Method, system and equipment for multi-parameter centralized monitoring of noninvasive ventilator and medium

Abstract

The invention relates to a multi-parameter centralized monitoring method, system and equipment for a noninvasive ventilator and a medium. The method comprises the steps that firstly, original parameter data and coding rules of noninvasive respirators of all brands are obtained, format unification processing is conducted on the original parameter data by constructing a standardized data mapping table, and a parameter data set is obtained; transmitting the parameter data set to a local gateway by adopting a low-power-consumption Bluetooth protocol to generate an alarm data set; analyzing the dynamic change trend of the parameters by combining the parameter data set and the alarm data set, and generating a monitoring report data set; and finally, extracting a continuous abnormal trend based on the monitoring report data set, performing priority ranking according to the emergency degree to generate an intervention instruction, and tracking the processing state of the intervention instruction to obtain a final decision data set. According to the method, data intercommunication of noninvasive respirators of different brands is realized, the hysteresis of manual inspection is avoided, the timeliness and accuracy of abnormal response are improved, and the safety of patients is guaranteed.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of medical equipment monitoring, and particularly relates to a method, system, device and medium for multi-parameter centralized monitoring of non-invasive ventilators. BACKGROUND

[0002] As an important device for assisted breathing treatment, non-invasive ventilators are increasingly widely used in general wards of respiratory departments, and the timeliness of their running state and alarm processing is directly related to the treatment effect and safety of patients. However, different brands of non-invasive ventilators on the current market adopt independent interface protocols and data formats in design, which leads to difficulty in realizing data sharing between different brands of devices. At the same time, due to the limitation of existing conditions, general wards lack a centralized monitoring system specially for multiple non-invasive ventilators, and the monitoring of running parameters, alarm information and the like of the devices mainly depends on regular manual patrol by medical staff. This decentralized management mode makes it difficult to discover and handle the situation of device alarm, mask falling off, parameter abnormality and the like in time, and there is a security risk of delaying the treatment opportunity. In addition, medical staff need to go back and forth between different wards for patrol, which not only increases the work burden but also reduces the work efficiency, and it is difficult to realize real-time and efficient supervision of all devices and timely response to non-invasive ventilator alarms, which exists a medical safety risk. SUMMARY

[0003] Therefore, it is necessary to provide a method, system, device and medium for multi-parameter centralized monitoring of non-invasive ventilators, which can realize data sharing of different brands of non-invasive ventilators, solve the problem of lagging behind depending on manual patrol, and improve the timeliness of non-invasive ventilator alarm response and the monitoring of the working state of non-invasive ventilators.

[0004] In a first aspect, the application provides a method for multi-parameter centralized monitoring of non-invasive ventilators, comprising:

[0005] Obtaining original parameter data and coding rules of non-invasive ventilators of different brands, constructing a standardized data mapping table to perform format unified processing on the original parameter data, and obtaining a parameter data set; the original parameter data includes parameter output format, transmission protocol, pressure and tidal volume.

[0006] Transmitting the parameter data set to a local gateway by using a low-power Bluetooth protocol, judging the parameter data based on a preset threshold, and triggering an abnormal alarm signal if the parameter exceeds the threshold, and generating an alarm data set including device identification and abnormal parameters.

[0007] Analyzing the dynamic change trend of the parameters by combining the parameter data set and the alarm data set, and generating a monitoring report data set including aggregated parameters, alarm information and trend characteristics.

[0008] Based on the monitoring report data set, the continuous abnormal trend is extracted, the intervention instructions are generated after being prioritized according to the emergency degree, the processing state of the intervention instructions is tracked, and the final decision data set is obtained.

[0009] In one of the embodiments, the standardized data mapping table is constructed to uniformly process the original parameter data, and the parameter data set is obtained, including:

[0010] The original parameter data is obtained, and the business fields are extracted therefrom to obtain the field set.

[0011] The regular expression is used to parse the field set, and the standardized field template is generated.

[0012] The standardized field template is used to construct the data mapping table in combination with the coding rules to determine the field correspondence relationship; the data mapping table is used to establish the corresponding mapping between the parameters of different brands of non-invasive ventilators and the unified standard parameters.

[0013] Based on the field correspondence relationship, the original parameter data is converted to obtain the parameter data set in a unified format.

[0014] The data integrity is judged by using the data verification algorithm on the parameter data set, and if the verification is passed, the structured parameter data set is generated; the data verification algorithm is the checksum algorithm and the hash algorithm.

[0015] The data normalization processing is performed on the structured parameter data set to obtain the normalized parameter data set.

[0016] In one of the embodiments, the parameter data set is transmitted to the local gateway by using the Bluetooth low energy protocol, and the parameter data is judged based on the preset threshold value; if the parameter exceeds the threshold value, an abnormal alarm signal is triggered, and an alarm data set including the device identifier and the abnormal parameter is generated, including:

[0017] The parameter data set including the device identifier and the parameter value is transmitted to the local gateway by using the Bluetooth low energy protocol.

[0018] The parameter value in the parameter data set is calculated by using the local gateway to calculate the dynamic fluctuation coefficient, and the parameter is determined to be abnormal or not in combination with the preset threshold value interval to obtain the judgment result.

[0019] If the judgment result shows that the dynamic fluctuation coefficient of the parameter value exceeds the preset threshold value interval, an abnormal alarm signal is generated.

[0020] The abnormal alarm signal, the dynamic fluctuation coefficient, and the device identifier are integrated to generate an alarm data set including the abnormal parameter details.

[0021] In one of the embodiments, the dynamic fluctuation coefficient is calculated by the following formula:

[0022]

[0023] wherein DFC represents a dynamic fluctuation coefficient, x i represents a parameter value at the i-th sampling time, represents a weighted mean value in a preset time window, ω i represents a time decay weight, ω i = e -k(m-i) , k represents a decay coefficient, and n represents a sampling number, represents a reference standard deviation.

[0024] In one of the embodiments, the dynamic change trend of the parameter is analyzed in combination with the parameter data set and the alarm data set, and a monitoring report data set including aggregated parameters, alarm information and trend characteristics is generated, including:

[0025] The time sequence parameters of the parameter data set and the alarm triggering conditions of the alarm data set are obtained, and the change rate is calculated by association to generate a rate sequence data; the alarm triggering conditions include a parameter preset threshold interval, an abnormal duration threshold and a device running state benchmark value.

[0026] The rate sequence data is grouped according to the device identifier, and the rate mean value, peak value and fluctuation rate in each window are calculated by using a sliding window algorithm, and the aggregated parameter set is aggregated.

[0027] The corresponding alarm start and end timestamps are matched for the aggregated parameter set according to the device identifier, the alarm duration is extracted by calculating the timestamp difference, and the trend characteristic data is integrated and generated.

[0028] In combination with the trend characteristic data and the alarm triggering condition, a trend curve is drawn with a time axis as the horizontal axis and a parameter change rate as the vertical axis, the alarm triggering threshold line and the abnormal period marker are superimposed, and a data visualization graph is generated.

[0029] If there is an abnormal point in the data visualization graph, the device identifier, time node and parameter fluctuation characteristics corresponding to the abnormal point are extracted, and the monitoring report data set is generated according to the alarm triggering condition.

[0030] In one of the embodiments, based on the monitoring report data set, a continuous abnormal trend is extracted, an intervention instruction is generated after priority sorting according to the emergency degree, and the processing state of the intervention instruction is tracked to obtain a final decision data set, including:

[0031] The parameter records exceeding the preset threshold in the continuous sampling period are filtered from the monitoring report data set, the corresponding continuous abnormal trend characteristics are extracted, and the abnormal trend data sequence is constructed according to the device identifier and the time dimension.

[0032] Based on the parameter fluctuation amplitude and the duration in the abnormal trend data sequence, the emergency degree is evaluated by using a quantitative formula to obtain a priority sorting matrix.

[0033] The hierarchical intervention instruction is generated according to the hierarchical division in the priority ranking matrix, and a structured intervention instruction set containing an execution subject, an execution time limit and an instruction number is formed.

[0034] The whole-process processing state of the structured intervention instruction set is tracked according to the instruction number, and instruction execution trajectory data containing a response time and an execution progress are collected.

[0035] A multi-dimensional performance evaluation model is constructed in combination with the instruction execution trajectory data and preset performance indicators, and dynamic decision support information containing details of uncompleted instructions is output; the performance evaluation model performs regression analysis with the execution progress as the dependent variable and the response time as the independent variable.

[0036] Based on the resource load data in the dynamic decision support information, resource allocation of the uncompleted instructions is optimized, and a final decision data set is generated.

[0037] In one embodiment, the quantitative formula is represented by the following formula:

[0038]

[0039] Wherein, Δx represents the parameter fluctuation amplitude, Δx max represents the historical maximum fluctuation amplitude, t represents the duration, t max represents the historical maximum duration, represents the parameter change rate, represents the historical maximum change rate, and α, β and γ represent weight coefficients.

[0040] In a second aspect, the application also provides a system for multi-parameter centralized monitoring of a non-invasive ventilator, which comprises:

[0041] A parameter standardization module is configured to acquire original parameter data and coding rules of each brand of non-invasive ventilator, construct a standardization data mapping table, and perform format uniform processing on the original parameter data to obtain a parameter data set; the original parameter data includes parameter output format, transmission protocol, pressure and tidal volume.

[0042] An abnormal alarm generation module is configured to transmit the parameter data set to a local gateway using a Bluetooth Low Energy protocol, judge the parameter data based on a preset threshold, and trigger an abnormal alarm signal if the parameter exceeds the threshold, and generate an alarm data set including device identification and abnormal parameters.

[0043] A monitoring report generation module is configured to analyze the dynamic change trend of the parameters in combination with the parameter data set and the alarm data set, and generate a monitoring report data set including aggregated parameters, alarm information and trend characteristics.

[0044] An intervention decision tracking module is configured to extract a continuous abnormal trend based on the monitoring report dataset, generate an intervention instruction after prioritizing according to the emergency degree, track the processing state of the intervention instruction, and obtain a final decision dataset.

[0045] In a third aspect, the present application also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the foregoing method when executing the computer program.

[0046] In a fourth aspect, the present application also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the foregoing method.

[0047] The method, system, computer device and storage medium for multi-parameter centralized monitoring of non-invasive ventilators described above first acquire original parameter data and coding rules of non-invasive ventilators of various brands, wherein the original parameter data covers parameter output format, transmission protocol, pressure and tidal volume, and the original parameter data is uniformly processed in format through a standardized data mapping table to obtain a parameter dataset; then the parameter dataset is transmitted to a local gateway using a Bluetooth Low Energy protocol, the parameter data is judged based on a preset threshold, and if the parameter exceeds the threshold, an abnormal alarm signal is triggered to generate an alarm dataset containing device identification and abnormal parameters; subsequently, the dynamic change trend of the parameter is analyzed by combining the parameter dataset and the alarm dataset to generate a monitoring report dataset including aggregated parameters, alarm information and trend characteristics; finally, a continuous abnormal trend is extracted based on the monitoring report dataset, an intervention instruction is generated after prioritizing according to the emergency degree, and the processing state of the intervention instruction is tracked to obtain a final decision dataset. This method realizes the intercommunication of data of non-invasive ventilators of different brands through standardized processing, solves the problem of non-uniform data format; realizes real-time monitoring and abnormal alarm of parameters through wireless transmission and centralized analysis, avoiding the hysteresis of manual patrol; generates an intervention instruction and tracks the processing state to improve the timeliness and accuracy of abnormal response. This process reduces the work burden of medical staff, reduces the safety risks caused by not timely processing of abnormal parameters, and improves the efficiency and standardization of non-invasive ventilator management. BRIEF DESCRIPTION OF DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed to be used in the embodiment or related art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0049] Figure 1A flowchart of a method for multi-parameter centralized monitoring of a noninvasive ventilator according to an embodiment of the present application is provided.

[0050] Figure 2 A structural block diagram of a multi-parameter centralized monitoring system for a noninvasive ventilator according to an embodiment of the present application is provided. DETAILED DESCRIPTION

[0051] For the purpose, technical solutions and advantages of the present application to be clearer, the present application is further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0052] In one embodiment, as shown in Figure 1 The method for multi-parameter centralized monitoring of a noninvasive ventilator according to the present application can include the following steps:

[0053] In step S101, the original parameter data and coding rules of noninvasive ventilators of various brands are acquired, a standardized data mapping table is constructed, the original parameter data is uniformly processed in format, and a parameter data set is obtained. The original parameter data includes parameter output format, transmission protocol, pressure and tidal volume.

[0054] Specifically, the original parameter data is collected from noninvasive ventilators of different brands through a device interface, and the coding rules of devices of various brands are collected to clearly define the definition, unit and format specification of parameters. Subsequently, according to these coding rules, key business fields in the original parameter data are extracted, such as parameter attribute information of the numerical range of pressure and the recording frequency of tidal volume. The specific monitoring values of core respiratory parameters such as pressure / peak pressure, exhaled tidal volume, air leakage, minute ventilation, respiratory rate and inspiration-expiration ratio are also covered, and regular expression is used to analyze these fields to generate a standardized field template. The template is used in combination with the coding rules to construct a data mapping table to establish the correspondence between parameters of devices of different brands and unified standard parameters, for example, the fields of "pressure" of different brand tables are uniformly mapped to standard fields. Based on this mapping relationship, the original parameter data is converted, and parameter values of different formats are adjusted to a unified format to form a preliminary parameter data set. Then, the parameter data set is checked for integrity by a checksum algorithm or a hash algorithm, and missing or incorrect data is removed to generate a structured parameter data set, and finally, normalization processing is performed on the parameter data set to make the parameter values in a unified data interval.

[0055] In step S102, the parameter data set is transmitted to a local gateway using a Bluetooth Low Energy protocol, and based on a preset threshold, the parameter data is judged. If the parameter exceeds the threshold, an abnormal alarm signal is triggered, and an alarm data set including the device identifier and the abnormal parameter is generated.

[0056] Specifically, the parameter data set is encapsulated as a data packet containing the unique identification of the device and the values of various parameters, and is transmitted wirelessly to the local gateway in the ward through the Bluetooth protocol. This transmission method can reduce the energy consumption of the device while ensuring the real-time nature of the data. After receiving the data packet, the local gateway calculates the dynamic fluctuation coefficient of the parameter values in the data packet. The coefficient quantifies the fluctuation of the parameter within a preset time window through a formula, and determines whether the parameter is abnormal in combination with a preset threshold interval. When the dynamic fluctuation coefficient exceeds the threshold interval, the system immediately triggers an abnormal alarm signal, which contains information such as the time of the abnormality and the type of the parameter. Finally, the abnormal alarm signal, the calculated dynamic fluctuation coefficient, and the corresponding device identification are integrated to form an alarm data set containing details such as the specific value of the abnormal parameter and the fluctuation amplitude.

[0057] Step S103: Analyze the dynamic change trend of the parameter by combining the parameter data set and the alarm data set, and generate a monitoring report data set including aggregated parameters, alarm information, and trend characteristics.

[0058] Specifically, first, the time series parameters are extracted from the parameter data set, i.e., the parameter values arranged in chronological order, and the alarm triggering conditions are obtained from the alarm data set, including the preset threshold interval of the parameter, the abnormal duration threshold, and the baseline value of the device in normal operation. The time series parameters and the alarm triggering conditions are associated and calculated to obtain the rate of change of the parameter over time, generating rate sequence data. Next, the rate sequence data is grouped by device identification, and a sliding window algorithm is used to divide each group of data into multiple consecutive time windows. The mean, peak, and fluctuation rate of the rate in each window are calculated, and these statistical values are aggregated to form an aggregated parameter set. Then, the alarm start and end timestamps are matched to the aggregated parameter set according to the device identification, and the alarm duration is extracted by calculating the difference between the timestamps. The alarm duration is integrated with the statistical values in the aggregated parameter set to generate trend characteristic data, such as the fluctuation trend of the parameter during abnormality, the correlation between the duration and the rate of change, etc. Trend curves are drawn based on the trend characteristic data and the alarm triggering conditions to visually display the parameter changes, and data visualization graphics are generated by superimposing the alarm triggering threshold line and the abnormal period markers. If there are abnormal points outside the normal range in the graphics, the device identification, time node, and parameter fluctuation characteristics corresponding to the abnormal points are extracted, and the monitoring report data set is generated based on the alarm information, including aggregated parameters (i.e., the summary statistics of multiple device parameters), alarm details, and trend characteristics.

[0059] Step S104: Extract the continuous abnormal trend based on the monitoring report data set, prioritize the intervention instructions according to the urgency, and generate the final decision data set while tracking the processing status of the intervention instructions.

[0060] First, the parameter records exceeding the preset threshold in continuous multiple sampling periods are screened out from the monitoring report data set, which reflect the continuous abnormal situation of the parameters. The corresponding continuous abnormal trend characteristics such as the type, fluctuation amplitude and duration of the abnormal parameters are extracted, and then these characteristics are sorted according to the equipment identifier and time dimension to construct the abnormal trend data sequence. Based on the data sequence, the emergency degree of each abnormal situation is calculated using a quantitative formula. The formula considers the ratio of parameter fluctuation amplitude to historical maximum fluctuation amplitude, the ratio of duration to historical maximum duration, and the ratio of parameter change rate to historical maximum change rate. The influence of each factor is adjusted by a weight coefficient, and a priority ranking matrix is generated accordingly. According to the hierarchical division in the matrix, a hierarchical intervention instruction is generated, which clearly defines the execution subject, completion time limit and unique instruction number of each instruction, forming a structured intervention instruction set. The processing state of the structured intervention instruction set is tracked throughout the process according to the instruction number, and information such as response time and current execution progress of the instruction is recorded to collect and form instruction execution trajectory data. A multi-dimensional performance evaluation model is constructed by combining the instruction execution trajectory data and the preset performance indicators. Through regression analysis, the relationship between execution progress and response time is studied, and dynamic decision support information containing details such as the specific content of uncompleted instructions and the reasons for non-execution is output. Based on the resource load data in the dynamic decision support information, such as the current task quantity and processing capacity of each execution subject, the uncompleted instructions are reconfigured for resources to generate optimized intervention instructions. At the same time, the processing state of these instructions is tracked to obtain updated execution data, and finally the final decision data set is integrated.

[0061] The method for multi-parameter centralized monitoring of the non-invasive ventilator, first acquires original parameter data and coding rules of each brand of non-invasive ventilator, wherein the original parameter data covers parameter output format, transmission protocol, pressure and tidal volume, and the original parameter data is uniformly processed in format through construction of a standardized data mapping table to obtain a parameter data set; then the parameter data set is transmitted to a local gateway by using a low-power Bluetooth protocol, the parameter data is judged based on a preset threshold, and if the parameter exceeds the threshold, an abnormal alarm signal is triggered to generate an alarm data set containing device identification and abnormal parameters; subsequently, the dynamic change trend of the parameter is analyzed in combination with the parameter data set and the alarm data set to generate a monitoring report data set including aggregated parameters, alarm information and trend characteristics; finally, a continuous abnormal trend is extracted based on the monitoring report data set, an intervention instruction is generated after priority sorting according to the emergency degree, and the processing state of the intervention instruction is tracked to obtain a final decision data set. The method realizes the intercommunication of data of different brands of non-invasive ventilators through standardized processing, solves the problem of non-uniform data format, realizes real-time monitoring and abnormal alarm of the parameter by means of wireless transmission and centralized analysis, avoids the hysteresis of manual patrol, and improves the timeliness and accuracy of abnormal response by generating an intervention instruction and tracking the processing state. This process reduces the work burden of medical staff, reduces the safety risk caused by unprocessed alarms, and improves the efficiency and standardization of non-invasive ventilator alarm management.

[0062] In one of the embodiments, the standardized data mapping table is constructed to uniformly process the original parameter data in format to obtain the parameter data set, which can include the following steps:

[0063] In step S201, the original parameter data is acquired, and business fields are extracted therefrom to obtain a field set.

[0064] Preferably, the original parameter data is untreated original information generated in real time by each brand of non-invasive ventilator during operation, and specifically includes parameter formats (such as numerical type and character type) output by the device, transmission protocols (such as private protocols of specific brands) for communication between the device and the outside, tidal volume (unit: mL) directly reflecting the patient's condition, pressure (cmH2O), minute ventilation (unit: L), respiratory rate (times / min), air leakage (unit: mL), and other key parameters. The business field is an independent information unit with actual business significance stripped from the original data, such as "specific numerical value of pressure", "measurement timestamp of tidal volume", "version number of device transmission protocol", and the like. During the extraction process, redundant invalid information (such as meaningless identification characters) is removed, and only the key fields related to device operation and patient monitoring are retained to finally form a set containing all valid business fields, i.e., the field set.

[0065] In step S202, the field set is parsed by using a regular expression to generate a standardized field template.

[0066] Further, the regular expression is a tool for matching text content by specific character patterns, which is used to accurately identify the format characteristics of each business field in the field set in this step, such as the specific numerical value of the pressure and respiratory rate (times / min), the representation form of the tidal volume unit (such as different writings of "mL", "ml", "milliliter"), the format of the time stamp (such as "YYYY-MM-DD HH:MM:SS"), etc. When parsing, the fields with different names or formats but the same meaning in different brands are unified into a standard form through the preset regular expression rules, for example, the "peak pressure" of brand A and the "pressure" of brand B are unified and parsed into "pressure", and the data type (such as numerical value type), unit (such as %), format requirement (such as retaining one decimal place) are clearly defined, and finally a standardized field template is formed, which provides a unified naming, format and rule standard for all business fields.

[0067] In step S203, the standardized field template is used in combination with the coding rule to construct a data mapping table to determine the field correspondence relationship; the data mapping table is used to establish the corresponding mapping between the parameters of non-invasive ventilators of different brands and the unified standard parameters.

[0068] The coding rule is the definition and description of the parameters of each brand equipment manufacturer, including the naming rule (such as abbreviation specification) of the parameter, the value range (such as the normal range of respiratory rate 16-20 times / min), the data coding format (such as binary, ASCII code) and other detailed information. When constructing the data mapping table, the standardized field template is used as a benchmark, and the coding rules of each brand are compared one by one to clearly define the mapping relationship between the original parameters of each brand and the corresponding fields in the standardized template, for example, the "TidalVol. (L)" (tidal volume, unit liter) of brand C corresponds to "tidal volume (mL)" in the standardized template, and the unit conversion relationship (1L = 1000mL) is marked; the "Respiratory Rate (abbreviated as RR)" of brand D corresponds to the "respiratory rate" in the standardized template. The mapping table is presented in the form of a table, with the original parameters of each brand on the left and the corresponding standardized parameters on the right, clearly recording the corresponding relationship of field name, unit and data type, and realizing the accurate mapping of parameters of different brands to the unified standard.

[0069] In step S204, the original parameter data is converted based on the field correspondence relationship to obtain a parameter data set in a unified format.

[0070] The conversion process strictly follows the field correspondence in the data mapping table to perform targeted processing on the original parameter data of each brand: for field names, replace them with standardized names according to the mapping table (e.g., change "RR" to "respiratory rate"); for units, unify them according to the conversion relationship in the mapping table (e.g., convert liters to milliliters); for data types, adjust them to the types specified in the standardized template (e.g., convert the string type "18" to the numeric type 18); and for formats, unify them to the preset form (e.g., unify the timestamp to "YYYY-MM-DD HH:MM:SS"). After the above conversion, the original parameter data, which was originally in a disordered format due to brand differences, is integrated into a unified format parameter data set that meets the requirements of the standardized template, completely eliminating the data format barriers between different brands.

[0071] In step S205, a data verification algorithm is used to judge the data integrity of the parameter data set, and if the verification is passed, a structured parameter data set is generated; the data verification algorithm is a checksum algorithm and a hash algorithm.

[0072] The structured parameter data set has a fixed table structure, with rows representing different sampling records and columns corresponding to standardized fields (such as "device identification", "pressure", "timestamp", etc.). Each row and column has a clear data correspondence, facilitating subsequent storage, query, and analysis.

[0073] In step S206, data normalization processing is performed on the structured parameter data set to obtain a normalized parameter data set.

[0074] First, the original parameter data is obtained, and business fields are extracted from it to form a field set. Regular expressions are used to parse the field set to generate a standardized field template. The standardized field template is used in conjunction with coding rules to build a data mapping table to determine the field correspondence. This data mapping table is used to establish a correspondence between the parameters of different brands of non-invasive ventilators and the unified standard parameters. Based on the above field correspondence, the original parameter data is converted to obtain a parameter data set in a unified format. Checksum algorithms and hash algorithms are used to judge the data integrity of the parameter data set, and if the verification is passed, a structured parameter data set is generated. Finally, data normalization processing is performed on the structured parameter data set to obtain a normalized parameter data set.

[0075] This embodiment achieves the format unification of original parameter data of different brands of non-invasive ventilators through steps such as extracting business fields, generating a standardized template, and building a mapping table, solving the data intercommunication problem. The use of verification algorithms ensures the integrity and accuracy of the data, providing a reliable foundation for subsequent analysis. Normalization processing puts the parameters in a unified data interval, facilitating cross-device and cross-time parameter comparison and analysis. This series of operations improves the usability and consistency of the data, providing high-quality data support for centralized monitoring and subsequent decision-making.

[0076] In one embodiment, the parameter data set is transmitted to the local gateway based on a preset threshold value using a low-power Bluetooth protocol, and if the parameter exceeds the threshold value, an abnormal alarm signal is triggered, and an alarm data set including the device identification and abnormal parameter is generated, which can include the following steps:

[0077] Step S301, a parameter data set containing device identification and parameter value is transmitted to the local gateway using a low-power Bluetooth protocol.

[0078] Preferably, the low-power Bluetooth protocol is a low-power, short-range wireless communication technology suitable for power-sensitive scenarios such as medical devices, which can reduce device power consumption while ensuring stable data transmission. The device identification is a unique identity code (such as a serial number or MAC address) for each non-invasive ventilator, used to distinguish different devices; the parameter value is the specific monitoring data after standardization, such as airway pressure 15 cmH2O, tidal volume 800 mL, etc. The parameter data set is encapsulated in the form of a data packet, containing device identification, parameter values, and time stamps, etc. information, and is sent wirelessly to the local gateway in the ward through the low-power Bluetooth protocol. The local gateway acts as a data aggregation node, responsible for receiving and temporarily storing the data sets transmitted by each device.

[0079] Step S302, the dynamic fluctuation coefficient of the parameter value in the parameter data set is calculated using the local gateway, and the parameter is determined to be abnormal in combination with the preset threshold interval to obtain a judgment result.

[0080] The preset threshold interval is the normal range of the dynamic fluctuation coefficient set according to the clinical standard and the device characteristics (such as the FC threshold interval of blood oxygen saturation is [0.2, 1.5]). After the local gateway calculates the dynamic fluctuation coefficient of each parameter according to the formula, it compares it with the corresponding threshold interval. If the coefficient is within the interval, it is determined to be normal, otherwise it is determined to be abnormal, forming a judgment result.

[0081] Step S303, if the judgment result shows that the dynamic fluctuation coefficient of the parameter value exceeds the preset threshold interval, an abnormal alarm signal is generated.

[0082] When the judgment result is abnormal, the system immediately triggers an abnormal alarm signal, which contains multi-dimensional information: device identification (clearly abnormal device), parameter type (such as "blood oxygen saturation"), abnormal timestamp, dynamic fluctuation coefficient specific value, and amplitude of exceeding threshold value (such as "FC = 2.3, exceeding the upper limit 1.5 by 0.8"). The signal is generated in a structured data form, which contains both machine-readable format (such as JSON) for system processing and concise text description (such as "Device ID: V123, blood oxygen saturation fluctuation abnormal, FC = 2.3") for medical staff to view, ensuring that abnormal information can be quickly identified and processed.

[0083] Step S304, integrate the abnormal alarm signal, dynamic fluctuation coefficient and device identifier to generate an alarm data set containing abnormal parameter details.

[0084] The embodiment realizes efficient transmission of parameter data through the low-power Bluetooth protocol, taking into account real-time performance and energy consumption control; the calculation of the dynamic fluctuation coefficient and the threshold determination ensure the accuracy of abnormal identification, avoiding false positives or false negatives; the generation of the alarm data set integrates key information, providing clear abnormal clues for subsequent monitoring and analysis. This series of operations realizes the timely discovery and recording of the parameter abnormalities of the non-invasive ventilator, reduces the lag of manual monitoring, lays the foundation for rapid response to abnormal situations, and helps to reduce safety risks.

[0085] In one of the embodiments, the dynamic fluctuation coefficient can be calculated by the following formula:

[0086]

[0087] wherein DFC represents the dynamic fluctuation coefficient, x i represents the parameter value at the i-th sampling time, represents the weighted mean value in the preset time window, ω i represents the time decay weight, ω i = e -k(n-i) , k represents the decay coefficient, and n represents the number of samples, represents the reference standard deviation, which is calculated based on the normal operation history data of each brand device for the corresponding parameter type in the past 30 days. The same brand and the same parameter type share a set of reference standard deviations.

[0088] Preferably, the length of the preset time window is set according to the parameter type and the clinical monitoring requirements, wherein the time window for pressure is 5 minutes, and the time window for tidal volume is 2 minutes; for the same type of parameters of different brands, the same time window length is adopted (such as all brand devices use a 5-minute time window for pressure).

[0089] The embodiment can accurately depict the fluctuation characteristics of the parameter in the time dimension, reflecting the influence weight of the recent data and combining the reference standard deviation of the historical normal data to calibrate the fluctuation degree, making the abnormal determination more consistent with the clinical reality; based on the comparison of the coefficient and the preset threshold interval, the normal fluctuation and abnormal change of the parameter can be effectively distinguished, reducing the false positives caused by instantaneous fluctuations, while ensuring that significant abnormalities are not missed, providing a reliable basis for the generation of subsequent abnormal alarm signals, improving the accuracy and reliability of the non-invasive ventilator parameter monitoring, and helping medical staff to timely grasp the device and patient status.

[0090] In one of the embodiments, the dynamic change trend of the parameter data set and the alarm data set is analyzed in combination to generate a monitoring report data set including aggregated parameters, alarm information and trend characteristics, which can include the following steps:

[0091] In step S401, the time series parameters of the parameter data set and the alarm trigger conditions of the alarm data set are obtained, and the change rate is calculated by correlation calculation to generate rate sequence data; the alarm trigger conditions include parameter preset threshold interval, abnormal duration threshold and device running state reference value.

[0092] The time series parameters are a set of parameter values arranged in time sequence in the parameter data set, for example, the tidal volume and respiratory rate of a device recorded every 5 seconds, which are arranged in time sequence to form a sequence. The alarm trigger condition is the basis for the system to determine parameter abnormalities and trigger alarms, wherein the parameter preset threshold interval is the numerical range of the normal operation of the parameter (20% above and below the parameter), the abnormal duration threshold is the minimum time that the parameter needs to be continuously above the threshold to trigger an alarm (for example, 30 seconds above the threshold), and the device running state reference value is the basic parameter of the device in normal operation (for example, the standard initial value of the tidal volume). In the correlation calculation, the time series parameters at the same time are compared with the reference value in the alarm trigger condition based on time, and the change rate of the parameter is calculated by the formula (current parameter value - previous parameter value) / time interval. The rate values are arranged in time sequence to generate rate sequence data, reflecting the speed of parameter change over time.

[0093] In step S402, the rate sequence data is grouped by device identifier, and the sliding window algorithm is used to calculate the rate mean, peak value and volatility in each window to aggregate the aggregated parameter set.

[0094] Preferably, the device identifier is a unique code to distinguish different non-invasive ventilators, and grouping by the device identifier ensures that the rate sequence data of the same device is processed. The sliding window algorithm divides the continuous rate sequence data into multiple fixed-length overlapping or non-overlapping windows (for example, the window length is set to 1 minute, and the window slides every 30 seconds), and calculates the data in each window: the rate mean is the average of all change rates in the window, reflecting the average speed of parameter change in the period; the peak value is the maximum change rate in the window, representing the most severe degree of parameter change; and the volatility is the dispersion degree (such as standard deviation) of the rate values in the window, reflecting the stability of parameter change. The calculation results are arranged by device identifier and window time to form the aggregated parameter set containing the rate characteristics of each device at different time periods.

[0095] In step S403, the aggregated parameter set is matched with the corresponding alarm start and end time stamps by device identifier, the alarm duration is extracted by calculating the time stamp difference, and the trend characteristic data is integrated and generated.

[0096] Further, the alarm start and end time stamp is the specific time when each alarm event starts and ends recorded in the alarm data set (such as 2024-05-2008:30:00 to 2024-05-2008:35:00). After matching the aggregated parameter set with the alarm start and end time stamp by device identification, the alarm duration is calculated by subtracting the start time stamp from the end time stamp (such as 5 minutes in the above example). Then, the average rate, peak value, and fluctuation rate in the aggregated parameter set are integrated with the corresponding alarm duration to form trend feature data, which embodies the parameter change trend features during the alarm period, such as the average rate of 2% per second and the peak value of 5% per second of a certain device within the 5-minute alarm duration.

[0097] In step S404, the trend curve is drawn with the time axis as the horizontal axis and the parameter change rate as the vertical axis, the alarm trigger threshold line and the abnormal period marker are superimposed, and the data visualization graph is generated by combining the trend feature data and the alarm trigger condition.

[0098] The alarm trigger threshold line is a change rate critical line converted from the parameter preset threshold interval in the alarm trigger condition (such as ±1% per second when the parameter preset threshold interval is 95%-100%), which is used to distinguish the normal and abnormal change rate ranges. The abnormal period marker is an interval annotation (such as marked with a shadow or a different color background) corresponding to the alarm start and end time on the trend curve, which clearly indicates the time period when the alarm occurs. Through the combination of these elements, the generated data visualization graph can enable medical staff to quickly and intuitively grasp the parameter change trend and abnormal period.

[0099] In step S405, if there is an abnormal point in the data visualization graph, the device identification, time node, and parameter fluctuation feature corresponding to the abnormal point are extracted, and the monitoring report data set is generated according to the alarm trigger condition.

[0100] Further, the abnormal point refers to a change rate point on the trend curve that exceeds the alarm trigger threshold line, or a rate value point that significantly deviates from the normal in the normal period. When extracting the abnormal point information, the corresponding device identification (which device), time node (specific occurrence time), and parameter fluctuation feature (such as the change rate value of the point, the deviation degree from the threshold value, the change of the rate before and after the point, etc.) need to be determined. These information are screened and verified in combination with the alarm trigger condition, such as judging whether the abnormal point meets the abnormal duration threshold, etc. Finally, the abnormal information that meets the conditions is integrated with the aggregated parameters (the summary statistics of multiple device parameters), the alarm information (alarm type, duration, etc.), and the trend feature (the overall trend of parameter change) to generate the monitoring report data set.

[0101] Specifically, the time series parameters (i.e. parameter values arranged in chronological order) of the parameter dataset and the alarm triggering conditions (including parameter preset threshold interval, abnormal duration threshold and device operating state reference value) of the alarm dataset are obtained, and the parameter change rate over time is calculated by associating the two to generate rate sequence data; after grouping the rate sequence data by device identifier, the sliding window algorithm (dividing continuous data into fixed length windows) is used to calculate the rate mean, peak and volatility in each window, and aggregate parameter sets are formed; the alarm start and end timestamps are matched to the aggregated parameter sets by device identifier, and the alarm duration is extracted by calculating the timestamp difference, and the trend feature data containing parameter change trend and alarm duration characteristics are generated by integrating the aggregated parameter sets; combined with the trend feature data and the alarm triggering condition, the trend curve is drawn with time axis as horizontal axis and parameter change rate as vertical axis, and the alarm triggering threshold line (used to visually display the normal and abnormal boundaries) and the abnormal period marker are superimposed to generate data visualization graphics; if there are abnormal points in the graphics, the corresponding device identifier, time node and parameter fluctuation characteristics are extracted, and the monitoring report dataset containing aggregated parameters, alarm information and trend characteristics is generated according to the alarm triggering condition.

[0102] The present embodiment accurately captures parameter change rate and trend by associating time series parameters with alarm triggering conditions, and the sliding window algorithm is used to realize fine extraction of parameter fluctuation characteristics, and the timestamp matching ensures accurate acquisition of alarm duration, and data visualization visually presents parameter abnormal state, and the finally generated monitoring report dataset integrates key information, providing comprehensive and clear device and parameter state reference for medical staff, which helps to quickly locate the abnormal root cause and improve the accuracy and efficiency of monitoring.

[0103] In one of the embodiments, based on the monitoring report dataset, the continuous abnormal trend is extracted, the intervention instructions are generated after priority sorting according to the emergency degree, and the processing state of the intervention instructions is tracked to obtain the final decision dataset, which can include the following steps:

[0104] Step S501, filtering the parameter records of the monitoring report dataset that exceed the preset threshold in consecutive sampling periods, extracting the corresponding continuous abnormal trend characteristics, and constructing an abnormal trend data sequence according to the device identifier and time dimension.

[0105] The monitoring report dataset contains parameter changes, alarm information, and trend characteristics of each device. The continuous sampling period refers to the number of consecutive times of data collection at fixed time intervals (e.g., every 10 seconds). The preset threshold is a critical value of the number of consecutive abnormal times set according to clinical standards (e.g., 5 consecutive periods). During screening, only parameter records with more than the threshold number of continuous sampling periods are retained, such as records of airway pressure of a certain device being lower than 5 cmH2O for 6 consecutive periods. The continuous abnormal trend characteristics are abnormal patterns of parameters extracted from these records, including fluctuation amplitude (e.g., the difference between airway pressure from 12 cmH2O to 6 cmH2O), fluctuation frequency (e.g., the frequency of a 1% decrease every 2 minutes), change slope (e.g., the rate of decrease per unit time), etc. According to the device identifier (to distinguish different ventilators) and the time dimension (sorted by sampling time), the continuous abnormal trend characteristics of the same device are arranged in chronological order to form an abnormal trend data sequence, which intuitively presents the development process of abnormal parameters of a single device.

[0106] In step S502, based on the parameter fluctuation amplitude and the duration of the abnormal trend data sequence, an emergency level evaluation is performed using a quantification formula to obtain a priority ranking matrix.

[0107] The parameter fluctuation amplitude is the maximum difference of the abnormal parameter from the normal range (e.g., when the tidal volume is 300 mL, the fluctuation amplitude is 200 mL, with a normal range of 500-1000 mL); the duration is the cumulative time of the parameter in an abnormal state (e.g., 10 minutes of continuous abnormality from 8:00 to 8:10). The quantification formula is: Emergency level = (parameter fluctuation amplitude / historical maximum fluctuation amplitude) x a + (duration / historical maximum duration) x b + (parameter change rate / historical maximum change rate) x c, where a, b, and c are weight coefficients (sum = 1, set according to parameter importance, e.g., a value of pressure is higher), and the historical maximum fluctuation amplitude, maximum duration, and maximum change rate are all based on historical abnormal data statistics. The emergency level score of each abnormal trend is calculated by the formula, arranged by device identifier and score, to form a priority ranking matrix, with rows representing devices and columns representing emergency level levels (e.g., level 1 is the highest and level 3 is the lowest), to clearly determine the order of intervention.

[0108] In step S503, hierarchical intervention instructions are generated according to the levels in the priority ranking matrix to form a structured intervention instruction set containing the execution subject, execution time limit, and instruction number.

[0109] Preferably, the priority ranking matrix divides abnormal trends into different levels (such as level 1, level 2, and level 3), with level 1 corresponding to emergency situations that need to be handled immediately (such as sudden pressure drop), level 2 corresponding to situations that need to be handled within a short period of time (such as persistent low tidal volume), and level 3 corresponding to situations that can be handled routinely. The hierarchical intervention instructions specify specific operation requirements according to the level, such as the instruction for level 1 requiring "immediate check of equipment connection and adjustment of parameters", and the instruction for level 2 requiring "verification in the patient room within 30 minutes". The execution subject is the medical staff responsible for handling (such as the responsible nurse or the attending physician), the execution time limit is the latest time for completing the instruction (such as the instruction for level 1 "response within 5 minutes"), and the instruction number is a unique identifier (such as "ZY-20240520-001") for subsequent tracking. Integrating these information into structured data (such as in table form) forms a set of structured intervention instructions, ensuring that the instructions are clear and executable.

[0110] Step S504, according to the instruction number, track the whole-process processing status of the set of structured intervention instructions, and collect instruction execution trajectory data including response time and execution progress.

[0111] The instruction number is a unique identifier for tracking, and the whole-process processing status includes "not received", "received", "processing", "completed", and "expired". The response time refers to the time interval from the generation of the instruction to the confirmation of receipt and the start of processing by the execution subject (such as instruction generated at 8:00, confirmed for processing at 8:03, response time is 3 minutes); the execution progress refers to the proportion of the completion of the instruction (such as "device check completed, parameter adjustment in progress" corresponds to 50% progress). By recording the time nodes and progress descriptions of each state in real time, these information is summarized according to the instruction number to form the instruction execution trajectory data.

[0112] Step S505, combine the instruction execution trajectory data with the preset performance indicators to construct a multi-dimensional performance evaluation model, and output dynamic decision support information including details of uncompleted instructions; the performance evaluation model performs regression analysis with execution progress as the dependent variable and response time as the independent variable.

[0113] The preset performance indicators are the standards for evaluating the efficiency of instruction execution, including standard response time (such as level 1 instruction ≤5 minutes), expected completion rate (such as ≥95% within 8 hours), average processing time (such as ≤30 minutes for routine instructions), etc. The multi-dimensional performance evaluation model studies the relationship between execution progress and response time through regression analysis, for example, analyzing "how many percentage points does the execution progress decrease when the response time increases by 1 minute", quantifying the correlation between the two. The dynamic decision support information output by the model includes details of uncompleted instructions: such as instruction number, corresponding equipment, current state (such as "processing stagnation"), uncompleted reason (such as "medical staff is busy"), estimated completion time, etc., providing real-time reference for resource allocation.

[0114] Step S506, based on the resource load data in the dynamic decision support information, optimizing the resource allocation of the unfinished instructions, generating the final decision data set.

[0115] The resource load data reflects the current working state of the execution subject, including the number of instructions being processed by each medical staff (such as nurse A currently processing 3 instructions), the average processing time (such as nurse B processing 1 instruction for an average of 15 minutes), the current idle state (such as doctor C having no task for the time being), etc. When optimizing resource allocation, according to the priority of the unfinished instructions and the resource load data, the execution subject is adjusted (such as assigning a level 1 unfinished instruction from the busy nurse A to the idle nurse D), or the reasonable time limit is extended (such as adjusting the time limit of a level 2 instruction from 30 minutes to 45 minutes due to the busy staff). The optimized instruction allocation scheme, the adjusted time limit, the expected effect, etc. are integrated to generate the final decision data set.

[0116] Specifically, the parameter records of the monitoring report data set that exceed the preset threshold in consecutive sampling periods are screened, the corresponding continuous abnormal trend characteristics (such as the amplitude, frequency and change pattern of parameter fluctuation, etc.) are extracted therefrom, and an abnormal trend data sequence is constructed according to the device identifier and the time dimension; based on the amplitude and duration of parameter fluctuation in the sequence, the emergency degree is evaluated using a quantitative formula (comprehensive factors such as the ratio of parameter fluctuation amplitude to historical maximum fluctuation amplitude, the ratio of duration to historical longest duration, etc.), and a priority ranking matrix is obtained; according to the hierarchical division in the matrix, a hierarchical intervention instruction is generated, forming a structured intervention instruction set containing the execution subject (such as specific medical staff), the execution time limit (such as responding within 30 minutes) and the instruction number (unique identifier); the whole-process processing state of the instruction set is tracked according to the instruction number, and instruction execution trajectory data containing response time (the time length from instruction generation to start processing) and execution progress (such as completion rate) are collected; a multi-dimensional performance evaluation model is constructed by combining the trajectory data and the preset performance indicators (such as standard response time, expected completion rate), and a regression analysis is performed with execution progress as the dependent variable and response time as the independent variable, outputting dynamic decision support information containing the details of unfinished instructions (such as the reason for not executing, the current blocking factors); based on the resource load data (such as the current task amount and processing capacity of each execution subject) in the information, the resource allocation of the unfinished instructions is optimized, and the final decision data set is generated.

[0117] This embodiment accurately locates the key problems that need to be intervened by screening continuous abnormal data and constructing a sequence; the quantitative evaluation and priority ranking ensure the pertinence and timeliness of the intervention instruction; the whole-process tracking and performance modeling realize the dynamic monitoring and optimization of instruction execution; the application of resource load data improves the rationality of resource allocation. Not only does it speed up the response to abnormalities and improve the efficiency of intervention execution, but it also provides scientific decision-making basis for medical staff, effectively reducing management costs and safety risks.

[0118] In one embodiment, the quantification formula is represented by the following formula:

[0119]

[0120] wherein Δx represents the parameter fluctuation amplitude, Δx max represents the historical maximum fluctuation amplitude, t represents the duration, t max represents the historical maximum duration, represents the parameter change rate, represents the historical maximum change rate, α, β, γ represent the weight coefficients, the value range is [0, 1] and α+β+γ=1, and the values are set by clinical experts combined with experimental data according to the clinical importance of the parameters (such as pressure α taking 0.4, tidal volume β taking 0.3, and change rate γ taking 0.3).

[0121] In this embodiment, the emergency degree of the abnormal trend is quantified from three dimensions of fluctuation intensity, abnormal duration, and change speed by the ratio of the parameter fluctuation amplitude to the historical maximum fluctuation amplitude, the ratio of the duration to the historical maximum duration, and the ratio of the parameter change rate to the historical maximum change rate, and the importance difference of different dimensions is reflected by the weight coefficients (such as a parameter with greater impact on patient vital signs can be given a higher weight). This multi-dimensional weighted calculation method not only avoids the one-sidedness of single index evaluation, but also ensures the objectivity and consistency of the evaluation standard through historical data calibration, so that the emergency degree evaluation is more in line with the clinical actual needs. The priority ranking matrix obtained based thereon can accurately distinguish the emergency level of different abnormal conditions, provide a scientific basis for the generation of graded intervention instructions, help medical staff to prioritize the processing of high emergency degree abnormalities, improve the pertinence and efficiency of intervention response, and reduce the safety risks caused by decision delay.

[0122] In one embodiment, as shown in Figure 2 The application also provides a non-invasive ventilator multi-parameter centralized monitoring system, which can include:

[0123] A parameter standardization module 601 is configured to acquire original parameter data and coding rules of each brand of non-invasive ventilator, construct a standardized data mapping table, perform format uniform processing on the original parameter data, and obtain a parameter data set. The original parameter data includes parameter output format, transmission protocol, pressure, and tidal volume.

[0124] An abnormal alarm generation module 602 is configured to transmit the parameter data set to a local gateway based on a low-power Bluetooth protocol, judge the parameter data based on a preset threshold, trigger an abnormal alarm signal if the parameter exceeds the threshold, and generate an alarm data set including device identification and abnormal parameters.

[0125] The monitoring report generation module 603 is configured to analyze the dynamic change trend of the parameters in combination with the parameter data set and the alarm data set, and generate a monitoring report data set including aggregated parameters, alarm information and trend characteristics.

[0126] The intervention decision tracking module 604 is configured to extract a continuous abnormal trend based on the monitoring report data set, generate an intervention instruction after priority sorting according to the emergency degree, and track the processing state of the intervention instruction to obtain a final decision data set.

[0127] The above-mentioned non-invasive ventilator multi-parameter centralized monitoring system, the parameter standardization module is responsible for obtaining the original parameter data (including parameter output format, transmission protocol, pressure and tidal volume) and coding rules of each brand of non-invasive ventilator, and performing format unified processing on the original parameter data by constructing a standardized data mapping table to obtain a parameter data set; the abnormal alarm generation module transmits the parameter data set to the local gateway by using the Bluetooth low energy protocol, judges the parameter data based on a preset threshold, and triggers an abnormal alarm signal if the parameter exceeds the threshold to generate an alarm data set containing the device identifier and the abnormal parameter; the monitoring report generation module analyzes the dynamic change trend of the parameters in combination with the parameter data set and the alarm data set, and generates a monitoring report data set including aggregated parameters, alarm information and trend characteristics; and the intervention decision tracking module extracts a continuous abnormal trend based on the monitoring report data set, generates an intervention instruction after priority sorting according to the emergency degree, and tracks the processing state of the intervention instruction to obtain a final decision data set.

[0128] In the embodiment, the parameter standardization module solves the problem of data intercommunication of different brands of equipment and provides a unified basis for subsequent processing; the abnormal alarm generation module realizes timely identification and alarm of parameter abnormalities and reduces safety hazards; the monitoring report generation module provides comprehensive monitoring information through trend analysis to assist accurate judgment; and the intervention decision tracking module promotes efficient execution and optimization of abnormal processing.

[0129] It should be understood that, although each step in the flowchart involved in each embodiment as described above is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other sequences. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or steps or stages in other steps.

[0130] In one embodiment, a computer device is provided, comprising a memory and a processor, the memory stores a computer program, and the processor implements the steps of the method, system, device and medium for non-invasive ventilator multi-parameter centralized monitoring as described above when executing the computer program.

[0131] In one embodiment, a computer readable storage medium is provided, which stores a computer program, and the computer program implements the steps of the method embodiments when executed by a processor.

[0132] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts are described in the part of the method embodiments. The above described device embodiments are only schematic, and the components described as separate components can or can not be physically separate, and the components displayed as a unit can or can not be a physical unit, i.e. can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present disclosure according to actual needs. Those skilled in the art can understand and implement it without creative labor.

[0133] The above described embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the patent scope of the present application. It should be pointed out that, for those skilled in the art, without departing from the concept of the present application, several modifications and improvements can be made, which are all within the protection scope of the present application.

Claims

1. A method for multi-parameter centralized monitoring of non-invasive ventilators, characterized in that, The method comprises: Obtaining original parameter data and coding rules of each brand of noninvasive ventilator, constructing a standardized data mapping table, performing format uniform processing on the original parameter data to obtain a parameter data set; the original parameter data comprises parameter output format, transmission protocol, pressure and tidal volume; Transmitting the parameter data set to a local gateway based on a low-power Bluetooth protocol, judging the parameter data based on a preset threshold, triggering an abnormal alarm signal if the parameter exceeds the threshold, and generating an alarm data set comprising device identification and abnormal parameters; Analyzing the dynamic change trend of the parameters by combining the parameter data set and the alarm data set, and generating a monitoring report data set comprising aggregated parameters, alarm information and trend characteristics; Based on the monitoring report data set, extracting a continuous abnormal trend, prioritizing according to the emergency degree to generate an intervention instruction, and tracking the processing status of the intervention instruction to obtain a final decision data set.

2. The method of claim 1, wherein, The method comprises: Obtaining original parameter data, extracting business fields from the original parameter data to obtain a field set; Using a regular expression to parse the field set to generate a standardized field template; Using the standardized field template to construct a data mapping table based on the coding rules to determine the field correspondence relationship; the data mapping table is used to establish the corresponding mapping between the parameters of different brands of noninvasive ventilators and the unified standard parameters; Based on the field correspondence relationship, converting the original parameter data to obtain a parameter data set in a unified format; Using a data verification algorithm to judge the data integrity of the parameter data set, and generating a structured parameter data set if the verification is passed; the data verification algorithm is a checksum algorithm and a hash algorithm; Performing data normalization processing on the structured parameter data set to obtain a normalized parameter data set.

3. The method of claim 1, wherein, The method comprises: Transmitting the parameter data set comprising device identification and parameter values to a local gateway based on a low-power Bluetooth protocol; Using the local gateway to calculate the dynamic fluctuation coefficient of the parameter values in the parameter data set, and determining whether the parameters are abnormal based on a preset threshold interval to obtain a judgment result; If the dynamic fluctuation coefficient of the parameter values exceeds the preset threshold interval, an abnormal alarm signal is generated; Integrating the abnormal alarm signal, the dynamic fluctuation coefficient and the device identification to generate an alarm data set comprising abnormal parameter details.

4. The method of claim 3, wherein, The dynamic fluctuation coefficient is calculated by the following formula: wherein DFC represents a dynamic fluctuation coefficient, x i represents a parameter value at the i-th sampling time, represents a weighted mean value within a preset time window, ω i represents a time decay weight, ω i = e -k(n-i) , k represents a decay coefficient, and n represents a sampling number, represents a reference standard deviation.

5. The method of claim 1, wherein, The method comprises: Analyzing the dynamic change trend of the parameters by combining the parameter data set and the alarm data set, and generating a monitoring report data set comprising aggregated parameters, alarm information and trend characteristics; Obtain the time sequence parameters of the parameter data set and the alarm trigger condition of the alarm data set, and perform correlation calculation on both to obtain a change rate, thereby generating rate sequence data; the alarm trigger condition includes a parameter preset threshold interval, an abnormal duration threshold, and a device operation state reference value; Group the rate sequence data according to the device identifier, and calculate the rate mean value, peak value, and fluctuation rate in each window by using a sliding window algorithm, thereby obtaining an aggregated parameter set; Match the aggregated parameter set with the corresponding alarm start and end time stamps according to the device identifier, calculate the time stamp difference to extract the alarm duration, and integrate to generate trend feature data; Draw a trend curve with the time axis as the horizontal axis and the parameter change rate as the vertical axis, superimpose the alarm trigger threshold line and the abnormal period marker based on the trend feature data and the alarm trigger condition, and generate a data visualization graph; If there is an abnormal point in the data visualization graph, extract the device identifier, time node, and parameter fluctuation feature corresponding to the abnormal point, and generate a monitoring report data set according to the alarm trigger condition.

6. The method of claim 1, wherein, Extract the continuous abnormal trend based on the monitoring report data set, prioritize according to the emergency degree to generate an intervention instruction, and track the processing state of the intervention instruction to obtain a final decision data set, including: Filter the parameter records that exceed the preset threshold in the continuous sampling period from the monitoring report data set, extract the corresponding continuous abnormal trend feature, and construct an abnormal trend data sequence according to the device identifier and the time dimension; Based on the parameter fluctuation amplitude and the duration in the abnormal trend data sequence, an emergency degree evaluation is performed using a quantitative formula to obtain a priority sorting matrix; Generate a hierarchical intervention instruction based on the hierarchical division in the priority sorting matrix, and form a structured intervention instruction set containing the execution subject, execution time limit, and instruction number; Track the whole-process processing state of the structured intervention instruction set according to the instruction number, and collect instruction execution trajectory data containing the response time and execution progress; Combine the instruction execution trajectory data with the preset performance indicators to construct a multi-dimensional performance evaluation model, and output dynamic decision support information containing the details of the uncompleted instructions; the performance evaluation model performs regression analysis with the execution progress as the dependent variable and the response time as the independent variable; Based on the resource load data in the dynamic decision support information, optimize the resource allocation of the uncompleted instructions to generate a final decision data set.

7. The method of claim 6, wherein, The quantitative formula is represented by the following formula: wherein Δx represents a parameter fluctuation amplitude, Δx max represents a historical maximum fluctuation amplitude, t represents a duration, t max represents a historical maximum duration, represents a parameter change rate, represents a historical maximum change rate, and α, β, and γ represent weight coefficients.

8. A system for multi-parameter centralized monitoring of non-invasive ventilators, characterized by, The system includes: A parameter standardization module is configured to obtain original parameter data of non-invasive ventilators of various brands and coding rules, construct a standardized data mapping table, perform format unification processing on the original parameter data, and obtain a parameter data set; the original parameter data includes parameter output format, transmission protocol, pressure, and tidal volume; An abnormal alarm generation module is configured to transmit the parameter data set to a local gateway using a Bluetooth Low Energy protocol, judge the parameter data based on a preset threshold, trigger an abnormal alarm signal if the parameter exceeds the threshold, and generate an alarm data set including a device identifier and an abnormal parameter; The monitoring report generation module is configured to analyze dynamic change trends of the parameters by combining the parameter data set and the alarm data set, and generate a monitoring report data set including aggregated parameters, alarm information and trend characteristics. The intervention decision tracking module is configured to extract continuous abnormal trends based on the monitoring report data set, generate intervention instructions after priority sorting according to emergency degrees, and track processing states of the intervention instructions to obtain a final decision data set. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor, when executing the computer program, implements the steps of the method in any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by the processor, implements the steps of the method in any one of claims 1 to 7.

Images