Early warning system for critical care in internal medicine based on multi-parameter monitoring

Abstract

This invention provides an early warning system for critical care nursing in internal medicine based on multi-parameter monitoring, belonging to the field of early warning system technology. It includes: a vital sign monitoring module for real-time collection of the patient's vital sign factors, taking the average value of the amplitude of each real-time vital sign factor as a benchmark reference value; a similarity early warning analysis module for establishing a mathematical model of the patient's self-monitoring, calculating the early warning similarity index corresponding to each real-time vital sign factor; calculating a reliability index by comparing the total short-term change value with the overall average change value of each real-time vital sign factor; and an early warning decision module for establishing two mechanisms to determine early warnings. This invention extracts various parameters from the patient's past multiple disease occurrence points and calculates the average value, and collects the median value of the disease occurrence point with the largest interval as a reference benchmark value, forming a unique benchmark reference value; thus, the abnormality judgment is entirely based on individual physical differences and historical response patterns, solving the problem of misjudgment due to uniform standards caused by physical differences.

Description

Technical Field

[0001] This invention relates to the field of early warning system technology, specifically to an early warning system for critical care nursing in internal medicine based on multi-parameter monitoring. Background Technology

[0002] Pneumonia, a respiratory disease caused by lung infection, is one of the most common severe illnesses in internal medicine. Patients with pneumonia exhibit abnormal changes in parameters such as respiratory rate (RESP), oxygen saturation (SpO2), peak airway pressure (Ppeak), arterial partial pressure of oxygen (PaO2), end-tidal carbon dioxide (EtCO2), and body temperature. These abnormalities indicate a potential risk of disease deterioration and not only affect subjective symptoms such as shortness of breath or fatigue, but more importantly, they may foreshadow life-threatening conditions. Furthermore, due to individual patient differences, the manifestations of these parameters during intensive care vary among patients. Considering individual patient characteristics and historical disease data as early warning factors can further improve the reliability of early warning systems.

[0003] In the prior art, CN118352080A discloses a risk early warning system and method for patients with post-ICU syndrome. This technology includes: screening risk early warning indicators for critically ill patients with post-ICU syndrome; constructing an early risk prediction model for post-ICU syndrome using a machine learning method based on a backpropagation neural network; constructing a comprehensive care model for patients with post-ICU syndrome and generating specific intervention plans for patients; and constructing a system for risk early warning and comprehensive management of patients with post-ICU syndrome. This approach enables medical staff to promptly identify potential or existing PICS risks in patients during nursing services, identify risk factors for post-ICU syndrome in advance, and guide the implementation of corresponding nursing measures, eliminating potential risks at the outset.

[0004] However, as can be seen from the above, existing technologies rely on static threshold alarms with a single parameter, lacking adaptability to individual patient differences and failing to integrate personal historical disease data to calculate the overall average change, thus making it impossible to personalize the abnormal judgment criteria. At the same time, there is a lack of a mutual exclusion and contradiction value verification mechanism, resulting in low data accuracy. For fluctuations in some physical parameters caused by non-pathological factors, it is impossible to combine other data for judgment and processing, thus leading to a high false alarm rate. Furthermore, the lack of a real-time vital sign factor credibility index calculation process results in insufficient sensitivity and a high false alarm rate in its early warning monitoring.

[0005] The information disclosed in the background section is only for enhancing the understanding of the background of the present invention, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide an early warning system for critical care medicine based on multi-parameter monitoring, in order to solve the problems mentioned in the background art. This invention extracts various parameters from multiple past illness episodes of the patient and calculates their average values ​​to form a unique benchmark reference value. In the similarity early warning analysis module, the real-time calculated early warning similarity index dynamically compares the current short-term change value with the patient's overall average change value, rather than using a fixed threshold. This allows anomaly judgment to be based entirely on individual physical differences and historical response patterns. The higher the deviation of the parameter change from the patient's past illness characteristics, the more sensitively the system captures personalized anomalies, effectively solving the core problem of misjudgment or missed judgment due to uniform standards caused by physical differences.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] The medical intensive care early warning system based on multi-parameter monitoring includes:

[0009] The vital signs monitoring module is used to collect the patient's real-time vital signs, including peak airway pressure, arterial blood oxygen partial pressure, end-tidal carbon dioxide, and body temperature. It eliminates contradictory values ​​among the real-time vital signs and collects the overall average change of each real-time vital sign at the patient's historical disease onset points. It extracts the sequence of real-time vital signs changes of the monitored patient at multiple disease onset points from the historical database and takes the average value of the amplitude of each real-time vital sign as the benchmark reference value.

[0010] The similarity early warning analysis module establishes a patient's self-monitoring mathematical model, extracts short-term changes of each real-time vital sign within a fixed time interval of K seconds, and substitutes the collected short-term change values ​​of each real-time vital sign into the self-monitoring mathematical model to calculate the early warning similarity index corresponding to each real-time vital sign. If the value of the warning similarity index is greater than or equal to 1, it is judged as an abnormal situation;

[0011] The credibility verification module collects the warning similarity index for each real-time vital sign factor. When more than one parameter becomes abnormal, it simultaneously acquires the short-term change value of each real-time vital sign factor within the corresponding abnormal time period, calculates the total short-term change value, and calculates the credibility index by comparing the total short-term change value with the overall mean change value of each real-time vital sign factor. ;

[0012] The early warning decision module establishes two mechanisms to determine early warnings, and the two mechanisms independently classify early warnings into low-level and high-level alarms. The median value of the adjacent historical disease occurrence nodes with the largest interval is collected and calibrated as the reference baseline value without abnormalities, and the results are output.

[0013] Furthermore, the vital signs monitoring module continuously collects the patient's real-time vital signs at a frequency of once per second;

[0014] The mutually exclusive contradictory values ​​are verified in real time by the system after collecting the patient's airway peak pressure, arterial blood oxygen partial pressure, end-tidal carbon dioxide and body temperature data. During this process, the system compares whether the combination of real-time vital signs violates the mutually exclusive condition within a fixed sampling period of t seconds. If a contradictory value is detected, it is marked as invalid data and the system excludes the invalid value.

[0015] Furthermore, the autonomous monitoring mathematical model for the early warning similarity index is calculated using the following formula:

[0016] ,

[0017] in:

[0018] include , , , ,and , , , These are the warning similarity indices for peak airway pressure, arterial partial pressure of oxygen, end-tidal carbon dioxide, and body temperature, respectively.

[0019] include , , , ,and 、 、 、 These are the short-term changes in peak airway pressure, arterial oxygen partial pressure, end-tidal carbon dioxide, and body temperature within a fixed time interval (seconds);

[0020] include 、 、 、 ,and 、 、 、 These represent the overall mean changes in peak airway pressure, arterial partial pressure of oxygen, end-tidal carbon dioxide, and body temperature at historical disease onset points in the monitored patients.

[0021] Furthermore, the calculation process of the autonomous monitoring mathematical model is as follows: real-time monitoring of four real-time vital signs, calculated once every K seconds. Obtain the corresponding data from the historical database of the monitored patients. The value is substituted into the formula of the autonomous monitoring mathematical model to obtain any... This real-time vital sign factor is marked as abnormal.

[0022] Furthermore, the short-time change value is calculated using the following formula:

[0023] ,

[0024] in:

[0025] This refers to the real-time measurement of a certain vital sign of the patient at the start of a fixed time interval of K seconds.

[0026] This refers to the real-time measurement of a certain vital sign of the patient at the end of a fixed time interval K seconds; therefore... and The time interval between them is always K, which is used to define the sampling window for short-term changes.

[0027] Furthermore, the overall average change is calculated using the following formula:

[0028] ,

[0029] in, The total number of the patient's historical medical events;

[0030] The total time of each symptom onset for the patient;

[0031] Index each patient's symptom episode;

[0032] For the first Parameter measurements at the start of the event;

[0033] For the first Parameter measurements at the end of the event;

[0034] For all Sum of the absolute changes in real-time vital signs for each event;

[0035] When a new historical event is added, the total number of events will be... and the overall mean change Perform a recalculation and update process.

[0036] Furthermore, the credibility index in the credibility verification module is calculated using the following formula:

[0037] ,

[0038] in, This is a credibility index;

[0039] It is the sum of short-term changes;

[0040] This represents the sum of the overall changes and their mean.

[0041] The inflation coefficient is used to determine the credibility index. Is it greater than ,like Then it is marked as a high-confidence multi-parameter anomaly event. If so, it is judged as a low-confidence multi-parameter anomaly.

[0042] Furthermore, the formula for calculating the sum of short-term changes is as follows:

[0043] ,

[0044] in:

[0045] , , , , which are the clinical weighting coefficients for peak airway pressure, arterial oxygen partial pressure, end-tidal carbon dioxide, and body temperature, respectively;

[0046] The sum of the overall change mean The calculation formula is:

[0047] ,

[0048] Retrieve the overall mean sequence of changes in the corresponding parameters from the historical database of the monitored patients, and set the clinical weighting coefficients to meet the following conditions: .

[0049] Furthermore, the early warning decision module comprises two mechanisms:

[0050] Real-time change warning: Real-time analysis of the credibility index. When the credibility index exceeds the set threshold, a warning is triggered directly, and the level of warning is determined as low-level or high-level based on the number of abnormal situations.

[0051] Node trend warning: When the confidence index is less than or equal to a set threshold, real-time vital signs that only show abnormalities are collected and abnormal processes are marked. The periodic position of the disease occurrence node corresponding to the real-time vital signs with the same value as the abnormal process in the patient's historical disease occurrence nodes is calculated. A low-level alarm or a high-level alarm is triggered according to the periodic position of the disease occurrence node of the abnormal process. The real-time change warning mechanism and the node trend warning mechanism are carried out synchronously.

[0052] Furthermore, in the aforementioned node trend early warning mechanism, the system periodically scans the changing trends of four parameters: peak airway pressure, arterial oxygen partial pressure, end-tidal carbon dioxide, and body temperature. Using a reference value without abnormalities as a reference, if the same parameter changes in the same direction relative to the reference value without abnormalities for four consecutive statistical cycles, a low-level alarm is triggered, the current trend data is recorded, and medical staff are notified to pay attention to potential risks. If one of the four parameters changes in the same direction relative to the reference value without abnormalities for six consecutive times, or if any two parameters simultaneously change in the same direction relative to the reference value without abnormalities for four consecutive times, a low-level alarm is triggered.

[0053] Compared with the prior art, the beneficial effects of the present invention are:

[0054] This solution collects patients' personal historical disease data through a vital signs monitoring module, establishing an overall average change centered on the individual—that is, extracting various parameters from the patient's past multiple disease occurrences and calculating the average value to form a unique benchmark reference value; in the similarity early warning analysis module, the real-time calculated early warning similarity index dynamically compares the current short-term change value with the patient's own overall average change value, rather than using a fixed threshold, so that the abnormality judgment is entirely based on individual physical differences and historical response patterns; the higher the degree of parameter change deviates from the patient's past disease characteristics, the more sensitively the system captures personalized abnormalities, effectively solving the core problem of misjudgment by uniform standards due to physical differences (e.g., in patients with strong tolerance, the actual risk is high even with small parameter fluctuations) or missed judgment (e.g., in patients with sensitive physical conditions, even small fluctuations are dangerous).

[0055] The system employs a vital signs monitoring module to collect multi-parameter data in real time every second and establishes a verification mechanism to eliminate mutually exclusive and contradictory values, ensuring data accuracy. Combined with a similarity early warning analysis module, it calculates an early warning similarity index based on the overall average and short-term changes in the patient's historical disease onset points, enabling personalized dynamic modeling and anomaly detection. Furthermore, a credibility verification module calculates a credibility index to weightedly verify short-term and long-term changes when multiple parameters are abnormal, improving the confidence assessment of abnormal events. Finally, the early warning decision module's dual mechanisms provide tiered urgency responses based on the credibility index and trend continuity, significantly improving early warning accuracy, reducing false alarms and missed alarms, and enhancing system adaptability and reliability. Attached Figure Description

[0056] Figure 1This is a schematic diagram of the principle of an early warning system for critical care nursing in internal medicine based on multi-parameter monitoring.

[0057] Figure 2 This is a schematic diagram of the operation process of the medical critical care nursing early warning system based on multi-parameter monitoring according to the present invention. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0059] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0060] Example:

[0061] Please see Figures 1-2 The present invention provides the following technical solution:

[0062] The medical intensive care early warning system based on multi-parameter monitoring includes the following functional modules:

[0063] The vital signs monitoring module is the core component of the entire early warning system, specifically designed for real-time acquisition and monitoring of patients' key vital signs parameters. These real-time vital signs include peak airway pressure, arterial partial pressure of oxygen, end-tidal carbon dioxide, and body temperature. This module not only continuously collects data once per second to ensure real-time and continuous monitoring, but also incorporates an efficient verification mechanism to eliminate potentially contradictory values ​​between parameters. Specifically, after real-time data acquisition, the system automatically compares parameter combinations at a fixed sampling period to detect violations of mutual exclusion conditions; if contradictory values ​​are identified, they are immediately marked as invalid data and excluded. Simultaneously, the module is also responsible for collecting the overall mean change of each real-time vital sign parameter across the patient's historical disease progression. This process involves extracting parameter change sequences from the historical database for the monitored patient across multiple disease progression points and generating a baseline reference value by averaging the changes in each parameter.

[0064] The mutually exclusive contradictory values ​​are verified in real time by the system after collecting the patient's airway peak pressure, arterial blood oxygen partial pressure, end-tidal carbon dioxide and body temperature data. During this process, the system compares whether the combination of real-time vital signs violates the mutually exclusive condition within a fixed sampling period of t seconds. If a contradictory value is detected, it is marked as invalid data and the system excludes the invalid value.

[0065] The mutually exclusive contradictory values ​​include:

[0066] Incompatible vital signs; for example: peak airway pressure consistently >35 cmH2O, indicating severe airway obstruction or decreased lung compliance, but EtCO2 ≤25 mmHg, far below the normal 35-45 mmHg, high-pressure ventilation usually leads to obstructed CO2 expulsion, and EtCO2 should be increased; if it is low, it indicates an air leak in the ETCO2 monitoring tubing or a sampling malfunction, and the function of the expiratory valve and the calibration of the CO2 sensor should be checked immediately.

[0067] High fever and metabolic CO2 production are disconnected: When body temperature is >39°C, an increased metabolic rate should lead to increased CO2 production, but EtCO2 does not rise or may even decrease. For every 1°C increase in body temperature, CO2 production increases by approximately 10%. If EtCO2 does not rise synchronously, it suggests hyperventilation or distorted EtCO2 monitoring.

[0068] Among them, abnormal device linkage refers to an anomaly where the data of related devices violates the logical linkage patterns of medical science; for example:

[0069] Conflict between ventilator and blood gas analysis: The ventilator shows Ppeak>30 cmH2O, indicating high-pressure ventilation, but arterial blood gas PaO2≤60 mmHg, which does not improve hypoxemia. High-pressure ventilation should improve oxygenation efficiency. If PaO2 does not respond, it suggests intrapulmonary shunting (such as ARDS) or blood gas sampling error, such as air bubbles.

[0070] CO2 monitoring conflict with ventilation status: The ventilator alarms "insufficient ventilation" and the tidal volume is too low, but the EtCO2 waveform is stable and the value is normal. Low tidal volume will inevitably lead to an increase in EtCO2. If the reading is normal, it indicates that the EtCO2 monitoring is delayed or the main / bypass sampling system is blocked.

[0071] Cross-parameter time difference conflict is an abnormal type of parameter change rate that exceeds the physiological limits of the human body; for example:

[0072] The rapid change in oxygenation and CO2 clearance is contradictory: PaO2 rises sharply from 80 mmHg to 150 mmHg within 2 minutes (with pure oxygen intervention), but EtCO2 simultaneously drops from 40 mmHg to 20 mmHg. The increase in PaO2 depends on the improvement of alveolar ventilation, while the rapid decrease in EtCO2 requires synchronous changes in pulmonary blood flow. This combination exceeds the physiological regulatory speed.

[0073] Delayed response to body temperature-driven CO2: Body temperature rises by 2°C within 10 minutes, but EtCO2 remains unchanged throughout, whereas it is expected to rise by 20-30%. CO2 production increases exponentially with rising body temperature, and the lack of response indicates a malfunction in the EtCO2 monitoring system.

[0074] A discrepancy between the condition and the numerical value represents an abnormality where the pathological condition is disconnected from typical clinical manifestations; for example:

[0075] Abnormal parameter combinations in ARDS patients: A known ARDS patient with lung compliance <30 ml / cmH2O, reported Ppeak <20 cmH2O and PaO2 >100 mmHg, and not using ECMO. Low compliance necessitates high-pressure ventilation, and high PaO2 is difficult to achieve with low Ppeak, suggesting ventilator pressure sensor drift or blood gas mislabeling.

[0076] Conflict between low temperature and high metabolism: Core body temperature ≤35℃, low temperature inhibits metabolism, but EtCO2 is consistently >50 mmHg. Low temperature should reduce CO2 production. High EtCO2 requires ruling out malignant hyperthermia or interference of CO2 monitoring equipment with electric blankets.

[0077] Extreme combination alerts are for abnormality types for which there is no actual clinical record data. For example:

[0078] Contradictory combination:

[0079] Ppeak > 40 cmH2O (extreme high pressure)

[0080] PaO2 < 55 mmHg (stubborn hypoxia)

[0081] EtCO2 15 mmHg (hypocapnia)

[0082] Body temperature 37℃ (normal metabolism).

[0083] The similarity warning analysis module, as a core component of the system, focuses on establishing a patient-specific self-monitoring mathematical model. This model assesses the risk of abnormalities in real-time vital signs, including peak airway pressure, arterial oxygen partial pressure, end-tidal carbon dioxide, and body temperature. During operation, the module first establishes a fixed monitoring cycle of K seconds. Within each cycle, it captures short-term changes in each vital sign—specifically, calculating the magnitude of the change by measuring the difference between the parameter value at the end of the interval and its initial value. Subsequently, the module inputs these short-term changes into the self-monitoring mathematical model.

[0084] The model calculates a warning similarity index for each real-time vital sign by comparing short-term changes with the overall average changes across the patient's historical disease progression. This index reflects the degree of abnormality of short-term changes relative to long-term historical trends. When the calculated warning similarity index is greater than or equal to 1, the system immediately determines that the parameter is abnormal, indicating that the current change has exceeded the normal range and requires immediate attention.

[0085] The autonomous monitoring mathematical model for the early warning similarity index is calculated using the following formula:

[0086] ,

[0087] in:

[0088] include , , , ,and , , , These are the warning similarity indices for peak airway pressure, arterial partial pressure of oxygen, end-tidal carbon dioxide, and body temperature, respectively. Each real-time vital sign factor is calculated independently in this formula. The larger the value of , until it exceeds the average value of the overall change, the larger the warning similarity index, that is, the higher the similarity. A result greater than 1 indicates that the parameter is abnormal at the current moment.

[0089] include , , , ,and 、 、 、 These are the short-term changes in peak airway pressure, arterial oxygen partial pressure, end-tidal carbon dioxide, and body temperature within a fixed time interval (seconds);

[0090] include 、 、 、 ,and 、 、 、 These represent the mean overall changes in peak airway pressure, arterial partial pressure of oxygen, end-tidal carbon dioxide, and body temperature at different historical disease onset points in the monitored patients, indicating the average magnitude of change at these historical disease onset points. When the value is close to zero (e.g., when historical changes are minimal), the formula may become unstable.

[0091] The formula above represents the average variation of this parameter across a patient's historical events (such as an acute respiratory failure episode). It is calculated by extracting a sequence of parameter variation values ​​from multiple event points in a historical database and averaging their magnitudes (i.e., the mean of all |historical variation values|). This serves as a baseline reference value, reflecting the patient's individualized normal fluctuation range.

[0092] The process of collecting the overall average changes of airway peak pressure, arterial blood oxygen partial pressure, end-tidal carbon dioxide and body temperature at the historical disease onset points of the monitored patients includes: extracting the real-time vital sign factor change value sequence of the monitored patients at multiple disease onset points from the historical database, and taking the average value of the amplitude of each real-time vital sign factor as the benchmark reference value.

[0093] The calculation process of the autonomous monitoring mathematical model is as follows: real-time monitoring of four real-time vital signs, calculated once every K seconds. Obtain the corresponding data from the historical database of the monitored patients. The value is substituted into the formula of the autonomous monitoring mathematical model to obtain any... This real-time vital sign factor is marked as abnormal.

[0094] Short-time changes are calculated using the following formula:

[0095] ,

[0096] in:

[0097] Let K be the real-time measurement value of a certain real-time vital sign of the patient at the beginning of a fixed time interval of K seconds.

[0098] This refers to the real-time measurement of a certain vital sign of the patient at the end of a fixed time interval K seconds; therefore... and The time interval between them is always K, which is used to define the sampling window for short-term changes.

[0099] The above formula, through absolute value processing, eliminates the influence of the direction of parameter change (such as increase or decrease), focusing solely on the magnitude of change. This ensures a more robust comparison with the historical average change, facilitating rapid anomaly detection in real-time monitoring. In practical applications, the system automatically performs this calculation every K seconds, relying on the high-quality data stream verified in step S1.

[0100] The overall average change is calculated using the following formula:

[0101] ,

[0102] in, The system automatically collects and continuously updates the patient's historical medical record data, which represents the total number of historical medical events. The larger the number, the more accurate and reliable the early warning system.

[0103] The total time of each disease event for the patient is the period from when any real-time vital signs or factors become abnormal until they begin to return to normal in historical cases.

[0104] Index each patient's symptom episode;

[0105] For the first Parameter measurements at the start of the event; For the first Parameter measurements at the end of the event;

[0106] For all Sum of the absolute changes in real-time vital signs for each event;

[0107] When a new historical event is added, the total number of events will be... and the overall mean change Perform a recalculation and update process.

[0108] As a key component of the system, the credibility verification module aims to verify the authenticity of multiple abnormal events: First, the module continuously collects the warning similarity index of each real-time vital sign factor; when the system detects an abnormality in two or more parameters (i.e., when the warning similarity index is ≥1, it is judged as abnormal), the module immediately synchronously obtains the short-term change value of each parameter within the time period corresponding to these abnormalities. These values ​​are calculated based on the difference between the end and the beginning of a fixed time interval of K seconds.

[0109] Subsequently, the module summarizes these short-term change values, applies a clinical weighting coefficient to the change of each parameter to reflect the medical importance of different parameters, and then performs a weighted summation to calculate the total short-term change value. Then, by comparing this total value with the sum of the overall change mean of each parameter retrieved from the patient's historical database as a long-term benchmark reference value, a confidence index is calculated to quantify the overall confidence level of abnormal events.

[0110] The credibility index in the credibility verification module is calculated using the following formula:

[0111] ,

[0112] in, This is a credibility index;

[0113] It is the sum of short-term changes;

[0114] This represents the sum of the overall changes and their mean.

[0115] The inflation coefficient is used to determine the credibility index. Is it greater than ,like Then it is marked as a high-confidence multi-parameter anomaly event. If so, it is judged as a low-confidence multi-parameter anomaly.

[0116] The formula for calculating the sum of short-term changes is:

[0117] ,

[0118] in:

[0119] , , , , which are the clinical weighting coefficients for peak airway pressure, arterial oxygen partial pressure, end-tidal carbon dioxide, and body temperature, respectively;

[0120] The sum of the overall change mean The calculation formula is:

[0121] ,

[0122] Retrieve the overall mean sequence of changes in the corresponding parameters from the historical database of the monitored patients, and set the clinical weighting coefficients to meet the following conditions: .

[0123] The early warning decision module establishes two mechanisms to determine early warnings, and the two mechanisms independently classify early warnings into low-level and high-level alerts.

[0124] The early warning decision module has two mechanisms:

[0125] Real-time change warning: Real-time analysis of the credibility index. When the credibility index exceeds the set threshold, a warning is triggered directly, and the level of warning is determined as low-level or high-level based on the number of abnormal situations.

[0126] Node trend warning: When the confidence index is less than or equal to a set threshold, real-time vital signs that only show abnormalities are collected and abnormal processes are marked. The periodic position of the disease occurrence node corresponding to the real-time vital signs with the same value as the abnormal process in the patient's historical disease occurrence nodes is calculated. A low-level alarm or a high-level alarm is triggered according to the periodic position of the disease occurrence node of the abnormal process. The real-time change warning mechanism and the node trend warning mechanism are carried out synchronously.

[0127] Specifically, when the credibility index is less than or equal to a preset threshold, the system automatically activates a node trend early warning mechanism. First, this mechanism accurately collects real-time vital signs that indicate abnormalities, such as peak airway pressure, arterial oxygen partial pressure, end-tidal carbon dioxide, or body temperature, and marks them as abnormal processes. Then, the system performs in-depth calculations of the corresponding positions of these abnormal processes within the patient's historical disease occurrence nodes: specifically, it retrieves the patient's past disease event records from the historical database, identifies instances with the same parameter type and value as the current abnormal process, thereby determining the relative position of that parameter within the disease occurrence node cycle, judging whether it is in the early, middle, or late stage. Based on this positional analysis within the cycle, the system intelligently triggers:

[0128] Low-level alerts indicate potential risks at early locations; high-level alerts, such as late-stage alerts, mark urgent events, ensuring that risk responses match individual patient history patterns; the entire process is synchronized in real time with an immediate change warning mechanism, with the two mechanisms operating independently but monitoring in parallel to seamlessly cover different types of abnormal scenarios.

[0129] In the aforementioned node trend early warning mechanism, the system periodically scans the changing trends of four parameters: peak airway pressure, arterial partial pressure of oxygen, end-tidal carbon dioxide, and body temperature. Compared to a reference value without abnormalities, if the same parameter is found to change in the same direction for four consecutive statistical cycles, a low-level alarm is triggered, the current trend data is recorded, and medical staff are notified to pay attention to potential risks. If one of the four parameters changes in the same direction for six consecutive times, or if any two parameters change in the same direction for four consecutive times, the alarm is upgraded to a high-level alarm.

[0130] In this process, the reference baseline value without abnormalities is taken from the median value of the adjacent historical disease occurrence nodes with the largest interval, that is, the middle of the longest non-disease time of the tested patient. The above-mentioned real-time vital signs at this time point are taken as the most standard safety value. Since this time point is in the healthiest period of the patient's entire nursing cycle, it is best to use this as the reference baseline value specific to the patient to reflect his best state.

[0131] Specifically, the changing trends of the four parameters are analyzed every 10 seconds, that is... , , , The four parameters can take positive or negative values. A low-level alarm is triggered when the same parameter changes in the same direction four times consecutively relative to a reference value without anomalies. , , , If any one of the parameters shows a positive value four times consecutively, or both are simultaneously negative, a low-level alarm is triggered. The system automatically records the current trend data and notifies medical staff to pay attention to potential risks. If the above parameters show six consecutive changes in the same direction, or two parameters change simultaneously four times consecutively, a high-level alarm is triggered. The node trend warning includes warnings for changes in the sum of short-term changes and warnings for changes in each independent parameter. The two mechanisms operate simultaneously. Changes in the same direction include situations where each real-time vital sign factor increases or decreases synchronously. For example, in airway peak pressure monitoring data, a high-level alarm is triggered if the airway peak pressure gradually increases six times consecutively. Or, a low-level alarm is triggered if end-tidal carbon dioxide decreases four times consecutively.

[0132] Specifically, if the total short-term change is calculated in the same direction for three consecutive times relative to a reference value without anomalies, a low-level alert is triggered. If it changes in the same direction five times consecutively, the risk increases significantly, and the alert is escalated to a high-level alert. If, during three consecutive checks of the "total short-term change" trend, the direction of change differs from the previous calculation (i.e., it is determined to be a non-uniform change), the system will reset the count of previous consecutive changes to zero and start the calculation again. Only a continuous change in one direction will accumulate and trigger an alert.

[0133] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0134] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0135] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0136] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A multi-parameter monitoring-based early warning system for critical care nursing in internal medicine, characterized in that, Including: The vital signs monitoring module is used to collect the patient's real-time vital signs, including peak airway pressure, arterial blood oxygen partial pressure, end-tidal carbon dioxide, and body temperature. It eliminates contradictory values ​​among the real-time vital signs and collects the overall average change of each real-time vital sign at the patient's historical disease onset points. It extracts the sequence of real-time vital signs changes of the monitored patient at multiple disease onset points from the historical database and takes the average value of the amplitude of each real-time vital sign as the benchmark reference value. The similarity early warning analysis module establishes a patient's self-monitoring mathematical model, extracts short-term changes of each real-time vital sign within a fixed time interval of K seconds, and substitutes the collected short-term change values ​​of each real-time vital sign into the self-monitoring mathematical model to calculate the early warning similarity index corresponding to each real-time vital sign. If the value of the warning similarity index is greater than or equal to 1, it is judged as an abnormal situation; The credibility verification module collects the warning similarity index for each real-time vital sign factor. When more than one parameter becomes abnormal, it simultaneously acquires the short-term change value of each real-time vital sign factor within the corresponding abnormal time period, calculates the total short-term change value, and calculates the credibility index by comparing the total short-term change value with the overall mean change value of each real-time vital sign factor. ; The early warning decision module establishes two mechanisms to determine early warnings, and the two mechanisms independently classify early warnings into low-level and high-level alarms. The median value of the adjacent historical disease occurrence nodes with the largest interval is collected and calibrated as the reference baseline value without abnormalities, and the results are output. The early warning decision module has two mechanisms: Real-time change early warning mechanism: Real-time analysis of the credibility index. When the credibility index exceeds the set threshold, an early warning is triggered directly, and the alarm is classified as a low-level alarm or a high-level alarm based on the number of abnormal situations. Node trend early warning mechanism: When the confidence index is less than or equal to a set threshold, real-time vital signs that only show abnormalities are collected and abnormal processes are marked. The periodic position of the disease occurrence node corresponding to the real-time vital signs with the same value as the abnormal process in the patient's historical disease occurrence nodes is calculated. A low-level alarm or a high-level alarm is triggered according to the periodic position of the disease occurrence node of the abnormal process. The real-time change early warning mechanism and the node trend early warning mechanism are carried out synchronously.

2. The medical intensive care early warning system based on multi-parameter monitoring according to claim 1, characterized in that, The vital signs monitoring module continuously collects the patient's real-time vital signs at a frequency of once per second. The mutually exclusive contradictory values ​​are verified in real time by the system after collecting the patient's airway peak pressure, arterial blood oxygen partial pressure, end-tidal carbon dioxide and body temperature data. During this process, the system compares whether the combination of real-time vital signs violates the mutually exclusive condition within a fixed sampling period of t seconds. If a contradictory value is detected, it is marked as invalid data and the system excludes the invalid value.

3. The medical intensive care early warning system based on multi-parameter monitoring according to claim 2, characterized in that, The autonomous monitoring mathematical model for the early warning similarity index is calculated using the following formula: ； in: include , , , ,and , , , These are the warning similarity indices for peak airway pressure, arterial partial pressure of oxygen, end-tidal carbon dioxide, and body temperature, respectively. include , , , ,and 、 、 、 These are the short-term changes in peak airway pressure, arterial oxygen partial pressure, end-tidal carbon dioxide, and body temperature within a fixed time interval (seconds); include 、 、 、 ,and 、 、 、 These represent the overall mean changes in peak airway pressure, arterial partial pressure of oxygen, end-tidal carbon dioxide, and body temperature at historical disease onset points in the monitored patients.

4. The medical intensive care early warning system based on multi-parameter monitoring according to claim 3, characterized in that: The calculation process of the autonomous monitoring mathematical model is as follows: real-time monitoring of four real-time vital signs, calculated once every K seconds. Obtain the corresponding data from the historical database of the monitored patients. The value is substituted into the formula of the autonomous monitoring mathematical model to obtain any... This real-time vital sign factor is marked as abnormal.

5. The medical intensive care early warning system based on multi-parameter monitoring according to claim 3, characterized in that: Short-time changes are calculated using the following formula: ； in: This refers to the real-time measurement of a certain vital sign of the patient at the start of a fixed time interval of K seconds. This refers to the real-time measurement of a certain vital sign of the patient at the end of a fixed time interval K seconds; therefore... and The time interval between them is always K, which is used to define the sampling window for short-term changes.

6. The medical intensive care early warning system based on multi-parameter monitoring according to claim 5, characterized in that: The overall average change is calculated using the following formula: ； in, The total number of the patient's historical medical events; The total time of each symptom onset for the patient; Index each patient's symptom episode; For the first Parameter measurements at the start of the event; For the first Parameter measurements at the end of the event; For all Sum of the absolute changes in real-time vital signs for each event; When a new historical event is added, the total number of events will be... and the overall mean change Perform a recalculation and update process.

7. The medical intensive care early warning system based on multi-parameter monitoring according to claim 3, characterized in that: The credibility index in the credibility verification module is calculated using the following formula: ； in, This is a credibility index; It is the sum of short-term changes; This represents the sum of the overall changes and their mean. The inflation coefficient is used to determine the credibility index. Is it greater than ,like Then it is marked as a high-confidence multi-parameter anomaly event. If so, it is judged as a low-confidence multi-parameter anomaly.

8. The medical intensive care early warning system based on multi-parameter monitoring according to claim 7, characterized in that: The formula for calculating the sum of short-term changes is: ； in: , , , , which are the clinical weighting coefficients for peak airway pressure, arterial oxygen partial pressure, end-tidal carbon dioxide, and body temperature, respectively; The sum of the overall change mean The calculation formula is: ； Retrieve the overall mean sequence of changes in the corresponding parameters from the historical database of the monitored patients, and set the clinical weighting coefficients to meet the following conditions: .

9. The medical intensive care early warning system based on multi-parameter monitoring according to claim 7, characterized in that: In the aforementioned node trend early warning mechanism, the system periodically scans the changing trends of four parameters: airway peak pressure, arterial blood oxygen partial pressure, end-tidal carbon dioxide, and body temperature. The reference value without abnormalities is used as a reference. If the same parameter changes in the same direction relative to the reference value without abnormalities for four consecutive statistical cycles, a low-level alarm is triggered, the current trend data is recorded, and medical staff are notified to pay attention to potential risks. If one of the four parameters changes in the same direction for six consecutive times relative to a reference value without anomalies, or if any two parameters change in the same direction for four consecutive times relative to a reference value without anomalies.

Images