Methods, systems, and electronic devices for monitoring a patient during a perioperative period

Abstract

The application discloses a kind of applied to the monitoring method, system and electronic equipment of perioperative patient, the system includes: wearable information acquisition device and risk prediction device, wherein, risk prediction device includes: user physiological information acquisition unit, for obtaining user physiological information by wearable information acquisition device, physiological information includes: electrocardiogram information, pulse wave information, blood vessel information;User comprehensive physiological information generation unit is used to generate user comprehensive physiological information based on predetermined algorithm according to physiological information, comprehensive physiological information includes: blood pressure information, cardiac output information;Prediction unit is used to input comprehensive physiological information into pre-trained multi-modal illness classification model, and obtain the user potential risk information output by model, multi-modal illness classification model is based on stack method to analyze comprehensive physiological information;Risk information sending unit is used to send user potential risk information to monitoring center.

Description

Technical Field

[0001] This application belongs to the field of medical technology, specifically relating to a monitoring method, system, and electronic device for perioperative patients. Background Technology

[0002] In the medical field, enhanced monitoring refers to the close and advanced hemodynamic monitoring of patients during general anesthesia, emergency care, and the diagnosis and treatment of critically ill patients, as required by their condition or surgery, in order to promptly detect dangerous situations, manage them in a timely manner, and reduce the patient's risks.

[0003] Advanced hemodynamic monitoring is a core technology for intensive care and enhanced monitoring, including parameters such as continuous blood pressure, cardiac output, and cardiac function status, which are currently mainly obtained through invasive vascular catheterization or cardiac catheterization.

[0004] Currently, a large number of surgical procedures are performed globally each year, and the perioperative mortality rate is nearly 1%. Therefore, perioperative patients are the primary group requiring intensive monitoring. However, existing intensive monitoring technologies for perioperative patients still have many shortcomings:

[0005] 1. Enhanced perioperative patient monitoring requires traversing different work environments, from the preoperative general surgical ward to the operating room, and then to the postoperative recovery room, intensive care unit, or surgical ward. Existing monitoring technologies can only meet the needs of one or two of these scenarios, and there are monitoring gaps during patient location transitions.

[0006] 2. Existing monitoring technologies can only analyze single monitoring parameters in isolation, such as whether low blood pressure or arrhythmia has occurred. They lack intelligent analysis of multiple monitoring parameters, thus failing to accurately predict patient risk.

[0007] In other words, current monitoring technologies for perioperative patients suffer from limitations such as limited application scenarios and an inability to comprehensively analyze monitoring data, making it difficult to identify the true risks of patients and even more difficult to accurately predict their potential risks. Summary of the Invention

[0008] The purpose of this application is to provide a monitoring method, system, and electronic device for perioperative patients to solve at least one of the aforementioned problems.

[0009] In a first aspect, embodiments of this application provide a monitoring system for perioperative patients, the system comprising: a wearable information acquisition device and a risk prediction device, wherein the risk prediction device includes:

[0010] The user physiological information acquisition unit is used to acquire user physiological information through a wearable information acquisition device, the physiological information including: electrocardiogram information, pulse wave information, and vascular information;

[0011] The user comprehensive physiological information generation unit is used to generate user comprehensive physiological information based on the physiological information according to a predetermined algorithm. The comprehensive physiological information includes: blood pressure information and cardiac output information.

[0012] The prediction unit is used to input the comprehensive physiological information into a pre-trained multimodal disease classification model and obtain the user's potential risk information output by the multimodal disease classification model. The user's potential risk information includes: blood pressure risk information, cardiac risk information, and respiratory risk information. The multimodal disease classification model analyzes the comprehensive physiological information based on a stacking method.

[0013] The risk information sending unit is used to send the user's potential risk information to the monitoring center.

[0014] Secondly, embodiments of this application provide a monitoring method for perioperative patients, the method comprising:

[0015] The user's physiological information is acquired through a wearable information acquisition device, including: electrocardiogram information, pulse wave information, and vascular information;

[0016] Based on a predetermined algorithm, comprehensive physiological information of the user is generated according to the physiological information, including: blood pressure information and cardiac output information;

[0017] The comprehensive physiological information is input into a pre-trained multimodal disease classification model, and the potential risk information of the user output by the multimodal disease classification model is obtained. The potential risk information of the user includes: blood pressure risk information, cardiac risk information, and respiratory risk information. The multimodal disease classification model analyzes the comprehensive physiological information based on a stacking method.

[0018] The potential risk information of the user is sent to the monitoring center.

[0019] Thirdly, embodiments of this application provide an electronic device including a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the method described in the second aspect.

[0020] In this embodiment, a comprehensive physiological information of the user is generated based on the user's physiological information acquired by the wearable information acquisition device using a predetermined algorithm. This comprehensive physiological information is then input into a pre-trained multimodal disease classification model, and the model outputs potential risk information of the user. This potential risk information is then sent to a monitoring center for processing. Compared to related technologies, the wearable information acquisition device in this embodiment can acquire multiple monitoring data of the patient in real time in different application scenarios. Simultaneously, by comprehensively analyzing the monitoring data through the multimodal disease classification model, the potential risks of the patient can be accurately predicted, improving the enhanced monitoring effect of perioperative patients and reducing patient risks. Attached Figure Description

[0021] Figure 1 This is a structural block diagram of a monitoring system for perioperative patients provided in an embodiment of this application;

[0022] Figure 2 This is a schematic diagram of a wearable information acquisition device 1 provided in an embodiment of this application;

[0023] Figure 3 This is a schematic diagram of a risk prediction device 2 provided in an embodiment of this application;

[0024] Figure 4 This is another structural schematic diagram of the risk prediction device 2 provided in the embodiments of this application;

[0025] Figure 5 This is a flowchart of a monitoring system for perioperative patients provided in an embodiment of this application;

[0026] Figure 6 This is a flowchart of a monitoring method for perioperative patients provided in an embodiment of this application;

[0027] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0028] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0029] Given that current perioperative patient monitoring technologies suffer from limitations such as limited application scenarios and an inability to comprehensively analyze monitoring data, they fail to identify the true risks of patients or accurately predict potential risks. Therefore, this application provides a perioperative patient monitoring solution based on wearable monitoring technology and a multimodal disease classification model. This solution enables intelligent monitoring of perioperative patients, predicting potential risks, improving the effectiveness of enhanced perioperative patient monitoring, and reducing patient risks.

[0030] The monitoring scheme for perioperative patients provided in this application will be described in detail below with reference to the accompanying drawings, through specific embodiments and application scenarios.

[0031] Figure 1 This is a structural block diagram of a monitoring system for perioperative patients provided in an embodiment of this application, such as... Figure 1 As shown, the system includes a wearable information acquisition device 1 and a risk prediction device 2. The wearable information acquisition device 1 is worn on a corresponding location of the user (i.e., the patient) to acquire the user's physiological information, such as electrocardiogram (ECG) information, pulse wave information, and vascular information. The risk prediction device 2 is used to predict the user's potential risks based on the physiological information acquired by the wearable information acquisition device 1. This can improve the effectiveness of perioperative patient monitoring and reduce patient risks.

[0032] The following combination Figure 2 and Figure 3 The wearable information acquisition device 1 and the risk prediction device 2 will be described in detail.

[0033] In one possible implementation, such as Figure 2 As shown, the wearable information acquisition device 1 includes: a wearable electrocardiogram sensor 11, a pulse sensor 12, and a wearable arterial blood flow sensor 13, wherein:

[0034] Wearable ECG sensor 11 is used to collect ECG information of the wearer (i.e., the patient);

[0035] Pulse sensor 12 is used to collect the wearer's pulse wave information;

[0036] Wearable arterial blood flow sensor 13 is used to collect vascular information from the wearer.

[0037] The wearable ECG sensor 11, pulse sensor 12, and wearable arterial blood flow sensor 13 enable the real-time acquisition of various patient monitoring data in different application scenarios.

[0038] In one possible implementation, the wearable ECG sensor is based on three-lead ECG sensor technology and is designed as a single patch integrating three-lead electrodes, which transmits the collected ECG signals to the risk prediction device 2 via wired or wireless means.

[0039] In one possible implementation, the pulse sensor can be integrated into a wristband and cuff for easy wear by the patient. This pulse sensor is based on pulse wave photoplethysmography (PPG) analysis technology.

[0040] In one possible implementation, the wearable arterial blood flow sensor is designed as a wristband or cuff for easy patient wearing. This wearable arterial blood flow sensor is based on optical Doppler and ultrasonic Doppler technology sensors.

[0041] In practice, the wearable information acquisition device 1 may also include sensors that can collect other physiological information of the patient, such as sensors that collect electrocardiogram, respiration, pulse oxygen saturation.

[0042] In one possible implementation, the sensors in the wearable information acquisition device 1 can be configured in combination or individually, depending on the actual situation, and this application does not impose any restrictions on this.

[0043] like Figure 3 As shown, the risk prediction device 2 includes: a user physiological information acquisition unit 21, a user comprehensive physiological information generation unit 22, a prediction unit 23, and a risk information transmission unit 24, wherein:

[0044] User physiological information acquisition unit 21 is used to acquire user physiological information through a wearable information acquisition device, the physiological information including: electrocardiogram information, pulse wave information, and vascular information.

[0045] The vascular information includes: vascular length information, blood flow velocity information, vascular pressure information, etc.

[0046] User comprehensive physiological information generation unit 22 is used to generate user comprehensive physiological information based on the physiological information according to a predetermined algorithm. The comprehensive physiological information includes: blood pressure information and cardiac output information.

[0047] In practice, the aforementioned comprehensive physiological information may also include: electrocardiogram information, respiratory information, pulse oxygen saturation information, etc.

[0048] The generation and monitoring of the aforementioned electrocardiogram (ECG), respiratory, and pulse oxygen saturation information can employ relevant technologies, and this application does not impose any limitations on these technologies. For example, electrical signals obtained from an ECG sensor can be converted into an electrocardiogram (ECG information) and periodic changes in impedance (respiratory signals). Based on signals from an optical pulse sensor at the fingertips, the amplitude of pulse fluctuations and pulse oxygen saturation (SpO2) can be detected. The generation of blood pressure and cardiac output information within the comprehensive physiological information is described in detail below.

[0049] In one possible implementation, the aforementioned user comprehensive physiological information generation unit 22 includes: a blood pressure information determination module 221 and a cardiac output information determination module 222, wherein:

[0050] The blood pressure information determination module 221 is used to determine blood pressure information based on the time difference between the electrocardiogram information and the pulse wave information, as well as the blood vessel length information;

[0051] The cardiac output information determination module 222 is used to determine the cardiac output information based on the blood pressure information and the blood flow velocity information.

[0052] The aforementioned predetermined algorithm can also be called a novel physiological information processing algorithm, namely, the blood pressure information determination module 221 determines the blood pressure information and the cardiac output information determination module 222 determines the calculation method of the cardiac output information.

[0053] In one possible implementation, the blood pressure information determination module 221 includes: a pulse wave velocity determination submodule 2211 and a blood pressure information determination submodule 2212, wherein:

[0054] The pulse wave conduction velocity determination submodule 2211 is used to determine the pulse wave conduction velocity based on the time difference between the electrocardiogram information and the pulse wave information, as well as the blood vessel length information;

[0055] The blood pressure information determination submodule 2212 is used to determine the blood pressure information based on the pulse wave conduction velocity and the vascular pressure information.

[0056] Specifically, the pulse wave velocity determination submodule 2211 calculates the pulse wave velocity based on the time difference T1 between the electrocardiogram signal and the pulse wave signal, and the vessel length D1 from the root of the aorta to the pulse wave detection position. Subsequently, the blood pressure information determination submodule 2212 calculates real-time continuous blood pressure based on the positive correlation between pulse wave velocity and vascular pressure. Then, the cardiac output information determination module 222 calculates the cardiac output CO based on the blood pressure increment per heartbeat P1 = SBP - DBP, the blood flow velocity increment per heartbeat V1, and the positive correlation between P1 and V1 and cardiac output CO. Here, SBP is systolic blood pressure, and DBP is diastolic blood pressure.

[0057] Specifically, this can be achieved using the following formula:

[0058] Systolic blood pressure SBP = SBP C *K SBP *T1 -1 ;

[0059] Diastolic blood pressure DBP = DBP C *K DBP *D1*T1 -1 ;

[0060] Cardiac output CO = (SBP - DBP) * V1 * K CO .

[0061] Among them, SBP C K is a reference value for calculating systolic blood pressure. SBP DBP is the coefficient for calculating systolic blood pressure. C K serves as a reference value for diastolic blood pressure calculation. DBP K is the coefficient for calculating diastolic blood pressure. CO Calculate the coefficient for cardiac output. SBP C K SBP DBP C K DBP K CO The parameters can be obtained based on relevant operations in existing technologies. For example, they can be obtained by measuring large sample sizes of blood pressure at different levels in different populations, calculating correlation coefficients, calculating regression equations, fitting model curves and performing statistical analysis, and then performing independent sample validation.

[0062] The prediction unit 23 is used to input the comprehensive physiological information into a pre-trained multimodal disease classification model and obtain the user's potential risk information output by the model. The user's potential risk information includes: blood pressure risk information, cardiac risk information, and respiratory risk information. The multimodal disease classification model analyzes the comprehensive physiological information based on the stacking method.

[0063] The risk information sending unit 24 is used to send the user's potential risk information to the monitoring center for processing.

[0064] In practice, the risk information sending unit 24 can send potential risk information of users to the monitoring center through wireless or wired communication networks.

[0065] The user comprehensive physiological information generation unit 22 generates user comprehensive physiological information based on a predetermined algorithm and the user physiological information acquired by the user physiological information acquisition unit 21 through the wearable information acquisition device. Subsequently, the prediction unit 23 inputs the comprehensive physiological information into a pre-trained multimodal disease classification model and obtains the user's potential risk information output by the model. Then, the risk information sending unit 24 sends the user's potential risk information to the monitoring center for processing. Compared with related technologies, the wearable information acquisition device in this embodiment can realize the real-time acquisition of multiple monitoring data of patients in different application scenarios. At the same time, the monitoring data can be comprehensively analyzed through the multimodal disease classification model, thereby accurately predicting the patient's potential risks, improving the enhanced monitoring effect of perioperative patients, and reducing patient risks.

[0066] In one possible implementation, see Figure 4 As shown, the risk prediction device 2 also includes a model training unit 25, used to train the multimodal disease classification model.

[0067] In practice, a multimodal disease classification model can be the MultimodalTag multimodal disease classification model, developed based on Paddle 2.0 (an open-source deep learning platform). This model uses real patient monitoring data and integrates text, image, and digit modalities for multimodal label classification. Input data can include: electrocardiogram, blood pressure, heart rate, respiratory rate, cardiac output, pulse wave, target vessel blood flow velocity curve, etc.

[0068] In one possible implementation, the model training unit 25 may include: a historical information acquisition module 251, a feature extraction module 252, a sequence learning module 253, and a training module 254, wherein:

[0069] The historical information acquisition module 251 is used to acquire historical user information, which includes: historical blood pressure information, historical cardiac output information, historical heart rate information, historical respiratory information, historical pulse oxygen saturation, and historical risk information.

[0070] The feature extraction module 252 is used to perform feature extraction operations on the historical comprehensive physiological information based on a pre-trained Residual Network (ResNet) model.

[0071] In one embodiment, a residual network is used to extract high-level semantic features from the image, while ERNIE 1.0 (a semantic understanding technology and platform) can be used to extract textual features.

[0072] The sequence learning module 253 is used to perform sequence learning operations on historical comprehensive physiological information after feature extraction based on a pre-trained Long Short-Term Memory (LSTM) network.

[0073] In one embodiment, a separate LSTM network can be used to perform sequence learning on the image features, text features, and data features extracted by the feature extraction module 252.

[0074] The training module 254 is used to send the historical comprehensive physiological information after the sequence learning operation to the multimodal disease classification model for training. In response to the difference between the output historical user potential risk information and the historical risk information being less than a predetermined threshold (e.g., 95%), the training of the multimodal disease classification model is completed.

[0075] The aforementioned multimodal disease classification model employs a stacking multimodal data model fusion technique for cross-fusion, ultimately combining text, image, and data features. The classifier for the prediction part of this multimodal disease classification model uses a Sigmoid multi-label classifier, with severe hypotension, arrhythmia, cardiac arrest, respiratory failure, and respiratory arrest as prediction targets.

[0076] Figure 5 This is a flowchart illustrating a monitoring system for perioperative patients according to an embodiment of this application, such as... Figure 5 As shown, the process includes:

[0077] Step S501: In the early perioperative period, the patient wears a wearable information acquisition device 1, which is connected to the risk prediction device 2 via wired or wireless means.

[0078] In step S502, the user comprehensive physiological information generation unit 22 in the risk prediction device 2 applies a novel physiological information processing algorithm to calculate the patient's continuous blood pressure, cardiac output, heart rate, respiratory rate, pulse oxygen saturation, etc. in real time based on the physiological information obtained in step S501.

[0079] In step S503, the prediction unit 23, based on the multimodal disease classification model, applies artificial intelligence and deep learning algorithms to further analyze and process the monitoring parameters obtained in step S502, and intelligently identify and predict the patient's dangerous situations or events.

[0080] In step S504, the risk information sending unit 24 sends the information predicted in step S503 to the monitoring center in a timely manner according to program instructions or temporary instructions.

[0081] The above-mentioned process steps S501-S504 can start from before the operation, continue during the operation, and last for several days after the operation, to achieve seamless and complete monitoring of the patient.

[0082] The monitoring system provided in this application embodiment can conveniently, continuously, and without interruption provide intelligent intensive care monitoring for perioperative patients. At the same time, it can accurately and intelligently identify various dangerous information and promptly push it to clinicians and other third-party monitors, providing close monitoring for perioperative patients. This enables clinical staff to promptly detect and handle abnormal patient conditions, thereby improving the safety level of patients during the perioperative period.

[0083] It should be noted that the risk prediction device 2 in the embodiments of this application can be a device, or it can be a component, integrated circuit, or chip in a terminal. The device can be a mobile terminal or a non-mobile terminal, and the embodiments of this application do not specifically limit it.

[0084] The risk prediction device 2 in this embodiment can be a device with an operating system. This operating system can be Android, iOS, or other possible operating systems; this embodiment does not specifically limit its use.

[0085] Based on a similar inventive concept, this application also provides a monitoring method for perioperative patients, which is preferably used to realize the function of the above-mentioned risk prediction device.

[0086] Figure 6 This is a schematic flowchart of a monitoring method for perioperative patients according to an embodiment of this application, such as... Figure 6 As shown, the method includes steps S601-S604, wherein:

[0087] Step S601 involves acquiring the user's physiological information via a wearable information acquisition device. This physiological information includes electrocardiogram (ECG) information, pulse wave information, and vascular information. Step S601 enables the functionality of the aforementioned user physiological information acquisition unit 21.

[0088] The vascular information may include: vascular length information, blood flow velocity information, and vascular pressure information.

[0089] Step S602 involves generating comprehensive user physiological information based on a predetermined algorithm, including blood pressure and cardiac output information. This step S602 enables the functionality of the aforementioned comprehensive user physiological information generation unit 22.

[0090] In one implementation, step S602 specifically includes: determining blood pressure information based on the time difference between the electrocardiogram information and the pulse wave information, and the blood vessel length information; subsequently, determining the cardiac output information based on the blood pressure information and the blood flow velocity information.

[0091] In determining blood pressure information, the pulse wave conduction velocity can be determined based on the time difference between the electrocardiogram information and the pulse wave information, as well as the blood vessel length information; then, the blood pressure information can be determined based on the pulse wave conduction velocity and the blood vessel pressure information.

[0092] Step S603 involves inputting the comprehensive physiological information into a pre-trained multimodal disease classification model and obtaining the user's potential risk information output by the model. This potential risk information includes blood pressure risk information, cardiac risk information, and respiratory risk information. The multimodal disease classification model analyzes the comprehensive physiological information based on a stacking method. This step S603 enables the functionality of the prediction unit 23 described above.

[0093] Step S604 involves sending the potential user risk information to the monitoring center for processing. This step S604 enables the functionality of the risk information sending unit 24 described above.

[0094] By generating comprehensive physiological information about a user based on physiological information acquired by a wearable information acquisition device using a predetermined algorithm, and then inputting this comprehensive physiological information into a pre-trained multimodal disease classification model, the model outputs potential risk information about the user. This potential risk information is then sent to a monitoring center for processing. Compared to related technologies, the wearable information acquisition device in this embodiment can achieve real-time acquisition of various patient monitoring data in different application scenarios. Simultaneously, by comprehensively analyzing the monitoring data through a multimodal disease classification model, it is possible to accurately predict the patient's potential risks, improve the effectiveness of perioperative patient monitoring, and reduce patient risks.

[0095] In one implementation, the method further includes training the multimodal disease classification model, which enables the functionality of the model training unit 25. Specifically, the training process includes:

[0096] (1) Obtain historical user information, which includes: historical blood pressure information, historical cardiac output information, historical heart rate information, historical respiratory information, historical pulse oxygen saturation, and historical risk information;

[0097] (2) Feature extraction operation is performed on the historical comprehensive physiological information based on the pre-trained residual network model;

[0098] (3) Based on a pre-trained long short-term memory network, sequence learning operation is performed on the historical comprehensive physiological information after feature extraction;

[0099] (4) The historical comprehensive physiological information after the sequence learning operation is sent to the multimodal disease classification model for training. If the difference between the output historical user potential risk information and the historical risk information is less than a predetermined threshold, the training of the multimodal disease classification model is completed.

[0100] The steps of the monitoring method for perioperative patients provided in this application embodiment can realize the functions of the corresponding units, corresponding modules and corresponding sub-modules in the risk prediction device 2 in the above system embodiment, and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0101] Optional, such as Figure 7 As shown, this application embodiment also provides an electronic device 700, including a processor 710 and a memory 720. The memory 720 stores programs or instructions that can run on the processor 710. For example, when the electronic device 700 is a terminal, when the program or instructions are executed by the processor 710, they implement the various processes described above in the embodiments of the monitoring method for perioperative patients, and achieve the same technical effect. To avoid repetition, further details are omitted here.

[0102] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes described above in the embodiments of the monitoring method for perioperative patients and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0103] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0104] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described monitoring method embodiment for perioperative patients, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0105] Furthermore, this application also provides a computer program product, which includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor. When the program or instructions are executed by the processor, they implement the various processes described above in the embodiments of the monitoring method for perioperative patients and achieve the same technical effects. To avoid repetition, they will not be described again here.

[0106] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited. Functions may be performed in the order shown or discussed, or may be performed substantially simultaneously or in reverse order depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0107] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0108] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A monitoring system applied to a patient in a perioperative period, characterized in that, The system comprises a wearable information acquisition device and a risk prediction device, wherein the risk prediction device comprises: a user physiological information acquisition unit configured to acquire user physiological information through the wearable information acquisition device, the user physiological information comprising electrocardiogram information, pulse wave information, and blood vessel information, the blood vessel information comprising blood vessel length information, blood flow velocity information, and blood vessel pressure information; a user comprehensive physiological information generation unit configured to generate user comprehensive physiological information based on a predetermined algorithm and according to the user physiological information, the user comprehensive physiological information comprising blood pressure information and cardiac output information; a prediction unit configured to input the user comprehensive physiological information into a pre-trained multi-modal disease classification model and acquire user potential risk information output by the multi-modal disease classification model, wherein the user potential risk information comprises blood pressure risk information, cardiac risk information, and respiratory risk information, and the multi-modal disease classification model analyzes the user comprehensive physiological information based on a stacking method; a risk information sending unit configured to send the user potential risk information to a monitoring center; wherein the user comprehensive physiological information generation unit comprises a blood pressure information determination module configured to determine blood pressure information according to a time difference between the electrocardiogram information and the pulse wave information and the blood vessel length information, and a cardiac output information determination module configured to determine the cardiac output information according to the blood pressure information and the blood flow velocity information; the blood pressure information determination module comprises a pulse wave transmission velocity determination submodule configured to determine pulse wave transmission velocity according to the time difference between the electrocardiogram information and the pulse wave information and the blood vessel length information, and a blood pressure information determination submodule configured to determine the blood pressure information according to the pulse wave transmission velocity and the blood vessel pressure information; wherein the cardiac output information determination module comprises a cardiac output calculation method based on a blood pressure increment P1=SBP-DBP and a blood flow velocity increment V1 of each heart beat, expressed as: Cardiac output ; wherein K CO represents a cardiac output calculation coefficient; SBP is systolic blood pressure, expressed as ; wherein SBP C is a systolic blood pressure calculation reference quantity, K SBP is a systolic blood pressure calculation coefficient, and T1 represents a pulse wave signal time difference; DBP is diastolic blood pressure, expressed in mmHg ; wherein DBP C is a diastolic blood pressure calculation reference quantity, K DBP is a diastolic blood pressure calculation coefficient, and D1 represents the length of the blood vessel from the aortic root of the heart to the position where the pulse wave is detected.

2. The system of claim 1, wherein, The risk prediction device further comprises a model training unit configured to train the multi-modal disease classification model, the model training unit comprises: a historical information acquisition module configured to acquire historical user information, the historical user information comprising historical blood pressure information, historical cardiac output information, historical heart rate information, historical respiratory information, historical pulse oxygen saturation, and historical risk information; a feature extraction module configured to perform feature extraction on the historical user information based on a pre-trained residual network model; a sequence learning module configured to perform sequence learning on the historical user information after the feature extraction based on a pre-trained long short-term memory network; a training module configured to send the historical user information after the sequence learning to the multi-modal disease classification model for training, and in response to a difference between output historical user potential risk information and the historical risk information being less than a predetermined threshold, the multi-modal disease classification model is trained to be complete.

3. The system of claim 1, wherein, The wearable information acquisition device comprises: a wearable electrocardiogram sensor configured to acquire the electrocardiogram information; a pulse sensor configured to collect the pulse wave information; a wearable arterial blood flow sensor configured to collect the blood vessel information.

4. A method of applying a monitoring to a patient during a perioperative period, characterized in that, The method comprises: acquiring user physiological information by a wearable information acquisition device, the user physiological information comprising electrocardiogram information, pulse wave information, and blood vessel information, the blood vessel information comprising blood vessel length information, blood flow velocity information, and blood vessel pressure information; generating user comprehensive physiological information based on a predetermined algorithm according to the user physiological information, the user comprehensive physiological information comprising blood pressure information and cardiac output information; inputting the user comprehensive physiological information into a pre-trained multi-modal condition classification model and acquiring user potential risk information output by the multi-modal condition classification model, wherein the user potential risk information comprises blood pressure risk information, cardiac risk information, and respiratory risk information, the multi-modal condition classification model analyzes the user comprehensive physiological information based on a stacking method, and the user potential risk information is sent to a monitoring center; wherein generating user comprehensive physiological information based on a predetermined algorithm according to the user physiological information comprises: determining blood pressure information according to a time difference between the electrocardiogram information and the pulse wave information and the blood vessel length information, and determining the cardiac output information according to the blood pressure information and the blood flow velocity information; determining blood pressure information according to a time difference between the electrocardiogram information and the pulse wave information and the blood vessel length information comprises: determining pulse wave conduction velocity according to a time difference between the electrocardiogram information and the pulse wave information and the blood vessel length information, and determining the blood pressure information according to the pulse wave conduction velocity and the blood vessel pressure information; wherein determining cardiac output information comprises calculating cardiac output according to a blood pressure increment P1 = SBP-DBP of each heartbeat and a blood flow velocity increment V1 of each heartbeat, and is expressed as: Cardiac output ; In the formula, K CO represents a cardiac output calculation coefficient; SBP is systolic blood pressure, expressed as ; wherein SBP C is a systolic blood pressure calculation reference quantity, K SBP is a systolic blood pressure calculation coefficient, and T1 represents a pulse wave signal time difference; DBP is diastolic blood pressure, expressed in mmHg ; wherein DBP C is a diastolic blood pressure calculation reference quantity, K DBP is a diastolic blood pressure calculation coefficient, and D1 represents the length of the blood vessel from the aortic root of the heart to the location where the pulse wave is detected.

5. The method of claim 4, wherein, The multi-modal condition classification model is trained in the following manner: acquiring historical user information, the historical user information comprising historical blood pressure information, historical cardiac output information, historical heart rate information, historical respiratory information, historical pulse oxygen saturation, and historical risk information; performing feature extraction operation on the historical user information based on a pre-trained residual network model; performing sequence learning operation on the historical user information after the feature extraction operation based on a pre-trained long short-term memory network; sending the historical user information after the sequence learning operation to the multi-modal condition classification model for training, and in response to a difference between output historical user potential risk information and the historical risk information being less than a predetermined threshold, the multi-modal condition classification model is trained to be complete.

6. An electronic device, comprising: a processor and a memory, the memory storing programs or instructions executable on the processor, the programs or instructions being executed by the processor to implement the steps of the monitoring method for a patient in a perioperative period according to any one of claims 4 to 5.

Images