NPPV bidirectional switching method and system based on acute exacerbation intelligent early warning

Abstract

The invention discloses an NPPV bidirectional switching method and system based on acute exacerbation intelligent early warning. The method comprises the steps that physiological indexes of a patient and running parameters of a breathing machine are continuously collected through a multi-source sensor; performing early warning according to the physiological indexes of the patient and the running parameters of the respirator; when the alarm signal is generated, the NPPV mode is automatically switched from low intensity to high intensity; and when the early warning release condition meets a preset condition, automatically switching the NPPV mode from high intensity to low intensity. According to the application, the re-hospitalization rate and the case fatality rate of COPD patients can be reduced, adverse reactions such as serious abdominal distension can be reduced, the problem that current NPPV treatment is not reasonable is solved, and the long-term use compliance of NPPV is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of noninvasive positive pressure ventilation (NPPV) management technology, and more specifically to a bidirectional switching method and system for NPPV based on intelligent early warning of acute exacerbations. Background Technology

[0002] Patients with chronic obstructive pulmonary disease (COPD) and chronic respiratory failure often require long-term home-based noninvasive positive pressure ventilation (NPPV) to maintain ventilatory function and improve quality of life. However, current NPPV treatment strategies suffer from a "single-intensity selection" flaw, making it difficult to balance clinical efficacy and safety, as shown below: In low-intensity NPPV strategies, the positive pressure support intensity is lower, which can reduce adverse reactions such as abdominal distension, severe air leakage, and dry mouth, nose, and throat, but the ventilation improvement effect is limited, and there is no significant advantage in reducing readmission rate and mortality.

[0003] While high-intensity NPPV strategies can significantly improve ventilation, reduce the work of breathing, alleviate respiratory muscle fatigue, and reduce readmission and mortality rates, long-term use can significantly increase the risks of abdominal distension, severe air leakage leading to patient-ventilator asynchrony, and reduce patient compliance.

[0004] Although recent studies have confirmed that a phased strategy of using high-intensity non-invasive ventilation (NPPV) during acute exacerbations of COPD (AECOPD) and low-intensity NPPV during stable phases is as effective as continuous high-intensity NPPV with fewer adverse reactions, this approach still has two key issues: 1) It relies on medical staff to manually determine the stage of the disease and manually adjust parameters, which cannot respond to the dynamic changes of the patient in real time; 2) Manual intervention may delay the optimal intervention period in the early stage of acute exacerbation, and it is difficult to reduce the treatment level in time after the condition is relieved.

[0005] Therefore, in summary, how to solve the above-mentioned technical problems is an urgent issue that needs to be addressed by those skilled in the art. Summary of the Invention

[0006] In view of this, the present invention provides a method and system for bidirectional switching of NPPV based on intelligent early warning of acute exacerbation, which aims to achieve intelligent switching of NPPV intensity according to changes in the patient's condition, thereby balancing clinical efficacy and safety, optimizing home-based NPPV management for COPD, and improving long-term patient compliance.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: First, this application provides a bidirectional switching method for NPPV based on intelligent early warning of acute exacerbations, the steps of which include: The patient's physiological indicators and ventilator operating parameters are continuously collected through multi-source sensors; Early warnings are issued based on the patient's physiological indicators and ventilator operating parameters; When an alarm signal is generated, the NPPV mode is automatically switched from low intensity to high intensity. When the conditions for lifting the warning are met, the NPPV mode will automatically switch from high intensity to low intensity; among which, In low-intensity or high-intensity modes, the positive pressure in the airway during inspiratory phase is dynamically adjusted based on the target tidal volume range and changes in physiological indicators.

[0008] In one optional embodiment, multiple sensors are used to continuously collect patient physiological indicators and ventilator operating parameters. The patient's physiological indicators include one or more of the following: heart rate (HR), blood pressure (BP), oxygen saturation (SpO2), and end-tidal carbon dioxide concentration (PETCO2). The ventilator operating parameters include one or more of the following: respiratory rate (RR), tidal volume (VT), minute ventilation (MV), and leakage volume (Vleak).

[0009] In one optional embodiment, the preset condition is that the warning information reaches the cancellation threshold and continues for more than 48 hours.

[0010] In one optional embodiment, an acute exacerbation early warning model is constructed, and the acute exacerbation early warning model is used to issue early warnings based on the patient's physiological indicators and ventilator operating parameters; The acute aggravation early warning model adopts a deep learning architecture that integrates 1D-CNN, LSTM, attention mechanism and MoE weighting algorithm to analyze multi-source sensor data in real time to generate early warning signals.

[0011] In one optional embodiment, dynamic adjustment of the positive airway pressure during inspiratory phase includes: The target tidal volume range for the low-intensity mode is 6–10 ml / kg, and the target tidal volume range for the high-intensity mode is 10–15 ml / kg. When the end-tidal carbon dioxide concentration (PETCO2) increases by more than 5% and the respiratory rate (RR) or heart rate (HR) increases by more than 5% for 24 hours, the inspiratory positive airway pressure (IPAP) is automatically increased until the end-tidal carbon dioxide concentration (PETCO2) returns to baseline. When the end-tidal carbon dioxide concentration (PETCO2) decreases by more than 5% and the respiratory rate (RR) and heart rate (HR) remain stable at baseline for 24 hours, the intra-inspiratory positive airway pressure (IPAP) is automatically reduced within the target VT range.

[0012] In one optional embodiment, the intra-inspiratory positive pressure (IPAP) is adjusted at a rate of 1–2 cmH2O per adjustment, with an interval of at least 2 hours between two consecutive dynamic adjustments.

[0013] In one optional embodiment, the upper limit of the positive airway pressure (IPAP) adjustment during the inspiratory phase in the low-intensity mode is 20 cmH2O; and the upper limit of the positive airway pressure (IPAP) adjustment during the inspiratory phase in the high-intensity mode is 30 cmH2O.

[0014] In one optional embodiment, when the positive airway pressure (IPAP) is automatically increased during the inspiratory phase, if the end-tidal carbon dioxide concentration (PETCO2) still does not improve after increasing the preset amount within the upper limit of the adjustment range, the increase operation is terminated.

[0015] Secondly, this application also provides an NPPV bidirectional switching system based on acute exacerbation intelligent early warning, which applies the NPPV bidirectional switching method based on acute exacerbation intelligent early warning as described above, including: Multi-sensor monitoring network: Used to continuously collect patient physiological indicators and ventilator operating parameters through multi-source sensors; the network integrates smart respiratory monitoring belt, smart wristband, millimeter-wave remote sensing, PETCO2 monitor and ventilator sensors to collect physiological and equipment data in real time; Intelligent early warning module: used to issue early warnings based on the patient's physiological indicators and ventilator operating parameters; The execution control module is used to respond to alarm signals, automatically switching the NPPV mode from low intensity to high intensity when an alarm signal is generated; respond to cancellation signals, automatically switching the NPPV mode from high intensity to low intensity when the warning cancellation conditions meet the preset conditions; and dynamically adjust the positive airway pressure (IPAP) parameters during inspiratory phase, including dynamically adjusting the positive airway pressure during inspiratory phase according to the target tidal volume range and changes in physiological indicators in low or high intensity modes.

[0016] In one optional embodiment, the system further includes a verification module for verifying the switching reliability under different respiratory mechanics using a simulated lung model; and for assessing the safety of barotrauma and histopathology using a COPD animal model.

[0017] This invention discloses a bidirectional switching method and system for NPPV based on intelligent early warning of acute exacerbations. Compared with the prior art, the bidirectional switching strategy provided by this invention can achieve precise, dynamic, and adaptive adjustment of NPPV treatment. It can not only reduce the readmission rate and mortality rate of COPD patients, but also reduce the occurrence of adverse reactions such as severe abdominal distension. This solves the problem of the current "irrational" NPPV treatment and significantly improves the long-term use compliance of NPPV.

[0018] Furthermore, this application achieves multi-dimensional data monitoring through multi-sensor fusion, breaking through the limitations of traditional single monitoring methods in terms of time and space coverage, and more comprehensively and meticulously reflecting the patient's physiological state and disease changes, providing a more sufficient basis for precise intervention. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1 This is a flowchart of the NPPV bidirectional switching method based on acute exacerbation intelligent early warning according to the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0023] Specifically, addressing the bottlenecks of existing COPD home-based NPPV treatment strategies, such as "irrationality, lack of dynamic adjustment mechanisms, and difficulty in balancing efficacy and adherence," this invention provides a bidirectional switching strategy for high and low intensities of NPPV based on an intelligent early warning model for acute exacerbations. The aim is to implement low-intensity NPPV when the patient's condition is stable, intelligently switch to high-intensity NPPV when an acute exacerbation risk warning is issued, and reverse the switching back to low-intensity NPPV after the condition improves. Specific embodiments of this invention are as follows: Example 1 This embodiment provides a bidirectional switching method for NPPV based on intelligent early warning of acute exacerbations. The core of this method lies in tightly coupling precise intelligent early warning of acute exacerbations with dynamic adjustment of NPPV intensity, achieving intelligent, personalized, and adaptive treatment. In this embodiment, the switching steps are as follows: Figure 1 ,include: The patient's physiological indicators and ventilator operating parameters are continuously collected through multi-source sensors; Early warnings are issued based on the patient's physiological indicators and ventilator operating parameters; When an alarm signal is generated, the NPPV mode is automatically switched from low intensity to high intensity. When the conditions for lifting the warning are met, the NPPV mode will automatically switch from high intensity to low intensity; among which, In low-intensity or high-intensity modes, the positive pressure in the airway during inspiratory phase is dynamically adjusted based on the target tidal volume range and changes in physiological indicators.

[0024] In this embodiment, Step 1: Continuously collect patient physiological indicators and ventilator operating parameters using multi-source sensors; Among them, the patient's physiological indicators include heart rate (HR), blood pressure (BP), blood oxygen saturation (SpO2), and end-tidal carbon dioxide concentration (PETCO2); the ventilator operating parameters include respiratory rate (RR), tidal volume (VT), minute ventilation (MV), and air leakage (Vleak).

[0025] Existing monitoring systems primarily focus on ventilator parameters, lacking systematic monitoring of key physiological indicators. This invention, through multi-sensor fusion, significantly enriches the dimensions of monitoring data, enabling a more comprehensive and detailed reflection of the patient's physiological state and disease changes, thus providing a more robust basis for precise assessment and intervention.

[0026] Step 2: Issue early warnings based on the patient's physiological indicators and ventilator operating parameters; This embodiment constructs an acute exacerbation early warning model, which is used to issue early warnings based on the patient's physiological indicators and ventilator operating parameters. The acute aggravation early warning model adopts a deep learning architecture that integrates 1D-CNN, LSTM, attention mechanism and MoE weighting algorithm to analyze multi-source sensor data in real time to generate early warning signals.

[0027] In one specific embodiment, the model design includes: 1D-CNN Feature Extraction: Leveraging the ability of one-dimensional convolutional neural networks (1D-CNN) to capture local features in multi-source data, this method performs feature extraction and preliminary classification on preprocessed multi-sensor data. Through multi-level processing using convolutional and pooling layers, deep feature representations are generated, providing rich input for subsequent models. Optimizing network structure and hyperparameter settings such as kernel size, number of convolutional layers, and learning rate improves the model's feature extraction and representation capabilities. LSTM Temporal Modeling: Deep features extracted by 1D-CNN are input into a Long Short-Term Memory (LSTM) network to capture long-term dependencies in the data and build a dynamic temporal model. The LSTM network, through memory units and gating mechanisms, effectively captures dynamic features in long-term sequences, which is crucial for predicting time-dependent acute exacerbations. Optimizing the number of layers and hidden units in the LSTM network improves model performance. Multi-head attention mechanism: Introducing an attention mechanism enables the model to focus on critical moments and important features, such as subtle changes in certain physiological indicators before acute exacerbations. By calculating feature weights, the role of high-impact features in disease prediction is highlighted, thereby ensuring more accurate and efficient information processing. Hybrid Expert Model (MoE) Weighted Approach: Based on MoE, this approach uses model averaging techniques to optimally weight different algorithms, achieving accurate prediction of acute exacerbations. The MoE architecture dynamically activates and integrates multiple "expert" sub-models (e.g., for infectious acute exacerbations, non-infectious acute exacerbations, etc.) according to different physiological states or subtypes of patients, and intelligently allocates weights through a gating network. This enables ultra-early, high-precision, and personalized prediction of COPD acute exacerbations, significantly improving the specificity and sensitivity of early warning. This model design better addresses the characteristics of COPD—high heterogeneity and complex and diverse causes of acute exacerbations—solving the current bottleneck of "inaccurate" early warning systems.

[0028] In some implementations, 1D-CNN, LSTM, multi-head attention mechanism and MoE are connected in sequence. MoE includes a gating network, multiple expert sub-models and a weighted fusion module. The outputs of the gating network and multiple expert sub-models are connected to the input of the weighted fusion module. In one embodiment, the multiple expert sub-models include an expert sub-model for acute exacerbations of infectious COPD and an expert sub-model for acute exacerbations of non-infectious COPD; An expert sub-model for acute exacerbations of infectious COPD was used to predict the probability of patients experiencing future acute exacerbations of infectious COPD. An expert sub-model for non-infectious COPD exacerbations was used to predict the probability of patients experiencing future non-infectious COPD exacerbations. Gated networks are used to obtain the weighting coefficients W1 for probability P1 and W2 for probability P2; The weighted fusion module is used to perform weighted fusion of probabilities P1 and P2 based on weighting coefficients W1 and W2 respectively, in order to obtain the probability P of the patient experiencing an acute exacerbation of COPD in the future. If the probability P is greater than the preset threshold, it is assumed that the patient will experience an acute exacerbation of COPD in the future, and an alarm signal is generated.

[0029] This model can capture changes in the condition more comprehensively and in detail. By further integrating with multi-source data, it can effectively analyze the dynamic evolution trend of physiological indicators and their coupling mechanism with acute exacerbations, thereby achieving ultra-early and high-precision early warning.

[0030] Step 3: When an alarm signal is generated, the NPPV mode is automatically switched from low intensity to high intensity; that is, when the acute exacerbation warning model triggers an alarm, the intelligent switching from low intensity NPPV to high intensity NPPV is automatically initiated; this mechanism ensures that the ventilation support intensity can be rapidly increased in the early stage of disease deterioration to cope with the physiological challenges brought about by acute exacerbation.

[0031] Step 4: When the warning cancellation conditions meet the preset conditions, automatically switch the NPPV mode from high intensity to low intensity; This application establishes an early warning information tracking program and a cancellation mechanism. When the early warning information reaches the cancellation threshold and persists for ≥48 hours, a reverse switch from high-intensity NPPV to low-intensity NPPV will be initiated. This ensures that ventilation intensity can be reduced in a timely manner after the patient's condition stabilizes or improves, thereby reducing adverse reactions and improving patient comfort and compliance.

[0032] The bidirectional switching strategy provided by this invention can achieve precise, dynamic, and adaptive adjustment of NPPV treatment, solving the problem of "irrationality" in current NPPV treatment and maximizing treatment effectiveness and comfort.

[0033] Furthermore, in a preferred embodiment, the present invention constructs intelligent operating procedures for low-intensity NPPV and high-intensity NPPV respectively, and provides a method for dynamically adjusting NPPV parameters based on real-time monitoring data. In a specific embodiment: 1) Low-intensity NPPV intelligent operation: The ventilator is started and operated based on clinically guided parameter settings; During the inspiratory phase, the intra-airway positive pressure (IPAP) is continuously optimized at an adjustment rate of 1-2 cmH2O per cycle (up to 20 cmH2O) to ensure that the target tidal volume (VT) remains stable at 6-10 ml / kg. When the end-tidal carbon dioxide concentration PETCO2 increases by >5% and the respiratory rate RR or heart rate HR increases by >5% and persists for more than 24 hours, the intra-inspiratory positive airway pressure (IPAP) is automatically increased at a rate of 1-2 cmH2O / blow. The adjustment is performed after ≥2 hours until the end-tidal carbon dioxide concentration PETCO2 returns to the baseline level. If the end-tidal carbon dioxide concentration (PETCO2) does not improve after increasing the positive intra-inspiratory airway pressure (IPAP) by a maximum of 5 cmH2O within the upper limit range, then it will not be increased further. When the end-tidal carbon dioxide concentration (PETCO2) drops below the baseline level by more than 5% and the respiratory rate (RR) and heart rate (HR) remain stable at the baseline level for more than 24 hours, the inspiratory positive airway pressure (IPAP) is automatically reduced at a rate of 1-2 cmH2O / breath within the target VT range of 6-10 ml / kg. The adjustment is performed again after ≥2 hours until the end-tidal carbon dioxide concentration (PETCO2) returns to the baseline level.

[0034] 2) High-intensity NPPV intelligent operation: The ventilator was started and operated with the low-intensity NPPV parameters set before the switch as a baseline. During the inspiratory phase, the intra-inspiratory positive airway pressure (IPAP) is continuously optimized at an adjustment rate of 1-2 cmH2O per cycle (upper limit 30 cmH2O) to ensure that the target VT is stable at 10-15 ml / kg. When the end-tidal carbon dioxide concentration (PETCO2) increases by >5% and the respiratory rate (RR) or heart rate (HR) increases by >5% for more than 24 hours, the inspiratory positive airway pressure (IPAP) is automatically increased at a rate of 1-2 cmH2O / bsp. Readjustment is performed after ≥2 hours until the end-tidal carbon dioxide concentration (PETCO2) returns to baseline. If the end-tidal carbon dioxide concentration (PETCO2) does not improve after increasing the inspiratory positive airway pressure (IPAP) by a maximum of 5 cmH2O within the upper limit, further increases are not permitted. When the end-tidal carbon dioxide concentration (PETCO2) drops below the baseline level by more than 5% and the respiratory rate (RR) or heart rate (HR) remains stable at the baseline level for more than 24 hours, the intra-inspiratory positive airway pressure (IPAP) is automatically reduced at a rate of 1-2 cmH2O / breath within the target VT range of 10-15 ml / kg. The adjustment is performed again after ≥2 hours until the end-tidal carbon dioxide concentration (PETCO2) returns to the baseline level.

[0035] Example 2: This application also provides an NPPV bidirectional switching system based on acute exacerbation intelligent early warning, which applies the NPPV bidirectional switching method based on acute exacerbation intelligent early warning as described above, including: Multi-sensor monitoring network: used to continuously collect patient physiological indicators and ventilator operating parameters through multi-source sensors; in this embodiment, the multi-sensor monitoring network integrates a non-invasive ventilator, a smart respiratory monitoring belt, a smart wristband, millimeter-wave remote sensing, and a PETCO2 monitor to achieve full-time, multi-dimensional coverage of key physiological indicators, overcoming the limitations of existing single monitoring methods; among which, Non-invasive ventilator: As a core device, it monitors ventilator parameters in real time during NPPV operation, including inspiratory positive airway pressure (IPAP), expiratory positive airway pressure (EPAP), respiratory rate (RR), inspiratory flow rate (IFR), tidal volume (VT), leakage volume (LVV), and minute ventilation (MV). These parameters directly reflect the effectiveness of ventilatory support and the patient's respiratory mechanics.

[0036] Intelligent respiratory monitoring strap: Used to monitor respiratory rate (RR) and tidal volume (VT) during NPPV weaning. This device solves the problem of data loss when patients are weaned from traditional ventilators, ensuring continuous monitoring and covering respiratory physiological data of patients in non-ventilated states.

[0037] Smart bracelet: Continuously monitors heart rate (HR), blood pressure (BP), and blood oxygen saturation (SpO2) 24 hours a day. The introduction of smart bracelets makes continuous monitoring of circulation and oxygenation status possible, providing important evidence for assessing the patient's overall physiological condition, especially during off-the-machine activities or sleep.

[0038] Millimeter-wave remote sensing: 24-hour continuous monitoring of heart rate (HR) and respiratory rate (RR). Millimeter-wave remote sensing technology provides a non-contact, imperceptible monitoring method, further enhancing patient compliance and data continuity, and is especially suitable for nighttime sleep monitoring, avoiding the constraints of wearable devices.

[0039] PETCO2 monitor: Connected to the ventilator tubing and mask, it monitors the end-tidal carbon dioxide concentration (PETCO2) during NPPV operation. PETCO2 is an important indicator for assessing ventilation efficiency and carbon dioxide retention; its introduction makes the assessment of patient ventilation status more comprehensive and accurate.

[0040] The sensors described in this application are not simply stacked, but meticulously designed to address the shortcomings of existing monitoring systems, such as insufficient data dimensionality and limited temporal and spatial coverage. For example, ventilator parameters are used for direct assessment of ventilation status, smart wristbands and millimeter-wave remote sensing are used for continuous monitoring of physiological status throughout the day, smart respiratory monitoring bands are used to fill data gaps during weaning, and end-tidal carbon dioxide concentration (PETCO2) is used to provide key indicators of ventilation efficiency. This complementary and partially redundant design (e.g., HR and RR can be monitored jointly by smart wristbands and millimeter-wave remote sensing) significantly improves the comprehensiveness, continuity, and reliability of the data, ensuring high-quality monitoring data can be obtained in various home scenarios.

[0041] Intelligent early warning module: used to issue early warnings based on the patient's physiological indicators and ventilator operating parameters; The execution control module is used to respond to alarm signals, automatically switching the NPPV mode from low intensity to high intensity when an alarm signal is generated; respond to cancellation signals, automatically switching the NPPV mode from high intensity to low intensity when the warning cancellation conditions meet the preset conditions; and adjust the positive airway pressure (IPAP) parameters during inspiratory phase, including dynamically adjusting the positive airway pressure during inspiratory phase according to the target tidal volume range and changes in physiological indicators in low or high intensity modes.

[0042] In this embodiment, the intelligent early warning module and the execution control module are implemented in the same way as in Embodiment 1, so they will not be described again here.

[0043] In one optional embodiment, the system further includes a verification module for verifying the switching reliability under different respiratory mechanics using a simulated lung model; and for assessing the safety of barotrauma and histopathology using a COPD animal model.

[0044] In one embodiment, the reliability of switching under different respiratory mechanics is verified by simulating a lung model, including: using the ASL5000 to actively simulate the lung and construct multiple combinations of airway resistance (5, 10, 20, 50 cmH2O / L*s) and elastic resistance (20, 40, 60, 80 cmH2O / L) to simulate a highly complex respiratory mechanics model.

[0045] Furthermore, high-intensity NPPV and low-intensity NPPV were implemented for 24 hours on this model to verify the reliability of NPPV operation. Simultaneously, a complex and variable scenario of important physiological indicators was simulated, setting multiple combinations of various early warning indicators reaching the early warning standard to verify the reliability of the switch from low-intensity NPPV to high-intensity NPPV. In addition, the reliability of the reverse switch from high-intensity NPPV to low-intensity NPPV was verified by simulating early warning information composed of different early warning indicators reaching the threshold cancellation standard for ≥48 hours.

[0046] In one embodiment, assessing the safety of barotrauma and histopathology using a COPD animal model includes: inducing a COPD rat model using smoke exposure and confirming successful model establishment through pulmonary function tests and changes in lung structure.

[0047] This invention utilizes a proprietary NPPV ventilation system for laboratory mice. COPD rat models and normal lung models underwent NPPV intervention for 24 hours using high-intensity strategies (IPAP 25 cmH2O) and low-intensity strategies (IPAP 15 cmH2O), respectively, with a blank control group included. Experimental observations will focus on barotrauma, verifying safety from multiple perspectives including histopathological damage, oxidative stress, apoptosis, tissue and plasma inflammatory factor expression, respiratory mechanics, ventilation, and oxygenation status.

[0048] Furthermore, this application has undergone large-scale clinical validation through a multicenter randomized controlled trial (RCT). Two hundred COPD patients undergoing home-based NPPV will be randomly assigned to an intervention group (receiving this intelligent two-way switching strategy) and a control group (receiving conventional low-intensity NPPV) for one year. The study will assess indicators such as 1-year readmission rate, mortality, number of acute exacerbations, incidence of adverse reactions, and patients' quality of life to clarify the clinical efficacy and safety of this strategy.

[0049] The beneficial effects of this invention include: 1. Significantly Improved Clinical Efficacy and Safety: Existing NPPV treatment strategies present a contradiction between the efficacy and adverse reactions of high-intensity NPPV. This invention innovatively proposes a bidirectional switching strategy combining low-intensity NPPV during stable periods and high-intensity NPPV during acute exacerbations. This strategy is based on important research findings published in JAMA by the research team, namely, the combined use of routine low-intensity NPPV and high-intensity NPPV during disease exacerbations is as effective as continuous use of high-intensity NPPV. Through RCT clinical validation, this strategy is expected to reduce the readmission rate of COPD patients by ≥20% and is projected to achieve a relative risk reduction of ≥15% in all-cause mortality. Simultaneously, by using low-intensity NPPV during stable periods, the occurrence of adverse reactions such as abdominal distension and severe gas leakage can be significantly reduced, thereby balancing clinical efficacy and safety.

[0050] 2. Achieving Precision and Adaptability in NPPV Treatment: The core of this strategy lies in leveraging the precise guidance of an intelligent early warning model for acute exacerbations to achieve automatic and dynamic adjustment of NPPV intensity. When the model warns of an acute exacerbation risk, the system automatically triggers a switch from low to high NPPV intensity, ensuring timely and adequate ventilatory support for the patient in the early stages of deterioration. Conversely, when the patient's condition stabilizes or improves, the system intelligently guides the NPPV intensity back down to a low level, avoiding overventilation and unnecessary side effects. This intelligent and adaptive intervention paradigm addresses the "inappropriate" nature of traditional NPPV treatment, maximizing both treatment effectiveness and comfort.

[0051] 3. Significantly Improved Long-Term Patient Adherence: Long-term high-intensity NPPV may reduce patient adherence due to adverse reactions. This invention, through intelligent bidirectional switching, significantly reduces the occurrence of adverse reactions, especially common discomforts such as abdominal distension, while ensuring efficacy. This will greatly improve patients' long-term adherence to NPPV, ensure treatment continuity and effectiveness, and ultimately improve patients' long-term prognosis and quality of life.

[0052] 4. Constructing a Closed-Loop Management System for "Monitoring-Early Warning-Treatment": This strategy is a crucial link in the closed-loop management system for COPD home-based NPPV (Non-Positive Pathway Vehicle) management. It deeply integrates a multi-sensor fusion real-time monitoring system, an intelligent early warning model for acute exacerbations, and NPPV treatment strategies, forming a complete intelligent home care pathway. This provides a new approach and technological path for COPD home management, potentially reducing reliance on and consumption of medical resources, optimizing resource allocation, and significantly improving the accessibility of COPD diagnosis and treatment in primary and remote areas. The NPPV high-low intensity bidirectional switching strategy proposed in this invention, as well as the intelligent early warning model upon which it relies, both have promising prospects for product commercialization. Its clinical validation results can provide a scientific basis for updating COPD treatment guidelines and formulating related health policies, and offer a technological paradigm for smart healthcare and chronic disease management, ultimately achieving significant economic and social benefits and broad translational application value.

[0053] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0054] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A bidirectional switching method for NPPV based on intelligent early warning of acute exacerbations, characterized in that, The patient's physiological indicators and ventilator operating parameters are continuously collected through multi-source sensors; Early warning system based on patient physiological indicators and ventilator operating parameters: When an alarm signal is generated, the NPPV mode is automatically switched from low intensity to high intensity. When the conditions for lifting the warning are met, the NPPV mode will automatically switch from high intensity to low intensity; among which, In low-intensity or high-intensity mode, the positive pressure in the airway during inspiratory phase is dynamically adjusted based on the target tidal volume range and changes in physiological indicators.

2. The NPPV bidirectional switching method based on acute exacerbation intelligent early warning as described in claim 1, characterized in that, The patient's physiological indicators and ventilator operating parameters are continuously collected by multi-source sensors. The patient's physiological indicators include one or more of heart rate (HR), blood pressure (BP), blood oxygen saturation (SpO2), and end-tidal carbon dioxide concentration (PETCO2). The ventilator operating parameters include one or more of respiratory rate (RR), tidal volume (VT), minute ventilation (MV), and leakage volume (Vleak).

3. The NPPV bidirectional switching method based on acute exacerbation intelligent early warning as described in claim 1, characterized in that, The preset condition is that the warning information reaches the cancellation threshold and continues for more than 48 hours.

4. The NPPV bidirectional switching method based on acute exacerbation intelligent early warning according to claim 1, characterized in that, An acute exacerbation early warning model was constructed, and the model was used to issue early warnings based on the patient's physiological indicators and ventilator operating parameters. The acute aggravation early warning model adopts a deep learning architecture that integrates 1D-CNN, LSTM, attention mechanism and MoE weighting algorithm to analyze multi-source sensor data in real time to generate early warning signals.

5. The NPPV bidirectional switching method based on acute exacerbation intelligent early warning according to claim 1, characterized in that, Dynamic adjustment of positive airway pressure during inspiratory phase includes: The target tidal volume range for the low-intensity mode is 6–10 ml / kg, and the target tidal volume range for the high-intensity mode is 10–15 ml / kg. When the end-tidal carbon dioxide concentration (PETCO2) increases by more than 5% and the respiratory rate (RR) or heart rate (HR) increases by more than 5% for 24 hours, the inspiratory positive airway pressure (IPAP) is automatically increased until the end-tidal carbon dioxide concentration (PETCO2) returns to baseline. When the end-tidal carbon dioxide concentration (PETCO2) decreases by more than 5% and the respiratory rate (RR) and heart rate (HR) remain stable at baseline for 24 hours, the intra-inspiratory positive airway pressure (IPAP) is automatically reduced within the target VT range.

6. The NPPV bidirectional switching method based on acute exacerbation intelligent early warning according to claim 5, characterized in that, During the inspiratory phase, the intra-inspiratory positive pressure (IPAP) is adjusted at a rate of 1–2 cmH2O per adjustment, with an interval of at least 2 hours between two consecutive dynamic adjustments.

7. The NPPV bidirectional switching method based on acute exacerbation intelligent early warning according to claim 5, characterized in that, In the low-intensity mode, the upper limit of the positive airway pressure (IPAP) adjustment during the inspiratory phase is 20 cmH2O; in the high-intensity mode, the upper limit of the positive airway pressure (IPAP) adjustment during the inspiratory phase is 30 cmH2O.

8. The NPPV bidirectional switching method based on acute exacerbation intelligent early warning according to claim 7, characterized in that, When the positive airway pressure (IPAP) is automatically increased during the inspiratory phase, if the end-tidal carbon dioxide concentration (PETCO2) does not improve after increasing the preset amount within the upper limit of adjustment, the increase operation is terminated.

9. An NPPV bidirectional switching system based on intelligent early warning of acute exacerbations, characterized in that, The NPPV bidirectional switching method based on acute exacerbation intelligent early warning, as described in any one of claims 1-8, includes: Multi-sensor monitoring network: used to continuously collect patient physiological indicators and ventilator operating parameters through multiple source sensors; Intelligent early warning module: used to issue early warnings based on the patient's physiological indicators and ventilator operating parameters; The execution control module is used to respond to alarm signals and automatically switch the NPPV mode from low intensity to high intensity when an alarm signal is generated; to respond to cancellation signals and automatically switch the NPPV mode from high intensity to low intensity when the warning cancellation conditions meet the preset conditions; and to adjust IPAP parameters, including dynamically adjusting the positive pressure in the inspiratory airway according to the target tidal volume range and changes in physiological indicators in low intensity or high intensity modes.

Images