Infusion cluster management system and method for real-time monitoring and AI early warning

Abstract

The invention discloses an infusion cluster management system and method for real-time monitoring and AI early warning, and relates to the technical field of infusion cluster management. The method comprises the steps that correlation verification of patient information and medicine information is executed; if the correlation verification is passed, after the patient information and the infusion medicine information are read, the fusion sensor group is activated, and infusion medicine monitoring is executed; receiving the real-time monitoring data set, performing data time sequence change analysis on the real-time monitoring data set, establishing a first monitoring early warning, initializing an anomaly recognition network according to patient information and infusion medicine information, performing anomaly recognition on the real-time monitoring data set based on the anomaly recognition network, and establishing a second monitoring early warning; and the first monitoring early warning and the second monitoring early warning are pushed to a hospital information system and medical care mobile terminal equipment, and early warning is reported. The technical problems that in the prior art, monitoring is not accurate in the infusion process, and early warning response is delayed are solved, and the technical effect of improving safety, accuracy and timeliness of the infusion process is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of infusion cluster management, and in particular to an infusion cluster management system and method with real-time monitoring and AI early warning. Background Art

[0002] As demand for healthcare services increases, the infusion process involves patient safety management. However, as this demand continues to grow, traditional infusion monitoring methods face increasing challenges. Traditional infusion monitoring often relies on manual inspections or patient notifications. This approach suffers from low monitoring accuracy, slow response times, and susceptibility to human error, making it difficult to meet the modern healthcare industry's demand for efficient and safe infusion monitoring. Summary of the Invention

[0003] The present application provides an infusion cluster management system and method with real-time monitoring and AI early warning, which is used to solve the technical problems of inaccurate infusion process monitoring and delayed early warning response in the existing technology.

[0004] In view of the above problems, the present application provides an infusion cluster management system and method with real-time monitoring and AI early warning.

[0005] The first aspect of the present application provides an infusion cluster management system with real-time monitoring and AI early warning, the system comprising:

[0006] The information entry module is used to scan the patient's wristband barcode and the infusion drug barcode through the PDA to perform the association verification of the patient information and the drug information; the monitoring configuration module is used to read the patient information and the infusion drug information if the association verification is passed, activate the fusion sensor group, and perform infusion drug monitoring. The fusion sensor group includes an infrared sensor, a pressure sensor, and a capacitive sensor; the AI ​​early warning module is used to receive the real-time monitoring data set, perform data time series change analysis on the real-time monitoring data set, establish a first monitoring early warning, initialize the abnormality recognition network according to the patient information and the infusion drug information, and then perform abnormality recognition on the real-time monitoring data set based on the abnormality recognition network to establish a second monitoring early warning; the interaction module is used to push the first monitoring early warning and the second monitoring early warning to the hospital information system and the medical mobile terminal device, and report the early warning.

[0007] The second aspect of the present application provides an infusion cluster management method with real-time monitoring and AI early warning, the method comprising:

[0008] Scan the barcode on the patient's wristband and the infusion drug barcode through the PDA to perform association verification of the patient information and the drug information; if the association verification is passed, after reading the patient information and the infusion drug information, activate the fusion sensor group to perform infusion drug monitoring, and the fusion sensor group includes an infrared sensor, a pressure sensor, and a capacitive sensor; receive a real-time monitoring data set, perform data time series change analysis on the real-time monitoring data set, establish a first monitoring warning, initialize the abnormality recognition network according to the patient information and the infusion drug information, perform abnormality recognition on the real-time monitoring data set based on the abnormality recognition network, and establish a second monitoring warning; push the first monitoring warning and the second monitoring warning to the hospital information system and the medical mobile terminal device, and report the warning.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] This application uses a PDA to scan the barcode on the patient's wristband and the infusion drug barcode to perform association verification of the patient information and the drug information; if the association verification is passed, after reading the patient information and the infusion drug information, the fusion sensor group is activated to perform infusion drug monitoring, and the fusion sensor group includes an infrared sensor, a pressure sensor, and a capacitive sensor; a real-time monitoring data set is received, and a data time series change analysis is performed on the real-time monitoring data set to establish a first monitoring warning. After initializing an abnormality recognition network based on the patient information and the infusion drug information, an abnormality recognition is performed on the real-time monitoring data set based on the abnormality recognition network to establish a second monitoring warning; the first monitoring warning and the second monitoring warning are pushed to the hospital information system and the medical mobile terminal device, and an early warning is reported. The present invention solves the technical problems of inaccurate infusion process monitoring and delayed early warning response in the existing technology. By integrating multiple sensors to monitor the infusion process in real time, AI algorithms to analyze real-time data and generate accurate early warnings, and an automated system to control the infusion process, the technical effect of improving the safety, accuracy and timeliness of the infusion process is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 A schematic diagram of the structure of an infusion cluster management system with real-time monitoring and AI early warning provided in an embodiment of the present application;

[0013] Figure 2A flow chart of a method for managing an infusion cluster with real-time monitoring and AI early warning provided in an embodiment of the present application.

[0014] Explanation of the accompanying drawings: information entry module 11, monitoring configuration module 12, AI early warning module 13, interaction module 14. DETAILED DESCRIPTION

[0015] This application provides an infusion cluster management system and method with real-time monitoring and AI early warning, aiming to solve the technical problems of inaccurate infusion process monitoring and delayed early warning response in the existing technology. By integrating multiple sensors to monitor the infusion process in real time, AI algorithms to analyze real-time data and generate accurate early warnings, and an automated system to control the infusion process, the technical effect of improving the safety, accuracy and timeliness of the infusion process is achieved.

[0016] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only some of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0017] It should be noted that any variations of the terms "include" and "have" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices.

[0018] Example 1, as Figure 1 As shown, the embodiment of the present application provides an infusion cluster management system with real-time monitoring and AI early warning, which includes:

[0019] The information entry module 11 is used to scan the barcode on the patient's wristband and the infusion drug barcode through the PDA to perform the association verification of the patient information and the drug information.

[0020] In this embodiment of the present application, the information entry module 11 uses a PDA device to scan the barcode on the patient's wristband and the infusion drug barcode to perform correlation verification between patient and drug information. Specifically, the PDA uses barcode scanning technology to read the barcode on the patient's wristband. The barcode contains the patient's basic information, such as name and medical record number. The PDA's built-in scanner uses optical recognition technology to read and enter the information from the barcode. This operation quickly and accurately obtains the patient's identity and reduces manual input errors. The PDA then scans the infusion drug barcode, which contains information such as the drug's name, strength, and dosage. The PDA also uses the same scanning method to read the drug's detailed data. Subsequently, based on the matching of the patient and drug information, correlation verification is performed. To ensure that the drug matches the patient's treatment needs, correlation verification uses a database matching method to compare the patient's medical record information with the drug's indications, dosage requirements, and other data. By reviewing the patient's medical record information, the drug's usage requirements, such as drug allergies, are verified, and the drug's suitability for the current patient's treatment is verified. If the patient and drug information match and the association verification is passed, subsequent treatment can continue; if a mismatch or potential risk is found, a warning will be issued immediately, prompting medical staff to check and adjust to ensure the patient's safety.

[0021] The monitoring configuration module 12 is used to activate the fusion sensor group and perform infusion drug monitoring after reading the patient information and infusion drug information if the association verification is passed. The fusion sensor group includes an infrared sensor, a pressure sensor, and a capacitive sensor.

[0022] In this embodiment of the present application, the monitoring configuration module 12 initiates the infusion drug monitoring process after association verification. Specifically, it first reads the patient information and infusion drug information that have passed association verification. This information provides the necessary data support for subsequent monitoring configuration, such as the patient's treatment requirements and drug characteristics. Next, based on the read information, the fusion sensor group is activated. The fusion sensor group is composed of a combination of different types of sensors, including infrared sensors, pressure sensors, and capacitive sensors.

[0023] Infrared sensors monitor the flow of liquid within infusion tubing. Using infrared technology, the sensors detect changes in liquid flow, such as a drop or stop in the liquid level, thereby providing real-time monitoring of the normal infusion process. Pressure sensors, installed at specific locations within the infusion tubing, primarily measure pressure changes during the infusion process. These pressure changes can indicate smooth infusion flow; abnormal pressure conditions indicate a blockage or other abnormality in the tubing. Capacitive sensors monitor changes in the conductivity of the liquid in the tubing to determine whether the liquid is flowing normally or whether there is air ingress. These sensors work simultaneously to collect real-time data and monitor various parameters during the infusion process.

[0024] The AI ​​early warning module 13 is used to receive a real-time monitoring data set, perform data time series change analysis on the real-time monitoring data set, establish a first monitoring early warning, initialize an abnormality recognition network based on the patient information and the infusion drug information, perform abnormality recognition on the real-time monitoring data set based on the abnormality recognition network, and establish a second monitoring early warning.

[0025] In this embodiment of the present application, the AI ​​early warning module 13 first receives a real-time monitoring data set and performs a time-series analysis on the data. Specifically, a forward search window is created with the current time node as the zero point. Monitoring data within this window is retrieved and compared for data fluctuations. This comparison is used to identify fluctuation anomalies in the data. Simultaneously, a trigger analysis is performed using a static comparison threshold to detect any abnormal triggering events. A first monitoring early warning is then established based on the fluctuation anomaly and the triggering anomaly.

[0026] After establishing the first monitoring alert, the anomaly recognition network is initialized based on the patient's individual information (such as age, medical record number, allergy history, etc.) and infusion drug information (such as drug name, dosage, etc.). By reading relevant information about the patient and the drug, personalized treatment parameters are customized for each patient. For example, some patients may need special adjustments to the drug infusion rate due to factors such as age and underlying diseases, or the contraindications of certain drugs require more stringent infusion methods. Utilizing this personalized input data, background support is provided for the anomaly recognition network, ensuring that the network can flexibly adapt to the specific needs of the patient and the characteristics of the drug.

[0027] Once the anomaly recognition network is initialized, it is used to identify anomalies in the real-time monitoring data set. This data is compared with the optimal infusion parameters generated based on patient and medication information to determine if there are any abnormal deviations. For example, if the real-time data deviates significantly from the set optimal infusion parameters, the anomaly recognition network will identify this deviation and generate a second monitoring alert.

[0028] Furthermore, in the system provided in the embodiment of the application, the AI ​​early warning module 13 performs data time series change analysis on the real-time monitoring data set, further comprising:

[0029] A forward search window is created with the current time node as the time zero point; monitoring data within the forward search window is obtained, and data fluctuation comparison of the real-time monitoring data set is performed based on the monitoring data to establish a fluctuation anomaly; a static comparison threshold is obtained, and trigger analysis of the real-time monitoring data set is performed using the static comparison threshold to establish a trigger anomaly; the first monitoring warning is established based on the fluctuation anomaly and the trigger anomaly.

[0030] In the embodiment of the present application, the current time node is first used as the time zero point, and a forward search window is created on this basis. This step facilitates the backtracking and analysis of historical monitoring data by determining a time starting point. Time series data analysis technology sets the time zero point and backtracks historical data within a certain time range from the current moment. The size of this time window is usually set according to the time range of the infusion to ensure that sufficient historical data can be obtained to provide support for subsequent comparison and analysis. Through this step, a period of historical data is obtained, which will serve as background information for subsequent analysis.

[0031] Next, the monitoring data is obtained from the forward search window. These data include key monitoring parameters such as liquid flow status, drip rate, pressure fluctuation, and liquid level change. These data are collected in real time by sensors (such as pressure sensors, infrared sensors, etc.) and represent the dynamic changes during the infusion process. By obtaining these real-time data, the basic conditions of the current infusion are determined. Based on these data, data fluctuation comparison is performed. This step compares the current monitoring data with the normal fluctuation pattern in the historical data to identify whether there are abnormal fluctuations in the data. For example, if the fluctuation range of the liquid level change in the historical data is certain, and the current data shows fluctuations beyond the normal range, this fluctuation will be identified as a potential anomaly. Through this step, a fluctuation anomaly is generated, that is, the deviation between the real-time data and the historical normal pattern is identified, indicating a possible infusion abnormality.

[0032] Next, a static comparison threshold, pre-set by technical experts, is obtained. This threshold is used to trigger analysis of real-time monitoring data, specifically to detect whether the real-time data exceeds the static threshold. If the real-time data exceeds the set threshold, it is marked as a trigger anomaly, indicating a possible infusion problem, such as an excessively fast infusion rate or poor fluid flow. This step establishes a trigger anomaly and further confirms abnormal data.

[0033] Finally, based on fluctuation anomalies and trigger anomalies, a first monitoring and early warning is established. This step combines fluctuation analysis of real-time data with trigger analysis of static thresholds to comprehensively determine whether there are potential infusion issues. Fluctuation anomalies reflect changes in the data's fluctuation range, while trigger anomalies highlight whether the data exceeds the set safety range. By integrating these two types of anomalies, potential risks during the infusion process are determined and a first monitoring and early warning is generated, reminding medical staff to pay attention to and make adjustments to the infusion process.

[0034] Furthermore, in the system provided in the embodiment of the application, the AI ​​warning module 13 performs anomaly identification on the real-time monitoring data set based on the anomaly identification network and establishes a second monitoring warning, further comprising:

[0035] An infusion adaptation database is called according to the infusion drug information; a matching analysis of the infusion adaptation database is performed based on the patient information to establish optimal infusion parameters; a deviation comparison between the real-time monitoring data set and the optimal infusion parameters is performed using the anomaly recognition network to establish a deviation comparison result; a matching channel is called to identify the matching impact of the deviation comparison result and the patient information, and the second monitoring warning is established according to the matching impact identification result.

[0036] In this embodiment of the present application, an infusion adaptation database is first accessed based on the infusion drug information. This database contains various drug information, such as the drug's indication, dosage, infusion rate, and infusion method. By accessing this database, detailed drug requirements are obtained, providing the necessary drug background information for subsequent matching analysis. For example, if a drug is indicated for a specific condition or requires a specific infusion rate, this information can be used to determine whether the drug is suitable for the patient.

[0037] Next, a matching analysis of the infusion adaptation database is performed based on the patient's information (such as age, gender, medical history, allergy history, etc.). The purpose of this step is to evaluate the suitability of the drug based on the patient's specific situation. By comparing the patient's information with the indications, dosage requirements, contraindications, etc. of the drugs in the database, ensure that the selected drug is consistent with the patient's treatment needs. If a drug is not suitable for the patient's health status, the drug infusion plan is adjusted according to the patient's individual characteristics (such as renal function, weight, age, etc.), and the optimal infusion parameters are generated, including the drug dosage, infusion rate, infusion method, etc. suitable for the patient, so as to ensure the individualization and safety of the treatment.

[0038] After obtaining the optimal infusion parameters, the initialized anomaly recognition network is used to compare deviations between the real-time monitoring data set and the optimal infusion parameters. Specifically, various parameters in the real-time monitoring data (such as drip rate and liquid level changes) are directly compared with the optimal infusion parameters determined based on patient information, identifying deviations and generating deviation comparison results.

[0039] The matching channel is then called to match the deviation comparison results with the patient information for impact identification. The purpose of this step is to analyze the potential impact of the deviation on the patient's health. Based on the patient information, it is determined whether this deviation poses a threat to the patient's safety. For example, if the infusion rate deviation in the deviation comparison result exceeds the preset threshold, and the patient has heart disease or other rate-sensitive health problems, the possible risks of this deviation are identified. The matching channel judges the deviation according to preset rules (such as whether the patient has heart disease, high blood pressure, etc.) to ensure that the warning is triggered only when the deviation has an actual impact on the patient. Finally, based on the matching impact identification results, a second monitoring warning is generated.

[0040] The interactive module 14 is used to push the first monitoring warning and the second monitoring warning to the hospital information system and the medical mobile terminal device, and report the warning.

[0041] In this embodiment of the present application, the interaction module 14 transmits the first and second monitoring warnings to the hospital information system via a standard communication protocol (e.g., WebSocket), ensuring that medical staff can view the relevant warnings promptly on their workstations. These warnings include the warning content and patient-related information (e.g., medical record number, treatment plan, etc.), helping medical staff quickly understand the situation and respond.

[0042] At the same time, the interaction module 14 uses wireless communication technologies (such as Wi-Fi and Bluetooth) to push the first and second monitoring warnings to the medical staff's mobile devices. These devices may include smartphones, tablets, etc., so that medical staff can receive warning notifications in real time regardless of their location. This process is usually completed through a push service (such as APN), ensuring that the warning information can be delivered to medical staff in a timely and accurate manner.

[0043] After receiving the warning, the interactive module 14 will remind the medical staff through the device's display interface, sound or vibration, etc., to ensure that the warning is not missed or ignored.

[0044] Furthermore, the system provided in the application embodiment also includes:

[0045] The linkage warning module is used to obtain the monitoring data of the smart bracelet worn by the patient, establish a linkage data set, perform a delayed linkage analysis on the linkage data set and the real-time monitoring data set, and establish a linkage feedback warning; the compensation module is used to receive the linkage feedback warning, compensate for the first monitoring warning and the second monitoring warning, and then issue a warning.

[0046] In an embodiment of the present application, the linkage warning module obtains the monitoring data of the smart bracelet worn by the patient, which includes physiological data collected by the bracelet, such as heart rate, blood pressure, body temperature, etc. These data are organized into a linkage data set, and a hysteresis linkage analysis is performed with the real-time monitoring data set during the infusion process (such as drip rate, liquid level change, infusion rate, etc.). The purpose of this analysis is to find the time difference between physiological data and infusion data, and to determine the optimal lag time between the two through techniques such as cross-correlation function. By adjusting the timeline of these data, the correlation between physiological data and the infusion process is captured, thereby identifying potential anomalies. Based on this analysis, a linkage feedback warning is established.

[0047] After generating a linked feedback alert, the compensation module corrects the previously generated first and second monitoring alerts based on this alert. Specifically, the compensation module first receives the linked feedback alert from the linked alert module. This feedback information includes the correlation between physiological status (such as heart rate and blood pressure fluctuations) and real-time monitoring data (such as infusion rate and drip rate). Suppose the linked feedback indicates that there is a time lag between the patient's heart rate fluctuation and the deviation from the infusion rate—information that was not fully captured by the original monitoring alert. Next, the original first monitoring alert (based on temporal change analysis) and second monitoring alert (based on the anomaly recognition network) are examined to see if any of these alerts failed to account for the linked data. If these alerts fail to fully reflect the potential risks of the patient's physiological changes, the compensation module will initiate corrections. The compensation module adjusts the original alert results based on the linked feedback alert data. For example, if the linked feedback indicates a mismatch between the patient's heart rate fluctuation and the infusion rate, and that this fluctuation could affect the patient's health, the compensation module will enhance the alert, ensuring that medical staff can detect this potential problem earlier. Compensation methods include adjusting the warning time window based on the linkage data, increasing the intensity of warnings for specific risks, or reassessing infusion parameters based on changes in physiological data. Finally, the revised warning information is transmitted to the hospital information system and medical staff's mobile devices via a push service.

[0048] Furthermore, in the system provided by the embodiment of the application, the linkage warning module performs a delayed linkage analysis on the linkage data set and the real-time monitoring data set, further comprising:

[0049] The linkage dataset and the real-time monitoring dataset, which have been time-series aligned using a cross-correlation function, are subjected to cross-correlation analysis to establish an optimal lag time; the linkage dataset and the real-time monitoring dataset are realigned using the optimal lag time; and linkage analysis of physiological data abnormalities is performed based on the realigned linkage dataset and the real-time monitoring dataset.

[0050] In an embodiment of the present application, the linkage data set and the real-time monitoring data set are first time-series aligned using a cross-correlation function. The cross-correlation function is a statistical method used to measure the similarity between two sets of signals, especially the lag relationship between them. Specifically, physiological data (such as heart rate, blood pressure, etc.) and infusion monitoring data (such as drip rate, liquid level changes, etc.) are first input as two time series. The cross-correlation function compares the two signals using a sliding window method. The sliding window method refers to moving a signal (such as physiological data) on the time axis by a certain step size (i.e., time offset) and calculating the correlation between the two sets of data at each time offset. This method can help identify the time difference between the two signals. By moving the time axis of the physiological data and calculating the correlation at each time offset, the best matching time between the physiological data and the infusion monitoring data is found. When calculating the similarity, the cross-correlation function uses the Pearson correlation coefficient to measure the similarity between the two signals. By calculating the correlation for each time offset, the cross-correlation function determines the optimal lag time, that is, the position with the greatest similarity under the time offset between the two sets of signals. Optimal lag time represents the optimal alignment between physiological data (such as heart rate) and infusion monitoring data (such as drip rate). It reflects the delay or advance of one signal (physiological data) relative to the other (infusion monitoring data). This lag time provides critical time alignment for subsequent analysis, enabling accurate matching of physiological and infusion monitoring data.

[0051] Next, the linkage dataset and the real-time monitoring dataset are realigned using the optimal lag time. This step is a practical application of the cross-correlation analysis results. This involves temporally aligning and synchronizing the data in the two datasets based on the calculated optimal lag time. This realignment ensures that the linkage data and the monitoring data are compared on the same timeline, ensuring that their trends and fluctuations accurately reflect the true relationship. This realignment can be accomplished through timestamp adjustment or interpolation algorithms to ensure time series consistency between the two datasets.

[0052] After the data alignment is completed, a linkage analysis of physiological data anomalies is performed based on the realigned linkage data set and the real-time monitoring data set. This analysis mainly identifies potential anomalies between the physiological state and the key parameters of the infusion process by examining the aligned data. For example, if there is an abnormal fluctuation in the relationship between the heart rate (data monitored by the smart bracelet) and the infusion rate, this anomaly is detected. The linkage analysis compares the change patterns between the two data sets to find, for example, patterns of lagged changes in physiological data, and whether these changes coincide with anomalies in the infusion process (such as too fast a rate, abnormal changes in liquid level, etc.). In this way, the abnormal linkage relationship between the physiological state and the infusion parameters can be identified in a timely manner, and relevant early warning information can be generated.

[0053] Furthermore, the system provided in the application embodiment also includes:

[0054] A prediction and warning module is used to perform joint time series anomaly prediction based on the real-time monitoring data set and the linkage data set, and establish a joint time series anomaly prediction result; an early warning update module is used to update the first monitoring early warning and the second monitoring early warning based on the joint time series anomaly prediction result.

[0055] In an embodiment of the present application, the prediction and warning module first performs a joint time series anomaly prediction through the real-time monitoring data set and the linkage data set. This process uses a long short-term memory network (LSTM) to perform time series prediction. LSTM is a deep learning model specifically used to process and predict time series data. It can capture long-term dependencies through memory units and is very suitable for processing the time series relationship between physiological data and infusion monitoring data. Specifically, the real-time monitoring data set (such as infusion rate, liquid level changes, etc.) and the linkage data set (such as heart rate, blood pressure and other physiological data) are used as input, and the LSTM network is used to train the historical data to learn the laws and potential abnormal patterns therein. The LSTM model predicts abnormal events that may occur in the future time period by identifying the time series patterns in the data. The prediction result is the joint time series anomaly prediction result, which provides a basis for subsequent warning updates, that is, the predicted potential health risks or treatment problems.

[0056] Next, the warning update module updates the primary and secondary monitoring warnings based on the combined time series anomaly prediction results. This update process is typically implemented using a rules engine. Based on pre-set business rules and algorithms, the rules engine dynamically adjusts the existing warning information, combining prediction results with current monitoring data. Specifically, the system compares the predicted time point of potential anomalies with current monitoring data (such as drip rate, fluid level changes, and the patient's physiological data) and updates the warning information based on the predicted anomaly type, intensity, and timing. For example, if the LSTM predicts that a patient may experience an abnormal heart rate or poor fluid flow in the upcoming time period, the rules engine will update the primary monitoring warning (generated based on time series analysis) and the secondary monitoring warning (generated based on the anomaly recognition network) based on this information, increasing the warning priority or adjusting the warning's response measures. The updated warning information provides more timely and accurate warnings, helping medical staff make adjustments before problems occur.

[0057] Through the above two steps, the prediction and warning module uses the LSTM network to predict anomalies in time series data, ensuring that potential risks can be identified in advance; the warning update module uses the rule engine to dynamically update warning information to ensure the accuracy and timeliness of the warning.

[0058] Furthermore, the system provided in the application embodiment also includes:

[0059] The cut-off solenoid valve is used to control the cut-off solenoid valve to cut off the infusion when the reported warning meets the preset braking conditions, and simultaneously send the cut-off information to the medical mobile terminal device.

[0060] In an embodiment of the present application, the function of the cut-off solenoid valve is to automatically cut off the infusion and notify the medical staff when the generated warning meets the preset braking conditions. Specifically, when the warning information (such as excessive dripping speed, abnormal liquid level, etc.) is generated based on the real-time monitoring data and the preset safety threshold, and the warning meets the preset braking conditions, the control system will trigger the cut-off solenoid valve. The preset braking conditions refer to standards set according to medical safety requirements. For example, when the infusion rate exceeds a predetermined range or the liquid level drops abnormally, it is automatically determined that the infusion process may pose a risk to the patient. At this time, the cut-off solenoid valve will receive a control signal, cut off the infusion flow path through the electronic control mechanism, immediately stop the infusion of the liquid, and prevent potential medical risks. At the same time, the cut-off information is synchronously sent to the medical mobile terminal device through wireless communication technology (such as Wi-Fi, Bluetooth, etc.) to ensure that medical staff can receive this critical notification in real time.

[0061] Furthermore, the system provided in the application embodiment also includes:

[0062] The image acquisition module is used to collect the liquid level information in the infusion bottle, generate a third monitoring warning based on the liquid level information in the bottle, and synchronously push the third monitoring warning to the hospital information system and the medical mobile terminal device.

[0063] In the embodiments of the present application, the image acquisition module uses a camera or liquid level sensor to capture real-time images of the infusion bottle to obtain information about the liquid level within the bottle. Liquid level monitoring is achieved using a high-definition camera, infrared sensor, or laser sensor. Image processing algorithms (such as edge detection and image segmentation) are used to analyze the liquid level in the images or videos captured by the camera. The liquid level information within the infusion bottle is obtained by calculating the distance between the liquid surface and the bottle opening.

[0064] Next, a third monitoring alert is generated based on the acquired bottle liquid level information. Specifically, the bottle liquid level information is compared with a preset safety level standard. If the liquid level falls below a safety threshold (e.g., near the bottom of the bottle), the liquid is deemed to be nearing exhaustion, and a third monitoring alert is generated. Finally, the generated third monitoring alert is synchronously pushed to the hospital information system (HIS) and the medical staff's mobile devices via wireless communication protocols (such as Wi-Fi, Bluetooth, 4G / 5G, etc.).

[0065] Furthermore, the system provided in the application embodiment also includes:

[0066] The acquisition frequency constraint module is used to configure the acquisition frequency of the fusion sensor group according to the patient information, and after the fusion sensor group is activated, perform infusion drug monitoring based on the acquisition frequency.

[0067] In this embodiment of the present application, the acquisition frequency constraint module configures the acquisition frequency of the fusion sensor group based on patient information. This configuration process is typically determined based on the patient's age, health status, and treatment needs. For example, preset rules can be used to adjust the data acquisition frequency based on the patient's age, disease type, and urgency.

[0068] For example, when configuring the acquisition frequency based on age group, for children or elderly patients, since they are more sensitive to the treatment process, a higher acquisition frequency (such as once per second) is set to detect potential abnormalities in a timely manner and ensure that every detail of the treatment process is monitored in real time. For adult patients, if their condition is stable and they do not have acute symptoms, a lower acquisition frequency (such as once per minute) is selected to reduce the burden on the equipment while meeting clinical monitoring needs.

[0069] After configuring the acquisition frequency, the fusion sensor group is activated. This group typically includes multiple sensors, such as infrared sensors, pressure sensors, and capacitive sensors, capable of monitoring multiple key parameters during the infusion process (such as drug flow rate, liquid level changes, and infusion pressure). Based on the configured acquisition frequency, these sensors are controlled to synchronously collect data at preset intervals. This way, the sensors will perform monitoring tasks at each time point according to the configured frequency, ensuring the accuracy and real-time nature of the data.

[0070] In the embodiments of the present application, in summary, the embodiments of the present application have at least the following technical effects:

[0071] This application uses a PDA to scan the barcode on the patient's wristband and the infusion drug barcode to perform association verification of the patient information and the drug information; if the association verification is passed, after reading the patient information and the infusion drug information, the fusion sensor group is activated to perform infusion drug monitoring, and the fusion sensor group includes an infrared sensor, a pressure sensor, and a capacitive sensor; a real-time monitoring data set is received, and a data time series change analysis is performed on the real-time monitoring data set to establish a first monitoring warning. After initializing an abnormality recognition network based on the patient information and the infusion drug information, an abnormality recognition is performed on the real-time monitoring data set based on the abnormality recognition network to establish a second monitoring warning; the first monitoring warning and the second monitoring warning are pushed to the hospital information system and the medical mobile terminal device, and an early warning is reported. The present invention solves the technical problems of inaccurate infusion process monitoring and delayed early warning response in the existing technology. By integrating multiple sensors to monitor the infusion process in real time, AI algorithms to analyze real-time data and generate accurate early warnings, and an automated system to control the infusion process, the technical effect of improving the safety, accuracy and timeliness of the infusion process is achieved.

[0072] Example 2, based on the same inventive concept as the infusion cluster management system with real-time monitoring and AI early warning in the previous embodiment, Figure 2 As shown, the embodiment of the present application provides an infusion cluster management method with real-time monitoring and AI early warning, the method comprising:

[0073] Scan the barcode on the patient's wristband and the infusion drug barcode through the PDA to perform association verification of the patient information and the drug information; if the association verification is passed, after reading the patient information and the infusion drug information, activate the fusion sensor group to perform infusion drug monitoring, and the fusion sensor group includes an infrared sensor, a pressure sensor, and a capacitive sensor; receive a real-time monitoring data set, perform data time series change analysis on the real-time monitoring data set, establish a first monitoring warning, initialize the abnormality recognition network according to the patient information and the infusion drug information, perform abnormality recognition on the real-time monitoring data set based on the abnormality recognition network, and establish a second monitoring warning; push the first monitoring warning and the second monitoring warning to the hospital information system and the medical mobile terminal device, and report the warning.

[0074] Furthermore, performing data time series change analysis on the real-time monitoring data set, the method further includes:

[0075] A forward search window is created with the current time node as the time zero point; monitoring data within the forward search window is obtained, and data fluctuation comparison of the real-time monitoring data set is performed based on the monitoring data to establish a fluctuation anomaly; a static comparison threshold is obtained, and trigger analysis of the real-time monitoring data set is performed using the static comparison threshold to establish a trigger anomaly; the first monitoring warning is established based on the fluctuation anomaly and the trigger anomaly.

[0076] Furthermore, based on the anomaly recognition network, anomaly recognition is performed on the real-time monitoring data set to establish a second monitoring warning. The method further includes:

[0077] An infusion adaptation database is called according to the infusion drug information; a matching analysis of the infusion adaptation database is performed based on the patient information to establish optimal infusion parameters; a deviation comparison between the real-time monitoring data set and the optimal infusion parameters is performed using the anomaly recognition network to establish a deviation comparison result; a matching channel is called to identify the matching impact of the deviation comparison result and the patient information, and the second monitoring warning is established according to the matching impact identification result.

[0078] Furthermore, the method further comprises:

[0079] Acquire the monitoring data of the smart bracelet worn by the patient, establish a linkage data set, perform a delayed linkage analysis on the linkage data set and the real-time monitoring data set, and establish a linkage feedback warning; after receiving the linkage feedback warning, compensate for the first monitoring warning and the second monitoring warning, and then report a warning.

[0080] Furthermore, performing a lagged linkage analysis on the linkage dataset and the real-time monitoring dataset, the method further includes:

[0081] The linkage dataset and the real-time monitoring dataset, which have been time-series aligned using a cross-correlation function, are subjected to cross-correlation analysis to establish an optimal lag time; the linkage dataset and the real-time monitoring dataset are realigned using the optimal lag time; and linkage analysis of physiological data abnormalities is performed based on the realigned linkage dataset and the real-time monitoring dataset.

[0082] Furthermore, the method further comprises:

[0083] Perform joint time series anomaly prediction based on the real-time monitoring data set and the linkage data set to establish a joint time series anomaly prediction result; and update the first monitoring warning and the second monitoring warning based on the joint time series anomaly prediction result.

[0084] Furthermore, the method further comprises:

[0085] When the alarm meets the preset braking conditions, the solenoid valve is controlled to cut off the infusion, and the cut-off information is simultaneously sent to the medical mobile device.

[0086] Furthermore, the method further comprises:

[0087] Collect the liquid level information in the infusion bottle, generate a third monitoring warning based on the liquid level information in the bottle, and synchronously push the third monitoring warning to the hospital information system and the medical mobile terminal device.

[0088] Furthermore, the method further comprises:

[0089] The acquisition frequency of the fusion sensor group is configured according to the patient information, and after the fusion sensor group is activated, the infusion drug monitoring is performed based on the acquisition frequency.

[0090] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0091] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

[0092] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. A real-time monitoring and AI early warning infusion cluster management system, characterized by: The management system includes: The information entry module is used to scan the barcode of the patient's wristband and the infusion drug barcode through the PDA to perform the association verification of the patient information and drug information; A monitoring configuration module is used to activate a fusion sensor group to perform infusion drug monitoring after reading patient information and infusion drug information if the association verification passes. The fusion sensor group includes an infrared sensor, a pressure sensor, and a capacitive sensor; An AI early warning module is configured to receive a real-time monitoring data set, perform a data time series change analysis on the real-time monitoring data set, establish a first monitoring early warning, initialize an abnormality recognition network based on the patient information and the infusion drug information, perform abnormality recognition on the real-time monitoring data set based on the abnormality recognition network, and establish a second monitoring early warning; The interactive module is used to push the first monitoring warning and the second monitoring warning to the hospital information system and medical mobile devices, and report the warning.

2. The infusion cluster management system with real-time monitoring and AI early warning according to claim 1, characterized in that: In the AI ​​early warning module, the real-time monitoring data set is analyzed for time series changes, including: Create a forward search window with the current time node as time zero; Acquire monitoring data within a forward search window, and perform data fluctuation comparison of the real-time monitoring data set based on the monitoring data to establish fluctuation anomalies; Obtaining a static comparison threshold, using the static comparison threshold to perform trigger analysis on the real-time monitoring data set, and establishing a trigger anomaly; The first monitoring and early warning is established according to the fluctuation anomaly and the triggering anomaly.

3. The real-time monitoring and AI early warning infusion cluster management system according to claim 1 is characterized in that: In the AI ​​early warning module, anomaly identification is performed on the real-time monitoring data set based on the anomaly identification network to establish a second monitoring early warning, including: calling an infusion adaptation database according to the infusion drug information; Performing a matching analysis on the infusion adaptation database based on the patient information to establish optimal infusion parameters; Using the anomaly recognition network to perform a deviation comparison between the real-time monitoring data set and the optimal infusion parameter, and establishing a deviation comparison result; The matching channel is called to identify the matching impact of the deviation comparison result and the patient information, and the second monitoring warning is established according to the matching impact identification result.

4. The real-time monitoring and AI early warning infusion cluster management system according to claim 1, characterized in that: The management system further comprises: A linkage warning module is used to obtain monitoring data from the smart bracelet worn by the patient, establish a linkage data set, perform a hysteresis linkage analysis on the linkage data set and the real-time monitoring data set, and establish a linkage feedback warning; The compensation module is used to receive the linkage feedback warning, compensate the first monitoring warning and the second monitoring warning, and then report the warning.

5. The real-time monitoring and AI early warning infusion cluster management system according to claim 4, characterized in that: In the linkage warning module, a delayed linkage analysis is performed on the linkage data set and the real-time monitoring data set, including: Performing cross-correlation analysis on the linkage dataset and the real-time monitoring dataset after time series alignment using a cross-correlation function to establish an optimal lag time; realigning the linkage dataset and the real-time monitoring dataset using the optimal lag time; A linkage analysis of abnormal physiological data is performed based on the realigned linkage data set and the real-time monitoring data set.

6. The real-time monitoring and AI early warning infusion cluster management system according to claim 5, characterized in that: The management system further comprises: A prediction and warning module, configured to perform joint time series anomaly prediction based on the real-time monitoring data set and the linkage data set, and establish a joint time series anomaly prediction result; An early warning update module is used to update the first monitoring early warning and the second monitoring early warning according to the joint time series anomaly prediction result.

7. The real-time monitoring and AI early warning infusion cluster management system according to claim 1, characterized in that: The management system further comprises: The cut-off solenoid valve is used to control the cut-off solenoid valve to cut off the infusion when the reported warning meets the preset braking conditions, and simultaneously send the cut-off information to the medical mobile terminal device.

8. The real-time monitoring and AI early warning infusion cluster management system according to claim 1, characterized in that: The management system further comprises: The image acquisition module is used to collect the liquid level information in the infusion bottle, generate a third monitoring warning based on the liquid level information in the bottle, and synchronously push the third monitoring warning to the hospital information system and the medical mobile terminal device.

9. The real-time monitoring and AI early warning infusion cluster management system according to claim 1, characterized in that: The management system further comprises: The acquisition frequency constraint module is used to configure the acquisition frequency of the fusion sensor group according to the patient information, and after the fusion sensor group is activated, perform infusion drug monitoring based on the acquisition frequency.

10. A real-time monitoring and AI early warning infusion cluster management method, characterized in that: The method is performed by a real-time monitoring and AI early warning infusion cluster management system according to any one of claims 1 to 9, comprising: Scan the barcode on the patient's wristband and the infusion drug barcode via PDA to verify the association between patient and drug information; If the association verification is passed, after reading the patient information and infusion drug information, the fusion sensor group is activated to perform infusion drug monitoring. The fusion sensor group includes an infrared sensor, a pressure sensor, and a capacitive sensor; receiving a real-time monitoring data set, performing a data time series change analysis on the real-time monitoring data set, establishing a first monitoring warning, initializing an anomaly recognition network based on the patient information and the infusion drug information, performing anomaly recognition on the real-time monitoring data set based on the anomaly recognition network, and establishing a second monitoring warning; The first monitoring warning and the second monitoring warning are pushed to the hospital information system and medical mobile devices, and warnings are reported.

Images