Intensive care method and system based on intelligent analysis

Abstract

The invention provides an intensive care method and system based on intelligent analysis, and the method comprises the steps: collecting multiple physiological data, including electrocardiogram, blood pressure and the like, of a critical patient in real time through a sensor network; cleaning and standardizing the data; calculating the weight of each physiological data by using a dynamic weight evaluation formula; predicting an abnormal trend; calculating a risk score of the patient by using a comprehensive risk assessment model according to the weight and the abnormal trend; when the score exceeds a threshold value, the system automatically gives an early warning and sends information to the medical staff, and provides preliminary treatment suggestions at the same time. The method combines a traditional intensive care technology with a modern sensor network and a data processing technology, and aims to improve the treatment efficiency and level of critical patients.

Description

Technical Field

[0001] The present application relates to the field of medical monitoring, and in particular to an intensive care method and system based on intelligent analysis. Background Art

[0002] Intensive care technology, also known as intensive care medicine, originated in the 1950s. At that time, with the advancement of medical technology, various complex surgeries and treatment methods were applied, but this was followed by an increase in postoperative complications and critical conditions of patients. In order to meet this challenge, the medical community began to explore how to better manage and treat these critically ill patients. The original Intensive Care Unit (ICU) was composed of a group of doctors and nurses with rich clinical experience. They used the medical equipment and means at the time to monitor and treat critically ill patients 24 hours a day.

[0003] As time goes by, critical care technology continues to develop, and various advanced monitoring equipment, treatment methods and drugs are widely used in clinical practice, greatly improving the success rate of critically ill patients. The core of critical care technology lies in the real-time monitoring of patients' vital signs. Modern I CU is equipped with various advanced monitoring equipment, such as electrocardiographs, ventilators, blood oxygen saturation monitors, blood pressure monitors, etc. These devices can monitor patients' heart rate, respiration, blood oxygen saturation, blood pressure and other vital signs in real time, providing doctors with accurate data support so that they can detect changes in the condition in a timely manner and take corresponding treatment measures.

[0004] However, traditional intensive care still has limitations. Although critical care medicine already has advanced monitoring systems and a variety of non-invasive and invasive treatments to provide timely interventions for patients, it is unproven that the integration of such systems represents the next step in bedside care. In addition, it is challenging to identify the root cause of clinical deterioration from the multitude of differential diagnoses, and occult features or coexisting conditions in the early stages of the disease can mask the primary problem, thereby delaying appropriate and timely management. Summary of the invention

[0005] The purpose of the present application is to overcome the defects in the above-mentioned prior art and provide a cable joint partial discharge monitoring method, system, device and storage medium.

[0006] The present application provides a critical care method based on intelligent analysis, comprising:

[0007] The physiological data of critically ill patients are collected in real time through a sensor network, including electrocardiogram, blood pressure, blood oxygen saturation, heart rate, respiratory rate, and body temperature;

[0008] Cleaning the collected physiological data, removing noise and outliers, and performing standardization processing;

[0009] Calculate the dynamic weight of each physiological data using the physiological data dynamic weight evaluation formula;

[0010] Use the abnormal trend prediction index formula to predict the abnormal trend of each physiological data;

[0011] calculating a comprehensive risk score for the patient using a comprehensive risk assessment model based on the dynamic weights and the abnormal trends;

[0012] The warning is triggered based on the comprehensive risk score. When the score exceeds the preset threshold, the system automatically sends a warning message to medical staff and provides preliminary treatment recommendations.

[0013] Optionally, the physiological data prediction value x^i+1 is calculated by using a time series prediction algorithm such as a long short-term memory network (LSTM).

[0014] Optionally, the sensor network uses wireless transmission to transmit the collected physiological data in real time to a central processing unit for processing and analysis.

[0015] Optionally, the data cleaning step includes removing outliers caused by equipment failure or improper operation, and removing noise by filtering, smoothing, etc.

[0016] Optionally, the dynamic weight evaluation formula is dynamically adjusted according to the volatility and importance of each physiological data to ensure the accuracy and reliability of the evaluation result.

[0017] The present application also provides an intensive care system based on intelligent analysis, comprising:

[0018] An acquisition module collects physiological data of critically ill patients in real time through a sensor network, wherein the physiological data includes electrocardiogram, blood pressure, blood oxygen saturation, heart rate, respiratory rate, and body temperature;

[0019] A preprocessing module cleans the collected physiological data, removes noise and outliers, and performs standardization processing;

[0020] A calculation module, using a physiological data dynamic weight evaluation formula to calculate the dynamic weight of each physiological data; using an abnormal trend prediction index formula to predict the abnormal trend of each physiological data; and using a comprehensive risk assessment model to calculate the patient's comprehensive risk score based on the dynamic weight and the abnormal trend;

[0021] The early warning module triggers an early warning based on the comprehensive risk score. When the score exceeds the preset threshold, the system automatically sends an early warning message to medical staff and provides preliminary treatment recommendations.

[0022] Optionally, the physiological data prediction value x^i+1 is calculated by using a time series prediction algorithm such as a long short-term memory network.

[0023] Optionally, the sensor network uses wireless transmission to transmit the collected physiological data in real time to a central processing unit for processing and analysis.

[0024] Optionally, the data cleaning step includes removing outliers caused by equipment failure or improper operation, and removing noise by filtering, smoothing, etc.

[0025] Optionally, the dynamic weight evaluation formula is dynamically adjusted according to the volatility and importance of each physiological data to ensure the accuracy and reliability of the evaluation result.

[0026] The beneficial effects of this application are:

[0027] The present application provides a critical care method based on intelligent analysis, including: real-time collection of physiological data of critically ill patients through a sensor network, the physiological data including electrocardiogram, blood pressure, blood oxygen saturation, heart rate, respiratory rate, and body temperature; cleaning the collected physiological data, removing noise and outliers, and performing standardization processing; using the physiological data dynamic weight evaluation formula to calculate the dynamic weight of each physiological data; using the abnormal trend prediction index formula to predict the abnormal trend of each physiological data; using the comprehensive risk assessment model according to the dynamic weight and the abnormal trend to calculate the patient's comprehensive risk score; triggering an early warning according to the comprehensive risk score, when the score exceeds the preset threshold, the system automatically sends an early warning message to the medical staff and provides preliminary treatment recommendations. The critical care method based on intelligent analysis of the present application is based on traditional critical care technology, combined with modern sensor networks and data processing technology, to improve the level of treatment of critically ill patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is a schematic diagram of the critical care process based on intelligent analysis in this application;

[0029] Figure 2 This is a schematic diagram of the CRAM model application process in this application;

[0030] Figure 3 Schematic diagram of the intensive care system based on intelligent analysis in this application. DETAILED DESCRIPTION

[0031] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that various forms of implementing the present disclosure should not be limited by the embodiments set forth herein. On the contrary, the embodiments are provided to more thoroughly understand the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0032] Please refer to Figure 1 As shown, the present application provides a critical care method based on intelligent analysis, comprising:

[0033] S101. Collect physiological data of critically ill patients in real time through a sensor network, wherein the physiological data includes electrocardiogram, blood pressure, blood oxygen saturation, heart rate, respiratory rate, and body temperature.

[0034] Electrocardiogram (ECG) sensors are attached to electrodes on the patient's chest and limbs to capture the heart's electrical activity. These electrodes are connected to an electrocardiograph via wires, which records and transmits ECG data in real time.

[0035] Blood pressure sensors use an arterial cannula and a pressure transducer, or a non-invasive blood pressure monitoring device such as a blood pressure cuff to measure blood pressure at regular intervals. Non-invasive devices measure blood pressure by automatically inflating and deflating, while the arterial cannula provides continuous pressure monitoring.

[0036] An oximetry sensor is a probe that clips onto a patient's finger or toe and uses spectroscopy to measure the oxygen content in the blood.

[0037] A heart rate sensor is used in conjunction with an electrocardiogram sensor, or by itself with a heart rate monitor strap, to measure heart rate by detecting the electrical signals or mechanical vibrations generated by your heartbeat.

[0038] Respiratory rate sensors use chest or abdominal strap sensors to detect chest or abdominal movement during breathing, or respiratory monitoring devices to measure respiratory airflow.

[0039] A temperature sensor is a thermometer placed under the patient's armpit, in the mouth or rectum, or a skin temperature sensor is used to continuously monitor body temperature.

[0040] Each sensor is connected to a data acquisition unit (such as a monitor), which is responsible for receiving and initially processing the sensor data.

[0041] The collected data is transmitted to the central monitoring system via wired or wireless means. Wired connections use standard medical interfaces (such as USB, RS-232, etc.), while wireless connections use Bluetooth, Wi-Fi, or dedicated medical wireless protocols. To ensure data compatibility and accuracy, sensor data follows specific medical data formats and standards (such as HL7, DICOM, etc.).

[0042] S102: Clean the collected physiological data, remove noise and outliers, and perform standardization.

[0043] In the critical care environment, the accuracy and reliability of physiological data are crucial for the treatment and care of patients. Therefore, cleaning, removing noise and outliers, and standardizing the collected physiological data are important steps in data preprocessing.

[0044] Data cleaning is the first step in data preprocessing, which aims to identify and correct errors, omissions, and inconsistencies in the data. For physiological data, data cleaning includes the following aspects:

[0045] Make sure that each data point has a corresponding timestamp and that there are no missing data points. For missing data, fill in the gaps through interpolation, regression, or other methods.

[0046] Check the data for obvious errors, such as unreasonable ranges (such as heart rate over 200 beats / min or blood pressure below 0 mmHg), and try to correct these errors. This is done by referring to other data sources or the patient's clinical condition.

[0047] During the data collection process, duplicate data may be generated due to equipment failure or operational errors. These duplicate data are identified and deleted to avoid interference with subsequent analysis.

[0048] Noise in physiological data is caused by equipment accuracy limitations, environmental factors, or physiological changes of the patient. Outliers are caused by equipment failure, operating errors, or sudden physiological events of the patient. Removing noise and outliers improves the accuracy and reliability of data.

[0049] For data containing noise, smoothing techniques such as moving average, exponential smoothing, etc. are used to reduce the effect of noise. These techniques smooth the data by calculating the average or weighted average of adjacent data points, thereby removing noise.

[0050] Outliers are significantly different from other data points in the data set. Use statistical methods (such as z-score, IQR, etc.) or machine learning algorithms (such as clustering, classification, etc.) to detect outliers. For detected outliers, choose to delete, replace, or retain them according to the specific situation. If the outliers are caused by sudden physiological events of the patient, then retaining these outliers will help doctors better understand the patient's condition.

[0051] Standardization is the process of converting data into a unified format and unit for subsequent analysis and comparison. For physiological data, standardization includes the following aspects:

[0052] Data format conversion: Convert data collected from different sources or different devices into a unified format and unit. For example, convert blood pressure data collected by different devices into a unified unit such as mmHg or kPa.

[0053] Data normalization: Convert the value range of data to a specific range (such as 0-1 or -1 to 1) for subsequent analysis and comparison. Normalization eliminates the dimensional differences between different data and makes the data more comparable.

[0054] Data standardization: Convert the data into a standard normal distribution with zero mean and unit variance. This helps reduce the correlation between data and improve the accuracy of data analysis.

[0055] S103: Calculate the dynamic weight of each physiological data using the physiological data dynamic weight evaluation formula.

[0056] In an intensive care environment, the patient's physiological data is an important basis for evaluating their health status and formulating treatment plans. However, different physiological data have different importance in different time periods. In order to more accurately reflect the patient's current condition, this application introduces a physiological data dynamic weight evaluation formula to calculate the dynamic weight of each physiological data.

[0057] The physiological data dynamic weight evaluation formula is:

[0058]

[0059] Among them, DWP i is the dynamic weight of the ith physiological data, Var(x i ) is the variance of the ith physiological data, Importance i is the importance coefficient of the ith physiological data, Δx i is the change of the ith physiological data in the most recent time window, and Max(Δx) is the maximum value of the change of all physiological data in the most recent time window.

[0060] For each physiological data, calculate its variance within a certain period of time. This is achieved through the variance calculation function in statistical software or programming language. According to clinical experience or expert opinion, set an importance coefficient for each physiological data. This coefficient is determined based on factors such as the reliability of the data and the degree of indication of the patient's health status.

[0061] For each physiological data, calculate its change in the most recent time window. This can be achieved by calculating the difference between adjacent data points or using other change calculation methods. Find the maximum value among all physiological data changes for normalization.

[0062] Substitute the above calculation results into the physiological data dynamic weight evaluation formula to calculate the dynamic weight of each physiological data.

[0063] S104. Predict the abnormal trend of each physiological data using an abnormal trend prediction index formula.

[0064] In order to provide detailed data about the Abnormal Trend Prediction Index formula (ATPI) and its related contents, this application explains the various components in the formula and provides some hypothetical example data to illustrate how to calculate ATPI.

[0065] The abnormal trend prediction index formula is:

[0066]

[0067] Among them, ATPI i is the abnormal trend prediction index of the ith physiological data, x i is the current value of the ith physiological data, μ i and σ i are the historical mean and standard deviation of the ith physiological data, x i-1 is the previous value of the ith physiological data, Δt is the time interval, is the predicted value of the i-th physiological data.

[0068] The physiological data predicted value Calculated using time series prediction algorithms such as long short-term memory networks.

[0069] Long Short-Term Memory (LSTM) is a special type of Recurrent Neural Network (RNN) that is particularly suitable for processing and predicting time series data. LSTM addresses the shortcomings of traditional RNN in the long-term dependency problem by introducing a gating mechanism (input gate, forget gate, and output gate). In addition to LSTM, there are other time series prediction algorithms to choose from, such as ARIMA (Autoregressive Integrated Moving Average), GRU (Gated Recurrent Unit), Prophet (a time series prediction tool developed by Facebook), etc.

[0070] The prediction model is trained using historical physiological data. During the training process, the data is divided into a training set and a validation set (sometimes a test set). The training set is used to train the model so that it can learn the underlying laws and patterns in the data. The validation set is used to evaluate the performance of the model so that parameter adjustments and model optimization can be made during the training process. The training process involves multiple iterations until the performance of the model on the validation set reaches a stable or optimal state.

[0071] After training is completed, the model is evaluated to determine its prediction performance. Evaluation indicators include prediction accuracy (such as mean square error MSE, mean absolute error MAE, etc.), prediction stability, model complexity, etc. By comparing the evaluation results of different models, the model with the best performance can be selected for subsequent application.

[0072] Finally, the trained model is used to calculate the predicted values ​​of the physiological data.

[0073] S105. Calculate the comprehensive risk score of the patient using a comprehensive risk assessment model according to the dynamic weight and the abnormal trend.

[0074] CRAM model definition: The CRAM model is a weighted sum formula used to calculate the patient's comprehensive risk score. The specific form of the model is as follows:

[0075] CRAM=Σ(DWP i ×ATPI i )(i=1~n)

[0076] Among them, CRAM is a comprehensive risk score, which is a numerical value that represents the patient's overall health risk or disease risk. Σ is a summation symbol, which represents the weighted sum of all physiological data, and n is the number of physiological data.

[0077] DWP i is the dynamic weight of the ith physiological data, which is a value between 0 and 1, indicating the importance of the physiological data in the comprehensive risk assessment. The dynamic weight can be dynamically adjusted according to the patient's specific situation, disease type, volatility of physiological data, and other factors.

[0078] ATPI i It is the abnormal trend prediction index of the ith physiological data, which is a numerical value that indicates the abnormal degree or trend of the current value of the physiological data compared with the historical value. ATPI is usually calculated by time series prediction algorithms (such as long short-term memory network LSTM), reflecting the future change trend of physiological data.

[0079] n is the number of physiological data, which represents the total number of physiological indicators monitored by the patient.

[0080] like Figure 2 As shown in the figure, the CRAM model application process is:

[0081] S201. Collect multiple physiological data of the patient, such as heart rate, blood pressure, body temperature, etc. These data should contain sufficient historical records to enable time series analysis and prediction.

[0082] S202: Determine the dynamic weight of each physiological data according to the patient's specific condition and disease type, which may involve the participation of medical experts to ensure the rationality and accuracy of the weight.

[0083] S203 uses a time series prediction algorithm (such as LSTM) to calculate the ATPI value of each physiological data. This usually involves steps such as data preprocessing, model training, and prediction value calculation.

[0084] S204 substitutes the DWP and ATPI of each physiological data into the CRAM model formula, performs weighted summation, and obtains the comprehensive risk score of the patient.

[0085] The patient's health status or disease risk is assessed based on the size of the CRAM value. The higher the CRAM value, the greater the patient's health risk or disease risk.

[0086] The CRAM value is provided to doctors, patients or health management agencies as decision support information to guide subsequent health management, disease treatment or preventive measures.

[0087] S106. Trigger an early warning based on the comprehensive risk score. When the score exceeds a preset threshold, the system automatically sends an early warning message to medical staff and provides preliminary treatment recommendations.

[0088] The system calculates a comprehensive risk score based on the collected patient physiological data through a comprehensive risk assessment model (such as CRAM).

[0089] This comprehensive risk score is compared with a preset threshold, which is derived from medical expert advice, clinical experience or statistical analysis to determine whether the patient's health condition is at risk or high risk. The setting of the threshold should take into account factors such as the severity of the disease, the patient's age, gender, and health condition.

[0090] Based on the comparison results of the comprehensive risk score and the preset threshold, an early warning message will be sent to the medical staff responsible for monitoring the patient, such as the attending physician, nurse or medical team.

[0091] Warning information usually includes key information such as the patient's name, ID, comprehensive risk score, physiological data exceeding the threshold and its abnormal trend, warning time, etc. This information helps medical staff quickly understand the patient's health status and take corresponding countermeasures.

[0092] Early warning information can be sent in a variety of ways, such as text messages, emails, instant messaging software (such as WeChat, DingTalk), internal notifications of medical information systems, etc. The choice of which delivery method depends on the actual situation of the hospital or medical institution and the preferences of medical staff.

[0093] like Figure 3As shown, the present application also provides an intensive care system based on intelligent analysis, comprising:

[0094] Acquisition module 301, collects physiological data of critically ill patients in real time through a sensor network, the physiological data including electrocardiogram, blood pressure, blood oxygen saturation, heart rate, respiratory rate, and body temperature;

[0095] The preprocessing module 302 cleans the collected physiological data, removes noise and abnormal values, and performs standardization processing;

[0096] The calculation module 303 calculates the dynamic weight of each physiological data using the physiological data dynamic weight evaluation formula; predicts the abnormal trend of each physiological data using the abnormal trend prediction index formula; and calculates the comprehensive risk score of the patient using the comprehensive risk assessment model according to the dynamic weight and the abnormal trend;

[0097] The early warning module 304 triggers an early warning based on the comprehensive risk score. When the score exceeds a preset threshold, the system automatically sends an early warning message to medical staff and provides preliminary treatment suggestions.

[0098] Furthermore, the physiological data prediction value x^i+1 is calculated by using a time series prediction algorithm such as a long short-term memory network.

[0099] Furthermore, the physiological data dynamic weight evaluation formula is:

[0100]

[0101] Wherein, DWP_i is the dynamic weight of the i-th physiological data, Var(x_i) is the variance of the i-th physiological data, Importance_i is the importance coefficient of the i-th physiological data, Δx_i is the change of the i-th physiological data in the most recent time window, and Max(Δx) is the maximum value of the change of all physiological data in the most recent time window.

[0102] Furthermore, the abnormal trend prediction index formula is:

[0103]

[0104] Among them, ATP I_i is the abnormal trend prediction index of the i-th physiological data, x_i is the current value of the i-th physiological data, μ_i and σ_i are the historical mean and standard deviation of the i-th physiological data, respectively, x_{i-1} is the previous value of the i-th physiological data, Δt is the time interval, and x^i+1 is the predicted value of the i-th physiological data.

[0105] Furthermore, the CRAM model is:

[0106]

[0107] Where CRAM is the comprehensive risk score and n is the number of physiological data.

[0108] Furthermore, the sensor network uses wireless transmission to transmit the collected physiological data in real time to a central processing unit for processing and analysis.

[0109] Furthermore, the data cleaning step includes removing outliers caused by equipment failure or improper operation, and removing noise by filtering, smoothing and other methods.

[0110] Furthermore, the dynamic weight evaluation formula is dynamically adjusted according to the volatility and importance of each physiological data to ensure the accuracy and reliability of the evaluation result.

[0111] Although the present invention is disclosed as above by the embodiments, it is not intended to limit the protection scope of the present invention. Any changes and modifications made by any technician familiar with the technology without departing from the concept and scope of the present invention should fall within the protection scope of the present invention.

Claims

1. A critical care method based on intelligent analysis, characterized in that: include: The physiological data of critically ill patients are collected in real time through a sensor network, including electrocardiogram, blood pressure, blood oxygen saturation, heart rate, respiratory rate, and body temperature; Cleaning the collected physiological data, removing noise and outliers, and performing standardization processing; Calculate the dynamic weight of each physiological data using the physiological data dynamic weight evaluation formula; Use the abnormal trend prediction index formula to predict the abnormal trend of each physiological data; calculating a comprehensive risk score for the patient using a comprehensive risk assessment model based on the dynamic weights and the abnormal trends; The warning is triggered based on the comprehensive risk score. When the score exceeds the preset threshold, the system automatically sends a warning message to medical staff and provides preliminary treatment recommendations.

2. The intensive care method based on intelligent analysis according to claim 1, characterized in that: The physiological data prediction value x^i+1 is calculated by using a time series prediction algorithm such as a long short-term memory network (LSTM).

3. The intensive care method based on intelligent analysis according to claim 1, characterized in that: The sensor network uses wireless transmission to transmit the collected physiological data to the central processing unit in real time for processing and analysis.

4. The intensive care method based on intelligent analysis according to claim 1, characterized in that: The data cleaning step includes removing outliers caused by equipment failure or improper operation, and removing noise by filtering, smoothing and other methods.

5. The intensive care method based on intelligent analysis according to claim 1, characterized in that: The dynamic weight evaluation formula is dynamically adjusted according to the volatility and importance of each physiological data to ensure the accuracy and reliability of the evaluation results.

6. An intensive care system based on intelligent analysis, characterized in that: include: An acquisition module collects physiological data of critically ill patients in real time through a sensor network, wherein the physiological data includes electrocardiogram, blood pressure, blood oxygen saturation, heart rate, respiratory rate, and body temperature; A preprocessing module cleans the collected physiological data, removes noise and outliers, and performs standardization processing; A calculation module, which uses a physiological data dynamic weight evaluation formula to calculate the dynamic weight of each physiological data; Use the abnormal trend prediction index formula to predict the abnormal trend of each physiological data; calculating a comprehensive risk score for the patient using a comprehensive risk assessment model based on the dynamic weights and the abnormal trends; The early warning module triggers an early warning based on the comprehensive risk score. When the score exceeds the preset threshold, the system automatically sends an early warning message to medical staff and provides preliminary treatment recommendations.

7. The intensive care system based on intelligent analysis according to claim 6, characterized in that: The physiological data prediction value x^i+1 is calculated by using a time series prediction algorithm such as a long short-term memory network.

8. The intensive care system based on intelligent analysis according to claim 1, characterized in that: The sensor network uses wireless transmission to transmit the collected physiological data to the central processing unit in real time for processing and analysis.

9. The intensive care system based on intelligent analysis according to claim 1, characterized in that: The data cleaning step includes removing outliers caused by equipment failure or improper operation, and removing noise by filtering, smoothing and other methods.

10. The intensive care system based on intelligent analysis according to claim 1, characterized in that: The dynamic weight evaluation formula is dynamically adjusted according to the volatility and importance of each physiological data to ensure the accuracy and reliability of the evaluation results.

Images