Parturient and fetus two-dimensional synergetic painless delivery anesthesia dynamic regulation and control monitoring method

Abstract

The invention provides a puerpera and fetus two-dimensional synergistic painless delivery anesthesia dynamic regulation and control monitoring method, and belongs to the technical field of painless delivery anesthesia. The method comprises the following steps: constructing a puerpera-fetus two-dimensional monitoring system, and synchronously collecting multidimensional signals such as puerpera side pain perception, circulatory respiration and drug metabolism and core signals such as fetus side heart rate and blood oxygen saturation; performing adaptive filtering preprocessing and attention mechanism driven multi-source data fusion analysis to obtain a puerpera anesthesia state, a fetus safety state and a collaborative correlation coefficient of the two states; constructing a regulation and control model by taking a parturient painless threshold-a fetus safety threshold as double constraints, outputting an initial anesthesia scheme, and performing real-time dynamic adjustment in combination with a delivery process and drug interaction; meanwhile, full-process data recording and three-level early warning are achieved, and the scheme is optimized through backtracking analysis after delivery. The problems that in the prior art, puerpera pain relieving is emphasized, and fetus safety cooperative monitoring and regulation lack dynamics are neglected are solved, and anesthesia precision and mother and infant safety are improved.

Description

Technical Field

[0001] This invention belongs to the field of painless childbirth anesthesia technology, and more specifically, it relates to a method for dynamic regulation and monitoring of painless childbirth anesthesia with dual-dimensional coordination between the mother and fetus. Background Technology

[0002] Painless childbirth is a key technology for alleviating labor pain and improving the childbirth experience, with precise control of the anesthesia protocol being crucial for ensuring a safe delivery. Currently, commonly used painless childbirth anesthesia methods primarily rely on the mother's pain perception for control, adjusting the dosage of anesthetic drugs by monitoring the mother's heart rate, blood pressure, and subjective pain scores (such as VAS scores). However, existing technologies have the following significant drawbacks:

[0003] 1. Limited monitoring dimensions, neglecting fetal safety coordination: Current methods only focus on the physiological state of the mother on one side, without establishing a coordinated monitoring mechanism between the mother and fetus. Anesthetic drugs can affect the physiological state of the fetus by crossing the placental barrier. If the dosage is adjusted only based on the mother's condition, it may lead to excessive anesthesia, causing risks such as abnormal fetal heart rate and intrauterine distress, or insufficient anesthesia, which may not effectively relieve the mother's pain.

[0004] 2. Lack of dynamism and individual adaptability in regulation: Existing protocols are mostly static dosing modes based on experience, or rely on manual adjustments by medical staff. They are difficult to respond in real time to the dynamic changes in the intensity of pain and the rate of drug metabolism during the labor process, and cannot accurately adapt to the individual differences of mothers with different gestational weeks, weights, and physical conditions, which can easily lead to poor anesthetic effects or safety risks.

[0005] 3. Insufficient signal processing accuracy and imperfect early warning mechanism: Physiological signals collected during childbirth are easily affected by power frequency interference, baseline drift, etc. Existing filtering methods have limited noise reduction effects, resulting in inaccurate signal feature extraction; at the same time, the lack of a hierarchical early warning system makes it impossible to identify abnormal changes in the maternal-fetal status in a timely manner, delaying the timing of intervention.

[0006] 4. Lack of consideration for drug interactions and postpartum optimization mechanisms: Some mothers need to use other drugs in combination due to pregnancy complications. The existing protocols do not consider the synergistic / antagonistic effects of drugs and anesthetics, which can easily lead to adverse reactions. In addition, there is no mechanism for full-process data backtracking and analysis, which cannot provide data support for optimizing anesthesia protocols for similar cases in the future.

[0007] Therefore, developing a painless childbirth anesthesia method that can achieve dual-dimensional collaborative monitoring of the mother and fetus, dynamic and precise control, and full-process safety early warning has become a technical problem that urgently needs to be solved in clinical practice. Summary of the Invention

[0008] To address the aforementioned technical problems, this invention provides a dynamic control and monitoring method for painless childbirth anesthesia with dual-dimensional coordination between the mother and fetus. This method solves the technical problems of existing painless childbirth anesthesia methods, such as single monitoring dimension, neglect of fetal safety coordination, lack of dynamic and individual adaptability in regulation, insufficient signal processing accuracy, imperfect early warning mechanism, and lack of consideration for drug interactions and postpartum optimization mechanism.

[0009] A method for dynamic control and monitoring of painless childbirth anesthesia with dual-dimensional coordination between the mother and fetus includes the following steps:

[0010] S1: Construct a maternal-fetal dual-dimensional monitoring system, simultaneously collecting multi-dimensional physiological signals from the maternal side and core physiological signals from the fetal side. The multi-dimensional physiological signals from the maternal side include at least pain perception signals, circulatory system signals, respiratory system signals, and anesthetic drug metabolism-related signals. The core physiological signals from the fetal side include at least fetal heart rate signals and fetal blood oxygen saturation signals.

[0011] S2: The two-dimensional physiological signals collected in step S1 are preprocessed. An adaptive filtering algorithm is used to remove interference noise from the signals. Then, a multi-source data fusion model is used to extract features and perform correlation analysis on the preprocessed two-dimensional signals to obtain the maternal anesthesia status assessment value, the fetal safety status assessment value, and the synergistic correlation coefficient between the two.

[0012] S3: Based on the evaluation results obtained in step S2, a two-dimensional collaborative anesthesia control model is constructed. The control model uses "maternal pain-free threshold - fetal safety threshold" as dual constraints and combines the parameters of the labor process stage to output the initial anesthesia administration plan. The initial anesthesia administration plan includes the type of anesthetic drug, the initial dosage and the administration rate.

[0013] S4: Real-time dynamic monitoring of changes in dual-dimensional physiological signals during labor. The real-time collected and processed maternal anesthesia status assessment value and fetal safety status assessment value are compared with preset dual constraint thresholds. When any assessment value exceeds the corresponding threshold range or the synergistic correlation coefficient fluctuates abnormally, the control model is triggered to make adaptive adjustments and output dynamically corrected anesthesia administration parameters.

[0014] S5: Records dual-dimensional physiological signal data, anesthesia administration parameter adjustment records, and maternal-fetal status assessment results throughout the entire labor process, forming a full-process data archive for anesthesia regulation and monitoring, while simultaneously outputting monitoring and early warning information to medical terminals in real time.

[0015] Preferably, in step S1, the pain perception signal is acquired through a multimodal acquisition method based on a combination of electromyography (EMG) signals and visual analog scale (VAS) pain scores. Specifically, the EMG signals of the lumbosacral region of the mother are acquired through an EMG sensor, and the real-time VAS score of the mother is acquired through an interactive terminal. The EMG signal feature parameters are correlated and calibrated with the VAS score to obtain a standardized pain perception signal.

[0016] Preferably, in step S1, the anesthetic drug metabolism-related signals include the concentration of anesthetic drugs in the mother's blood, liver and kidney function indicators, and blood gas analysis indicators, which are collected in real time by a portable non-invasive blood component monitoring device at a collection frequency of not less than 1 Hz.

[0017] Preferably, in step S2, the multi-source data fusion model adopts a deep learning model based on the attention mechanism. The input layer of the model is a preprocessed two-dimensional physiological signal feature vector. The hidden layer strengthens the correlation features between the maternal pain signal and the fetal heart rate signal through the attention weight allocation module. The output layer simultaneously outputs the maternal anesthesia status assessment value, the fetal safety status assessment value, and the co-correlation coefficient. The co-correlation coefficient is used to characterize the correlation strength of the effect of anesthetic drugs on the physiological status of the maternal and fetal patients.

[0018] Preferably, in step S3, the dual constraint condition of "maternal pain-free threshold - fetal safety threshold" is determined in the following way: based on the historical labor anesthesia case database, machine learning algorithms are used to mine the range of maternal pain-free thresholds for different gestational weeks, weights and labor stages, while the range of fetal safety thresholds is determined in combination with fetal physiological development standards. The dual thresholds are dynamically corrected through a threshold calibration module to ensure individual adaptability of the thresholds.

[0019] Preferably, in step S4, the adaptive adjustment specifically includes: when the maternal anesthesia status assessment value is lower than the pain-free threshold (i.e., increased pain perception) and the fetal safety status assessment value is within the safe range, gradually increasing the anesthesia administration rate within the preset dose range; when the fetal safety status assessment value exceeds the safe threshold (e.g., abnormal fetal heart rate), immediately reducing the anesthesia administration dose and triggering the fetal status enhancement monitoring mode to increase the frequency of fetal physiological signal acquisition.

[0020] Preferably, step S4 also includes a mechanism for regulating the interaction between anesthetic drugs: when the mother needs to use other drugs in combination due to complications, the interaction coefficient between the target drug and the anesthetic drug is queried through the drug interaction database, and the coefficient is input into the regulation model to make compensatory adjustments to the anesthetic drug administration parameters to avoid the risk of excessive anesthesia or fetal toxicity caused by drug synergy.

[0021] Preferably, in step S5, the monitoring and early warning information is divided into three levels: Level 1 early warning corresponds to a slight fluctuation in the synergistic correlation coefficient, prompting medical staff to pay attention; Level 2 early warning corresponds to a single-dimensional assessment value exceeding the threshold, triggering the control model to automatically adjust and prompting medical staff to review; Level 3 early warning corresponds to both dual-dimensional assessment values ​​exceeding the threshold, immediately suspending anesthesia administration and issuing an audible and visual alarm, while simultaneously pushing emergency treatment suggestions to the medical staff terminal.

[0022] Preferably, the adaptive filtering algorithm in step S2 adopts a hybrid algorithm combining Kalman filtering and wavelet threshold denoising. First, random noise is eliminated by Kalman filtering, and then baseline drift and power frequency interference are eliminated by wavelet threshold denoising to ensure the accuracy of signal preprocessing.

[0023] Preferably, it also includes a postpartum retrospective analysis module: based on the full-process data archive formed in step S5, data mining algorithms are used to analyze the correlation between anesthesia administration parameters and the mother's childbirth experience and the fetus's postpartum status, providing data support for optimizing anesthesia plans for similar mothers in the future.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] This invention constructs a dual-dimensional collaborative monitoring system to ensure the safety of both mother and baby: Breaking through the limitations of existing single-dimensional monitoring, it simultaneously collects physiological signals from both the mother and fetus. Through fusion analysis, a collaborative correlation model between the two is established, enabling anesthesia control to simultaneously address the mother's need for pain relief and the fetus's safety. This significantly reduces the incidence of fetal distress caused by anesthetic drugs, solving the core problem of existing technologies neglecting fetal safety. Experimental verification shows that the incidence of fetal distress using this method is only 2%, far lower than the 8% of traditional methods.

[0026] Dynamic adaptive regulation enhances anesthesia precision and individual adaptability: Based on dual-constraint thresholds and a machine learning model, this invention enables personalized initial configuration and real-time dynamic adjustment of anesthesia protocols. It can accurately adapt to individual differences in mothers at different gestational weeks, weights, and stages of labor, while also responding to dynamic changes in physiological signals and drug interactions during labor. Compared to traditional manual adjustment methods, this invention achieves a 98% accuracy rate in adjusting anesthesia parameters, increases the effective pain relief rate for mothers to 96%, and reduces the number of adjustments by 47.7%, significantly reducing the workload of medical staff.

[0027] High-precision signal processing and hierarchical early warning reduce safety risks: A hybrid algorithm combining Kalman filtering and wavelet threshold denoising is adopted to effectively eliminate signal interference, improve the signal-to-noise ratio by ≥20dB, and ensure the accuracy of signal analysis; a three-level early warning system is constructed, which can accurately trigger different levels of early warning based on changes in the maternal-fetal status and synergistic correlation coefficient, and push targeted treatment suggestions to achieve early identification and early intervention of risks and avoid delaying treatment.

[0028] Full-process data backtracking and scheme optimization form a closed-loop management: This invention records full-process data through a hybrid architecture of local + cloud storage. After delivery, it uses association rule mining to achieve backtracking analysis of anesthesia scheme and delivery effect. The optimization rules are stored in the knowledge base to provide data support for subsequent similar cases, forming a closed-loop management of "monitoring-control-early warning-backtracking-optimization", which promotes the continuous iteration and upgrading of painless childbirth anesthesia technology.

[0029] Non-invasive data collection and convenient interaction enhance maternal comfort: The collection of physiological signals from the maternal side uses non-invasive sensors, avoiding the risk of infection from invasive procedures; pain scores can be actively recorded through an easily accessible interactive terminal, improving maternal participation and the childbirth experience, and has good prospects for clinical promotion. Attached Figure Description

[0030] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0031] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0032] Please see Figure 1 This invention provides a method for dynamic control and monitoring of painless childbirth anesthesia in a two-dimensional coordinated manner between the mother and fetus. The invention is further described in detail below with reference to specific embodiments, so that those skilled in the art can fully understand and implement it. It should be noted that the following embodiments are only for explaining the invention and are not intended to limit the scope of protection of the invention.

[0033] The core of the dynamic regulation and monitoring method for painless childbirth anesthesia with dual-dimensional coordination between the mother and fetus disclosed in this invention lies in constructing a dual-dimensional monitoring and coordinated regulation system for the mother and fetus. This system enables dynamic and precise regulation of anesthetic administration and full-process safety monitoring, solving the technical problems of existing painless childbirth anesthesia that only focus on relieving the mother's pain, neglecting the coordinated monitoring of fetal safety, and lacking dynamic and individualized regulation. The following is a detailed description of the specific implementation method.

[0034] In this embodiment, the system implementing the method of the present invention includes: a dual-dimensional signal acquisition module, a signal preprocessing module, a multi-source data fusion analysis module, a dual-dimensional collaborative anesthesia control module, a real-time monitoring and early warning module, a data storage and retrospective analysis module, and a medical terminal (including a display unit, an alarm unit, and an interaction unit). Each module interacts with the other via industrial Ethernet or wireless local area network (5G / Wi-Fi 6), with a data transmission latency of ≤50ms to ensure real-time performance.

[0035] The core equipment selections for each module are as follows:

[0036] Dual-dimensional signal acquisition module: On the mother's side, there is a lumbosacral electromyography sensor, a multi-parameter monitor (model: acquires circulatory system signals: heart rate, blood pressure, pulse oximetry; respiratory system signals: respiratory rate, tidal volume), and a portable non-invasive blood component monitoring device (acquires anesthetic drug concentration, liver and kidney function indicators, and blood gas analysis indicators); on the fetal side, there is a fetal heart rate monitor (acquires fetal heart rate signals) and a fetal oxygen saturation monitor (acquires fetal oxygen saturation signals); the interactive terminal uses a tablet computer to obtain the mother's VAS score.

[0037] Signal preprocessing module and data fusion analysis module: Employs an edge computing gateway, with built-in adaptive filtering algorithms and attention-based deep learning models to achieve real-time signal preprocessing and fusion analysis.

[0038] Dual-dimensional collaborative anesthesia control module: linked with the anesthesia infusion pump, it realizes automatic adjustment of anesthesia administration parameters through serial communication.

[0039] Real-time monitoring and early warning module: The medical terminal is equipped with an industrial-grade display and an audible and visual alarm. Early warning information is pushed to the medical staff's mobile APP (based on Android / iOS development).

[0040] Data storage and backtracking analysis module: It adopts a hybrid storage architecture of local server + cloud storage (Alibaba Cloud OSS) to ensure data security and traceability.

[0041] Specific implementation steps:

[0042] Step S1: Construct a maternal-fetal dual-dimensional monitoring system and simultaneously collect dual-dimensional physiological signals.

[0043] Signal Acquisition Preparation: After the mother enters the delivery room, electromyography (EMG) sensor electrodes are attached to the L3-L5 segment of the lumbosacral region, ensuring good contact between the electrodes and the skin (impedance ≤5kΩ). The blood pressure cuff of the multi-parameter monitor is attached to the mother's upper limb, the pulse oximeter is clipped to the fingertip, and the respiration sensor is fixed to the chest and abdomen. The sensor of the non-invasive blood component monitoring device is attached to the inner side of the mother's forearm. On the fetal side, the fetal heart rate monitor sensor is fixed by locating the fetal heart rate position using an ultrasound probe. The fetal oxygen saturation sensor is attached to the fetal scalp through a vaginal probe (suitable for mothers with cervical dilation ≥2cm). At the same time, the interactive terminal (tablet computer) is placed within easy reach of the mother, and the VAS score entry interface is opened (0-10 points, 0 points for no pain, 10 points for the most severe pain).

[0044] Synchronous signal acquisition: Activate all acquisition devices and set the acquisition frequency: electromyography (EMG) signals at 200Hz, circulatory and respiratory system signals at 10Hz, anesthetic drug metabolism-related signals at 1Hz (meeting the requirement of an acquisition frequency not lower than 1Hz), and fetal heart rate and fetal blood oxygen saturation signals at 5Hz. Timestamp synchronization technology ensures that the acquisition time error of all signals is ≤10ms, achieving synchronous acquisition of two-dimensional signals. The specific physiological signals on the maternal side include: pain perception signals (electromyography + VAS score), circulatory system signals (heart rate 60-100 beats / min, systolic blood pressure 90-140 mmHg, diastolic blood pressure 60-90 mmHg, pulse oxygen saturation 95%-100%), respiratory system signals (respiratory rate 12-20 breaths / min, tidal volume 500-800 mL), and anesthetic drug metabolism-related signals (such as ropivacaine concentration 0-4 μg / mL, liver function indicator ALT 7-40 U / L, kidney function indicator Cr 53-106 μmol / L, blood gas analysis indicator pH 7.35-7.45); the specific core physiological signals on the fetal side include: fetal heart rate 110-160 beats / min, and fetal oxygen saturation ≥30% (normal range for intrauterine fetal blood oxygen saturation).

[0045] Multimodal acquisition and calibration of pain perception signals: Electromyography (EMG) sensors collect EMG signals from the lumbosacral region of the mother in real time, and extract characteristic parameters such as peak value, integrated electromyography (IEMG) value, and root mean square value (RMS) of the EMG signals; at the same time, the mother enters a real-time VAS score every 5 minutes through an interactive terminal, or actively enters a score when the pain intensifies.

[0046] A linear regression algorithm was used to correlate and calibrate the electromyographic signal characteristic parameters with the VAS score, and a calibration model was established: VAS_pred=a×RMS+b×IEMG+c (where a, b, and c are calibration coefficients, which were obtained by training with data from 50 pre-labor cases, and the calibration error is ≤0.5 points). This model converts the electromyographic signal into a standardized pain perception signal, thereby achieving an objective quantitative assessment of the pain state.

[0047] Step S2: Two-dimensional physiological signal preprocessing and multi-source data fusion analysis

[0048] Signal preprocessing: A hybrid adaptive filtering algorithm combining Kalman filtering and wavelet threshold denoising is used to preprocess the acquired two-dimensional physiological signals. Specific process:

[0049] ① Kalman filtering to eliminate random noise: The state equation of the Kalman filter is set as X(k)=A×X(k-1)+B×u(k)+w(k), and the observation equation is Z(k)=H×X(k)+v(k), where A is the state transition matrix (set as an identity matrix), B is the control matrix (0), H is the observation matrix (identity matrix), the variance of the process noise w(k) is Q=0.01, and the variance of the observation noise v(k) is R=0.05. This algorithm is used to perform preliminary denoising on signals containing random noise such as electromyography signals and heart rate signals.

[0050] ② Wavelet thresholding for baseline drift and power frequency interference removal: The db4 wavelet is used to decompose the Kalman-filtered signal into three levels. A threshold λ = σ × √(2 × lnN) (where σ is the noise standard deviation and N is the signal length) is set. The high-frequency coefficients obtained from the decomposition are thresholded (soft thresholding function: w' = sign(w)(|w|-λ)). Then, the signal is reconstructed through inverse wavelet transform to remove baseline drift (such as slow baseline fluctuations in electromyography signals) and power frequency interference (50Hz). After preprocessing, the signal-to-noise ratio is improved by ≥20dB, ensuring signal quality.

[0051] Multi-source data fusion analysis: A CNN-LSTM deep learning model based on an attention mechanism is used to extract features and perform correlation analysis on the preprocessed two-dimensional signals. Model construction and training:

[0052] ① Input layer: Normalize the preprocessed physiological signals (mapped to the range [0,1]) and construct feature vectors with a dimension of 12 (maternal side: pain perception signal, heart rate, blood pressure, pulse oxygen saturation, respiratory rate, tidal volume, anesthetic drug concentration, liver and kidney function indicators, blood gas analysis indicators; fetal side: fetal heart rate, fetal oxygen saturation, fetal movement signal).

[0053] ② Hidden layer: It includes a CNN feature extraction submodule and an LSTM temporal modeling submodule. The CNN submodule uses two convolutional layers (3×1 kernel size, 32 and 64 kernels respectively) to extract the spatial features of the signal, and the LSTM submodule uses two LSTM layers (128 hidden units) to extract the temporal features of the signal.

[0054] The attention weight allocation module calculates the attention weight of each feature using the Softmax function, strengthening the correlation between maternal pain signal (weight 0.25) and fetal heart rate signal (weight 0.2) and weakening the influence of secondary features;

[0055] ③ Output layer: A fully connected layer is used to output three results: maternal anesthesia status assessment value (0-10 points, 0 points for insufficient anesthesia, 10 points for excessive anesthesia, 4-6 points for ideal anesthesia status), fetal safety status assessment value (0-10 points, 0 points for extreme danger, 10 points for safety, ≥8 points for safe range), and co-correlation coefficient (0-1, the larger the value, the stronger the correlation between the effect of anesthetic drugs on the physiological status of the mother and fetus, the normal range is 0.3-0.7).

[0056] The model was trained using data from 1,000 historical childbirth anesthesia cases, with a training set to test set ratio of 8:2. The prediction error of the evaluation value on the test set was ≤0.3 points, and the prediction error of the co-correlation coefficient was ≤0.05.

[0057] Step S3: Construct a two-dimensional collaborative anesthesia regulation model and output the initial anesthesia administration regimen.

[0058] Dual-constraint threshold determination: Based on a historical database of childbirth anesthesia cases (containing data from 5000 women at different gestational weeks, weights, and stages of labor), a random forest machine learning algorithm was used to mine the range of pain-free thresholds for the mother and the range of safety thresholds for the fetus. Specifically:

[0059] ① Painless threshold range for mothers: By exploring the correlation between pain perception signals and anesthetic drug dosage in mothers at different gestational weeks (28-42 weeks), weight (45-80kg), and labor stages (latent phase, active phase, second stage of labor), the painless threshold range for different individuals (the pain perception signal range corresponding to a maternal anesthesia status assessment score of 4-6) was determined.

[0060] ② Fetal safety threshold range: Based on the fetal physiological development standards in the clinical application guidelines for electronic fetal monitoring, determine the fetal safety threshold range (fetal safety status assessment value ≥8 points corresponds to a fetal heart rate of 110-160 beats / min and a fetal blood oxygen saturation ≥30%).

[0061] ③ Threshold calibration: The threshold calibration module combines the current mother's individual information (39 weeks of gestation, weight 65kg, active labor stage) with the fetus's basic physiological parameters (fetal heart rate 140 beats / min, fetal blood oxygen saturation 60%) to dynamically correct the dual thresholds, thereby obtaining the current mother's personalized pain-free threshold range (mother's anesthesia status assessment value 4.2-5.8 points) and the fetal safety threshold range (fetal safety status assessment value ≥8.2 points).

[0062] Initial anesthesia administration protocol output: The dual-dimensional collaborative anesthesia control model uses the maternal pain threshold and fetal safety threshold as dual constraints, combined with current labor stage parameters (active phase, cervical dilation 5cm), to screen suitable anesthetic drugs from the anesthetic drug library (including commonly used painless labor anesthetics such as ropivacaine and sufentanil), and calculates the initial dosage and administration rate. In this embodiment, the screened anesthetic drug is 0.75% ropivacaine, the initial dosage is 10mg, and the administration rate is 8mL / h. The initial anesthesia administration protocol is transmitted to the anesthesia infusion pump through the control module to initiate anesthesia administration.

[0063] Step S4: Real-time dynamic monitoring and adaptive adjustment of anesthesia drug administration parameters

[0064] Real-time monitoring and threshold comparison: Two-dimensional physiological signals during labor are collected in real time. The preprocessing and fusion analysis process in step S2 is repeated to obtain real-time assessment values ​​of maternal anesthesia status, fetal safety status, and co-correlation coefficients. The real-time assessment values ​​are compared with the dual-constraint thresholds determined in step S3, and the fluctuation of the co-correlation coefficient is monitored (fluctuation amplitude ≥ 0.2 is considered abnormal).

[0065] Adaptive adjustment: The control model is triggered to perform adaptive adjustments based on the comparison results.

[0066] ① When the maternal anesthesia status assessment score is 3.5 (below the lower limit of the pain-free threshold of 4.2, indicating increased pain perception) and the fetal safety status assessment score is 8.5 (within the safe range), the anesthesia administration rate is gradually increased within the preset dose range (8-15 mL / h), increasing by 1 mL / h each time, maintaining the increase for 3 minutes, and the maternal anesthesia status is reassessed until the assessment score enters the ideal range of 4.2-5.8. Finally, the administration rate is adjusted to 10 mL / h.

[0067] ② When the fetal safety status assessment score drops to 7.8 (below the lower limit of the safety threshold of 8.2, and the fetal heart rate drops to 105 beats / min), and the maternal anesthesia status assessment score is 5.0 (within the ideal range), immediately reduce the anesthetic drug dosage to 8mg, reduce the drug administration rate to 6mL / h, and trigger the fetal status enhanced monitoring mode, increase the fetal physiological signal acquisition frequency to 10Hz, and continuously monitor changes in fetal heart rate and blood oxygen saturation until the fetal safety status assessment score rises back to ≥8.2.

[0068] Regulation of anesthetic drug interactions: If a pregnant woman requires the combined use of labetalol for gestational hypertension, the interaction coefficient between labetalol and ropivacaine is found to be 0.15 (synergistic effect, enhancing anesthetic efficacy) using a drug interaction database (which includes interaction data for 1000 commonly used clinical drugs and anesthetic drugs). This coefficient is then input into the regulation model to compensate for the anesthetic drug administration parameters, further reducing the administration rate to 5 mL / h to avoid the risk of excessive anesthesia or fetal toxicity due to drug synergy.

[0069] Step S5: Full-process data recording and monitoring early warning

[0070] The entire process data archive is formed by recording two-dimensional physiological signal data (stored once every 1 second) throughout the entire labor process, anesthesia administration parameter adjustment records (including adjustment time, parameters before adjustment, parameters after adjustment, and reasons for adjustment), and maternal-fetal status assessment results (recorded once every 30 seconds). The data is stored on local servers and cloud storage to form a complete data archive for anesthesia control and monitoring. The archive contains a unique maternal identifier (medical record number) for easy subsequent query and retrieval.

[0071] Level 3 Early Warning Information Output: Based on real-time assessment results and collaborative correlation coefficients, Level 3 early warning information is output to the medical staff terminal.

[0072] ① Level 1 warning: When the co-correlation coefficient fluctuates to 0.8 (out of the normal range of 0.3-0.7), but the assessment values ​​of the mother and fetus are both within the threshold range, a Level 1 warning is triggered. The medical terminal displays a yellow warning message "Abnormal fluctuation in co-correlation coefficient, please pay attention to monitoring", without sound alarm;

[0073] ② Level 2 warning: When the fetal safety status assessment value drops to 7.8 points (exceeding the safety threshold range), a level 2 warning is triggered. The medical terminal displays an orange warning message "The fetal safety status is abnormal and the anesthesia parameters have been automatically adjusted". At the same time, a low-frequency sound alarm (frequency 500Hz, lasting 2 seconds) is issued to prompt medical staff to review and adjust the plan.

[0074] ③ Level 3 warning: When the maternal anesthesia status assessment value drops to 2.5 points (insufficient anesthesia) and the fetal safety status assessment value drops to 6.0 points (extremely dangerous), a level 3 warning is triggered. Anesthesia administration is immediately suspended, and the medical staff terminal displays a red warning message "Abnormal maternal-fetal dual-dimensional status, take emergency measures immediately". A high-frequency audible and visual alarm is issued (frequency 1000Hz, lasting 5 seconds). At the same time, emergency treatment suggestions (such as stopping medication, oxygen inhalation, and adjusting the maternal position) are pushed to the medical staff's mobile APP.

[0075] Step S6: Postpartum Retrospective Analysis

[0076] After delivery, the postpartum retrospective analysis module, based on the full-process data archive formed in step S5, uses the Apriori algorithm to analyze the correlation between anesthetic administration parameters (such as ropivacaine dosage and administration rate) and the mother's delivery experience (postpartum VAS score, delivery duration) and the fetus's postpartum status (Apgar score, umbilical artery blood gas analysis). For example, it mines the association rule that "when the initial ropivacaine dose is 10mg and the administration rate is 8-10mL / h, the mother's postpartum VAS score is ≤3 points and the fetal Apgar score is ≥9 points." This rule is stored in the knowledge base to provide data support for optimizing anesthesia protocols for similar mothers (39 weeks gestation, 65kg weight, active delivery).

[0077] Implementation effect verification:

[0078] One hundred women in labor were randomly divided into an experimental group (using the method of this invention) and a control group (using traditional painless childbirth anesthesia methods, monitoring only the mother's pain and vital signs, and manually adjusting the anesthetic dosage), with 50 cases in each group. The results showed that the effective pain relief rate (VAS score ≤ 3) in the experimental group was 96%, significantly higher than the 82% in the control group; the incidence of fetal distress in the experimental group was 2%, significantly lower than the 8% in the control group; the average number of anesthesia parameter adjustments was 3.2 times in the experimental group and 6.5 times in the control group, and the adjustment accuracy rate (the proportion of mothers whose anesthesia state entered the ideal range after adjustment) was 98% in the experimental group, significantly higher than the 75% in the control group. These results indicate that the method of this invention can achieve coordinated monitoring and precise control of both the mother and fetus, improving the effectiveness of painless childbirth and fetal safety.

[0079] This invention discloses a dynamic monitoring and control method for painless childbirth anesthesia with dual-dimensional coordination between the mother and fetus. The core of this method lies in constructing a dual-dimensional monitoring and coordinated control system for the mother and fetus, addressing the shortcomings of traditional painless childbirth anesthesia which focuses on maternal pain relief while neglecting fetal safety and lacking dynamic and individualized monitoring and control. This method simultaneously collects multi-dimensional signals from the mother (pain perception, circulation and respiration, drug metabolism, etc.) and core signals from the fetus (heart rate, blood oxygen saturation, etc.). Through adaptive filtering preprocessing and attention-driven multi-source data fusion analysis, it obtains the mother's anesthesia status, the fetal safety status, and the synergistic correlation coefficient between the two. A control model is constructed using the "maternal pain relief threshold - fetal safety threshold" as dual constraints, outputting an initial anesthesia plan and dynamically adjusting it in real time according to the progress of labor, while also incorporating drug interaction control mechanisms. A three-level early warning system enables full-process safety monitoring, forming a complete data archive and optimizing subsequent plans based on postpartum retrospective analysis. Implementation and verification show that this method can significantly improve the effectiveness of maternal pain relief, reduce the incidence of fetal distress, achieve precise and coordinated control of anesthesia, and ensure maternal and infant safety.

[0080] The embodiments of the present invention are given for illustrative and descriptive purposes only, and are not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of the invention, and to enable those skilled in the art to understand the invention and to design various embodiments with various modifications suitable for a particular purpose.

Claims

1. A method for dynamically regulating and monitoring painless delivery anesthesia of maternal-fetal two-dimensional coordination, characterized in that, Includes the following steps: S1: Construct a maternal-fetal dual-dimensional monitoring system, simultaneously collecting multi-dimensional physiological signals from the maternal side and core physiological signals from the fetal side. The multi-dimensional physiological signals from the maternal side include at least pain perception signals, circulatory system signals, respiratory system signals, and anesthetic drug metabolism-related signals. The core physiological signals from the fetal side include at least fetal heart rate signals and fetal blood oxygen saturation signals. S2: The two-dimensional physiological signals collected in step S1 are preprocessed. An adaptive filtering algorithm is used to remove interference noise from the signals. Then, a multi-source data fusion model is used to extract features and perform correlation analysis on the preprocessed two-dimensional signals to obtain the maternal anesthesia status assessment value, the fetal safety status assessment value, and the synergistic correlation coefficient between the two. S3: Based on the evaluation results obtained in step S2, a two-dimensional collaborative anesthesia control model is constructed. The control model uses the maternal pain-free threshold and the fetal safety threshold as dual constraints, and combines the parameters of the labor process stage to output the initial anesthesia administration plan. The initial anesthesia administration plan includes the type of anesthetic drug, the initial dosage, and the administration rate. S4: Real-time dynamic monitoring of changes in dual-dimensional physiological signals during labor. The real-time collected and processed maternal anesthesia status assessment value and fetal safety status assessment value are compared with preset dual constraint thresholds. When any assessment value exceeds the corresponding threshold range or the synergistic correlation coefficient fluctuates abnormally, the control model is triggered to make adaptive adjustments and output dynamically corrected anesthesia administration parameters. S5: Records dual-dimensional physiological signal data, anesthesia administration parameter adjustment records, and maternal-fetal status assessment results throughout the entire labor process, forming a full-process data archive for anesthesia regulation and monitoring, while simultaneously outputting monitoring and early warning information to medical terminals in real time.

2. The method of claim 1, wherein, In step S1, the pain perception signal is acquired through a multimodal acquisition method combining electromyographic signals and visual analog scale (VAS) pain scores, specifically as follows: Electromyography (EMG) sensors are used to collect EMG signals from the lumbosacral region of postpartum women. Simultaneously, the real-time VAS score of the postpartum women is obtained through an interactive terminal. The characteristic parameters of the EMG signals are correlated and calibrated with the VAS scores to obtain standardized pain perception signals.

3. The method of claim 1, wherein, In step S1, the anesthetic drug metabolism-related signals include the concentration of anesthetic drugs in the mother's blood, liver and kidney function indicators, and blood gas analysis indicators, which are collected in real time using a portable non-invasive blood component monitoring device at a frequency of not less than 1 Hz.

4. The method of claim 1, wherein, In step S2, the multi-source data fusion model adopts a deep learning model based on the attention mechanism. The input layer of the model is a preprocessed two-dimensional physiological signal feature vector. The hidden layer strengthens the correlation features between the maternal pain signal and the fetal heart rate signal through the attention weight allocation module. The output layer simultaneously outputs the maternal anesthesia status assessment value, the fetal safety status assessment value, and the co-correlation coefficient. The co-correlation coefficient is used to characterize the correlation strength of the effect of anesthetic drugs on the physiological status of the maternal and fetal patients.

5. The method of claim 1, wherein, In step S3, the dual constraints of the maternal pain-free threshold and the fetal safety threshold are determined in the following way: Based on the historical labor anesthesia case database, machine learning algorithms are used to mine the range of maternal pain-free thresholds for different gestational weeks, weights and labor stages. At the same time, the range of fetal safety thresholds is determined by combining fetal physiological development standards. The dual thresholds are dynamically corrected through a threshold calibration module to ensure individual adaptability of the thresholds.

6. The method according to claim 1, characterized in that, In step S4, the adaptive adjustment specifically includes: when the maternal anesthesia status assessment value is lower than the pain-free threshold and the fetal safety status assessment value is within the safe range, gradually increasing the anesthesia administration rate within the preset dose range; when the fetal safety status assessment value exceeds the safe threshold, immediately reducing the anesthesia administration dose and triggering the fetal status enhanced monitoring mode to increase the frequency of fetal physiological signal acquisition.

7. The method according to claim 1, characterized in that, Step S4 also includes a mechanism for regulating the interaction between anesthetic drugs: when a pregnant woman needs to use other drugs in combination due to complications, the interaction coefficient between the target drug and the anesthetic drug is queried through the drug interaction database, and the coefficient is input into the regulation model to make compensatory adjustments to the anesthetic drug administration parameters to avoid the risk of excessive anesthesia or fetal toxicity caused by drug synergy.

8. The method according to claim 1, characterized in that, In step S5, the monitoring and early warning information is divided into three levels: Level 1 early warning corresponds to a slight fluctuation in the synergistic correlation coefficient, prompting medical staff to pay attention; Level 2 early warning corresponds to a single-dimensional assessment value exceeding the threshold, triggering the control model to automatically adjust and prompting medical staff to review; Level 3 early warning corresponds to both dual-dimensional assessment values ​​exceeding the threshold, immediately suspending anesthesia administration and issuing an audible and visual alarm, while simultaneously pushing emergency treatment suggestions to the medical staff terminal.

9. The method according to claim 1, characterized in that, The adaptive filtering algorithm in step S2 adopts a hybrid algorithm that combines Kalman filtering and wavelet threshold denoising. First, random noise is eliminated by Kalman filtering, and then baseline drift and power frequency interference are eliminated by wavelet threshold denoising to ensure the accuracy of signal preprocessing.

10. The method according to claim 1, characterized in that, It also includes a postpartum retrospective analysis module: based on the full-process data archive formed in step S5, data mining algorithms are used to analyze the correlation between anesthesia administration parameters and the mother's childbirth experience and the fetus's postpartum status, providing data support for optimizing anesthesia plans for similar mothers in the future.

Images