Clinical rehabilitation data management and evaluation system for rhabdomyolysis patient

Abstract

The invention provides a clinical rehabilitation data management and evaluation system for a rhabdomyolysis patient, and relates to a medical information management system. The system comprises an index acquisition module, a rate detection module, a curvature analysis module, a node identification module and a trend construction module. According to the method, the rate difference sequence of creatine kinase and myohemoglobin and the curvature sequence of urine sodium and serum creatinine are compositely calculated, the limitation of traditional single-parameter static analysis is broken through, the dynamic capture of a multi-index collaborative trend is realized, and a joint evolution label system is constructed by combining the consistency judgment of creatine kinase and blood potassium concentration change directions; according to the method, multi-system interaction risk nodes are accurately positioned, statistics is performed on direction consistency proportional distribution of myoglobin and serum creatinine in an association time period, discrete parameter association is converted into a dynamic visual model, and the problems of data discreteness misjudgment and collaborative evolution mode splitting in the traditional technology are solved. And a multi-dimensional dynamic evaluation framework of the rhabdomyolysis rehabilitation process is provided for clinic.

Description

Technical Field

[0001] The present invention relates to the technical field of medical information management systems, and in particular to a clinical rehabilitation data management and evaluation system for patients with rhabdomyolysis. Background Art

[0002] The technical field of medical information management systems encompasses the systematic collection, structured management, and analytical processing of the vast amounts of heterogeneous data generated during clinical medical care. With the core goals of improving clinical decision-making efficiency, optimizing resource allocation, and enabling intelligent disease management, this technical field covers multiple key areas, including patient medical record archiving, dynamic monitoring of health indicators, tracking of disease progression, and management of individualized treatment pathways. By integrating clinical information resources such as electronic health records, physiological parameter monitoring data, and test results, medical information management systems establish a cross-departmental and cross-process data linkage mechanism, enabling refined data support and information coordination throughout the patient's medical care process, ensuring the efficient operation of all aspects of clinical work under a unified data platform.

[0003] Among them, a clinical rehabilitation data management and evaluation system for rhabdomyolysis patients refers to the key physiological data and disease change information involved in the clinical treatment and rehabilitation process of rhabdomyolysis patients. It periodically collects and archives data such as patient muscle enzyme level changes, urine volume records, and renal function test values ​​in a unified format, and continuously tracks and dynamically classifies various parameters based on the set indicator classification standards. It combines the timeline of disease evolution to perform time series data comparison and trend pre-assessment. The system generally uses multi-period data point extraction, physiological indicator correlation mapping, data standardization comparison and other means to assist in the fine-grained identification of patient status and clinical information evaluation, providing data support for the formulation of treatment plans.

[0004] Existing technologies for data management of patients with rhabdomyolysis still suffer from structural deficiencies in the collaborative analysis of multidimensional parameters and the identification of dynamic trends. Traditional methods rely on periodic data collection and independent archiving of parameters, resulting in a break in the temporal correlation of key indicators such as creatine kinase and myoglobin, and failing to reflect the interplay between different biochemical indicators during the recovery process. Existing systems limit monitoring of rate changes to static comparisons within a single time window and fail to establish a continuous analysis framework for adjacent rate differences, causing the identification of abnormal fluctuation points to lag behind the actual disease progression. Regarding trend analysis, existing technologies employ single-parameter linear regression models and lack curvature features for indicators such as urine sodium concentration and serum creatinine, making it impossible to capture the coordinated timing of sudden changes in multi-parameter trends. Furthermore, existing methods lack a mechanism for verifying the direction of change across indicators, making it difficult to identify the joint risk evolution pattern of creatine kinase and serum potassium concentration, leading to reliance on fragmented data for clinical decision-making. Regarding data presentation, existing systems present test results in discrete reports, failing to transform multi-parameter trend correlations into temporally continuous visualizations, limiting the clinical applicability of dynamic rehabilitation assessments. These shortcomings make traditional technologies unable to meet the complex monitoring needs of multi-system interactions during the recovery process of rhabdomyolysis. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the existing technology and propose a clinical rehabilitation data management and evaluation system for patients with rhabdomyolysis.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] A clinical rehabilitation data management and evaluation system for patients with rhabdomyolysis includes:

[0008] The indicator collection module obtains the creatine kinase value, myoglobin value, urine sodium concentration, blood creatinine level, lactate dehydrogenase value, and blood potassium concentration of patients with rhabdomyolysis at time points during the recovery period, organizes the data in chronological order, and generates a patient continuous indicator distribution table;

[0009] The rate detection module extracts the time change rate of the creatine kinase value and the myoglobin value based on the patient continuous indicator distribution table, establishes a difference sequence between adjacent rates, determines the abnormal points of the rate change amplitude, and generates a recovery cycle rate abnormality record set;

[0010] The curvature analysis module extracts the temporal trend changes of urine sodium concentration, lactate dehydrogenase value, and blood creatinine level based on the abnormal recovery cycle rate record set and the patient continuous indicator distribution table, constructs a curvature sequence for the continuous data, identifies the trend mutation time point, and generates a curvature linkage inflection point set;

[0011] The node identification module locates the intersection time node of the curvature linkage inflection point set and the recovery cycle rate abnormality record set, calls the node creatine kinase value and blood potassium concentration, determines the consistency of the change direction, establishes a joint trend label based on the indicator change direction, and generates a risk joint evolution node label set.

[0012] Optionally, the patient continuous indicator distribution table includes creatine kinase value, myoglobin value, urine sodium concentration, blood creatinine level, lactate dehydrogenase value, and blood potassium concentration; the recovery cycle rate abnormality record set includes creatine kinase rate difference sequence, myoglobin rate difference sequence, and abnormal change amplitude point; the curvature linkage inflection point set includes urine sodium concentration curvature sequence, lactate dehydrogenase curvature sequence, blood creatinine curvature sequence, and trend mutation time point; the risk joint evolution node label set includes joint time node, indicator change direction consistency judgment result, and joint trend label.

[0013] Optionally, the indicator collection module includes:

[0014] The data acquisition submodule obtains the creatine kinase value, myoglobin value, urine sodium concentration, blood creatinine level, lactate dehydrogenase value, and blood potassium concentration at time points during the recovery period of patients with rhabdomyolysis. The data are aggregated by patient number and the indicators are assigned to the corresponding time points based on the time records to generate time series data for the six indicators.

[0015] The time arrangement submodule extracts the time nodes of each patient based on the time series data of the six indicators, calls the indicator values ​​corresponding to the time points, arranges the data in chronological order, calculates the interval values ​​between adjacent times, establishes a sequentially arranged data set, and generates a continuous time series structure value;

[0016] The indicator distribution submodule summarizes the six indicator values ​​at the time nodes according to the continuous time series structure value, forms an indicator arrangement table based on time advancement, calculates the value changes within the time period, establishes the distribution relationship of the indicators over time, and generates a patient continuous indicator distribution table.

[0017] Optionally, the specific calculation formula for arranging the data in chronological order is:

[0018]

[0019] Where Δt i represents the weighted time interval between the i-th time point and the i+1-th time point, T i Represents the time value of the i-th time node, T i+1 Represents the time value of the i+1th time node, V ij represents the original value of the i-th time node under the j-th indicator, T jRepresents the weighted coefficient corresponding to the jth indicator (note that T j With T i different, does not represent a time value), m represents the total number of indicators, j is the indicator index from 1 to m, and i is the index of the current time point.

[0020] Optionally, the rate detection module includes:

[0021] The rate extraction submodule obtains the creatine kinase value and the myoglobin value in the patient continuous indicator distribution table based on the patient continuous indicator distribution table, locates the continuous records of the indicators in chronological order, constructs a rate sequence based on the numerical relationship between adjacent time points, and generates a time-varying rate value set;

[0022] The difference construction submodule extracts the numerical changes of adjacent rates based on the sequential relationship between the rate values ​​in the time change rate value set, constructs a difference sequence in chronological order, and obtains a rate change difference sequence;

[0023] The abnormality identification submodule calls the change information in the rate change difference sequence, identifies the abnormal amplitude according to the rate change amplitude threshold, marks the corresponding time node and records it, and establishes a recovery cycle rate abnormality record set.

[0024] Optionally, the specific calculation formula for identifying the abnormal amplitude based on the rate change amplitude threshold is:

[0025]

[0026] Where, ΔR i Represents the rate change intensity index at the i-th moment of the rehabilitation cycle, r i,j represents the actual observed value of the rate of the jth call item at the i-th moment in the recovery cycle, represents the arithmetic mean of all call rate observations at time i in the recovery cycle, σ i,j Represents the observed value of the call item rate at the i-th moment in the recovery cycle and the corresponding The standard deviation calculated from the difference between i,j represents the rate credibility weight factor of the jth call item at the i-th time in the recovery cycle, n represents the total number of call items involved in the calculation in the recovery cycle, and ∈ represents a very small positive real number constant set to avoid division by zero errors.

[0027] Optionally, the curvature analysis module includes:

[0028] The trend extraction submodule obtains the abnormal recovery cycle rate record set and the patient continuous indicator distribution table, extracts the time series of urine sodium concentration, lactate dehydrogenase value and blood creatinine level, constructs the trend sequence of indicators in chronological order, and generates the indicator time trend sequence value;

[0029] The curvature calculation submodule calls the indicator data of adjacent time nodes based on the indicator time trend sequence value, analyzes the change pattern within the time period, sequentially constructs the curvature sequence corresponding to the indicator, and generates the indicator curvature change trend value;

[0030] The inflection point identification submodule determines the change direction of adjacent nodes according to the curvature change trend value of the indicator, identifies the curvature mutation position, screens the co-occurrence nodes of the three indicators, and obtains the curvature linkage inflection point set.

[0031] Optionally, the node identification module includes:

[0032] The inflection point extraction submodule obtains the time series corresponding to the curvature linkage inflection point set and the rehabilitation cycle rate abnormality record set based on the two sets, determines whether the data of the two sets exist at the same timestamp, and generates an intersection time node set by matching the timestamp intersection between the sets;

[0033] The parameter linkage determination submodule calls the creatine kinase value and the blood potassium concentration value corresponding to the time point of the intersection time node concentration, determines whether the change direction of the creatine kinase value is the same as the change direction of the blood potassium concentration, and screens the time nodes with consistent change directions through the direction consistency judgment operation to generate a parameter consistency node sequence;

[0034] The risk label generation submodule establishes a joint trend direction label based on the change direction of the creatine kinase value and the change direction of the blood potassium concentration corresponding to each time node in the parameter consistency node sequence, and generates a risk joint evolution node label set by pairing the trend label with the time node set.

[0035] Optionally, the system further comprises:

[0036] The trend construction module calls the risk joint evolution node label set, extracts the myoglobin value and blood creatinine level within the associated time period, calculates the proportion of indicators with consistent change directions, marks the trend range accordingly, and generates a recovery stage fluctuation trend channel map;

[0037] The recovery stage fluctuation trend channel map includes consistent indicator ratios and trend range markings within the associated time period.

[0038] Optionally, the trend building module includes:

[0039] The data extraction submodule obtains the time range annotated in the risk joint evolution node label set, collects the myoglobin value and blood creatinine level within the corresponding time period, aligns the time granularity, calculates the total number of samples and the two types of indicator values ​​at the corresponding time points, and generates a set of myoglobin and blood creatinine alignment indicators;

[0040] The indicator comparison submodule extracts indicator values ​​at consecutive time points based on the myoglobin and blood creatinine alignment indicator set, determines whether the change directions of the two indicators are consistent, calculates the proportion of time periods with consistent directions, and generates an indicator change consistency rate;

[0041] The trend generation submodule marks the time period within the consistent direction interval according to the consistency rate of the indicator changes, extracts the maximum and minimum values ​​of myoglobin and blood creatinine to construct upper and lower boundaries, connects the boundary nodes to construct a channel map, and obtains the recovery stage fluctuation trend channel map.

[0042] Compared with the prior art, the advantages and positive effects of the present invention are:

[0043] In the present invention, by compositely calculating the rate difference sequence of creatine kinase and myoglobin and the curvature sequence of urine sodium and blood creatinine, the limitations of traditional single-parameter static analysis are broken through, and the dynamic capture of the collaborative trend of multiple indicators is achieved. Combined with the consistency judgment of the direction of change of creatine kinase and blood potassium concentration, a joint evolution label system is constructed to accurately locate the risk nodes of multi-system interaction, and the distribution of the directional consistency ratio of myoglobin and blood creatinine in the associated time period is statistically analyzed. The discrete parameter association is converted into a dynamic visualization model. This method solves the problems of misjudgment of data discreteness and the separation of collaborative evolution patterns in traditional technologies, and provides a multi-dimensional dynamic evaluation framework for the clinical process of rhabdomyolysis recovery. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 is a system flow chart of the present invention;

[0045] Figure 2 This is a system block diagram of the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0047] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0048] See also Figure 1 The embodiment of the present invention provides a clinical rehabilitation data management and evaluation system for patients with rhabdomyolysis, including:

[0049] The indicator collection module obtains the creatine kinase value, myoglobin value, urine sodium concentration, blood creatinine level, lactate dehydrogenase value, and blood potassium concentration of patients with rhabdomyolysis at time points during the recovery cycle, organizes the data in chronological order, and generates a patient continuous indicator distribution table;

[0050] The rate detection module extracts the time-varying rates of creatine kinase and myoglobin values ​​based on the patient's continuous indicator distribution table, establishes a difference sequence between adjacent rates, determines the amplitude abnormalities of rate changes, and generates a record set of abnormal rates during the rehabilitation cycle.

[0051] The curvature analysis module extracts the temporal trend changes of urine sodium concentration, lactate dehydrogenase value, and blood creatinine level based on the abnormal recovery cycle rate record set and the patient continuous indicator distribution table. It constructs a curvature sequence for the continuous data, identifies the time points of trend mutation, and generates a set of curvature linkage inflection points.

[0052] The node identification module locates the intersection time node based on the curvature linkage inflection point set and the recovery cycle rate anomaly record set. It then calls the node's creatine kinase value and blood potassium concentration to determine the consistency of the change direction. It then establishes a joint trend label based on the change direction of the indicator to generate a risk joint evolution node label set.

[0053] The trend construction module calls the risk joint evolution node label set, extracts the myoglobin value and blood creatinine level in the associated time period, counts the proportion of indicators with consistent change directions, marks the trend range accordingly, and generates a fluctuation trend channel map for the recovery stage.

[0054] The patient continuous indicator distribution table includes creatine kinase value, myoglobin value, urine sodium concentration, blood creatinine level, lactate dehydrogenase value, and blood potassium concentration. The recovery cycle rate abnormality record set includes creatine kinase rate difference sequence, myoglobin rate difference sequence, and abnormal change amplitude points. The curvature linkage inflection point set includes urine sodium concentration curvature sequence, lactate dehydrogenase curvature sequence, blood creatinine curvature sequence, and trend mutation time points. The risk joint evolution node label set includes joint time nodes, indicator change direction consistency judgment results, and joint trend labels. The recovery stage fluctuation trend channel map includes the proportion of consistent indicators in the associated time period and trend range annotations.

[0055] The indicator collection module includes:

[0056] The data acquisition submodule obtains the creatine kinase value, myoglobin value, urine sodium concentration, blood creatinine level, lactate dehydrogenase value, and blood potassium concentration at time points during the recovery period of patients with rhabdomyolysis. The data are aggregated by patient number and the indicators are assigned to the corresponding time points based on the time records to generate time series data for the six indicators.

[0057] The data acquisition submodule screens patients diagnosed with rhabdomyolysis by connecting to the medical information system, traverses all test records during their recovery period using the patient number as the primary key, and extracts the values ​​of six indicators, including creatine kinase, based on the execution time of the doctor's order. For example, the creatine kinase values ​​of patient number P001 on the first, third, fifth, and seventh days of treatment were 5200 U / L, 3800 U / L, 2800 U / L, and 1800 U / L, respectively, and the myoglobin values ​​were 850 ng / mL, 620 ng / mL, and 450 ng / mL, respectively. , 300ng / mL. If the data for a certain time node is missing, the mean value is filled based on the test results of the adjacent time points. For example, if the myoglobin value of patient P002 on day 3 is missing, the intermediate value of 750ng / mL, which is 900ng / mL on day 1 and 600ng / mL on day 5, is taken as the completion value. Finally, the indicator values ​​of different time nodes are integrated into a structured table file according to the patient number. The table fields contain the patient number, time node and the test values ​​of the six indicators, which are stored as a data set arranged in ascending time order.

[0058] The time arrangement submodule extracts the time nodes of each patient based on the time series data of the six indicators, calls the indicator values ​​corresponding to the time points, arranges the data in chronological order, calculates the interval values ​​between adjacent times, establishes a sequentially arranged data set, and generates a continuous time series structure value;

[0059] The specific calculation formula for arranging data in chronological order is:

[0060]

[0061] Where Δt i represents the weighted time interval between the i-th time point and the i+1-th time point, T i Represents the time value of the i-th time node, T i+1 Represents the time value of the i+1th time node, V ij represents the original value of the i-th time node under the j-th indicator, T j Represents the weighted coefficient corresponding to the jth indicator (note that T j With T i Different, does not represent a time value), m represents the total number of indicators, j is the indicator index from 1 to m, and i is the index of the current time point;

[0062] Detailed explanation of the formula and the process of formula calculation and derivation:

[0063] Parameter description and acquisition method:

[0064] T i Obtained through system logs or sensor records.

[0065] T i+1 Obtained through system logs or sensor records.

[0066] V i,j For example, milliseconds, seconds, etc., are collected through real-time monitoring systems.

[0067] T j It is set based on the importance or influence of the indicator, usually determined through expert evaluation or historical data analysis.

[0068] m represents the number of indicators involved in the calculation.

[0069] Specific numerical settings and basis:

[0070] Set time node T i =120 minutes, T i+1 =150 minutes, indicating that the interval between the two time points is 30 minutes.

[0071] The number of indicators m=3, which means that there are three indicators involved in the calculation.

[0072] Original value of the indicator:

[0073] V i,1 =100 (unit: milliseconds), obtained through system response time monitoring.

[0074] V i,2 =200 (unit: times), obtained through request count.

[0075] V i,3 =300 (unit: milliseconds), obtained through processing time monitoring.

[0076] Weighting coefficient:

[0077] T1=1.0, indicating that the indicator is a benchmark indicator.

[0078] T2=1.5, indicating that the importance of this indicator is 1.5 times that of the benchmark indicator.

[0079] T3=2.0, indicating that the importance of this indicator is twice that of the benchmark indicator.

[0080] Formula calculation process:

[0081] Compute the sum of time values:

[0082] T i+1 +Ti =150+120=270;

[0083] Calculate the molecular part:

[0084]

[0085] Calculate the denominator:

[0086]

[0087] Calculate the final result:

[0088]

[0089] Result interpretation:

[0090] The results show that between the current time node i and the next time node i+1, taking into account the original values ​​of each indicator and its weighting coefficient, the calculated weighted time interval value is approximately 0.27. This value is used to establish a sequentially arranged data set and generate a continuous time series structure value.

[0091] The indicator distribution submodule summarizes the six indicator values ​​at the time nodes based on the continuous time series structure value, forms an indicator arrangement table based on time advancement, calculates the value changes within the time period, establishes the distribution relationship of the indicators over time, and generates a patient continuous indicator distribution table;

[0092] The indicator distribution submodule expands patient P001's six indicators into a longitudinal numerical sequence by time node, marking the test value at each time point and the change range between adjacent time points. For example, the creatine kinase value dropped from 5200 U / L to 3800 U / L, a change range of 26.9%. At the same time, judgment rules are set. If the single drop in creatine kinase exceeds 30% or the myoglobin value is continuously above 500 ng / mL, a warning flag is triggered. The normal reference range for blood potassium concentration is set to 3.5 to 5.0 mmol / L. A value below 3.5 is marked as hypokalemia, and a value above 5.0 is marked as hyperkalemia. For example, if the blood potassium value of patient P006 on day 3 was 5.2 mmol / L, it was classified as hyperkalemia and an abnormal flag was recorded. Finally, a detailed distribution table containing the time node, original value, change range, and warning flag is formed, which is stored in matrix form and associated with the patient number.

[0093] The rate detection module includes:

[0094] The rate extraction submodule obtains the creatine kinase and myoglobin values ​​from the patient continuous indicator distribution table, locates the continuous records of the indicators in chronological order, constructs a rate sequence based on the numerical relationship between adjacent time points, and generates a time-varying rate value set;

[0095] The rate extraction submodule extracts the values ​​of creatine kinase and myoglobin from the distribution data table containing continuous health indicators. The data table must contain fields such as patient number, test time, indicator type, and specific value. All records are arranged in chronological order, and then the same type of indicators are paired according to continuous time points. For example, if a patient underwent creatine kinase testing on March 1, March 2, and March 3, respectively, and the test values ​​were 140, 160, and 180, respectively, then March 1 and March 2 constitute the first pairing, and March 2 and March 3 constitute the second pairing. For each pairing, the value change between the two time points is calculated, and then the two are combined to form the second pairing. The interval length between time points is used to determine the rate of change within that period. For example, if the test is conducted once a day, the rate of change for the first group is 20 units per day, and the rate of change for the second group is also 20 units per day. The above operations are performed step by step during all consecutive test time periods to form an ordered rate sequence. Myoglobin also needs to undergo the same processing steps. To ensure the accuracy of the calculation, abnormal test intervals or missing data must be excluded. For example, if the continuous test interval exceeds three days or the test data deviates abnormally, the data should be cleaned or interpolated before performing the rate calculation. Finally, the two indicators each form a complete set of change rates, which are recorded in chronological order to form a rate data set that can be used for subsequent analysis.

[0096] The difference construction submodule extracts the numerical changes of adjacent rates based on the sequential relationship between the rate values ​​in the time-varying rate value set, constructs a difference sequence in chronological order, and obtains a rate-varying difference sequence;

[0097] After obtaining the rate value set, the difference construction submodule begins to extract the numerical changes between any two adjacent groups of rates and construct a difference set. In this process, the rate sequence is compared one by one in chronological order, and the difference between each two adjacent rates is calculated. For example, if the first rate is 20 and the second rate is 25, the difference is 5. This operation is continued for the second and third rates, and so on. During the construction process, special attention should be paid to whether the time intervals are equidistant. If the detection time intervals between two rates are different, they should be adjusted using a standardized method to ensure that the difference has reference value. For example, an intermediate rate point is inserted in a larger interval through linear interpolation to keep the front and back spacing consistent. Then the difference extraction operation is performed. Each difference is sorted in the corresponding chronological order to form a difference sequence. At the same time, the starting rate and ending rate values ​​corresponding to each difference are recorded to ensure that the original detection change path can be traced back in subsequent modules. After generation, this sequence can be used to identify abnormal change trends or rhythm mutations.

[0098] The anomaly identification submodule calls the change information in the rate change difference sequence, identifies the abnormal amplitude based on the rate change amplitude threshold, marks the corresponding time node and records it, and establishes a recovery cycle rate abnormality record set;

[0099] The specific calculation formula for identifying abnormal amplitude based on the rate change amplitude threshold is:

[0100]

[0101] Where, ΔR i Represents the rate change intensity index at the i-th moment of the rehabilitation cycle, r i,j represents the actual observed value of the rate of the jth call item at the i-th moment in the recovery cycle, represents the arithmetic mean of all call rate observations at time i in the recovery cycle, σ i,j Represents the observed value of the call item rate at the i-th moment in the recovery cycle and the corresponding The standard deviation calculated from the difference between i,j represents the rate credibility weight factor of the jth call item at the i-th moment in the recovery cycle, n represents the total number of call items involved in the calculation in the recovery cycle, and ∈ represents a very small positive real constant set to avoid division by zero errors;

[0102] Detailed explanation of the formula and the process of formula calculation and derivation:

[0103] At the i-th moment of the rehabilitation cycle, the call rate data of n=3 submodules are collected, which are: r i,1 =120 times / min, r i,2 =100 times / min, r i,3 =80 times / minute.

[0104] Calculate average rate

[0105]

[0106] Calculate the standard deviation σ of each submodule i,j , calculated using the Welford iteration method. The collected rate data are: 110, 120, 130.

[0107] Step 1: M1=110, S1=0;

[0108] Step 2:

[0109] Step 3:

[0110] The standard deviation is:

[0111] Similarly, calculate:

[0112] σ i,2 =8,σi,3 =5.

[0113] Set the confidence weighting factor w i,j , set according to the stability and historical performance indicators of the submodule. According to the WRR weight setting method in QoS configuration, the weight value range is 1-10, and the higher the weight, the higher the priority. Setting:

[0114] w i,1 =1.0,w i,2 =0.8, w i,3 =0.6.

[0115] Setting ∈ = 0.01 is used to prevent the denominator from being zero.

[0116] Substitute the above values ​​into the formula to calculate ΔR i :

[0117]

[0118] The results show that the rate variation intensity index at moment i in the recovery cycle is 0.229, indicating a certain degree of fluctuation in the submodule call rate. This index can be used to identify abnormal amplitudes, mark and record corresponding time nodes, and establish a record set of abnormal recovery cycle rates.

[0119] The curvature analysis module includes:

[0120] The trend extraction submodule obtains the abnormal recovery cycle rate record set and the patient continuous indicator distribution table, extracts the time series of urine sodium concentration, lactate dehydrogenase value and blood creatinine level, constructs the trend sequence of indicators in chronological order, and generates the indicator time trend sequence value;

[0121] The trend extraction submodule is used to extract data records of abnormal rates and continuous test indicators of patients from the rehabilitation cycle. First, it is necessary to obtain the patient groups whose rehabilitation rates deviate from the statistical mean in all rehabilitation records through the rehabilitation database. The deviation standard can be set to a range greater than or less than two standard deviations of the group mean. For example, if the overall average gait recovery rate increases by 0.6 levels per week, a case only improves by 0.3 levels within two weeks, which is judged as an abnormal record. Then, the continuous measurement data of urine sodium concentration, lactate dehydrogenase value and blood creatinine level are obtained from the electronic medical records or laboratory information system of such cases, and sorted according to the time nodes to construct a time series. On this basis, The sliding window method is used to smooth the value of each indicator to improve the stability of the data series. For example, the values ​​of five consecutive detection time points are assigned different weights, with the weights of values ​​at more recent times set to larger and those at more distant time points set to smaller. The trend value of the node is finally weighted to generate the trend expression of the indicator at the current time node. For example, if the urine sodium measurement values ​​of a patient are 120, 130, 145, 155 and 160 units for five consecutive days, the trend value on the fifth day is approximately 147 units after applying the weighted average method. All time nodes are processed in this way to form trend sequences for the three types of indicators and obtain complete time trend information.

[0122] The curvature calculation submodule calls the indicator data of adjacent time nodes based on the indicator time trend sequence value, analyzes the change pattern within the time period, sequentially constructs the curvature sequence corresponding to the indicator, and generates the indicator curvature change trend value;

[0123] The curvature calculation submodule is based on the trend sequence, selects the trend values ​​of adjacent time nodes for difference calculation, analyzes the numerical change pattern of the indicator in a continuous time period, and estimates the degree of change curvature of the current node through a comprehensive comparison of the direction and amplitude of the previous and subsequent differences, forming a curvature value sequence. For example, if the urine sodium trend value first rises and then falls for three consecutive days, it indicates that there is a reversal of the change direction at the current node, and the ratio of the two change amplitudes needs to be evaluated. For example, if it rises from 140 to 150 and then falls to 135 units, it can be considered that the curvature of the node is obvious. By processing all time nodes of urine sodium, lactate dehydrogenase and blood creatinine through this method, a curvature change trend data set for each indicator is generated.

[0124] The inflection point identification submodule determines the change direction of adjacent nodes based on the curvature change trend value of the indicator, identifies the location of curvature mutation, screens the co-occurrence nodes of the three indicators, and obtains the curvature linkage inflection point set;

[0125] Based on the curvature trend sequence, the inflection point identification submodule determines the direction of curvature change of adjacent time nodes in turn. If there is a sudden change from negative to positive or from positive to negative, and the amplitude of the change exceeds the set limit, for example, if the threshold is set to 10 units, if the curvature before and after a node is -15 and 12 respectively, it can be marked as a sudden change point. The sudden change points of the three indicators are summarized one by one, and the locations where the curvature sudden changes occur in all three indicators at the same time point are further screened out and marked as co-occurrence nodes. These nodes are aggregated into a curvature linkage inflection point set, which represents the time position where the curvature of the three indicators jointly undergoes abnormal changes.

[0126] The node identification module includes:

[0127] The inflection point extraction submodule obtains the time series corresponding to the curvature linkage inflection point set and the rehabilitation cycle rate anomaly record set, determines whether the data of the two sets exist at the same timestamp, and generates the intersection time node set by matching the timestamp intersection between the sets;

[0128] The inflection point extraction submodule is based on the curvature linkage inflection point set and the rehabilitation cycle rate anomaly record set. First, for the time series data collected from the electromyographic signal, a frequency of one sample per second is selected, and signal points are continuously extracted to analyze their changing trends. The difference method is applied to the values ​​before and after each point to estimate the rate of change. The local curvature of the signal is calculated based on the difference between the positions before and after the point. A judgment threshold of 0.08 is set to extract points with obvious change characteristics in the time series as the curvature inflection point set. Then, the gait speed data is extracted from the wearable device. The speed mean and fluctuation range are calculated by setting three consecutive time windows. Those with larger speed changes are classified into the rate anomaly record set. The fluctuation size is evaluated through normalization processing to determine whether the gait speed at a certain moment has abnormal characteristics. After obtaining the time point sets corresponding to the two sets, a traversal method is used to determine whether there are overlapping time points. All time points that appear in both sets are extracted to form a new intersection time point set for further calculation in subsequent modules.

[0129] The parameter linkage judgment submodule calls the creatine kinase value and blood potassium concentration value corresponding to the time point in the intersection time node concentration to determine whether the change direction of the creatine kinase value is the same as the change direction of the blood potassium concentration. Through the direction consistency judgment operation, the time nodes with consistent change directions are screened to generate a parameter consistency node sequence;

[0130] The parameter linkage judgment submodule calls the creatine kinase value and blood potassium concentration value corresponding to the time points in the intersection time node set. At each intersection time point, the corresponding creatine kinase measurement result and blood potassium test result are extracted. The data value of each time point is compared with the data value of the previous time point. By comparing the increase and decrease direction of the current value with the previous value, it is judged whether the value is increasing, decreasing, or remaining unchanged. For example, at a certain time point, the creatine kinase value increases from 180 to 220, which corresponds to an increase. At the same time, the blood potassium value increases from 4.3 to 4.7, which is also judged to be an increase. Time points with consistent change directions of these two parameters are included in the consistency node sequence. If the change directions are inconsistent, the time point is excluded. Through repeated comparison and screening, a set of all qualified time nodes is obtained.

[0131] The risk label generation submodule establishes a joint trend direction label based on the change direction of the creatine kinase value and the blood potassium concentration corresponding to each time node in the parameter consistency node sequence. By pairing the trend label with the time node set, a risk joint evolution node label set is generated.

[0132] The risk label generation submodule combines the change direction of the creatine kinase value and the change direction of the blood potassium concentration corresponding to each time node in the parameter consistency node sequence, and establishes a unified trend description mark. For example, when both parameters show an upward trend, they are marked as "++", if both show a downward trend, they are marked as "--", and if one rises and the other falls, they are marked as "+-". Each trend mark is paired with its corresponding time point to form a risk joint evolution label set, which represents the change pairing of the two indicators at each time node and can be used for subsequent further trend modeling or early warning classification.

[0133] Trend building blocks include:

[0134] The data extraction submodule obtains the time range annotated by the risk joint evolution node label set, collects the myoglobin values ​​and blood creatinine levels within the corresponding time period, aligns the time granularity, removes missing samples, calculates the total number of samples and the two types of indicator values ​​at the corresponding time points, and generates a set of aligned myoglobin and blood creatinine indicators;

[0135] The data extraction submodule must first read the time range information marked by the risk joint evolution node label set. This label set usually comes from pathological state evolution data or medical model recognition results. The label set records the health risk status of each patient in a specific time period. The time field is usually expressed in the form of "start time" and "end time". First, the time field is uniformly formatted to ensure that there is no error in the subsequent docking time series. When collecting the myoglobin and blood creatinine index values ​​within the time period, by docking with the experimental data platform, the laboratory examination records that meet the time range are extracted, including the examination time and the original values ​​of the two indicators corresponding to each record. They are further organized into time series indicator pairs according to the patient dimension. In the data alignment link, a unified reference time axis is established with a granularity of minutes or hours, and the original data is mapped to the axis. The time window range is set at each time point to match the actual detection time to fill in the missing or unified intervals. For example, the granularity is set per hour. The closest test value is collected based on every hour. After completing the data filling or interpolation operation, only records with both indicator values ​​at the same time point are retained as valid samples for subsequent analysis. In the process of eliminating samples with missing data, a threshold can be set to determine whether the number of valid records for each patient in a certain time period meets the specified standard. For example, it is set that the valid time points must account for more than 80% of the theoretical time points. Samples that do not meet the conditions are removed. Then the total number of remaining samples is counted, and the two indicator values ​​are sorted according to the unified time point and output in a dictionary or table structure. Each time point corresponds to a set of myoglobin and blood creatinine values. For example, from 08:00 on April 1, 2025 to 08:00 on April 2, 2025, a total of 24 hours, after eliminating the missing data, the remaining 20 hours of complete data constitute the alignment indicator set. For example, at 09:00 on April 1, 2025, the corresponding myoglobin value is 87.1 and the blood creatinine value is 135.6.

[0136] The indicator comparison submodule is based on the alignment indicator set of myoglobin and blood creatinine, extracts indicator values ​​at consecutive time points, determines whether the change directions of the two indicators are consistent, calculates the proportion of time periods with consistent directions, and generates the indicator change consistency rate;

[0137] After obtaining the alignment indicator set, the indicator comparison submodule creates change analysis segments based on each pair of consecutive time points. For example, 08:00 and 09:00 constitute a segment. The change trends of the two indicators within each segment are extracted, and the difference between the indicator values ​​at the previous time point and the next time point is compared. For example, if myoglobin is 85.6 at 08:00 and 87.1 at 09:00, it is considered to be increasing. If serum creatinine increases from 134.2 to 135.6 during the same time period, the two change directions are determined to be the same. The number of segments with the same direction in all time periods is counted and then the ratio is calculated with the total number of segments to obtain the proportion of segments with the same change direction. For example, if a total of 23 consecutive time periods are generated and the two indicators change in the same direction in 18 of them, the consistency rate is 18 divided by 23 multiplied by the percentage, which is approximately 78.26%. This ratio is used as the final indicator change consistency rate for subsequent trend channel map construction.

[0138] The trend generation submodule marks the time periods within the consistent direction interval based on the consistency rate of indicator changes, extracts the maximum and minimum values ​​of myoglobin and blood creatinine to construct upper and lower boundaries, connects the boundary nodes to construct a channel map, and obtains the fluctuation trend channel map of the recovery stage;

[0139] After obtaining the consistency rate of indicator changes, the trend generation submodule screens out time periods with higher directional consistency rates and numbers all consecutive consistent time periods. In each time period, all values ​​of myoglobin and blood creatinine are extracted, and the maximum and minimum values ​​in the segment are calculated, which are used as the upper and lower boundaries of the indicator fluctuations in that segment. In the map construction stage, the minimum points in each time period are connected to form a lower boundary curve, and the maximum points form an upper boundary curve. The closed area structure is drawn through visualization tools to form a trend channel map. For example, between 10:00 and 14:00 on April 1, 2025, the minimum value of myoglobin in the interval is 88 and the maximum value is 110. The minimum value of blood creatinine in the same time period is 137 and the maximum value is 151. When drawing, time is used as the horizontal axis and the indicator value is used as the vertical axis. The upper and lower boundary curves are drawn for each time period and the middle area is filled to form a complete recovery stage fluctuation trend map.

[0140] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A clinical rehabilitation data management and evaluation system for patients with rhabdomyolysis, characterized by: The system comprises: The indicator collection module obtains the creatine kinase value, myoglobin value, urine sodium concentration, blood creatinine level, lactate dehydrogenase value, and blood potassium concentration of patients with rhabdomyolysis at time points during the recovery cycle, organizes the data in chronological order, and generates a patient continuous indicator distribution table; A rate detection module extracts the time-varying rates of creatine kinase and myoglobin values ​​based on the patient's continuous indicator distribution table, establishes a difference sequence between adjacent rates, determines abnormal points in the rate variation, and generates a recovery cycle rate abnormality record set; A curvature analysis module, based on the abnormal recovery cycle rate record set and the patient continuous indicator distribution table, extracts the time trend changes of urine sodium concentration, lactate dehydrogenase value and blood creatinine level, constructs a curvature sequence for the continuous data, identifies the trend mutation time point, and generates a curvature linkage inflection point set; The node identification module locates the intersection time node of the curvature linkage inflection point set and the recovery cycle rate abnormality record set, calls the node creatine kinase value and blood potassium concentration, and determines whether (creatine kinase value and blood potassium concentration) can be omitted, and whether the consistency of the change direction can be added. Combined with the indicator change direction, a joint trend label is established to generate a risk joint evolution node label set.

2. The clinical rehabilitation data management and evaluation system for rhabdomyolysis patients according to claim 1, characterized in that: The patient continuous indicator distribution table includes creatine kinase value, myoglobin value, urine sodium concentration, blood creatinine level, lactate dehydrogenase value, and blood potassium concentration. The recovery cycle rate abnormality record set includes creatine kinase rate difference sequence, myoglobin rate difference sequence, and abnormal change amplitude points. The curvature linkage inflection point set includes urine sodium concentration curvature sequence, lactate dehydrogenase curvature sequence, blood creatinine curvature sequence, and trend mutation time point. The risk joint evolution node label set includes joint time node, indicator change direction consistency judgment result, and joint trend label.

3. The clinical rehabilitation data management and evaluation system for rhabdomyolysis patients according to claim 1, characterized in that: The indicator collection module includes: The data acquisition submodule obtains the creatine kinase value, myoglobin value, urine sodium concentration, blood creatinine level, lactate dehydrogenase value, and blood potassium concentration at time points during the recovery period of patients with rhabdomyolysis. The data are aggregated by patient number and the indicators are classified into corresponding time points based on time records to generate time series data for the six indicators. The time arrangement submodule extracts the time nodes of each patient based on the time series data of the six indicators, calls the indicator values ​​corresponding to the time points, arranges the data in chronological order, calculates the interval values ​​between adjacent times, establishes a sequentially arranged data set, and generates a continuous time series structure value; The indicator distribution submodule summarizes the six indicator values ​​at the time nodes according to the continuous time series structure value, forms an indicator arrangement table based on time advancement, calculates the value changes within the time period, establishes the distribution relationship of the indicators over time, and generates a patient continuous indicator distribution table.

4. The clinical rehabilitation data management and evaluation system for rhabdomyolysis patients according to claim 3, characterized in that: The specific calculation formula for arranging data in chronological order is: Where Δt i represents the weighted time interval between the i-th time point and the i+1-th time point, T i Represents the time value of the i-th time node, T i+1 Represents the time value of the i+1th time node, V i,j represents the original value of the i-th time node under the j-th indicator, T j Represents the weighting coefficient corresponding to the j-th indicator, m represents the total number of indicators, j is the indicator index from 1 to m, and i is the index of the current time point.

5. The clinical rehabilitation data management and evaluation system for rhabdomyolysis patients according to claim 3, characterized in that: The rate detection module includes: The rate extraction submodule obtains the creatine kinase value and the myoglobin value in the patient continuous indicator distribution table based on the patient continuous indicator distribution table, locates the continuous records of the indicators in chronological order, constructs a rate sequence based on the numerical relationship between adjacent time points, and generates a time-varying rate value set; A difference construction submodule extracts the numerical changes of adjacent rates based on the sequential relationship between the rate values ​​in the time change rate value set, constructs a difference sequence in chronological order, and obtains a rate change difference sequence; The abnormality identification submodule calls the change information in the rate change difference sequence, identifies the abnormal amplitude according to the rate change amplitude threshold, marks the corresponding time node and records it, and establishes a recovery cycle rate abnormality record set.

6. The clinical rehabilitation data management and evaluation system for rhabdomyolysis patients according to claim 5, characterized in that: The specific calculation formula for identifying abnormal amplitude based on the rate change amplitude threshold is: Where, ΔR i Represents the rate change intensity index at the i-th moment of the rehabilitation cycle, r i,j represents the actual observed value of the rate of the jth call item at the i-th moment in the recovery cycle, represents the arithmetic mean of all call rate observations at time i in the recovery cycle, σ i,j Represents the observed value of the call item rate at the i-th moment in the recovery cycle and the corresponding The standard deviation calculated from the difference between i,j represents the rate credibility weight factor of the jth call item at the i-th time in the recovery cycle, n represents the total number of call items involved in the calculation in the recovery cycle, and ∈ represents a very small positive real number constant set to avoid division by zero errors.

7. The clinical rehabilitation data management and evaluation system for rhabdomyolysis patients according to claim 5, characterized in that: The curvature analysis module includes: A trend extraction submodule obtains the abnormal recovery cycle rate record set and the patient continuous indicator distribution table, extracts the time series of urine sodium concentration, lactate dehydrogenase value and blood creatinine level, constructs the trend sequence of indicators in chronological order, and generates the indicator time trend sequence value; The curvature calculation submodule, based on the indicator time trend sequence value, calls the indicator data of adjacent time nodes, analyzes the change pattern within the time period, sequentially constructs the curvature sequence corresponding to the indicator, and generates the indicator curvature change trend value; The inflection point identification submodule determines the change direction of adjacent nodes according to the curvature change trend value of the indicator, identifies the curvature mutation position, screens the co-occurrence nodes of the three indicators, and obtains the curvature linkage inflection point set.

8. The clinical rehabilitation data management and evaluation system for rhabdomyolysis patients according to claim 1 or 7, characterized in that: The node identification module includes: The inflection point extraction submodule obtains the time series corresponding to the curvature linkage inflection point set and the rehabilitation cycle rate abnormality record set based on the two sets, determines whether the data of the two sets exist at the same timestamp, and generates an intersection time node set by matching the timestamp intersection between the sets; The parameter linkage determination submodule calls the creatine kinase value and the blood potassium concentration value corresponding to the time point of the intersection time node concentration, determines whether the change direction of the creatine kinase value is the same as the change direction of the blood potassium concentration, and screens the time nodes with consistent change directions through the direction consistency judgment operation to generate a parameter consistency node sequence; The risk label generation submodule establishes a joint trend direction label based on the change direction of the creatine kinase value and the change direction of the blood potassium concentration corresponding to each time node in the parameter consistency node sequence, and generates a risk joint evolution node label set by pairing the trend label with the time node set.

9. The clinical rehabilitation data management and evaluation system for rhabdomyolysis patients according to claim 1, characterized in that: The system further comprises: A trend construction module calls the risk joint evolution node label set, extracts the myoglobin value and blood creatinine level within the associated time period, calculates the proportion of indicators with consistent change directions, marks the corresponding trend range, and generates a recovery stage fluctuation trend channel map; The recovery stage fluctuation trend channel map includes consistent indicator ratios and trend range markings within the associated time period.

10. The clinical rehabilitation data management and evaluation system for rhabdomyolysis patients according to claim 9, characterized in that: The trend building module includes: The data extraction submodule obtains the time range annotated in the risk joint evolution node label set, collects the myoglobin value and blood creatinine level within the corresponding time period, aligns the time granularity, calculates the total number of samples and the two types of indicator values ​​at the corresponding time points, and generates a set of aligned myoglobin and blood creatinine indicators; The indicator comparison submodule extracts indicator values ​​at consecutive time points based on the myoglobin and blood creatinine alignment indicator set, determines whether the change directions of the two indicators are consistent, and calculates the proportion of time periods with consistent directions (myoglobin values ​​and blood creatinine levels), which can be omitted or added, to generate the indicator change consistency rate; The trend generation submodule marks the time periods within the consistent direction interval according to the consistency rate of the indicator changes, extracts the maximum and minimum values ​​of myoglobin and blood creatinine to construct upper and lower boundaries, connects the boundary nodes to construct a channel map, and obtains the fluctuation trend channel map of the recovery stage.

Images