Medical ventilation apparatus and ventilation monitoring method

Abstract

A medical ventilation device and a ventilation monitoring method, the ventilation monitoring method comprising: acquiring at least one parameter data capable of representing a patient's occurrence of a human-machine confrontation event, the parameter data including at least one type of device ventilation parameter data and patient physiological parameter data; enabling an algorithm module associated with the at least one parameter data to obtain an identification result of the human-machine confrontation event of the patient in the ventilation process, wherein the algorithm module includes multiple algorithm modules, and different algorithm modules perform calculations based on different data combinations formed by the at least one parameter data to obtain the identification result; determining the human-machine confrontation event of the patient according to the identification results obtained by the different algorithm modules respectively; and outputting the human-machine confrontation event of the patient. The ventilation monitoring method applied to the medical ventilation device can more accurately determine whether a human-machine confrontation event occurs.

Description

Technical Field

[0001] This invention relates to the field of medical devices, specifically to medical ventilation equipment and ventilation monitoring methods. Background Technology

[0002] Patient-ventilator asynchrony, also known as human-ventilator asynchrony, refers to the asynchrony between the delivery of air by the ventilation equipment (such as a ventilator) and the patient's spontaneous breathing during mechanical ventilation. Patient-ventilator asynchrony can affect the patient's condition, causing discomfort, such as increased work of breathing and prolonged on-the-go ventilation time. Severe patient-ventilator asynchrony can lead to further lung damage and increased mortality. Therefore, it is crucial to identify patient-ventilator asynchrony during mechanical ventilation and to alert healthcare professionals to this phenomenon.

[0003] Currently, the method for identifying patient-ventilator asynchrony involves real-time monitoring of the patient's ventilation waveform and automatic identification of asynchrony phenomena based on waveform characteristics. The monitored ventilation waveforms typically include three parameters: airway pressure, flow rate, and volume. However, this method relies on a fixed monitoring object, resulting in drawbacks such as insufficient accuracy and limited types of asynchrony that can be identified. Furthermore, the identification results are easily affected by waveform signal strength and monitoring quality, hindering accurate identification of asynchrony. Summary of the Invention

[0004] According to a first aspect, the present invention provides a medical ventilation device, comprising:

[0005] Gas source interface, used to connect to a gas source;

[0006] The patient interface is used to connect to the patient's respiratory system;

[0007] A breathing circuit is used to connect the gas source interface and the patient interface to deliver gas supplied by the gas source to the patient.

[0008] A respiratory support device is used to provide respiratory support power to control the delivery of gas supplied by a gas source to the patient;

[0009] A processor is configured to acquire at least one type of parameter data capable of characterizing a patient-machine aggression event, the parameter data including at least one type of device ventilation parameter data and patient physiological parameter data;

[0010] The processor is further configured to, upon receiving the parameter data, activate an algorithm module associated with the parameter data to obtain the identification result of the patient's human-ventilator aggression event during ventilation; wherein, the algorithm module includes multiple modules, and different algorithm modules perform calculations based on different data combinations formed by at least one parameter data to obtain the identification result of the patient's human-ventilator aggression event during ventilation; the identification result includes: the human-ventilator aggression event is identified or the human-ventilator aggression event is not identified;

[0011] And based on the recognition results obtained by different algorithm modules, determine and output the human-machine confrontation event that occurred to the patient.

[0012] According to a second aspect, the present invention provides a ventilation monitoring method, comprising the steps of:

[0013] Acquire at least one type of parameter data that can characterize a patient experiencing a patient-machine aggression event, the parameter data including at least one type of device ventilation parameter data and patient physiological parameter data;

[0014] An algorithm module associated with the at least one parameter data is activated to obtain the identification result of the patient's human-ventilator aggression event during ventilation. The algorithm module includes multiple modules, and different algorithm modules perform calculations based on different data combinations formed by the at least one parameter data to obtain the identification result. The identification result includes: the human-ventilator aggression event is identified or the human-ventilator aggression event is not identified.

[0015] Based on the recognition results obtained by each of the different algorithm modules, the human-machine confrontation event that occurred to the patient is determined;

[0016] Output the human-machine interaction events that occurred with the patient.

[0017] According to a third aspect, the present invention provides a computer-readable storage medium, characterized in that it includes a program that can be executed by a processor to implement the method described in any of the preceding aspects.

[0018] In the above embodiments, corresponding algorithm modules are activated based on at least one parameter data, and the patient-ventilator asynchrony events occurring during ventilation are determined by comprehensively considering the recognition results of each algorithm module. Compared with existing methods:

[0019] (1) The output human-machine confrontation events are based on at least one parameter data, so more types of human-machine confrontation can be identified.

[0020] (2) Different algorithm modules can use different algorithms to identify human-machine confrontation, and then combine the identification results of multiple algorithm modules to confirm the human-machine confrontation event, so that the identification accuracy of the human-machine confrontation event is higher and the robustness is better. Attached Figure Description

[0021] Figure 1 A schematic diagram of a medical ventilation device according to one embodiment;

[0022] Figure 2 This is a schematic diagram of an algorithm module and its data combination according to one embodiment;

[0023] Figure 3 This is a waveform diagram of the corresponding parameter data when an invalid triggering event occurs, according to one embodiment.

[0024] Figure 4 for Figure 3 Waveform characteristics of airway pressure and airway flow at the moment of occurrence of invalid triggering event;

[0025] Figure 5 This is a waveform diagram of the corresponding parameter data when a dual-trigger event occurs in one embodiment.

[0026] Figure 6 This is a waveform diagram of the corresponding parameter data when an invalid triggering event occurs, according to another embodiment.

[0027] Figure 7 This is a waveform diagram of the corresponding parameter data when an invalid triggering event occurs in another embodiment;

[0028] Figure 8 This is a schematic diagram of a judgment module corresponding to various types of human-machine confrontation events in one embodiment;

[0029] Figure 9 This is a schematic diagram illustrating the marking of human-machine adversarial events on a waveform according to one embodiment.

[0030] Figure 10 This is an embodiment of a display interface for displaying prompt information and human-machine confrontation statistics;

[0031] Figure 11 A flowchart of a ventilation monitoring method according to one embodiment;

[0032] 10. Gas source interface;

[0033] 20. Respiratory assist device;

[0034] 30. Breathing circuit; 30a. Inspiratory pathway; 30b. Expiratory pathway;

[0035] 31. Carbon dioxide receiver; 32. One-way valve;

[0036] 40. Patient interface;

[0037] 50. Processor;

[0038] 60. Parameter measuring device;

[0039] 70. Monitor. Detailed Implementation

[0040] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of this application. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to this application are not shown or described in the specification. This is to avoid obscuring the core parts of this application with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0041] Furthermore, the features, operations, or characteristics described in the specification can be combined in any suitable manner to form various embodiments. At the same time, the steps or actions in the method description can be rearranged or adjusted in a manner obvious to those skilled in the art. Therefore, the various orders in the specification and drawings are only for the clear description of a particular embodiment and do not imply a necessary order, unless otherwise stated that a particular order must be followed.

[0042] The serial numbers assigned to components in this document, such as "first" and "second," are used only to distinguish the described objects and have no sequential or technical meaning. The terms "connection" and "linkage" used in this application, unless otherwise specified, include both direct and indirect connections (linkages).

[0043] Currently, existing technologies for identifying human-machine aggression primarily focus on recognizing aggression characteristics in terms of airway pressure or flow rate. However, judging solely from pressure or flow rate cannot identify false triggers or reverse triggers. Furthermore, interference in the flow rate or pressure waveform can affect the accuracy of identifying invalid triggers and premature switching events. If transdiaphragmatic pressure and diaphragmatic electrophysiological signals are primarily used to determine human-machine aggression, the results are highly susceptible to the quality of signal monitoring. For example, transdiaphragmatic pressure monitoring requires esophageal and intragastric pressure measuring accessories, which must be placed in appropriate locations within the patient's body. The accuracy and stability of pressure monitoring results heavily depend on the measuring accessories themselves and the adherence to standardized operating procedures. During signal monitoring, the intensity of signal noise, involuntary swallowing by the patient, or changes in body position can easily lead to misinterpretations. Similarly, diaphragmatic electrophysiological monitoring also requires appropriate measuring accessories and placement procedures. If the patient's spontaneous breathing ability is weak, the diaphragmatic electrophysiological signal intensity will be low, and the signal periodicity will be unclear, resulting in poor signal performance for identifying human-machine aggression. Furthermore, relying solely on transdiaphragmatic pressure or diaphragmatic electrical activity cannot identify human-human interaction events such as low patient flow rate or slow pressure rise time.

[0044] The parameter data referred to in this invention includes analog data and digital data, which can be structured data or unstructured data. The signal referred to below is also included in the parameter data referred to in this invention.

[0045] The feature extraction referred to in this invention includes the process of calculating signals or data. This calculation includes performing operations on the signals to obtain new intermediate signals. Possible methods include basic mathematical operations, such as signal addition and subtraction, as well as various forms of mathematical calculations. Extracting new parameters from the signals to assist in human-machine adversarial recognition, such as calculating the mean, variance, standard deviation, central moments of each order, and standard moments of each order, also falls within the scope of feature extraction. Performing various preprocessing or postprocessing on the signals, such as filtering and normalization, also falls within the scope of feature extraction.

[0046] The carbon dioxide module referred to in this invention is a module used to monitor carbon dioxide concentration. Its monitoring principle is mostly based on infrared absorption spectroscopy technology, which uses the relationship between carbon dioxide concentration in the gas and absorption rate to calculate the carbon dioxide concentration in the patient's exhaled gas.

[0047] Please refer to Figure 1 The embodiment shown provides a medical ventilation device (e.g., a ventilator, anesthesia machine, etc.) which includes an air source interface 10, a respiratory assist device 20, a breathing circuit 30, a patient interface 40, and a processor 50.

[0048] The gas source interface 10 is used to connect to a gas source (not shown in the figure) to provide gas. This gas can typically be oxygen or air. In some embodiments, the gas source can be a compressed gas cylinder or a central gas supply source, supplying gas to the medical ventilation equipment through the gas source interface 10. The supplied gas types include oxygen (O2) and air. The gas source interface 10 may include conventional components such as a pressure gauge, pressure regulator, flow meter, pressure reducing valve, and proportional control protection device, used to control the flow rate of various gases (e.g., oxygen and air). The gas input into the gas source interface 10 enters the breathing circuit 30 and mixes with the existing gas in the breathing circuit 30 to form a gas mixture.

[0049] The respiratory assist device 20 is used to power the patient's involuntary breathing and maintain airway patency. It drives the gas input from the gas source interface 10 and the mixed gas in the breathing circuit 30 into the patient's respiratory system, and guides the patient's exhaled air into the breathing circuit 30, thereby improving ventilation and oxygenation and preventing hypoxia and carbon dioxide accumulation in the patient's body. In a specific embodiment, the respiratory assist device 20 typically includes a mechanical ventilation module, whose airflow channel is connected to the breathing circuit 30. During surgery, when the patient has not resumed spontaneous breathing, the mechanical ventilation module provides the power for breathing.

[0050] The breathing circuit 30 includes an inspiratory pathway 30a, an expiratory pathway 30b, and a carbon dioxide absorber 31. The inspiratory pathway 30a and expiratory pathway 30b are connected to form a closed loop. The carbon dioxide absorber 31 is disposed on the tubing of the expiratory pathway 30b. A mixture of fresh air and gas introduced by the air source interface 10 is input through the inlet of the inspiratory pathway 30a and provided to the patient through a patient interface 40 disposed at the outlet of the inspiratory pathway 30a. The patient interface 40 can be a face mask, a nasal cannula, or an endotracheal tube. In a preferred embodiment, a one-way valve 32 is provided on the inspiratory pathway 30a, which opens during the inspiratory phase and closes during the expiratory phase. A one-way valve 32 is also provided on the expiratory pathway 30b, which closes during the inspiratory phase and opens during the expiratory phase. The inlet of the expiratory pathway 30b is connected to the patient interface 40. When the patient exhales, the exhaled gas enters the carbon dioxide absorber 31 through the expiratory pathway 30b. The carbon dioxide in the exhaled gas is filtered out by the substances in the carbon dioxide absorber 31, and the gas after the carbon dioxide is removed is recirculated into the inspiratory pathway 30a. In some embodiments, a flow sensor and / or a pressure sensor are also provided in the breathing circuit 30 to detect the gas flow rate and / or the pressure in the tubing, respectively.

[0051] The processor 50 is used to execute instructions or programs to control various control valves in the breathing assist device 20, the air source interface 10 and / or the breathing circuit 30, or to process the received data to generate the required calculation or judgment results, or to generate visual data or graphics and output the visual data or graphics to the display 70 for display.

[0052] In this embodiment, the processor 50 receives at least one type of parameter data measured by an external device that can characterize a patient-machine interaction event. The parameter data includes at least one type of device ventilation parameter data and patient physiological parameter data. The device ventilation parameter data includes control parameter data set by the medical ventilation device itself, as well as ventilation-related parameter data monitored by the medical ventilation device itself. These parameter data include numerical data, waveform data, etc. Similarly, the physiological parameter data also includes numerical data, waveform data, etc. For example, the device ventilation parameter data includes at least one of the airway pressure, airway flow rate, and gas volume of the medical ventilation device. The physiological parameter data includes at least one of the patient's esophageal pressure, intragastric pressure, transdiaphragmatic pressure, carbon dioxide concentration, and diaphragmatic electrophysiology during ventilation. The external device can be various sensors placed inside the patient's body or various plug-ins or modules installed on the medical ventilation device, such as an esophageal pressure sensor placed in the patient's esophagus, or a carbon dioxide module for measuring carbon dioxide concentration. In other embodiments, the medical ventilation device itself may also include a parameter measuring device 60 for acquiring the aforementioned parameter data. It should be noted that in this invention, device ventilation parameter data and patient physiological parameters are two types of parameter data, while airway pressure and the like are one type of parameter data. If the processor 50 receives airway pressure and airway flow rate measured by an external device, then it will obtain two types of parameter data of the same type.

[0053] Upon receiving the aforementioned parameter data, the processor 50 activates the algorithm module associated with the received parameter data to obtain the identification results of patient-ventilator asynchrony events during ventilation. Multiple algorithm modules are included, and different algorithm modules perform calculations based on different data combinations formed from at least one parameter data to obtain the identification results of patient-ventilator asynchrony events during ventilation. The relationship between the algorithm module and the parameter data upon which its calculation is based can be defined as an association. The aforementioned data combination refers to a combination in a broad sense; a combination can be defined as having only one parameter data, meaning that the algorithm module can perform calculations based on a data combination with only one parameter data. In some embodiments, the relationship between the algorithm module and its corresponding data combination is as follows: Figure 2As shown, the data combination corresponding to algorithm module A includes only airway pressure as a parameter, while the data combination corresponding to algorithm module D includes both airway pressure and gas volume. In this embodiment, the processor 50 will only activate the corresponding algorithm module when at least one of the acquired parameter data meets the algorithm module's requirements for the parameter data combination. For example, Figure 2 For example, if processor 50 obtains the parameter data of airway pressure, then algorithm module A is activated, and algorithm module B is not activated based on airway pressure. In other embodiments, the requirements for parameter data in data combination also include the validity requirements of the parameter data, similarly... Figure 2 For example, if processor 50 acquires airway pressure, it will determine the validity of the airway pressure. If the airway pressure is valid data, processor 50 will activate algorithm module A. The validity requirements for the above parameter data can include requirements such as data range and periodicity. For example, if the acquired airway pressure is not within a preset range, the airway pressure is determined to be invalid.

[0054] Figure 2 The diagram illustrates one way in which parameter data is combined. In other embodiments, the data combination of each algorithm module may include airway flow rate and / or airway pressure, or any other combination of any number of parameter data, making the characteristics of human-machine confrontation more obvious and easier to obtain.

[0055] The aforementioned algorithm modules can be corresponding computer programs stored in memory. The number of algorithm modules can be increased or decreased, and they can be manually enabled or disabled. The recognition result calculated by the algorithm module based on parameter data can include whether a human-machine confrontation event was recognized or not. In other words, Figure 2 Algorithm module A can determine whether a human-machine confrontation event has occurred based on airway pressure and a certain calculation method, while algorithm module C can determine whether a human-machine confrontation event has occurred based on airway pressure and airway flow rate, using a different calculation method. Although both algorithm modules A and C use airway pressure as a parameter, their calculation methods can be independent, meaning that the recognition results of algorithm modules A and C can be independent.

[0056] In some embodiments, the identification result also includes the type of the identified human-machine interaction event. That is, algorithm module A can not only identify whether a human-machine interaction event has occurred based on airway pressure, but also identify the type of human-machine interaction that has occurred based on airway pressure. The types of human-machine interaction include one or more of the following: invalid triggering events, double triggering events, false triggering events, reverse triggering events, trigger delay events, early switching events, delayed switching events, and insufficient flow rate events. The following explanation uses algorithm modules C and D as examples to illustrate how the algorithm modules obtain the above identification results.

[0057] like Figure 3 The image shows the corresponding airway pressure and airway flow waveforms generated by processor 50 based on airway pressure and airway flow rate. If a patient experiences a patient-ventilator asynchrony, corresponding "traces" will be left on the airway pressure and airway flow waveforms, manifesting as abnormal waveform characteristics. Furthermore, the abnormal waveform characteristics vary depending on the type of patient-ventilator asynchrony event. Taking an invalid triggering event as an example... Figure 4 The image shows the waveform characteristics of airway pressure and airway flow when an invalid triggering event occurs. Figure 4 The upper middle part shows the airway pressure waveform. Figure 4 The lower center shows the airway flow rate waveform. If, at a certain moment, both the airway flow rate waveform and the airway pressure waveform suddenly increase and decrease simultaneously, and certain waveform features meet preset threshold conditions during the waveform changes, it indicates that an invalid triggering event has occurred in the patient. These waveform features may include the amplitude of the waveform change, the rate of change (first derivative), the second derivative, and the duration of the change. Algorithm module C extracts features from the airway pressure and airway flow rate waveforms, which involves calculating the parameter data. If the extracted features meet preset threshold conditions, algorithm module C will obtain the identification result of an invalid triggering event in the patient. In some embodiments, it can be set that if both the airway flow rate waveform and the airway pressure waveform features simultaneously meet the corresponding threshold conditions, algorithm module C will obtain the identification result of an invalid triggering event in the patient. In other embodiments, it can also be that within a certain time range where one of the airway flow rate waveform features and the airway pressure waveform features meets the corresponding threshold conditions, the other waveform feature also meets the corresponding threshold conditions, algorithm module C will obtain the identification result of an invalid triggering event in the patient.

[0058] Figure 5The processor 50 generates corresponding airway pressure waveforms and gas volume waveforms based on airway pressure and gas volume. Taking a double-trigger event as an example, the algorithm module D can calculate the duration of the patient's respiratory cycle based on the obtained airway pressure. If the patient's expiratory time within a respiratory cycle is too short and less than a preset threshold, a double-trigger event can be identified within that respiratory cycle. In this embodiment, the preset threshold can be the result of multiple related threshold conditions. For example, the average inspiratory time and average expiratory time of multiple (e.g., 12) respiratory cycles prior to the current respiratory cycle can be calculated. If the expiratory time of the current respiratory cycle is less than the minimum value among the average inspiratory time, the average expiratory time, and a fixed time threshold (e.g., 500ms), then the algorithm module D identifies a double-trigger event as occurring within the current respiratory cycle. In addition, the algorithm module D can also combine gas volume to determine whether a double-trigger event exists. Because a very short expiratory phase exists in a respiratory cycle with a double-trigger event, resulting in a smaller exhaled tidal volume, algorithm module D can calculate the exhaled tidal volume and / or inhaled tidal volume based on the obtained gas volume. When the exhaled tidal volume is detected to be less than a threshold (e.g., 1 / 2 of the inhaled tidal volume in the current respiratory cycle), or the result of subtracting the exhaled tidal volume from the inhaled tidal volume is greater than IBW*k (where IBW is the ideal body weight and k is a coefficient threshold, e.g., k=1ml / kg), then algorithm module D identifies a double-trigger event as having occurred.

[0059] The above provides an example illustrating how the algorithm modules obtain the identification results of human-machine confrontation events. Algorithm module C, in addition to identifying invalid trigger events, can also identify other types of human-machine confrontation events based on airway pressure and airway flow rate. For a specific type of human-machine confrontation event, multiple algorithm modules can be used to determine whether that type of event has occurred. For example, besides algorithm module C, algorithm modules E, G, H, I, and K can also identify whether invalid trigger events have occurred. These multiple algorithm modules can use different calculation methods based on slightly different parameter data. Therefore, the threshold conditions, threshold time, and thresholds mentioned above and below can have different set values ​​or preset values ​​depending on the algorithm module and / or the type of human-machine confrontation event being identified. This avoids different algorithm modules repeating the same judgment logic (calculation and judgment method). This invention identifies human-machine confrontation events from different perspectives, resulting in better robustness and higher accuracy of the identification results.

[0060] For example, Figure 6In the illustrated embodiment, the airway pressure change caused by the patient's invalid triggering event is relatively small. If airway pressure is relied upon alone to determine whether an invalid triggering event has occurred, the judgment may be inaccurate due to the insufficient characteristics of airway pressure. Figure 6 The study also incorporates esophageal pressure to jointly determine whether an invalid triggering event has occurred. When a patient is breathing spontaneously, the respiratory muscles, such as the diaphragm and intercostal muscles, actively contract, causing a decrease in intrapleural pressure. Clinically, monitoring this esophageal pressure is approximately equivalent to monitoring the patient's intrapleural pressure. Therefore, identifying changes in esophageal pressure can distinguish the patient's spontaneous breathing status. Figure 6 The middle arrow indicates the timing of changes in esophageal pressure caused by the patient's spontaneous inspiration, along with slight changes in airway pressure. These two points represent the moments when invalid triggering events occur. In this embodiment, the characteristic of the patient's spontaneous inspiration phase can be obtained by identifying the downward pressure swing of the esophageal pressure. Combining this characteristic with the identified airway pressure characteristics, invalid triggering events can be determined. That is, if within a threshold time after identifying airway pressure characteristics related to invalid triggering events, and then spontaneous breathing effort is detected through changes in esophageal pressure, then it can be determined that an invalid triggering event has occurred. Two algorithm modules can be set up to obtain the identification results of human-machine interaction events based on airway pressure and esophageal pressure respectively, or... Figure 2 In the illustrated embodiment, an algorithm module E is provided, which can comprehensively determine whether a human-machine confrontation event (such as an invalid triggering event) has occurred based on airway pressure and esophageal pressure, thereby improving the accuracy of the judgment.

[0061] For example, Figure 7 The illustrated embodiment also incorporates carbon dioxide concentration data to jointly determine whether an invalid triggering event has occurred. In some embodiments, an external carbon dioxide module can be used to monitor the patient's carbon dioxide concentration. During the patient's exhalation phase, the carbon dioxide waveform curve will show an upward movement; during the patient's inhalation phase, the carbon dioxide waveform curve will show a downward movement. Therefore, when a significant downward trend in the carbon dioxide waveform is detected (e.g., the magnitude and duration of the decline meet certain thresholds), it indicates that the patient is making an inspiratory effort. If the medical ventilation device does not trigger ventilation at this time, it indicates an invalid triggering event. Figure 7 The middle arrow corresponds to the moment when an invalid triggering event occurred. It is evident that identifying invalid triggers through carbon dioxide waveform characteristics can be combined with existing methods using airway pressure, airway flow rate, and gas volume curves, thereby improving accuracy. By extracting features from the carbon dioxide waveform using the algorithm module—that is, calculating changes in carbon dioxide concentration—a clear waveform decline characteristic can be detected, thus yielding the identification result of the human-machine confrontation event. For example, in… Figure 2 In the illustrated embodiment, algorithm module H is enabled to work with other algorithm modules to determine whether an invalid triggering event has occurred.

[0062] The processor 50 is also used to determine the human-machine confrontation event that the patient has experienced based on the recognition results obtained by different algorithm modules. In some embodiments, the human-machine confrontation events are distinguished according to different types, that is, to determine which types of human-machine confrontation events the patient has experienced. A schematic diagram of different algorithm modules comprehensively judging whether a certain type of human-machine confrontation event has occurred is shown below. Figure 8 As shown in the diagram, each type of human-machine confrontation event has a corresponding judgment module. Each judgment module includes algorithm modules capable of identifying the same type of human-machine confrontation event. The output of each judgment module is whether the corresponding human-machine confrontation event has occurred or not. In other words, the judgment module determines whether the corresponding type of human-machine confrontation event has occurred based on the recognition results of each algorithm module capable of identifying the same type of human-machine confrontation event. The types of human-machine confrontation events experienced by the patient can be determined based on the output results of each judgment module, which in turn are based on the recognition results of the internal algorithm modules. For example, in the invalid trigger judgment module, some algorithm modules may identify that an invalid trigger event has occurred, while others may identify that no invalid trigger event has occurred. If the final output of the invalid trigger judgment module is that an invalid trigger event has occurred, it means that the combined judgment result of the algorithm modules used by the processor 50 is that the patient has experienced an invalid trigger event. Judgment modules corresponding to other types of human-machine confrontation events will also obtain corresponding output results, thereby determining the types of human-machine confrontation events experienced by the patient.

[0063] In some embodiments, the output of the judgment module can be determined in the following manner:

[0064] The credibility of the first algorithm module is obtained. The first algorithm module is the one that identifies a human-machine confrontation event. For example, in the invalid trigger judgment module, if algorithm modules C, E, and G identify invalid trigger events, these three algorithm modules are defined as the first algorithm module. In other embodiments, the first algorithm module is the one that identifies the type of human-machine confrontation event corresponding to its judgment module. Then, based on the credibility of each first algorithm module, it is determined whether the same type of human-machine confrontation event identified by each first algorithm module has occurred. In this embodiment, the credibility of algorithm modules C, E, and G is used to determine whether an invalid trigger event has occurred. The credibility can be related to at least one of the following: signal-to-noise ratio, feature clarity, and regularity of the parameter data associated with the algorithm module. For example, when using airway pressure and airway flow rate to identify invalid trigger events, the clarity of the feature can be determined based on the magnitude of the identified pressure drop and the change in the first derivative of the flow rate. That is, the greater the pressure drop and the greater the change in the first derivative of the flow rate, the higher the feature clarity and the higher the credibility. When using physiological signals from patients, such as esophageal pressure or diaphragmatic electrical activity, for human-machine interaction recognition, factors such as signal monitoring quality and signal-to-noise ratio can affect the reliability of the corresponding algorithm modules. For example, when a weak signal fluctuation amplitude or irregular periodic movement is detected, the reliability is low; conversely, if the signal fluctuation exceeds a certain threshold and a regular fluctuation cycle is detected within a recent period, the reliability is high.

[0065] Based on the credibility of the first algorithm module, the processor 50 generates a credibility score corresponding to the credibility of the first algorithm module. Then, combining this score with the corresponding weight coefficient, the credibility score of the first algorithm module is corrected to obtain a corrected credibility score. For different types of human-machine adversarial events, different algorithm modules have pre-set corresponding weight coefficients. For example, the weight coefficients can be determined at least based on the correlation between the algorithm module and the type of human-machine adversarial event. For instance, when identifying delayed trigger events, considering that diaphragmatic electrical signals generally precede esophageal pressure signals, a higher weight coefficient is given to algorithm modules that use diaphragmatic electrical signals for identification. The advantage of this is that the algorithm results can be corrected to a certain extent based on clinical consensus or the essential characteristics of the parameters, making the algorithm results more credible. Based on the corrected credibility score, the sum of the credibility scores of each first algorithm module is calculated, and then it is determined whether the sum of the credibility scores is greater than a preset threshold. If it is greater than the preset threshold, the same type of human-machine adversarial event identified by each first algorithm module has occurred; otherwise, the same type of human-machine adversarial event identified by each first algorithm module has not occurred. The sum of the credibility scores of each of the first algorithm modules mentioned above refers to the sum of the credibility scores of the first algorithm modules within the same judgment module. If the sum of the credibility scores of each of the first algorithm modules within the judgment module is greater than a preset threshold, the output result of the judgment module is that a human-machine confrontation event of the corresponding type has occurred.

[0066] The above method takes into account the credibility of the recognition results of each first algorithm module. When the total credibility of each first algorithm module in the judgment module meets certain conditions, the output result of the judgment module is the occurrence of the corresponding type of human-machine confrontation event.

[0067] In other embodiments, the occurrence of the same type of human-computer interaction event identified by each algorithm module is determined based on the ratio and / or quantity relationship between the first algorithm module and the second algorithm module. The first algorithm module is defined as the one that identifies a human-computer interaction event, and the second algorithm module is defined as the one that does not identify a human-computer interaction event. For example, in the invalid trigger judgment module, if algorithm modules C, E, and G identify invalid trigger events, these three algorithm modules are defined as first algorithm modules. If algorithm modules H, I, and K do not identify invalid trigger events, these three algorithm modules are second algorithm modules. In this example, there are three first algorithm modules. If a judgment module is predefined to have two or more first algorithm modules, then the output of that judgment module is the occurrence of a corresponding type of human-computer interaction event. Therefore, the output of the invalid trigger judgment module is that an invalid trigger event has occurred. The advantage of this embodiment is that it can minimize false identification, as excessive false identification can lead to user information fatigue and weaken the original prompting effect.

[0068] After identifying the human-machine interaction event that occurred to the patient, the processor 50 also outputs the human-machine interaction event to the display interface of the monitor 70 or other display device.

[0069] In some embodiments, the processor 50 marks the identified human-machine adversarial event near a corresponding feature of the corresponding parameter data waveform within the display interface, for example, such as... Figure 9 As shown, the occurrence of human-machine interaction (HMI) events is indicated by triangular symbols and / or the name of the HMI (Paw for airway pressure, Pes for esophageal pressure). The triangular symbols indicate the position of the corresponding feature on the waveform, and the string below the triangular symbol represents the type name of the HMI (IE for invalid trigger event, DT for double trigger event, RT for reverse trigger event). The marking method can also use any symbol, color, or string to distinguish different types of HMI events. The advantage of this is that it allows for a direct association between the type of HMI event and the corresponding feature on the waveform. Experienced doctors can directly judge the accuracy of the recognition results from the marking results. The markings on the waveform also correspond to the changes in the monitored value of the HMI incidence rate, allowing doctors to clearly understand the meaning of the monitored value. For general medical staff, the waveform annotations can also be used for learning. Without waveform annotations, only the monitored value of the HMI incidence rate is available, and doctors cannot determine the accuracy of the recognition algorithm. Another advantage of directly annotating results on the waveform interface is that users can still observe the latest ventilation waveform without freezing the waveform or switching to other interfaces. In addition, users do not need to perform any additional operations to see the recognition results and waveform characteristics, making it convenient and easy to use.

[0070] In some embodiments, such as Figure 10 As shown, based on the identification results of human-computer interaction events, the incidence rate of each type of event is statistically analyzed. This statistical analysis can be based on the incidence rate over a recent period or on the incidence rate over a certain number of recent breathing cycles. When the incidence rate of a certain type of human-computer interaction event exceeds a certain threshold, a notification is displayed in a specific area of ​​the main interface indicating that too many events of that type have occurred, along with suggested actions. Taking invalid triggers as an example... Figure 10 This demonstration shows a notification method used when the occurrence rate of invalid triggering events exceeds 10%. This method primarily displays two pieces of information: first, it notifies the user of the type of human-computer interaction event that has occurred excessively; second, it provides operational suggestions, such as... Figure 10 The prompt message indicates that the user should lower the threshold setting based on the trigger sensitivity threshold in the current ventilation parameters.

[0071] Compared to existing technologies, this approach offers the advantage of not alerting the user immediately upon the occurrence of human-computer interaction (HCI) conflict, but rather only when the frequency of HCI conflict exceeds a certain level. Frequent alerts can cause user fatigue or visual strain. Furthermore, another advantage over existing technologies is that they rely solely on abnormal monitoring parameters to identify HCI events and provide operational prompts (US9027552). However, clinical definitions and identification of HCI conflict are based on waveform features. The algorithm module in this invention further extracts features from the waveform of the parameter data to determine HCI conflict, thereby providing targeted operational prompts to improve such situations. These prompts are closer to clinical doctor operations, and the results and guidance information are more meaningful.

[0072] The present invention also provides a ventilation monitoring method, such as Figure 11 As shown, the steps include:

[0073] Step 1000: Obtain at least one parameter data that can characterize the human-machine confrontation event that occurred in the patient.

[0074] The parameter data includes at least one type of equipment ventilation parameter data and patient physiological parameter data. For example, ventilation parameter data includes at least one of the following: airway pressure, airway flow rate, and gas volume of the medical ventilation equipment; physiological parameter data includes at least one of the following: esophageal pressure, intragastric pressure, transdiaphragmatic pressure, carbon dioxide concentration, and diaphragmatic electrical activity of the patient during ventilation. It should be noted that equipment ventilation parameter data and patient physiological parameters are two types of parameter data, while airway pressure, etc., is only one type of parameter data. If airway pressure and airway flow rate are obtained, then two types of parameter data are obtained within the same category.

[0075] The aforementioned parameter data can be measured by external devices. These external devices can be various sensors placed inside the patient's body or various plug-ins or modules installed on the medical ventilation equipment, such as an esophageal pressure sensor placed in the patient's esophagus or a carbon dioxide module for measuring carbon dioxide concentration. In other embodiments, the medical ventilation equipment itself may also include a parameter measuring device 60 for acquiring the aforementioned parameter data.

[0076] Step 2000: Activate an algorithm module associated with at least one parameter data to obtain the identification results of patient-ventilator asynchrony events during ventilation. The algorithm module includes multiple modules, each performing calculations based on different data combinations formed from the at least one parameter data to obtain the identification results of patient-ventilator asynchrony events during ventilation.

[0077] In this step, the algorithm module and the parameter data upon which its calculations are based can be defined as related. The aforementioned data combination refers to a combination in a broad sense; a combination can be defined as having only one type of parameter data, meaning the algorithm module can perform calculations based on a data combination with only one type of parameter data. In some embodiments, the relationship between the algorithm module and its corresponding data combination is as follows: Figure 2 As shown, the data combination corresponding to algorithm module A includes only airway pressure, while the data combination corresponding to algorithm module D includes both airway pressure and gas volume. In this embodiment, the corresponding algorithm module will only be activated when at least one obtained parameter meets the algorithm module's requirements for the parameter data combination. For example, using... Figure 2 For example, if airway pressure is obtained, algorithm module A is activated, and algorithm module B is not activated based on airway pressure. In other embodiments, the requirements for parameter data in data combination also include the validity requirements of the parameter data, similarly... Figure 2 For example, if airway pressure is obtained, its validity can be determined. If the airway pressure is valid data, algorithm module A is activated. The validity requirements for the above parameter data can include data range and periodicity. For instance, if the obtained airway pressure is not within a preset range, it is determined that the airway pressure is invalid.

[0078] Figure 2 The diagram illustrates one way in which parameter data is combined. In other embodiments, the data combination of each algorithm module may include airway flow rate and / or airway pressure, or any other combination of any number of parameter data, making the characteristics of human-machine confrontation more obvious and easier to obtain.

[0079] The aforementioned algorithm modules can be corresponding computer programs stored in memory, and the number of algorithm modules can be increased or decreased. The recognition result calculated by the algorithm module based on parameter data can include whether a human-machine confrontation event was recognized or not; that is, Figure 2 Algorithm module A can determine whether a human-machine confrontation event has occurred based on airway pressure and a certain calculation method, while algorithm module C can determine whether a human-machine confrontation event has occurred based on airway pressure and airway flow rate, using a different calculation method. Although both algorithm modules A and C use airway pressure as a parameter, their calculation methods can be independent; that is, the recognition results of algorithm modules A and C can be independent.

[0080] In some embodiments, the identification result also includes the type of the identified human-machine interaction event. That is, algorithm module A can not only identify whether a human-machine interaction event has occurred based on airway pressure, but also identify the type of human-machine interaction that has occurred based on airway pressure. The types of human-machine interaction include one or more of the following: invalid triggering event, double triggering event, false triggering event, reverse triggering event, trigger delay event, premature switching event, delayed switching event, and insufficient flow rate event. The following explanation uses algorithm modules C and D as examples to illustrate how the algorithm modules obtain the above identification results.

[0081] like Figure 3 The image shows the corresponding airway pressure and airway flow waveforms generated based on airway pressure and flow rate. If a patient experiences a patient-ventilator asynchrony, corresponding "traces" will be left on the airway pressure and flow rate waveforms, manifesting as abnormal waveform characteristics. Furthermore, the abnormal waveform characteristics vary depending on the type of patient-ventilator asynchrony event. Taking an invalid triggering event as an example... Figure 4 The image shows the waveform characteristics of airway pressure and airway flow when an invalid triggering event occurs. Figure 4 The upper middle part shows the airway pressure waveform. Figure 4 The lower center shows the airway flow rate waveform. If, at a certain moment, both the airway flow rate waveform and the airway pressure waveform suddenly increase and decrease simultaneously, and certain waveform features meet preset threshold conditions during the waveform changes, it indicates that an invalid triggering event has occurred in the patient. These waveform features may include the amplitude of the waveform change, the rate of change (first derivative), the second derivative, and the duration of the change. Algorithm module C extracts features from the airway pressure and airway flow rate waveforms, which involves calculating the parameter data. If the extracted features meet preset threshold conditions, algorithm module C will obtain the identification result of an invalid triggering event in the patient. In some embodiments, it can be set that if both the airway flow rate waveform and the airway pressure waveform features simultaneously meet the corresponding threshold conditions, algorithm module C will obtain the identification result of an invalid triggering event in the patient. In other embodiments, it can also be that within a certain time range where one of the airway flow rate waveform features and the airway pressure waveform features meets the corresponding threshold conditions, the other waveform feature also meets the corresponding threshold conditions, algorithm module C will obtain the identification result of an invalid triggering event in the patient.

[0082] Figure 5The waveforms shown in the image are the corresponding airway pressure waveform and gas volume waveform generated based on airway pressure and gas volume. Taking a double-trigger event as an example, algorithm module D can calculate the duration of the patient's respiratory cycle based on the obtained airway pressure. If the patient's expiratory time within a respiratory cycle is too short and less than a preset threshold, a double-trigger event can be identified within that respiratory cycle. In this embodiment, the preset threshold can be the result of multiple related threshold conditions. For example, the average inspiratory time and average expiratory time of multiple (e.g., 12) respiratory cycles prior to the current respiratory cycle can be calculated. If the expiratory time of the current respiratory cycle is less than the minimum value among the average inspiratory time, the average expiratory time, and a fixed time threshold (e.g., 500ms), then algorithm module D identifies a double-trigger event as occurring within the current respiratory cycle. In addition, algorithm module D can also combine gas volume to determine whether a double-trigger event exists. Because a very short expiratory phase exists in a respiratory cycle with a double-trigger event, resulting in a smaller exhaled tidal volume, algorithm module D can calculate the exhaled tidal volume and / or inhaled tidal volume based on the obtained gas volume. When the exhaled tidal volume is detected to be less than a threshold (e.g., 1 / 2 of the inhaled tidal volume in the current respiratory cycle), or the result of subtracting the exhaled tidal volume from the inhaled tidal volume is greater than IBW*k (where IBW is the ideal body weight and k is a coefficient threshold, e.g., k=1ml / kg), then algorithm module D identifies a double-trigger event as having occurred.

[0083] The above provides an example illustrating how the algorithm modules obtain the identification results of human-machine confrontation events. Algorithm module C, in addition to identifying invalid trigger events, can also identify other types of human-machine confrontation events based on airway pressure and airway flow rate. For a specific type of human-machine confrontation event, multiple algorithm modules can be used to determine whether that type of event has occurred. For example, besides algorithm module C, algorithm modules E, G, H, I, and K can also identify whether invalid trigger events have occurred. These multiple algorithm modules can use different calculation methods based on slightly different parameter data. Therefore, the threshold conditions, threshold time, and thresholds mentioned above and below can have different set values ​​or preset values ​​depending on the algorithm module and / or the type of human-machine confrontation event being identified. This avoids different algorithm modules repeating the same judgment logic (calculation and judgment method). This invention identifies human-machine confrontation events from different perspectives, resulting in better robustness and higher accuracy of the identification results.

[0084] For example, Figure 6In the illustrated embodiment, the airway pressure change caused by the patient's invalid triggering event is relatively small. If airway pressure is relied upon alone to determine whether an invalid triggering event has occurred, the judgment may be inaccurate due to the insufficient characteristics of airway pressure. Figure 6 The study also incorporates esophageal pressure to jointly determine whether an invalid triggering event has occurred. When a patient is breathing spontaneously, the respiratory muscles, such as the diaphragm and intercostal muscles, actively contract, causing a decrease in intrapleural pressure. Clinically, monitoring this esophageal pressure is approximately equivalent to monitoring the patient's intrapleural pressure. Therefore, identifying changes in esophageal pressure can distinguish the patient's spontaneous breathing status. Figure 6 The middle arrow indicates the timing of changes in esophageal pressure caused by the patient's spontaneous inspiration, along with slight changes in airway pressure. These two points represent the moments when invalid triggering events occur. In this embodiment, the characteristic of the patient's spontaneous inspiration phase can be obtained by identifying the downward pressure swing of the esophageal pressure. Combining this characteristic with the identified airway pressure characteristics, invalid triggering events can be determined. That is, if within a threshold time after identifying airway pressure characteristics related to invalid triggering events, and then spontaneous breathing effort is detected through changes in esophageal pressure, then it can be determined that an invalid triggering event has occurred. Two algorithm modules can be set up to obtain the identification results of human-machine interaction events based on airway pressure and esophageal pressure respectively, or... Figure 2 In the illustrated embodiment, an algorithm module E is provided, which can comprehensively determine whether a human-machine confrontation event (such as an invalid triggering event) has occurred based on airway pressure and esophageal pressure, thereby improving the accuracy of the judgment.

[0085] For example, Figure 7 The illustrated embodiment also incorporates carbon dioxide concentration data to jointly determine whether an invalid triggering event has occurred. In some embodiments, an external carbon dioxide module can be used to monitor the patient's carbon dioxide concentration. During the patient's exhalation phase, the carbon dioxide waveform curve will show an upward movement; during the patient's inhalation phase, the carbon dioxide waveform curve will show a downward movement. Therefore, when a significant downward trend in the carbon dioxide waveform is detected (e.g., the magnitude and duration of the decline meet certain thresholds), it indicates that the patient is making an inspiratory effort. If the medical ventilation device does not trigger ventilation at this time, it indicates an invalid triggering event. Figure 7 The middle arrow corresponds to the moment when an invalid triggering event occurred. It is evident that identifying invalid triggers through carbon dioxide waveform characteristics can be combined with existing methods using airway pressure, airway flow rate, and gas volume curves, thereby improving accuracy. By extracting features from the carbon dioxide waveform using the algorithm module—that is, calculating changes in carbon dioxide concentration—a clear waveform decline characteristic can be detected, thus yielding the identification result of the human-machine confrontation event. For example, in… Figure 2 In the illustrated embodiment, algorithm module H is enabled to work with other algorithm modules to determine whether an invalid triggering event has occurred.

[0086] Step 3000: Based on the recognition results obtained by different algorithm modules, determine the human-machine confrontation event that occurred in the patient.

[0087] In some embodiments, step 3000 may include:

[0088] Step 3100: Based on the recognition results of each algorithm module that can identify the same type of human-machine confrontation event, determine whether the same type of human-machine confrontation event identified by each algorithm module has occurred.

[0089] In some embodiments, such as Figure 8 As shown in the figure, each type of human-machine adversarial event has a corresponding judgment module. Each judgment module includes algorithm modules capable of identifying the same type of human-machine adversarial event. The output of the judgment module is whether the corresponding human-machine adversarial event has occurred or not. In other words, the judgment module determines whether the corresponding type of human-machine adversarial event has occurred based on the recognition results of the algorithm modules capable of identifying the same type of human-machine adversarial event. Based on this, step 3100 specifically includes:

[0090] Step 3110: Obtain the credibility of the first algorithm module. The first algorithm module is the algorithm module whose identification result is the identification of a human-machine confrontation event. For example, in the invalid trigger judgment module, if the identification results of algorithm modules C, E, and G are all invalid trigger events, then these three algorithm modules are defined as the first algorithm module. In other embodiments, the first algorithm module is the algorithm module that identifies the type of human-machine confrontation event corresponding to its judgment module.

[0091] Step 3120: Based on the credibility of each first algorithm module, determine whether the same type of human-machine confrontation event identified by each first algorithm module has occurred.

[0092] In this embodiment, the determination of whether an invalid triggering event has occurred is based on the reliability of algorithm modules C, E, and G. Reliability can be correlated with at least one of the following: signal-to-noise ratio, feature clarity, and regularity of the parameter data associated with the algorithm module. For example, when identifying invalid triggering events using airway pressure and airway flow rate, the clarity of a feature can be determined based on the magnitude of the identified pressure drop and the change in the first derivative of the flow rate. That is, the greater the pressure drop and the greater the change in the first derivative of the flow rate, the higher the feature clarity and the higher the reliability. However, when using patient physiological signals such as esophageal pressure or diaphragmatic electroencephalography (EEG) for human-machine interaction identification, signal monitoring quality and signal-to-noise ratio will affect the reliability of the corresponding algorithm module. For example, when a weak signal fluctuation or irregular periodic movement is detected, the reliability is low; conversely, if the signal fluctuation exceeds a certain threshold and a regular fluctuation cycle is detected within a recent period, the reliability is high.

[0093] In some embodiments, step 3120 may include:

[0094] Based on the credibility of the first algorithm module, a credibility score corresponding to the credibility of the first algorithm module is generated.

[0095] By combining the weight coefficients corresponding to the first algorithm module, the credibility score of the first algorithm module is corrected to obtain the corrected credibility score.

[0096] In this context, different algorithm modules are pre-set with corresponding weight coefficients for different types of human-machine adversarial events. For example, the weight coefficients can be determined based on the degree of correlation between the algorithm module and the type of human-machine adversarial event. For instance, when identifying delayed trigger events, considering that diaphragmatic electrical signals generally precede esophageal pressure signals, a higher weight coefficient is given to the algorithm module that uses diaphragmatic electrical signals for identification. The advantage of doing so is that the algorithm results can be modified to a certain extent based on clinical consensus or the essential characteristics of the parameters, making the algorithm results more reliable.

[0097] Based on the revised credibility score, the sum of the credibility scores of each first algorithm module is calculated.

[0098] Determine whether the sum of the credibility scores is greater than a preset threshold. If it is greater than the preset threshold, then the same type of human-machine confrontation event identified by each first algorithm module has occurred; otherwise, the same type of human-machine confrontation event identified by each first algorithm module has not occurred.

[0099] The sum of the credibility scores of each of the first algorithm modules mentioned above refers to the sum of the credibility scores of the first algorithm modules within the same judgment module. If the sum of the credibility scores of each of the first algorithm modules within the judgment module is greater than a preset threshold, the output result of the judgment module is that a human-machine confrontation event of the corresponding type has occurred.

[0100] The above method takes into account the credibility of the recognition results of each first algorithm module. When the total credibility of each first algorithm module in the judgment module meets certain conditions, the output result of the judgment module is the occurrence of the corresponding type of human-machine confrontation event.

[0101] In some embodiments, step 3100 specifically includes: determining whether the same type of human-computer interaction event identified by each algorithm module has occurred based on the ratio and / or quantity relationship between the first algorithm module and the second algorithm module. The first algorithm module is the one that identifies a human-computer interaction event, and the second algorithm module is the one that does not identify a human-computer interaction event. For example, in the invalid trigger judgment module, if algorithm modules C, E, and G identify invalid trigger events, these three algorithm modules are defined as first algorithm modules; if algorithm modules H, I, and K do not identify invalid trigger events, these three algorithm modules are second algorithm modules. In this example, there are three first algorithm modules. If a judgment module is predefined to have two or more first algorithm modules, then the output of that judgment module is the occurrence of a corresponding type of human-computer interaction event. Therefore, the output of the invalid trigger judgment module is that an invalid trigger event has occurred. The advantage of this embodiment is that it can minimize misidentification, as excessive misidentification can lead to user information fatigue and weaken the original prompting effect.

[0102] Step 3200: Determine the types of human-machine interaction events that occurred in the patient based on whether different types of human-machine interaction events occurred.

[0103] The types of human-machine interaction events that the patient experienced can be determined based on the output results of each judgment module, which in turn are determined based on the recognition results of each internal algorithm module.

[0104] Step 4000: Output the human-machine interaction events that occurred in the patient.

[0105] In some embodiments, the identified human-machine adversarial events are marked near corresponding features of the corresponding parameter data waveform, for example, such as... Figure 9As shown, the occurrence of human-machine confrontation events is indicated by a triangular symbol and the name of the event. The triangular symbol indicates the position of the corresponding feature on the waveform, and the string below the triangular symbol represents the type name of the event (IE is short for invalid trigger event, DT for double trigger event, and RT for reverse trigger event). The marking method can also use any symbol, color, or string to distinguish different types of human-machine confrontation events. The advantage of this is that it can correlate the type of human-machine confrontation event with the corresponding feature on the waveform. Experienced doctors can directly judge the accuracy of the recognition results from the marking results. The markings on the waveform are also consistent with the changes in the monitored value of the human-machine confrontation event incidence rate, allowing doctors to clearly understand the meaning of the monitored value. For general medical staff, the waveform annotations can also be used for learning. Without waveform annotations, only the monitored value of the human-machine confrontation event incidence rate is available, and doctors cannot understand whether the recognition algorithm is accurate. Another advantage of directly annotating results on the waveform interface is that users can still observe the latest ventilation waveform without freezing the waveform or switching to other interfaces. In addition, users do not need to perform any additional operations to see the recognition results and waveform characteristics, making it convenient and easy to use.

[0106] In some embodiments, such as Figure 10 As shown, based on the identification results of human-computer interaction events, the incidence rate of each type of event is statistically analyzed. This statistical analysis can be based on the incidence rate over a recent period or on the incidence rate over a certain number of recent breathing cycles. When the incidence rate of a certain type of human-computer interaction event exceeds a certain threshold, a notification is displayed in a specific area of ​​the main interface indicating that too many events of that type have occurred, along with suggested actions. Taking invalid triggers as an example... Figure 10 This demonstration shows a notification method used when the occurrence rate of invalid triggering events exceeds 10%. This method primarily displays two pieces of information: first, it notifies the user of the type of human-computer interaction event that has occurred excessively; second, it provides operational suggestions, such as... Figure 10 The prompt message indicates that the user should lower the threshold setting based on the trigger sensitivity threshold in the current ventilation parameters.

[0107] This invention employs multiple algorithm modules based on at least one parameter data, and uses different algorithms to comprehensively analyze and judge human-machine confrontation events, making the identification results of human-machine confrontation events more accurate and robust.

[0108] The above examples illustrate the present invention and are only intended to aid in understanding the invention, not to limit it. Those skilled in the art can make variations to the specific embodiments described above based on the spirit of the invention.

Claims

1. A medical ventilation device, characterized in that, include: Gas source interface, used to connect to a gas source; The patient interface is used to connect to the patient's respiratory system; A breathing circuit is used to connect the gas source interface and the patient interface to deliver gas supplied by the gas source to the patient. A respiratory support device is used to provide respiratory support power to control the delivery of gas supplied by a gas source to the patient; Processor, used for: Acquire at least two types of parameter data, which can characterize patient-ventilator asynchrony events, and the types of the at least two types of parameter data include at least one type of equipment ventilation parameter data and patient physiological parameter data; Based on the received at least two parameter data, at least two algorithm modules associated with the at least two parameter data are activated, and the identification results of human-machine asynchrony events during ventilation are obtained through the at least two algorithm modules respectively; The at least two algorithm modules are different algorithm modules. Each algorithm module performs calculations based on a data combination formed by at least one parameter data associated with the algorithm module from the at least two parameter data to obtain the identification results of the patient's human-machine asynchrony event during ventilation. The data combinations corresponding to different algorithm modules are different. The identification result includes: the human-machine confrontation event was identified or the human-machine confrontation event was not identified; And based on the recognition results obtained by each of the at least two algorithm modules, determine and output the human-machine interaction event that occurred to the patient, including: The credibility of the first algorithm module among the at least two algorithm modules is obtained, wherein the first algorithm module is the algorithm module whose recognition result is the recognition of the human-machine adversarial event among the at least two algorithm modules; the credibility is related to at least one of the signal-to-noise ratio, feature clarity and regularity of the parameter data associated with the first algorithm module; Obtain the weight coefficients of the first algorithm module, and correct the credibility score corresponding to the credibility of the first algorithm module based on the weight coefficients; When the overall credibility score of each of the first algorithm modules meets the conditions, the corresponding type of human-machine confrontation event that occurred to the patient is determined and output.

2. The medical ventilation device as described in claim 1, characterized in that, The at least two algorithm modules include: An algorithm module associated with one of the at least two parameter data; and / or, An algorithm module associated with any combination of the at least two parameter data.

3. The medical ventilation device as described in claim 1, characterized in that, Enable at least two algorithm modules associated with the at least two parameter data, including: When it is determined that the acquired at least two types of parameter data meet the data combination requirements of the algorithm module for parameter data, the algorithm module is activated.

4. The medical ventilation device as described in claim 3, characterized in that, The requirements for parameter data in the data combination include the validity requirements for the parameter data.

5. The medical ventilation device as described in claim 1, characterized in that, The recognition results of the algorithm module also include the type of human-machine confrontation event identified.

6. The medical ventilation device as described in claim 5, characterized in that, The types of human-machine confrontation events include one or more of the following: invalid triggering events, double triggering events, false triggering events, reverse triggering events, trigger delay events, early switching events, switching delay events, and events with too low flow rate.

7. The medical ventilation device as described in claim 5, characterized in that, Based on the recognition results obtained by each of the at least two algorithm modules, the human-machine confrontation event experienced by the patient is determined, including: Based on the recognition results of each algorithm module that can identify the same type of human-machine confrontation event, determine whether the same type of human-machine confrontation event identified by each algorithm module has occurred; Based on whether different types of human-machine interaction events occur, the types of human-machine interaction events that occurred in the patient are determined.

8. The medical ventilation device as described in claim 1, characterized in that, When the overall credibility score of each of the first algorithm modules meets the conditions, the corresponding type of human-machine interaction event occurring to the patient is determined and output, including: Based on the revised credibility score, calculate the sum of the credibility scores of each first algorithm module; Determine whether the sum of the credibility scores is greater than a preset threshold. If it is greater than the preset threshold, then the same type of human-machine confrontation event identified by each of the first algorithm modules has occurred, and the human-machine confrontation event that occurred to the patient is determined and output.

9. The medical ventilation device as described in claim 8, characterized in that, If the sum of the credibility scores is less than or equal to a preset threshold, then the same type of human-machine confrontation event identified by the first algorithm module has not occurred.

10. The medical ventilation device as described in claim 1, characterized in that, The weighting coefficients are determined at least based on the degree of correlation between the algorithm module and the type of human-machine confrontation event.

11. The medical ventilation device as described in claim 7, characterized in that, Based on the recognition results of each algorithm module capable of identifying the same type of human-machine adversarial event, determine whether the same type of human-machine adversarial event identified by each algorithm module has occurred, including: Based on the ratio and / or quantity relationship between the first algorithm module and the second algorithm module, it is determined whether the same type of human-machine confrontation event identified by each algorithm module has occurred. The first algorithm module is the algorithm module that has identified the human-machine confrontation event, and the second algorithm module is the algorithm module that has not identified the human-machine confrontation event.

12. The medical ventilation device as described in claim 1, characterized in that, Obtain at least two types of parameter data, including: Receive the parameter data measured by an external device, or The medical ventilation device further includes a parameter measuring device to acquire at least one parameter data that can characterize a patient experiencing a human-machine interaction event, including the parameter measuring device acquiring the parameter data through measurement.

13. The medical ventilation device as described in claim 1, characterized in that, The ventilation parameter data includes at least one of airway pressure, airway flow rate, and gas volume; The physiological parameters include at least one of esophageal pressure, intragastric pressure, transdiaphragmatic pressure, carbon dioxide concentration, and diaphragmatic electrical activity.

14. A ventilation monitoring method, characterized in that, Including the following steps: Acquire at least two types of parameter data, which can characterize patient-ventilator asynchrony events, and the types of the at least two types of parameter data include at least one type of device ventilation parameter data and patient physiological parameter data; At least two algorithm modules associated with the at least two parameter data are enabled, and the identification results of human-ventilator asynchrony events during ventilation are obtained through the at least two algorithm modules respectively; Wherein, the at least two algorithm modules are different algorithm modules, and each algorithm module performs calculations based on a data combination formed by at least one parameter data associated with the algorithm module from the at least two parameter data to obtain a recognition result, the recognition result including: the human-machine confrontation event is recognized or the human-machine confrontation event is not recognized; Based on the recognition results obtained by each of the at least two algorithm modules, the human-machine confrontation event experienced by the patient is determined, including: The credibility of the first algorithm module among the at least two algorithm modules is obtained, wherein the first algorithm module is the algorithm module whose recognition result is the recognition of the human-machine adversarial event among the at least two algorithm modules; the credibility is related to at least one of the signal-to-noise ratio, feature clarity and regularity of the parameter data associated with the first algorithm module; Obtain the weight coefficients of the first algorithm module, and correct the credibility score corresponding to the credibility of the first algorithm module based on the weight coefficients; When the overall credibility score of each of the first algorithm modules meets the conditions, the corresponding type of human-machine confrontation event that occurred to the patient is determined; Output the human-machine interaction events that occurred with the patient.

15. The ventilation monitoring method as described in claim 14, characterized in that, The at least two algorithm modules include: An algorithm module associated with one of the at least two parameter data; And / or, An algorithm module associated with any combination of the at least two parameter data.

16. The ventilation monitoring method as described in claim 14, characterized in that, Enable at least two algorithm modules associated with the at least two parameter data, including: When it is determined that at least two types of parameter data satisfy the requirements of the algorithm module for parameter data, the algorithm module is then activated.

17. The ventilation monitoring method as described in claim 16, characterized in that, The requirements for parameter data in the data combination include the validity requirements for the parameter data.

18. The ventilation monitoring method as described in claim 14, characterized in that, The recognition results of the algorithm module also include the type of human-machine confrontation event identified.

19. The ventilation monitoring method as described in claim 18, characterized in that, The types of human-machine confrontation events include one or more of the following: invalid triggering events, double triggering events, false triggering events, reverse triggering events, trigger delay events, early switching events, switching delay events, and events with too low flow rate.

20. The ventilation monitoring method as described in claim 18, characterized in that, Based on the recognition results obtained by each of the different algorithm modules, the human-machine confrontation event experienced by the patient is determined, including: Based on the recognition results of each algorithm module that can identify the same type of human-machine confrontation event, determine whether the same type of human-machine confrontation event identified by each algorithm module has occurred; Based on whether different types of human-machine interaction events occur, the types of human-machine interaction events that occurred in the patient are determined.

21. The ventilation monitoring method as described in claim 14, characterized in that, When the overall credibility score of each of the first algorithm modules meets the conditions, the corresponding type of human-machine interaction event occurring to the patient is determined, including: Based on the corrected credibility score, calculate the sum of the credibility scores of each of the first algorithm modules; Determine whether the sum of the credibility scores is greater than a preset threshold. If it is greater than the preset threshold, then the same type of human-machine confrontation event identified by each of the first algorithm modules has occurred.

22. The ventilation monitoring method as described in claim 21, characterized in that, If the sum of the credibility scores is less than or equal to a preset threshold, then the same type of human-machine confrontation event identified by the first algorithm module has not occurred.

23. The ventilation monitoring method as described in claim 14, characterized in that, The weighting coefficients are determined at least based on the degree of correlation between the algorithm module and the type of human-machine confrontation event.

24. The ventilation monitoring method as described in claim 20, characterized in that, Based on the recognition results of each algorithm module capable of identifying the same type of human-machine adversarial event, determine whether the same type of human-machine adversarial event identified by each algorithm module has occurred, including: Based on the ratio and / or quantity relationship between the first algorithm module and the second algorithm module, it is determined whether the same type of human-machine confrontation event identified by each algorithm module has occurred. The first algorithm module is the algorithm module that has identified the human-machine confrontation event, and the second algorithm module is the algorithm module that has not identified the human-machine confrontation event.

25. The ventilation monitoring method as described in claim 14, characterized in that, The ventilation parameter data includes at least one of airway pressure, airway flow rate, and gas volume; The physiological parameters include at least one of esophageal pressure, intragastric pressure, transdiaphragmatic pressure, carbon dioxide concentration, and diaphragmatic electrical activity.

26. A computer-readable storage medium, characterized in that, Includes a program that can be executed by a processor to implement the method as described in any one of claims 14-25.

Images