Method and apparatus for intelligent cardiopulmonary monitoring by electrocardiogram, respiratory acoustics, thoracic acceleration, temperature, and hemoglobin oxygen saturation

Abstract

This invention relates to a patch and a method for determining a respiratory quality level index and noninvasive blood pressure in neonates, infants, and adult patients by combining parameters extracted from electrocardiogram (ECG), chest and cardiac motion and acceleration, respiratory acoustics, cardiac acoustics, temperature, and transcutaneous oxygen saturation of hemoglobin. The device includes a microprocessor configured to: calculate cross-information between heart rate variability from ECG, respiratory acoustics, chest motion, tidal volume variability, and transcutaneous oxygen saturation of hemoglobin; combine at least three extracted parameters from cross-information, ECG, respiratory acoustics, chest motion, temperature, and transcutaneous oxygen saturation of hemoglobin into a respiratory quality level index using an adaptive neurofuzzy inference system or any other classifier; and combine at least three extracted parameters from cross-information, ECG, cardiac acoustics, and pulse wave conduction time using an adaptive neurofuzzy inference system to determine noninvasive blood pressure.

Description

Technical Field

[0001] Continuous monitoring of respiratory rate, heart rate, arterial blood pressure, oxygen saturation, and temperature are fundamental parameters for reducing morbidity and mortality in hospitalized patients and those in primary care settings. Several algorithms have been developed using these parameters to identify patients requiring urgent medical attention.

[0002] This invention generally relates to an apparatus and method for determining a subject's respiratory quality and effort, respiratory rate (RespRate), heart rate, oxygen saturation, temperature, and continuous noninvasive blood pressure. More specifically, this invention relates to a wired or wireless device and method for obtaining an index representing the probability of respiratory deterioration, referred to as the Smart Respiratory Index (SRI). Background Technology

[0003] When acutely ill patients require hospitalization, or when a patient's condition acutely deteriorates in the hospital, evidence suggests that early detection and a rapid and effective clinical response are crucial for improving patient outcomes. Disease severity can be quantified by measuring a combination of simple physiological parameters such as respiratory rate, oxygen saturation, temperature, arterial blood pressure, and pulse rate. In cases of acute illness in a hospital setting, this combination of parameters is a good predictor of patient mortality and length of hospital stay.

[0004] In this sense, breathing is fundamental to life. The lungs are responsible for breathing, which is the process of supplying the body with oxygen and removing carbon dioxide from the body.

[0005] Respiratory distress is caused by oxygenation failure (insufficient oxygen intake) or ventilation failure (insufficient carbon dioxide removal).

[0006] These forms of respiratory failure can be characterized by irregular respiratory rates and / or abnormal breathing patterns, accompanied by low hemoglobin oxygen saturation (SpO2), which can lead to organ and tissue hypoxia. Continuous monitoring of SpO2 and respiratory rate (RespRate) and their patterns is crucial for detecting the onset of respiratory failure. When SpO2 falls below 90%, the risk of hypoxia increases very rapidly. The time available to correct this change is very short. This is why monitoring SpO2 is so important. However, not all patients admitted to general wards are monitored with SpO2 monitors. Furthermore, most hospital wards fail to properly monitor RespRate, and when this vital sign is measured, it is often done at a low frequency. Patients in general wards are typically checked only every six to eight hours, and there is often significant human error because nurses often observe a patient's breathing for 15 seconds and then multiply the result by 4 to obtain the RespRate per minute. The World Health Organization recommends counting the number of breaths in a full minute. If the RespRate is above or below the acceptable range of 12-16 breaths per minute, it indicates a breathing problem.

[0007] Current devices for measuring SpO2 and RespRate are very expensive and are connected to patients via cables, so these devices are only used in special departments such as intensive care units (ICUs).

[0008] Among patients with a RespRate between 25 and 29 breaths per minute, 21% died in the hospital. The best way to alert patients to their clinical condition and reduce complications is through continuous monitoring of RespRate and SpO2. The coronavirus pandemic (Covid-19) has clearly demonstrated the necessity of respiratory monitoring outside of hospital settings, in nursing homes, and even in home care settings.

[0009] Joseph et al.'s U.S. patent application, "Acoustic sensor and ventilation monitoring system" (US 2020 / 0054277 A1), discloses a method for monitoring respiration using an acoustic measurement device. After a detailed description of respiratory pathophysiology, the authors propose a device integrating two components. One component is attached to the patient, while the second component is attached to the first. The first component consists of a sound transducer, an accelerometer, and a transmitter. The second component consists of a rechargeable battery. Several sensors can be integrated into the device to measure temperature, heart rate, and oxygen saturation. The device can also be connected to a smartwatch. Joseph et al., utilizing information from different sensors, proposed a risk factor index to determine respiratory function in mammals due to pathophysiological changes and drug or alcohol abuse. Joseph et al. described a quadratic equation for calculating their respiratory risk factor index. The main differences between this patent application and that of Joseph et al. are, firstly, that Joseph et al. only used an accelerometer to assess body movement, while this patent application uses an accelerometer to estimate respiratory rate; and secondly, that the formulas used for the respiratory risk index and the smart respiratory index are significantly different. Furthermore, in the patent application by Joseph et al., respiratory rate is estimated by respiratory acoustics, while in this patent application, respiratory rate is estimated by thoracic acceleration measured by an accelerometer, the first derivative of the respiratory sound envelope, and the variation of the RR interval of the ECG.

[0010] Both Joseph's patent application and this patent application use sensors such as microphones and accelerometers to assess breathing sounds and movement.

[0011] However, the parameters analyzed in the two works and the derivation formula used by Joseph et al. to calculate the risk factor index are completely different from the SRI in this disclosure. Joseph et al. used accelerometers to assess patients' body movements, while in this disclosure, accelerometers are used to calculate respiratory rate.

[0012] Joseph et al. used breath sounds to calculate respiratory rate and TV, while in this disclosure, the envelope is used to assess the variability of TV.

[0013] Finally, for Joseph et al., physical movement and verbal accumulation received the highest scores, while bradykinesia and lack of movement had the lowest possible scores.

[0014] US Patent Application No. 2015 / 0313484 A1, entitled "Portable device with multiple integrated sensors for vital sign scanning," discloses a portable device with multiple integrated sensors. This patent application differs from it because it does not use photoplethysmography (PPG) and defines a respiratory quality index.

[0015] The U.S. patent application “Mobile frontend system for comprehensive cardiac diagnosis”, US2015 / 0065814 A1, is significantly different from this patent application because the purpose of US2015 / 0065814 is the comprehensive diagnosis of cardiac problems.

[0016] The U.S. patent application “Mesh network personal emergency appliance”, US 2008 / 0001735 A1, discloses a system including one or more wireless nodes forming a wireless mesh network. This application differs in that it does not form a wireless mesh network.

[0017] The U.S. patent application “Monitoring, predicting and treating clinical episodes”, US2008 / 0275349 A1, discloses a device for sensing physiological parameters of a subject and sensing large-amplitude body movements, which is significantly different from this patent application, which does not disclose sensing large-amplitude body movements.

[0018] The U.S. patent application "Physiological acoustic monitoring system" (US8821415B2) discloses a method for assessing respiratory rate using acoustic signals. However, that application only uses one or two acoustic sensors (microphones), so this application is completely different because it also uses an electrocardiogram (ECG) and an accelerometer.

[0019] US Patent 6918878 B2 discloses a method for determining a patient's respiratory rate, comprising several parts. Respiratory rate can be determined by measuring the S2 split of the heart.

[0020] S2 splitting can be identified by observing the timing of heart sounds. Other respiratory-related information, such as respiratory phase and the occurrence of apnea, can also be identified. This type of respiratory monitor can be used to monitor subacute and outpatient patients. The sensor for the respiratory monitor and the electrode for the ECG monitor can be combined into a single probe.

[0021] This patent application does not include the S2 split for respiratory rate estimation, and is therefore different from the US 6918878 patent application.

[0022] US Patent 2018 / 0214090 A1, entitled "System and method for monitoring respiratory rate measurements," discloses a system and method for determining multi-parameter confidence in respiratory rate measurements using multiple physiological parameter inputs. This disclosure differs from this patent application because it combines respiratory acoustics and ECG RR interval changes (R peaks).

[0023] PCT application WO2020 / 071926 A1 discloses a sensor for monitoring numerous parameters, while this application differs in that it includes tidal volume variability, pulse wave conduction time, and uses an adaptive neurofuzzy inference system to classify the index level of respiratory quality and determine non-invasive blood pressure, thereby improving the health monitoring of the device.

[0024] US Patent “Lung function monitoring from heart signals”, US 2022 / 0167856 A1 uses a bioimpedance sensor, while this application does not use such a sensor.

[0025] US Patent 2011 / 0224565 A1, entitled "Method of predicting acute cardiopulmonary events and survivability of a patient," provides a method for generating an artificial network capable of predicting patient survival, which is not included in this application.

[0026] PCT application WO2019 / 036805 A1 discloses a method and system for activity classification using multiple sensors such as GPS and video cameras, which are not included in this application.

[0027] Canadian Patent Application No. 2 811 326, “System for early detection of life-threatening conditions of persons,” discloses a system for detecting surgical bleeding, which is not included in any embodiment of this application. Summary of the Invention

[0028] The present invention is a patch (1) comprising at least sensors for electrocardiogram (ECG), respiratory rate (RespRacc) measured by an accelerometer, respiratory rate (RespR) measured by a microphone, oxygen saturation (SatO2), and temperature (t), and attached to the patient’s chest. The ECG (2) is further processed using a heart rate (HR) extraction algorithm, such as, but not necessarily, a fast Fourier transform (FFT) (11), to extract the HR; the QRS amplitude (QRSv) and RRv (11) of the ECG are also determined by FFT (spectral analysis).

[0029] Breathing sounds from the microphone (3) are subjected to envelope waveform breathing extraction formula (12) to obtain the respiratory rate (called RespR). Formula (13) is used to calculate tidal volume variability (TVv). The acceleration signal of the thoracic cavity (4) is analyzed by a Hilbert Transform model, which estimates the breathing rate called RespRacc. As breathing deteriorates, the relationship between the four parameters HRv, QRSv, RespR, and RespRacc may change, so cross-mutual information (6) is calculated and the variable CMIbreath is generated. The parameters extracted from the measurements are fed into a classifier, which may be, but is not necessarily, an Adaptive Neural Fuzzy Inference System (ANFIS) (16). The output of the classifier is called the Smart Breath Index (SRI) (17). The patch is made by Figure 4The hardware components shown include electrodes attached to the patient (22, 26, 32). ECG (27), SpO2 (24), microphone (25), temperature (31), and acceleration (23) signals are processed by a microprocessor (29) and then transmitted to an external display via Bluetooth Low Energy (BLE) (33) or UART / USB (34). The battery (30) is charged via USB (34).

[0030] RespR is estimated using the envelope formula.

[0031] like Figure 5 As shown, the acoustic signal recorded from the microphone is input into a spline function or other curve function that allows for the evaluation of the amplitude envelope. See also Figure 4 In the example, RespR can be calculated by counting the peaks in the Aenvelope curve.

[0032] Estimation of tidal volume variability (TVv)

[0033] Previously, the relationship between respiratory airflow F and the energy E of respiratory (tracheal) sound was optimally expressed as follows: A power-law fit is obtained, where k and α are constants, and different research groups have proposed different values ​​for this exponent. This sound amplitude-airflow relationship has been used for respiratory monitoring, particularly for qualitative and quantitative assessment of respiratory airflow and for estimation of continuous respiratory rate.

[0034] However, we found that the estimation can be improved by adding the derivative of the envelope of the breathing sound to the equation, and therefore the flow equation is defined as follows:

[0035]

[0036] Therefore, the volume can be estimated as the integral of the airflow F over time during the inhalation period.

[0037]

[0038] In this paper, tidal volume variability is defined as the change over time. For example, if the tidal volume increases from 10 to 12, the tidal volume variability is 20%.

[0039] RespRacc is determined by applying the Hilbert transform to the acceleration signal.

[0040] A novel method for extracting respiration from acceleration signals using Hilbert vibrational decomposition (HVD) is proposed. It is shown that the maximum energy component of the acquired acceleration signal is proportional to the respiration signal.

[0041] Determination of the Smart Respiratory Index (SRI)

[0042] In one implementation, SRI is a linear or quadratic function of RespRacc, HR, HRv, QRSv, TVv, and SpO2:

[0043] SRI=k1*RespRacc+k2*HR+k3*HRv+k4*QRSv+k5*TVv+k6*SpO2+k7* RespR*HR*RRV*QRSv*TVv*SpO2,

[0044] Among them, the constants k1 to k7 should be within the following range:

[0045] 0.30 < k1 < 0.7,

[0046] 0.2 < k2 < 0.4,

[0047] 0.1 < k3 < 0.3,

[0048] 0.01 < k4 < 0.2,

[0049] 0.01 < k5 < 0.2,

[0050] 0.1 < k6 < 0.7,

[0051] 0.01 < k7 < 1.

[0052] SRI is defined by combining parameters using a classifier.

[0053] In the second embodiment, such as Figure 1 As shown, the device uses a classifier, such as, but not necessarily, an ANFIS model, to combine parameters to define the SRI. Parameters extracted from at least four sensors (ECG, respiratory sounds, thoracic acceleration, and SpO2) are used as input to the Adaptive Neural Fuzzy Inference System (ANFIS).

[0054] In the third embodiment, such as Figure 1 As shown, the device uses an ANFIS model to combine parameters to define the SRI. Parameters extracted from at least four sensors (ECG, respiratory sounds, chest acceleration, and pulse oximeter) are used as inputs to the Adaptive Neural Fuzzy Inference System (ANFIS).

[0055] In the fourth embodiment, such as Figure 1As shown, the device uses an ANFIS model to combine parameters to define the SRI. Parameters extracted from at least four sensors (ECG, respiratory sounds, chest acceleration, pulse oximeter) along with patient demographic data (age, sex, height, weight) and patient clinical data (chronic obstructive pulmonary disease, asthma, sympathetic nervous system disorders, atrial fibrillation, beta-blockers, pacemaker) are used as inputs to the Adaptive Neurofuzzy Inference System (ANFIS).

[0056] Assessment of continuous non-invasive blood pressure.

[0057] The time between the R peak and the first heart sound on an electrocardiogram (ECG) is called the pulse wave conduction time (PWTT). When blood pressure rises, PWTT decreases due to increased vascular resistance. Conversely, decreased arterial blood pressure prolongs PWTT. Therefore, there is a negative correlation between arterial blood pressure and PWTT. Cross-information between heart rate variability and heart sound time series improves the accuracy of arterial blood pressure estimation. PWTT, cross-information between heart rate variability and heart sound time series, patient sex, age, and BMI were used as inputs to the ANFIS model.

[0058] ANFIS Overview

[0059] ANFIS is a hybrid of fuzzy logic systems and neural networks, which does not assume any mathematical function governing the relationship between inputs and outputs. ANFIS employs a data-driven approach, where training data determines the system's behavior.

[0060] The five layers of ANFIS have the following functions:

[0061] Each cell in layer 1 stores three parameters to define a bell-shaped membership function. Each cell is connected to exactly one input cell and calculates the membership degree of the obtained input value.

[0062] Each rule is represented by a unit in the second layer. Each unit is connected to those units in the previous layer that are derived from the antecedents of that rule. The input to a unit is the membership degree, which is multiplied to determine the degree to which the represented rule is satisfied.

[0063] In the third layer, for each rule, there exists a unit that calculates its relative satisfaction using a normalized equation. Each unit is connected to all rule units in the second layer.

[0064] The cells in layer 4 are connected to all input cells and to exactly one cell in layer 3. Each cell computes the output of one rule.

[0065] One of the output units in the 5th layer calculates the final output by summing all the outputs from the 4th layer.

[0066] The standard learning procedure from neural network theory is applied to ANFIS. Backpropagation is used to learn the antecedent parameters, i.e., membership functions, and least squares estimation is used to determine the coefficients of the linear combination in the consequent of the rules. One step in the learning procedure has two pass-through processes. In the first pass-through (forward pass), the input pattern is propagated, and the optimal consequent parameters are estimated by iterative least mean squares estimation, while the antecedent parameters are fixed in the current loop through the training set. In the second pass-through (backward pass), the pattern is propagated again, and in this pass-through, backpropagation is used to modify the antecedent parameters, while the consequent parameters remain fixed. The procedure is then iterated for the desired number of training epochs. If the antecedent parameters are appropriately chosen initially based on expert knowledge, one training epoch is usually sufficient, because the LMS algorithm determines the optimal consequent parameters in one pass-through, and if the antecedent is not significantly changed by using gradient descent, the LMS calculation of the consequent will not lead to another result. For example, in a two-input, two-rule system, the rule is defined as follows:

[0067] If x is A and y is B, then f1 = p1x + q1y + r1.

[0068] Where p, q, and r are linear, they are called consequent parameters or simply consequents. The most common is first-order f, because higher-order Sugeno fuzzy models introduce significant complexity with little apparent advantage.

[0069] Number of categories.

[0070] The input to the ANFIS system is fuzzified into multiple predetermined categories. The number of categories should be greater than or equal to two. The number of categories can be determined by different methods. In traditional fuzzy logic, the categories are defined by experts. This method can only be applied if it is obvious to the experts where the boundary between two categories can be placed. ANFIS optimizes the placement of the boundary; however, if the initial values ​​of the parameters defining the categories are close to the optimal values, the gradient descent method will reach its minimum faster. By default, ANFIS initial boundary is chosen by dividing the interval from the minimum to the maximum value of all data into n equidistant intervals, where n is the number of categories. The number of categories can also be chosen by plotting the data as a histogram and visually determining the appropriate number of categories, by sorting as done by fuzzy inductive reasoning (FIR), by various clustering methods, or by Markov models. The default ANFIS setting was chosen for this invention, and it was shown that more than three categories would lead to instability during the validation phase; therefore, two or three categories were used.

[0071] Enter the number.

[0072] Both the number of categories and the number of inputs increase the complexity of the model, i.e., the number of parameters. For example, a system with 4 inputs, each of which is fuzzified into 3 categories, consists of 36 antecedent (nonlinear) parameters and 405 consequent (linear) parameters, calculated using the following two formulas:

[0073] Antecedent = Number of categories × Number of inputs × 3

[0074] Successor = Number of Categories 输入数 × (input number + 1)

[0075] The number of input-output pairs should typically be much larger than the number of parameters (at least 10 times the number) in order to obtain meaningful solutions for the parameters.

[0076] Stability standards.

[0077] Unfortunately, there is no defined stability criterion for neural fuzzy systems. The most useful tools for ensuring stability are the experience gained from working with a neural fuzzy system (such as ANFIS) in the context of a specific dataset, and testing with extreme data, such as data obtained through simulation.

[0078] Number of training epochs.

[0079] ANFIS uses the root mean square error (RMSE) to validate training results, and the RMSE validation error can be calculated after each training epoch based on a set of validation data. A training epoch is defined as updating both the antecedent and consequent parameters. Increasing the number of training epochs generally reduces the training error. Attached Figure Description

[0080] Figure 1 The extracted parameters are fed into ANFIS, a hybrid of neural networks and fuzzy logic systems. The input consists of at least three of the following parameters: HR, RRv, QRSv, RespR, TVv, RespRacc, pulse oximetry, cross-information between HR, RRv, and RespRacc (CMIbreath), and demographic data such as sex, age, and body mass index (BMI). The output of the ANFIS model is the Smart Respiratory Index (SRI), a unitless number from 0 to 100, where decreasing values ​​correspond to a deterioration in the patient's respiratory function.

[0081] Figure 2The extracted parameters are fed into ANFIS, a hybrid of neural networks and fuzzy logic systems. Inputs include PWTT, cross-information between R-position sequences and heart sound locations, and demographic data such as sex, age, and body mass index (BMI). The output of the ANFIS model is mean arterial pressure (MAP).

[0082] Figure 3 The diagram shows how the breathing patch is attached to the subject's chest and its location. The monitor is connected via Bluetooth.

[0083] Figure 4 The patch consists of an amplifier for ECG, an accelerometer, a microphone, a temperature sensor, a saturation sensor, a radio transmitter module (such as a Bluetooth Low Energy module), a USB-C port, a battery, and three electrodes that are also used to attach the patch to the patient.

[0084] Figure 5 This figure shows a Core Safe device with sensors.

[0085] Figure 6 This figure shows the pulse wave conduction time scheme and heart sounds used to estimate mean arterial pressure.

[0086] Figure 7 This table shows the relationship between clinical status and the Smart Respiratory Index (SRI). The SRI is a progressive scale, where 100 corresponds to normal respiratory function, while a decreasing value reflects a decline in respiratory function, 75 corresponds to mild deterioration, 50 corresponds to severe deterioration, and 0 is presented when respiratory arrest occurs.

Claims

1. A device configured to determine respiratory quality and noninvasive blood pressure in neonates, infants, and adult patients by combining parameters extracted from electrocardiogram, chest and cardiac motion and acceleration, respiratory acoustics, cardiac acoustics, temperature, and transcutaneous oxygen saturation of hemoglobin. Its characteristics include the following steps: a) Using sensors to measure electrocardiograms; b) Use sensors to measure chest wall and heart movement and acceleration; c) Using sensors to measure respiratory and cardiac acoustics; d) Use a sensor to measure temperature; e) Measure transcutaneous oxygen saturation of hemoglobin using a sensor positioned on the patient's upper thoracic cavity; f) Calculate heart rate variability and QRS amplitude variability based on the electrocardiogram using the Choi-Williams distribution or fast Fourier transform; g) Calculate the respiratory rate based on chest wall movement and acceleration, and based on the RR variability of the electrocardiogram; h) Calculate tidal volume variability based on the amplitude of the respiratory acoustics and the first derivative of the amplitude envelope of the respiratory acoustics; i) Calculate the pulse wave conduction time based on the time difference between the R peak of the electrocardiogram and the first heart sound of the cardiac acoustics; j) Calculate the cross-information between the heart rate variability, respiratory acoustics, chest wall movement, tidal volume variability, and transcutaneous oxygen saturation of hemoglobin in the electrocardiogram. k) Using an adaptive neurofuzzy inference system or any other classifier, combine at least three extracted parameters from cross-mutual information, electrocardiogram, respiratory acoustics, chest movement, temperature, and transcutaneous oxygen saturation of hemoglobin into a respiratory quality level index. l) Using an adaptive neurofuzzy inference system, at least three extracted parameters from cross-mutual information, electrocardiogram, cardiac acoustics, and pulse wave conduction time are combined to determine non-invasive arterial blood pressure.

2. The apparatus according to claim 1, wherein, Step a is characterized by: measurement by a patch consisting of two or more electrodes to determine an electrocardiogram located on the upper or lower thorax of the subject.

3. The apparatus according to claim 1, wherein, Step b is characterized by recording thoracic acceleration using an accelerometer integrated in the patch in order to calculate respiratory rate based on the acceleration.

4. The apparatus according to claim 1, wherein, Step c is characterized by recording respiratory acoustics, inhalation, and exhalation using one or two microphones integrated into the patch in order to calculate the respiratory rate.

5. The apparatus according to claim 1, wherein, Step h is characterized by estimating the flow rate using the following formula: ， Where F is the flow rate and A is the amplitude of the respiratory acoustics, therefore, the tidal volume variability (TVv) is the integral of the flow rate over time: 。 6. The apparatus according to claim 1, wherein, Step g is characterized in that the breathing frequency is calculated by applying a Kalman filter to the acceleration signal to extract the breathing frequency from the acceleration acquired by the accelerometer.

7. The apparatus according to claim 1, wherein, Step k is characterized in that: heart rate variability, QRS amplitude variability extracted from electrocardiogram, tidal volume variability, cross-information, and transcutaneous oxygen saturation of hemoglobin calculated from respiratory acoustics are used as inputs to an adaptive neurofuzzy inference system or another classifier, the output of which is the respiratory quality index.

8. The apparatus according to claim 1, wherein, Step j is characterized in that the formula for the respiratory quality index is a linear or quadratic function of respiratory rate, heart rate, heart rate variability, QRS variability, tidal volume variability, and transcutaneous oxygen saturation of hemoglobin. SRI=k1*RespRacc+k2*HR+k3*HRv+k4*QRSv+k5*TVv+k6*SpO2+k7* RespR*HR*RRV*QRSv*TVv*SpO2, Among them, the constants k1 to k7 should be within the following range: 0.30 ＜ k1 ＜ 0.7, 0.2 ＜ k2 ＜ 0.4, 0.1 ＜ k3 ＜ 0.3, 0.01 ＜ k4 ＜ 0.2, 0.01 ＜ k5 ＜0.2, 0.1＜k6＜0.7, 0.01 ＜ k7 ＜ 1。 9. The apparatus according to claim 1, wherein step 1 is characterized by using an adaptive neural fuzzy inference system to combine at least three extracted parameters from cross-mutual information, electrocardiogram, cardiac acoustics, and pulse wave conduction time to determine non-invasive blood pressure, changes in non-invasive blood pressure, and trends.

10. The apparatus of claims 1 to 9, wherein the apparatus is integrated in a wired or wireless patch, the patch comprising interconnected electrocardiogram sensors and amplifiers, one or two microphones, an accelerometer, a transcutaneous oxygen saturation sensor for hemoglobin, a temperature sensor, a battery, and a radio transmitter, the radio transmitter being, for example, a Bluetooth Low Energy (BLE) module, which transmits data to an external receiver and display device having a central processing unit and a keyboard.

11. A method for determining respiratory quality and noninvasive blood pressure in neonates, infants, and adult patients by using the apparatus according to claim 1, combining parameters extracted from electrocardiogram, chest and cardiac motion and acceleration, respiratory acoustics, cardiac acoustics, temperature, and hemoglobin transcutaneous oxygen saturation. Its characteristics include the following steps: a) Using sensors to measure electrocardiograms; b) Use sensors to measure chest wall and heart movement and acceleration; c) Using sensors to measure respiratory and cardiac acoustics; d) Use a sensor to measure temperature; e) Measure transcutaneous oxygen saturation of hemoglobin using a sensor positioned on the patient's upper thoracic cavity; f) Calculate heart rate variability and QRS amplitude variability based on the electrocardiogram using the Choi-Williams distribution or fast Fourier transform; g) Calculate the respiratory rate based on chest wall movement and acceleration, and based on the RR variability of the electrocardiogram; h) Calculate tidal volume variability based on the amplitude of the respiratory acoustics and the first derivative of the amplitude envelope of the respiratory acoustics; i) Calculate the pulse wave conduction time based on the time difference between the R peak of the electrocardiogram and the first heart sound of cardiac acoustics; j) Calculate the cross-information between the heart rate variability, respiratory acoustics, chest wall movement, tidal volume variability, and transcutaneous oxygen saturation of hemoglobin in the electrocardiogram. k) Using an adaptive neurofuzzy inference system or any other classifier, combine at least three extracted parameters from the cross-mutual information, electrocardiogram, respiratory acoustics, chest movement, temperature, and transcutaneous oxygen saturation of hemoglobin into a respiratory quality level index. l) Using an adaptive neurofuzzy inference system, at least three extracted parameters from cross-mutual information, electrocardiogram, cardiac acoustics, and pulse wave conduction time are combined to determine noninvasive blood pressure.

12. The method according to claim 11, wherein, Step h is characterized by estimating the flow rate using the following formula: Where F is the flow rate and A is the amplitude of the respiratory acoustics, therefore, the tidal volume variability (TVv) is the integral of the flow rate over time: 。 13. The method according to claim 11, wherein, Step g is characterized in that the breathing frequency is calculated by applying a Kalman filter to the acceleration signal to extract the breathing frequency from the acceleration acquired by the accelerometer.

14. The method according to claim 11, wherein, Step k is characterized in that: heart rate variability, QRS amplitude variability extracted from electrocardiogram, tidal volume variability, cross-information, and transcutaneous oxygen saturation of hemoglobin calculated from respiratory acoustics are used as inputs to an adaptive neurofuzzy inference system or another classifier, the output of which is the respiratory quality index.

15. The method according to claim 11, wherein, Step j is characterized in that the formula for the respiratory quality index is a linear or quadratic function of respiratory rate, heart rate, heart rate variability, QRS variability, tidal volume variability, and transcutaneous oxygen saturation of hemoglobin. SRI=k1*RespRacc+k2*HR+k3*HRv+k4*QRSv+k5*TVv+k6*SpO2+k7*RespR*HR*RRV*QRSv*TVv*SpO2, Among them, the constants k1 to k7 should be within the following range: 0.30 ＜ k1 ＜ 0.7, 0.2 ＜ k2 ＜ 0.4, 0.1 ＜ k3 ＜ 0.3, 0.01 ＜ k4 ＜ 0.2, 0.01 ＜ k5 ＜0.2, 0.1＜k6＜0.7, 0.01 ＜ k7 ＜ 1。 16. The method according to claim 11, wherein step 1 is characterized by using an adaptive neural fuzzy inference system to combine at least three extracted parameters from cross-mutual information, electrocardiogram, cardiac acoustics, and pulse wave conduction time to determine non-invasive blood pressure, changes in non-invasive blood pressure, and trends.

17. The method according to claims 11 to 16, wherein the method is integrated into a wired or wireless patch, the patch comprising interconnected electrocardiogram sensors and amplifiers, one or two microphones, an accelerometer, a transcutaneous oxygen saturation sensor for hemoglobin, a temperature sensor, a battery, and a radio transmitter, the radio transmitter being, for example, a Bluetooth Low Energy (BLE) module, which transmits data to an external receiver and display device having a central processing unit and a keyboard.

Images