Intelligent rheumatology and immunology patient personalized care path management system

Abstract

The present application relates to the technical field of rheumatology and immunology nursing, in particular to an intelligent rheumatology and immunology patient personalized nursing path management system, which comprises a joint activity-inflammation data acquisition unit, a path generation and dose binding unit, a cross-modal verification feedback unit and a closed-loop priority regulation unit. The present application fuses and acquires multi-modal data such as joint angle and inflammatory factors through a nine-axis inertial sensor and biomarker detection, identifies the morning stiffness period through a time series segmentation model, dynamically adjusts the prednisone dose in combination with a hormone dose response rule base, and verifies the effectiveness of rehabilitation movements based on a dynamic time warping algorithm. When the medicine taking time deviation is more than 1 hour or the movement execution is abnormal, the sensor enhanced collection and dose increment mechanism is triggered, and at the same time, the nursing instruction priority is optimized through a conflict arbitration mechanism, realizing the whole process intelligentization of "data acquisition-path generation-verification feedback-priority regulation" and providing precise nursing solutions for rheumatology and immunology patients.

Description

Technical Field

[0001] The present invention relates to the field of rheumatology and immunology nursing technology, and in particular to an intelligent rheumatology and immunology patient personalized nursing pathway management system. Background Art

[0002] Rheumatology and immunology nursing is an important technology. In the clinical care of rheumatic and immunology diseases, accurate monitoring of joint movement function, dynamic adjustment of drug dosage and personalized rehabilitation training pathways are crucial to controlling the progression of inflammation and improving patients' quality of life.

[0003] With the development of wearable sensors and bioassay technologies, the real-time collection of nursing data has significantly improved. However, traditional nursing systems, in practical applications, suffer from a core problem: insufficient dynamic correlation analysis capabilities. Existing solutions often rely on single-dimensional data, ignoring the complex coupling between joint range of motion, inflammatory factor levels, and drug metabolism dynamics. When a patient's morning stiffness duration fluctuates or inflammatory factors suddenly increase, traditional systems are unable to quickly identify anomalies based on multi-source data. The root cause lies in the lack of modeling capabilities for the dynamic chain of "inflammatory load-joint function-drug response." Drug dosage adjustments are often based on fixed thresholds, failing to fully consider the impact of individual physiological characteristics such as weight and liver function on drug metabolism. Furthermore, rehabilitation training pathways often fail to precisely align the patient's morning stiffness period with peak drug blood concentrations, significantly reducing training effectiveness due to inappropriate timing. This lag in data correlation and model construction reduces treatment synergy, and nursing plans are slow to respond to changes in individual physiological status, making them difficult to adapt to the characteristics of rheumatic and immunological diseases, which have a fluctuating course and significant individual differences. As a result, traditional nursing models lack dynamism and individualization, resulting in a waste of nursing resources. To address this technical issue, we have developed an intelligent personalized nursing pathway management system for rheumatology and immunology patients. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent personalized nursing pathway management system for rheumatology and immunology patients to solve the problems raised in the above background technology.

[0005] 1. Because traditional systems rely on single-dimensional data and lack the ability to dynamically correlate and analyze multi-source data, they are unable to quickly identify abnormal patient conditions. Therefore, this case uses a joint activity-inflammation data acquisition unit to collect multimodal data, calculate dynamic activity scores and inflammatory load indexes, and can accurately assess patient conditions based on multidimensional data, improving abnormality identification capabilities.

[0006] 2、Due to the traditional system drug dose adjustment is based on fixed threshold, individual physiological characteristics are not considered, and the rehabilitation training path is not properly formulated, therefore, the case combines the path generation and dose binding unit, time series segmentation model and hormone dose response rule base to generate personalized rehabilitation training interval and dynamically adjust drug dose, which can accurately match individual needs and improve treatment synergy and effect.

[0007] To achieve the above purpose, an intelligent rheumatology and immunology patient personalized nursing path management system is provided, which comprises the following units:

[0008] The joint activity-inflammation data acquisition unit acquires a three-dimensional joint angle change sequence through a nine-axis inertial sensor, outputs a dynamic activity score after Kalman filtering, and detects interleukin-6 and tumor necrosis factor- , combined with the current prednisone daily dose to calculate the inflammation load index;

[0009] The path generation and dose binding unit inputs the dynamic activity score into the time series segmentation model, identifies the morning stiffness starting time window, and generates the next day's rehabilitation training recommended interval, matches the preset hormone dose response rule base according to the 24-hour fluctuation curve of the inflammation load index, and triggers the prednisone daily dose gradient adjustment instruction when the inflammation load index exceeds the threshold for 6 consecutive hours;

[0010] The cross-modal verification feedback unit obtains the actual dispensing time through the intelligent medicine box, and when the deviation from the prednisone daily dose gradient adjustment instruction exceeds hours, the joint sensor enhanced acquisition mode is activated, and the dynamic activity score is compared with the dynamic time warping distance of the standard action template in real time during the rehabilitation training period. If the distance exceeds the threshold and the inflammation load index rises by or more simultaneously, a hormone dose increment signal is sent to the path generation and dose binding unit.

[0011] As a further improvement of the technical solution, the specific implementation method of the time series segmentation model in the path generation and dose binding unit comprises:

[0012] A morning stiffness feature extractor is constructed based on a time convolution network, and the dilation coefficient is set to , wherein n is the network layer number, and k is the convolution kernel width, which is 3;

[0013] The dynamic activity score sequence is phase segmented, and when the standard deviation of the joint angle in the adjacent 30-minute window decreases by 40% or more and maintains for more than 2 hours, it is determined as the morning stiffness starting window;

[0014] The formula for generating the next day's rehabilitation training recommended interval is: ; wherein, is the predicted value of the onset time of morning stiffness, , Buffer time for experience, This is the recommended interval for rehabilitation training the next day.

[0015] As a further improvement of this technical solution, the method for constructing the hormone dose response rule library includes:

[0016] Establish the calculation formula for dynamic threshold: ;in, For the moment The dynamic threshold of is the patient's baseline inflammatory load index, is the inflammatory load index, The inflammatory load index is expressed as any time within 24 hours. ;in, is the current daily dose of prednisone, is the dose-influencing factor, is the half-life decay coefficient of prednisone, The time since the last dose of the medicine, is interleukin-6, Tumor necrosis factor- , used to reflect the immediate inflammatory state and the effect of drug intervention. The patient's baseline inflammatory load index is obtained based on the patient's individual historical baseline value and is used to reflect the basic inflammatory level of the individual's physiological characteristics;

[0017] The generation rule of gradient adjustment instructions is:

[0018] when More than 6 consecutive hours When the patient is diagnosed with leukemia, the prednisone dose is increased by 5 mg every 24 hours until the cumulative increase reaches 15 mg;

[0019] when Less than 0.8 for 12 consecutive hours The prednisone dose was reduced by 2.5 mg every 24 hours, and the sampling rate of the triggered joint sensor was increased to 15 Hz after the dose reduction.

[0020] As a further improvement of this technical solution, the path generation and dose binding unit also includes a multi-constraint dose optimization mechanism:

[0021] a. Before the gradient adjustment instruction takes effect, verify the following constraints:

[0022] b. The cumulative daily dose of prednisone is less than or equal to 1 mg / kg, calculated based on the patient's body weight;

[0023] c. The fluctuation of alanine aminotransferase, a liver function indicator, within the past 7 days is less than 30%;

[0024] If any of a, b, and c is not satisfied, the alternative decision process is initiated:

[0025] The 5-mg prednisone increment was replaced with an equivalent intravenous interleukin-6 inhibitor regimen; and the rehabilitation training period was adjusted to 2 hours after the peak blood drug concentration.

[0026] As a further improvement of this technical solution, the dynamic time warping distance calculation method in the cross-modal verification feedback unit includes:

[0027] Construct a joint angle change template library for standard rehabilitation movements. Each template contains a 15-keyframe 3D Euler angle sequence, which is used to describe a set of angle values ​​for the rigid body posture in 3D space.

[0028] The formula for calculating the dynamic time warping distance between the patient's motion and the template in real time is:

[0029] ;in, is the path penalty coefficient, is the final dynamic time warping distance, is the optimal alignment path, To align point pairs, is the patient's joint angle vector, is the template joint angle vector, is the Euclidean distance;

[0030] when When the dynamic time warping distance threshold is greater than the inflammatory load index rising rate is greater than or equal to 2pg / ml / h, it is determined that the action execution is abnormal. The dynamic time warping distance threshold is used to determine the critical value of the degree of difference between the patient's rehabilitation action and the standard template.

[0031] As a further improvement of the present technical solution, the activation logic of the joint sensor enhanced acquisition mode includes:

[0032] When the medication collection time deviation exceeds 1 hour, the sampling rate of the nine-axis inertial sensor is increased from 10Hz to 20Hz, the electromyographic signal acquisition module around the joint is activated, and the muscle activation delay time is synchronously analyzed. In enhanced mode, the joint stability index is generated every 5 minutes. When the joint stability index is less than 0.7, a training period compression instruction is sent to the path generation and dose binding unit, reducing the single training time by 25%;

[0033] Among them, the method for generating the joint stability index is as follows:

[0034] The joint stability index is constructed by fusing nine-axis sensor data and electromyographic signals. The nine-axis sensor data includes angular velocity variance. And acceleration fluctuation coefficient , the electromyogram includes muscle activation synchrony , the formula for calculating the joint stability index is: ; wherein, = actual synergistic contraction index / ideal synergistic contraction index, used to measure the coordination of muscle activation around the joint, the actual synergistic contraction index is obtained by calculating the activation time difference and intensity ratio of the antagonistic muscle through the surface electromyogram, and the ideal synergistic contraction index is obtained based on the benchmark value established by the healthy population data.

[0035] As a further improvement of the technical solution, the triggering mechanism of the hormone dose increment signal includes the following steps:

[0036] Establish the mapping relationship between the increment signal intensity and the multi-modal data:

[0037] When is in the interval , and the synchronous rise rate of the inflammation load index is greater than or equal to 15%, a low-intensity increment signal is triggered, and the low-intensity increment signal is +2.5mg / 6h.

[0038] When , and the synchronous rise rate of the inflammation load index is greater than or equal to 25%, a high-intensity increment signal is triggered, and the high-intensity increment signal is +5mg / 6h.

[0039] During the validity period of the increment signal, the real-time exercise monitoring mode of the patient's APP is forcibly opened:

[0040] If the joint range of motion is detected to drop by more than 30%, immediately suspend the dose adjustment and start the emergency video call link;

[0041] If the joint range of motion returns to more than 90% of the baseline value, resume the increment process and shorten the monitoring interval to every 15 minutes.

[0042] As a further improvement of the technical solution, it also includes a closed-loop priority control unit, which is as follows:

[0043] Establish a nursing instruction conflict arbitration mechanism. When the hormone dose adjustment and the rehabilitation training period conflict in time, calculate the expected inflammation suppression efficiency of the hormone adjustment and the joint function improvement efficiency of the rehabilitation training, respectively.

[0044] If the expected inflammation suppression efficiency of the hormone adjustment is more than 1.2 times the joint function improvement efficiency of the rehabilitation training, delay the training and perform the dose increment, and generate a compensatory training plan at the same time.​​

[0045] If the joint function improvement efficiency of rehabilitation training is greater than or equal to the expected inflammation inhibition efficiency of hormone adjustment, the original 60-minute training will be divided into two 25-minute modules and one 10-minute module, which will be placed 0.5 hours before and 1.5 hours after the dose adjustment respectively. The two 25-minute modules are the main training of 50 minutes, and the 10-minute module is a buffer time of 10 minutes for preparation and relaxation before and after training.

[0046] As a further improvement of this technical solution, the calculation formulas for the expected inflammation suppression efficiency of hormone adjustment and the joint function improvement efficiency of rehabilitation training are as follows:

[0047] Expected inflammation suppression efficiency of the hormone regulation:

[0048] ;in, is the difference between the current inflammatory load index and the average of the previous 6 hours, is the predicted value of prednisone blood concentration, calculated based on the pharmacokinetic model;

[0049] The efficiency of the rehabilitation training in improving joint function:

[0050] ;in, The rate of change of dynamic activity score before and after training, The coordination of muscle activation is calculated by the phase synchronization of electromyographic signals.

[0051] As a further improvement of this technical solution, the closed-loop priority control unit updates the parameters of the time series segmentation model every 30 days, as follows:

[0052] The expansion coefficient of the time series segmentation model is refitted based on historical data, and the convolution kernel width is adjusted using the Bayesian optimization algorithm;

[0053] The threshold of the hormone dose response rule library is modified according to the individual pharmacokinetic characteristics of the patient. The modification formula is:

[0054] ;in, It is a new generation of inflammatory load threshold. is the current cycle threshold, For the Peak daily inflammatory load, and The weights are assigned based on historical data.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] In the intelligent personalized nursing pathway management system for rheumatology and immunology patients, the pathway generation and dose binding unit uses a time series segmentation model to accurately identify the morning stiffness period, and combines it with a hormone dose response rule library to dynamically adjust the prednisone dose, thereby improving the timing matching of rehabilitation training and shortening the lag time for drug dose adjustment. The cross-modal verification feedback unit triggers sensor enhanced acquisition and dose increment signals through medication collection time monitoring and dynamic time regularization algorithm, shortening the response time for abnormal joint stability warning, improving the accuracy of identifying abnormal movement execution, and enhancing the dynamic adaptability and treatment effect of rheumatic and immunological disease care. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Fig. 1 It is a workflow diagram of the present invention;

[0058] Fig. 2 It is an overall block diagram of the present invention.

[0059] The meaning of each number in the figure is:

[0060] 1. Joint activity-inflammation data acquisition unit; 2. Path generation and dose binding unit; 3. Cross-modal verification feedback unit; 4. Closed-loop priority control unit. DETAILED DESCRIPTION

[0061] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0062] Patients with rheumatic and immunological diseases often face the dual contradiction of "inflammation control" and "joint function protection." Existing nursing systems suffer from three major pain points: static threshold rigidity, multimodal data fragmentation, and the lack of arbitration for treatment conflicts. This invention, through a three-dimensional closed loop of biomechanics, biochemical indicators, and medication behavior, achieves morning stiffness time series prediction based on a fully convolutional neural network, dynamic threshold-driven hormone dosage binding for inflammatory load, and intelligent priority arbitration under cross-modal verification. This provides an intelligent personalized nursing pathway management system for rheumatology and immunology patients. Please refer to [the "Career Management System"]. Figs. 1-2 As shown, it includes the following units:

[0063] The joint activity-inflammation data acquisition unit 1 collects the three-dimensional joint angle change sequence through the nine-axis inertial sensor and outputs the dynamic activity score after Kalman filtering. The nine-axis sensor can fully capture the three-dimensional movement of the joint. The Kalman filter can effectively suppress noise and improve the accuracy of the activity score. It also detects interleukin-6 and tumor necrosis factor- , combined with the current prednisone daily dose to calculate the inflammation load index, interleukin-6 and tumor necrosis factor- is a key inflammatory factor, combined with the dose to quantify the inflammation load, using electrochemiluminescence immunoassay, through subcutaneous microdialysis probe continuous collection of interstitial fluid, the detection cycle is 1 hour, the temperature compensation module is built in the chip, eliminate the influence of body temperature fluctuation on the test result, the inflammation load index ; wherein, is the current prednisone daily dose, is the dose impact factor, is the prednisone half-life attenuation coefficient, is the time after the last medication, is interleukin-6, is tumor necrosis factor- , for reflecting the immediate inflammatory state and the effect of drug intervention, the patient baseline inflammation load index is obtained according to the individual historical baseline value of the patient, for reflecting the basic inflammation level of individual physiological characteristics.

[0064] The path generation and dose binding unit 2 inputs the dynamic activity score into the time series segmentation model to identify the morning stiffness starting time window, and the specific implementation method of the time series segmentation model in the path generation and dose binding unit 2 includes:

[0065] The morning stiffness feature extractor is constructed based on the time convolution network, and the inflation coefficient is set to , is the network layer number, , the convolution kernel width is 3, the activity trend of different time scales is captured, the phase segmentation is carried out on the dynamic activity score sequence, the window size is 30 minutes, the step is 5 minutes, the standard deviation change rate of joint angle of adjacent windows is calculated, when the standard deviation decreases by more than or equal to 40% and the continuous time length is more than 2 hours, it is marked as the morning stiffness starting window, and the next day rehabilitation training recommended interval is generated, and the formula for generating the next day rehabilitation training recommended interval is: ; wherein, is the morning stiffness starting time prediction value, , is the experience buffer length, is the next day rehabilitation training recommended interval, according to the 24-hour fluctuation curve of the inflammation load index, the pre-set hormone dose response rule library is matched, the hormone dose response rule library includes the dose adjustment threshold corresponding to different inflammation load levels, the threshold is personalized according to the patient's weight and disease severity, when the inflammation load index exceeds the threshold for 6 hours continuously, the mean value of the inflammation load index is calculated, compared with the threshold of the rule library in real time, the prednisone daily dose gradient adjustment instruction is triggered, the continuous inflammation load exceeding the standard needs to be intervened in time, and the gradient adjustment avoids the side effects caused by dose mutation.

[0066] The method for constructing a hormone dose response rule library includes:

[0067] Establish the calculation formula for dynamic threshold: ;in, For the moment The dynamic threshold of is the patient's baseline inflammatory load index, is the inflammatory load index, Indicates any time within 24 hours;

[0068] The rules for generating the daily dose gradient adjustment instructions for prednisone are:

[0069] when More than 6 consecutive hours When the patient is diagnosed with leukemia, the prednisone dose is increased by 5 mg every 24 hours until the cumulative increase reaches 15 mg;

[0070] when Less than 0.8 for 12 consecutive hours When the dose of prednisone was reduced by 2.5 mg every 24 hours, and the sampling rate of the joint sensor was increased to 15 Hz after the dose reduction, the rationality of the dose adjustment was improved through multi-constraint verification, while ensuring the effect of inflammation control.

[0071] Among them, the multi-constraint verification improvement also includes a multi-constraint dose optimization mechanism:

[0072] a. Before the gradient adjustment instruction takes effect, verify the following constraints:

[0073] b. The cumulative daily dose of prednisone should be less than or equal to 1 mg / kg. Based on the patient's weight, a dose exceeding 1 mg / kg increases the risk of liver injury. The cumulative dose should be dynamically limited based on body weight, as follows:

[0074] Obtain patient weight data in real time and calculate the maximum allowable dose. ; Record the total daily dose of prednisone in history. When the cumulative dose after this adjustment is greater than When the patient is in a coma, a restraint alarm is triggered to avoid hormonal side effects caused by drug overdose.

[0075] c. The fluctuation of alanine aminotransferase (ALT) in the liver function indicator should be less than 30% in the past 7 days. When liver function is abnormal, hormone metabolism capacity is reduced, and transaminase fluctuations should be monitored to assess liver tolerance, as follows:

[0076] Blood samples were collected from patients twice a week to test the alanine aminotransferase value and the fluctuation range in the past 7 days was calculated. ;in, It is the fluctuation range of alanine aminotransferase value in the past 7 days, which is used to evaluate the stability of liver function. The patient's current (most recent) alanine aminotransferase test value reflects the immediate liver function status. The patient's alanine aminotransferase baseline value for the past 7 days is taken as the average value or stable period test value during this period. When the liver function is judged to be unstable, the increase of hormone dose is prohibited, and the risk of liver damage is identified in advance.

[0077] If any of a, b, and c is not met, the alternative plan decision-making process is initiated. When hormonal adjustment is limited, it is necessary to switch to biologics to maintain inflammatory control. At the same time, the training timing is optimized. According to the drug equivalence conversion table (e.g., 1 mg prednisone ≈ 0.4 mg interleukin-6 inhibitor), the alternative dose is calculated and sent to the medical order system. The peak time of the interleukin-6 inhibitor blood concentration is predicted by the population pharmacokinetic model. The 5 mg prednisone increment is replaced with an equivalent interleukin-6 inhibitor intravenous injection regimen, and the rehabilitation training period is adjusted to 2 hours after the peak blood concentration to avoid physical fatigue in the early stage of the drug taking effect affecting the training effect and reduce training interruptions due to drug side effects.

[0078] The cross-modal verification feedback unit 3 obtains the actual medication collection time through the smart medicine box. When the deviation from the prednisone daily dose gradient adjustment instruction exceeds Activate the joint sensor enhanced acquisition mode every hour, integrating accelerometer, gyroscope, and magnetometer data. Improve posture calculation accuracy through the Kalman filter algorithm and extend the single data acquisition window from 10 minutes to 30 minutes, covering the critical period for drug effectiveness. During rehabilitation training, the activation logic of the joint sensor enhanced acquisition mode includes:

[0079] When the deviation in medication collection time is greater than 1 hour, the deviation in medication collection time may cause fluctuations in blood drug concentration. Enhanced monitoring is required to capture changes in joint function. The sampling rate of the nine-axis inertial sensor is increased from 10Hz to 20Hz to enhance the ability to capture subtle joint movements. The electromyographic signal acquisition module around the joint is activated, and a bandpass filter (20-450Hz) is used to extract electromyographic activity features and calculate the muscle activation delay time. The muscle activation delay time is the time difference from the start of the movement to the electromyographic signal exceeding the threshold, which is used to identify early muscle dysfunction. In enhanced mode, a joint stability index is generated every 5 minutes. The control ability of the joint in movement needs to be quantitatively evaluated to provide a basis for training adjustments. The nine-axis sensor data and electromyographic signals are integrated to construct the joint stability index. The nine-axis sensor data includes angular velocity variance and acceleration fluctuation coefficient , EMG signals include muscle activation synchronization , the calculation formula of joint stability index is: ;in, = Actual co-contraction index / ideal co-contraction index, used to measure the coordination of activation of muscle groups around joints. The actual co-contraction index is obtained by calculating the activation time difference and strength ratio of antagonistic muscles through surface electromyographic signals. The ideal co-contraction index is obtained based on the baseline value established based on healthy population data. The current window data is integrated every 5 minutes, and the stability index is generated through edge computing. When the joint stability index is less than 0.7, the training period compression logic is triggered, and the new training duration is automatically calculated. The training period compression instruction is sent to the path generation and dose binding unit 2, and the single duration is reduced by 25%.

[0080] Real-time comparison of the dynamic activity score and the dynamic time regularization distance of the standard action template. If the distance exceeds the threshold and the inflammatory load index rises simultaneously , a hormone dose increment signal is sent to the path generation and dose binding unit 2. The dynamic time warping distance calculation method includes:

[0081] A joint angle change template library for standard rehabilitation movements was constructed. Rehabilitation physicians performed standard movements wearing nine-axis inertial sensors, collecting data from more than 100 repetitive movements. A clustering algorithm was used to segment the continuous movement sequence, extracting a 15-keyframe 3D Euler angle sequence. The 3D Euler angles for each keyframe were averaged to construct a standard template matrix consisting of 15 keyframes × 3 angles. This matrix is ​​used to describe a set of angle values ​​for the rigid body's posture in 3D space. The 3D Euler angles include pitch, yaw, and roll angles, providing a reliable benchmark for movement evaluation.

[0082] The patient's movements may have different time scales from the standard template, so a dynamic time warping algorithm is needed to achieve elastic alignment. The formula for calculating the dynamic time warping distance between the patient's movements and the template in real time is:

[0083] ;in, is the path penalty coefficient, which is used to control the penalty intensity of time axis expansion. The final dynamic time warping distance measures the degree of difference between the patient's movement and the standard template. The larger the value, the more serious the deviation. is the optimal alignment path, which represents the optimal time correspondence between the patient action sequence and the template sequence. To align the point pairs, the first Frame and template sequence Matching points of the frame, is the patient's joint angle vector, i.e. The 3D Euler angles of the patient's joints at the frame time, is the template joint angle vector, i.e. The preset Euler angles of the standard rehabilitation action. The Euclidean distance is used to calculate the instantaneous difference between the patient and the template in the joint angle space. The threshold is determined by the test data of healthy people. A sliding window with a window size of 2 hours and a step length of 30 minutes is used to calculate the change rate of the inflammatory load index. ;in, Represents the current moment The detected interleukin-6 concentration value represents the level of inflammation in real time. For the current moment The interleukin-6 concentration value 2 hours ago is used as the interleukin-6 concentration value detected by the historical reference baseline. When the dynamic time warping distance threshold is greater than the dynamic time warping distance threshold and the rising rate of the inflammatory load index is greater than or equal to 2pg / ml / h, it is judged as abnormal action execution. The dynamic time warping distance threshold is used to determine the critical value of the degree of difference between the patient's rehabilitation action and the standard template, thereby improving the recognition accuracy of the dynamic time warping algorithm for rehabilitation actions.

[0084] The triggering mechanism for the hormone dose increment signal includes the following steps:

[0085] Establish a mapping relationship between incremental signal intensity and multimodal data. The rate of increase in inflammatory load and joint function status need to be evaluated in a coordinated manner to accurately match the dose adjustment intensity:

[0086] Calculate the rising rate of inflammatory load and use exponential smoothing method to calculate the dynamic rising rate of inflammatory load index. ;in, is the dynamic rising rate of the inflammatory load index, for Inflammatory load index at each moment, for Push forward the inflammatory load index 6 hours in real time and compare the current rate with the preset threshold. In the range When the synchronous increase rate of the inflammatory load index is greater than or equal to 15%, a low-intensity incremental signal is triggered. is a low-intensity incremental threshold, used to trigger the inflammatory load range of a low-dose increment, wherein the low-intensity incremental signal is +2.5 mg / 6 h;

[0087] when When the synchronous rising rate of the inflammatory load index is greater than or equal to 25%, a high-intensity incremental signal is triggered, and the high-intensity incremental signal is +5mg / 6h. is the high-intensity increment threshold, which is the critical value of inflammatory load used to trigger high-dose increments;

[0088] During the period when the incremental signal is in effect, the real-time motion monitoring mode of the patient-side APP is forced to be turned on. During the dose adjustment period, changes in joint function must be closely monitored to prevent acute injuries caused by drug side effects or increased inflammation:

[0089] The joint sensor sampling rate is temporarily increased to 50Hz, and the joint range of motion is collected every 2 minutes to calculate the difference from the baseline value. If a sudden drop in joint range of motion of more than 30% is detected, the dose adjustment is immediately suspended and an emergency video call link is initiated. If the joint range of motion recovers to more than 90% of the baseline value, the incremental process is resumed and the monitoring interval is shortened to once every 15 minutes, shortening the average time for adjustment and recovery, and improving treatment safety and clinical efficacy.

[0090] Closed-loop priority control unit 4 is as follows:

[0091] Hormone adjustment and rehabilitation training may conflict in time. The expected benefits of both interventions need to be quantitatively assessed to determine priority. Hormone adjustment instructions and rehabilitation training plans should be time-aligned. Conflict is identified when the following conditions are met: Hormone adjustment instructions include execution time and duration, while rehabilitation training plans include start and end times:

[0092] ;in is the execution time, is the start time, For duration, For the end time, a nursing instruction conflict arbitration mechanism is established. When the hormone dose adjustment conflicts with the rehabilitation training period, the expected inflammation suppression efficiency of hormone adjustment and the joint function improvement efficiency of rehabilitation training are calculated respectively. The calculation formulas for the expected inflammation suppression efficiency of hormone adjustment and the joint function improvement efficiency of rehabilitation training are as follows:

[0093] Expected Inflammation Suppression Efficiency of Hormonal Adjustment:

[0094] ;in, is the difference between the current inflammatory load index and the average of the previous 6 hours, is the predicted value of prednisone blood concentration, calculated based on the pharmacokinetic model;

[0095] Efficiency of rehabilitation training in improving joint function:

[0096] ;in, The rate of change of dynamic activity score before and after training, The muscle activation coordination is calculated by the phase synchronization of the electromyographic signal. The muscle activation coordination is obtained as follows:

[0097] First, multi-channel EMG signals are collected using the peri-articular EMG signal acquisition module. After preprocessing through bandpass filtering and Kalman filtering, the preprocessed signals are subjected to Hilbert transform to obtain the instantaneous phase. The phase difference between adjacent channels is calculated, and the phase synchronization of a single pair of muscles is quantified by the phase locking value. Then, weights are set based on the proportion of muscle physiological function, and the weighted average of multiple pairs of phase locking values ​​is performed to obtain the muscle activation coordination degree.

[0098] If the expected inflammation suppression efficiency of hormone adjustment is greater than 1.2 times the joint function improvement efficiency of rehabilitation training, then the training will be postponed and the dose will be increased. Rehabilitation training will be postponed until 2 hours after the end of hormone adjustment. At the same time, a compensatory training program will be generated, and 15 minutes of joint flexibility training will be added.

[0099] If the joint function improvement efficiency of rehabilitation training is greater than or equal to the expected inflammation inhibition efficiency of hormone adjustment, the original 60-minute training will be divided into two 25-minute modules and one 10-minute module, which will be placed 0.5 hours before and 1.5 hours after the dose adjustment respectively. The two 25-minute modules are 50 minutes of main training, and the 10-minute module is a buffer time of 10 minutes for preparation and relaxation before and after training. The time window for efficiency calculation is dynamically adjusted according to the patient's current inflammatory state to avoid loss of treatment effect due to improper sequence. Postponing training may affect the progress of rehabilitation, and an equivalent compensation plan needs to be designed to maintain treatment continuity. Based on the patient's current joint function status, targeted training movements are selected from the preset template library, and the compensation coefficient is calculated according to the delay time. ; In order to extend the duration and adjust the training intensity, the compensatory training should be spread over 24 hours, 10-15 minutes each time, to avoid fatigue caused by concentrated training.

[0100] The closed-loop priority control unit 4 updates the parameters of the time series segmentation model every 30 days, as follows:

[0101] The expansion coefficient of the time series segmentation model is refitted based on historical data, and the convolution kernel width is adjusted using the Bayesian optimization algorithm;

[0102] The threshold of the hormone dose response rule library is modified according to the individual pharmacokinetic characteristics of the patient. The modification formula is:

[0103] ;in, It is a new generation of inflammatory load threshold. is the current cycle threshold, For the Peak daily inflammatory load, and It is the weight assigned based on historical data. The new generation inflammatory load threshold refers to the inflammatory load threshold used for the next cycle after correction by historical data. It is an updated value based on the current cycle threshold and the integration of the inflammatory load peak data of the past 30 days. The current cycle threshold refers to the critical value of the inflammatory load used for hormone dose adjustment decisions within the current 30-day cycle, which is used to determine whether the inflammatory load index triggers dose adjustment.

[0104] The present invention integrates nine-axis inertial sensors with biomarker detection to collect multimodal data such as joint angles and inflammatory factors, identifies the morning stiffness period through a time series segmentation model, dynamically adjusts the prednisone dose in combination with a hormone dose response rule library, and verifies the effectiveness of rehabilitation movements based on a dynamic time warping algorithm. When the medication collection time deviation exceeds 1 hour or the movement execution is abnormal, the sensor-enhanced collection and dose increment mechanism are triggered. At the same time, the priority of nursing instructions is optimized through a conflict arbitration mechanism, realizing the intelligentization of the entire process of "data collection-path generation-verification feedback-priority control", and providing precise nursing plans for patients in the rheumatology and immunology department.

[0105] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and improvements fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. Intelligent personalized nursing pathway management system for rheumatology and immunology patients, characterized by: The following units are included: The joint activity-inflammation data acquisition unit (1) collects the three-dimensional joint angle change sequence through the nine-axis inertial sensor, outputs the dynamic activity score after Kalman filtering, and detects interleukin-6 and tumor necrosis factor- , the inflammatory load index was calculated in combination with the current daily dose of prednisone; The path generation and dose binding unit (2) inputs the dynamic activity score into the time series segmentation model, identifies the starting time window of morning stiffness, and generates the recommended interval for rehabilitation training the next day. It matches the preset hormone dose response rule library according to the 24-hour fluctuation curve of the inflammatory load index. When the inflammatory load index exceeds the threshold for 6 consecutive hours, the prednisone daily dose gradient adjustment instruction is triggered; The cross-modal verification feedback unit (3) obtains the actual medication collection time through the smart medicine box, and when the deviation from the prednisone daily dose gradient adjustment instruction exceeds Hours, activate the joint sensor enhanced acquisition mode, and compare the dynamic activity score with the dynamic time regularization distance of the standard action template in real time during the rehabilitation training period. If the distance exceeds the threshold and the inflammatory load index rises simultaneously, it will be greater than or equal to , then a hormone dose increment signal is sent to the path generation and dose binding unit (2); The method for constructing the hormone dose response rule library includes: Establish the calculation formula for dynamic threshold: ;in, For the moment The dynamic threshold of is the patient's baseline inflammatory load index, is the inflammatory load index, The inflammatory load index is expressed as any time within 24 hours. ;in, is the current daily dose of prednisone, is the dose-influencing factor, is the half-life decay coefficient of prednisone, The time since the last dose of the medicine, is interleukin-6, Tumor necrosis factor- , used to reflect the immediate inflammatory state and the effect of drug intervention. The patient's baseline inflammatory load index is obtained based on the patient's individual historical baseline value and is used to reflect the basic inflammatory level of the individual's physiological characteristics; The generation rule of gradient adjustment instructions is: when More than 6 consecutive hours When the patient is diagnosed with leukemia, the prednisone dose is increased by 5 mg every 24 hours until the cumulative increase reaches 15 mg; when Less than 0.8 for 12 consecutive hours When the prednisone dose is reduced by 2.5 mg every 24 hours, the sampling rate of the joint sensor is increased to 15 Hz after the dose reduction. The dynamic time warping distance calculation method in the cross-modal verification feedback unit (3) includes: Construct a joint angle change template library for standard rehabilitation movements. Each template contains a 15-keyframe 3D Euler angle sequence, which is used to describe a set of angle values ​​for the rigid body posture in 3D space. The formula for calculating the dynamic time warping distance between the patient's motion and the template in real time is: ;in, is the path penalty coefficient, is the final dynamic time warping distance, is the optimal alignment path, To align point pairs, is the patient's joint angle vector, is the template joint angle vector, is the Euclidean distance; when When the dynamic time warping distance threshold is greater than the threshold and the inflammatory load index rising rate is greater than or equal to 2pg / ml / h, it is determined that the action execution is abnormal. The dynamic time warping distance threshold is used to determine the critical value of the degree of difference between the patient's rehabilitation action and the standard template; The activation logic of the joint sensor enhanced acquisition mode includes: When the time deviation of taking medicine is greater than 1 hour, the sampling rate of the nine-axis inertial sensor is increased from 10Hz to 20Hz, the electromyographic signal acquisition module around the joint is activated, and the muscle activation delay time is synchronously analyzed. In the enhanced mode, the joint stability index is generated every 5 minutes. When the joint stability index is less than 0.7, a training period compression instruction is sent to the path generation and dose binding unit (2), and the single time length is reduced by 25%; Among them, the method for generating the joint stability index is as follows: The joint stability index is constructed by fusing nine-axis sensor data and electromyographic signals. The nine-axis sensor data includes angular velocity variance. and acceleration fluctuation coefficient , EMG signals include muscle activation synchronization , the calculation formula of joint stability index is: ;in, = Actual co-contraction index / ideal co-contraction index, used to measure the coordination of muscle activation around joints. The actual co-contraction index is obtained by calculating the activation time difference and intensity ratio of antagonistic muscles through surface electromyographic signals. The ideal co-contraction index is obtained based on the baseline value established based on data from healthy people.

2. The intelligent personalized nursing pathway management system for rheumatology and immunology patients according to claim 1 is characterized in that: The specific implementation method of the time series segmentation model in the path generation and dose binding unit (2) includes: A morning stiffness feature extractor is constructed based on a temporal convolutional network, and its expansion coefficient is set to , is the network layer number, k , the convolution kernel width is 3; The dynamic activity scoring sequence was phase-segmented. When the standard deviation of the joint angle within the adjacent 30-minute window decreased by 40% or more and lasted for more than 2 hours, it was determined to be the onset window of morning stiffness. The formula for generating the recommended interval for next-day rehabilitation training is: ;in, is the predicted value of the onset time of morning stiffness, , Buffer time for experience, This is the recommended interval for rehabilitation training the next day.

3. The intelligent personalized nursing pathway management system for rheumatology and immunology patients according to claim 1 is characterized in that: The path generation and dose binding unit (2) also includes a multi-constraint dose optimization mechanism: a. Before the gradient adjustment instruction takes effect, verify the following constraints: b. The cumulative daily dose of prednisone is less than or equal to 1 mg / kg, calculated based on the patient's body weight; c. The fluctuation of alanine aminotransferase, a liver function indicator, within the past 7 days is less than 30%; If any of a, b, and c is not satisfied, the alternative decision process is initiated: The 5-mg prednisone increment was replaced with an equivalent intravenous interleukin-6 inhibitor regimen; and the rehabilitation training period was adjusted to 2 hours after the peak blood drug concentration.

4. The intelligent personalized nursing pathway management system for rheumatology and immunology patients according to claim 1 is characterized in that: The triggering mechanism of the hormone dose increment signal includes the following steps: Establish a mapping relationship between incremental signal strength and multimodal data: when In the range When the synchronous increase rate of the inflammatory load index is greater than or equal to 15%, a low-intensity incremental signal is triggered. is a low-intensity incremental threshold, used to trigger the inflammatory load range of a low-dose increment, wherein the low-intensity incremental signal is +2.5 mg / 6 h; when When the synchronous increase rate of the inflammatory load index is greater than or equal to 25%, a high-intensity incremental signal is triggered, and the high-intensity incremental signal is +5mg / 6h. is the high-intensity increment threshold, which is the critical value of inflammatory load used to trigger high-dose increments; While the incremental signal is in effect, the real-time motion monitoring mode of the patient-side APP is forced to be turned on: If a sudden decrease in joint range of motion greater than 30% is detected, immediately suspend dose adjustment and initiate an emergency video call link; If the joint range of motion recovers to more than 90% of the baseline value, resume the incremental process and shorten the monitoring interval to every 15 minutes.

5. The intelligent personalized nursing pathway management system for rheumatology and immunology patients according to claim 1 is characterized in that: It also includes a closed-loop priority control unit (4), which is as follows: Establish a conflict arbitration mechanism for nursing instructions. When the adjustment of hormone dosage conflicts with the rehabilitation training period, calculate the expected inflammation suppression efficiency of hormone adjustment and the joint function improvement efficiency of rehabilitation training respectively. If the expected inflammation suppression efficiency of hormone adjustment is greater than 1.2 times the joint function improvement efficiency of rehabilitation training, then the training will be postponed and the dose will be increased, and a compensatory training plan will be generated at the same time; If the joint function improvement efficiency of rehabilitation training is greater than or equal to the expected inflammation inhibition efficiency of hormone adjustment, the original 60-minute training will be divided into two 25-minute modules and one 10-minute module, which will be placed 0.5 hours before and 1.5 hours after the dose adjustment respectively. The two 25-minute modules are the main training of 50 minutes, and the 10-minute module is a buffer time of 10 minutes for preparation and relaxation before and after training.

6. The intelligent personalized nursing pathway management system for rheumatology and immunology patients according to claim 5 is characterized in that: The calculation formulas for the expected inflammation suppression efficiency of hormone adjustment and the joint function improvement efficiency of rehabilitation training are as follows: Expected inflammation suppression efficiency of the hormone regulation: ;in, is the difference between the current inflammatory load index and the average of the previous 6 hours, is the predicted value of prednisone blood concentration, calculated based on the pharmacokinetic model; The efficiency of the rehabilitation training in improving joint function: ;in, The rate of change of dynamic activity score before and after training, The coordination of muscle activation is calculated by the phase synchronization of electromyographic signals.

7. The intelligent personalized nursing pathway management system for rheumatology and immunology patients according to claim 6 is characterized in that: The closed-loop priority control unit (4) updates the parameters of the time series segmentation model every 30 days, as follows: The expansion coefficient of the time series segmentation model is refitted based on historical data, and the convolution kernel width is adjusted using the Bayesian optimization algorithm; The threshold of the hormone dose response rule library is modified according to the individual pharmacokinetic characteristics of the patient. The modification formula is: ;in, It is a new generation of inflammatory load threshold. is the current cycle threshold, For the Peak daily inflammatory load, and The weights are assigned based on historical data.

Images