Method for constructing a volume management regimen for a patient on maintenance hemodialysis and use thereof

Abstract

This invention relates to the field of hemodialysis nursing technology, specifically to a method and application for constructing a volume management program for maintenance hemodialysis patients. The method includes: constructing an information anchoring layer; completing the construction of a basic patient volume management profile; synchronously collecting multi-dimensional data on patient basic information, dialysis treatment data, and physiological indicators; generating an individual characteristic map; and, based on the individual characteristic map, differentiating patients' cognitive abilities, management compliance, and risk levels to complete the hierarchical classification of needs. This invention constructs an individual patient characteristic map by integrating multi-dimensional clinical diagnosis and treatment data and conducting stratified assessments, matching differentiated intervention strategies, setting up a tiered volume management knowledge system, linking peers, families, and medical staff to form continuous positive support, and tracking daily management behaviors in real time and dynamically adjusting the intervention direction. This effectively improves the situation in routine volume management where guidance is fragmented, individual information is not fully collected, belief reinforcement measures are lacking, and dynamic feedback mechanisms are insufficient.

Description

Technical Field

[0001] This invention relates to the field of hemodialysis nursing technology, specifically to a method for constructing and applying a volume management program for maintenance hemodialysis patients. Background Technology

[0002] Maintenance hemodialysis is the core treatment for end-stage renal disease patients to maintain life. Volume management is a key aspect of dialysis nursing, which directly affects the patient's cardiac function, dialysis quality, and long-term survival. In clinical practice, related management work is carried out through health education, indicator monitoring, and other methods.

[0003] Current clinical volume management is mostly routine and fragmented guidance, without first comprehensively collecting information on the patient's individual disease, cognition, and management needs. It also lacks belief reinforcement measures and dynamic behavioral feedback mechanisms throughout the management process. As a result, patients cannot form a systematic understanding of volume management, lack sustained motivation when performing management behaviors, and find it difficult to improve their self-efficacy. Consequently, the control of interdialysis weight gain is poor, and it is difficult to effectively reduce dialysis-related acute complications. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method for constructing and applying a volume management program for maintenance hemodialysis patients, which solves the problems of fragmented management guidance, incomplete collection of individual information, lack of belief reinforcement measures, and insufficient dynamic feedback mechanisms in existing technologies.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for constructing a volume management program for maintenance hemodialysis patients, comprising: Construct an information anchoring layer to complete the basic profile of patient capacity management. Collect multi-dimensional data synchronously from patient basic information, dialysis treatment data, and physiological indicator data to generate individual feature maps. Based on the individual feature maps, distinguish patients’ cognitive abilities, management compliance, and risk levels to complete the hierarchical division of needs. A knowledge adaptation layer is constructed, and a knowledge content gradient library is established. Capacity management knowledge is divided into three gradients: basic cognition, advanced operation, and emergency response. The corresponding gradient knowledge is matched according to the required level. Knowledge transmission is promoted by a parallel mode of group common transmission and individual blind spot targeted transmission. The pace and depth of knowledge transmission are dynamically adjusted based on the results of phased cognitive tests. We build a belief-driven layer, which integrates three positive empowerment pathways: peer experience transmission, family emotional support, and medical and nursing value guidance. We regularly organize experience exchanges among patients with similar management levels, involve family members in daily management, and ensure that the medical and nursing team conveys the value of capacity management throughout the process. A behavioral evolution layer is constructed, relying on online recording channels to retain the trajectory of patients' daily management behaviors, and periodically conduct behavioral execution effect assessments. By comparing the trajectory data with preset standards and norms, the types and causes of behavioral execution deviations are identified. Based on the deviation analysis results, the information anchoring layer, knowledge adaptation layer, and belief-driven layer are linked in reverse to dynamically adjust the intervention strategies at each layer, forming an iterative cycle of deviation identification strategy optimization and behavior improvement.

[0006] Furthermore, the patient's basic information includes age, gender, education level, dialysis duration, comorbidities, and family support status; The dialysis treatment data includes dialysis frequency, dialysis duration, ultrafiltration volume, dialysate concentration, and vascular access type; The physiological data include predialysis blood pressure, postdialysis blood pressure, dry weight, interdialysis weight gain, cardiothoracic ratio, serum albumin, and hemoglobin.

[0007] Furthermore, the multi-dimensional data synchronization and collection includes the following steps: Establish standardized data collection templates and set the collection time points and collection cycles for data in each dimension; Patients fill in basic information through mobile devices or self-service terminals, and the system automatically extracts dialysis treatment data and physiological indicator data simultaneously. Perform integrity and consistency checks on the collected data from each dimension, identify missing data items, and trigger supplementary data collection processes; The validated data are linked and integrated according to the patient's unique identifier to generate a structured patient dataset.

[0008] Furthermore, the generation of individual feature maps includes the following steps: Discrete and continuous features are extracted from the patient dataset. Discrete features are vectorized using one-hot encoding, and continuous features are processed using interval discretization. A three-dimensional feature vector is constructed based on patient cognitive ability scores, management compliance scores, and risk level scores; The three-dimensional feature vectors are mapped to a high-dimensional feature space, and the patient groups are classified using a clustering algorithm. For each individual patient, an individual feature map containing both group and individual labels is generated by combining the common characteristics of their group category with their own specific indicators.

[0009] Furthermore, the hierarchical division of requirements includes the following steps: Patients were classified into three levels based on cognitive ability scores: low cognitive, medium cognitive, and high cognitive; three levels based on management compliance scores: low compliance, medium compliance, and high compliance; and three levels based on risk level scores: low risk, medium risk, and high risk. The hierarchical classification results of the three dimensions are combined to generate nine types of needs; for each type of need, different intervention intensity, intervention frequency, knowledge depth, and belief reinforcement methods are preset.

[0010] Furthermore, the knowledge content gradient base includes the following hierarchical content: The basic cognitive level includes the principles of dialysis, the mechanism of fluid balance, the significance of volume management, the principles of ultrafiltration, and the principles of dietary and water restriction; Advanced operational levels include methods for monitoring weight between dialysis sessions, dry weight assessment procedures, methods for calculating fluid intake, the relationship between exercise and volume management, and early identification of complications; The emergency response levels include management of hypotension during dialysis, management of hypertension during dialysis, identification of acute heart failure symptoms, and emergency management of volume overload.

[0011] Furthermore, the phased cognitive test includes the following steps: A cognitive assessment questionnaire is sent to the patient at preset time intervals. The questionnaire covers the key knowledge points recently conveyed. The system automatically calculates the patient's correct answer rate and answer time. The accuracy rate is compared with a preset threshold, and knowledge points that are below the threshold are marked as weak areas of mastery. Based on the distribution of weak areas, determine the direction and depth of content adjustments for the next stage of knowledge transfer.

[0012] Furthermore, the real-time tracking of the behavioral trajectory includes the following steps: Patients record their daily fluid intake, food types, actual weight, and subjective feelings via mobile devices. The system automatically calculates the weight gain during the interdialysis period and compares it with the target weight gain range; Generate daily behavior completion scores and phased behavior trend charts; When a patient's fluid intake exceeds the preset range or their weight gain exceeds the target upper limit for several consecutive days, a deviation warning signal is triggered.

[0013] Furthermore, the dynamic adjustment of intervention strategies at each level includes the following steps: When the behavioral evolution layer identifies a patient’s behavioral deviation, it transmits the deviation type and its cause to the information anchoring layer to reassess the patient’s characteristic profile. If the reassessment results show changes in the patient's cognitive abilities, the knowledge transfer gradient of the knowledge adaptation layer will be adjusted. If the reassessment results show that the patient's belief support is insufficient, then the empowerment pathways of the relatively weak links in the belief-driven layer should be strengthened. If the reassessment results show that the patient's risk level has increased, the monitoring frequency of the information anchoring layer will be increased and the intervention plan will be updated accordingly.

[0014] This invention also provides an application method for a volume management scheme for maintenance hemodialysis patients, applicable to patient management scenarios in hemodialysis centers of medical institutions, including: Data acquisition module, evaluation and analysis module, intervention implementation module, and tracking and feedback module; The data acquisition module is used to collect basic patient information, dialysis treatment data, and physiological indicator data and store them in the database; The assessment and analysis module is used to generate individual characteristic profiles based on data in the database, classify demand levels, assess knowledge mastery, and determine behavioral deviations. The intervention execution module is used to perform knowledge transfer, belief reinforcement, and strategy adjustment based on the output of the assessment and analysis module. The tracking and feedback module is used to record patient behavior and generate feedback data for the evaluation and analysis module.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention integrates multi-dimensional clinical data to construct individual patient characteristic profiles and conducts stratified assessments. It matches differentiated intervention strategies, establishes a tiered capacity management knowledge system, and links peers, families, and healthcare providers to form continuous positive support. It tracks daily management behaviors in real time and dynamically adjusts the direction of intervention. This effectively improves the situation in routine capacity management where guidance is fragmented, individual information is not fully collected, belief reinforcement measures are lacking, and dynamic feedback mechanisms are insufficient. It helps patients establish a systematic and coherent understanding of capacity management, maintain stable motivation for self-management, rationally control weight gain between dialysis sessions, reduce the occurrence of dialysis-related acute complications, and realize the transformation of capacity management from homogeneous education to individualized and dynamic management, adapting to the actual management needs of patients with different cognitive levels, compliance status, and risk levels. Attached Figure Description

[0016] Figure 1 This is a diagram of the four-layer linkage closed-loop architecture of the present invention; Figure 2 A general flowchart for the capacity management scheme of this invention is provided. Figure 3 This is a flowchart illustrating the multi-dimensional data synchronization and collection process of the present invention. Figure 4 This is a flowchart of the individual feature map generation process of the present invention; Figure 5 This is a flowchart of the behavioral deviation identification and strategy adjustment process of the present invention; Figure 6 This is a pie chart showing the distribution of patient needs types according to 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] This implementation method is designed for clinical scenarios involving volume management of maintenance hemodialysis patients. It constructs a hierarchical, data-driven, and dynamically iterative volume management solution, which fully covers the entire volume management process from data collection, feature modeling, hierarchical intervention to behavioral closed-loop optimization. It solves problems such as fragmented guidance, cognitive gaps, insufficient motivation, and delayed feedback under the existing management model. The following provides a detailed explanation in conjunction with data sources, algorithm implementation, parameter calibration, and module collaboration.

[0019] Please see Figures 1-6 This invention provides a method for constructing a volume management program for maintenance hemodialysis patients, comprising: constructing an information anchoring layer to complete the construction of a basic patient volume management profile; synchronously collecting multi-dimensional data on patient basic information, dialysis treatment data, and physiological indicator data to generate an individual feature map; and differentiating patients' cognitive abilities, management compliance, and risk levels based on the individual feature map to complete the hierarchical division of needs; constructing a knowledge adaptation layer to establish a knowledge content gradient library; dividing volume management knowledge into three gradients: basic cognition, advanced operations, and emergency treatment; matching corresponding gradient knowledge according to the hierarchical division of needs; and promoting knowledge transfer using a parallel mode of group commonality transmission and individual blind spot targeted transmission, based on the results of phased cognitive tests. The system dynamically adjusts the pace and depth of knowledge transmission; it constructs a belief-driven layer, aggregating three positive empowerment pathways: peer experience transmission, family emotional support, and medical value guidance. It regularly organizes experience exchanges among patients at similar management levels, integrates family members into daily management processes, and ensures the medical team delivers capacity management value throughout the process. It also constructs a behavior evolution layer, relying on online recording channels to retain patients' daily management behavior trajectories. Periodic evaluations of behavior execution effectiveness are conducted, and by comparing trajectory data with preset standards and norms, the types and causes of behavioral deviations are identified. Based on the deviation analysis results, the system reversely links the information anchoring layer, knowledge adaptation layer, and belief-driven layer, dynamically adjusting intervention strategies at each layer to form an iterative cycle of deviation identification, strategy optimization, and behavior improvement.

[0020] Specifically, this solution is based on a four-layer architecture. All aspects rely on three data sources: the hospital's clinical database, the dialysis treatment system, and patient mobile terminals. Data is collected automatically from daily clinical practice, self-reported by patients, and aggregated from real-time equipment monitoring, ensuring that the data accurately reflects the clinical scenario. The information anchoring layer, as the data foundation, is responsible for cleaning raw data, extracting features, and modeling, outputting standardized patient profiles. The knowledge adaptation layer, based on the profile's hierarchical results and combined with learning patterns, generates gradient knowledge content to achieve precise knowledge delivery. The belief-driven layer connects patients, families, and medical staff to build a psychological and emotional support system, compensating for the limitations of single-knowledge intervention. The behavior evolution layer collects patients' daily behavioral data in real time, identifies deviations through algorithms, and drives the iteration of the strategies in the first three layers, forming a closed-loop management system. The four layers interact in real time through standardized data interfaces. Feature data output by the information anchoring layer is synchronously pushed to the knowledge adaptation layer and the belief-driven layer. Deviation results from the behavior evolution layer are fed back to the information anchoring layer for remodeling, achieving dynamic linkage across the entire chain. This avoids policy rigidity, ensures the individual adaptability of management solutions from the data source, and directly improves the inefficient intervention problem caused by "one-size-fits-all" guidance in conventional management.

[0021] In one specific embodiment, the patient's basic information includes age, gender, education level, dialysis history, comorbidities, and family support status; dialysis treatment data includes dialysis frequency, dialysis duration, ultrafiltration volume, dialysate concentration, and vascular access type; physiological indicators include predialysis blood pressure, postdialysis blood pressure, dry weight, interdialysis weight gain, cardiothoracic ratio, serum albumin, and hemoglobin.

[0022] Specifically, the data dimensions were selected based on routine clinical practice in the diagnosis and treatment of end-stage renal disease dialysis, covering three core dimensions: patient background, treatment process, and physiological status. The data sources were clearly defined: basic patient information was extracted from the hospital's HIS system and patient admission records; dialysis treatment data was automatically generated by the dialysis machine monitoring module and dialysis record system; and physiological indicator data were obtained through manual testing by medical staff before and after dialysis and analysis by laboratory equipment. All data were incorporated into a standardized hospital database for unified management, avoiding data fragmentation and loss. Information such as age and dialysis duration was used to assess the patient's physiological tolerance and treatment familiarity; ultrafiltration volume and dry weight were directly related to core parameters of volume balance; and serum albumin and hemoglobin reflected the patient's nutritional and anemia status. These three types of data complemented each other, comprehensively depicting the individual characteristics related to patient volume management. The comprehensiveness of the data dimensions directly determined the accuracy of subsequent profiling, providing a reliable data foundation for stratified intervention and avoiding deviations in intervention direction due to missing dimensions.

[0023] In one specific embodiment, the multi-dimensional data synchronous collection includes the following steps: establishing a standardized data collection template and setting the collection time and collection cycle for each dimension of data; patients fill in basic information through mobile devices or self-service terminals, and the system automatically extracts dialysis treatment data and physiological indicator data synchronously; performing integrity verification and consistency verification on the collected data of each dimension, identifying missing data items and triggering the supplementary collection process; and associating and integrating the verified data according to the patient's unique identifier to generate a structured patient dataset.

[0024] Specifically, the data collection process is optimized based on clinical data collection standards. A unified collection template is first established, and the collection time and cycle parameters are calibrated in conjunction with dialysis treatment patterns: basic information is collected initially in a single session; dialysis treatment data is collected once before and after each dialysis session, with the cycle consistent with the dialysis frequency; physiological indicators such as blood pressure and weight are collected before and after each dialysis session, while serum albumin and hemoglobin are collected every two weeks, with parameter values ​​based on the "Hemodialysis Quality Control Standards" to ensure timely collection. Data sources are divided into three parties: patients fill in subjective information such as education level and family support status through a mobile app; the dialysis system interface automatically captures treatment data such as dialysis frequency and ultrafiltration volume; and the laboratory interface synchronizes physiological indicator data to reduce manual input errors. The data verification process uses preset rule logic verification. Integrity verification is performed by checking if fields are not empty, and consistency verification is performed by comparing logical relationships between indicators. For example, interdialysis weight gain must not exceed 5% of dry weight. Verification thresholds are derived from clinical big data statistics. If data is abnormal, a pop-up reminder is triggered on the mobile app to supplement or correct the data. Finally, by associating and integrating the data with the patient's unique hospital number, a structured dataset is generated. The dataset has a unified field format, no redundancy, and a high data qualification rate. The improved data quality directly reduces the modeling error of subsequent models, provides reliable input for feature map generation, and avoids feature distortion caused by dirty data.

[0025] In a specific embodiment, generating an individual feature map includes the following steps: extracting discrete and continuous features from the patient dataset, vectorizing the discrete features using one-hot encoding, and discretizing the continuous features using interval discretization; constructing a three-dimensional feature vector based on the patient's cognitive ability score, management compliance score, and risk level score; mapping the three-dimensional feature vector to a high-dimensional feature space, and dividing the patient group into categories using a clustering algorithm; and generating an individual feature map containing group labels and individual labels for each individual patient, combining the common features of their group category with their own specific indicators.

[0026] Specifically, feature map generation relies on conventional machine learning feature processing and clustering methods, fine-tuned based on the characteristics of medical data. In the data preprocessing stage, discrete features such as gender and comorbidity type are encoded using one-hot encoding, converting non-numerical features into 0-1 numerical vectors. For example, male gender is encoded as [1,0] and female as [0,1]. Continuous features such as age and dialysis duration are discretized into intervals based on clinical patient distribution patterns. For example, age is divided into three groups: 18-44 years, 45-64 years, and ≥65 years; dialysis duration is divided into three groups: <1 year, 1-3 years, and >3 years. This process eliminates dimensional differences, facilitating subsequent calculations. Three-dimensional scores are quantified using clinical scales. Cognitive ability scores use the Simplified Mental State Examination (SMSA), compliance scores combine dietary and fluid restriction rates with weight monitoring adherence, and risk level scores are based on the probability of volume-related complications. All scores are set to a range of 0-100 points, constructing a three-dimensional feature vector. ; In the formula, Scoring cognitive abilities To manage compliance scores, For risk level scoring, the vector dimension is determined by the number of scoring indicators, and the unit of measurement is uniformly points.

[0027] Clustering uses the K-means algorithm, which calculates sample similarity based on Euclidean distance: ; In the formula, The distance between the feature vectors of two patients; the smaller the distance, the more similar the individual characteristics. Number of clusters. The parameter is set to 3, calibrated through clinical pre-experiments, corresponding to low, medium, and high characteristic groups respectively, avoiding excessive clustering leading to group fragmentation and insufficient clustering leading to coarse classification. The algorithm iterates until the cluster center no longer changes, outputting the group classification result. Then, it combines individual-specific indicators such as comorbidity type and vascular access type to generate an atlas containing group and individual labels. For example, a patient who is 60 years old, has been on dialysis for 2 years, has a cognitive score of 75, a compliance score of 70, and a risk score of 65, is clustered into the medium characteristic group, and the individual comorbidity hypertension label is superimposed to form a complete atlas. This process uses automatic modeling through algorithms, reducing subjective human judgment, accurately calculating feature similarity, and ensuring that group division is consistent with clinical reality. Improved atlas accuracy can directly reduce the mismatch rate of subsequent stratified interventions, making interventions more tailored to individual patient conditions.

[0028] In a specific embodiment, the needs hierarchy classification includes the following steps: classifying patients into three levels—low cognitive, medium cognitive, and high cognitive—based on cognitive ability scores; classifying patients into three levels—low compliance, medium compliance, and high compliance—based on management compliance scores; and classifying patients into three levels—low risk, medium risk, and high risk—based on risk level scores; combining the classification results of the three dimensions to generate nine needs types; and pre-setting differentiated intervention intensity, intervention frequency, knowledge depth, and belief reinforcement methods for each needs type.

[0029] Specifically, the demand hierarchy is based on the logic of stratified management of clinical patients. Scoring thresholds are determined using a large sample of dialysis patients' data: cognitive ability score <60 indicates low cognition, 60-80 indicates medium cognition, and >80 indicates high cognition; management compliance score <65 indicates low compliance, 65-85 indicates medium compliance, and >85 indicates high compliance; risk level score <50 indicates low risk, 50-75 indicates medium risk, and >75 indicates high risk. The threshold values ​​balance patient distribution with the feasibility of clinical intervention, avoiding an excessively small or large number of patients in any particular stratum. After independent stratification across these three dimensions, they are combined in pairs to form nine demand categories, covering all cognitive-compliance-risk combinations without any blind spots. Intervention parameters are graded, with intervention intensity divided into low, medium, and high levels; frequency divided into once a week, twice a week, and once a day; knowledge depth is matched to the gradient library hierarchy; and belief reinforcement methods correspond to three empowerment paths. For example, for high-risk patients with low adherence, the intervention intensity is set to high, the frequency is once daily, the knowledge depth covers the entire gradient, and family support and medical guidance are strengthened; for low-risk patients with high cognition, the intervention intensity is set to low, the frequency is once a week, the knowledge depth focuses on basic cognition, and the emphasis is on peer experience exchange. The differentiated setting of stratified parameters can accurately match the intervention needs of different patients, avoiding the waste of resources from high-intensity interventions and the insufficient effect of low-intensity interventions. Improved intervention suitability directly optimizes management efficiency and improves the problem of "heavy intervention, light suitability" in routine management.

[0030] In one specific embodiment, the knowledge content gradient base includes the following hierarchical levels: the basic cognitive level includes dialysis principles, fluid balance mechanisms, the significance of volume management, ultrafiltration principles, and dietary and water restriction principles; the advanced operational level includes interdialysis weight monitoring methods, dry weight assessment procedures, fluid intake calculation methods, the relationship between exercise and volume management, and early identification of complications; and the emergency response level includes management of hypotension during dialysis, management of hypertension during dialysis, identification of acute heart failure manifestations, and emergency treatment of volume overload.

[0031] Specifically, the knowledge gradient base is built upon hemodialysis nursing guidelines and patient learning patterns. The difficulty gradient progresses gradually from theory to practice, and from routine to emergency situations, adapting to the learning abilities of patients at different cognitive levels. The basic cognitive level uses simple language and simplified principles, suitable for patients with low cognitive abilities, addressing the "why" question. The advanced operational level focuses on practical steps and quantifiable methods, suitable for patients with medium to high cognitive abilities, addressing the "how" question. The emergency response level focuses on emergency scenarios and key points for rapid response, suitable for patients with medium to high risk, addressing the "what to do in case of an abnormality" question. The gradient difficulty coefficient is determined by content complexity: 1.0 for the basic level, 1.8 for the advanced level, and 2.5 for the emergency level. These coefficients are set based on the difficulty of understanding the knowledge points and the complexity of the practical operations, avoiding excessive jumps in difficulty that could cause learning difficulties for patients. This gradient content allows patients to learn step by step, improving knowledge absorption efficiency, reducing learning frustration, strengthening capacity management fundamentals from a cognitive perspective, and reducing the probability of behavioral errors due to insufficient understanding.

[0032] In one specific embodiment, the phased cognitive test includes the following steps: sending a cognitive assessment questionnaire to the patient at preset time intervals, the questionnaire covering the key knowledge points recently conveyed; the system automatically calculates the patient's correct answer rate and answer time; the correct answer rate is compared with a preset threshold, and knowledge points below the threshold are marked as weak areas of mastery; based on the distribution of weak areas of mastery, the direction and depth of the content adjustment for the next stage of knowledge transmission are determined.

[0033] Specifically, the cognitive testing mechanism is designed based on the forgetting curve of adult learning. The testing cycle parameters are combined with the difficulty level of the knowledge points: 7 days for basic cognitive level, 14 days for advanced operational level, and 10 days for emergency response level. The cycle settings match the rhythm of memory consolidation, avoiding too short intervals that increase the burden and too long intervals that lead to forgetting. The assessment questionnaire consists of 10 questions randomly selected from the knowledge base, in multiple-choice and true / false formats, ensuring quick completion. The formula for calculating the accuracy rate is: ; In the formula, For accuracy, To answer the number of questions correctly, The total number of questions is set, with a 70% accuracy threshold. This threshold is determined by clinical learning qualification standards; scores below the threshold are considered weaknesses. For example, if a patient answers 6 out of 10 questions correctly on a basic cognitive test (accuracy 60% < 70%), the "dietary and water restriction principle" is marked as a weakness. The frequency of this knowledge point will be increased, and the explanation method will be simplified. The system automatically adjusts its knowledge delivery strategy based on the distribution of weaknesses, addressing specific gaps in knowledge, avoiding ineffective knowledge delivery, improving the accuracy of knowledge delivery, continuously reinforcing the patient's understanding, reducing cognitive blind spots, and laying a solid cognitive foundation for behavioral execution.

[0034] In one specific embodiment, real-time tracking of behavioral trajectories includes the following steps: the patient records daily fluid intake, food type, actual weight, and subjective feelings via a mobile device; the system automatically calculates the weight gain value between dialysis sessions and compares it with the target growth range; a daily behavior completion score and a phased behavior trend chart are generated; when the patient's fluid intake exceeds the preset range or the weight gain value exceeds the target upper limit for several consecutive days, a deviation warning signal is triggered.

[0035] Specifically, behavioral tracking relies on lightweight mobile data collection tools, with data self-recorded daily by patients, simplifying the process and reducing the burden of recording. Fluid intake includes all liquid foods such as water, soup, porridge, and fruit, accurate to the ml; weight is measured daily in the morning on an empty stomach, after emptying the bladder and bowels, accurate to 0.1 kg, minimizing measurement error. The formula for calculating weight gain between dialysis sessions is: ; In the formula, For weight gain value, Weight before this dialysis, The target weight gain is set at 5% of dry body weight, based on the dialysis volume management standards. For example, for a patient with a dry body weight of 50 kg, the upper limit is 2.5 kg. The behavioral completion score formula is: ; In the formula, To score completion To record completeness scores, Weighting for data compliance score , The weighting is determined by clinical importance, with compliance receiving a higher weight. The alert trigger condition is set at three consecutive days of abnormality, and parameters are calibrated based on the cumulative risk pattern to avoid false alerts from single fluctuations. For example, if a patient has a dry weight of 50kg, and their fluid intake exceeds 2000ml for three consecutive days with a daily weight gain exceeding 0.8kg, the system will automatically send an alert to the patient and medical staff. Real-time tracking makes behavioral data visible and provides early warnings of abnormalities, solving the problems of uncontrollable behavior and delayed problem detection in routine management. Improved timeliness of behavioral control can effectively control weight gain between dialysis sessions and reduce the risk of volume overload.

[0036] In a specific embodiment, dynamically adjusting the intervention strategies at each layer includes the following steps: when the behavioral evolution layer identifies behavioral deviations in a patient, the deviation type and cause are transmitted to the information anchoring layer to reassess the patient's characteristic profile; if the reassessment results show changes in the patient's cognitive abilities, the knowledge transfer gradient of the knowledge adaptation layer is adjusted; if the reassessment results show insufficient belief support in the patient, the empowerment pathways of relatively weak links in the belief-driven layer are strengthened; if the reassessment results show an increased risk level in the patient, the monitoring frequency of the information anchoring layer is increased and the intervention plan is updated synchronously.

[0037] Specifically, the strategy adjustment is based on a four-layer closed-loop linkage logic. Modules transmit data in real time through standardized interfaces, with a response latency of less than 1 hour, ensuring timely strategy adjustments. Deviation causes are categorized into three types: insufficient cognition, weak belief, and increased risk. Deviation data includes type, duration, and degree of impact, and is fed back to the information anchoring layer in real time, triggering feature map remodeling and updating cognition, compliance, and risk scores. If the cognition score improves by more than 5 points, the knowledge adaptation layer adjusts the gradient upwards, for example, from basic cognition to advanced operations. If belief support is insufficient, such as low family involvement, the belief-driven layer increases the frequency of communication with family members and pushes home care guidance. If the risk level score increases by more than 10 points, the monitoring frequency increases from once a week to once a day, and the intervention intensity is increased simultaneously. For example, if a patient's excessive fluid intake is due to insufficient cognition, after deviation feedback, the cognitive score is reassessed and decreases, the knowledge gradient is adjusted downwards, and basic cognition explanations are strengthened. One week later, cognition improves and behavior meets the target. Closed-loop linkage allows strategies to dynamically adapt to the patient's condition, avoiding strategy rigidity, continuously optimizing intervention adaptability, forming a virtuous cycle of "deviation-adjustment-improvement", and improving the patient's self-management ability in the long term.

[0038] This invention also provides an application method for a volume management scheme for maintenance hemodialysis patients. The method is applied to patient management scenarios in hemodialysis centers of medical institutions, including: a data acquisition module, an assessment and analysis module, an intervention execution module, and a tracking and feedback module. The data acquisition module collects basic patient information, dialysis treatment data, and physiological indicator data and stores them in a database. The assessment and analysis module generates individual characteristic profiles based on the data in the database, classifies needs levels, assesses knowledge mastery, and determines behavioral deviations. The intervention execution module performs knowledge transfer, belief reinforcement, and strategy adjustment based on the output results of the assessment and analysis module. The tracking and feedback module records patient behavior trajectories and generates feedback data for the assessment and analysis module.

[0039] Specifically, the application method relies on the modular expansion of the existing dialysis management system in medical institutions. Four main modules are deployed on the hospital's intranet, enabling data interoperability and collaborative operation, adapting to the scenario of managing batches of patients in dialysis centers. The data acquisition module connects to the HIS system, dialysis machines, and a mobile app, unifying multi-source data into a single database with a data synchronization delay of less than 30 minutes. The assessment and analysis module integrates feature modeling, clustering algorithms, and scoring calculation models to automatically complete patient assessments, with a single patient assessment taking less than 10 seconds, reducing manual workload. The intervention execution module outputs personalized intervention lists to guide medical staff in conducting education, family collaboration, and patient communication. The tracking and feedback module collects behavioral data in real time and pushes it to the assessment and analysis module, forming a closed loop. The module collaboration process is: data acquisition → assessment and analysis → intervention execution → tracking and feedback → data feedback and reassessment, with a high degree of automation and convenient clinical operation. For example, after dialysis centers adopt this solution, medical staff no longer need to manually organize data or perform manual stratification. The system automatically outputs intervention plans, significantly improving management efficiency. At the same time, patient compliance and capacity achievement rates improve simultaneously, and the probability of complications decreases. It is well-suited to the needs of large-scale and standardized clinical management and has good clinical applicability.

[0040] In summary, this invention integrates multi-dimensional clinical data to construct individual patient characteristic profiles and conducts stratified assessments. It matches differentiated intervention strategies, establishes a tiered capacity management knowledge system, and fosters continuous positive support through collaboration with peers, families, and healthcare professionals. It also tracks daily management behaviors in real time and dynamically adjusts intervention directions. This effectively improves upon the fragmented guidance, incomplete collection of individual information, lack of belief reinforcement measures, and insufficient dynamic feedback mechanisms in routine capacity management. It helps patients establish a systematic and coherent understanding of capacity management, maintain stable motivation for self-management, rationally control interdialysis weight gain, and reduce the occurrence of dialysis-related acute complications. This transforms capacity management from homogeneous education to individualized and dynamic control, adapting to the actual management needs of patients with different cognitive levels, compliance statuses, and risk levels.

[0041] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0042] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for constructing a volume management protocol for maintenance hemodialysis patients, characterized in that, include: Construct an information anchoring layer to complete the basic profile of patient capacity management. Collect multi-dimensional data synchronously from patient basic information, dialysis treatment data, and physiological indicator data to generate individual feature maps. Based on the individual feature maps, distinguish patients’ cognitive abilities, management compliance, and risk levels to complete the hierarchical division of needs. A knowledge adaptation layer is constructed, and a knowledge content gradient library is established. Capacity management knowledge is divided into three gradients: basic cognition, advanced operation, and emergency response. The corresponding gradient knowledge is matched according to the required level. Knowledge transmission is promoted by a parallel mode of group common transmission and individual blind spot targeted transmission. The pace and depth of knowledge transmission are dynamically adjusted based on the results of phased cognitive tests. We build a belief-driven layer, which integrates three positive empowerment pathways: peer experience transmission, family emotional support, and medical and nursing value guidance. We regularly organize experience exchanges among patients with similar management levels, involve family members in daily management, and ensure that the medical and nursing team conveys the value of capacity management throughout the process. A behavioral evolution layer is constructed, relying on online recording channels to retain the trajectory of patients' daily management behaviors, and periodically conduct behavioral execution effect assessments. By comparing the trajectory data with preset standards and norms, the types and causes of behavioral execution deviations are identified. Based on the deviation analysis results, the information anchoring layer, knowledge adaptation layer, and belief-driven layer are linked in reverse to dynamically adjust the intervention strategies at each layer, forming an iterative cycle of deviation identification strategy optimization and behavior improvement.

2. The method for constructing a volume management program for maintenance hemodialysis patients according to claim 1, characterized in that, The patient's basic information includes age, gender, education level, dialysis duration, comorbidities, and family support status; The dialysis treatment data includes dialysis frequency, dialysis duration, ultrafiltration volume, dialysate concentration, and vascular access type; The physiological data include predialysis blood pressure, postdialysis blood pressure, dry weight, interdialysis weight gain, cardiothoracic ratio, serum albumin, and hemoglobin.

3. The method for constructing a volume management program for maintenance hemodialysis patients according to claim 1, characterized in that, The multi-dimensional data synchronization and collection includes the following steps: Establish standardized data collection templates and set the collection time points and collection cycles for data in each dimension; Patients fill in basic information through mobile devices or self-service terminals, and the system automatically extracts dialysis treatment data and physiological indicator data simultaneously. Perform integrity and consistency checks on the collected data from each dimension, identify missing data items, and trigger supplementary data collection processes; The validated data are linked and integrated according to the patient's unique identifier to generate a structured patient dataset.

4. The method for constructing a volume management program for maintenance hemodialysis patients according to claim 1, characterized in that, The generation of individual feature maps includes the following steps: Discrete and continuous features are extracted from the patient dataset. Discrete features are vectorized using one-hot encoding, and continuous features are processed using interval discretization. A three-dimensional feature vector is constructed based on patient cognitive ability scores, management compliance scores, and risk level scores; The three-dimensional feature vectors are mapped to a high-dimensional feature space, and the patient groups are classified using a clustering algorithm. For each individual patient, an individual feature map containing both group and individual labels is generated by combining the common characteristics of their group category with their own specific indicators.

5. The method for constructing a volume management program for maintenance hemodialysis patients according to claim 1, characterized in that, The demand hierarchy division includes the following steps: Patients were classified into three levels based on cognitive ability scores: low cognitive, medium cognitive, and high cognitive; three levels based on management compliance scores: low compliance, medium compliance, and high compliance; and three levels based on risk level scores: low risk, medium risk, and high risk. The hierarchical classification results of the three dimensions are combined to generate nine types of needs; for each type of need, different intervention intensity, intervention frequency, knowledge depth, and belief reinforcement methods are preset.

6. The method for constructing a volume management program for maintenance hemodialysis patients according to claim 1, characterized in that, The knowledge content gradient base includes the following hierarchical content: The basic cognitive level includes the principles of dialysis, the mechanism of fluid balance, the significance of volume management, the principles of ultrafiltration, and the principles of dietary and water restriction; Advanced operational levels include methods for monitoring weight between dialysis sessions, dry weight assessment procedures, methods for calculating fluid intake, the relationship between exercise and volume management, and early identification of complications; The emergency response levels include management of hypotension during dialysis, management of hypertension during dialysis, identification of acute heart failure symptoms, and emergency management of volume overload.

7. The method for constructing a volume management program for maintenance hemodialysis patients according to claim 1, characterized in that, The phased cognitive test includes the following steps: A cognitive assessment questionnaire is sent to the patient at preset time intervals. The questionnaire covers the key knowledge points recently conveyed. The system automatically calculates the patient's correct answer rate and answer time. The accuracy rate is compared with a preset threshold, and knowledge points that are below the threshold are marked as weak areas of mastery. Based on the distribution of weak areas, determine the direction and depth of content adjustments for the next stage of knowledge transfer.

8. The method for constructing a volume management program for maintenance hemodialysis patients according to claim 1, characterized in that, The real-time tracking of the behavior trajectory includes the following steps: Patients record their daily fluid intake, food types, actual weight, and subjective feelings via mobile devices. The system automatically calculates the weight gain during the interdialysis period and compares it with the target weight gain range; Generate daily behavior completion scores and phased behavior trend charts; When a patient's fluid intake exceeds the preset range or their weight gain exceeds the target upper limit for several consecutive days, a deviation warning signal is triggered.

9. The method for constructing a volume management program for maintenance hemodialysis patients according to claim 1, characterized in that, The dynamic adjustment of intervention strategies at each level includes the following steps: When the behavioral evolution layer identifies a patient’s behavioral deviation, it transmits the deviation type and its cause to the information anchoring layer to reassess the patient’s characteristic profile. If the reassessment results show changes in the patient's cognitive abilities, the knowledge transfer gradient of the knowledge adaptation layer will be adjusted. If the reassessment results show that the patient's belief support is insufficient, then the empowerment pathways of the relatively weak links in the belief-driven layer should be strengthened. If the reassessment results show that the patient's risk level has increased, the monitoring frequency of the information anchoring layer will be increased and the intervention plan will be updated accordingly.

10. A method for applying a volume management protocol for maintenance hemodialysis patients, characterized in that, The method for constructing a maintenance hemodialysis patient volume management scheme as described in any one of claims 1 to 9 is applied to the patient management scenario of a hemodialysis center in a medical institution, including: Data acquisition module, evaluation and analysis module, intervention implementation module, and tracking and feedback module; The data acquisition module is used to collect basic patient information, dialysis treatment data, and physiological indicator data and store them in the database; The assessment and analysis module is used to generate individual characteristic profiles based on data in the database, classify demand levels, assess knowledge mastery, and determine behavioral deviations. The intervention execution module is used to perform knowledge transfer, belief reinforcement, and strategy adjustment based on the output of the assessment and analysis module. The tracking and feedback module is used to record patient behavior and generate feedback data for the evaluation and analysis module.

Images