preset baseline heart rate refers to the standard heart rate reference value for the patient at rest before surgery. This baseline value is determined by averaging multiple heart rate measurements taken within 24 hours before surgery. The default preset baseline heart rate is 70 beats per minute, based on the physiological standard of normal heart rate range in adults. The default preset blood pressure weighting coefficient is 0.6, and the default preset heart rate weighting coefficient is 0.4. This weighting allocation is determined based on physiological studies of the impact of blood pressure and heart rate on the circulatory system, with blood pressure changes having a more significant impact on hemodynamics. The values ​​obtained through this calculation method can accurately reflect the degree of deviation of the circulatory system from the baseline state, providing key parameters for comprehensive risk assessment.

[0066] The generation of various risk intensity streams follows a unified time base and data format, ensuring the consistency and accuracy of subsequent integral calculations. The risk intensity streams for body position, anesthesia, hemostatic agents, and circulatory parameters together constitute a complete set of risk intensity streams. This set comprehensively covers the main influencing factors of intraoperative VTE risk, providing reliable basic data for subsequent duration integral calculations. This systematic intensity stream generation mechanism achieves an effective conversion from discrete monitoring data to continuous risk signals, laying a solid foundation for dynamic risk assessment.

[0067] Step 300: Integrate the risk intensity flow of the risk factor within a preset time window to obtain the cumulative integral value of the risk factor, specifically as follows... Figure 3 As shown.

[0068] Based on the multi-dimensional risk intensity stream generated in step 200, a unified framework for calculating duration and intensity integrals is constructed. This framework converts instantaneous risk signals into cumulative risk quantification by performing time-domain integration on the intensity streams of various risk factors. The integral calculation process follows a unified time window setting and numerical calculation method to ensure comparability among risk factors and consistency in subsequent fusion calculations. The cumulative integral values ​​of risk factors include positional risk integral, anesthesia risk integral, hemostatic drug risk integral, and circulatory parameter risk integral.

[0069] First, a unified time benchmark for integral calculation is determined, namely the preset time window length. This parameter controls the time range of the risk accumulation effect. The default value for the preset time window length is usually set to 2 to 4 hours, specifically 3 hours. This duration is determined based on clinical observation studies of the cumulative effects of VTE risk factors, effectively capturing the short- to medium-term risk accumulation effect while avoiding historical data interference caused by excessively long windows. The selection of the time window reflects the balance between sensitivity to recent events and retention of historical information in VTE risk assessment, ensuring that the integral results reflect both the current risk status and the persistent impact of risk factors.

[0070] Based on a unified time reference, the postural risk intensity flow is integrally calculated to obtain the postural risk integral. Specifically, the postural risk integral is calculated by continuously integrating the postural risk intensity flow over the time interval from the current time t minus a preset time window length to the current time t. The calculation of the postural risk integral fully considers the continuous nature of postural changes, transforming discrete postural states into a continuous cumulative risk quantity through integration, reflecting the time-cumulative effect of postural factors on lower extremity venous return. This integral result directly reflects the degree of VTE risk accumulation caused by postural factors within a specific time window, providing basic data for subsequent comprehensive risk assessment.

[0071] Corresponding to the positional risk integral, the anesthesia risk intensity flow is integrally calculated to obtain the anesthesia risk integral. Specifically, the anesthesia risk integral is calculated by continuously integrating the anesthesia risk kernel function value within the time interval from the current time t minus the preset time window length to the current time t. The calculation of the anesthesia risk integral considers the pharmacokinetic characteristics of anesthetic drugs, converting drug concentration changes into a cumulative risk through integration, reflecting the time-additive effect of anesthetic drugs on coagulation function and vascular tone. This integration process uses the same time window and calculation method as the positional risk integral, ensuring comparability between different risk factors and laying the foundation for subsequent multi-factor risk fusion.

[0072] Based on the anesthesia risk score, the risk intensity flow of hemostatic drugs is further integrated to obtain the hemostatic drug risk score. The calculation method for the hemostatic drug risk score is as follows: within the time interval from the current time t minus the preset time window length to the current time t, the gain function value of the hemostatic drug is continuously integrated. The calculation of the hemostatic drug risk score reflects the time-cumulative characteristics of the hemostatic drug's enhancing effect on the coagulation system. Through integration, the time decay process of the drug effect is converted into a cumulative risk. This integral calculation uses the same time benchmark and numerical method as the aforementioned risk factors, ensuring the consistency of hemostatic drug risk with other risk factors on the time scale, and providing standardized input data for comprehensive risk assessment.

[0073] The calculation of the circulatory parameter risk integral is a crucial component of the integral framework. The method involves continuously integrating the circulatory parameter risk intensity flow value over the time interval from the current time t minus a preset time window length to the current time t. This integration process converts instantaneous changes in hemodynamic parameters into a cumulative risk, reflecting the persistent impact of circulatory system abnormalities on VTE risk. The circulatory parameter risk integral uses a unified calculation framework with other risk factor integrals, ensuring the coordination and consistency of various physiological parameters in risk assessment.

[0074] All integration calculations employ the trapezoidal rule for numerical integration. This method approximates the integral value by adding the intensity flow values ​​at two adjacent time points, dividing by 2, multiplying by a preset time step, and then summing over all time steps. The default preset time step is set to 1 minute, determined based on real-time monitoring accuracy requirements and computational resource balance considerations. This ensures both the accuracy of the integration calculation and the real-time performance of the system. The application of the trapezoidal rule provides a unified numerical calculation standard for all risk factors, ensuring the accuracy and reliability of the integration results.

[0075] The trapezoidal rule is used for numerical integration because it achieves a good balance between computational accuracy and efficiency. Compared to the simple rectangular rule, it offers higher accuracy, and compared to complex higher-order integration methods, it provides better real-time performance. The advantages of the trapezoidal rule lie in its simple and stable algorithm, low computational complexity, and suitability for real-time system applications.

[0076] The unified integral calculation framework described above outputs the cumulative integral values ​​of risk factors. These integral values ​​quantify the cumulative effect of duration and intensity, providing standardized input data for subsequent time-weighted dynamic risk trajectory calculations. The consistency of the integral values ​​of each risk factor in terms of time scale, numerical precision, and calculation method ensures the scientific rigor and accuracy of the subsequent risk fusion and assessment process.

[0077] Step 400: Calculate the dynamic risk trajectory based on the cumulative integral value of the risk factors.

[0078] Based on the cumulative integral values ​​of risk factors obtained in step 300, a time-kernel-weighted dynamic risk trajectory calculation framework is constructed. This framework transforms discrete risk integral values ​​into continuous dynamic risk trajectory signals by introducing a time decay mechanism and feature weight optimization. The core concept of dynamic risk trajectory calculation is based on the time sensitivity characteristics of VTE risk factors, assigning higher weights to recent risk events while gradually reducing the impact of historical events through an exponential decay function, thereby achieving a precise quantitative assessment of the current VTE risk status.

[0079] First, a unified time-weighted calculation framework is established. This framework uses the risk integral values ​​of body position, anesthesia, hemostatic drugs, and circulatory parameters output from step 300 as basic input data. The calculation of the dynamic risk trajectory adopts a multi-factor weighted summation method. By comprehensively evaluating all risk factors, the contribution of each risk factor is determined by a specific time-weighted algorithm. The specific calculation process of this algorithm is as follows: the cumulative integral value corresponding to the risk factor, the preset feature weight vector, and the time decay kernel function value are multiplied to obtain the integral feature term. The integral feature term is then integrated within a time interval from 0 to a preset time window.

[0080] In the time-weighted calculation process, the pre-defined feature weight vector plays a crucial balancing role. This weight vector, obtained through machine learning training on large-scale clinical data, fully reflects the differentiated contributions of various risk factors to the probability of VTE. Specifically, the pre-defined feature weight for positional risk is set at 0.25, reflecting the important role of position as a fundamental risk factor and the sustained impact of prolonged improper positioning on lower extremity venous return. The pre-defined feature weight for anesthesia risk is set at 0.35, the highest weight value, reflecting the significant impact of anesthetic drugs on vascular tone and coagulation function, and highlighting the core role of anesthetic factors in intraoperative VTE risk formation. The pre-defined feature weight for hemostatic agent risk is set at 0.20, reflecting the importance of hemostatic drugs' enhancing effect on the coagulation system while considering their relative limitations and time-sensitive characteristics. The pre-defined feature weight for circulatory parameter risk is also set at 0.20, reflecting the important value of hemodynamic parameters as a comprehensive indicator in VTE risk assessment and demonstrating the direct impact of circulatory system status on thrombosis risk. The above weighting scheme was determined based on clinical epidemiological study data on the contribution of large-scale VTE occurrence, ensuring the scientific validity and rationality of each risk factor in the comprehensive assessment.

[0081] The reason for using machine learning training methods to determine the preset feature weight vector is that different risk factors contribute significantly differently to the probability of VTE occurrence. Traditional empirical weight allocation methods lack objectivity and accuracy, and cannot fully reflect the true impact of each risk factor. The advantage of machine learning training methods is that they can automatically learn the optimal weight combination for each risk factor based on large-scale clinical data, ensuring the objectivity and scientific nature of the weight allocation. The design with the highest weight for anesthesia risk reflects the central role of anesthetic drugs in the formation of intraoperative VTE risk, a conclusion supported by numerous clinical studies. The weight allocation for positional risk, hemostatic agent risk, and circulatory parameter risk fully considers the clinical importance and impact characteristics of each factor, ensuring the balance and accuracy of the comprehensive risk assessment.

[0082] The machine learning training steps for the feature weight vector include four stages: data preparation, feature engineering, model training, and weight extraction. The data preparation stage involves collecting a large-scale clinical dataset containing VTE occurrence outcomes. The dataset should include complete perioperative monitoring data from at least 10,000 patients, with the VTE incidence rate maintained within a reasonable range of 5% to 15%, ensuring a balance between positive and negative samples. The dataset needs to include positional monitoring data, anesthesia medication records, hemostatic drug usage, circulatory parameter monitoring values, and confirmed VTE outcomes. All data must undergo quality review and annotation verification by medical experts. The feature engineering stage preprocesses and extracts features from the raw monitoring data, transforming continuous monitoring data into risk intensity streams and integral features consistent with the algorithm of this invention, ensuring consistency between the training data and the actual application scenario. The model training stage uses a gradient boosting decision tree algorithm to model the VTE probability. This algorithm can automatically learn the importance weights of each feature. The training process uses 10-fold cross-validation to ensure the model's generalization ability. The model performance metrics require an AUC value of no less than 0.85, and sensitivity and specificity of no less than 0.80. In the weight extraction stage, the feature importance scores of each risk factor are extracted from the trained model. The final feature weight vector is obtained through normalization, ensuring that the sum of all weights equals 1.0, which meets the mathematical requirements of probability weights.

[0083] The time decay kernel function, as a core component of dynamic risk trajectory calculation, is designed to fully consider the time decay characteristics of VTE risk factors and actual clinical needs. This function is defined as an exponential function with a base of the natural constant e, an exponent of a negative preset decay parameter, and a time window length. It utilizes the mathematical properties of exponential decay to gradually weaken the impact of historical risk events. The default value of the preset decay parameter is set to 0.5 per hour. This parameter value is determined based on clinical observation studies of the time decay characteristics of VTE risk factors, ensuring both high sensitivity to recent risk events and appropriate amnesia of long-term historical events, thus achieving a balance between timeliness and effectiveness in risk assessment. The introduction of the time decay kernel function enables the dynamic risk trajectory to accurately reflect the temporal evolution of VTE risk, giving full attention to recently occurring high-risk events while avoiding excessive interference from historical low-risk states in current assessment results.

[0084] The aforementioned time-weighted calculation framework effectively transforms static risk integrals into dynamic risk trajectories. This transformation process fully reflects the time sensitivity and multi-factor comprehensiveness of VTE risk assessment. The physical significance of time-weighted calculation lies in establishing a risk weight distribution pattern centered on the current moment and gradually decreasing towards historical time. This allows the dynamic risk trajectory to keenly capture real-time changes in risk status while maintaining a reasonable consideration of the cumulative effects of historical risks. This design ensures that the risk assessment results are both real-time, continuous, and stable, providing a reliable data foundation for subsequent risk situation analysis and clinical decision support.

[0085] This step ultimately outputs a continuous dynamic risk trajectory signal, which comprehensively reflects the VTE risk change trend based on real-time monitoring data. This provides high-quality dynamic risk assessment data for the standardized scale adaptive fusion in the subsequent step 500, ensuring the timeliness and accuracy of the entire VTE risk assessment system.

[0086] Step 500: The baseline scores of the standardized scales in the standardized data are fused with the dynamic risk trajectory to obtain the total risk trajectory.

[0087] The total risk trajectory is the adaptive fusion factor multiplied by the baseline score, plus the dynamic risk term, which is the difference between 1 and the adaptive fusion factor multiplied by the dynamic risk trajectory.

[0088] The adaptive fusion factor is dynamically adjusted based on data completeness and clinical scenario. It is calculated by multiplying the preset initial fusion weight by an exponential function with the product of a negative preset adjustment parameter (base e) and the data completeness index as the exponent, and then multiplying by a preset scenario adjustment factor.

[0089] The default value for the initial fusion weight is typically set to 0.7, a value determined based on clinical validation studies balancing the effects of static assessment and dynamic monitoring. The data completeness index is calculated by summing the products of the default weights for all risk factors and the availability indicator function, then dividing by the sum of the default weights for all risk factors. Specifically, the default weights for positional risk (0.2), anesthesia risk (0.3), hemostatic agent risk (0.3), and circulatory parameter risk (0.2) are determined based on clinical studies of the importance of each risk factor in VTE risk assessment. The availability indicator function has a value of 1 when the risk intensity flow corresponding to a risk factor is not zero, and a value of 0 when it is zero. The default value for the preset adjustment parameter is 1.5, determined based on a systematic analysis of the impact of data completeness on fusion effectiveness, and is used to control the degree of influence of data completeness on the fusion weight.

[0090] The preset scenario adjustment factor is set according to different clinical scenarios: 1.0 for the preoperative assessment stage, 0.8 for the intraoperative monitoring stage, and 0.6 for the postoperative ICU stage. This allocation is determined based on medical practice experience regarding the reliability of dynamic monitoring data at different stages and the needs of clinical decision-making.

[0091] This step outputs the fused overall risk trajectory, combining the advantages of static assessment and dynamic monitoring.

[0092] Step 600: Calculate the instantaneous risk probability based on the total risk trajectory, and determine whether to trigger a VTE risk warning based on an adaptive threshold.

[0093] Based on the total risk trajectory obtained in step 500, calculate the instantaneous risk probability and adaptive threshold:

[0094] The instantaneous risk probability is calculated by inputting the linear combination result into the sigmoid activation function. The linear combination result is a preset constant term plus a first risk coefficient term plus a second risk coefficient term; where the first risk coefficient term is the product of the preset first risk coefficient and the total risk trajectory value, and the second risk coefficient term is the product of the preset second risk coefficient and the time derivative of the total risk trajectory. The preset constant term, the preset first risk coefficient, and the preset second risk coefficient are linear combination coefficients, with default values ​​of -2.0, 3.5, and 1.2, respectively. This parameter combination is determined through logistic regression analysis training on historical VTE event data.

[0095] The adaptive threshold is calculated by multiplying the preset adaptive adjustment factor by the sliding standard deviation of the total risk trajectory, and then adding the preset base threshold. The default value of the preset base threshold is typically set to 0.6, determined based on clinical evidence-based medicine studies of VTE risk thresholds. The default value of the preset adaptive adjustment factor ranges from 0.1 to 0.3, specifically 0.2, and is determined based on statistical analysis of the sensitivity of risk variability to threshold adjustment. The sliding standard deviation of the total risk trajectory is calculated by first squared the difference between the total risk trajectory value at each time point within the preset sliding window and the average total risk trajectory value within the sliding window, then calculating the average, and finally taking the square root. The total risk trajectory value refers to the value of the fused total risk trajectory output in step 500 at each time point, which comprehensively reflects the overall VTE risk level based on standardized scale baseline scores and dynamic monitoring data. The average total risk trajectory value within the sliding window is the arithmetic mean of all total risk trajectory values ​​within the preset sliding window time range, reflecting the average risk level over the recent period. The default value for the preset sliding window size is 30 time points, and the interval between adjacent time points is the preset time step. This window size is determined based on data analysis of the statistical characteristics of risk variability.

[0096] Whether a VTE risk warning is triggered is determined by triggering conditions, including persistent and acute triggers. A persistent trigger occurs when the total risk trajectory value exceeds an adaptive threshold for a duration exceeding a preset duration threshold. The default value of this preset duration threshold is typically 15 minutes, determined based on clinical guidelines for assessing the persistence of VTE risk. An acute trigger occurs when the rate of change of the total risk trajectory exceeds a preset risk rate of change threshold and the total risk trajectory value approaches the adaptive threshold. The default value of this preset risk rate of change threshold is 0.1 per minute, determined based on clinical observation studies of acute VTE risk change characteristics. The rate of change of the total risk trajectory refers to the magnitude of change in the total risk trajectory value per unit time. Specifically, it is calculated by dividing the difference between the current and previous total risk trajectory values ​​by a preset time step. This rate of change reflects the instantaneous speed and trend of VTE risk change. The total risk trajectory value approaching the adaptive threshold means that the difference between the current total risk trajectory value and the adaptive threshold is less than a preset approximation tolerance. The default value of this preset approximation tolerance is 10% of the adaptive threshold, determined based on clinical needs for the accuracy of risk threshold identification.

[0097] The peak window identification method involves identifying the peak moment and determining the peak window when the triggering condition is met. The peak moment refers to the point within a preset search time window after the triggering condition is met, where the total risk trajectory value reaches a local maximum. This moment marks a critical node where VTE risk reaches a stage of peak. The peak moment is identified by searching backwards for all total risk trajectory values ​​within a preset search time window, starting from the moment the triggering condition is met. The point with the largest value is identified as the peak moment. The default value for this preset search time window is 60 minutes, determined based on clinical statistical analysis of the timing characteristics of VTE risk peaks. The peak window ranges from the peak moment minus 30 minutes to the peak moment plus 60 minutes. This time range is determined based on clinical epidemiological studies of the persistence characteristics of VTE risk peaks, ensuring complete coverage of critical periods before and after the risk peak.

[0098] This step outputs the risk situation assessment results, including instantaneous risk probability, trigger status, and peak window.

[0099] Step 700: If a VTE risk warning is triggered, generate individualized intervention recommendations.

[0100] Risk stratification and intervention mapping are based on different ranges of instantaneous risk probability, with corresponding strategies developed for each range. Low risk is defined as an instantaneous risk probability less than a preset low-risk threshold of 0.3, and basic preventive measures are adopted for low risk. Medium risk is defined as an instantaneous risk probability between a preset lower limit of medium risk (0.3) and a preset upper limit of medium risk (0.7), and mechanical prophylaxis plus early mobilization is adopted. High risk is defined as an instantaneous risk probability greater than or equal to a preset high-risk threshold of 0.7, and pharmacological prophylaxis plus mechanical prophylaxis plus close monitoring is adopted. These risk stratification thresholds are determined based on international clinical guidelines and evidence-based medicine for VTE risk classification.

[0101] Intervention recommendations are generated based on risk levels and patient characteristics, producing specific intervention measures. Mechanical prophylaxis includes compression stockings and intermittent pneumatic compression devices; pharmacological prophylaxis includes anticoagulants such as low molecular weight heparin and warfarin; monitoring measures include D-dimer retesting and lower extremity vascular ultrasound examination.

[0102] Resource capacity-aware scheduling performs capacity-aware resource allocation by invoking the hospital resource management system. The selection of scheduling schemes is based on minimizing a cost function, which includes resource cost and delay penalty. Resource cost is the sum of the products of preset weights for all required resources and their costs. The preset weights for nursing staff are 0.4, medical equipment is 0.3, and medication resources are all 0.3. This weight allocation is determined based on operational management research on hospital resource allocation optimization. The delay penalty is a preset delay penalty coefficient multiplied by the delay time of the scheduling scheme. The default value of this preset delay penalty coefficient is 10 per minute, determined based on clinical quality management standards for the timeliness requirements of VTE intervention. The feasible scheduling scheme set includes all possible resource allocation schemes, and the required resource set includes all types of resources needed to complete the intervention. The preset delay penalty coefficient controls the sensitivity to delays.

[0103] This step outputs individualized intervention recommendations and actionable time scheduling plans.

[0104] Step 800: Push the intervention suggestions generated in step 700 through edge computing nodes with low latency.

[0105] The warning system sends risk warning information to responsible medical staff, including risk level and probability value, analysis of the contribution of major risk factors, recommended intervention measures and time windows.

[0106] The implementation confirmation form from medical staff is collected, and the implementation time, completion rate, patient response, and effect evaluation of the intervention measures are recorded.

[0107] The audit chain construction forms a complete audit chain that integrates early warning, intervention, and execution processes, ensuring traceability.

[0108] This step achieves closed-loop control from risk identification to intervention execution.

[0109] This embodiment utilizes an innovative technical paradigm of event-driven, time-weighted, and intensity- and duration-integrated methods to achieve dynamic real-time assessment and precise intervention of VTE risk without replacing existing standardized scales, demonstrating significant clinical application value.

[0110] Example 2

[0111] See Figure 4 As shown, a VTE system execution system incorporating a standardized scale is provided. This system stores computer-readable instructions, which, when read, execute the aforementioned VTE system execution method incorporating a standardized scale. The VTE system execution system incorporating a standardized scale includes:

[0112] The data processing module 101 collects the patient's raw data and performs standardization processing on the raw data to obtain standardized data;

[0113] Risk intensity module 102 constructs a risk intensity stream for multiple risk factors based on the standardized data;

[0114] The cumulative integration module 103 performs integration calculations on the risk intensity flow of the risk factor within a preset time window to obtain the cumulative integration value of the risk factor.

[0115] The risk trajectory module 104 calculates the dynamic risk trajectory based on the cumulative integral value of the risk factors.

[0116] The risk fusion module 105 fuses the baseline scores of the standardized scales in the standardized data with the dynamic risk trajectory to obtain the total risk trajectory.

[0117] The risk assessment module 106 calculates the instantaneous risk probability based on the total risk trajectory and determines whether to trigger a VTE risk warning based on an adaptive threshold.

[0118] The embodiments of the present invention have been described above. However, the embodiments are not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make more equivalent embodiments under the guidance of the present embodiments, and all of them are within the protection scope of the present embodiments.

Claims

1. A method for implementing a VTE system using standardized scales, characterized in that, Includes the following steps: Collect the patient's raw data and standardize the raw data to obtain standardized data; Based on the standardized data, a risk intensity stream for multiple risk factors is constructed; Risk factors include positional risk, anesthesia risk, hemostatic agent risk, and circulatory parameter risk; risk intensity flows include positional risk intensity flow, anesthesia risk intensity flow, hemostatic agent risk intensity flow, and circulatory parameter risk intensity flow. The postural risk intensity flow is the product of the postural risk kernel function and the postural indicator function; where the postural indicator function outputs a value of 1 when the current postural state is a specified postural state, and outputs a value of 0 otherwise. The specified postural state includes supine, semi-recumbent and sitting positions. The postural risk kernel function sets a preset risk coefficient for each specified postural state. The calculation of the anesthesia risk intensity flow is based on the anesthesia risk kernel function, which is the ratio of the current dose to the reference dose multiplied by the preset drug type weight, and then multiplied by the exponential function with the product of the negative preset drug elimination constant and the drug administration time as the exponent. The drug administration time is the difference between the drug administration time and time t. The risk intensity flow of hemostatic drugs is calculated through the hemostatic drug gain function. The hemostatic drug gain function accumulates and sums all medication events. The contribution of each medication is the preset medication gain amplitude multiplied by an exponential function with the natural constant e as the base and the negative time difference to the preset duration of action as the exponent, and then multiplied by a step function. The cyclic parameter risk intensity flow is the product of the preset blood pressure weighting coefficient and the blood pressure term, plus the product of the preset heart rate weighting coefficient and the heart rate term. The blood pressure term is the absolute value of the difference between the current blood pressure and the preset baseline blood pressure divided by the preset baseline blood pressure. The heart rate term is the absolute value of the difference between the current heart rate and the preset baseline heart rate divided by the preset baseline heart rate. The risk intensity flow of the risk factor is integrated within a preset time window to obtain the cumulative integral value of the risk factor; Calculate the dynamic risk trajectory based on the cumulative integral value of the aforementioned risk factors; The baseline scores of the standardized scales in the standardized data are fused with the dynamic risk trajectory to obtain the total risk trajectory; The instantaneous risk probability is calculated based on the total risk trajectory, and a VTE risk warning is triggered based on the adaptive threshold.

2. The VTE system execution method combining standardized scales according to claim 1, characterized in that, The cumulative score of risk factors includes the risk score for body position, the risk score for anesthesia, the risk score for hemostatic agents, and the risk score for circulatory parameters, among which: The method for calculating the postural risk integral is as follows: within the time interval from the current time t minus the preset time window length to the current time t, the postural risk intensity flow is continuously integrated. The method for calculating the anesthesia risk score is as follows: within the time interval from the current time t minus the preset time window length to the current time t, the anesthesia risk kernel function value is continuously integrated; The method for calculating the risk integral of hemostatic drugs is as follows: within the time interval from the current time t minus the preset time window length to the current time t, the value of the hemostatic drug gain function is continuously integrated; The method for calculating the integral of the cyclic parameter is as follows: within the time interval from the current time t minus the preset time window length to the current time t, the cyclic parameter intensity flow value is continuously integrated.

3. The VTE system execution method combining standardized scales according to claim 2, characterized in that, Methods for calculating dynamic risk trajectories include: For each risk factor, the cumulative integral value, the preset feature weight vector, and the time decay kernel function value corresponding to the risk factor are multiplied together to obtain an integral feature term. The integral feature term is then integrated within a time interval from 0 to the length of a preset time window. The time decay kernel function is an exponential function with the natural constant as the base and the product of a negative preset decay parameter and the time window length as the exponent.

4. The VTE system execution method combining standardized scales according to claim 1, characterized in that, The total risk trajectory is the adaptive fusion factor multiplied by the baseline score, plus a dynamic risk term; the dynamic risk term is the difference between 1 and the adaptive fusion factor multiplied by the dynamic risk trajectory; where the adaptive fusion factor is the preset initial fusion weight multiplied by an exponential function with a negative preset adjustment parameter as the base of the natural constant and the data completeness index as the exponent, and then multiplied by a preset scenario adjustment factor; the calculation method of the data completeness index includes summing the products of the preset weights of multiple risk factors and the availability indicator function, and then dividing by the sum of the preset weights of multiple risk factors; the availability indicator function has a value of 1 when the risk intensity flow corresponding to the risk factor is not 0, and a value of 0 when it is 0.

5. The VTE system execution method combining a standardized scale according to claim 4, characterized in that, Methods for calculating instantaneous risk probability include: The linear combination result is input into the sigmoid activation function; the linear combination result is a preset constant term plus a first risk coefficient term plus a second risk coefficient term; where the first risk coefficient term is the product of the preset first risk coefficient and the total risk trajectory value, and the second risk coefficient term is the product of the preset second risk coefficient and the time derivative of the total risk trajectory.

6. The VTE system execution method combining a standardized scale according to claim 5, characterized in that, The adaptive threshold is the product of the preset adaptive adjustment factor and the sliding standard deviation of the total risk trajectory, plus the preset base threshold. The sliding standard deviation of the total risk trajectory is calculated as follows: calculate the square of the difference between the total risk trajectory value at each time point within the preset sliding window and the average value of the total risk trajectory within the sliding window, then calculate the average value, and finally take the square root.

7. The VTE system execution method combining a standardized scale according to claim 6, characterized in that, Determining whether a VTE risk warning has been triggered includes: Whether to trigger a VTE risk warning is determined by triggering conditions. Triggering conditions include continuous triggering and acute triggering. Continuous triggering is when the total risk trajectory value is greater than the adaptive threshold for a duration exceeding a preset duration threshold. Acute triggering is when the total risk trajectory change rate is greater than a preset risk change rate threshold and the total risk trajectory value approaches the adaptive threshold. The total risk trajectory change rate is the difference between the total risk trajectory value at the current time and the total risk trajectory value at the previous time divided by a preset time step. The total risk trajectory value approaches the adaptive threshold when the difference between the current total risk trajectory value and the adaptive threshold is less than a preset approach tolerance.

8. The VTE system execution method combining standardized scales according to claim 1, characterized in that, The raw data included body position angle, anesthetic drug dosage, drug type identification, blood transfusion rate, hemostatic drug dosage, blood pressure, heart rate, blood oxygen saturation, and baseline scores of standardized scales.

9. A VTE system execution system incorporating standardized scales, characterized in that, For storing computer-readable instructions, and when the computer-readable instructions are read, executing the VTE system execution method incorporating a standardized scale as described in any one of claims 1-8; the VTE system execution system incorporating a standardized scale includes: The data processing module collects the patient's raw data and performs standardization processing on the raw data to obtain standardized data; The risk intensity module constructs risk intensity streams for multiple risk factors based on the standardized data. The cumulative integration module performs integration calculations on the risk intensity flow of the risk factors within a preset time window to obtain the cumulative integration value of the risk factors. The risk trajectory module calculates the dynamic risk trajectory based on the cumulative integral value of the risk factors. The risk fusion module fuses the baseline scores of the standardized scales in the standardized data with the dynamic risk trajectory to obtain the total risk trajectory; The risk assessment module calculates the instantaneous risk probability based on the total risk trajectory and determines whether to trigger a VTE risk warning based on an adaptive threshold.

Images