A hemodialysis information management and monitoring system

Abstract

The application relates to the technical field of medical information processing, and discloses a hemodialysis information management and monitoring system, which is used for solving the problems that in the traditional method, hemodialysis machine high-frequency operation data and patient low-frequency physiological data cannot be effectively time-sequentially corresponded and associated analyzed, and the real-time linkage regulation and control capability of the dialysis equipment is lacked; the system firstly connects the dialysis equipment operation parameters and the patient physiological parameters, sequentially completes session filing, time correspondence, state association, risk judgment, equipment regulation and execution review, and updates the patient individual tolerance parameters and the key monitoring interval according to the review result, is used for multi-source data association analysis and equipment linkage control in the hemodialysis process, and realizes timely identification, accurate regulation and closed-loop management of hemodynamics abnormalities.

Description

Technical Field

[0001] The present invention relates to the technical field of medical information processing, and specifically relates to a hemodialysis information management and monitoring system. Background Art

[0002] During the hemodialysis process, the patient's body fluid is continuously removed by ultrafiltration, and physiological states such as blood pressure, heart rate and circulatory load change dynamically with the treatment process; if monitoring and intervention are not timely, hemodynamic abnormalities such as intradialytic hypotension are easily induced; in the prior art, there have been related schemes for dialysis process monitoring and early warning; for example, the published invention patent CN108986419B discloses a data alarm method for hemodialysis, which detects dialysis parameters in real time, compares the difference between the detection results and the parameters at corresponding time points of previous dialysis, and issues an alarm when the difference exceeds a threshold; for another example, the published invention patent CN113289098B discloses a dialysate monitoring method and device based on a hemodialysis apparatus, which generates early warning information through dialysis risk assessment, and prompts medical staff to adjust the temperature, flow rate or electrolyte composition of the dialysate; Although the above schemes can improve the abnormality prompting capability during the dialysis process to a certain extent, they generally still mainly stay at the levels of parameter monitoring, threshold judgment or risk reminding; in the actual hemodialysis process, the operating parameters of the dialysis machine are usually high-frequency data collected continuously, while the physiological signs such as the patient's blood pressure and heart rate are mostly low-frequency data collected at intervals, and the two belong to data of different sources and different frequencies; in the prior art, without effective time sequence correspondence and correlation analysis of the above data, it is often difficult to accurately reflect the change relationship between the dialysis dehydration process and the human physiological response, so that it is difficult to detect the early signs of the patient's hemodynamic abnormality in time, and problems such as delayed early warning and high false alarm rate are prone to occur; on the other hand, existing systems mainly focus on monitoring and alarming, and after identifying risks, medical staff are usually still required to manually judge and adjust the dialysis equipment, lacking the capability of real-time linked regulation and control of the dialysis equipment according to changes in risks, which easily leads to a time delay between early warning information and equipment adjustment, and affects the timeliness and accuracy of clinical intervention; therefore, how to realize effective correlation analysis of multi-source data during the hemodialysis process and improve the timeliness of equipment regulation and control on the basis of risk identification has become an urgent technical problem to be solved in the field. Summary of the Invention

[0003] In view of the deficiencies of the prior art, the present invention provides a hemodialysis information management and monitoring system, which solves the problems existing in traditional methods that it is difficult to perform effective time sequence correspondence and correlation analysis on high-frequency operating data of the dialysis machine and low-frequency physiological data of the patient, and the capability of real-time linked regulation and control of the dialysis equipment is lacking.

[0004] To achieve the above objectives, the present invention provides the following technical solutions: A hemodialysis information management and monitoring system, comprising: Data access and archiving module: used to collect dialysis equipment parameters and patient physiological parameters, and to create data archives; Time alignment module: used to perform time-correlation processing on dialysis equipment parameters and patient physiological parameters; Status correlation module: used to perform correlation analysis between dialysis equipment parameters and patient physiological parameters after time correspondence; Risk assessment module: used to determine the risk status and trend of change based on the results of correlation analysis; Equipment control module: Used to generate control instructions for dialysis equipment based on risk status and changing trends; The execution review module is used to obtain and review the equipment feedback and monitoring feedback after adjustment.

[0005] Preferably, the data access and filing module includes: A dialysis session is established based on the blood pump operating status and the first non-invasive blood pressure result. A session identifier is generated and a session master table is established, which includes the patient's identity identifier, bed identifier, dialysis equipment identifier, monitoring equipment identifier, session start time, and basic prescription information. Connect the dialysis equipment's operating parameters and physiological monitoring parameters, write the connected data into the raw acquisition buffer, and establish a raw data record table; A parameter mapping table is established based on the device protocol field encoding, parameter units, and clinical semantics. Link interruption determination, interruption continuation, and session merging are performed on the data from the device side and the physiological side.

[0006] Preferably, the timing alignment module includes: Synchronous detection and original timestamp correction of raw data from dialysis and monitoring equipment; Based on the device correlation values ​​within the sliding window and adjacent real non-invasive blood pressure measurement results, a continuous blood pressure sequence is deduced, individual hysteresis time parameters are identified, and time-shifted matching is performed on the device parameter stream and physiological parameter stream. Continuous physiological simulations should be stopped when key equipment parameters are missing or the interval between physiological reference data exceeds the limit.

[0007] Preferably, the state association module includes: The synchronous records were organized into a status-related record table, and information on ultrafiltration changes, pulse pressure changes, transmembrane pressure changes, and heart rate changes were extracted according to different observation windows of different durations. Trend separation was performed on the estimated systolic blood pressure and heart rate sequences, and volume-pressure difference markers, pulse pressure narrowing markers, compensation enhancement markers, and compensation decay markers were generated by combining the device action characteristics. Add dehydration stage markers and retain original equipment parameters, baseline physiological parameters, and finite state markers when the reliability of physiological reference data is insufficient.

[0008] Preferably, trend separation is performed on the estimated systolic blood pressure sequence and heart rate sequence, and volume-pressure difference markers, pulse pressure narrowing markers, compensatory enhancement markers, and compensatory attenuation markers are generated in combination with device action characteristics, including: Modal decomposition was performed on systolic blood pressure estimation data and heart rate data to extract low-frequency main trends; Map the ultrafiltration operation curve, the cumulative ultrafiltration volume change curve, and the transmembrane pressure change curve to the same observation window; By combining the relationship between pulse pressure changes, mean arterial pressure changes, heart rate changes, and cumulative ultrafiltration volume changes, volume-pressure difference markers, pulse pressure narrowing markers, compensatory enhancement markers, and compensatory attenuation markers are generated according to numerical conditions and trends.

[0009] Preferably, the risk assessment module includes: Based on the state association sequence, the volume compensation consumption divergence index, vascular compliance decay index and compensation exhaustion slope are extracted within the sliding decision window and normalized and weighted to form a risk confidence index. The risk threshold for each stage is dynamically set based on the target dehydration completion rate, and the risk status is determined by combining the duration and rate of change. An intervention urgency index is formed based on the risk confidence index, risk change trend, estimated systolic blood pressure and remaining dry weight deficit. Temporary encrypted monitoring is implemented for the yellow observation status. If the data is unreliable, the previous effective risk status is maintained and marked as pending review.

[0010] Preferably, risk thresholds for each stage are dynamically set based on the target dehydration completion rate, and the risk status is determined by combining the duration and rate of change, including: The dehydration process is divided into early, middle, and final stages based on the target dehydration completion rate, and corresponding risk thresholds are set accordingly. The risk threshold in the mid-term stage tightens linearly with the completion rate of the target dehydration. The risk status is determined based on the duration of the risk confidence index exceeding the stage risk threshold and the smooth rate of change.

[0011] Preferably, the equipment control module includes: Based on the risk level, corresponding control strategies are invoked. Under moderate risk conditions, the ultrafiltration rate and dialysate sodium concentration are jointly adjusted, while under severe risk conditions, the ultrafiltration rate and dialysate temperature are protected and adjusted. The control content is written into the control data packet and sent out via protocol framing. Based on the equipment acknowledgments, resend status, and deviations between target parameters and actual execution status, manual takeover or anomaly review is performed.

[0012] Preferably, the control content is written into the control data packet and sent via protocol framing, including: The control parameters are organized into control data packets, which include the dialysis session identifier, target parameter name, target parameter value, transition time, priority, generation time, validity period, and checksum. The edge gateway completes instruction framing according to the communication protocol corresponding to the target device. The effective time limit is determined based on the transition time and observation time. The priority of medium-risk control instructions is lower than that of high-risk control instructions. The instruction framing result is converted into a control message that the device can recognize and then sent.

[0013] Preferably, the execution review module includes: Based on the equipment feedback and monitoring feedback after the control measures take effect, an independent review data sequence is formed, and time correspondence, status association and risk assessment are re-executed according to the short-term review window and the total review window. Based on the review and judgment criteria, implement step-by-step restoration of control, secondary regulation, or manual emergency takeover; The review information is written into the review record form, and the patient's individual tolerance parameters and key monitoring intervals are updated based on the review record.

[0014] Compared with the prior art, the present invention provides a hemodialysis information management and monitoring system, which has the following beneficial effects: 1. This invention, by establishing a session record, unifying time correspondence, and associating multiple duration states between high-frequency operating parameters of dialysis equipment and low-frequency physiological parameters of patients, can more accurately reflect the correspondence between changes in equipment operation and physiological responses such as blood pressure and heart rate during continuous dehydration, reducing judgment bias caused by different data sources and sampling frequencies; and by continuously assessing the risks of state changes, after identifying abnormal trends, it can link the dialysis equipment to perform graded adjustments of ultrafiltration rate, dialysate sodium concentration, or dialysate temperature, and then combine the verification and parameter write-back update after adjustment, which can shorten the time interval between early warning and treatment, reduce the lag and error caused by relying solely on manual judgment, thereby achieving timely identification of hemodynamic abnormalities and closed-loop control of equipment during hemodialysis.

[0015] 2. This invention, by continuously acquiring equipment and monitoring feedback after the issuance of control commands, and re-performing risk reviews, control restoration, and secondary control judgments according to short-term and total review windows, avoids directly using a single control action as the treatment endpoint, reducing subsequent risk omissions caused by equipment execution deviations, short-term fluctuations, or inaccurate single judgments. Simultaneously, by writing the review results into the review record and using them for subsequent treatment parameter correction, the risk thresholds, ultrafiltration rate adjustment ranges, and key monitoring intervals at different dehydration stages can be gradually adjusted according to the individual patient's tolerance, thereby achieving continuous optimization and individualized control during dialysis treatment. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the structure of a hemodialysis information management and monitoring system according to the present invention; Figure 2 This is a schematic diagram showing the time correspondence between high-frequency equipment parameters and low-frequency physiological parameters of the present invention; Figure 3 This is a schematic diagram of the state association and risk determination logic of the present invention; Figure 4 This is a curve showing the dynamic adjustment of the risk threshold during the dehydration stage of this invention. Figure 5 This is a schematic diagram of the closed-loop control and execution verification of the equipment in this invention; Figure 6 This is a schematic diagram of the deployment structure of the clinical dialysis monitoring system of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1: Figure 1 - Figure 5 A hemodialysis information management and monitoring system is presented, including: Data access and archiving module: used to collect dialysis equipment parameters and patient physiological parameters, and to create data archives; Time alignment module: used to perform time-correlation processing on dialysis equipment parameters and patient physiological parameters; Status correlation module: used to perform correlation analysis between dialysis equipment parameters and patient physiological parameters after time correspondence; Risk assessment module: used to determine the risk status and trend of change based on the results of correlation analysis; Equipment control module: Used to generate control instructions for dialysis equipment based on risk status and changing trends; The execution review module is used to obtain and review the equipment feedback and monitoring feedback after adjustment.

[0019] This system can be deployed in a clinical dialysis monitoring environment consisting of a bedside hemodialysis treatment terminal, an independent multi-parameter physiological monitor, a medical edge computing core gateway, and an in-hospital electronic medical record interface. The system operates for patients receiving maintenance hemodialysis treatment and works collaboratively with dialysis equipment performing extracorporeal circulation and ultrafiltration dehydration, monitoring equipment collecting physiological signs such as non-invasive blood pressure and heart rate, and edge gateways responsible for cross-protocol data processing and control distribution. The information accessed by the system includes at least the high-frequency operating parameters continuously output by the dialysis equipment and the low-frequency physiological parameters periodically uploaded by the monitoring equipment. The system is used to establish a unified time correlation and dynamic state link for the patient's treatment process within the same dialysis session and generate control instructions for the dialysis equipment when risk conditions are met. The system output results include at least the dialysis session data archive, the synchronous state sequence corresponding to the completion time, the risk status identifier, the dialysis equipment control instructions, and the review results after control.

[0020] Specifically, such as Figure 2 As shown: The data access and record-keeping module is used to synchronously access dialysis equipment operation information and patient physiological monitoring information at the beginning of a single hemodialysis treatment, and form a session-level data archive for that treatment process; the edge computing core gateway continuously monitors the operation status of the dialysis equipment and the upload status of the monitoring equipment. When it detects that the blood flow rate reaches 100mL / min after the blood pump is started and is maintained for 5 seconds, and the first non-invasive blood pressure measurement result has been received within 120 seconds before and after this point, it is determined that the current treatment has entered the formal record-keeping state; 100mL / min is set as the record-keeping trigger flow rate because when it is lower than this value, it is usually still in the initial blood drawing stage, the tubing filling stage, or the equipment output ramp-up stage, and the blood flow, pressure, and ultrafiltration are still in the initial blood drawing stage, the tubing filling stage, or the equipment output ramp-up stage. The flow parameters fluctuate greatly, making it unsuitable as the starting point for a dialysis session. To accommodate the differences in initial blood flow settings among different patients, the trigger flow rate for record establishment can be selected from 80 mL / min to 120 mL / min. The continuous maintenance time is set to 5 seconds because the device parameters can be updated in 1-second cycles. When the flow conditions are met at 5 consecutive sampling points, false triggers caused by transient pulsations, short-term drops, or switching jitter can be eliminated. The continuous maintenance time can be selected from 3 seconds to 10 seconds. The allowable interval between the first non-invasive blood pressure measurement result and the record establishment time is set to 120 seconds to ensure that the initial physiological data can characterize the starting state of the current dialysis session and to avoid using historical measurement results with excessively long time intervals as the initial baseline data point. After the registration conditions are met, the core gateway generates a dialysis session identifier corresponding to this treatment and establishes a session master table. The session master table includes at least the following fields: patient identity identifier, bed identifier, dialysis equipment identifier, monitoring equipment identifier, session start time, and basic prescription information. The basic prescription information includes at least the following fields: dry weight, target dehydration volume, prescription duration, basic sodium concentration of dialysate, and basic temperature of dialysate. Dry weight refers to the target weight when the patient's excess water has been largely removed and the circulating volume is in a relatively stable state. It serves as a reference benchmark for setting the dehydration volume and calculating the remaining dehydration volume during dialysis treatment, defining the basic operating boundary of this treatment, and providing background parameters for subsequent time correspondence, state association, risk assessment, and equipment adjustment. Preparatory data accessed before the session master table is established is temporarily stored in the pending registration cache area. After the session is established, it is merged into the current session according to the dialysis session identifier to avoid data from different treatment periods being mixed in the same data link. After the session master table is established, the core gateway initiates the device data acquisition process and the physiological data acquisition process respectively. The device data acquisition process reads the operating parameters of the dialysis equipment through the serial communication interface, industrial bus interface, or the data communication interface open by the dialysis equipment. The acquired data includes at least the following parameters: extracorporeal blood volume, instantaneous ultrafiltration rate, cumulative ultrafiltration volume, dialysate conductivity, dialysate temperature, venous pressure, and transmembrane pressure. The sampling period for the device parameters can be selected as 1 second. 1 second is chosen as the preferred sampling period because the main mechanical switching cycles related to fluid balance or flow regulation inside conventional dialysis equipment are usually in the range of 1 second to 1.2 seconds. Using 1 second sampling can cover the main changes in equipment operation during continuous dehydration. When the sampling period is greater than 1.2 seconds, it may miss key state changes by spanning a complete switching process. When the sampling period is significantly less than 0.5 seconds, it will significantly increase the transmission, buffering, and processing burden, while the actual gain for bedside real-time monitoring is limited. Therefore, 1 second can be selected as the sampling period on the device side. The physiological data acquisition process receives physiological parameters such as systolic blood pressure, diastolic blood pressure, mean arterial pressure, heart rate, and blood oxygen saturation uploaded by monitoring equipment through the hospital's medical information interface. The non-invasive blood pressure sampling cycle can be selected from 600s to 1200s, with 900s being the preferred value. 600s is set as the lower limit to avoid excessive inflation and deflation of the cuff-type non-invasive blood pressure measurement, which could continuously interfere with the patient's local circulation and increase discomfort. 1200s is set as the upper limit to ensure that at least three real physiological baseline data points are obtained within the subsequent one-hour observation window, thus supporting subsequent time-corresponding processing. 900s is selected as the preferred value to balance the frequency of routine bedside non-invasive blood pressure measurement, patient comfort, and the requirements of subsequent processing for the density of physiological baseline data points. Heart rate and blood oxygen saturation are received according to the original output cycle of the monitoring equipment and a time correspondence is established with the most recent non-invasive blood pressure record during writing. When the monitoring equipment only reports physiological data packets periodically, heart rate and blood oxygen saturation are received along with the physiological data packet. The data received for the first time is not directly written to the analysis data area, but is first written to the raw acquisition buffer area, and a raw data record table is created simultaneously. The raw data record table records data according to data source, sampling time, field type, and verification status. Each record includes at least the following fields: dialysis session identifier, source device identifier, original timestamp, parameter name, parameter value, quality flag, and storage index information. The quality flag can be divided into four categories: valid, placeholder, verification failed, and link interruption. Valid means that the data has passed the interface verification and can enter subsequent processing. Placeholder means that the position is temporarily filled by the most recent valid value. Verification failed means that the data frame has been received but the field integrity or value validity has not passed the verification. Link interruption means that the corresponding source data has not been received within the specified time. Through the cooperation of the raw acquisition buffer area and the raw data record table, the high-frequency mechanical operation parameters output by the dialysis equipment and the low-frequency physiological parameters output by the monitoring equipment can be clearly distinguished in subsequent processing, and a traceable data foundation can be provided for time correspondence, anomaly backtracking, and result verification. To avoid confusion in subsequent processing caused by different brands of equipment, different protocol fields, or different parameter naming methods, the system further establishes a parameter mapping table. The parameter mapping table maps the same physical quantities to a unified field index based on the equipment protocol field code, parameter unit, and clinical semantics. For example, the blood pump flow rate field recorded in different dialysis equipment protocols is uniformly mapped to the blood flow rate field index, and the non-invasive systolic blood pressure field recorded in different monitoring equipment protocols is uniformly mapped to the systolic blood pressure field index. After processing through the parameter mapping table, data from different sources can be stored under a unified field structure, which facilitates subsequent content to be time-corresponding and status-analyzed in a unified format. In terms of anomaly handling, the system sets up interruption judgment mechanisms for both the equipment side and the physiological side. For dialysis equipment data, if no dialysis equipment running frames are received for three consecutive equipment sampling cycles, a equipment data interruption flag is written. Since the equipment sampling cycle can be selected as 1 second, three consecutive sampling cycles correspond to 3 seconds of no running frames being returned. This duration usually exceeds the duration of single frame jitter, short-term buffer delay, or instantaneous link congestion, which can better distinguish between occasional transmission anomalies and continuous interruptions. The equipment data interruption judgment threshold can be selected from 2 to 5 seconds. For physiological data, if no valid measurement results are received for two consecutive non-invasive blood pressure sampling cycles, a physiological acquisition interruption flag is written. Based on the preferred sampling cycle of 900 seconds, two consecutive cycles correspond to 30 minutes without obtaining new real non-invasive blood pressure results. A single missing result may be due to cuff loosening, a single measurement failure, nursing operation interference, or upload delay, and it is not appropriate to directly judge the link interruption. However, two consecutive missing cycles are enough to affect the subsequent processing and risk judgment. Therefore, two physiological sampling cycles are used as the physiological acquisition interruption judgment condition. When physiological data is missing for a single time but has not yet reached the interruption judgment condition, the system calls the most recent valid measurement value as a placeholder reference and writes a placeholder quality mark simultaneously. The placeholder reference is only used to maintain the continuity of the session record and the integrity of the cache structure, and is not used as a low-frequency physiological baseline data point in subsequent risk judgment and time correspondence processing. The placeholder duration can be selected to be no more than one non-invasive blood pressure sampling cycle. When the preferred sampling cycle value is 900s, the generation of a new placeholder reference stops after the placeholder duration exceeds 900s, and waits for the next real measurement result to be restored. Through this processing method, the substitution value can be avoided from being mistakenly introduced into the subsequent core processing link as real physiological input without destroying the integrity of the session-level data structure. After the above processing, the system generates session-level data archives, raw high-frequency equipment parameter streams, and raw low-frequency physiological parameter streams, and uses these three as inputs for subsequent time-corresponding processing. By completing session establishment, parameter acquisition, field unification, raw archiving, quality marking, and anomaly detection, it can provide a data foundation with clear sources, unified structure, clear boundaries, and traceability for subsequent time-series alignment, state correlation, risk assessment, and equipment control.

[0021] Specifically, such as Figure 2As shown: The timing alignment module is used to perform time-correlation processing on dialysis equipment parameters and patient physiological parameters. After completing session registration, the system continuously receives raw high-frequency equipment parameter streams and raw low-frequency physiological parameter streams. When raw data is continuously acquired for 30 seconds, unified time-correlation processing is initiated. The 30-second initialization window is set because this duration can complete the initial clock synchronization detection and accumulate continuous equipment operation data at the beginning of the session, providing basic data for subsequent time correction and physiological hysteresis correction. If the initialization window is too short, the number of network jitter samples will be insufficient, and the clock offset estimation stability will be poor. If the initialization window is too long, the start time of the subsequent risk identification link will be delayed. The time-correlation processing first performs clock offset compensation and then performs physiological hysteresis correction to eliminate system clock deviations between different hardware terminals and causal misalignments between dialysis equipment actions and human physiological responses, respectively. Dialysis equipment and monitoring equipment are typically located on different hardware terminals, and their internal clock sources drift. Network transmission paths, protocol parsing times, and buffer delays also introduce additional time deviations. Therefore, the core gateway sends synchronization probe messages to both the dialysis and monitoring equipment within a 30-second initialization window, recording the sending time, node response time, and return packet arrival time of each probe message. The number of synchronization probe messages can be selected from 50 to 150, with 100 being the preferred value. The choice of 100 is to ensure that this number can cover short-term network jitter samples while also taking into account the initialization waiting time. Long; when probe messages are sent at intervals of approximately 300ms, 100 messages can be sent and retrieved within a 30s window without significantly interfering with the reception of the original data; the system calculates the round-trip time of the messages based on the recorded results and discards probe results with round-trip times exceeding 50ms; 50ms is set as the rejection threshold because the round-trip delay of a single message in the wired local area network within the hospital is usually less than 10ms under normal communication conditions. When it exceeds 50ms, it is mostly related to link congestion, instantaneous crosstalk, or retransmission. If it continues to be included in the calculation, it will significantly amplify the clock deviation estimation error; After removing abnormal detection results, the system uses Kalman filtering to estimate the one-way transmission delay and node local clock offset for the remaining valid detection results. It then uses the session clock of the core gateway as a unified benchmark to uniformly correct the timestamps of the original data records. The unified time error control target can be set to no more than 5ms. 5ms is chosen as the preferred control target because when dialysis equipment parameters are collected in the previous module at 1-second intervals, a time error controlled within 5ms results in a time deviation ratio of less than 0.5% relative to a single sampling period, which is insufficient to determine the sequence of critical equipment actions and physiological responses. A significant time inversion can meet the time consistency requirements of bedside real-time monitoring scenarios. If 10 consecutive synchronization probe messages are lost during the detection phase, the corresponding link is considered to have severe interference. The criterion of losing 10 consecutive messages is because a sudden loss of 10% out of 100 synchronization probe messages is significantly beyond the range of occasional single-packet loss, indicating that the current link does not have stable time correction capabilities. In this case, the data within the corresponding time period is marked as untrusted and temporarily excluded from the high-level risk identification process. Normal processing will resume after a stable synchronization result is obtained again. After clock offset compensation, the system performs dynamic time-correlation processing on high-frequency device data and low-frequency physiological data. Simply unifying the clock is insufficient to eliminate the causal misalignment between device actions and human responses, as blood pressure changes typically exhibit physiological lag relative to ultrafiltration. Therefore, the system uses a continuous 3600s data point as a sliding processing window. The 3600s window length is chosen because, under the condition of a 900s sampling period for non-invasive blood pressure, a 3600s window can typically cover four real non-invasive blood pressure measurements and reflect a relatively complete continuous dehydration process, facilitating the establishment of a time correspondence between device operating status and physiological changes. If the window is too short, the number of real physiological reference data points will be insufficient, hindering the formation of a stable physiological change trajectory; if the window is too long, too many historical states will be introduced, reducing the targeting of the current treatment stage. Within this sliding window, the system extracts correlation features such as transmembrane pressure change rate, blood pump speed perturbation, and cumulative ultrafiltration progress from high-frequency equipment data to characterize the intensity of dialysis equipment operation and the trend of intravascular volume changes. Simultaneously, using adjacent real non-invasive blood pressure measurements as boundary reference data points, the system extrapolates the physiological trajectory within a 900-second measurement interval. The system can employ a state-space extrapolation method, reconstructing a continuous blood pressure estimation sequence within the measurement window second-by-second based on the continuous change relationship between adjacent boundary reference data points. The inputs for state-space extrapolation include at least the second derivative of transmembrane pressure, the first-order change in heart rate, and the slope of relative blood volume decrease. The second derivative of transmembrane pressure reflects the acceleration characteristics of changes in extracorporeal circulation fluid resistance, the first-order change in heart rate reflects the short-term sympathetic compensation trend, and the slope of relative blood volume decrease corresponds to the trend of circulating blood volume changes during continuous dehydration. By incorporating these inputs into the continuous estimation process, a continuous physiological change curve can be formed within the non-invasive blood pressure measurement interval, ensuring that the physiological data and equipment data have consistent sampling intervals. To further determine the individual lag time parameter between dialysis action and physiological response, the system performs a mutual information search on the ultrafiltration work curve and the relative blood volume decrease curve within the range of 0s to 1800s. The shift corresponding to the peak mutual information is determined as the individual lag time parameter for the current dialysis stage. The search range of 0s to 1800s is set because dialysis hypotension precursors usually do not all appear immediately after ultrafiltration. Instead, the lag time parameter is searched within 30 minutes, which can cover the commonly seen compensatory delay interval in clinical practice without introducing excessively long irrelevant historical segments. When the mutual information reaches its peak, it indicates that the shifted equipment action curve and physiological change curve have the strongest information correlation. Therefore, the corresponding shift can be used to characterize the equipment-physiological response delay relationship in the current dialysis stage. Subsequently, based on this individual lag time parameter, the system performs time shift matching on the high-frequency equipment parameter stream and the estimated physiological parameter stream, so that each second-level time slice simultaneously corresponds to the equipment action state and the human response state. To ensure the reliability of continuous simulation results, the system sets data integrity constraints during the dynamic time correspondence process. Key equipment parameters include at least instantaneous ultrafiltration rate, transmembrane pressure, and blood flow. If two or more key equipment parameters are missing consecutively within a time window, the equipment action chain within that time window is deemed incomplete, and continuous physiological simulation is discontinued. Using the consecutive absence of two or more key equipment parameters as a downgrade condition is because even with the absence of a single key equipment parameter, the remaining parameters can still maintain basic action state judgment; however, the absence of two or more key equipment parameters makes it difficult to fully express the correspondence between ultrafiltration action and pressure state, thus no longer meeting the continuous simulation conditions. On the other hand, if the time interval between real non-invasive blood pressure reference data points exceeds 1200s, then this time window will not be continuously extrapolated. The maximum allowable interval is 1200s because when the preferred sampling period for non-invasive blood pressure is 900s, 1200s is equivalent to reserving 300s of measurement and upload fluctuation margin on the basis of the normal sampling period. After exceeding this time length, the real physiological reference data points are too sparse and no longer suitable for supporting continuous blood pressure trajectory reconstruction. When any degradation condition is triggered, the system only performs clock offset compensation for this time window, does not generate a continuous physiological estimation sequence, and waits for the next valid physiological reference data point to arrive before restoring the complete dynamic time correspondence processing. After the above processing, the system outputs a second-level synchronization status sequence. Each record in this sequence corresponds to a unified time coordinate, a unified sampling interval, and a device-physiological correspondence corrected by individual hysteresis time parameters, and serves as the direct input for subsequent status association processing. By completing clock offset compensation, continuous physiological trajectory reconstruction, individual hysteresis time parameter identification, and abnormal degradation processing, a stable, traceable, and clinically corresponding unified time basis can be established between high-frequency dialysis equipment parameters and low-frequency patient physiological parameters.

[0022] Specifically, such as Figure 3 As shown: The state association module is used to perform correlation analysis on the parameters of dialysis equipment and patient physiological parameters after time-correspondence. After obtaining the second-level synchronous state sequence, the system first organizes the synchronous records of each moment into a state association record table, and then analyzes the correspondence between dialysis equipment actions and patient physiological changes based on different observation durations. The state association record table uses a second-level time index as the recording basis, and the content written includes at least the following fields: blood flow rate, instantaneous ultrafiltration rate, cumulative ultrafiltration volume, dialysate conductivity, dialysate temperature, venous pressure, transmembrane pressure, estimated systolic pressure, estimated diastolic pressure, estimated mean arterial pressure, heart rate, blood oxygen saturation, relative blood volume decrease slope, target dehydration completion percentage, and remaining dry weight deficit. By first forming a state association record table with a unified structure, and then extracting the correlation information within different observation durations, the equipment action information and patient physiological information that have completed time-correspondence processing can be incorporated into the same continuous analysis link. After establishing the state association record table, the system constructs short-term, medium-term, and long-term observation windows, and extracts corresponding state information within each window. The short-term observation window can be set to 300 seconds. This value is chosen because, under second-level synchronized state sequence conditions, 300 seconds can cover 5 minutes of continuous recording, reflecting rapid changes caused by dialysis equipment adjustments, body position changes, or short-term cyclical fluctuations, without misjudging single noise fluctuations as stable trends due to excessively short observation times. The medium-term observation window can be set to 900 seconds because this duration is consistent with the absence of... The optimal sampling period for non-invasive blood pressure measurement is consistent, which facilitates the observation of the correlation between ultrafiltration and blood pressure changes within a complete low-frequency physiological measurement cycle. The long observation window can be selected as 1800s. The value of 1800s is chosen because this duration covers two consecutive non-invasive blood pressure measurement cycles, which can more completely reflect the changing trends of volume redistribution and continuous depletion of compensatory reserve. The system calculates the ultrafiltration rate change amplitude, pulse pressure change, transmembrane pressure peak-to-trough difference, heart rate change trend, and estimated correlation between systolic blood pressure and cumulative ultrafiltration volume within the above observation window, and uses them as the basis for subsequent status judgment. To reduce the impact of patient turning over, coughing, slight limb movements, short-term cuff disturbances, or local signal fluctuations on the correlation analysis results, the system performs modal decomposition processing on the estimated systolic blood pressure and heart rate sequences, decomposing the original sequences into high-frequency perturbation components and low-frequency main trend components. The high-frequency perturbation components are not directly used for state determination, while the low-frequency main trend components serve as the main input for state correlation analysis. The reason for processing the estimated systolic blood pressure and heart rate sequences simultaneously is that hemodynamic changes during dialysis usually manifest as both pressure changes and compensatory heart rate changes. If only a single parameter is used, short-term compensatory adjustment information may be missed. The system also projects the ultrafiltration action curve, cumulative ultrafiltration volume change curve, and transmembrane pressure change curve into the same observation window, forming a combination of device action features for the corresponding window. This combination of device action features is then used together with the low-frequency main trend components for subsequent state determination. After generating the status information within the window, the system generates a status association record for the corresponding window based on the correlation link between equipment dehydration intensity, blood volume changes, and cardiovascular compensation performance. Each status association record, in addition to the original parameters, also includes status markers such as volume-pressure difference markers, pulse pressure narrowing markers, compensation enhancement markers, and compensation decay markers. The volume-pressure difference marker indicates whether there is a significant deviation between changes in dehydration action and pressure feedback. The pulse pressure narrowing marker indicates whether the difference between systolic and diastolic blood pressure continues to decrease. The compensation enhancement marker indicates whether the body continues to maintain circulatory stability by increasing heart rate or maintaining the main pressure trend as cumulative dehydration progresses. The compensation decay marker indicates whether the original compensatory capacity begins to weaken as dehydration continues. These status markers are not directly output as risk conclusions but rather as intermediate results for subsequent risk assessment processing. The generation of status markers adopts a combined judgment method of numerical conditions and change trends. In this module, 5 mmHg is used as the judgment value for short-term pressure changes, and 5 beats / min is used as the judgment value for sustained heart rate changes. 5 mmHg is used as the judgment value because the change amplitude is usually higher than the short-term random fluctuation range of conventional non-invasive blood pressure, and can be distinguished from general measurement noise. 5 beats / min is used as the judgment value because when it is lower than this amplitude, it is more susceptible to single heart rate fluctuations or occasional interference, while when it reaches more than 5 beats / min, it can better reflect sustained compensatory heart rate increases or decreases. The pulse pressure narrowing marker can be determined by combining the pulse pressure changes within the mid-time observation window. The pulse pressure is obtained by subtracting the estimated diastolic pressure from the estimated systolic pressure. When the pulse pressure decreases by more than 5 mmHg relative to the initial value of the window, a pulse pressure narrowing marker can be written. The conditions for generating the compensatory enhancement marker can be selected as follows: within the 900s observation window, the first-order rate of change of the cumulative ultrafiltration volume is continuously greater than zero, the estimated main trend of systolic pressure changes by no more than 5 mmHg relative to the initial value of the window, and the heart rate increases by more than 5 beats / min compared to the initial value of the window. Among the above conditions, the continuous first-order rate of change of the cumulative ultrafiltration volume is used to indicate that the dehydration process is still ongoing; the estimated main trend of systolic pressure changes by no more than 5 mmHg is used to indicate that the short-term pressure is still in a basically stable range; and the heart rate increases by more than 5 beats / min is used to indicate that the body has a continuous compensatory increase in heart rate. The conditions for generating the compensatory attenuation marker can be selected as follows: while the cumulative ultrafiltration volume continues to increase, the pulse pressure decreases by more than 5 mmHg relative to the initial value of the window, and the heart rate change rate is no longer greater than zero in the later part of the window, or the heart rate no longer continues to increase compared to the previous comparison point. This condition can distinguish between a compensatory state that can still maintain pressure stability and a compensatory state that has begun to lose its ability to maintain stability. The volume-pressure difference marker can be generated based on the degree of deviation between the trend of cumulative ultrafiltration volume change and the trend of estimated mean arterial pressure change within the mid-time observation window. For example, if the cumulative ultrafiltration volume continues to increase, and the main trend of estimated mean arterial pressure decreases by more than 5 mmHg relative to the initial value of the window, and this decrease is not accompanied by the writing of the compensatory enhancement marker, the volume-pressure difference marker can be written. The above judgment method can eliminate the interference of short-term fluctuations and retain the continuous change information related to hemodynamic instability. To ensure the comparability of status association results for the same patient at different dehydration stages, the system also adds a treatment stage identifier to each status association record based on the target dehydration completion rate. The treatment stage identifier can be divided into early stage, intermediate stage, and final stage. When the target dehydration completion rate is below 30%, it is marked as the early stage; when the target dehydration completion rate is between 30% and 85%, it is marked as the intermediate stage; and when the target dehydration completion rate is above 85%, it is marked as the final stage. 30% is used as the upper limit for the early stage because when the target dehydration completion rate is low, there is usually still considerable space for interstitial fluid to return to the blood vessels, and the body's compensatory capacity is relatively sufficient. 85% is used as the lower limit for the final stage because as the target dehydration nears its end, the ability for further return weakens, and slight ultrafiltration changes are more likely to trigger physiological instability. To improve adaptability, the cutoff value for the early stage can be selected as 25% to 35%, and the cutoff value for the final stage can be selected as 80% to 90%. This application uses 30% and 85%, respectively. The system also checks the reliability of physiological estimation results. If the proportion of placeholder reference data in the basic reference data used to form the physiological estimation sequence within a certain observation window exceeds 50%, the physiological estimation results within that window are considered to be unreliable. The 50% threshold is used because when the placeholder reference data exceeds half of the basic reference data within the window, the actual measurement results are insufficient to support stable correlation analysis, and continuing to output complete state correlation records would amplify the bias caused by the substitute data. In this case, the system only retains the original equipment parameters, basic physiological parameters, and finite state markers within that window, and does not output complete state correlation records. Complete state correlation processing is restored after sufficient real physiological reference data is subsequently obtained. After the above processing, the system generates a state association sequence with a unified time index, unified stage identifier, and consistent state label expression method for subsequent risk assessment processing. By completing multi-window association analysis, noise removal, state label generation, stage identifier appending, and reliability check, the time-corresponding dialysis equipment parameters and patient physiological parameters can be further organized into association results that can be directly called for risk assessment.

[0023] Specifically, such as Figure 3 and Figure 4 As shown: The risk assessment module is used to determine the risk status and trend based on the state correlation results; the system makes continuous judgments on the state correlation sequence, rather than making isolated comparisons of parameter values ​​at a single moment; to balance the timeliness and stability of risk identification, the system adopts an 1800s sliding judgment window consistent with the aforementioned long-term observation window, and updates the judgment results in a rolling cycle of 300s; the 1800s sliding judgment window is chosen because this duration covers two consecutive non-invasive blood pressure measurement cycles, which can reflect a relatively complete volume compensation change range; the 300s risk assessment cycle is chosen because this duration can update the risk status in a timely manner without misjudging short-term disturbances as a continuous deterioration trend due to excessively rapid updates; Within each 1800s sliding decision window, the system extracts three core quantitative indicators: volume compensation consumption divergence, vascular compliance decay, and compensation exhaustion slope. The volume compensation consumption divergence is used to characterize the deviation between the cumulative ultrafiltration rate change curve and the estimated low-frequency main trend curve of systolic blood pressure. The vascular compliance decay is used to characterize the coupling decay between the transmembrane pressure peak-to-trough difference and pulse pressure changes. The compensation exhaustion slope is used to characterize the turning speed when heart rate changes from enhancement to unsustainability. The system first normalizes the above three indicators, then inputs them into the risk assessment model, and outputs a risk confidence index between 0 and 1. The closer the risk confidence index is to 1, the higher the probability that the current patient will experience hemodynamic instability. The risk assessment model can be selected from a normalized weighted fusion model, a support vector regression model, or other assessment models capable of outputting continuous risk scores; this application preferably adopts a normalized weighted fusion model; the risk confidence index is formed by the following formula: Risk confidence index = w1 × volume compensation consumption divergence index + w2 × vascular compliance decline index + w3 × compensation exhaustion slope + w4 × remaining dry weight gap correction amount; Wherein, w1, w2, w3, and w4 represent the weighting coefficients corresponding to the volume compensation consumption divergence index, the vascular compliance decline index, the compensation depletion slope, and the remaining dry weight deficit correction amount, respectively. Each weighting coefficient is used to characterize the contribution of the corresponding index to the risk confidence index, and all input quantities are normalized to the range of 0 to 1. The weight of the volume compensation consumption divergence index can be selected from 0.35 to 0.50, with a preferred value of 0.40; the weight of the vascular compliance decline index can be selected from 0.20 to 0.35, with a preferred value of 0.30. The weight of the compensatory exhaustion slope can be selected from 0.15 to 0.30, with a preferred value of 0.20; the weight of the remaining dry weight deficit correction can be selected from 0.05 to 0.15, with a preferred value of 0.10. This setting is because the deviation between the equipment's dehydration intensity and the main trend of physiological pressure, as well as the decrease in vascular compliance, have a stronger direct characterization effect on hemodynamic instability; heart rate changes are used to characterize the rate of compensatory changes; the remaining dry weight deficit is used to reflect the dehydration pressure that still needs to be completed in the current treatment stage, so a relatively low correction weight is taken. The risk confidence index does not use a fixed threshold trigger but dynamically tightens with the progress of dialysis. To maintain consistency with the treatment stage identifiers mentioned earlier, the system divides the system into early, intermediate, and final stages based on the target dehydration completion rate, and sets risk thresholds for each stage. For the early stage, where the target dehydration completion rate is below 30%, the risk threshold can be selected from 0.75 to 0.90, with 0.85 being the preferred value. For the intermediate stage, where the target dehydration completion rate is 30% to 85%, the risk threshold tightens linearly with the dehydration completion rate. For the final stage, where the target dehydration completion rate is above 85%, the risk threshold can be selected from 0.40 to 0.55, with 0.45 being the preferred value. The threshold for the early stage is set at 0.85 because the interstitial fluid refill space is relatively large in the early stages of treatment, and the body has sufficient compensatory capacity, allowing for a more relaxed threshold. The threshold for the final stage is set at 0.45 because the ability for further refill weakens in the final stages of treatment, and even slight changes in dehydration can trigger physiological instability; therefore, the threshold needs to be lowered. The risk threshold for the intermediate stage is determined as follows: Mid-term risk threshold = 0.85 - Normalized progress × 0.40 Wherein, normalized progress = (target dehydration completion rate - 30%) ÷ (85% - 30%); Through the above processing, the risk threshold can be smoothly transitioned from 0.85 to 0.45, avoiding abrupt changes in the judgment boundary at the stage switching point; The system compares the current risk confidence index with the current stage risk threshold and determines the risk status based on the risk change trend. The risk change trend is represented by a smoothed change rate, which can be calculated by smoothing the difference in risk confidence indices from 20 consecutive second-level sampling points. Setting 20 consecutive seconds as the continuous judgment time is because, under second-level update conditions, if all 20 consecutive sampling points exceed the stage threshold, it can essentially exclude instantaneous fluctuations caused by short-term postural changes, cuff disturbances, or single abnormal heartbeats, while not significantly delaying the identification of true risk events. The smoothed change rate threshold can be selected from 0.005 / s to 0.015 / s, with a preferred value of 0.01 / s. The optimal value is s because this slope corresponds to a risk increment of approximately 0.20 within 20 seconds, which can reflect a continuous deterioration trend without making the actual risk too difficult to trigger in a timely manner due to an excessively high threshold. If the risk confidence index is not lower than the current stage risk threshold for 20 consecutive seconds and the smoothing rate of change is not less than 0.01 / s, then a continuous deterioration trend is determined, and the risk status is set to red. If the risk confidence index is higher than the current stage risk threshold but the duration is less than 20 seconds, or the duration reaches 20 seconds but the smoothing rate of change is less than 0.01 / s, then the risk status is set to yellow (observation). If the risk confidence index is lower than the current stage risk threshold, then the green (stable) status is maintained. For sessions entering the red risk range, the system further generates an intervention urgency index based on the risk confidence index, risk trend, current estimated systolic blood pressure, and remaining dry weight deficit. The intervention urgency index can be a normalized value between 0 and 1, obtained by weighting the normalized risk confidence index, risk trend, deviation of the current estimated systolic blood pressure from the 90 mmHg reference value, and remaining dry weight deficit. To facilitate subsequent equipment control, the system classifies the intervention urgency index: when the intervention urgency index is between 0.60 and 0.85 and the current estimated systolic blood pressure is not lower than 85 mmHg, it is classified as moderate risk; when the intervention urgency index is higher than 0.85, or the current estimated systolic blood pressure is lower than 85 mmHg, it is classified as severe risk. 85 mmHg is used as the pressure threshold for severe risk because this pressure level is close to the significantly low perfusion risk range; 0.60 is used as the lower limit for moderate risk to distinguish between a mild observation state and a risk state requiring intervention preparation. When the patient is in the yellow observation state for three consecutive risk assessment cycles, the system will not trigger device control, but will temporarily adjust the non-invasive blood pressure measurement frequency from once every 15 minutes to once every 5 minutes within the limited observation period to increase the update density of low-frequency physiological data. The three consecutive risk assessment cycles refer to three consecutive 300-second risk assessment cycles. The 5-minute interval is chosen as the temporary increased measurement frequency because it provides a higher physiological update density than the conventional 15-minute interval, while being significantly lower than continuous high-frequency measurement, thus avoiding long-term impact of repeated cuff inflation on local circulation in the limbs. The limited observation period can be selected from 15 minutes to 30 minutes, with 15 minutes being the preferred value. If the risk status recovers to the green stable state within the observation period, the non-invasive blood pressure measurement frequency will return to the conventional frequency. If the risk status further rises to the red risk state, subsequent device adjustment processing will proceed directly. If a complete set of core quantitative indicators cannot be formed within a certain risk assessment period due to unreliable data, the system will not directly determine the safety status. Instead, it will maintain the risk status of the previous valid risk assessment period and require continued sampling to complete the data. To prevent the long-term use of the old risk status when data is continuously missing, the system can select an upper limit for maintaining the previous valid risk status to two consecutive risk assessment periods, corresponding to 600 seconds. If a valid output cannot be formed after more than two consecutive risk assessment periods, a "pending review" mark will be added while maintaining the previous valid risk status, and the data for that period will be sent for subsequent manual review or enhanced data collection and processing. This mechanism can avoid erroneously resolving the risk status due to data loss in a single assessment period, and also avoid continuing to use outdated risk judgment results when data is insufficient for a long period of time. After the above processing, the system outputs the risk status, risk change trend, and intervention urgency index. By continuously judging the status correlation sequence, dynamically tightening the risk threshold with the progress of dialysis, implementing temporary monitoring encryption for the yellow observation status, and setting a limited maintenance and pending review mechanism for the data missing status, the risk judgment results can simultaneously take into account sensitivity, stability, and clinical feasibility.

[0024] Specifically, such as Figure 5 As shown: The equipment control module generates control instructions for dialysis equipment based on risk status, risk change trends, and intervention urgency index. After receiving the risk status output by the risk assessment module, the system no longer repeats the risk classification judgment, but directly calls the matching control strategy library according to the corresponding risk level to generate parameter adjustment instructions for the dialysis equipment. The control objects include at least parameters such as ultrafiltration rate, dialysate sodium concentration, and dialysate temperature. By directly converting the risk assessment result into equipment parameter adjustment content, the time interval between risk identification and equipment response can be shortened, enabling the assessment result to be transmitted to the underlying equipment control link in a timely manner. When the risk assessment module outputs a moderate risk status, the system generates a joint control command to adjust the ultrafiltration rate and dialysate sodium concentration. The system first reads the current ultrafiltration rate of the dialysis equipment and determines the ultrafiltration rate reduction ratio based on the current intervention urgency index. The ultrafiltration rate reduction ratio can be selected from 20% to 50% of the current value. To ensure the control range is consistent with the risk level, it can be determined in segments according to the intervention urgency index: when the intervention urgency index is 0.60 to 0.70, the ultrafiltration rate reduction ratio can be selected from 20% to 30%; when the intervention urgency index is 0.70 to 0.80, the ultrafiltration rate reduction ratio can be selected from 30% to 40%; when the intervention urgency index is 0.80 to 0.85, the ultrafiltration rate reduction ratio can be selected from 40% to 50%. In typical moderate-risk scenarios, the preferred value is a 40% reduction, corresponding to a target ultrafiltration rate of 60% of the current value. A 40% reduction is chosen because this level can significantly slow the rate of fluid loss in a short time without completely interrupting the dialysis process, thus preserving some dehydration capability. The transition time for the target ultrafiltration rate can be selected from 180s to 300s, with 180s being the preferred value. A 180s transition time is chosen because it avoids cyclical fluctuations caused by sudden changes in ultrafiltration parameters, while also preventing significant delays in risk control. Within this transition time, the system gradually reduces the ultrafiltration rate in a step-by-step manner, rather than jumping directly to the target value, to reduce the impact of parameter mutations on the cyclic state and equipment stability. Under moderate risk conditions, the system also synchronously adjusts the dialysate sodium concentration to enhance plasma osmotic pressure and promote the return of interstitial fluid to the blood vessels. The target dialysate sodium concentration can be increased from the baseline value by 3 mmol / L to 7 mmol / L. To ensure that the adjustment range is consistent with the risk level, it can be determined in segments according to the intervention urgency index: when the intervention urgency index is 0.60 to 0.70, the target dialysate sodium concentration can be increased from the baseline value by 3 mmol / L to 4 mmol / L; when the intervention urgency index is 0.70 to 0.80, the target dialysate sodium concentration can be increased from the baseline value by 4 mmol / L to 5 mmol / L; when the intervention urgency index is 0.80 to 0.85, the target dialysate sodium concentration can be increased from the baseline value by 5 mmol / L to 7 mmol / L; when the baseline value is 138 mmol / L, it is preferable to increase it to 143 mmol / L. The preferred target value is 1 mmol / L because this value can create a moderate increase in osmotic pressure during the moderate-risk phase, while still remaining below the safe upper limit of 145 mmol / L. The sodium concentration increase rate can be selected from 0.5 mmol / L / min to 1.0 mmol / L / min, with 1.0 mmol / L / min being the preferred value. The increase rate of 1.0 mmol / L / min is chosen because this rate can achieve effective regulation in a short time without increasing the risk of sodium load fluctuations due to excessively rapid increases. The system can select a safe upper limit of 145 mmol / L for dialysate sodium concentration. When the calculated target value is higher than 145 mmol / L, it will be automatically truncated to 145 mmol / L to avoid increasing the risk of osmotic complications due to excessively high sodium concentration. If the dialysis equipment does not support remote adjustment of dialysate sodium concentration, the system will retain the ultrafiltration rate reduction command and write an auxiliary adjustment flag that requires manual confirmation. When the risk assessment module outputs a severe risk status, the system generates an emergency protection control command. This command prioritizes preventing further dehydration from worsening the circulatory condition, thus rapidly reducing the ultrafiltration rate to zero without stopping the extracorporeal blood volume pump. The transition time for the ultrafiltration rate to zero can be selected from 2 seconds to 10 seconds, with 2 seconds being the preferred value. A transition time of 2 seconds is chosen because, in a severe risk state, life support safety takes precedence over the completion of dialysis goals, requiring the rapid elimination of factors contributing to circulatory deterioration caused by continued dehydration. Simultaneously, the system slightly lowers the dialysate temperature; the target dialysate temperature can be selected from 35.0℃ to 35.5℃. ℃; When the baseline temperature is 36.5℃, it is preferable to lower it to 35.0℃; Lowering it from 36.5℃ to 35.0℃ is because a temperature drop of about 1.0℃ to 1.5℃ can create a mild cooling stimulus, causing peripheral vasoconstriction and increasing central circulation perfusion pressure, without significantly increasing the patient's shivering or additional metabolic burden; The target temperature should not be set lower than the lower limit allowed by the equipment, and should be constrained by the minimum temperature adjustment step that the bedside equipment can perform; If the dialysis equipment does not support remote adjustment of dialysate temperature, the system will only perform control to rapidly reduce the ultrafiltration rate to zero, and simultaneously write a temperature auxiliary processing flag that requires manual confirmation; After generating various control parameters, the system organizes the control content into control data packets. These control data packets include at least the following fields: dialysis session identifier, target parameter name, target parameter value, transition time, priority, generation time, validity period, and checksum. Moderate-risk control instructions have lower priority than high-risk control instructions. The validity period is determined by both the transition time and the observation time, and can be extended by a risk assessment cycle after the transition time ends to prevent expired control instructions from being repeatedly executed. The edge gateway frames the instructions according to the communication protocol of the target device and converts the control data packets into control messages that the device can recognize before sending them out. By first generating control data packets with a unified structure and then converting them into control messages according to the device protocol, the control logic can be decoupled from the specific device communication method, facilitating compatibility with different dialysis devices. After the command is issued, the system confirms the device's response. The device confirmation waiting time can be selected from 300ms to 500ms, with 500ms being the preferred value. A confirmation waiting time of 500ms is chosen because the bedside control link should typically complete control confirmation within sub-seconds; this duration covers a normal round-trip confirmation process without causing excessive delays in anomaly detection. If no device confirmation is received within the confirmation waiting time, the system retransmits the control message according to the retransmission mechanism. The retransmission interval can be selected from 100ms to 300ms, with 200ms being the preferred value. The number of retransmissions can be selected from 2 to 4, with 2 being the preferred value. The value is set to 3 times; the retransmission interval is set to 200ms and the number of retransmissions is set to 3 times because this combination can complete automatic recovery in a short time, while avoiding continuous high-frequency retransmissions that cause continuous congestion to the communication link; the above parameters are applicable to the local control link between the bedside gateway and the dialysis equipment; if a cross-network segment or control link forwarded through the hospital middleware is used, the confirmation waiting time and retransmission interval can be relaxed according to the average round-trip delay of the link; if no valid acknowledgment is received after 3 consecutive retransmissions, the system will switch to manual takeover mode, stop sending control commands, and trigger the bedside audible and visual alarm and the nurse station high-priority prompt. After receiving the equipment confirmation acknowledgment, the system also checks the deviation between the equipment's execution status and the target parameters. For ultrafiltration rate control, if the actual ultrafiltration rate is still higher than 50 mL / h or higher than 5% of the target value after the transition period ends when the target ultrafiltration rate requirement is reduced to zero, the ultrafiltration execution deviation is determined to be excessive. For dialysate sodium concentration control, if the deviation between the actual dialysate sodium concentration and the target value continuously exceeds 1 mmol / L, the execution deviation is determined to be excessive. For dialysate temperature control, if the deviation between the actual dialysate temperature and the target value continuously exceeds 0.3℃, the execution deviation is determined to be excessive. L / h, 1 mmol / L, and 0.3℃ are used as the execution deviation limits for ultrafiltration rate, sodium concentration, and temperature, respectively. This is because these deviation ranges are higher than the control resolution and short-term execution fluctuation range of conventional dialysis equipment, and can better distinguish between transient adjustment errors and substantial execution anomalies. When the equipment's execution feedback for multiple target parameters is inconsistent, the system takes the high-priority control parameter that has been successfully executed as the current effective control state and writes the parameters that were not successfully executed into the anomaly review record. When any execution deviation exceeds the limit, the system generates an execution anomaly flag and transfers the current session to the anomaly handling branch in the subsequent review module. After the above processing, the system outputs the issued control plan, equipment confirmation receipt, and real-time execution status, and uses them as input for subsequent execution review. By generating different equipment control content according to moderate and severe risks, setting transition ranges and safety limits for control parameters, setting a limited retry mechanism for communication receipts, and setting clear abnormal judgment boundaries for execution deviations, the equipment control results can simultaneously take into account the timeliness of response, the stability of parameter adjustment, equipment compatibility, and clinical safety.

[0025] Specifically, such as Figure 5 As shown: The verification module is used to obtain and verify the equipment feedback and monitoring feedback after regulation; after the dialysis equipment receives and begins to execute the regulation command, the system enters the verification state; in the verification state, an independent post-regulation verification data sequence is formed. Historical data before regulation is not directly merged with new data after regulation to avoid the difference in state before and after regulation interfering with the verification conclusion; the system continues to sample equipment parameters at 1 second and temporarily increases the non-invasive blood pressure measurement frequency from once every 15 minutes to once every 3 minutes, while continuously receiving heart rate and blood oxygen saturation data; the non-invasive blood pressure measurement frequency is temporarily increased to once every 3 minutes because this frequency can obtain multiple real physiological measurement results within a short observation period to determine whether the regulation action has changed the circulatory state; at the same time, this frequency is only maintained during the short verification period, and the normal frequency is restored after the observation period ends, thereby avoiding repeated cuff inflation causing continuous interference to local circulation; The system constructs a short-term review window and a total review window. The short-term review window can be selected from 180s to 300s, with 180s being the preferred value. The total review window can be selected from 8min to 15min, with 10min being the preferred value. The short-term review window is set to 180s because this duration is consistent with the temporary encrypted measurement cycle of non-invasive blood pressure, which facilitates timely updates to the review judgment after each new measurement result. The total review window is set to 10min because, under the condition that non-invasive blood pressure is measured every 3 minutes, more than 3 real blood pressure measurement results after adjustment can usually be obtained within 10 minutes. Combined with continuous equipment parameters, heart rate, and blood oxygen saturation data, a complete judgment is formed on the early effect of a single adjustment action. Within the total review window, the system re-executes time correspondence processing, state association processing, and risk assessment processing, but the review judgment only uses the newly generated equipment feedback and monitoring feedback data after the adjustment takes effect to form the review sequence. The system determines the effectiveness of current control measures based on the risk and physiological status at the end of the overall review window. If, at the end of the overall review window, the risk confidence index drops below 0.60, the current estimated systolic blood pressure rises above 95 mmHg, and the smoothed rate of change of the risk confidence index within the most recent short-term review window is not greater than 0, then the current control measures are deemed effective. 0.60 is used as the risk confidence threshold for passing the review because this value is below the aforementioned high-risk trigger range and can serve as a safe threshold for a return to normalcy after leaving the high-risk state. 95 mmHg is used as the systolic blood pressure recovery threshold because this pressure level is higher than the aforementioned risk reference values ​​of 85 mmHg and 90 mmHg, providing a certain safety margin for cyclical recovery. The additional condition of a smoothed rate of change not greater than 0 is to confirm that the risk trend has shifted from continuous deterioration to stabilization or decline, rather than a short-term decline followed by a resurgence. When all three conditions are met simultaneously, the system generates a recovery control plan. The recovery control plan does not directly restore the equipment parameters to the original prescription values, but instead uses a gradual increase method to restore the ultrafiltration rate. The recovery range can be selected from 5% to 15% of the original prescription ultrafiltration rate per recovery cycle, with 10% being preferred. The recovery cycle can be selected from 3 minutes to 10 minutes, with 5 minutes being preferred. The recovery range of 10% and the recovery cycle of 5 minutes are chosen because this combination can restore dialysis efficiency while avoiding sudden parameter increases that could cause cycle fluctuations again. Each recovery cycle begins timing after the previous recovery instruction takes effect and a short-term review window is completed. The system only allows the next recovery cycle if the risk status remains green and stable within the most recent short-term review window, and the current estimated systolic blood pressure does not drop below 90 mmHg again. If the risk confidence index rises above 0.60 again during the recovery process, or the current estimated systolic blood pressure drops again, the recovery will be stopped immediately, and the current ultrafiltration rate will be maintained. By setting the recovery range, recovery cycle, and recovery prerequisites, it is possible to avoid a rapid return to the original prescription after the control takes effect, which could lead to secondary instability. If the risk confidence index does not drop below 0.60 by the end of the final review window, or the current estimated systolic blood pressure remains below 90 mmHg, or a new equipment malfunction occurs during the review period, the current control is deemed insufficient, and the process enters the secondary control branch. The secondary control branch performs enhanced treatment based on the current risk level. When the session is in a moderate risk state and the initial control is insufficient, the system continues to maintain the already reduced ultrafiltration rate and further reduces the ultrafiltration rate based on the original control. The further reduction can be 10% to 20% of the original adjusted ultrafiltration rate. With equipment support, the target sodium concentration of the dialysate can also be increased by 1 mmol / L. The concentration is set to 2 mmol / L, but not exceeding the safety limit of 145 mmol / L. This setting is because if the initial regulation is insufficient but the risk level has not yet reached a high level, the circulation status can still be improved by moderately increasing ultrafiltration reduction and osmotic pressure, without immediately entering a complete ultrafiltration shutdown process. If the session is already in a high-risk state, or if the moderate risk level has not improved after secondary regulation, the system will forcibly maintain the ultrafiltration pause state and retain the implemented high-risk protection control, while triggering manual emergency takeover. By distinguishing between the two situations of insufficient initial regulation and ineffective secondary regulation, the system can differentiate between the state where automatic adjustment can continue and the state where manual takeover must be initiated. The system establishes an execution review log and writes the review results into a review record table. The review record table includes at least the following fields: dialysis session identifier, changes in risk status, control instruction content, equipment confirmation receipt, physiological response after regulation, abnormal events, manual takeover identifier, and execution result of the recovery control plan. An intermediate review record is written at the end of each short review window, and a review conclusion record is written at the end of the overall review window. If the equipment feedback and monitoring feedback are inconsistent, for example, if the equipment receipt shows that the control has been executed but the physiological status has not improved as expected, the system writes a control effectiveness pending review mark in the review record table and lowers the priority of the corresponding regulation strategy when updating individual parameters in subsequent updates. Through the above records, the entire process of a regulation action from issuance to execution, and from physiological feedback to review conclusion, can be traced back. The system also updates the patient's individual tolerance parameters based on the review record sheet. At the start of the next dialysis session, the system can call up the valid review records from the previous session to adjust the patient's risk threshold tightening rate, ultrafiltration rate reduction ratio, and dialysate sodium concentration increase rate at different dehydration stages. This ensures that the parameters are no longer fixed values ​​but form an individualized update mechanism. If the review record shows that yellow observation or red risk occurs in the same dehydration progress interval for two consecutive treatments, the system will set that dehydration progress interval as a key monitoring interval before the next dialysis session. The key monitoring interval can be defined as a range of 5% to 10% before and after the target dehydration completion rate. The risk threshold tightened in advance can be selected to be reduced by 0.03 to 0.10 from the original threshold, and the reduced threshold is not lower than the aforementioned lower limit of the risk threshold for the end stage. This setting is made because repeated risk states in similar dehydration progress intervals for the same patient usually indicate that the interval is the patient's weakest tolerance. Tightening the threshold in advance helps to identify the same risk in the next session earlier. After the above processing, the system outputs the adjusted equipment feedback, monitoring feedback, review conclusion, recovery control plan and abnormal takeover indicator, and writes the relevant results into the review record table. By performing short-term enhanced acquisition after adjustment, re-performing time correspondence and risk assessment, performing recovery or secondary adjustment according to the review results, and writing the review results back into the patient's individual tolerance parameter record, the dialysis control process can form a closed-loop processing link from risk assessment, equipment adjustment to effect verification and parameter inheritance.

[0026] Example 2: Based on Example 1, the specific application process of a hemodialysis information management and monitoring system is further explained: Specifically, such as Figure 6The following example illustrates the process using a patient receiving maintenance hemodialysis. Prior to treatment, the patient's basic prescription information was entered by medical staff: prescription duration 4 hours, dry weight 62.0 kg, pre-treatment weight 64.6 kg, target fluid removal volume 2.6 L, baseline dialysate sodium concentration 138 mmol / L, and baseline dialysate temperature 36.5℃. After treatment began, the edge computing core gateway continuously monitored the dialysis equipment's operating status and the uploaded status. Once the blood pump was activated, the blood flow rate reached 100 mL / min and remained at that level for 5 seconds, and data was received within 120 seconds before and after this point. Upon receiving the first non-invasive blood pressure measurement result, the system determines that the dialysis session has entered the formal record-keeping stage and generates a corresponding dialysis session identifier. Subsequently, the system writes the patient's identity identifier, bed identifier, dialysis equipment identifier, monitoring equipment identifier, session start time, and basic prescription information into the session master table, and simultaneously initiates equipment parameter acquisition and physiological parameter acquisition. On the equipment side, high-frequency parameters such as blood flow, instantaneous ultrafiltration rate, cumulative ultrafiltration volume, venous pressure, and transmembrane pressure are continuously input at 1-second intervals; on the physiological side, low-frequency parameters such as non-invasive blood pressure, heart rate, and blood oxygen saturation are input at predetermined frequencies, forming a raw data record link that corresponds one-to-one with this dialysis session. After the session is established and raw data is continuously collected for 30 seconds, the system initiates unified time correspondence processing. Within the initialization window, the core gateway sends synchronization detection messages to the dialysis equipment and monitoring equipment respectively, estimates the offset between the local clock of the equipment and the session clock, and corrects the raw timestamp using the session clock as a unified reference. After clock offset compensation is completed, the system uses 3600 seconds of continuous data as the sliding processing range to perform dynamic time correspondence between high-frequency equipment actions and low-frequency physiological responses. During the processing, the system extracts correlation quantities such as transmembrane pressure change rate, cumulative ultrafiltration progress, and relative blood volume decrease slope, and reconstructs the continuous blood pressure estimation sequence at intermediate moments by combining adjacent real non-invasive blood pressure measurement results. Subsequently, by identifying the hysteresis time parameter, the equipment action curve and the physiological response curve are translated and matched under a unified time coordinate to obtain a second-level synchronization state sequence. Each record in this sequence corresponds to the equipment action state and human response state at the same moment, providing direct input for subsequent state correlation and risk assessment. After the synchronous state sequence is generated, the system organizes the records at each time point into a state association record table and constructs short-term observation windows, medium-term observation windows, and long-term observation windows respectively. For example, at 115 minutes after the start of this treatment, the target dehydration completion rate is approximately 68%. The system extracts the following association information within the corresponding observation window: the cumulative ultrafiltration volume continues to increase, the peak-to-trough difference of transmembrane pressure is larger than the previous window, the estimated systolic blood pressure main trend begins to slowly decrease, the pulse pressure is smaller than the initial value of the window, and the heart rate initially rises and then tends to level off. The system performs high-frequency perturbation stripping on the estimated systolic blood pressure sequence and heart rate sequence, retains the low-frequency main trend component, and further generates the state label of the corresponding window. Since the cumulative ultrafiltration volume continues to increase, the pulse pressure continuously decreases, and the heart rate rising trend weakens during this period, the system writes the compensatory attenuation label as valid. Also, since the mean arterial pressure main trend begins to deviate from the original stable range, the system simultaneously writes the volume-pressure difference label. Subsequently, the system combines the current target dehydration completion rate, adds a medium-term stage label to the window, and writes the above state results into the state association sequence. After the state association sequence is formed, the system enters continuous risk assessment. Taking the aforementioned intermediate stage as an example, the system extracts three core quantitative indicators within an 1800s sliding assessment window: volume compensation consumption divergence index, vascular compliance decay index, and compensation depletion slope. These indicators are then normalized and weighted to obtain the current risk confidence index. Since the target dehydration completion rate is between 30% and 85%, the system calculates the current stage risk threshold according to the dynamic tightening rules of the intermediate stage. If the system determines that the current risk confidence index is not lower than the stage risk threshold for 20 consecutive seconds and the smoothing rate of change is not less than 0.01 / s, the risk status is set to red. If the risk confidence index exceeds the stage risk threshold, the system will not be considered red. If the value is high but the duration is less than 20 seconds, or the duration reaches 20 seconds but the smoothing rate of change is low, the risk status is set to yellow observation. Taking the above session as an example, the system continuously detected that the risk confidence index was rising between 115 minutes and 118 minutes, and exceeded the current stage risk threshold for 20 consecutive seconds. Therefore, the current status of the session was determined to be red risk. At the same time, the system formed an intervention urgency index based on the risk confidence index, risk change trend, current estimated systolic blood pressure, and remaining dry weight deficit. Since the estimated systolic blood pressure was still higher than 85 mmHg at that time, the intervention urgency index fell into the intermediate risk range. Therefore, the system further entered the intermediate risk control strategy. Once the system determines that the current state is at a moderate risk, the equipment control module does not implement a blanket shutdown. Instead, it calls upon the moderate risk control strategy library to generate joint control measures for adjusting the ultrafiltration rate and the dialysate sodium concentration. Specifically, the system first reads the current ultrafiltration rate of the dialysis equipment, and based on the current intervention urgency index, lowers the ultrafiltration rate by 40%, gradually transitioning to the target value in a stepwise manner within 180 seconds. Subsequently, the system gradually increases the dialysate sodium concentration from the baseline value of 138 mmol / L to 143 mmol / L, at an increase rate of 1.0 mmol / L / m³. The dialysis device is allowed to operate within a specified timeframe, but the upper limit is 145 mmol / L. The control data is organized into a unified control data packet, containing the dialysis session identifier, target parameter name, target parameter value, transition time, priority, generation time, validity period, and checksum. This packet is then converted into a control message by the edge gateway according to the device protocol and sent out. Upon receiving the message, the dialysis device returns an acknowledgment, and the system continues to monitor its execution status. If no valid acknowledgment is received within the specified waiting time, the packet is resent at predetermined retransmission intervals and retries. If no valid acknowledgment is still received, manual intervention is triggered. After the control command takes effect, the system does not end with the completion of the control command, but immediately enters the execution review state. During the review, equipment parameters continue to be sampled at 1-second intervals, and the frequency of non-invasive blood pressure measurement is temporarily increased from once every 15 minutes to once every 3 minutes. Heart rate and blood oxygen saturation are continuously input. The system uses a short review window of 180 seconds and a total review window of 10 minutes to re-execute time correspondence, status correlation, and risk assessment on the newly collected data after the control. If, at the end of the 10-minute observation period, the risk confidence index drops below 0.60, the current estimated systolic blood pressure rises above 95 mmHg, and the smoothed change rate of the risk confidence index within the most recent short review window is not greater than 0, the system will continue to perform the following checks. If the current control is deemed effective, then, taking the above session as an example, approximately 9 minutes after the implementation of moderate-risk control, the system detected that the estimated systolic blood pressure gradually increased from 92 mmHg before control to 97 mmHg, the risk confidence index dropped from the high-risk zone to 0.54, and the rate of change changed from positive to negative. Therefore, the system determined that the control was effective and generated a recovery control plan. The recovery control plan does not directly restore the ultrafiltration rate to the original prescription value, but gradually increases it by 10% every 5 minutes. Only when the risk status remains green and stable within the most recent short-term review window and the estimated systolic blood pressure does not drop below 90 mmHg again is the next round of recovery allowed. In the subsequent stages of the same dialysis session, as the target dehydration completion rate approaches the end stage, the system again continuously assesses the synchronization status sequence. If the risk confidence index rises again during the later stages of treatment, and the current estimated systolic blood pressure has dropped below 85 mmHg, the risk assessment module directly classifies the session as high-risk. At this point, the equipment control module invokes the emergency protection control strategy, reducing the ultrafiltration rate to zero within 2 seconds without stopping the extracorporeal blood volume. Simultaneously, the dialysate temperature is lowered from 36.5℃ to 35.0℃ to create a mild temperature drop. Warm stimulation increases central circulation perfusion pressure; if, after the implementation of this severe risk control, the risk confidence index still does not drop below 0.60 by the end of the final review window, or the current estimated systolic blood pressure is still below 90 mmHg, or an equipment malfunction occurs during the review period, the system determines that the current control is insufficient and enters the secondary control branch; for situations where there is already severe risk and no improvement after secondary control, the system maintains the ultrafiltration paused state, maintains the implemented severe risk protection control, and simultaneously triggers manual emergency takeover, thereby smoothly connecting automatic control with manual bedside treatment; Throughout the session, the system continuously builds an execution review log and writes information such as changes in risk status, control command content, equipment confirmation receipts, physiological responses after regulation, abnormal events, manual takeover indicators, and the execution results of the recovery control plan into the review record table. An intermediate review record is written at the end of each short review window, and a review conclusion record is written at the end of the overall review window. If the equipment receipt shows that the control has been executed, but the physiological status has not improved accordingly, the system writes a control effectiveness pending review mark in the review record table and lowers the priority of this control strategy in subsequent similar situations. Before the next dialysis treatment begins, the system retrieves the valid review records from the previous session and updates the patient's individual tolerance parameters. For example, if the review record shows that two consecutive treatments show yellow observation or red risk in the range of 70% to 80% of the target dehydration completion rate, the system sets this range as the key monitoring range for the next dialysis treatment and tightens the risk threshold in advance within this range. At the same time, it appropriately lowers the trigger point of ultrafiltration rate and the increase of dialysate sodium concentration in advance, so that subsequent treatments no longer use completely fixed control parameters, but form a dynamic adjustment mode that updates according to the patient's individual tolerance characteristics. In this case, the system completes session registration, time correspondence, status organization, risk identification, equipment adjustment, execution review, and parameter write-back and update within a single dialysis session. Each step uses the output of the previous step as the input of the next step, which can not only detect and deal with hemodynamic abnormalities in a timely manner during treatment, but also feed back the effect of the current adjustment to the parameter setting of the next treatment, thus forming a closed-loop operation mechanism for continuous optimization for the same patient.

[0027] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0028] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented in software, the above embodiments can be implemented in whole or in part by a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions of the embodiments of this application are implemented in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted wirelessly or wiredly from one website, computer, server, or data center to another website, computer, server, or data center. Wired methods include optical fiber, twisted pair, coaxial cable, etc. Wireless methods include infrared, microwave, etc. Available media include any available media that can be accessed by a computer or data storage devices such as servers and data centers that contain one or more sets of available media. Available media can be magnetic media (floppy disks, hard disks, magnetic tapes), optical media (DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0029] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0030] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A hemodialysis information management and monitoring system, characterized in that, include: Data access and archiving module: used to collect dialysis equipment parameters and patient physiological parameters, and to create data archives; Time alignment module: used to perform time-correlation processing on dialysis equipment parameters and patient physiological parameters; Status correlation module: used to perform correlation analysis between dialysis equipment parameters and patient physiological parameters after time correspondence; Risk assessment module: used to determine the risk status and trend of change based on the results of correlation analysis; Equipment control module: Used to generate control instructions for dialysis equipment based on risk status and changing trends; The execution review module is used to obtain and review the equipment feedback and monitoring feedback after adjustment.

2. The hemodialysis information management and monitoring system according to claim 1, characterized in that, The data access and filing module includes: A dialysis session is established based on the blood pump operating status and the first non-invasive blood pressure result. A session identifier is generated and a session master table is established, which includes the patient's identity identifier, bed identifier, dialysis equipment identifier, monitoring equipment identifier, session start time, and basic prescription information. Connect the dialysis equipment's operating parameters and physiological monitoring parameters, write the connected data into the raw acquisition buffer, and establish a raw data record table; A parameter mapping table is established based on the device protocol field encoding, parameter units, and clinical semantics. Link interruption determination, interruption continuation, and session merging are performed on the data from the device side and the physiological side.

3. The hemodialysis information management and monitoring system according to claim 1, characterized in that, The timing alignment module includes: Synchronous detection and original timestamp correction of raw data from dialysis and monitoring equipment; Based on the device correlation values ​​within the sliding window and adjacent real non-invasive blood pressure measurement results, a continuous blood pressure sequence is deduced, individual hysteresis time parameters are identified, and time-shifted matching of device parameter streams and physiological parameter streams is performed. Continuous physiological simulations should be stopped when key equipment parameters are missing or the interval between physiological reference data exceeds the limit.

4. The hemodialysis information management and monitoring system according to claim 1, characterized in that, The state association module includes: The synchronous records were organized into a status-related record table, and information on ultrafiltration changes, pulse pressure changes, transmembrane pressure changes, and heart rate changes were extracted according to different observation windows of different durations. Trend separation was performed on the estimated systolic blood pressure and heart rate sequences, and volume-pressure difference markers, pulse pressure narrowing markers, compensation enhancement markers, and compensation decay markers were generated by combining the device action characteristics. Add dehydration stage markers and retain original equipment parameters, baseline physiological parameters, and finite state markers when the reliability of physiological reference data is insufficient.

5. The hemodialysis information management and monitoring system according to claim 4, characterized in that, Trend separation was performed on the estimated systolic blood pressure and heart rate sequences, and volume-pressure difference markers, pulse pressure narrowing markers, compensatory enhancement markers, and compensatory attenuation markers were generated by combining equipment action characteristics. Modal decomposition was performed on systolic blood pressure estimation data and heart rate data to extract low-frequency main trends; Map the ultrafiltration operation curve, the cumulative ultrafiltration volume change curve, and the transmembrane pressure change curve to the same observation window; By combining the relationship between pulse pressure changes, mean arterial pressure changes, heart rate changes, and cumulative ultrafiltration volume changes, volume-pressure difference markers, pulse pressure narrowing markers, compensatory enhancement markers, and compensatory attenuation markers are generated according to numerical conditions and trends.

6. The hemodialysis information management and monitoring system according to claim 1, characterized in that, The risk assessment module includes: Based on the state association sequence, the volume compensation consumption divergence index, vascular compliance decay index and compensation exhaustion slope are extracted within the sliding decision window and normalized and weighted to form a risk confidence index. The risk threshold for each stage is dynamically set based on the target dehydration completion rate, and the risk status is determined by combining the duration and rate of change. An intervention urgency index is formed based on the risk confidence index, risk change trend, estimated systolic blood pressure and remaining dry weight deficit. Temporary encrypted monitoring is implemented for the yellow observation status. If the data is unreliable, the previous effective risk status is maintained and marked as pending review.

7. The hemodialysis information management and monitoring system according to claim 6, characterized in that, Based on the target dehydration completion rate, dynamic risk thresholds are set for each stage, and the risk status is determined by combining the duration and rate of change, including: The dehydration process is divided into early, middle, and final stages based on the target dehydration completion rate, and corresponding risk thresholds are set accordingly. The risk threshold in the mid-term stage tightens linearly with the completion rate of the target dehydration. The risk status is determined based on the duration of the risk confidence index exceeding the stage risk threshold and the smooth rate of change.

8. The hemodialysis information management and monitoring system according to claim 1, characterized in that, The equipment control module includes: Based on the risk level, corresponding control strategies are invoked. Under moderate risk conditions, the ultrafiltration rate and dialysate sodium concentration are jointly adjusted, while under severe risk conditions, the ultrafiltration rate and dialysate temperature are protected and adjusted. The control content is written into the control data packet and sent out via protocol framing. Based on the equipment acknowledgments, resend status, and deviations between target parameters and actual execution status, manual takeover or anomaly review is performed.

9. The hemodialysis information management and monitoring system according to claim 8, characterized in that, The control content is written into the control data packet and sent out via protocol framing, including: The control parameters are organized into control data packets, which include the dialysis session identifier, target parameter name, target parameter value, transition time, priority, generation time, validity period, and checksum. The edge gateway completes instruction framing according to the communication protocol corresponding to the target device. The effective time limit is determined based on the transition time and observation time. The priority of medium-risk control instructions is lower than that of high-risk control instructions. The instruction framing result is converted into a control message that the device can recognize and then sent.

10. A hemodialysis information management and monitoring system according to claim 1, characterized in that, The execution review module includes: Based on the equipment feedback and monitoring feedback after the control measures take effect, an independent review data sequence is formed, and time correspondence, status association and risk assessment are re-executed according to the short-term review window and the total review window. Based on the review and judgment criteria, implement step-by-step restoration of control, secondary regulation, or manual emergency takeover; The review information is written into the review record form, and the patient's individual tolerance parameters and key monitoring intervals are updated based on the review record.

Images