An AI-based dialysis plan optimization system

Abstract

The present invention relates to the field of dialysis technology, and specifically to an AI-based dialysis scheme optimization system, the system comprising: based on continuous dialysis cycle data, analyzing changes in urea nitrogen concentration and dialysate discharge volume, obtaining dynamic characteristics of cycle nitrogen, and then optimizing sodium concentration settings, combining abnormal linkage of blood flow rate, heart rate and systolic pressure, outputting physiological linkage warning signals, adjusting ultrafiltration volume and dialysis duration, and generating equipment control instructions. The present invention, by realizing multi-parameter trend induction and real-time linkage control, can analyze the complex coupling relationship between multi-dimensional physiological signals in a continuous dialysis cycle, automatically identify key metabolic states and dynamic changes in parameters, drive adaptive responses of core parameters such as sodium concentration and ultrafiltration operation, and output dynamic control instructions for hemodynamics and volume abnormalities based on the risk signal generation mechanism, so that the dialysis process has the capabilities of intelligent discrimination, automatic sorting and rapid matching, and forms an optimization strategy for individual conditions.

Description

Technical Field

[0001] The present invention relates to the field of dialysis technology, and in particular to an AI-based dialysis scheme optimization system. Background Art

[0002] Dialysis technology involves the use of medical methods such as hemodialysis or peritoneal dialysis, with the help of dialysis equipment and dialysate to remove metabolic waste and excess water from the body, thereby replacing part of the kidney function in patients with renal dysfunction or failure. It includes the selection of dialysis methods, the setting of dialysis parameters, the collection and evaluation of patient health data, and the dynamic adjustment of the dialysis process. It covers the analysis of physiological parameters related to dialysis treatment, the formulation of treatment plans, and the monitoring of the treatment process. Among them, the traditional dialysis plan optimization system refers to the medical staff manually analyzing the patient's various health indicators such as blood pressure, weight, and blood biochemical parameters based on the patient's previous medical records, physical sign monitoring results, and accumulated experience. In combination with the conventional dialysis parameter adjustment principles, a treatment plan is formulated, including dialysis cycle, duration, and dialysate composition. The plan is usually optimized through regular follow-up, empirical judgment, and simple statistical analysis.

[0003] Existing technologies mostly use fixed-cycle data collection and static parameter settings, making it difficult to cope with the problems caused by changes in multiple factors during the dialysis process. In particular, when multiple parameters fluctuate rapidly or risk signals suddenly appear, there is a lack of global linkage analysis and real-time judgment mechanisms, resulting in delayed parameter adjustment and insufficient adaptation to individual differences. There are deficiencies in risk warning and response speed. In actual operation, it is easy to cause control mismatch and omission of abnormal conditions due to rigid processes or untimely information updates, affecting the dynamic adjustment of dialysis plans and efficacy assurance. Summary of the Invention

[0004] The purpose of the present invention is to solve the shortcomings of the existing technology and propose an AI-based dialysis plan optimization system.

[0005] To achieve the above objectives, the present invention adopts the following technical solutions: an AI-based dialysis plan optimization system, comprising:

[0006] The nitrogen dynamic feature recognition module analyzes the changes in urea nitrogen concentration based on continuous dialysis cycle data, compares the increase and decrease directions of each cycle, and combines the changes in dialysate discharge volume to calculate the correlation between urea nitrogen concentration fluctuations and dialysate discharge volume, determine the relationship type, and obtain the cycle nitrogen dynamic feature;

[0007] The sodium control parameter generation module analyzes the synchronous changes of the dialysate discharge and the nitrogen balance trend based on the cyclic nitrogen dynamic characteristics, compares the cyclic trends, calculates the sodium concentration adjustment range, dynamically optimizes the sodium concentration setting, and obtains the sodium concentration adjustment range;

[0008] The linkage risk identification module analyzes the impact of continuous changes in blood flow rate based on the sodium concentration adjustment range, compares the fluctuation trends of heart rate and systolic blood pressure, determines the correlation of synchronous fluctuations, identifies abnormal linkage periods, and obtains physiological linkage warning signals;

[0009] The ultrafiltration baseline setting module determines the risk type based on the physiological linkage warning signal, analyzes the changes in blood volume before and after the dialysis cycle, compares the corresponding relationship between the risk judgment amount and the blood volume, adjusts the ultrafiltration volume and dialysis time, and obtains the ultrafiltration parameter control baseline.

[0010] The present invention has improvements in that the dynamic characteristics of cyclical nitrogen include metabolic fluctuation indicators, cyclical correlation factors, and trend classification identifiers; the sodium concentration adjustment range includes parameter setting range, adjustment priority, and change type; the physiological linkage warning signal includes a risk identifier, a synchronous reaction marker, and an abnormal event type; and the ultrafiltration parameter control baseline includes a baseline control parameter, a cyclical adaptation category, and a risk control element.

[0011] The present invention is improved in that the steps of obtaining the periodic nitrogen dynamic characteristics are specifically as follows:

[0012] S111: Based on the continuous dialysis cycle data, the changes in urea nitrogen concentration at each stage are analyzed. By comparing the concentration increase and decrease directions of adjacent cycles, the change trend of metabolites is determined and the concentration change sequence is obtained;

[0013] S112: comparing the concentration change sequence with the change process of the dialysate discharge volume in the corresponding cycle, analyzing the synchronization and deviation between the two, determining the correlation of the changes within the cycle, and obtaining a cycle correlation feature identifier;

[0014] S113: calling the cycle-related feature identifier, combining the various change trends within the cycle, classifying the cycle feature type, optimizing the cycle feature expression, and obtaining the cycle nitrogen dynamic feature.

[0015] The present invention is improved in that the step of obtaining the sodium concentration adjustment range is specifically as follows:

[0016] S211: Based on the cyclic nitrogen dynamic characteristics, comparing the change trends of the dialysate discharge volume series and the urea nitrogen concentration series in each time period, determining whether the change directions of the two are consistent, selecting time segments with consistent change trends, and obtaining the number of segments with consistent trends;

[0017] S212: comparing the dialysate discharge change rate and the nitrogen balance trend change based on the number of segments with consistent trends, determining the difference between the two, identifying the time period with the optimal difference, and obtaining the optimal adjustment amplitude sequence;

[0018] S213: Dynamically optimize the sodium concentration setting according to the optimal adjustment range sequence to match the current nitrogen balance change and obtain the sodium concentration adjustment range.

[0019] The present invention is improved in that the steps of obtaining the physiological linkage warning signal are specifically as follows:

[0020] S311: Based on the sodium concentration adjustment range, comparing the change directions of blood flow velocity, heart rate, and systolic blood pressure in each time period, determining the synchronization of each parameter, calculating the correlation result, and obtaining the heart rate-pressure difference consistency coefficient;

[0021] S312: Based on the heart rate-pressure difference consistency coefficient, a collaborative screening is performed on the changes in blood flow rate in each time period and the fluctuations in heart rate and systolic pressure to obtain the linkage abnormality identification strength, identify the time period when there is linkage abnormality between blood flow rate and physiological parameters, and obtain a physiological linkage warning signal.

[0022] The present invention is improved in that the step of obtaining the ultrafiltration parameter control baseline is specifically as follows:

[0023] S411: Based on the physiological linkage warning signal, optimizing the blood volume monitoring information of the associated stages before and after the dialysis cycle, comparing the blood volume change trends corresponding to each time period, and calculating the blood volume change rate sequence of the continuous time period;

[0024] S412: Based on the blood volume change rate sequence, according to the risk identifier, identifying the time period corresponding to the associated risk type, comparing it with the risk trend discriminant, and obtaining the ultrafiltration adjustment coefficient;

[0025] S413: Based on the ultrafiltration adjustment coefficient, adjust the ultrafiltration volume parameter of the current cycle and the dialysis duration setting parameter of the dialysis device, optimize the current parameter combination, and obtain the ultrafiltration parameter control baseline.

[0026] The present invention is improved in that the system further comprises:

[0027] The intelligent instruction output module analyzes the effect of the ultrafiltration parameter control baseline in the current cycle, calculates the control action priority, determines the top parameter item, adjusts the dialysis equipment control parameters, and obtains the equipment control instruction;

[0028] The device control instruction includes an action instruction code, a priority number, and an implementation parameter set.

[0029] The present invention is improved in that the step of obtaining the device control instruction is specifically as follows:

[0030] S511: Based on the ultrafiltration parameter control baseline, analyzing the linkage status between blood flow rate, dialysate sodium concentration, dialysis duration, heart rate, systolic blood pressure, and blood volume change rate, comparing the correlation strength between each parameter and the risk trend discriminant, optimizing the coordination relationship between the parameters, and obtaining the parameter priority ranking result;

[0031] S512: Determine the first parameter item in the parameter priority ranking result, adjust the control parameters of the dialysis equipment, coordinate the current dialysate discharge, blood flow rate change trend and sodium concentration setting, and obtain equipment control instructions.

[0032] Compared with the prior art, the advantages and positive effects of the present invention are:

[0033] In the present invention, by realizing multi-parameter trend induction and real-time linkage regulation, the complex coupling relationship between multi-dimensional physiological signals can be analyzed in a continuous dialysis cycle, the dynamic changes of key metabolic states and parameters can be automatically identified, and the core parameters such as sodium concentration and ultrafiltration operation can be driven to respond adaptively. According to the risk signal generation mechanism, dynamic control instructions are output for hemodynamic and volume abnormalities, so that the dialysis process has the capabilities of intelligent judgment, automatic sorting and rapid matching, forming an optimization strategy for individual status, effectively enhancing the timeliness and sensitivity of regulation, improving the foresight of risk prevention and control, and expanding the boundaries of dialysis treatment data-driven and individual feature adaptation. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a system flow chart of the present invention;

[0035] Figure 2 This is a flow chart for obtaining the dynamic characteristics of periodic nitrogen in the present invention;

[0036] Figure 3 This is a flow chart for obtaining the sodium concentration adjustment range in the present invention;

[0037] Figure 4 This is a flow chart for obtaining physiological linkage warning signals in the present invention;

[0038] Figure 5 This is a flow chart for obtaining the ultrafiltration parameter control baseline in the present invention;

[0039] Figure 6 This is a flow chart for obtaining device control instructions in the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0041] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings and are only for the convenience of describing the present invention and simplifying the description. They do not indicate or imply that the devices or elements referred to must have a specific direction, be constructed and operate in a specific direction, and therefore should not be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0042] Example

[0043] See also Figure 1 The present invention provides a technical solution: an AI-based dialysis plan optimization system, comprising:

[0044] S1: Based on continuous dialysis cycle data, analyze the changes in urea nitrogen concentration, compare the increase and decrease directions of urea nitrogen concentration in each cycle, combine the changes in dialysate discharge volume in each cycle, calculate the correlation between urea nitrogen concentration fluctuations and dialysate discharge volume, determine the relationship type, and obtain the cycle nitrogen dynamic characteristics;

[0045] S2: Based on the dynamic characteristics of cyclic nitrogen, the synchronous characteristics of the change of dialysate discharge volume with the nitrogen balance trend are analyzed. By comparing the trend changes within the cycle, the adjustment range of the dialysate sodium concentration parameter is calculated, and the sodium concentration setting is dynamically optimized to match the current nitrogen balance changes to obtain the sodium concentration adjustment range;

[0046] S3: Based on the adjustment range of sodium concentration, analyze the impact of continuous changes in blood flow rate, compare the fluctuation trends of heart rate and systolic blood pressure in the same time period, determine the correlation between changes in blood flow rate and synchronous fluctuations of heart rate and systolic blood pressure, identify periods of abnormal linkage between blood flow rate and physiological parameters, and obtain physiological linkage warning signals;

[0047] S4: Based on the physiological linkage warning signal, determine the risk type it reflects, analyze the blood volume change rate before and after the dialysis cycle, compare the correspondence between the risk trend judgment value and the blood volume change, adjust the ultrafiltration volume parameters and dialysis duration settings, optimize the current parameter combination, and obtain the ultrafiltration parameter control baseline;

[0048] S5: Based on the ultrafiltration parameter control baseline, analyze its effect in the current cycle, calculate the priority order of each control action, determine the parameter item with the highest priority, and adjust the control parameters of the dialysis equipment accordingly to obtain the equipment control instructions.

[0049] The dynamic characteristics of cyclical nitrogen include metabolic fluctuation indicators, cyclical correlation factors, and trend classification identifiers. The sodium concentration adjustment range includes parameter setting range, adjustment priority, and change type. The physiological linkage warning signal includes risk identifier, synchronous reaction marker, and abnormal event type. The ultrafiltration parameter control baseline includes baseline control parameters, cyclical adaptation category, and risk management elements. The equipment control instructions include action instruction code, priority number, and implementation parameter set.

[0050] In S1, the change in urea nitrogen concentration refers to the numerical change in the patient's blood urea nitrogen (BUN) concentration during different dialysis cycles (such as between one or more dialysis sessions) during dialysis treatment, which is manifested as an increase or decrease; the direction of increase or decrease refers to the comparison of urea nitrogen concentration before and after each dialysis cycle or between cycles, whether it is an increase or decrease, reflecting the trend of metabolite clearance; the change process refers to the continuous change trajectory of urea nitrogen concentration on the time axis, which is the process information of continuous changes between static sampling points; the judgment relationship type refers to the qualitative or quantitative judgment of whether the fluctuation of urea nitrogen concentration and dialysate output are positively correlated, negatively correlated, or have no obvious correlation, reflecting the coupling relationship between the two parameters.

[0051] In S2, the change in nitrogen balance trend refers to the "cyclic nitrogen dynamic characteristics" obtained based on S1, which judges the overall change trend of urea nitrogen in continuous dialysis cycles, whether it is continuously positive (accumulation), negative (clearance) or fluctuating (no stable trend); the trend change within the cycle refers to the specific change direction and amplitude of the values ​​of key parameters (such as urea nitrogen concentration and dialysate output) over time in the same dialysis cycle; the amplitude of the sodium concentration parameter that needs to be adjusted refers to the range and direction in which the sodium concentration in the dialysate needs to be adjusted. The adjustment range is given by AI after a comprehensive analysis of nitrogen dynamics and fluid changes, and is not a specific value; optimizing the sodium concentration setting refers to automatically or manually adjusting the target concentration of sodium in the dialysate based on the AI ​​decision results to make it more in line with the actual needs of the patient's current metabolism and electrolyte balance.

[0052] In S3, the fluctuation trend in the same time period refers to the changing trends of the blood flow rate, heart rate, systolic blood pressure and other parameters collected during the dialysis process, whether they change synchronously or are unrelated to each other; the period of abnormal linkage refers to the AI's judgment that within a certain time period, the blood flow rate and at least one physiological parameter (such as heart rate or systolic blood pressure) simultaneously experience abnormal fluctuations. This "abnormal linkage" potentially indicates a potential risk event.

[0053] In S4, risk type refers to the classification and determination of risk properties based on early warning signals, combined with historical data and real-time changes, such as volume load risk, electrolyte imbalance risk, circulatory dynamics abnormality, etc.; corresponding situation refers to the quantitative or qualitative relationship between the risk trend judgment value and blood volume change, judging whether the change mode of blood volume when the risk signal appears is highly consistent with the risk category; ultrafiltration volume parameter refers to the ultrafiltration target that needs to be adjusted during dialysis, that is, the amount of water that needs to be removed from the patient's body, which is the core control item of the dialysis machine; dialysis duration setting refers to adjusting the total duration of the dialysis course according to the current risk and blood volume changes to control the intensity and safety of treatment; risk trend judgment value refers to the comprehensive signal result reflecting risk events or potential risk trends identified after comprehensive analysis, trend calculation, and linkage judgment of multiple key physiological parameters collected throughout the dialysis process.

[0054] In S5, the first parameter item refers to the specific parameter that ranks first among all the parameters that need to be adjusted in this dialysis cycle after priority sorting and most needs to be adjusted immediately (ultrafiltration volume setting parameter, dialysis duration setting parameter, blood flow rate setting parameter, etc.); the control parameter refers to the dialysis operation parameter that is finally determined to be adjusted on the equipment, such as sodium concentration, ultrafiltration volume, dialysis duration, etc. The parameters can directly affect the dialysis process.

[0055] See also Figure 2 , the specific steps for obtaining the dynamic characteristics of periodic nitrogen are:

[0056] S111: Based on the continuous dialysis cycle data, the changes in urea nitrogen concentration at each stage are analyzed. By comparing the concentration increase and decrease directions of adjacent cycles, the change trend of metabolites is determined and the concentration change sequence is obtained;

[0057] First, for each dialysis cycle, a blood test is performed before the start of the cycle to collect the urea nitrogen concentration value and record it as the cycle start value. After the cycle ends, the urea nitrogen concentration is collected again and recorded as the cycle end value. Then, the cycle end value is subtracted from the cycle start value. If the result is a positive number, it is judged that the urea nitrogen concentration of the cycle is on an increasing trend. If the result is a negative number, it is judged to be a decreasing trend. If the result is zero, it is judged to be no obvious change trend. When performing this operation, each cycle is processed independently to form a set of urea nitrogen concentration change trend mark sequences. For example, the starting value of cycle one is 20mmol / L and the ending value is 15mmol / L, which is judged to be a decreasing trend. The starting value of cycle two is 15mmol / L and the ending value is 17mmol / L, which is judged to be an increasing trend. The starting value of cycle three is 17mmol / L and the ending value is 14mmol / L, which is judged to be a decreasing trend. The starting value of cycle four is 14mmol / L and the ending value is 12mmol / L, which is judged to be a decreasing trend. The trend mark sequence is formed as decrease, increase, decrease, decrease. Next, the trends of two adjacent cycles are compared pairwise. If the trends of the two adjacent cycles are the same, such as both decrease or both increase, it is determined that the trend continues. If the trends are opposite, such as one decreases and the other increases, it is determined that the trend is reversed. All cycle comparisons are completed one by one to form a trend comparison sequence. For example, the trends of cycle one and cycle two are decrease and increase, which are determined as trend reversal. The trends of cycle two and cycle three are increase and decrease, which are determined as trend reversal. The trends of cycle three and cycle four are decrease and decrease, which are determined as trend continuation. Finally, the trend comparison sequence is reversal, reversal, and continuation. During the execution process, if the absolute value of the change in urea nitrogen concentration in a single cycle is less than 1mmol / L, the cycle change trend is determined to be an invalid trend and is not included in the trend sequence. The 1mmol / L threshold is set with reference to the error range of routine hemodialysis testing, forming a urea nitrogen change trend sequence and a trend comparison sequence.

[0058] S112: comparing the concentration change sequence with the change process of the dialysate discharge volume in the corresponding cycle, analyzing the synchronization and deviation between the two, determining the correlation of the changes within the cycle, and obtaining a cycle correlation feature identifier;

[0059] For each dialysis cycle, the dialysate discharge data is extracted. During the dialysis process, the equipment records the dialysate discharge volume every hour to form a complete discharge volume time series. During the execution process, the discharge volume time series is first read, and then the urea nitrogen change trend is mapped to the dialysis time axis. Each hour is used as an analysis window to determine whether the dialysate discharge volume in the window is increasing, decreasing or stable. At the same time, the direction of the urea nitrogen change trend in the window is determined. If the two trend directions are consistent, such as the discharge volume decreases and the urea nitrogen concentration decreases, or the discharge volume increases and the urea nitrogen concentration increases, it is determined to be a synchronous segment. If the trend directions are opposite, it is determined to be a divergent segment. The trend judgment of all windows in the entire cycle is completed one by one, and the total length of the synchronous segment time and the total length are calculated. The proportion of cycle time is used to calculate the percentage of synchronization. If the synchronization segment time accounts for more than 70%, it is determined to be a high synchronization cycle. If the synchronization ratio is between 30% and 70%, it is determined to be a medium synchronization cycle. If the synchronization ratio is less than 30%, it is determined to be a low synchronization cycle. For example, the hourly data of cycle five discharge are 0.9 liters, 0.8 liters, 0.7 liters, 0.6 liters, 0.7 liters, and 0.8 liters. The urea nitrogen concentration decreases in the first three hours and increases in the last three hours. The first three hours are determined to be a synchronization segment, and the last three hours are determined to be a divergence segment. The total cycle length is 6 hours, and the synchronization segment accounts for 3 hours, accounting for 50%, which is determined to be a medium synchronization cycle. Combined with the trend comparison sequence, a cycle-related feature identifier is formed for this cycle.

[0060] S113: Calling the period-related feature identifier, combining the changing trends within the period, classifying the period feature type, optimizing the period feature expression, and obtaining the period nitrogen dynamic feature;

[0061] Integrate other trend data within the cycle, extract urea nitrogen concentration trend, dialysate discharge trend, blood flow rate trend, heart rate trend, systolic blood pressure trend, collect time series data for each parameter within the cycle, map it to a unified dialysis time axis, divide it into hourly windows, and judge whether the trend in each window is increasing, decreasing, or stable. The specific operation is to compare the difference between the starting value and the ending value of the current window. If the difference exceeds the corresponding clinical judgment threshold, it is classified as an obvious trend. If the difference is lower than the threshold, it is classified as stable. After executing all window trend classifications, a trend classification matrix is ​​formed. For example, the urea nitrogen trend is decreasing, decreasing, and stable, and the discharge trend is increasing, decreasing , increase, the blood flow rate trend is stable, stable, and decrease, the heart rate trend is increase, increase, and decrease, and the systolic blood pressure trend is stable, decrease, and stable. Then the trend classification matrix is ​​matched with the cycle-related feature identifier, and the cycle feature type is classified according to the preset rules. For example, if the urea nitrogen concentration trend decreases as a whole and the synchronization feature is high synchronization, it is classified as a benign clearance cycle. If the urea nitrogen trend fluctuates repeatedly and the synchronization feature is low synchronization, it is classified as a fluctuation clearance cycle. If the urea nitrogen trend continues to rise or the systolic blood pressure continues to decrease, it is classified as a risk warning cycle. Finally, the cycle feature type and the trend classification matrix are jointly encoded to form the cycle nitrogen dynamic feature.

[0062] See also Figure 3 , the specific steps for obtaining the sodium concentration adjustment range are:

[0063] S211: Based on the dynamic characteristics of periodic nitrogen, compare the change trends of the dialysate discharge volume series and the urea nitrogen concentration series in each time period to determine whether the change directions of the two are consistent, select time segments with consistent change trends, and obtain the number of segments with consistent trends;

[0064] Extract the urea nitrogen concentration trend sequence and the dialysate discharge trend sequence from the processed cyclic nitrogen dynamic characteristics, and divide each dialysis cycle into a unified time period. The dialysis cycle is generally divided into one time period per hour. If the total duration of a cycle is four hours, it is divided into four time periods. All time periods are processed in sequence. First, the urea nitrogen concentration change trend in each time period is processed for the urea nitrogen concentration sequence. The urea nitrogen concentration value at the start time of the period is directly compared with the urea nitrogen concentration value at the end time. If the end value is greater than the start value, it is judged to be an increasing trend. If it is less than, it is judged to be a decreasing trend. If they are equal, it is judged to be a stable trend. Then, for the dialysate discharge sequence in the corresponding time period, the same processing method is used to compare the discharge volume at the start time of the period with the discharge volume at the end time. If the end value is greater than the start value, it is judged to be an increasing trend. If it is less than, it is judged to be a decreasing trend. If they are equal, it is determined to be a steady trend. After completing the trend judgment, the urea nitrogen concentration trend and the excretion trend are judged one by one. If the two trends are the same, that is, they increase or decrease or stabilize at the same time, the time period is determined to be a trend-consistent segment. Otherwise, it is determined to be a trend-inconsistent segment. For example, in the four time periods within the cycle, the urea nitrogen concentration trends are decreasing, increasing, stabilizing, and decreasing, and the excretion trends are decreasing, increasing, stabilizing, and increasing. The corresponding trend-consistent segments are consistent in the first three time periods and inconsistent in the last time period. The number of statistically consistent segments is three segments. During the execution process, if the change amplitude in a certain time period is less than 0.5mmol / L or the excretion change is less than 0.1L / h during trend judgment, it is determined to be a steady trend. The threshold is set based on the range of conventional physiological fluctuations during dialysis to obtain the number of trend-consistent segments.

[0065] S212: comparing the dialysate discharge change rate and the nitrogen balance trend change based on the number of segments with consistent trends, determining the difference between the two, identifying the time period with the optimal difference, and obtaining the optimal adjustment amplitude sequence;

[0066] The time period information corresponding to the trend-consistent segments is extracted to obtain the dialysate discharge rate change sequence and the urea nitrogen balance trend change sequence. The discharge rate change is the ratio of the discharge change value in each consistent segment to the time period length, and the nitrogen balance trend change is the ratio of the urea nitrogen concentration change value in each consistent segment to the time period length. The two change rates corresponding to each consistent segment are calculated in turn, and then the absolute difference between the discharge rate change rate and the nitrogen balance trend change is calculated for each consistent segment. If the discharge rate change rate in a consistent segment is 0.3L / h and the nitrogen balance trend change is -1.5mmol / L / h, the absolute value of the difference between the two is 1.8. All consistent segments are processed in turn to generate a difference sequence. Next, the difference sequence is compared to screen out the segment with the smallest difference. If there are multiple segments with the same difference, the segment with the larger discharge rate change rate is selected first. The discharge rate change rate in the judgment standard is less than 0.1 L / h is regarded as a low change rate segment and is not selected as the optimal segment. After the optimal segment is selected, the discharge change rate and nitrogen balance trend change corresponding to the segment are respectively used as an element in the optimal adjustment amplitude sequence. If there are multiple optimal segments, the data corresponding to all optimal segments are added to the adjustment amplitude sequence in sequence to finally form the optimal adjustment amplitude sequence. For example, if the consistent segment difference sequence is 1.8, 0.9, and 2.3, the discharge change rates are 0.3L / h, 0.4L / h, and 0.2L / h, respectively, and the nitrogen balance trend changes are -1.5mmol / L / h, -0.5mmol / L / h, and -2.5mmol / L / h, respectively, then the second segment corresponding to the difference of 0.9 is selected as the optimal segment, and the corresponding discharge change rate is recorded as 0.4L / h and the nitrogen balance trend change is -0.5mmol / L / h to form the optimal adjustment amplitude sequence.

[0067] S213: Dynamically optimize the sodium concentration setting based on the optimal adjustment amplitude sequence to match the current nitrogen balance changes using the formula:

[0068] ;

[0069] Get the sodium concentration adjustment range ,in, Indicates the The individual adjustment range within a time period, that is, the basic adjustment range of multiple time periods, represents the number of time periods in the optimal adjustment amplitude sequence, Indicates the The nitrogen balance trend changes during the period.

[0070] The sodium concentration adjustment range refers to the range of adjustment required for the target sodium concentration setting in the dialysate during the dialysis cycle, based on the dynamic change trend of urea nitrogen (nitrogen balance trend) and the analysis of multiple parameters such as dialysate output. That is, the degree to which the sodium concentration in the current cycle should be higher or lower than the basic setting level. It is used to dynamically match the patient's metabolic state and electrolyte balance. It is a quantitative representation of the adjustment direction and range, and is used to guide the equipment to dynamically adjust the sodium concentration parameters, so that the electrolyte balance and hemodynamic state during the dialysis process are more consistent with the patient's metabolic state.

[0071] Call the optimal adjustment amplitude sequence data obtained in the previous section. The sodium concentration adjustment amplitude of the first section in the original collected data is , the nitrogen balance trend changes to , paragraph 2 , , paragraph 3 , , paragraph 4 , , for the participants of different physical dimensions, the normalized adjustment range is used Changes in nitrogen balance trends Calculation is performed, and the corresponding values ​​after normalization are as follows: , , , ; , , , , according to the formula, first calculate the average of the normalized sodium concentration adjustment range:

[0072] ;

[0073] The average of the normalized nitrogen balance trend changes is then calculated:

[0074] ;

[0075] The final combined calculation of sodium concentration adjustment range is:

[0076] ;

[0077] Result calculation:

[0078] ;

[0079] The results show that the normalized amplitude of the dialysate sodium concentration in this cycle needs to be adjusted based on the current level. This value is directly used as the adjustment range of sodium concentration in this cycle, which is used to update the control parameters of the dialysis equipment. The adjustment parameters are kept in line with the dynamic nitrogen balance of the current cycle to ensure the consistency of the adjustment process and the comparability of the cycles. The benefit of the formula is that the adjustment range after normalization is With trend changes The balanced combination eliminates the influence of different dimensions and makes the output results consistent and valuable for period comparison.

[0080] See also Figure 4 ,The specific steps for obtaining physiological linkage warning signals are:

[0081] S311: Based on the sodium concentration adjustment range, compare the change direction of blood flow velocity, heart rate and systolic blood pressure in each period, judge the synchronization of each parameter, calculate the correlation result, and obtain the heart rate-pressure difference consistency coefficient;

[0082] Extract the time period corresponding to each adjustment amplitude in the sodium concentration adjustment amplitude sequence, extract the continuous monitoring data of blood flow rate, heart rate, and systolic pressure for each time period, divide the time period into one minute units, compare the starting value and the ending value of each section to determine the direction of parameter change, and if the ending value is greater than the starting value, it is judged as rising, if it is less than the starting value, it is judged as falling, and if they are equal, it is stable. Then, in the same time period, determine whether the change directions of the three parameters are consistent. If they are consistent, it is marked as complete synchronization. If any two of them are consistent, it is partial synchronization. If the three are inconsistent, it is asynchronous. Count the number of synchronization segments, and then analyze the heart rate. The heart rate-pressure difference consistency is calculated for heart rate and systolic blood pressure. The difference between the absolute value of heart rate change and the absolute value of systolic blood pressure change is calculated for each segment. A segment with a difference less than 5 is considered to be consistent. The ratio of the number of consistent segments to the total number of segments is counted as the consistency ratio. A ratio ≥80% is high consistency, 50%-80% is medium consistency, and <50% is low consistency. For example, a time period is divided into 20 segments, with 12 completely synchronized segments, 5 partially synchronized segments, 3 unsynchronized segments, 16 segments with consistent heart rate-pressure difference, and a consistency ratio of 80%. The final heart rate-pressure difference consistency coefficient is high consistency.

[0083] S312: Based on the heart rate-pressure difference consistency coefficient, the synergy screening is performed on the changes in blood flow velocity and the fluctuations in heart rate and systolic blood pressure in each period using the formula:

[0084] ;

[0085] Get linkage abnormality identification strength , identify the time period when there is abnormal linkage between blood flow rate and physiological parameters, and obtain physiological linkage warning signals, among which, Representative The blood flow rate during the period, Representative The blood flow rate during the period, Representative Heart rate during the period, Representative Systolic blood pressure during the period, Representative Heart rate pressure difference consistency coefficient of the time period, Represents the mean of the heart rate pressure difference consistency coefficient for all periods, Represents the total number of time periods collected during linkage anomaly judgment.

[0086] The abnormal linkage indicator strength is a key numerical result used to quantitatively determine whether there is abnormal linkage between blood flow rate, heart rate, and systolic blood pressure within a certain period of time. It is a comprehensive indicator used to quantitatively measure whether there is abnormal synchronization or abnormal linkage between changes in blood flow rate and changes in heart rate and systolic blood pressure within a certain period of time. It is a direct trigger reference value for judging "whether a physiological linkage warning signal needs to be issued". The larger the value, the stronger the linkage between blood flow rate changes and heart rate / systolic blood pressure and the higher the degree of abnormality. The smaller the value, the weaker the linkage or the normal fluctuation.

[0087] The blood flow rate changes and heart rate and systolic pressure fluctuations in each period are screened collaboratively. First, three consecutive periods in the current dialysis cycle are collected. Blood flow rate , heart rate , systolic blood pressure , heart rate pressure difference consistency coefficient The sampling period is set to 5 minutes, and the sampling values ​​are as follows: blood flow rate Unit: mL / min, heart rate Unit bpm, systolic blood pressure The unit is mmHg, and the normalized values ​​are as follows:

[0088] 、 ;

[0089] 、 ;

[0090] 、 ;

[0091] 、 ;

[0092] 、 ;

[0093] 、 ;

[0094] 、 ;

[0095] 、 ;

[0096] 、 ;

[0097] ;

[0098] ;

[0099] ;

[0100] Calculate the mean heart rate pressure difference consistency coefficient:

[0101] ;

[0102] Execute the formula calculation process, first calculate the product of the blood flow rate change increment and the heart rate pressure difference change:

[0103] ;

[0104] Expand:

[0105] , no increment, recorded as 0;

[0106] , ;

[0107] , ;

[0108] accumulation:

[0109] ;

[0110] Calculate the cumulative absolute difference of the heart rate pressure difference fluctuation term:

[0111] ;

[0112] Denominator calculation:

[0113] ;

[0114] Calculate the cumulative difference between blood flow velocity and heart rate:

[0115] ;

[0116] Expand:

[0117] ;

[0118] Final linkage abnormality identification strength calculate:

[0119] ;

[0120] The results show that the linkage abnormality indicator strength in the three consecutive periods of the current dialysis cycle is According to the physiological linkage warning signal judgment benchmark interval [0.2, 0.4], the current result is at the lower edge of the interval. It is determined that a physiological linkage warning signal is generated at this stage, and subsequent dynamic tracking of linkage risks is triggered. The formula dynamically adjusts the influence of the denominator by superimposing the trend of the product term of the blood flow rate change increment and the heart rate pressure difference change, combining the heart rate pressure difference fluctuation normalization term, enhancing the linkage risk sensitivity between the blood flow rate and heart rate change trends, and effectively improving the reliability and transparency of the physiological linkage warning signal generation and judgment.

[0121] See also Figure 5 , the steps for obtaining the ultrafiltration parameter control baseline are as follows:

[0122] S411: Based on the physiological linkage warning signal, optimize the blood volume monitoring information of the relevant stages before and after the dialysis cycle, compare the blood volume change trends corresponding to each time period, and calculate the blood volume change rate sequence of continuous time periods;

[0123] Extract blood volume monitoring information within 30 minutes before and after the warning signal occurs, divide the time period into one-minute units, extract the corresponding blood volume value, read the first and last blood volume values ​​of each segment in turn to determine the direction of change. If the tail value is greater than the first value, it is judged to be rising; if it is less than the first value, it is falling; if it is equal, it is judged to be stable. Then compare the blood volume change trends of each time period, and compare the change directions of adjacent time periods one by one. If the direction of three or more consecutive time periods is consistent, it is determined to be a continuous change trend segment. Count the number and duration of such continuous segments, and then calculate the blood volume change rate of each time period. The tail value minus the first value and then divided by the time period length is obtained. The rate value is determined as a rapid increase if it is greater than 0.02% / min, a stable state if it is greater than -0.02% / min and less than 0.02% / min, and a rapid decrease if it is less than -0.02% / min. For example, if the blood volume values ​​for five consecutive one-minute segments are 70%, 69.5%, 69%, 68.5%, and 68%, and the change trend is consistently downward, and the calculated rates are -0.5%, -0.5%, -0.5%, and -0.5%, respectively, it is determined to be a continuous rapid decrease trend segment. The blood volume change rates of each consecutive time segment are summarized in sequence to form a blood volume change rate sequence for the consecutive time segments.

[0124] S412: Based on the blood volume change rate sequence, identify the time period corresponding to the associated risk type according to the risk identifier, and compare it with the risk trend discriminant using the formula:

[0125] ;

[0126] Get the ultrafiltration adjustment coefficient ,in, Indicates the The rate of change of blood volume in the stage Indicates the The trend item of the risk trend discriminant of the stage, is the risk trend sensitive factor, is the mean rate of change of blood volume during the dialysis cycle, represents the total number of analysis phases;

[0127] Call the constructed blood volume change rate sequence, compare the rate data of each stage with the risk trend discrimination value one by one, identify the stage corresponding to each risk, combine the preset risk trend sensitivity factor, calculate the degree of difference between it and the current rate, amplify the difference through square processing, and then introduce the square root of the mean term of the blood volume change rate in the current cycle to enhance the intervention weight of the overall rate change level of the current cycle, forming an integrated adjustment factor. Taking actual data as an example, let the three-stage normalized blood volume change rate sequence be , the corresponding risk trend discriminant is , risk trend sensitive factor The average periodic rate is , substituting into the formula we can get:

[0128] ;

[0129] ;

[0130] ;

[0131] The results show that the deviation between blood volume changes and risk trends in the current dialysis cycle is small, and the overall regulation demand intensity is in the medium range. It is suitable to intervene by fine-tuning the ultrafiltration parameters, which can serve as the decision-making basis for parameter adjustment in subsequent steps. The formula emphasizes the degree of deviation between the rate and trend in different stages by means of squared differences, and uses the square root term as a harmonizing factor for the overall fluctuation of the cycle, effectively avoiding the excessive impact of local deviations on overall regulation.

[0132] S413: Based on the ultrafiltration adjustment coefficient, adjust the ultrafiltration volume parameter of the current cycle of the dialysis device and the dialysis duration setting parameter, optimize the current parameter combination, and obtain the ultrafiltration parameter control baseline.

[0133] The ultrafiltration adjustment coefficient is a comprehensive quantitative indicator used to guide the adjustment amplitude and direction of the ultrafiltration volume parameters and dialysis duration setting parameters in the current cycle of the dialysis device. "Ultrafiltration" during the dialysis process refers to the process of removing excess water from the patient's body through the dialysis device. The ultrafiltration volume parameter determines how much fluid needs to be removed in a dialysis cycle, and the dialysis duration setting affects the rate and smoothness of fluid removal. However, in actual practice, the patient's blood volume change rate will be affected by factors such as individual metabolic status, cardiovascular response capacity, and vascular capacity adaptability. There will be fluctuations and even risk trends at different stages. If a larger ultrafiltration volume or a shorter dialysis duration is blindly set, it will induce risks such as hypotension, arrhythmia, and insufficient organ perfusion. In essence, as a dimensionless normalization coefficient, it is directly used to dynamically guide the "regulation intensity" of ultrafiltration parameter adjustment to ensure that the equipment control strategy is more in line with the patient's real-time physiological state and risk trend changes within the cycle.

[0134] The device's set ultrafiltration target volume and dialysis duration for the current cycle are read, and the specific value of the ultrafiltration adjustment coefficient is extracted. If the adjustment coefficient is 1.1, it indicates that the original ultrafiltration target volume should be increased by 10%. The adjusted ultrafiltration volume is multiplied by the current ultrafiltration target volume to obtain the adjusted ultrafiltration volume. If the original ultrafiltration volume is 2000 mL, the adjusted ultrafiltration volume is 2200 mL. The dialysis duration is then adjusted. If the adjustment coefficient is greater than 1.05 and there are more than five consecutive rapid decline trend segments in the blood volume change rate sequence, the dialysis duration is extended by 10%. If the original setting is 240 minutes, it is adjusted to 264 minutes. If the adjustment coefficient is less than 0.95 and the proportion of consecutive rapid increase trend segments in blood volume exceeds 50%, the dialysis duration is shortened by 10%. If there are no obvious abnormal trends, the original duration is maintained. The adjusted ultrafiltration volume and adjusted dialysis duration form the optimized parameter combination for the current cycle and serve as the baseline for ultrafiltration parameter control.

[0135] See also Figure 6 , the specific steps for obtaining device control instructions are:

[0136] S511: Based on the ultrafiltration parameter control baseline, analyze the linkage between blood flow rate, dialysate sodium concentration, dialysis duration, heart rate, systolic blood pressure, and blood volume change rate, compare the correlation between each parameter and the risk trend discriminant, optimize the coordination relationship between parameters, and obtain the parameter priority ranking results;

[0137] First, extract the time series data corresponding to the six parameters of blood flow rate, dialysate sodium concentration, dialysis time, heart rate, systolic blood pressure, and blood volume change rate in the current cycle, and uniformly map them to the same time axis. Divide the analysis window into a time period of 5 minutes. For each time period, calculate the change trend of each parameter in the current segment. If the end value of the segment is higher than the beginning value of the segment, it is judged as rising, if it is lower, it is judged as falling, and if they are equal, it is stable. Then, read the direction of risk trend change in the corresponding time period for the risk trend judgment amount, and make synchronous judgments in the same time period window. Compare the consistency of each parameter change trend with the risk trend direction. If the change direction is consistent, assign a value of 1, otherwise assign a value of 0. Process all parameters in turn, and accumulate the consistency assignment results for each time period to obtain the results of each parameter. The consistency score between the parameter and the risk trend discriminant within the segment is then summed up for all time periods within the entire cycle to calculate the total consistency score of each parameter. The total consistency score is then divided by the total number of time periods to obtain the correlation strength percentage of each parameter. The correlation strength interval judgment standard is set, ≥80% is a strong correlation, 50%-80% is a medium correlation, and <50% is a weak correlation. For example, if the total consistency score of blood flow rate is 18 and the total number of time periods is 20, the correlation strength is 90%, which is classified as a strong correlation. The remaining parameters are classified in turn and then sorted from high to low according to the correlation strength percentage. If the sorting result is blood flow rate > blood volume change rate > systolic blood pressure > dialysate sodium concentration > heart rate > dialysis time, the parameter priority sorting result is formed.

[0138] S512: Determine the first parameter item in the parameter priority ranking result, adjust the control parameters of the dialysis equipment, coordinate the current dialysate discharge volume, blood flow rate change trend and sodium concentration setting, and obtain equipment control instructions;

[0139] First, read the sorting result list and extract the first-ranked parameter. If the first parameter is blood flow rate, read the current device blood flow rate control parameter and compare its current setting value with the historical optimal setting value of the previous cycle. If the current value deviates from the historical optimal value by more than 10%, adjust the current setting value by 5% toward the historical optimal value. For example, the current setting value is 300mL / min, the historical optimal value is 330mL / min, and the current deviation is -9%, which is less than the 10% threshold. Maintain the current setting. If the deviation is 15%, adjust the current setting value +5%, that is, adjust it to 315mL / min. If the first parameter is dialysate sodium concentration, execute the same adjustment logic. If it is systolic pressure , then adjust the blood flow rate or sodium concentration of the device to indirectly regulate blood pressure. During the adjustment process, coordinate the current trend of dialysate discharge. If the discharge trend continues to decline and the heart rate trend increases, the blood flow rate adjustment range is limited to ±5% first. If the discharge trend increases and the heart rate is stable, the adjustment range is relaxed to ±10%. At the same time, adjust the sodium concentration setting according to the dialysate sodium concentration target value. If the current setting is lower than 138mmol / L and needs to be increased, adjust it by +2mmol / L each time, and the adjustment shall not exceed the upper limit of 142mmol / L. Generate a device control instruction containing the current adjusted blood flow rate, sodium concentration setting, and dialysate discharge coordination parameters.

[0140] An AI-based dialysis plan optimization method, comprising:

[0141] S1: Based on continuous dialysis cycle data, analyze the changes in urea nitrogen concentration, compare the increase and decrease directions of urea nitrogen concentration in each cycle, combine the changes in dialysate discharge volume in each cycle, calculate the correlation between urea nitrogen concentration fluctuations and dialysate discharge volume, determine the relationship type, and obtain the cycle nitrogen dynamic characteristics;

[0142] S2: Based on the dynamic characteristics of cyclic nitrogen, the synchronous characteristics of the change of dialysate discharge volume with the nitrogen balance trend are analyzed. By comparing the trend changes within the cycle, the adjustment range of the dialysate sodium concentration parameter is calculated, and the sodium concentration setting is dynamically optimized to match the current nitrogen balance changes to obtain the sodium concentration adjustment range;

[0143] S3: Based on the adjustment range of sodium concentration, analyze the impact of continuous changes in blood flow rate, compare the fluctuation trends of heart rate and systolic blood pressure in the same time period, determine the correlation between changes in blood flow rate and synchronous fluctuations of heart rate and systolic blood pressure, identify periods of abnormal linkage between blood flow rate and physiological parameters, and obtain physiological linkage warning signals;

[0144] S4: Based on the physiological linkage warning signal, determine the risk type it reflects, analyze the blood volume change rate before and after the dialysis cycle, compare the correspondence between the risk trend judgment value and the blood volume change, adjust the ultrafiltration volume parameters and dialysis duration settings, optimize the current parameter combination, and obtain the ultrafiltration parameter control baseline;

[0145] S5: Based on the ultrafiltration parameter control baseline, analyze its effect in the current cycle, calculate the priority order of each control action, determine the parameter item with the highest priority, and adjust the control parameters of the dialysis equipment accordingly to obtain the equipment control instructions.

[0146] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. An AI-based dialysis plan optimization system, characterized in that: include: The nitrogen dynamic feature recognition module analyzes the changes in urea nitrogen concentration based on continuous dialysis cycle data, compares the increase and decrease directions of each cycle, and combines the changes in dialysate discharge volume to calculate the correlation between urea nitrogen concentration fluctuations and dialysate discharge volume, determine the relationship type, and obtain the cycle nitrogen dynamic feature; The sodium control parameter generation module analyzes the synchronous changes of the dialysate discharge and the nitrogen balance trend based on the cyclic nitrogen dynamic characteristics, compares the cyclic trends, calculates the sodium concentration adjustment range, dynamically optimizes the sodium concentration setting, and obtains the sodium concentration adjustment range; The linkage risk identification module analyzes the impact of continuous changes in blood flow rate based on the sodium concentration adjustment range, compares the fluctuation trends of heart rate and systolic blood pressure, determines the correlation of synchronous fluctuations, identifies abnormal linkage periods, and obtains physiological linkage warning signals; The ultrafiltration baseline setting module determines the risk type based on the physiological linkage warning signal, analyzes the blood volume changes before and after the dialysis cycle, compares the corresponding relationship between the risk judgment amount and the blood volume, adjusts the ultrafiltration volume and dialysis time, and obtains the ultrafiltration parameter control baseline; The intelligent instruction output module analyzes the effect of the ultrafiltration parameter control baseline in the current cycle, calculates the control action priority, determines the top parameter item, adjusts the dialysis equipment control parameters, and obtains the equipment control instruction; The device control instruction includes an action instruction code, a priority number, and an implementation parameter set; The steps for obtaining the device control instruction are specifically as follows: S511: Based on the ultrafiltration parameter control baseline, analyzing the linkage status between blood flow rate, dialysate sodium concentration, dialysis duration, heart rate, systolic blood pressure, and blood volume change rate, comparing the correlation strength between each parameter and the risk trend discriminant, optimizing the coordination relationship between the parameters, and obtaining the parameter priority ranking result; S512: Determine the first parameter item in the parameter priority ranking result, adjust the control parameters of the dialysis equipment, coordinate the current dialysate discharge, blood flow rate change trend and sodium concentration setting, and obtain equipment control instructions.

2. The AI-based dialysis plan optimization system according to claim 1, characterized in that: The cyclical nitrogen dynamic characteristics include metabolic fluctuation indicators, cycle correlation factors, and trend classification identifiers; the sodium concentration adjustment range includes parameter setting range, adjustment priority, and change type; the physiological linkage warning signal includes risk identifiers, synchronous reaction markers, and abnormal event types; and the ultrafiltration parameter control baseline includes baseline control parameters, cycle adaptation categories, and risk management elements.

3. The AI-based dialysis plan optimization system according to claim 1, characterized in that: The steps for obtaining the periodic nitrogen dynamic characteristics are specifically as follows: S111: Based on the continuous dialysis cycle data, the changes in urea nitrogen concentration at each stage are analyzed. By comparing the concentration increase and decrease directions of adjacent cycles, the change trend of metabolites is determined and the concentration change sequence is obtained; S112: comparing the concentration change sequence with the change process of the dialysate discharge volume in the corresponding cycle, analyzing the synchronization and deviation between the two, determining the correlation of the changes within the cycle, and obtaining a cycle correlation feature identifier; S113: calling the cycle-related feature identifier, combining the various change trends within the cycle, classifying the cycle feature type, optimizing the cycle feature expression, and obtaining the cycle nitrogen dynamic feature.

4. The AI-based dialysis plan optimization system according to claim 1, characterized in that: The steps for obtaining the sodium concentration adjustment range are specifically as follows: S211: Based on the cyclic nitrogen dynamic characteristics, comparing the change trends of the dialysate discharge volume series and the urea nitrogen concentration series in each time period, determining whether the change directions of the two are consistent, selecting time segments with consistent change trends, and obtaining the number of segments with consistent trends; S212: comparing the dialysate discharge change rate and the nitrogen balance trend change based on the number of segments with consistent trends, determining the difference between the two, identifying the time period with the optimal difference, and obtaining the optimal adjustment amplitude sequence; S213: Dynamically optimize the sodium concentration setting according to the optimal adjustment range sequence to match the current nitrogen balance change and obtain the sodium concentration adjustment range.

5. The AI-based dialysis plan optimization system according to claim 1, characterized in that: The steps for obtaining the physiological linkage warning signal are specifically as follows: S311: Based on the sodium concentration adjustment range, comparing the change directions of blood flow velocity, heart rate, and systolic blood pressure in each time period, determining the synchronization of each parameter, calculating the correlation result, and obtaining the heart rate-pressure difference consistency coefficient; S312: Based on the heart rate-pressure difference consistency coefficient, a collaborative screening is performed on the changes in blood flow rate in each time period and the fluctuations in heart rate and systolic pressure to obtain the linkage abnormality identification strength, identify the time period when there is linkage abnormality between blood flow rate and physiological parameters, and obtain a physiological linkage warning signal.

6. The AI-based dialysis plan optimization system according to claim 1, characterized in that: The steps for obtaining the ultrafiltration parameter control baseline are specifically as follows: S411: Based on the physiological linkage warning signal, optimizing the blood volume monitoring information of the associated stages before and after the dialysis cycle, comparing the blood volume change trends corresponding to each time period, and calculating the blood volume change rate sequence of the continuous time period; S412: Based on the blood volume change rate sequence, according to the risk identifier, identifying the time period corresponding to the associated risk type, comparing it with the risk trend discriminant, and obtaining the ultrafiltration adjustment coefficient; S413: Based on the ultrafiltration adjustment coefficient, adjust the ultrafiltration volume parameter of the current cycle and the dialysis duration setting parameter of the dialysis device, optimize the current parameter combination, and obtain the ultrafiltration parameter control baseline.

Images