Dialysis scheme adaptive interface system and method based on individual parameters of patient

Abstract

The invention relates to the technical field of dialysis process monitoring, in particular to a dialysis scheme self-adaptive interface system and method based on individual parameters of a patient, according to physiological state data and blood purification process data of the patient, the blood purification abnormal condition of the patient under the current dialysis scheme is analyzed; the acute complication risk of the patient under the current dialysis scheme is predicted; in combination with the physiological state data and the dialysis strategy data of the patient, analyzing the re-adaptation abnormal risk of purified blood and the body of the patient after dialysis under the current dialysis scheme; the assessment of the dialysis treatment safety risk of the patient under the current dialysis scheme is realized; therefore, self-adaptive adjustment and early warning can be carried out on the dialysis scheme of the patient. The reliability of dialysis treatment is improved, the treatment safety of a patient is guaranteed, and the occurrence risk of acute adverse events is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of dialysis process monitoring, and in particular to a dialysis scheme adaptive interface system and method based on patient individual parameters. BACKGROUND

[0002] End-stage renal disease, as a major chronic disease threatening human health, its treatment process is highly dependent on regular hemodialysis to maintain life. Dialysis treatment replaces the damaged kidney function by artificial means, removes metabolic waste, removes excess water and corrects electrolyte imbalance, and the accuracy and safety of the treatment scheme are directly related to the quality of life and long-term prognosis of patients. In the long-term dialysis process, patients will not only be continuously affected by toxin accumulation and excessive fluid load, gradually destroying the homeostasis, leading to increased cardiovascular burden; and dialysis treatment itself as a non-physiological intervention method, the setting of its treatment parameters will directly cause the dramatic changes of blood composition and volume, leading to acute hemodynamic instability in patients during dialysis, chronic complications during the interdialytic period, and long-term survival rate decline; if the treatment scheme cannot be dynamically adjusted according to the individual physiological state of the patient, it may cause serious complications such as intradialytic hypotension, muscle spasm and heart failure, not only increasing the patient's pain, but also endangering life safety.

[0003] However, the prior art does not dynamically correlate and analyze the real-time collected physiological parameters and dialysis strategy settings when monitoring and warning the dialysis process, for example, when the patient's blood volume rapidly decreases, if the synergistic effect of the current ultrafiltration rate setting and dialysate composition is not considered, it is difficult to accurately judge the evolution trend of hypotension risk and the degree of physical tolerance, thus leading to misjudgment of treatment safety. At the same time, the prior art also lacks comprehensive consideration of the purification of blood after dialysis and the re-adaptation process of the body, making it difficult for the monitoring results to fully reflect the long-term impact of the treatment scheme.

[0004] In order to solve these problems, the present application designs a dialysis scheme adaptive interface system and method based on patient individual parameters. SUMMARY

[0005] In order to overcome the defects and deficiencies of the prior art, the present application provides a dialysis scheme adaptive interface system and method based on patient individual parameters, which constructs blood purification anomaly analysis, acute complication risk prediction and re-adaptation abnormality risk assessment models by real-time collection of patient physiological data and dialysis process data, and realizes dynamic warning and adaptive adjustment of the dialysis scheme based on the comprehensive safety risk assessment results.

[0006] In order to achieve the above purpose, the present application adopts the following technical solutions: In a first aspect, the embodiments of the present application provide a dialysis scheme self-adaptive interface method based on individual parameters of a patient, comprising the following steps: S1, collecting physiological state data and blood purification process data of the patient, and acquiring dialysis strategy data under a current dialysis scheme; S2, analyzing abnormal conditions of blood purification of the patient under the current dialysis scheme in combination with the physiological state data and the blood purification process data of the patient, and predicting an acute complication risk of the patient under the current dialysis scheme according to an analysis result of the abnormal conditions of blood purification; S3, analyzing abnormal risk of readaptation of purified blood to the patient's body after dialysis under the current dialysis scheme in combination with the physiological state data of the patient and the dialysis strategy data; S4, evaluating a safety risk of dialysis treatment of the patient under the current dialysis scheme based on a prediction result of the acute complication risk of the patient under the current dialysis scheme and an analysis result of the abnormal risk of readaptation of the purified blood to the patient's body after dialysis under the current dialysis scheme; S5, self-adaptively adjusting and warning the dialysis scheme of the patient based on an evaluation result of the safety risk of dialysis treatment of the patient under the current dialysis scheme.

[0007] As an embodiment of the present application, the step S2 of analyzing the abnormal conditions of blood purification of the patient under the current dialysis scheme in combination with the physiological state data and the blood purification process data of the patient comprises the following specific steps: S21, extracting a blood volume relative change rate and an arterial end blood pressure from the physiological state data, and extracting a venous end urea concentration, a dialysate flow rate and a blood flow rate from the blood purification process data; S22, calculating an effective Kt / V value of the current dialysis scheme based on a single-chamber variable-volume urea kinetics model, taking the venous end urea concentration, the blood flow rate, the dialysate flow rate and a dry body weight of the patient as inputs; S23, calculating a ratio of an absolute difference between the effective Kt / V value and a preset target Kt / V value to the preset target Kt / V value as a purification efficiency deviation; S24, calculating a linear slope of the blood volume relative change rate in a current monitoring period as a blood volume decline slope, and taking a ratio of the blood volume decline slope to a preset safety maximum slope as a blood volume decline risk; S25, performing weighted summation on the purification efficiency deviation and the blood volume decline risk to obtain the abnormal conditions of blood purification of the patient under the current dialysis scheme.

[0008] As an embodiment of the present application, the step S2 of predicting the acute complication risk of the patient under the current dialysis scheme according to the analysis result of the abnormal conditions of blood purification comprises the following specific steps: S26, a gradient boosting tree machine learning prediction model is constructed, and the blood purification abnormality of the patient under the current dialysis scheme, the average heart rate of the patient in the current monitoring period, the average blood pressure variability, and the diabetes history in the patient static file are taken as input features of the prediction model; wherein the gradient boosting tree machine learning prediction model outputs the first risk probability of occurrence of intradialytic hypotension in the next monitoring period and the second risk probability of occurrence of muscle spasm; S27, taking the maximum value of the first risk probability and the second risk probability as the acute complication risk of the patient under the current dialysis scheme.

[0009] As an embodiment of the present application, in step S3, the physiological state data and the dialysis strategy data of the patient are combined to analyze the abnormal risk of the purified blood after dialysis and the patient's body re-adaptation under the current dialysis scheme, including the following specific steps: S31, the average blood pressure and the average heart rate of the patient in the current monitoring period in the physiological state data and the low-to-high frequency power ratio in the heart rate variability are obtained, and the preset total ultrafiltration volume, the treatment remaining time, the dialysate sodium concentration and the dialyzer urea clearance rate in the dialysis strategy data are obtained; S32, a multiple linear regression prediction model is established, and the average blood pressure, the average heart rate, the preset total ultrafiltration volume, the treatment remaining time and the dialysate sodium concentration are taken as input variables to predict the blood pressure recovery value of the patient in the monitoring period after dialysis; S33, a classification model based on random forest is constructed, and the average heart rate, the preset total ultrafiltration volume, the treatment remaining time, and the low-to-high frequency power ratio in the heart rate variability are taken as input features to predict the state probability of occurrence of severe post-dialysis fatigue; S34, the absolute difference between the predicted blood pressure recovery value and the individualized target blood pressure is calculated, and the ratio of the absolute difference to the individualized target blood pressure is taken as the cardiovascular re-adaptation risk of the patient; S35, the state probability of occurrence of severe post-dialysis fatigue is taken as the fatigue risk of the patient; the cardiovascular re-adaptation risk and the fatigue risk are weighted and summed to obtain the abnormal risk of the purified blood after dialysis and the patient's body re-adaptation under the current dialysis scheme.

[0010] As an embodiment of the present application, in step S33, the state probability of occurrence of severe post-dialysis fatigue is predicted, including the following specific contents: S331, historical dialysis records of multiple patients are extracted from a historical medical database, each record containing physiological state data during dialysis, dialysis strategy data and post-dialysis self-evaluation results of fatigue; according to the standardized fatigue evaluation scale, the records whose post-dialysis self-evaluation results reach or exceed the preset severity threshold are marked as positive samples, i.e. "occur severe fatigue", and are assigned a value of one, and the remaining records are marked as negative samples and assigned a value of zero; S332, extracting objective features related to fatigue risk from the historical dialysis records, the objective features including but not limited to: average heart rate during dialysis, heart rate variability low frequency to high frequency power ratio within a certain time window before the end of dialysis, percentage of total ultrafiltration volume of this dialysis relative to dry body weight, total duration of dialysis treatment, urea clearance rate of dialyzer, and solute clearance index estimated according to urea kinetics model; S333, using the objective features related to fatigue risk as input features and using the positive and negative sample labels as training targets, performing supervised training on the classification model using a random forest algorithm to obtain a trained severe post-dialysis fatigue state prediction model; S334, during the current dialysis treatment, real-time acquiring or calculating the specific values of the objective features related to fatigue risk and inputting them into the trained severe post-dialysis fatigue state prediction model; the model outputs the probability of the current patient developing severe post-dialysis fatigue under the current dialysis regimen after processing the input features.

[0011] As an embodiment of the present application, the step S4 of evaluating the patient's dialysis treatment safety risk under the current dialysis regimen includes the following specific steps: S41, extracting the acute complication risk of the patient under the current dialysis regimen and the abnormal re-adaptation risk of the purified blood after dialysis to the patient's body; S42, weighting and summing the acute complication risk of the patient and the abnormal re-adaptation risk of the purified blood after dialysis to the patient's body to obtain the patient's dialysis treatment safety risk under the current dialysis regimen.

[0012] As an embodiment of the present application, the step S5 of performing self-adaptive adjustment warning on the patient's dialysis regimen based on the patient's dialysis treatment safety risk evaluation result under the current dialysis regimen includes the following specific steps: S51, extracting the patient's dialysis treatment safety risk under the current dialysis regimen; S52, presetting a dialysis treatment safety risk threshold, when the patient's dialysis treatment safety risk under the current dialysis regimen is greater than the dialysis treatment safety risk threshold, generating a dialysis regimen self-adaptive adjustment warning; when the patient's dialysis treatment safety risk under the current dialysis regimen is less than or equal to the dialysis treatment safety risk threshold, continuing to use the current dialysis regimen for the patient's dialysis treatment.

[0013] In a second aspect, the embodiments of the present application also provide a dialysis regimen self-adaptive interface system based on patient individual parameters, comprising: a data acquisition module for acquiring physiological state data of the patient, blood purification process data, and simultaneously acquiring dialysis strategy data under the current dialysis regimen; A complication risk prediction module is configured to analyze blood purification abnormality of the patient under the current dialysis scheme in combination with physiological state data and blood purification process data of the patient, and predict the acute complication risk of the patient under the current dialysis scheme according to the analysis result of the blood purification abnormality; A re-adaptation abnormality analysis module is configured to analyze the re-adaptation abnormality risk of the purified blood after dialysis and the patient's body under the current dialysis scheme in combination with physiological state data and dialysis strategy data of the patient; A dialysis risk detection module is configured to evaluate the safety risk of the patient under the current dialysis scheme based on the prediction result of the acute complication risk of the patient under the current dialysis scheme and the analysis result of the re-adaptation abnormality risk of the purified blood after dialysis and the patient's body under the current dialysis scheme; A scheme adjustment early warning module is configured to perform adaptive adjustment early warning on the dialysis scheme of the patient based on the evaluation result of the safety risk of the patient under the current dialysis scheme.

[0014] Compared with the prior art, the present application has the following advantages and beneficial effects: 1. The present application can accurately identify the deviation of dialysis sufficiency and safety by real-time analysis of blood purification abnormality, thereby ensuring the quality basis of dialysis treatment, reducing clinical complications caused by insufficient or excessive purification, and improving the reliability of dialysis treatment; 2. The present application can realize early warning of safety hazards in the dialysis process by prospectively predicting the risk of acute complications, thereby ensuring the safety of the patient and reducing the risk of acute adverse events; 3. The present application can comprehensively consider the long-term impact of dialysis treatment by evaluating the re-adaptation abnormality risk of the purified blood and the body, thereby providing a basis for individualized scheme optimization and ensuring the long-term life quality and prognosis of the patient. BRIEF DESCRIPTION OF DRAWINGS

[0015] Other features, objects and advantages of the present application will become more apparent from the following detailed description of non-limiting embodiments made with reference to the accompanying drawings: Figure 1 FIG. 1 is a schematic diagram of the overall process of a dialysis scheme adaptive interface method based on individual parameters of a patient according to the present application; Figure 2 FIG. 2 is a workflow diagram of step S2 in the dialysis scheme adaptive interface method based on individual parameters of a patient according to the present application; Figure 3 FIG. 3 is a workflow diagram of step S3 in the dialysis scheme adaptive interface method based on individual parameters of a patient according to the present application; Figure 4 FIG. 4 is a structural schematic diagram of a dialysis scheme adaptive interface system based on individual parameters of a patient according to the present application. DETAILED DESCRIPTION

[0016] The technical solutions of the present application will be described in detail below with reference to the drawings and specific embodiments. It should be understood that the embodiments and specific features in the embodiments are detailed descriptions of the technical solutions of the present application, and are not limitations of the technical solutions of the present application. In the case of no conflict, the technical features in the embodiments and the embodiments can be combined with each other.

[0017] Embodiment 1

[0018] As shown in the embodiment, the present embodiment provides a dialysis scheme adaptive interface method based on patient individual parameters, which specifically comprises the following steps: Figure 1 S1, collecting physiological state data and blood purification process data of the patient, and obtaining dialysis strategy data under the current dialysis scheme; S2, combining the physiological state data and the blood purification process data of the patient, analyzing the blood purification abnormality of the patient under the current dialysis scheme, and predicting the acute complication risk of the patient under the current dialysis scheme according to the blood purification abnormality analysis result; S3, combining the physiological state data and the dialysis strategy data of the patient, analyzing the re-adaptation abnormality risk of the purified blood after dialysis and the patient's body under the current dialysis scheme; S4, based on the acute complication risk prediction result of the patient under the current dialysis scheme and the re-adaptation abnormality risk analysis result of the purified blood after dialysis and the patient's body, evaluating the safety risk of the patient's dialysis treatment under the current dialysis scheme; S5, based on the safety risk evaluation result of the patient's dialysis treatment under the current dialysis scheme, self-adapting and adjusting the dialysis scheme of the patient.

[0019] In the embodiment, as shown in the embodiment, the step S2 combines the physiological state data and the blood purification process data of the patient to analyze the blood purification abnormality of the patient under the current dialysis scheme, which comprises the following specific steps: Figure 2 S21, combining the physiological state data and the blood purification process data of the patient, analyzing the blood purification abnormality of the patient under the current dialysis scheme, and obtaining the blood purification abnormality analysis result of the patient under the current dialysis scheme; ​S21, extract the relative change rate of blood volume and arterial blood pressure from the physiological state data, extract the venous end urea concentration, dialysate flow rate and blood flow rate from the blood purification process data; wherein the relative change rate of blood volume is calculated by monitoring the change of hematocrit in blood in real time through optical sensor or ultrasonic sensor, specifically by emitting light signals or acoustic signals of specific wavelength, detecting the change of its propagation characteristics in blood, and converting the signal change into blood volume relative change value combined with the preset calibration curve, the arterial blood pressure is continuously monitored by the pressure sensor on the arterial branch of the dialysis pipeline, the venous end urea concentration is detected in real time by the urea sensor installed on the dialysate circuit, the sensor works based on urease-electrochemical detection principle or optical detection principle, the dialysate flow rate and blood flow rate are measured in real time by electromagnetic flowmeter or ultrasonic flowmeter on the dialysate pipeline and blood pipeline respectively; S22, based on the single-chamber variable-volume urea kinetics model, taking the venous end urea concentration, blood flow rate, dialysate flow rate and patient dry body weight as input, calculating the effective Kt / V value of the current dialysis scheme; In this embodiment, the construction process of the single-chamber variable-volume urea kinetics model in step S22 is as follows: S221, the human body is regarded as a single-chamber of uniform mixing, and its volume linearly decreases during dialysis due to ultrafiltration; the core differential equation of the model is based on the law of conservation of mass, that is, the change rate of urea in the body is equal to the difference between its generation rate and dialysis removal rate, wherein the core differential equation is expressed as: the derivative of the product of urea distribution volume and urea concentration with respect to time is equal to the urea generation rate minus the product of dialyzer urea removal rate and blood urea concentration; wherein the initial value of urea distribution volume is estimated based on patient body weight, height and body composition, and linearly decreases during dialysis with ultrafiltration volume; the dialyzer urea removal rate is based on the standard removal rate curve provided by the dialyzer manufacturer, and is corrected in real time according to the current blood flow rate, dialysate flow rate and measured venous end urea concentration; S222, the core differential equation is solved by numerical method, specifically by four-order Runge-Kutta method for discrete numerical integration, and the dynamic attenuation process of urea concentration is simulated in real time by time step; the urea generation rate and initial distribution volume are obtained by fitting and backstepping through least square method based on multiple venous end urea concentration measurement values collected at different time points in the same dialysis treatment, so as to realize individual parameter calibration of the model; S223, input the real-time collected venous end urea concentration, blood flow rate, dialysate flow rate and ultrafiltration rate into the model calibrated by individual parameters, and calculate the current effective Kt / V value in real time to represent the instantaneous dialysis adequacy, wherein Kt / V is an important indicator of dialysis adequacy, representing the ratio of the amount of urea removed by the dialyzer to the urea distribution volume in the patient's body during dialysis. Specifically, K represents the clearance rate, t represents the dialysis time, and V represents the distribution volume of urea. The higher the Kt / V value, the better the dialysis effect, and the more effectively the toxins in the body can be removed, thereby improving the patient's health status.

[0020] S23, calculate the ratio of the absolute difference between the effective Kt / V value and the preset target Kt / V value to the preset target Kt / V value as the purification efficiency deviation; wherein the determination of the preset target Kt / V value is based on clinical guidelines and individual patient conditions, and the target value is usually in the range of 1.2-1.4. For patients with better residual renal function, the target value can be appropriately reduced, and for patients with malnutrition or high metabolic state, the target value needs to be increased. The purification efficiency deviation can quantify the deviation degree of the current dialysis scheme from the expected target in terms of execution effect. When the deviation exceeds 0.1, it indicates that the purification efficiency is abnormal, and attention needs to be paid and the dialysis parameters may need to be adjusted. The calculation of the purification efficiency deviation provides a quantitative basis for subsequent blood purification abnormality analysis, enabling the system to objectively evaluate the execution effect of the dialysis scheme and timely detect abnormal conditions; S24, calculate the linear slope of the blood volume relative change rate in the current monitoring period as the blood volume decline slope; and calculate the ratio of the blood volume decline slope to the preset safe maximum slope as the blood volume decline risk. In this embodiment, the blood volume decline slope is calculated using the linear regression method, with time as the independent variable and the blood volume relative change rate as the dependent variable. The slope value is obtained by least squares fitting. The preset safe maximum slope is based on a large amount of clinical data statistics, and the default value in this embodiment is -1.0% / min. The blood volume decline risk can reflect the influence degree of the current ultrafiltration rate on the stability of the patient's blood volume. When the risk value exceeds 0.8, it indicates that there is a high risk of hypotension, and the ultrafiltration parameters need to be adjusted in a timely manner. This risk assessment provides early warning for preventing hypotension during dialysis.

[0021] S25, weight and sum the purification efficiency deviation and the blood volume decline risk to obtain the blood purification abnormality of the patient under the current dialysis scheme. In this embodiment, the weight setting process first collects a large amount of historical dialysis data, including purification efficiency deviation, blood volume decline risk and actual clinical complication records, determines the predictive value of each indicator for complications through logistic regression analysis, and determines the final weight by combining expert scoring method and analytic hierarchy process, so as to balance the relationship between dialysis adequacy and treatment safety, and provide quantitative support for clinical decision-making; In this embodiment, step S2, based on the analysis results of abnormal blood purification, predicts the risk of acute complications for patients under the current dialysis regimen, including the following specific steps: S26. Construct a gradient boosting tree machine learning prediction model, taking the patient's abnormal blood purification status under the current dialysis protocol, the patient's average heart rate, average blood pressure variability during the current monitoring period, and the patient's history of diabetes in the static patient file as input features of the prediction model; wherein, the output features of the gradient boosting tree machine learning prediction model are the first risk probability of hypotension during dialysis and the second risk probability of muscle spasm in the next monitoring period. In this embodiment, the construction process of the gradient boosting tree machine learning prediction model in step S26 is as follows: S261. Extract samples from historical dialysis records. Each sample contains multiple feature dimensions, including but not limited to: real-time calculated abnormal blood purification conditions, patient's heart rate, blood pressure variability, whether there is a history of diabetes, blood volume decrease slope, dialysate sodium concentration, and whether hypotension and muscle cramps actually occurred during the current dialysis session. The labeling results are binary variables used for supervised learning. S262. Start with a basic constant prediction value. In this embodiment, the basic constant prediction value is assumed to be the log odds of the proportion of positive examples in the training samples. Perform multiple rounds of iteration. In each round, first calculate the negative gradient between the current model's prediction value for all training samples and the true label, that is, the derivative of the loss function with respect to the current prediction value. S263. Using the negative gradient as a new pseudo-response variable, a regression decision tree is used to fit the current pseudo-response variable. In this embodiment, the regression decision tree is constructed by traversing all features and possible split points with the goal of minimizing the sum of squared residuals after fitting. S264. Determine an optimal weight coefficient for this newly generated regression decision tree through line search so that the loss function of the overall model after adding the tree is minimized on the training data. S265. Multiply the new tree by the optimal weight coefficient and add it to the current model to update the predicted value; S266. Repeat steps S262-S265 until the preset number of iterations is reached or the loss function converges. The final model is a weighted combination of multiple weak regression decision trees. When a new real-time feature vector is input, the model sums the outputs of all trees and converts them into the first risk probability of hypotension during dialysis and the second risk probability of muscle cramps in the next monitoring cycle through a sigmoid function. To prevent overfitting, this embodiment introduces regularization constraints during training, including limiting the depth of each tree, randomly selecting some features for segmentation when constructing the tree, and reducing the weight coefficient of each tree. S27. In this embodiment, the maximum value between the first risk probability and the second risk probability is taken as the risk of acute complications for the patient under the current dialysis plan. By comparing the risk probabilities of different complications, the highest risk value is selected as the overall evaluation index based on the principle of conservatism. The first risk probability refers to the probability of hypotension during dialysis, which is obtained by transforming the corresponding output node in the gradient boosting tree model through the sigmoid function. The second risk probability refers to the probability of muscle spasm, which is calculated through another output node. The decision logic of taking the maximum value is based on the conservative principle of clinical risk management, ensuring that no potential serious complication risk is underestimated, and that potential acute complication risks can be detected in a timely manner, providing a time window for preventive intervention.

[0022] In this embodiment, as Figure 3 As shown, step S3 combines the patient's physiological status data and dialysis strategy data to analyze the abnormal risk of post-dialysis purified blood and the patient's readjustment under the current dialysis protocol, including the following specific steps: S31. Obtain the patient's average blood pressure and average heart rate within the current monitoring period from the physiological status data, as well as the low-frequency to high-frequency power ratio in heart rate variability. Simultaneously, obtain the preset ultrafiltration total, remaining treatment time, dialysate sodium concentration, and dialyzer urea clearance rate from the dialysis strategy data. The average blood pressure is calculated using the arithmetic mean of all blood pressure readings within the monitoring period. The average heart rate is also calculated using the arithmetic mean. The low-frequency to high-frequency power ratio in heart rate variability is obtained through spectral analysis. Specifically, a fast Fourier transform or autoregressive model is used to perform frequency domain analysis on the heart rate signal, extracting the power spectral density of the low-frequency band (0.04-0.15Hz) and the high-frequency band (0.15-0.4Hz), and calculating their ratio. The preset ultrafiltration total is directly obtained from the dialysis machine settings. The remaining treatment time is calculated by subtracting the treated time from the total planned dialysis time. The dialysate sodium concentration is read from the dialysis machine parameter settings. The dialyzer urea clearance rate is obtained by interpolation based on the dialyzer model and current blood flow rate and dialysate flow rate from the clearance rate curve provided by the manufacturer. S32. Establish a multiple linear regression prediction model, using the average blood pressure, average heart rate, preset total ultrafiltration volume, remaining treatment time, and dialysate sodium concentration as input variables, to predict the patient's blood pressure recovery value during the monitoring period after dialysis ends. In this embodiment, the construction process of the multiple linear regression prediction model in step S32 is as follows: S321. Determine the dependent and independent variables of the model, wherein the dependent variable is the blood pressure recovery value measured at a specific time point after dialysis, and the independent variables are multiple predictive variables obtained in real time from the current treatment, including but not limited to: the patient's average blood pressure, average heart rate, preset total ultrafiltration volume, remaining treatment time, and dialysate sodium concentration during the current monitoring period. S322. Based on a training dataset containing a large number of historical dialysis records, where each record contains the measured values ​​of independent variables and corresponding dependent variables; the least squares method is used to estimate the model parameters, with the goal of finding a set of regression coefficients that minimizes the sum of squared residuals between the model's predicted values ​​and the measured blood pressure values; this process is achieved by solving the normal equation, that is, the solution of the regression coefficients is equal to the inverse matrix of the transpose of the design matrix of the independent variables multiplied by itself, and then multiplied by the result of multiplying the transpose of the design matrix by the dependent variable vector; S323. In this embodiment, multicollinearity diagnosis is performed on the trained model. The variance inflation factor is calculated to ensure the independence of each variable. If severe collinearity exists, ridge regression or principal component regression is used for correction. Finally, a determined multiple linear regression equation is obtained, which is in the form that the predicted blood pressure recovery value is equal to the intercept term plus the sum of the products of each variable and its corresponding regression coefficient. The real-time variable values ​​during the current dialysis process are substituted into this equation to calculate the predicted value of the patient's blood pressure recovery value during the monitoring period after the end of dialysis, which is used for subsequent risk assessment. S33. Construct a classification model based on random forest, using the average heart rate, preset total ultrafiltration volume, remaining treatment time, and the low-frequency to high-frequency power ratio in heart rate variability as input features, to predict the probability of fatigue after severe dialysis. In this embodiment, the prediction of the probability of post-dialysis fatigue in step S33 includes the following specific details: S331. Extract historical dialysis records of multiple patients from the historical medical database. Each record contains physiological status data during dialysis, dialysis strategy data, and fatigue self-assessment results recorded after dialysis. Based on the standardized fatigue assessment scale, record records whose post-dialysis self-assessment results reach or exceed the preset severity threshold are marked as positive samples, i.e., "severe fatigue occurred", and assigned a value of one. The remaining records are marked as negative samples and assigned a value of zero. S332. Extract objective features related to fatigue risk from the historical dialysis records. The objective features include, but are not limited to: the average heart rate during dialysis, the ratio of low-frequency to high-frequency heart rate variability within a specific time window before the end of dialysis, the percentage of the preset total ultrafiltration volume of this dialysis relative to dry body weight, the total duration of dialysis treatment, the urea clearance rate of the dialyzer, and the solute clearance index estimated based on the urea kinetic model. S333. Using objective features related to fatigue risk as input features and the positive and negative sample labels as training targets, the random forest algorithm is used to supervise the training of the classification model to obtain a well-trained fatigue state prediction model after severe dialysis. S334. During the current dialysis treatment, the specific values ​​of objective features related to fatigue risk are acquired or calculated in real time and input into the trained severe dialysis post-fatigue state prediction model; after processing the input features, the model outputs the probability of the current patient experiencing severe dialysis post-fatigue state under the current dialysis plan.

[0023] S34. Calculate the absolute difference between the predicted blood pressure recovery value and the individualized target blood pressure. Use the ratio of the absolute difference to the individualized target blood pressure as the patient's cardiovascular readjustment risk. The individualized target blood pressure is determined based on the patient's usual baseline blood pressure. In this embodiment, the average of the home blood pressure measured one week before dialysis is used. When the predicted blood pressure is lower than the target blood pressure, the difference is taken as a positive value, reflecting the risk of hypotension. When the predicted blood pressure is higher than the target blood pressure, the difference is taken as an absolute value, reflecting the risk of hypertension. The ratio calculation uses the absolute difference divided by the target blood pressure to quantify the ability of the cardiovascular system to recover to an ideal state after dialysis, providing an important basis for assessing the long-term impact of dialysis treatment on the cardiovascular system.

[0024] S35. The probability of experiencing severe post-dialysis fatigue is taken as the patient's fatigue risk; the cardiovascular readaptation risk and fatigue risk are weighted and summed to obtain the risk of abnormal readaptation of purified blood and the patient's body after dialysis under the current dialysis protocol; the weight coefficients are determined based on clinical expert experience and statistical analysis results. The Delphi method is used to collect scores on the importance of the two risks from multiple nephrologists, and the weight vector is calculated using the analytic hierarchy process. At the same time, the weight coefficients are corrected by combining the correlation analysis between the two risks and the actual occurrence of abnormal readaptation in historical data; thus, it can comprehensively reflect the recovery status of the patient's physiological system after dialysis and provide an important reference for optimizing the dialysis protocol.

[0025] In this embodiment, step S4 assesses the safety risks of patient dialysis treatment under the current dialysis protocol, including the following specific steps: S41. Extract the risk of acute complications in patients under the current dialysis protocol and the risk of abnormal readjustment of purified blood and the patient's body after dialysis; S42. The patient's acute complication risk, post-dialysis purified blood risk, and readjustment abnormality risk are weighted and summed to obtain the patient's dialysis treatment safety risk under the current dialysis protocol. The weight coefficients are determined based on risk severity and clinical priority. By analyzing the impact of acute complications and readjustment abnormalities on patient prognosis in historical complication data, a multi-criteria decision analysis method is used to determine the weight allocation. At the same time, the evaluation of the importance of the two risks by clinical experts is combined for calibration, which can balance the relationship between immediate safety and long-term prognosis and provide a scientific basis for the dynamic adjustment of the dialysis protocol.

[0026] In this embodiment, step S5 involves adaptively adjusting and issuing early warnings for the patient's dialysis plan based on the patient's dialysis treatment safety risk assessment results under the current dialysis protocol. This includes the following specific steps: S51. Extract the safety risks of patient dialysis treatment under the current dialysis protocol; S52. A preset safety risk threshold for dialysis treatment is established. When the safety risk of dialysis treatment for a patient under the current dialysis plan exceeds the safety risk threshold, an adaptive adjustment warning for the dialysis plan is generated. When the safety risk of dialysis treatment for a patient under the current dialysis plan is less than or equal to the safety risk threshold, the current dialysis plan is continued to be used for dialysis treatment. The safety risk threshold is set based on clinical safety standards and historical data analysis. By collecting a large amount of dialysis treatment safety risk data and the correspondence between actual complication occurrences, ROC curve analysis is used to determine the optimal cutoff point as the safety risk threshold for dialysis treatment.

[0027] Example 2

[0028] like Figure 4 As shown, this embodiment provides an adaptive interface system for dialysis protocols based on individual patient parameters, including: The data acquisition module is used to collect patients' physiological status data, blood purification process data, and dialysis strategy data under the current dialysis protocol. The complication risk prediction module is used to analyze abnormal blood purification conditions under the current dialysis protocol by combining the patient's physiological status data and blood purification process data; and to predict the risk of acute complications under the current dialysis protocol based on the analysis results of abnormal blood purification conditions. The readaptation anomaly analysis module is used to analyze the risk of readaptation anomalies in the patient's body by combining the patient's physiological status data and dialysis strategy data; The dialysis risk detection module is used to assess the safety risks of dialysis treatment for patients under the current dialysis protocol based on the prediction results of the patient's acute complication risk under the current dialysis protocol and the analysis results of abnormal risk of the purified blood and the patient's body readjustment after dialysis. The protocol adjustment early warning module is used to adaptively adjust and issue early warnings for patients' dialysis protocols based on the patient's dialysis treatment safety risk assessment results under the current dialysis protocol.

[0029] The steps for implementing the corresponding functions of each parameter and each unit module in the adaptive interface system for dialysis protocols based on individual patient parameters of the present invention described above can be referred to the parameters and steps in the embodiments of the adaptive interface method for dialysis protocols based on individual patient parameters described above, and will not be repeated here.

[0030] The various embodiments in this invention are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on its differences from other embodiments. In particular, the embodiments for IoT devices and media are relatively simple in description because they are fundamentally similar to the method embodiments; relevant parts can be referred to the descriptions in the method embodiments.

[0031] The systems, media, and methods provided in the embodiments of the present invention are in one-to-one correspondence. Therefore, the systems and media also have similar beneficial technical effects as their corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the systems and media will not be repeated here.

[0032] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0033] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0034] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0035] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0036] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0037] Computer-readable media include both permanent and non-permanent, removable and non-removable media that can store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0038] It should also be noted that 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. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0039] The above are merely embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. An adaptive interface method for dialysis protocols based on individual patient parameters, characterized in that, Includes the following steps: S1. Collect patient physiological status data and blood purification process data, and obtain dialysis strategy data under the current dialysis protocol; S2. Analyze abnormal blood purification conditions in patients under the current dialysis protocol by combining the patient's physiological status data and blood purification process data; Based on the analysis results of abnormal blood purification, the risk of acute complications in patients under the current dialysis regimen is predicted; S3. Combining patient physiological data and dialysis strategy data, analyze the abnormal risk of purified blood after dialysis and the patient's readjustment under the current dialysis protocol; S4. Based on the prediction results of acute complication risks of patients under the current dialysis protocol and the analysis results of abnormal readjustment risks of purified blood after dialysis and the patient's body, assess the safety risks of dialysis treatment under the current dialysis protocol. S5. Based on the patient's dialysis treatment safety risk assessment results under the current dialysis protocol, adaptively adjust and issue early warnings for the patient's dialysis protocol.

2. The adaptive interface method for dialysis protocols based on individual patient parameters according to claim 1, characterized in that, Step S2 combines the patient's physiological status data and blood purification process data to analyze abnormal blood purification conditions under the current dialysis protocol, including the following specific steps: S21. Extract the relative change rate of blood volume and arterial blood pressure from physiological state data, and extract venous urea concentration, dialysate flow rate and blood flow velocity from blood purification process data. S22. Based on the single-compartment variable volume urea kinetic model, the effective Kt / V value of the current dialysis protocol is calculated using the urea concentration at the venous end, blood flow rate, dialysate flow rate and patient dry weight as inputs. The Kt / V value is the ratio of the amount of urea removed by the dialyzer to the volume of urea distributed in the patient's body. S23. Calculate the ratio of the absolute difference between the effective Kt / V value and the preset target Kt / V value to the preset target Kt / V value, and use it as the deviation of the purification efficiency. S24. Calculate the linear slope of the relative rate of change of blood volume within the current monitoring period, as the slope of blood volume decrease; and take the ratio of the slope of blood volume decrease to the preset safe maximum slope as the risk of blood volume decrease. S25. Weighted summation of purification efficiency deviation and blood volume decrease risk to obtain the abnormal blood purification status of the patient under the current dialysis protocol.

3. The adaptive interface method for dialysis protocols based on individual patient parameters according to claim 2, characterized in that, In step S2, based on the analysis results of abnormal blood purification, the risk of acute complications for patients under the current dialysis regimen is predicted, including the following specific steps: S26. Construct a gradient boosting tree machine learning prediction model, taking the patient's abnormal blood purification status under the current dialysis protocol, the patient's average heart rate, average blood pressure variability during the current monitoring period, and the patient's history of diabetes in the static patient file as input features of the prediction model; wherein, the output features of the gradient boosting tree machine learning prediction model are the first risk probability of hypotension during dialysis and the second risk probability of muscle spasm in the next monitoring period. S27. Take the maximum value between the first risk probability and the second risk probability as the patient's risk of acute complications under the current dialysis regimen.

4. The adaptive interface method for dialysis protocols based on individual patient parameters according to claim 3, characterized in that, In step S3, the patient's physiological status data and dialysis strategy data are combined to analyze the abnormal risk of post-dialysis purified blood and the patient's readjustment under the current dialysis protocol. This includes the following specific steps: S31. Obtain the patient’s average blood pressure and average heart rate during the current monitoring period from the physiological status data, as well as the low-frequency and high-frequency power ratio in the heart rate variability. At the same time, obtain the preset ultrafiltration total, remaining treatment time, dialysate sodium concentration, and dialyzer urea clearance rate from the dialysis strategy data. S32. Establish a multiple linear regression prediction model, using the average blood pressure, average heart rate, preset total ultrafiltration volume, remaining treatment time, and dialysate sodium concentration as input variables, to predict the patient's blood pressure recovery value during the monitoring period after dialysis ends. S33. Construct a classification model based on random forest, using the average heart rate, preset total ultrafiltration volume, remaining treatment time, and the low-frequency to high-frequency power ratio in heart rate variability as input features, to predict the probability of fatigue after severe dialysis. S34. Calculate the absolute difference between the predicted blood pressure recovery value and the individualized target blood pressure, and use the ratio of the absolute difference to the individualized target blood pressure as the patient's cardiovascular readjustment risk. S35. The probability of experiencing fatigue after severe dialysis is taken as the patient's fatigue risk. The cardiovascular readaptation risk and fatigue risk are weighted and summed to obtain the abnormal readaptation risk of purified blood and the patient's body after dialysis under the current dialysis protocol.

5. The adaptive interface method for dialysis protocols based on individual patient parameters according to claim 4, characterized in that, Step S4 involves assessing the safety risks of dialysis treatment for patients under the current dialysis protocol, including the following specific steps: S41. Extract the risk of acute complications in patients under the current dialysis protocol and the risk of abnormal readjustment of purified blood and the patient's body after dialysis; S42. The patient's risk of acute complications and the risk of abnormal readjustment of purified blood after dialysis are weighted and summed with the patient's body readjustment risk to obtain the patient's dialysis treatment safety risk under the current dialysis protocol.

6. The adaptive interface method for dialysis protocols based on individual patient parameters according to claim 5, characterized in that, Step S5 involves adaptively adjusting and issuing early warnings for the patient's dialysis plan based on the patient's dialysis treatment safety risk assessment results under the current dialysis protocol. This includes the following specific steps: S51. Extract the safety risks of patient dialysis treatment under the current dialysis protocol; S52. Preset dialysis treatment safety risk threshold. When the patient's dialysis treatment safety risk under the current dialysis plan is greater than the dialysis treatment safety risk threshold, generate an adaptive adjustment warning for the dialysis plan. If the safety risk of dialysis treatment for a patient under the current dialysis regimen is less than or equal to the dialysis treatment safety risk threshold, the current dialysis regimen shall continue to be used for dialysis treatment.

7. An adaptive interface system for dialysis protocols based on individual patient parameters, implemented based on the adaptive interface method for dialysis protocols based on individual patient parameters as described in any one of claims 1-6, characterized in that, The system includes: The data acquisition module is used to collect patients' physiological status data, blood purification process data, and dialysis strategy data under the current dialysis protocol. The complication risk prediction module is used to analyze abnormal blood purification conditions under the current dialysis protocol by combining the patient's physiological status data and blood purification process data; and to predict the risk of acute complications under the current dialysis protocol based on the analysis results of abnormal blood purification conditions. The readaptation anomaly analysis module is used to analyze the risk of readaptation anomalies in the patient's body by combining the patient's physiological status data and dialysis strategy data; The dialysis risk detection module is used to assess the safety risks of dialysis treatment for patients under the current dialysis protocol based on the prediction results of the patient's acute complication risk under the current dialysis protocol and the analysis results of abnormal risk of the purified blood and the patient's body readjustment after dialysis. The protocol adjustment early warning module is used to adaptively adjust and issue early warnings for patients' dialysis protocols based on the patient's dialysis treatment safety risk assessment results under the current dialysis protocol.

Images