A method and system for predicting the risk of intradialytic hypotension

Abstract

The present application relates to the medical health technical field, especially relate to a kind of method and system for predicting low blood pressure risk in dialysis, the system includes: data acquisition module, for obtaining dialysis equipment operating parameter and patient sign parameter;Two-way coupling modeling module, for based on dialysis equipment operating parameter and patient sign parameter, construct the two-way coupling model between patient and dialysis equipment;Closed loop feature extraction module, for based on two-way coupling model, assess the closed loop response characteristic parameter of human body and dialysis equipment;Risk prediction module, for based on the dynamic change of closed loop response characteristic parameter, construct low blood pressure risk prediction model;Risk warning module, for realizing low blood pressure risk real-time early warning in dialysis process.The present application quantitatively depicts the influence of dialysis equipment operating abnormality on patient hemodynamics and the disturbance feedback of patient blood volume change to equipment fluid control loop, effectively improves the accuracy and reliability of low blood pressure risk early warning.

Description

Technical Field

[0001] This invention relates to the field of medical and health technology, and in particular to a method and system for predicting the risk of low blood pressure during dialysis. Background Technology

[0002] Hemodialysis is one of the main treatment methods for patients with end-stage renal disease. Because the kidneys have lost their original metabolic function, metabolic waste products cannot be excreted normally. To remove metabolic waste and harmful substances from the blood, dialysis equipment is used to remove metabolic waste and excess water from the patient's body to maintain electrolyte and acid-base balance. However, hemodialysis is often accompanied by various complications, among which hypotension is relatively common. Hypotension during dialysis can not only cause acute discomfort symptoms such as dizziness, sweating, nausea, and convulsions, but in severe cases, it can even lead to critical events such as cerebral ischemia and loss of consciousness, forcing the interruption of dialysis and affecting the adequacy of dialysis and the patient's long-term prognosis.

[0003] Existing methods for predicting the risk of hypotension during dialysis mainly rely on one-way causal models or static threshold determination methods, such as analyzing the operating status of dialysis equipment based on its operating parameters, or making trend judgments based on the patient's physiological parameters. These methods have the following problems:

[0004] (1) Focusing only on the equipment or patient-only factors without considering the dynamic coupling characteristics between the dialysis equipment and the human body, in the actual dialysis process, the fluid control of the dialysis equipment will directly affect the patient's hemodynamic state, while changes in the patient's blood volume and vascular access resistance will in turn affect the control loop of the equipment, making the system exhibit bidirectional nonlinear feedback characteristics.

[0005] (2) Traditionally, a fixed threshold is used to determine hypotension. However, the physical characteristics and hypotension tolerance of different patients vary significantly. The fixed threshold is difficult to reflect individual differences and is prone to misjudgment or omission.

[0006] Therefore, existing technologies cannot capture the reverse disturbance feedback generated by changes in the patient's physiological state on the internal fluid control loop of the dialysis device, as well as the patient's tolerance to hypotension, resulting in insufficient accuracy in predicting hypotension events and insufficient risk control capabilities during dialysis. Summary of the Invention

[0007] To overcome the defects and shortcomings of existing technologies, this invention provides a method and system for predicting the risk of hypotension during dialysis. By quantitatively characterizing the impact of abnormal operation of dialysis equipment on patient hemodynamics and the disturbance feedback of changes in patient blood volume on the equipment fluid control loop, the accuracy and reliability of hypotension risk early warning are effectively improved.

[0008] To achieve the above objectives, the present invention adopts the following technical solution:

[0009] In a first aspect, the present invention provides a system for predicting the risk of hypotension during dialysis, comprising:

[0010] The data acquisition module is used to acquire the operating parameters of the dialysis equipment and the patient's vital signs.

[0011] The bidirectional coupling modeling module is used to construct a bidirectional coupling model between the patient and the dialysis equipment based on the operating parameters of the dialysis equipment and the patient's vital signs. The bidirectional coupling model includes a forward response sub-model, a backward mapping sub-model, and a mutual information gain sub-model.

[0012] The closed-loop feature extraction module is used to evaluate the closed-loop response feature parameters of the human body and dialysis equipment based on the bidirectional coupling model. The closed-loop response feature parameters include the stability margin of the control loop and the spectral coherence of the blood pump drive current signal and the arterial pressure signal of the vascular access.

[0013] The risk prediction module is used to build a hypotension risk prediction model based on the dynamic changes of closed-loop response characteristic parameters. By analyzing the decay trend of control loop stability margin and the abnormal increase of spectral coherence, it predicts the risk of hypotension during dialysis.

[0014] The risk warning module is used to dynamically adjust the hypotension risk threshold based on the patient's pre-dialysis vital signs and generate a hypotension risk warning signal by combining the hypotension risk prediction results, so as to realize real-time warning of hypotension risk during dialysis.

[0015] Furthermore, the specific steps for constructing the positive response sub-model include:

[0016] Obtain the operating parameters of the dialysis equipment and the vital signs parameters of the patients, and construct the operating parameter sequence of the dialysis equipment and the vital signs parameter sequence of the patients, respectively;

[0017] The dialysis equipment operating parameter sequences and patient vital sign parameter sequences were preprocessed, including detrending and standardization.

[0018] Using the pre-processed sequence of dialysis equipment operating parameters as exogenous input variables and the pre-processed sequence of patient vital signs parameters as response variables, a causal response relationship between abnormal dialysis equipment operation and changes in patient blood pressure fluctuations and vascular access resistance was established based on the RBF-ARX model.

[0019] The parameters of the RBF-ARX model are divided into linear and nonlinear parameters using the SNPOM nonlinear parameter optimization method. The linear parameters are optimized by the least squares method and the nonlinear parameters are optimized by the Levenberg-Marquardt method, thus obtaining a positive response sub-model of the effect of dialysis equipment state changes on the human blood pressure system.

[0020] Furthermore, the specific steps for constructing the inverse mapping sub-model include:

[0021] Hemodynamic disturbance characteristic parameters are extracted based on patient vital signs parameters. These parameters include the rate of change of blood volume, the amplitude of arteriovenous pressure difference, and its spectral energy distribution.

[0022] The feedback relationship between patient hemodynamic disturbances and the fluid control loop of dialysis equipment is established based on the state-space model. The characteristic parameters of hemodynamic disturbances are used as input variables, and the flow regulation response and transmembrane pressure control deviation of dialysis equipment are used as output variables.

[0023] By recursively estimating the state vector and covariance matrix of the state-space model using the extended Kalman filter algorithm, and combining this with a multi-step prediction residual correction strategy to improve estimation accuracy, an inverse mapping sub-model that dynamically represents the impact of changes in patient state on the stability of equipment fluid control is obtained.

[0024] Furthermore, the specific execution steps of the mutual information gain sub-model include:

[0025] The dynamic response signal of the blood pressure system output by the positive response sub-model and the fluid control disturbance response signal output by the inverse mapping sub-model are obtained, and the two sets of signals are synchronously registered and time-frequency resampled to eliminate the influence of sampling frequency difference and phase drift.

[0026] The mutual information gain of two sets of signals at different time scales is calculated by using a sliding time window mutual information analysis method, and a time delay correction term is introduced to generate a time delay mutual information matrix, which is used to characterize the coupling strength and time delay characteristics between changes in the patient's physiological state and the dynamic response of the device control.

[0027] Furthermore, the specific execution steps of the closed-loop feature extraction module include:

[0028] Based on the time delay mutual information matrix, the eigenvalues ​​corresponding to the dominant coupling mode are extracted. The ratio of the imaginary part to the real part of the eigenvalue in the complex plane is used as the stability margin of the control loop to characterize the energy exchange stability between the human hemodynamic system and the fluid control system of the dialysis equipment.

[0029] Based on the time delay mutual information matrix, key delay time is extracted as a correction value, and time-shifted registration is performed on the blood pump drive current signal and the arterial pressure signal of the vascular pathway.

[0030] The spectral coherence of the time-shift corrected blood pump drive current signal and the vascular access arterial pressure signal was calculated using the weighted coherence spectrum estimation method, which was used to characterize the dynamic coupling relationship between blood pump load changes and vascular access impedance response.

[0031] Furthermore, the specific execution steps of the risk prediction module include:

[0032] Closed-loop response trend parameters are extracted based on closed-loop response characteristic parameters combined with a sliding time window. The closed-loop response trend parameters include stability margin decay rate and spectral coherence increase rate.

[0033] A hypotension risk prediction model based on long short-term memory network was constructed, which takes risk-sensitive feature sequences as input and outputs the hypotension risk coefficient during dialysis.

[0034] Furthermore, the specific execution steps of the risk warning module include:

[0035] Obtain the patient's static vital signs before the current dialysis session and muscle activity signals during dialysis, and calculate the rate of decrease in muscle activity.

[0036] The static vital signs parameters and the rate of decline in muscle activity are input into the Bayesian hierarchical regression model. The posterior distribution of the model is updated in real time through variational inference to obtain the hypotension tolerance threshold that reflects the patient's current dialysis status.

[0037] The low blood pressure risk coefficient output by the risk prediction module is compared with the low blood pressure tolerance threshold. When the low blood pressure risk coefficient is greater than the low blood pressure tolerance threshold, a low blood pressure risk warning is issued.

[0038] Secondly, the present invention provides a method for predicting the risk of hypotension during dialysis, comprising:

[0039] Obtain the operating parameters of the dialysis equipment and the patient's vital signs;

[0040] A bidirectional coupling model between patients and dialysis equipment is constructed based on the operating parameters of dialysis equipment and the vital signs parameters of patients. The bidirectional coupling model includes a positive response sub-model, a reverse mapping sub-model, and a mutual information gain sub-model.

[0041] The closed-loop response characteristic parameters of the human body and dialysis equipment are evaluated based on a two-way coupling model. The closed-loop response characteristic parameters include the stability margin of the control loop and the spectral coherence of the blood pump drive current signal and the arterial pressure signal of the vascular access.

[0042] A hypotension risk prediction model is constructed based on the dynamic changes of closed-loop response characteristic parameters. By analyzing the decay trend of control loop stability margin and the abnormal increase of spectral coherence, the risk of hypotension during dialysis is predicted.

[0043] The hypotension risk threshold is dynamically adjusted based on the patient's pre-dialysis vital signs and the hypotension risk prediction results are combined to generate a hypotension risk warning signal, thereby achieving real-time warning of hypotension risk during dialysis.

[0044] Thirdly, the present invention provides an electronic device comprising: a processor and a memory, wherein the memory stores a computer program that can be called by the processor, and the processor executes a method for predicting the risk of low blood pressure during dialysis by calling the computer program stored in the memory.

[0045] Fourthly, the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform a method for predicting the risk of low blood pressure during dialysis.

[0046] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0047] (1) By constructing a bidirectional coupling model between the patient and the dialysis equipment, this invention quantitatively characterizes the impact of abnormal equipment operation on the patient's hemodynamics and the disturbance feedback of changes in the patient's blood volume on the equipment's fluid control loop, thereby realizing closed-loop dynamic analysis of equipment status changes and human physiological responses, which significantly improves the accuracy of hypotension risk assessment.

[0048] (2) This invention extracts the closed-loop response characteristics between the operation of the dialysis device and the physiological response of the human body, quantifies the stability margin of the control loop and the spectral coherence of the blood pump drive current signal and the arterial pressure signal of the vascular pathway, and identifies the potential instability trend of the system before the blood pressure drops suddenly, so as to realize early warning of the risk of hypotension.

[0049] (3) This invention constructs an adaptive risk warning mechanism based on hypotension tolerance, and assesses the patient's hypotension tolerance threshold in the current dialysis state by combining the patient's physical signs before the current dialysis. This solves the problem that the fixed threshold method cannot reflect individual differences among patients, and achieves precise and individualized hypotension risk warning. Attached Figure Description

[0050] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0051] Figure 1 This is a schematic diagram of the structure of a system for predicting the risk of hypotension during dialysis, provided by an embodiment of the present invention;

[0052] Figure 2 This is a schematic diagram of the dialysis device structure provided in an embodiment of the present invention;

[0053] Figure 3 This is a flowchart illustrating a method for predicting the risk of hypotension during dialysis, provided by an embodiment of the present invention.

[0054] Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0055] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0056] Please see Figure 1 , Figure 1 This is a schematic diagram of a system for predicting the risk of hypotension during dialysis, provided by an embodiment of the present invention, comprising:

[0057] The data acquisition module 210 is used to acquire the operating parameters of the dialysis equipment and the vital signs parameters of the patient. The operating parameters of the dialysis equipment include ultrafiltration rate, dialysate flow rate, blood flow rate and conductivity, and the vital signs parameters of the patient include blood volume change rate, blood pressure fluctuation rate, heart rate variability, arterial pressure and venous pressure of the vascular access.

[0058] The bidirectional coupling modeling module 220 is used to construct a bidirectional coupling model between the patient and the dialysis equipment based on the operating parameters of the dialysis equipment and the patient's vital signs. The bidirectional coupling model includes a forward response sub-model, a backward mapping sub-model, and a mutual information gain sub-model.

[0059] The closed-loop feature extraction module 230 is used to evaluate the closed-loop response feature parameters of the human body and dialysis equipment based on the bidirectional coupling model. The closed-loop response feature parameters include the stability margin of the control loop and the spectral coherence of the blood pump drive current signal and the arterial pressure signal of the vascular access.

[0060] The risk prediction module 240 is used to construct a hypotension risk prediction model based on the dynamic changes of closed-loop response characteristic parameters. By analyzing the decay trend of control loop stability margin and the abnormal increase of spectral coherence, the risk of hypotension during dialysis is predicted.

[0061] The risk warning module 250 is used to dynamically adjust the low blood pressure risk threshold based on the patient's pre-dialysis vital signs and generate a low blood pressure risk warning signal by combining the low blood pressure risk prediction results, so as to realize real-time warning of low blood pressure risk during dialysis.

[0062] In this embodiment of the invention, the data acquisition module 210 is used to acquire the operating parameters of the dialysis equipment and the patient's vital signs parameters. The operating parameters of the dialysis equipment include ultrafiltration rate, dialysate flow rate, blood flow rate and conductivity. The patient's vital signs parameters include blood volume change rate, blood pressure fluctuation rate, heart rate variability, arterial pressure and venous pressure in the vascular access. The operating parameters of the dialysis equipment are acquired in real time through the built-in sensors and control system of the dialysis machine, while the patient's vital signs parameters are acquired through a blood pressure monitor, a blood volume monitoring system and a heart rate variability analysis device.

[0063] In this embodiment of the invention, the bidirectional coupling modeling module 220 is used to construct a bidirectional coupling model between the patient and the dialysis equipment based on the dialysis equipment operating parameters and the patient's vital signs parameters. The bidirectional coupling model includes a positive response sub-model, a reverse mapping sub-model, and a mutual information gain sub-model.

[0064] By constructing a positive response sub-model to characterize the positive impact mechanism of changes in dialysis equipment operating status on changes in the patient's blood pressure system and vascular access resistance, quantitative modeling of hypotension-inducing factors can be achieved from the equipment end. Please refer to [link to relevant documentation]. Figure 2 , Figure 2 This is a schematic diagram of the dialysis device structure provided in an embodiment of the present invention. The dialysis device mainly includes a blood circulation monitoring system and a dialysate monitoring system. The blood circulation monitoring system includes a blood pump, a hepatic cord pump, an arterial and venous pressure monitoring system, and an air monitoring system. The dialysate monitoring system includes a temperature control system, a solution preparation system, a degassing system, a conductivity monitoring system, an ultrafiltration monitoring system, and a blood leakage monitoring system. During dialysis, the operating status of the dialysis device directly determines the flow balance and pressure distribution in the blood circulation system. When the device's operating status malfunctions, such as an excessively high ultrafiltration rate, transmembrane pressure fluctuations, or excessive blood pump suction intensity, it can cause dynamic responses in the patient's arterial and venous pressure pathways, leading to blood pressure fluctuations and even the risk of hypotension. Therefore, by constructing a causal response relationship model between the dialysis device's operating parameters and the patient's blood pressure response, explicit modeling of the impact of device-side anomalies on the human hemodynamic system can be achieved, providing a foundational forward propagation feature for subsequent bidirectional coupling modeling and risk prediction. The specific steps for constructing the positive response sub-model include:

[0065] Obtain the operating parameters of the dialysis equipment and the vital signs parameters of the patients, and construct the operating parameter sequence of the dialysis equipment and the vital signs parameter sequence of the patients, respectively;

[0066] The dialysis equipment operating parameter sequences and patient vital sign parameter sequences were preprocessed, including detrending and standardization.

[0067] The preprocessed sequence of dialysis equipment operating parameters was used as an exogenous input variable, and the preprocessed sequence of patient vital signs parameters was used as a response variable. The causal response relationship between abnormal dialysis equipment operation and changes in patient blood pressure fluctuations and vascular access resistance was established based on the RBF-ARX model. The RBF-ARX model is an autoregressive model with exogenous input.

[0068] The parameters of the RBF-ARX model are divided into linear and nonlinear parameters by the SNPOM nonlinear parameter optimization method. The linear parameters are optimized by the least squares method and the nonlinear parameters are optimized by the Levenberg-Marquardt method, so as to obtain the positive response sub-model of the effect of dialysis equipment status changes on the human blood pressure system.

[0069] In actual dialysis, changes in patient blood volume, blood pressure fluctuations, and dynamic changes in vascular access resistance all have feedback effects on the fluid control system of the dialysis equipment. For example, a sudden drop in blood volume may lead to an imbalance between blood pump suction and cardiac perfusion, causing abnormal blood pump load and transmembrane pressure control deviation. By constructing an inverse mapping sub-model to quantitatively characterize the feedback relationship between the human body and the equipment, the system can identify potential equipment instability risks, providing an important information source for hypotension risk prediction. The specific steps for constructing the inverse mapping sub-model include:

[0070] Hemodynamic disturbance characteristic parameters are extracted based on patient vital signs parameters. These parameters include the rate of change of blood volume, the amplitude of arteriovenous pressure difference, and its spectral energy distribution.

[0071] The feedback relationship between patient hemodynamic disturbances and the fluid control loop of dialysis equipment is established based on the state-space model. The characteristic parameters of hemodynamic disturbances are used as input variables, and the flow regulation response and transmembrane pressure control deviation of dialysis equipment are used as output variables.

[0072] The state vector and covariance matrix of the state-space model are recursively estimated by the extended Kalman filter algorithm, and the estimation accuracy is improved by combining a multi-step prediction residual correction strategy. This yields an inverse mapping sub-model that dynamically represents the impact of changes in patient state on the stability of equipment fluid control.

[0073] The main purpose of constructing the mutual information gain sub-model is to quantify the dynamic coupling relationship and time lag characteristics between changes in the patient's physiological state and the control response of the dialysis equipment. This provides a reliable basis for subsequent closed-loop feature extraction and hypotension risk prediction. Furthermore, mutual information gain analysis, based on the statistical dependence of the signal rather than fixed model assumptions, can adapt to different patients, different hemodynamic characteristics, and various dialysis equipment operating conditions, effectively improving the model's generalization ability. The specific execution steps of the mutual information gain sub-model include:

[0074] The dynamic response signal of the blood pressure system output by the positive response sub-model and the fluid control disturbance response signal output by the inverse mapping sub-model are obtained, and the two sets of signals are synchronously registered and time-frequency resampled to eliminate the influence of sampling frequency difference and phase drift.

[0075] The mutual information gain of two sets of signals at different time scales is calculated by using a sliding time window mutual information analysis method, and a time delay correction term is introduced to generate a time delay mutual information matrix, which is used to characterize the coupling strength and time delay characteristics between changes in the patient's physiological state and the dynamic response of the device control.

[0076] In this embodiment of the invention, the closed-loop feature extraction module 230 is used to evaluate the closed-loop response feature parameters of the human body and dialysis equipment based on a bidirectional coupling model. The closed-loop response feature parameters include the stability margin of the control loop and the spectral coherence of the blood pump drive current signal and the arterial pressure signal of the vascular access.

[0077] The closed-loop feature extraction module can quantitatively characterize the closed-loop dynamic characteristics between the patient's hemodynamic system and the dialysis equipment's fluid control system, providing closed-loop indicators and characteristic parameters for hypotension risk prediction. Through further analysis of the output signal of the bidirectional coupling model, it can identify the dominant coupling mode, critical delay time, and energy exchange stability in the closed-loop system between the human body and the equipment, achieving high-precision monitoring of hypotension risk during dialysis. The specific execution steps of the closed-loop feature extraction module include:

[0078] Based on the time delay mutual information matrix, the eigenvalues ​​corresponding to the dominant coupling mode are extracted. The ratio of the imaginary part to the real part of the eigenvalue in the complex plane is used as the stability margin of the control loop to characterize the energy exchange stability between the human hemodynamic system and the fluid control system of the dialysis equipment.

[0079] Based on the time delay mutual information matrix, key delay time is extracted as a correction value, and time-shifted registration is performed on the blood pump drive current signal and the arterial pressure signal of the vascular pathway.

[0080] The spectral coherence of the time-shift corrected blood pump drive current signal and the vascular access arterial pressure signal was calculated using the weighted coherence spectrum estimation method, which was used to characterize the dynamic coupling relationship between blood pump load changes and vascular access impedance response.

[0081] In this embodiment of the invention, the risk prediction module 240 is used to construct a hypotension risk prediction model based on the dynamic changes of closed-loop response characteristic parameters. By analyzing the decay trend of control loop stability margin and the abnormal increase of spectral coherence, the risk of hypotension during dialysis is predicted.

[0082] The risk prediction module transforms the quantitative closed-loop response parameters obtained by the closed-loop feature extraction module into real-time predictive indicators of patient hypotension risk, enabling early warning of sudden drops in cardiac blood pressure during dialysis. Through time-series analysis and deep learning modeling of the closed-loop system's characteristic parameters, complex human-device interactions are quantified into interpretable risk coefficients. A hypotension risk prediction model is constructed using a Long Short-Term Memory (LSTM) network, which effectively learns the time dependence and nonlinear dynamic characteristics of the closed-loop system. The LSTM network retains past state information and captures long-term dependencies, accurately predicting the patient's blood pressure trends in the near future. The specific execution steps of the risk prediction module include:

[0083] Closed-loop response trend parameters are extracted based on closed-loop response characteristic parameters combined with a sliding time window. The closed-loop response trend parameters include stability margin decay rate and spectral coherence increase rate.

[0084] A hypotension risk prediction model based on long short-term memory network was constructed, which takes risk-sensitive feature sequences as input and outputs the hypotension risk coefficient during dialysis.

[0085] In this embodiment of the invention, the risk warning module 250 is used to dynamically adjust the low blood pressure risk threshold based on the patient's pre-dialysis vital signs parameters, and generate a low blood pressure risk warning signal in combination with the low blood pressure risk prediction results, so as to realize real-time warning of low blood pressure risk during dialysis.

[0086] The risk warning module dynamically assesses the patient's physiological state before each dialysis session and adjusts the hypotension risk threshold in real time, without relying on fixed blood pressure standards. This accurately reflects changes in the patient's individual hypotension tolerance, enabling personalized and dynamic risk warnings. It effectively improves the accuracy of predicting hypotension events during dialysis and enhances the ability to intervene promptly. The specific execution steps of the risk warning module include...

[0087] Acquire the patient's static vital signs parameters before the current dialysis session and muscle activity signals during dialysis and calculate the rate of decrease in muscle activity. The static vital signs parameters include resting blood pressure, blood volume estimate, heart rate variability and vascular compliance index.

[0088] The static vital signs parameters and the rate of decline in muscle activity are input into the Bayesian hierarchical regression model. The posterior distribution of the model is updated in real time through variational inference to obtain the hypotension tolerance threshold that reflects the patient's current dialysis status.

[0089] The low blood pressure risk coefficient output by the risk prediction module is compared with the low blood pressure tolerance threshold. When the low blood pressure risk coefficient is greater than the low blood pressure tolerance threshold, a low blood pressure risk warning is issued.

[0090] Please see Figure 3, Figure 3 This is a schematic diagram of the overall process of a method for predicting the risk of hypotension during dialysis provided by an embodiment of the present invention, which specifically includes the following steps:

[0091] S100: Obtain the operating parameters of the dialysis equipment and the patient's vital signs parameters. The operating parameters of the dialysis equipment include ultrafiltration rate, dialysate flow rate, blood flow rate and conductivity. The patient's vital signs parameters include blood volume change rate, blood pressure fluctuation rate, heart rate variability, and arterial and venous pressure in the vascular access.

[0092] S200. Based on the dialysis equipment operating parameters and patient vital signs parameters, a bidirectional coupling model between the patient and the dialysis equipment is constructed. The bidirectional coupling model includes a forward response sub-model, a backward mapping sub-model, and a mutual information gain sub-model. The specific steps for constructing the forward response sub-model include:

[0093] Obtain the operating parameters of the dialysis equipment and the vital signs parameters of the patients, and construct the operating parameter sequence of the dialysis equipment and the vital signs parameter sequence of the patients, respectively;

[0094] The dialysis equipment operating parameter sequences and patient vital sign parameter sequences were preprocessed, including detrending and standardization.

[0095] Using the pre-processed sequence of dialysis equipment operating parameters as exogenous input variables and the pre-processed sequence of patient vital signs parameters as response variables, a causal response relationship between abnormal dialysis equipment operation and changes in patient blood pressure fluctuations and vascular access resistance was established based on the RBF-ARX model.

[0096] The parameters of the RBF-ARX model are divided into linear and nonlinear parameters by the SNPOM nonlinear parameter optimization method. The linear parameters are optimized by the least squares method and the nonlinear parameters are optimized by the Levenberg-Marquardt method, so as to obtain the positive response sub-model of the effect of dialysis equipment status changes on the human blood pressure system.

[0097] The specific steps for constructing the inverse mapping sub-model include:

[0098] Hemodynamic disturbance characteristic parameters are extracted based on patient vital signs parameters. These parameters include the rate of change of blood volume, the amplitude of arteriovenous pressure difference, and its spectral energy distribution.

[0099] The feedback relationship between patient hemodynamic disturbances and the fluid control loop of dialysis equipment is established based on the state-space model. The characteristic parameters of hemodynamic disturbances are used as input variables, and the flow regulation response and transmembrane pressure control deviation of dialysis equipment are used as output variables.

[0100] The state vector and covariance matrix of the state-space model are recursively estimated by the extended Kalman filter algorithm, and the estimation accuracy is improved by combining a multi-step prediction residual correction strategy. This yields an inverse mapping sub-model that dynamically represents the impact of changes in patient state on the stability of equipment fluid control.

[0101] The specific execution steps of the mutual information gain sub-model include:

[0102] The dynamic response signal of the blood pressure system output by the positive response sub-model and the fluid control disturbance response signal output by the inverse mapping sub-model are obtained, and the two sets of signals are synchronously registered and time-frequency resampled to eliminate the influence of sampling frequency difference and phase drift.

[0103] The mutual information gain of two sets of signals at different time scales is calculated by using a sliding time window mutual information analysis method, and a time delay correction term is introduced to generate a time delay mutual information matrix, which is used to characterize the coupling strength and time delay characteristics between changes in the patient's physiological state and the dynamic response of the device control.

[0104] S300. Based on a bidirectional coupling model, the closed-loop response characteristic parameters of the human body and dialysis equipment are evaluated. These parameters include the control loop stability margin and the spectral coherence of the blood pump drive current signal and the arterial pressure signal in the vascular access, including:

[0105] Based on the time delay mutual information matrix, the eigenvalues ​​corresponding to the dominant coupling mode are extracted. The ratio of the imaginary part to the real part of the eigenvalue in the complex plane is used as the stability margin of the control loop to characterize the energy exchange stability between the human hemodynamic system and the fluid control system of the dialysis equipment.

[0106] Based on the time delay mutual information matrix, key delay time is extracted as a correction value, and time-shifted registration is performed on the blood pump drive current signal and the arterial pressure signal of the vascular pathway.

[0107] The spectral coherence of the time-shift corrected blood pump drive current signal and the vascular access arterial pressure signal was calculated using the weighted coherence spectrum estimation method, which was used to characterize the dynamic coupling relationship between blood pump load changes and vascular access impedance response.

[0108] S400: A hypotension risk prediction model is constructed based on the dynamic changes of closed-loop response characteristic parameters. By analyzing the decay trend of control loop stability margin and the abnormal increase in spectral coherence, the risk of hypotension during dialysis is predicted, including:

[0109] Closed-loop response trend parameters are extracted based on closed-loop response characteristic parameters combined with a sliding time window. The closed-loop response trend parameters include stability margin decay rate and spectral coherence increase rate.

[0110] A hypotension risk prediction model based on long short-term memory network was constructed, which takes risk-sensitive feature sequences as input and outputs the hypotension risk coefficient during dialysis.

[0111] S500 dynamically adjusts the hypotension risk threshold based on the patient's pre-dialysis vital signs and generates a hypotension risk warning signal based on the hypotension risk prediction results, achieving real-time warning of hypotension risk during dialysis, including:

[0112] Obtain the patient's static vital signs before the current dialysis session and muscle activity signals during dialysis, and calculate the rate of decrease in muscle activity.

[0113] The static vital signs parameters and the rate of decline in muscle activity are input into the Bayesian hierarchical regression model. The posterior distribution of the model is updated in real time through variational inference to obtain the hypotension tolerance threshold that reflects the patient's current dialysis status.

[0114] The low blood pressure risk coefficient output by the risk prediction module is compared with the low blood pressure tolerance threshold. When the low blood pressure risk coefficient is greater than the low blood pressure tolerance threshold, a low blood pressure risk warning is issued.

[0115] The parameters and steps in the above-described method for predicting the risk of hypotension during dialysis can be referred to the parameters and steps of each unit module in the above-described embodiment of a system for predicting the risk of hypotension during dialysis, which are used to implement the corresponding functions, and will not be repeated here.

[0116] Please refer to Figure 4 The embodiments of the present invention also provide an electronic device 300, including a memory 320 for storing a computer program 322; and a processor 310 for executing the computer program 322 to implement a method for predicting the risk of hypotension during dialysis as described in any of the above embodiments.

[0117] It should be noted that... Figure 4 This is a structural diagram of an electronic device 300 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of the invention.

[0118] Specifically, the electronic device 300 may include at least one processor 310, at least one memory 320, a power supply 330, a communication interface 340, an input / output interface 350, and a communication bus 360. The memory 320 stores a computer program 322, which is loaded and executed by the processor 310 to implement relevant steps in the method for predicting the risk of low blood pressure during dialysis disclosed in any of the foregoing embodiments. Alternatively, the electronic device 300 in the embodiments of the present invention may specifically be an electronic computer.

[0119] In embodiments of the present invention, the power supply 330 is used to provide operating voltage for each hardware device on the electronic device 300; the communication interface 340 can create a data transmission channel between the electronic device 300 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of the present invention, and is not specifically limited here; the input / output interface 350 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0120] In addition, the memory 320, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 321, computer program 322, etc., and the storage method can be temporary storage or permanent storage.

[0121] The operating system 321 is used to manage and control the various hardware devices and computer programs on the electronic device 300, and may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program 322 capable of performing a method for predicting the risk of hypotension during dialysis, which is executed by the electronic device 300 as disclosed in any of the foregoing embodiments, the computer program 322 may further include a computer program 322 capable of performing other specific tasks.

[0122] Embodiments of the present invention also provide a computer-readable storage medium for storing a computer program 322, which, when executed by a processor 310, implements a method for predicting the risk of hypotension during dialysis as described in any of the above embodiments.

[0123] For example, computer-readable storage media can be read-only memory (ROM), random access memory (RAM), compact disc read-only memory (CD-ROM), magnetic tape, floppy disk, and optical data storage devices.

[0124] The terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0125] The above description is merely a preferred embodiment of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this application is not limited to the technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions claimed in this invention.

Claims

1. A system for predicting the risk of hypotension during dialysis, characterized in that, include: The data acquisition module is used to acquire the operating parameters of the dialysis equipment and the patient's vital signs. The bidirectional coupling modeling module is used to construct a bidirectional coupling model between the patient and the dialysis equipment based on the operating parameters of the dialysis equipment and the patient's vital signs. The bidirectional coupling model includes a forward response sub-model, a backward mapping sub-model, and a mutual information gain sub-model. The closed-loop feature extraction module is used to evaluate the closed-loop response feature parameters of the human body and dialysis equipment based on the bidirectional coupling model. The closed-loop response feature parameters include the stability margin of the control loop and the spectral coherence of the blood pump drive current signal and the arterial pressure signal of the vascular access. The risk prediction module is used to build a hypotension risk prediction model based on the dynamic changes of closed-loop response characteristic parameters. By analyzing the decay trend of control loop stability margin and the abnormal increase of spectral coherence, it predicts the risk of hypotension during dialysis. The risk warning module is used to dynamically adjust the hypotension risk threshold based on the patient's pre-dialysis vital signs and generate a hypotension risk warning signal in combination with the hypotension risk prediction results, so as to realize real-time warning of hypotension risk during dialysis. The specific steps for constructing the positive response sub-model include: Obtain the operating parameters of the dialysis equipment and the vital signs parameters of the patients, and construct the operating parameter sequence of the dialysis equipment and the vital signs parameter sequence of the patients, respectively; The dialysis equipment operating parameter sequences and patient vital sign parameter sequences were preprocessed, including detrending and standardization. Using the pre-processed sequence of dialysis equipment operating parameters as exogenous input variables and the pre-processed sequence of patient vital signs parameters as response variables, a causal response relationship between abnormal dialysis equipment operation and changes in patient blood pressure fluctuations and vascular access resistance was established based on the RBF-ARX model. The parameters of the RBF-ARX model are divided into linear and nonlinear parameters by the SNPOM nonlinear parameter optimization method. The linear parameters are optimized by the least squares method and the nonlinear parameters are optimized by the Levenberg-Marquardt method, so as to obtain the positive response sub-model of the effect of dialysis equipment status changes on the human blood pressure system. The specific steps for constructing the inverse mapping sub-model include: Hemodynamic disturbance characteristic parameters are extracted based on patient vital signs parameters. These parameters include the rate of change of blood volume, the amplitude of arteriovenous pressure difference, and its spectral energy distribution. The feedback relationship between patient hemodynamic disturbances and the fluid control loop of dialysis equipment is established based on the state-space model. The characteristic parameters of hemodynamic disturbances are used as input variables, and the flow regulation response and transmembrane pressure control deviation of dialysis equipment are used as output variables. The state vector and covariance matrix of the state-space model are recursively estimated by the extended Kalman filter algorithm, and the estimation accuracy is improved by combining a multi-step prediction residual correction strategy. This yields an inverse mapping sub-model that dynamically represents the impact of changes in patient state on the stability of equipment fluid control. The specific execution steps of the mutual information gain sub-model include: The dynamic response signal of the blood pressure system output by the positive response sub-model and the fluid control disturbance response signal output by the inverse mapping sub-model are obtained, and the two sets of signals are synchronously registered and time-frequency resampled to eliminate the influence of sampling frequency difference and phase drift. The mutual information gain of two sets of signals at different time scales is calculated by using the mutual information analysis method with a sliding time window and a time delay correction term is introduced to generate a time delay mutual information matrix, which is used to characterize the coupling strength and time delay characteristics between changes in the patient's physiological state and the dynamic response of the device control. The specific execution steps of the closed-loop feature extraction module include: Based on the time delay mutual information matrix, the eigenvalues ​​corresponding to the dominant coupling mode are extracted. The ratio of the imaginary part to the real part of the eigenvalue in the complex plane is used as the stability margin of the control loop to characterize the energy exchange stability between the human hemodynamic system and the fluid control system of the dialysis equipment. Based on the time delay mutual information matrix, key delay time is extracted as a correction value, and time-shifted registration is performed on the blood pump drive current signal and the arterial pressure signal of the vascular pathway. The spectral coherence of the time-shift corrected blood pump drive current signal and the vascular access arterial pressure signal was calculated using the weighted coherence spectrum estimation method, which was used to characterize the dynamic coupling relationship between blood pump load changes and vascular access impedance response.

2. The system for predicting the risk of hypotension during dialysis according to claim 1, characterized in that, The specific execution steps of the risk prediction module include: Closed-loop response trend parameters are extracted based on closed-loop response characteristic parameters combined with a sliding time window. The closed-loop response trend parameters include stability margin decay rate and spectral coherence increase rate. A hypotension risk prediction model based on long short-term memory network was constructed, which takes risk-sensitive feature sequences as input and outputs the hypotension risk coefficient during dialysis.

3. The system for predicting the risk of hypotension during dialysis according to claim 1, characterized in that, The specific execution steps of the risk warning module include: Obtain the patient's static vital signs before the current dialysis session and muscle activity signals during dialysis, and calculate the rate of decrease in muscle activity. The static vital signs parameters and the rate of decline in muscle activity are input into the Bayesian hierarchical regression model. The posterior distribution of the model is updated in real time through variational inference to obtain the hypotension tolerance threshold that reflects the patient's current dialysis status. The low blood pressure risk coefficient output by the risk prediction module is compared with the low blood pressure tolerance threshold. When the low blood pressure risk coefficient is greater than the low blood pressure tolerance threshold, a low blood pressure risk warning is issued.

4. A method for predicting the risk of hypotension during dialysis, applied to a system for predicting the risk of hypotension during dialysis according to any one of claims 1-3, characterized in that, The method includes: Obtain the operating parameters of the dialysis equipment and the patient's vital signs; A bidirectional coupling model between patients and dialysis equipment is constructed based on the operating parameters of dialysis equipment and the vital signs parameters of patients. The bidirectional coupling model includes a positive response sub-model, a reverse mapping sub-model, and a mutual information gain sub-model. The closed-loop response characteristic parameters of the human body and dialysis equipment are evaluated based on a two-way coupling model. The closed-loop response characteristic parameters include the stability margin of the control loop and the spectral coherence of the blood pump drive current signal and the arterial pressure signal of the vascular access. A hypotension risk prediction model is constructed based on the dynamic changes of closed-loop response characteristic parameters. By analyzing the decay trend of control loop stability margin and the abnormal increase of spectral coherence, the risk of hypotension during dialysis is predicted. The hypotension risk threshold is dynamically adjusted based on the patient's pre-dialysis vital signs and status parameters, and a hypotension risk warning signal is generated in combination with the hypotension risk prediction results to achieve real-time warning of hypotension risk during dialysis. The specific steps for constructing the positive response sub-model include: Obtain the operating parameters of the dialysis equipment and the vital signs parameters of the patients, and construct the operating parameter sequence of the dialysis equipment and the vital signs parameter sequence of the patients, respectively; The dialysis equipment operating parameter sequences and patient vital sign parameter sequences were preprocessed, including detrending and standardization. Using the pre-processed sequence of dialysis equipment operating parameters as exogenous input variables and the pre-processed sequence of patient vital signs parameters as response variables, a causal response relationship between abnormal dialysis equipment operation and changes in patient blood pressure fluctuations and vascular access resistance was established based on the RBF-ARX model. The parameters of the RBF-ARX model are divided into linear and nonlinear parameters by the SNPOM nonlinear parameter optimization method. The linear parameters are optimized by the least squares method and the nonlinear parameters are optimized by the Levenberg-Marquardt method, so as to obtain the positive response sub-model of the effect of dialysis equipment status changes on the human blood pressure system. The specific steps for constructing the inverse mapping sub-model include: Hemodynamic disturbance characteristic parameters are extracted based on patient vital signs parameters. These parameters include the rate of change of blood volume, the amplitude of arteriovenous pressure difference, and its spectral energy distribution. The feedback relationship between patient hemodynamic disturbances and the fluid control loop of dialysis equipment is established based on the state-space model. The characteristic parameters of hemodynamic disturbances are used as input variables, and the flow regulation response and transmembrane pressure control deviation of dialysis equipment are used as output variables. The state vector and covariance matrix of the state-space model are recursively estimated by the extended Kalman filter algorithm, and the estimation accuracy is improved by combining a multi-step prediction residual correction strategy. This yields an inverse mapping sub-model that dynamically represents the impact of changes in patient state on the stability of equipment fluid control. The specific execution steps of the mutual information gain sub-model include: The dynamic response signal of the blood pressure system output by the positive response sub-model and the fluid control disturbance response signal output by the inverse mapping sub-model are obtained, and the two sets of signals are synchronously registered and time-frequency resampled to eliminate the influence of sampling frequency difference and phase drift. The mutual information gain of two sets of signals at different time scales is calculated by using the mutual information analysis method with a sliding time window and a time delay correction term is introduced to generate a time delay mutual information matrix, which is used to characterize the coupling strength and time delay characteristics between changes in the patient's physiological state and the dynamic response of the device control. The specific steps for evaluating the closed-loop response characteristic parameters of the human body and dialysis equipment include: Based on the time delay mutual information matrix, the eigenvalues ​​corresponding to the dominant coupling mode are extracted. The ratio of the imaginary part to the real part of the eigenvalue in the complex plane is used as the stability margin of the control loop to characterize the energy exchange stability between the human hemodynamic system and the fluid control system of the dialysis equipment. Based on the time delay mutual information matrix, key delay time is extracted as a correction value, and time-shifted registration is performed on the blood pump drive current signal and the arterial pressure signal of the vascular pathway. The spectral coherence of the time-shift corrected blood pump drive current signal and the vascular access arterial pressure signal was calculated using the weighted coherence spectrum estimation method, which was used to characterize the dynamic coupling relationship between blood pump load changes and vascular access impedance response.

5. An electronic device, comprising: A processor and a memory, wherein the memory stores a computer program that can be called by the processor; characterized in that the processor executes, as described in claim 4, a method for predicting the risk of hypotension during dialysis by calling the computer program stored in the memory.

6. A computer-readable storage medium, characterized in that, The device stores instructions that, when executed on a computer, cause the computer to perform a method for predicting the risk of low blood pressure during dialysis as described in claim 4.

Images