Method for early warning of multiple organ dysfunction in patients with extremely severe burns

Abstract

This invention relates to the field of multi-organ dysfunction (MOD) early warning technology, specifically to an MOD early warning method for patients with severe burns. The method includes the following steps: acquiring heart rate, mean arterial pressure, respiratory rate, blood oxygen saturation, lactate concentration, and urine output per unit time; calculating the rate of lactate change at adjacent time points to form a sequence; comparing the rate of change across time intervals to obtain the difference in change; combining heart rate and blood oxygen saturation to calculate circulatory load; and counting the number of early warning triggers that meet risk conditions. In this invention, by calculating the rate of change of lactate concentration at adjacent time points and combining it with interval direction statistics, tissue metabolic changes show trend characteristics; the rate of change of urine output combined with mean arterial pressure forms a record of renal perfusion changes, providing a continuous characterization of perfusion status; heart rate and blood oxygen saturation are used to calculate circulatory load; multiple physiological signals are linked for judgment; and recurring abnormalities are expressed through the number of triggers. This improves the stability and predictability of MOD risk identification in critically burned patients.

Description

Technical Field

[0001] This invention relates to the field of early warning technology for multiple organ dysfunction, and in particular to a method for early warning of multiple organ dysfunction in patients with severe burns. Background Technology

[0002] The field of multi-organ dysfunction (MOD) early warning technology refers to the medical monitoring technology field that monitors and identifies the risk of multiple organ dysfunction in critically ill patients during disease progression. This technology field mainly revolves around core aspects such as continuous collection of patients' physiological status, identification of pathological changes, and risk trend assessment. By systematically analyzing patients' vital signs, laboratory test indicators, and disease severity scores, dynamic monitoring of organ function changes can be achieved. Its technical methods include continuous monitoring of clinical indicators such as blood pressure, heart rate, respiratory rate, blood oxygen saturation, urine output, lactate concentration, creatinine level, platelet count, and inflammatory factor levels. Combined with assessment systems such as sequential organ failure scores, acute physiological and chronic health scores, and multi-organ dysfunction scores, a comprehensive judgment is made on the trend of changes in the patient's condition to identify the risk of organ failure in patients. The traditional method for predicting multiple organ dysfunction (MOD) in patients with severe burns refers to a clinical assessment approach that anticipates the risk of MOD arising from systemic inflammatory response, oxidative stress, and microcirculatory disturbances that may occur after burns in these patients. Regarding the technical aspect of identifying organ function impairment risk in burn patients, the traditional method involves recording vital signs such as body temperature, heart rate, respiratory rate, mean arterial pressure, urine output, and blood oxygen saturation, while simultaneously measuring laboratory indicators such as blood lactate concentration, serum creatinine level, serum bilirubin concentration, platelet count, and white blood cell count. This is combined with sequential organ failure scores or acute physiological and chronic health scores to conduct phased assessments of the patient's organ function status, serving as the primary basis for identifying the risk of MOD in patients with severe burns.

[0003] Current technologies primarily rely on vital signs, laboratory indicators, and scoring results for phased assessments. In practice, data is often recorded separately by collection batch or observation point, resulting in inconsistent time bases for various information entering the judgment process. Data from the same patient collected at different minutes or even hours are often juxtaposed for reference, seemingly providing abundant information but making it difficult to confirm the chronological order and interrelationships of abnormalities. Clinically, situations easily arise where lactate levels are persistently elevated while urine output changes are not simultaneously examined, or where a brief rise in mean arterial pressure masks continued inadequate perfusion, leading to risk signals being misinterpreted as short-term fluctuations. Another shortcoming is that the judgment focuses on whether a single value exceeds the limit or whether the score reaches a certain range, lacking detailed characterization of the rate of continuous change, the degree of repetition, and cross-indicator shifts, especially when the condition is still in the early stages of a turning point. At that time, a single test result may still be within an acceptable range, and the total score may not have increased significantly, but the trend of organ damage has gradually accumulated. Clinicians can only wait for more obvious abnormalities to appear before raising their vigilance, thus delaying the intervention point. Furthermore, routine assessments emphasize summarizing the overall severity and lack close connection between metabolic status, renal perfusion status, and circulatory load status. In the face of severe burns with intense inflammatory response, rapid changes in microcirculation, and significant diurnal fluctuations in the course of the disease, it is easy to see abnormalities but find it difficult to determine whether the abnormalities are expanding. For example, a rise in lactate may be due to temporary stress or may correspond to a precursor to perfusion deterioration. Without continuous correlation judgment, the early warning results are prone to being too coarse, and the direction of resource allocation may also be scattered, thus affecting the efficiency of early identification and the accuracy of treatment of critically ill patients. Summary of the Invention

[0004] To achieve the above objectives, the present invention employs the following technical solution: a method for early warning of multiple organ dysfunction in patients with severe burns, comprising the following steps: S1: Acquire time data of heart rate, mean arterial pressure, respiratory rate, blood oxygen saturation, lactate concentration, and urine output per unit time; uniformly sort the time data of the monitoring equipment and establish corresponding relationships; construct a continuous monitoring information structure; and obtain a continuous monitoring record sequence for burn patients. S2: Read the lactate concentration records of adjacent time nodes in the continuous monitoring record sequence of the burn patient, calculate the time difference and calculate the lactate change rate, arrange all lactate change rates by time, and obtain the lactate change rate sequence. S3: Read the lactate change rate record of the continuous time interval of the lactate change rate sequence, compare the lactate change rate of adjacent intervals and mark the direction of change, calculate the specific numerical difference of the lactate change rate of adjacent intervals, and obtain the lactate change rate difference. S4: Read the unit time urine volume record and mean arterial pressure record of the continuous monitoring record sequence of the burn patient, calculate the urine volume change rate and generate the renal perfusion change record by combining the mean arterial pressure change, arrange all renal perfusion change records, and obtain the renal perfusion change sequence. S5: Read the difference in lactate change rate and the renal perfusion change sequence, and read the heart rate and blood oxygen saturation recorded by the bedside monitor. Extract the renal perfusion change records at the same time point as the difference in lactate change rate and perform matching and comparison. Count the number of all time points that meet the preset risk conditions and obtain the number of times the multi-organ dysfunction warning is triggered.

[0005] As a further aspect of the present invention, the continuous monitoring recording sequence includes a monitoring timeline, multi-source indicator sites, and patient status segments; the lactate change rate sequence includes rate fluctuation characteristics, rate inflection points, and rate sustained segments; the lactate change rate difference includes the magnitude of metabolic acceleration, the magnitude of metabolic deceleration, and the intensity of directional shift; the renal perfusion change sequence includes a perfusion stability phase, a perfusion fluctuation phase, and a perfusion decline phase; and the number of multi-organ dysfunction warning triggers includes the risk trigger frequency, the period of concentrated risk, and the degree of risk accumulation.

[0006] As a further aspect of the present invention, the process of calculating the lactate change rate specifically involves: reading the monitoring device time data corresponding to two adjacent lactate concentration records in the lactate concentration time data and calculating the time difference; dividing the lactate concentration difference between the two adjacent lactate concentration records by the time difference to obtain the lactate change rate for a single time interval.

[0007] As a further aspect of the present invention, all lactic acid concentration records are sorted from smallest to largest according to the time data of the monitoring equipment, and the lactic acid change rate of adjacent lactic acid concentration records is calculated one by one, and arranged in the order of the time data of the monitoring equipment to form the lactic acid change rate sequence.

[0008] As a further aspect of the present invention, the specific steps of S1 are as follows: S101: Acquire heart rate, mean arterial pressure, respiratory rate and blood oxygen saturation monitoring records collected by bedside monitors in burn intensive care unit, extract the internal time stamp field of the monitoring records and form a timestamp sequence, perform sorting processing on the timestamp sequence and rearrange the vital signs monitoring records according to the sorting results, and at the same time establish a corresponding data frame index structure to obtain the vital signs time series matrix. S102: Obtain the lactate concentration record from the blood gas analyzer and the urine volume monitoring record from the urinary catheter monitoring device, extract the record time identifier field and call the vital signs time series matrix as a time reference, perform time alignment judgment on the lactate concentration record and the urine volume monitoring record, map the matching record to the corresponding index position and write it into a unified data frame structure to obtain the multi-source monitoring data index structure. S103: Based on the multi-source monitoring data index structure, extract the monitoring data of heart rate, mean arterial pressure, respiratory rate, blood oxygen saturation, lactate concentration and urine output per unit time at the index position, integrate the data frames according to the index order and establish a continuous data linked list structure, set placeholders for missing field positions and form a continuously arranged data structure to obtain the continuous monitoring record sequence of burn patients.

[0009] As a further aspect of the present invention, the specific steps of S2 are as follows: S201: Based on the continuous monitoring record sequence of the burn patient, read the lactate concentration records of adjacent time nodes, extract the lactate concentration numerical field of the time node and the corresponding time record and establish a sequential index structure, perform difference value calculation for lactate concentration records at adjacent index positions, and perform time interval calculation in combination with the corresponding time record, and associate and map the difference value record with the time interval record to form a change rate expression data frame structure, and obtain the lactate rate expression matrix. S202: Extract the index position change rate expression data frame according to the lactate rate expression matrix, perform change rate calculation processing on the lactate concentration difference record and time interval record within each data frame, and associate the change rate record with the corresponding time node index to form a continuously arranged change rate data frame structure. At the same time, establish a sequence identification structure based on the time index to obtain the lactate rate index set. S203: Extract all change rate records based on the lactic acid rate index set, perform sequential sorting processing on the time node index corresponding to the change rate record, serialize and integrate the sorted change rate records according to the time index, and establish a continuous data chain structure to obtain the lactic acid change rate sequence.

[0010] As a further aspect of the present invention, the specific steps of S3 are as follows: S301: Read the lactate change rate record of continuous time interval according to the lactate change rate sequence, extract the change rate numerical field of continuous time interval and establish a sequential index structure, perform numerical comparison processing on the change rate record of adjacent intervals, and mark the direction of the change trend of the time interval according to the comparison result. Associate the direction mark record with the time interval index and form a direction expression data frame structure to obtain the rate direction mark matrix. S302: Extract time interval direction identifier records based on the rate direction identifier matrix, perform direction statistical processing on continuous time interval direction identifier data frames, classify and organize different direction identifier records, establish an association expression structure with the statistical results and the corresponding time interval index to form a trend expression data frame set, and generate a metabolic trend expression matrix. S303: Extract time interval trend expression records based on the metabolic trend expression matrix, perform change difference value calculation processing on continuous time intervals in combination with the change rate numerical field, and arrange and organize the difference records according to the time order to establish a continuous change difference data sequence structure to obtain the lactate change rate difference value.

[0011] As a further aspect of the present invention, the specific steps of S4 are as follows: S401: Based on the continuous monitoring record sequence of the burn patient, read the urine volume record and mean arterial pressure record per unit time, extract the urine volume value field and mean arterial pressure value field of the time node and establish a unified time index structure, perform urine volume change rate calculation for adjacent time node urine volume records, and associate the change rate record with the corresponding time node index to form a urine volume change expression data frame structure, and obtain the urine volume rate expression matrix. S402: Extract the urine volume change rate record at the time node according to the urine volume rate expression matrix, and simultaneously extract the mean arterial pressure record at the corresponding time node. Compare the urine volume change rate record and the mean arterial pressure record with the preset threshold respectively, and establish a renal perfusion status identifier based on the comparison result. Associate the status identifier record with the corresponding time node index to generate a renal perfusion status record set. S403: Based on the renal perfusion status record set, extract all time node renal perfusion change records, perform time-order sorting processing on the time node status records, and serialize and integrate the records according to the time index to establish a continuous change data sequence structure, thereby obtaining the renal perfusion change sequence.

[0012] As a further aspect of the present invention, the specific steps of S5 are as follows: S501: Based on the difference in lactate change rate and the renal perfusion change sequence, read the heart rate and blood oxygen saturation monitoring information recorded by the bedside monitor in the burn intensive care unit, extract the heart rate monitoring value and blood oxygen saturation monitoring value at the time node and establish a unified time index structure, perform circulatory load value calculation on the heart rate monitoring value and blood oxygen saturation monitoring value, and associate the circulatory load value with the corresponding time node index to form a circulatory state expression data frame structure, and obtain the circulatory load record set; S502: Extract the time node cyclic load value records according to the cyclic load record set, and simultaneously extract the time node records corresponding to the lactate change rate difference and renal perfusion change sequence. Perform risk condition comparison processing on the cyclic load value records, lactate change rate difference and renal perfusion change records, and establish a risk status identification structure based on the comparison results. Associate the risk status identification with the time node index to generate an organ risk identification sequence. S503: Based on the organ risk identifier sequence, extract all time node risk status identifier records, perform statistical counting processing on identifiers that meet preset risk conditions, and arrange and aggregate the risk identifier records in sequence according to the time node index to establish a risk trigger counting data structure and obtain the number of multi-organ dysfunction warning triggers.

[0013] As a further aspect of the present invention, the lactate change rate difference is obtained by reading two adjacent lactate change rate records and calculating the numerical difference between the two lactate change rate records, and then arranging them in the order of the monitoring device time data to form a lactate change rate difference record for the corresponding time interval.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, the rate of change of lactate concentration at adjacent time points is calculated and combined with interval directional statistics to show the trend characteristics of tissue metabolic changes. The rate of change of urine volume is combined with mean arterial pressure to form a record of renal perfusion changes, and the perfusion status is continuously characterized. Heart rate and blood oxygen saturation are used to calculate circulatory load. Multiple physiological signals are linked for judgment, and repetitive abnormalities are expressed by the number of triggers. The stability and prospectiveness of multi-organ dysfunction risk identification in critically burned patients are improved. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the steps of the present invention; Figure 2 This is a detailed schematic diagram of S1 of the present invention; Figure 3 This is a detailed schematic diagram of S2 of the present invention; Figure 4 This is a detailed schematic diagram of S3 of the present invention; Figure 5 This is a detailed schematic diagram of S4 of the present invention; Figure 6 This is a detailed schematic diagram of S5 of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0018] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0019] Please see Figure 1 This invention provides a method for early warning of multiple organ dysfunction in patients with severe burns, comprising the following steps: S1: Obtain heart rate, mean arterial pressure, respiratory rate and blood oxygen saturation monitoring records from bedside monitors in the burn intensive care unit, obtain lactate concentration records from blood gas analyzers, obtain urine volume monitoring records per unit time from urinary catheter monitoring devices, uniformly sort the time records of the monitoring devices and establish the corresponding relationship of the monitoring records to form a continuous arrangement of monitoring information structure, and obtain a continuous monitoring record sequence for burn patients. S2: Based on the continuous monitoring record sequence of burn patients, read the lactate concentration records of adjacent time nodes, calculate the lactate change rate according to the time records of adjacent time nodes, and arrange all lactate change rates in chronological order to obtain the lactate change rate sequence. S3: Read the lactate change rate record of continuous time intervals according to the lactate change rate sequence, compare the lactate change rate of adjacent intervals and mark the change direction, perform statistical processing on the change direction record of time intervals, generate result data to characterize the state of tissue metabolic change, and obtain the lactate change rate difference. S4: Based on the continuous monitoring record sequence of burn patients, read the urine volume record and mean arterial pressure record per unit time, calculate the rate of change of urine volume according to the time record, and generate the renal perfusion change record by combining the change of mean arterial pressure. Arrange the renal perfusion change records of all time points in sequence to obtain the renal perfusion change sequence. S5: Based on the difference in lactate change rate and the renal perfusion change sequence, read the heart rate and blood oxygen saturation monitoring information recorded by the bedside monitor in the burn intensive care unit. Calculate the heart rate monitoring information and blood oxygen saturation monitoring information to generate a circulatory load record. Then, compare and process the difference in lactate change rate and the renal perfusion change sequence, and statistically analyze the time nodes that meet the preset risk conditions to obtain the number of times the multi-organ dysfunction warning is triggered.

[0020] The continuous monitoring record sequence includes a monitoring timeline, multi-source indicator sites, and patient status segments; the lactate change rate sequence includes rate fluctuation characteristics, rate inflection points, and rate sustained segments; the lactate change rate difference includes the magnitude of metabolic acceleration, the magnitude of metabolic deceleration, and the intensity of directional shift; the renal perfusion change sequence includes the perfusion stabilization phase, the perfusion fluctuation phase, and the perfusion decline phase; the number of multiple organ dysfunction warning triggers includes the frequency of risk triggers, the period of concentrated risk, and the degree of risk accumulation.

[0021] Please see Figure 2 The specific steps of S1 are as follows: S101: Acquire heart rate, mean arterial pressure, respiratory rate and blood oxygen saturation monitoring records collected by bedside monitors in burn intensive care unit, extract the internal time stamp field of the monitoring records and form a timestamp sequence, perform sorting processing on the timestamp sequence and rearrange the vital signs monitoring records according to the sorting results, and at the same time establish a corresponding data frame index structure to obtain the vital signs time series matrix. Bedside monitors in the burn intensive care unit transmit monitoring records in real time to the monitoring data server via a wired network. The acquisition frequency is set to once every 10 seconds, and the collected data includes four vital signs: heart rate, mean arterial pressure, respiratory rate, and blood oxygen saturation. The raw records are stored in structured text format, with each record containing the device number, patient number, acquisition time identifier, heart rate value, mean arterial pressure value, respiratory rate value, and blood oxygen saturation value. When reading data, the acquisition time identifier field is first extracted from each record and converted into a standard timestamp sequence. For example, 08:00:10 on June 1, 2025 is converted into the timestamp 1748736010, and this is sequentially formed into a timestamp sequence. Then, all timestamp sequences are sorted in ascending order, and the corresponding vital sign records are rearranged according to the sorting result. For example, if four records are collected for a patient in the first minute, the original order may have a device cache write delay, as shown in Table 1.

[0022] Table 1 Original Record of Vital Signs Collection Collection time heart rate beats per minute Mean arterial pressure (mmHg) respiratory rate (breathing rate) per minute blood oxygen saturation percentage 08:00:30 118 72 24 96 08:00:10 115 70 22 97 08:00:40 120 74 25 95 08:00:20 117 71 23 96 After converting the collection time into timestamps and sorting in ascending order, the data sequence was rearranged to 08:00:10, 08:00:20, 08:00:30, and 08:00:40. A data frame index structure was then built according to the sorting order, with the first record indexed as 1, the second as 2, and so on. The resulting vital signs time series matrix is ​​as follows: Index 1 corresponds to a heart rate of 115 beats per minute, mean arterial pressure of 70 mmHg, respiratory rate of 22 breaths per minute, and oxygen saturation of 97%; Index 2 corresponds to 117 beats per minute, 71 mmHg, 23 breaths per minute, and 96%; Index 3 corresponds to 118 beats per minute, 72 mmHg, 24 breaths per minute, and 96%; Index 4 corresponds to 120 beats per minute, 74 mmHg, 25 breaths per minute, and 95%. This forms a vital signs time series matrix arranged in chronological order.

[0023] S102: Obtain the lactate concentration record from the blood gas analyzer and the urine volume monitoring record from the urinary catheter monitoring device, extract the record time identifier field and call the vital signs time series matrix as the time reference, perform time alignment judgment on the lactate concentration record and the urine volume monitoring record, map the matching record to the corresponding index position and write it into a unified data frame structure to obtain the multi-source monitoring data index structure. The blood gas analyzer outputs a lactate concentration test record every 30 minutes, and the urinary catheter monitoring device outputs a urine volume record per unit time every 60 minutes. When reading these records, the detection time field is extracted and converted into a timestamp, and then the vital signs time series matrix generated in step S101 is used as a time reference. Subsequently, time alignment is performed record by record. Specifically, the timestamp of the lactate concentration record is read and compared with two adjacent timestamps in the vital signs time series. When the lactate detection time falls between two timestamps, the lactate record is mapped to the next time node index position. For example, if the lactate detection record time is 08:00:25, and its timestamp is between the vital signs records 08:00:20 and 08:00:30, then it is mapped to index 3. Assuming the blood gas analyzer detects a lactate concentration of 3.2 mmol / L, this value is written to the data frame field corresponding to index 3. The urine volume monitoring device recorded a urine volume of 45 ml between 08:00 and 09:00. This record is time-stamped at 09:00. This record is then mapped to the corresponding vital sign time node index position near the timestamp 09:00. For interval-cumulative monitoring data (e.g., urine volume per unit time), the time stamp preferably uses the end time of the monitoring interval to ensure the data time sequence is consistent with the actual collection process. After writing to a unified data frame, the data at each index position forms the following structure example: Index 1 contains only vital sign records; Index 2 contains only vital sign records; Index 3 contains a heart rate of 118 beats per minute, mean arterial pressure of 72 mmHg, respiratory rate of 24 breaths per minute, oxygen saturation of 96%, and lactate concentration of 3.2 mmol / L; Index 4 contains a heart rate of 120 beats per minute, mean arterial pressure of 74 mmHg, respiratory rate of 25 breaths per minute, oxygen saturation of 95%, and urine volume of 45 ml. Through this time mapping process, records from different monitoring devices are uniformly written into data frames with the same index structure, thereby forming a multi-source monitoring data index structure.

[0024] S103: Based on the multi-source monitoring data index structure, extract the monitoring data of heart rate, mean arterial pressure, respiratory rate, blood oxygen saturation, lactate concentration and urine output per unit time at the index position, integrate the data frames according to the index order and establish a continuous data linked list structure, set placeholders for missing field positions and form a continuous data structure to obtain the continuous monitoring record sequence of burn patients. Based on a multi-source monitoring data index structure, vital sign records, lactate concentration records, and urine output monitoring records stored at each index location are read sequentially. During reading, an integration operation is performed according to the index number order, writing heart rate, mean arterial pressure, respiratory rate, blood oxygen saturation, lactate concentration, and urine output per unit time into a unified data frame. If a lactate concentration or urine output record is missing at a certain index location, a missing placeholder is written to the corresponding field. For example, if only a vital sign record exists at index 1, the corresponding data frame will contain a heart rate of 115 beats per minute, a mean arterial pressure of 70 mmHg, a respiratory rate of 22 breaths per minute, and a blood oxygen saturation of 97%, while placeholders are written to both the lactate and urine output fields. If a lactate concentration record of 3.2 mmol / L exists at index 3, this value is written to that field. If a urine output record of 45 ml exists at index 4, this value is written to the urine output field. Subsequently, the data frames are linked together sequentially according to the index order to form a continuous data linked list structure. For example, the first node of the linked list stores the data frame at index 1, the second node stores the data frame at index 2, the third node stores the data frame at index 3, and the fourth node stores the data frame at index 4. Once the linked list structure is formed, each node contains a complete time-based data structure. Taking index 3 as an example, its recorded data includes a heart rate of 118 beats per minute, a mean arterial pressure of 72 mmHg, a respiratory rate of 24 breaths per minute, an oxygen saturation of 96%, a lactate concentration of 3.2 mmol / L, and a missing urine output marker. By continuously adding new monitoring record nodes through the linked structure, a continuous monitoring record sequence for burn patients can be obtained.

[0025] Please see Figure 3 The specific steps of S2 are as follows: S201: Based on the continuous monitoring record sequence of burn patients, read lactate concentration records at adjacent time nodes, extract the lactate concentration numerical field of time nodes and the corresponding time records and establish a sequential index structure, perform difference value calculation for lactate concentration records at adjacent index positions, and perform time interval calculation in combination with the corresponding time records, and associate and map the difference value records with the time interval records to form a change rate expression data frame structure, and obtain the lactate rate expression matrix. In the continuous monitoring record sequence of burn patients, time nodes containing lactate concentration fields are read sequentially, and the lactate concentration value and corresponding time record are extracted from each record. First, a sequential index structure is established; for example, the first lactate record is indexed as 1, and the second as 2. Then, the difference in lactate concentration between adjacent index positions is calculated. Specifically, the lactate concentration value of the subsequent time node is read and subtracted from the lactate concentration value of the previous time node. For example, index 1 has a lactate concentration of 2.8 mmol / L, and the detection time is 08:00; index 2 has a lactate concentration of 3.2 mmol / L, and the detection time is 08:30. The difference is calculated to obtain a lactate concentration difference of 0.4 mmol / L. Then, the two detection time records are read and the time interval is calculated; in this example, the time interval is 30 minutes. An association mapping is established between the lactate concentration difference of 0.4 mmol / L and the 30-minute time interval, forming a data frame expressing the rate of change. Subsequent records are processed in the same manner. For example, the difference between lactate at 08:30 (3.2 mmol / L) and 09:00 (3.5 mmol / L) is 0.3 mmol / L, with the time interval remaining 30 minutes. The processed data frames show the following lactate rate expression: index intervals 1 and 2 correspond to a difference of 0.4 mmol / L at a time interval of 30 minutes; index intervals 2 and 3 correspond to a difference of 0.3 mmol / L at a time interval of 30 minutes. All data frames are then sequentially combined to form a lactate rate expression matrix.

[0026] S202: Extract the index position change rate expression data frame based on the lactate rate expression matrix, perform change rate calculation processing on the lactate concentration difference record and time interval record within each data frame, and associate the change rate record with the corresponding time node index to form a continuously arranged change rate data frame structure. At the same time, establish a sequence identification structure based on the time index to obtain the lactate rate index set. The lactate rate expression matrix is ​​read line by line from the change rate expression data frames, and the lactate change rate is calculated. During calculation, the lactate concentration difference record and time interval record are read from the data frame. The lactate change rate per unit time is obtained by dividing the difference value by the time interval. For example, if the difference value is 0.4 mmol / L and the time interval is 30 minutes, the lactate change rate per unit time is 0.013 mmol / L per minute. If the difference value is 0.3 mmol / L and the time interval is 30 minutes, the change rate is 0.010 mmol / L per minute. The calculated change rate records are written into a new data frame structure and associated with the corresponding time node index. For example, the change rate record for the 08:30 node is 0.013 mmol / L per minute, and the change rate record for the 09:00 node is 0.010 mmol / L per minute. Subsequently, a sequence identifier structure is established based on the time index, with the index numbers arranged in chronological order. For example, index 1 represents the rate record for the 08:30 node, and index 2 represents the rate record for the 09:00 node. All change rate data frames are arranged and combined according to this index structure to obtain the lactate rate index set.

[0027] S203: Extract all change rate records based on the lactate rate index set, perform sequential sorting processing on the time node index corresponding to the change rate record, serialize and integrate the sorted change rate records according to the time index and establish a continuous data chain structure to obtain the lactate change rate sequence. All rate change records are read from the lactate rate index set and sorted sequentially according to the time node index. During sorting, rate records are read in chronological order, for example, 08:30 node: 0.013 mmol / L / min, 09:00 node: 0.010 mmol / L / min, 09:30 node: 0.015 mmol / L / min. The sorted rate change records are then written into a continuous sequence structure. Specifically, a continuous data link structure is established, with each node recording the rate change at one time node. The first node of the link records 0.013 mmol / L / min, the second node records 0.010 mmol / L / min, and the third node records 0.015 mmol / L / min. As new lactate detection data is added, the difference value is calculated and a new rate of change node is appended. For example, if the lactate level is 3.9 mmol / L at 10:00 and 3.7 mmol / L at 09:30, the difference is 0.2 mmol / L, and the time interval is 30 minutes, then the rate of change is 0.0067 mmol / L per minute. This rate is recorded in the 4th node of the linked list. By continuously appending data, a complete time series is formed, thus obtaining the lactate rate of change sequence.

[0028] Please see Figure 4 The specific steps of S3 are as follows: S301: Read the lactate change rate record of continuous time interval according to the lactate change rate sequence, extract the change rate numerical field of continuous time interval and establish a sequential index structure, perform numerical comparison processing on the change rate record of adjacent intervals, and mark the direction of the change trend of the time interval according to the comparison result. Associate the direction mark record with the time interval index and form a direction expression data frame structure to obtain the rate direction mark matrix. The system reads the rate of change records for continuous time intervals based on the lactate change rate sequence, for example, reading 0.013 mmol / L / min at 08:30 and 0.010 mmol / L / min at 09:00. First, the rate of change values ​​for each interval are extracted and a sequential index structure is established, with index 1 representing the rate record for the first interval and index 2 representing the rate record for the second interval. Then, numerical comparisons are performed on adjacent interval rate records. When the value of a later record is greater than that of a previous record, the trend for that interval is marked as upward; when the value of a later record is less than that of a previous record, the trend is marked as downward. For example, comparing 0.013 mmol / L / min with 0.010 mmol / L / min results in a downward trend. Then comparing 0.010 mmol / L / min with 0.015 mmol / L / min results in an upward trend. The trend direction markers are then associated with the corresponding time interval indices; for example, the interval from 08:30 to 09:00 is marked as downward, and the interval from 09:00 to 09:30 is marked as upward. All interval direction identifiers are organized in index order into a set of direction expression data frames, thus forming a rate direction identifier matrix.

[0029] S302: Extract direction identifier records for each time interval based on the rate direction identifier matrix, perform direction statistical processing on the direction identifier data frames of continuous time intervals, classify and organize different direction identifier records, establish an association expression structure with the statistical results and the corresponding time interval index to form a trend expression data frame set, and generate a metabolic trend expression matrix. The direction records for each time interval are read based on the rate direction identifier matrix, and statistical processing is performed on the direction identifiers of consecutive intervals. During the statistics, the number of upward direction records and the number of downward direction records are calculated separately. For example, the statistical result for five consecutive time intervals is 3 upward direction records and 2 downward direction records. Then, the different direction identifiers are classified and organized, with upward direction records grouped into one category and downward direction records into another, and the statistical results are written into a trend expression data frame. For example, the statistical result for the time interval from 08:30 to 11:00 is 3 upward direction records and 2 downward direction records. An association expression structure is established between this statistical result and the corresponding time interval index. The data frame content includes the interval start time, interval end time, number of upward direction records, and number of downward direction records. All time interval statistical results are combined in chronological order to form a trend expression data frame set; for example, the first interval is counted as 2 upward direction records and 1 downward direction record, and the second interval is counted as 3 upward direction records and 2 downward direction records. This generates a metabolic trend expression matrix.

[0030] S303: Extract time interval trend expression records based on the metabolic trend expression matrix, perform change difference value calculation processing on continuous time intervals in combination with the change rate numerical field, and arrange and organize the difference records according to the time order to establish a continuous change difference data sequence structure to obtain the lactate change rate difference value. The trend records for each time interval are read sequentially from the metabolic trend expression matrix, and the number of increases and decreases are extracted. Then, the difference in value is calculated by subtracting the number of decreases from the number of increases. For example, if an interval has 3 increases and 2 decreases, the difference in value is 1; if another interval has 1 increase and 3 decreases, the difference in value is -2. The difference records for each time interval are arranged and organized with their corresponding time interval indices; for example, interval 1 has a difference of 1, interval 2 has a difference of -2, and interval 3 has a difference of 2. Then, the difference records are sequentially written into a continuous difference data sequence structure according to time order; for example, the first node of the linked list records a difference of 1, the second node records a difference of -2, and the third node records a difference of 2. This continuous difference sequence represents the degree of difference in the rate of lactate change across different time intervals, thus obtaining the difference in the rate of lactate change.

[0031] Please see Figure 5 The specific steps of S4 are as follows: S401: Based on the continuous monitoring record sequence of burn patients, read the urine volume record and mean arterial pressure record per unit time, extract the urine volume value field and mean arterial pressure value field of time node and establish a unified time index structure, perform urine volume change rate calculation for urine volume records of adjacent time nodes, and associate the change rate record with the corresponding time node index to form a urine volume change expression data frame structure, and obtain the urine volume rate expression matrix. Urine volume and mean arterial pressure (MAP) records were read from the continuous monitoring series of burn patients, and the urine volume and MAP values ​​at each time point were extracted. First, a unified time index structure was established, for example, 08:00 node index number 1, 09:00 node index number 2. Then, the rate of change of urine volume was calculated for adjacent time points. The calculation involved reading the urine volume at the next time point, subtracting the urine volume at the previous time point, and then dividing by the time interval between the two records. For example, if the urine volume from 08:00 to 09:00 was 45 ml, and the urine volume from 09:00 to 10:00 was 35 ml, with a time interval of 60 minutes, the difference in urine volume was -10 ml. Dividing this difference by 60 minutes yielded a rate of change of urine volume of -0.167 ml per minute. The difference of 15 ml between the urine volume of 50 ml from 10:00 to 11:00 and the urine volume of 35 ml from 09:00 to 10:00, with a time interval of 60 minutes, is calculated. The rate of change is 0.25 ml per minute. These rate records are then associated with their corresponding time node indices and written into the urine volume change expression data frame. For example, the rate recorded at the 09:00 node is -0.167 ml per minute, and the rate recorded at the 10:00 node is 0.25 ml per minute, thus obtaining the urine volume rate expression matrix.

[0032] S402: Extract the urine volume change rate records at time nodes based on the urine volume rate expression matrix, and simultaneously extract the mean arterial pressure records at the corresponding time nodes. Compare the urine volume change rate records and mean arterial pressure records with preset thresholds respectively, and establish renal perfusion status identifiers based on the comparison results. Associate the status identifier records with the corresponding time node indexes to generate a renal perfusion status record set. The urine volume rate of change records at each time point are read sequentially based on the urine volume rate expression matrix, and the mean arterial pressure record at the same time point is read simultaneously. For example, at 09:00, the urine volume rate of change is -0.167 ml / min, and the mean arterial pressure is 72 mmHg. Then, a status comparison process is performed. First, a baseline value for renal perfusion is established. This baseline value is determined through clinical statistics. In nearly 200 patients in the burn intensive care unit, the incidence of insufficient renal perfusion was found to be 78% when the mean arterial pressure was below 65 mmHg. Therefore, a mean arterial pressure of 65 mmHg is set as the perfusion baseline value. Simultaneously, a urine volume rate of change below -0.10 ml / min is considered a period of decreased urine volume. When the mean arterial pressure (MAP) is below 65 mmHg and the rate of change in urine volume is below -0.10 ml / min, it is recorded as a state of declining renal perfusion. When the MAP is above 65 mmHg and the rate of change in urine volume is positive, it is recorded as a state of stable perfusion. When the MAP is above 65 mmHg but the rate of change in urine volume is below -0.10 ml / min, it is recorded as a state of fluctuating perfusion. For example, at 09:00, the MAP is 72 mmHg, which is higher than the baseline of 65 mmHg, but the rate of change in urine volume is -0.167 ml / min, which is below the decline threshold, so it is determined to be a state of fluctuating perfusion. At 10:00, the MAP is 75 mmHg and the rate of change in urine volume is 0.25 ml / min, so it is recorded as a state of stable perfusion. The state identifiers are associated with the time node indexes to generate a set of renal perfusion state records.

[0033] S403: Extract all time-node renal perfusion change records based on the renal perfusion status record set, perform time-order sorting processing on the time-node status records, and serialize and integrate the records according to the time index to establish a continuous change data sequence structure and obtain the renal perfusion change sequence. All time-point status records are read from the renal perfusion status record set and arranged in chronological order. For example, the status at 09:00 is perfusion decline, at 10:00 is perfusion stable, and at 11:00 is perfusion stable. Then, a serialization and integration operation is performed, writing each time-point status record into a continuous change data sequence structure. Specifically, a continuous linked list structure is established, with the first node recording 09:00 perfusion decline, the second node recording 10:00 perfusion stable, and the third node recording 11:00 perfusion stable. As new urine volume records are generated, a judgment operation is performed and nodes are added. For example, at 12:00, the urine volume change rate is -0.20 ml / min and the mean arterial pressure is 63 mmHg, lower than the baseline value; therefore, this node is recorded as a perfusion decline state and added to the linked list. This sequential arrangement and linked list appending method forms a continuous renal perfusion change sequence.

[0034] Please see Figure 6 The specific steps of S5 are as follows: S501: Based on the difference in lactate change rate and the renal perfusion change sequence, read the heart rate and blood oxygen saturation monitoring information recorded by the bedside monitor in the burn intensive care unit, extract the heart rate monitoring value and blood oxygen saturation monitoring value at time nodes and establish a unified time index structure, perform circulatory load value calculation on the heart rate monitoring value and blood oxygen saturation monitoring value, and associate the circulatory load value with the corresponding time node index to form a circulatory state expression data frame structure, and obtain the circulatory load record set; Cyclic load value The calculation formula is: ; in, This is the baseline heart rate value. This is the baseline value for blood oxygen saturation. and These are conversion factors used to eliminate differences in dimensions. and To reflect the weighting coefficients of the impact of heart rate deviation and blood oxygenation decline on circulatory load, the calculated circulatory load values ​​are associated with the corresponding time node indices to form a circulatory state expression data frame structure, resulting in a circulatory load record set. The system reads heart rate and blood oxygen saturation monitoring information recorded by the bedside monitor and extracts heart rate and blood oxygen saturation values ​​at each time point. For example, a patient's heart rate is 118 beats per minute and blood oxygen saturation is 95% at 09:00. A unified time index structure is then established to map the monitoring records to the time points of the lactate change rate difference and renal perfusion change sequence. Circulatory load is then calculated. Circulatory load is calculated by combining the heart rate value and the degree of deviation of blood oxygen saturation from the normal range. Clinical statistics define the normal heart rate range as 60 to 100 beats per minute and the normal blood oxygen saturation range as 95 to 100%. When the heart rate exceeds 100 beats per minute, it is recorded as a heart rate deviation value, calculated by subtracting 100 from the actual heart rate. When blood oxygen saturation is below 95%, it is recorded as a blood oxygen decrease value, calculated by subtracting the actual blood oxygen saturation from 95. For example, at 09:00, the heart rate is 118 beats per minute, so the heart rate deviation value is 18; the blood oxygen saturation is 95%, so the blood oxygen decrease value is 0. The heart rate deviation value and the blood oxygen decrease value are then summed to obtain the circulatory load value of 18. As another example, at 09:30, the heart rate is 125 beats per minute and the blood oxygen saturation is 93%, so the heart rate deviation value is 25 and the blood oxygen decrease value is 2. The sum of these two values ​​gives the circulatory load value of 27. The calculated circulatory load values ​​are then associated with the corresponding time node indices to form a circulatory state expression data frame structure, thus obtaining the circulatory load record set.

[0035] S502: Extract the time node cyclic load value records from the cyclic load record set, and simultaneously extract the time node records corresponding to the lactate change rate difference and renal perfusion change sequence. Perform risk condition comparison processing on the cyclic load value records, lactate change rate difference and renal perfusion change records, and establish a risk status label structure based on the comparison results. Associate the risk status labels with the time node index to generate an organ risk label sequence. The circulatory load values ​​at each time point are read from the circulatory load record set, and the time point records corresponding to the lactate rate of change difference and the renal perfusion change sequence are extracted simultaneously. For example, at 09:30, the circulatory load value is 27, the lactate rate of change difference is 2, and the renal perfusion status is declining. Risk condition comparison processing is then performed. First, a risk judgment threshold is set. Statistical analysis of monitoring data from 120 severe burn patients shows that when the circulatory load value exceeds 20 and the lactate rate of change difference is greater than 1, and renal perfusion is declining, the probability of multiple organ dysfunction reaches 82%. Therefore, a circulatory load value of 20 is set as the risk threshold, and a lactate rate of change difference of 1 is set as the trend threshold. When the circulatory load value is greater than 20, the lactate rate of change difference is greater than 1, and the renal perfusion status is declining, it is recorded as a risk state. For example, at 09:30, the circulatory load value is 27, which is greater than 20, the lactate rate of change difference is 2, which is greater than 1, and the renal perfusion status is declining; therefore, this time point is recorded as a risk state. A correlation record is established between the risk state identifier and the time point index to generate an organ risk identifier sequence.

[0036] S503: Extract all time node risk status identifier records based on organ risk identifier sequence, perform statistical counting processing on identifiers that meet preset risk conditions, and arrange and aggregate the risk identifier records in sequence according to the time node index to establish a risk trigger counting data structure and obtain the number of multi-organ dysfunction warning triggers. Based on the organ risk identifier sequence, all time-node risk status records are read, and statistical counting is performed on records that meet the risk conditions. For example, if a risk status record appears 3 times in 6 consecutive monitoring time nodes, then all risk identifier records are arranged in time node index order, and an aggregation operation is performed to accumulate the number of risk nodes. For example, if the risk conditions are met at 09:30, 10:00, and 10:30, the cumulative count is 3. Subsequent node records are then read; for example, if the risk conditions are not met at 11:00, the cumulative count remains 3; if the risk conditions are met again at 11:30, the cumulative count is updated to 4. The statistical results form a risk trigger count data structure; for example, if a risk status is triggered 4 times in 12 consecutive monitoring time nodes, this statistical value represents the number of multiple organ dysfunction warning triggers.

[0037] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for early warning of multiple organ dysfunction in patients with severe burns, characterized in that, Includes the following steps: S1: Acquire time data of heart rate, mean arterial pressure, respiratory rate, blood oxygen saturation, lactate concentration, and urine output per unit time; uniformly sort the time data of the monitoring equipment and establish corresponding relationships; construct a continuous monitoring information structure; and obtain a continuous monitoring record sequence for burn patients. S2: Read the lactate concentration records of adjacent time nodes in the continuous monitoring record sequence of the burn patient, calculate the time difference and calculate the lactate change rate, arrange all lactate change rates by time, and obtain the lactate change rate sequence. S3: Read the lactate change rate record of the continuous time interval of the lactate change rate sequence, compare the lactate change rate of adjacent intervals and mark the direction of change, calculate the specific numerical difference of the lactate change rate of adjacent intervals, and obtain the lactate change rate difference. S4: Read the unit time urine volume record and mean arterial pressure record of the continuous monitoring record sequence of the burn patient, calculate the urine volume change rate and generate the renal perfusion change record by combining the mean arterial pressure change, arrange all renal perfusion change records, and obtain the renal perfusion change sequence. S5: Read the difference in lactate change rate and the renal perfusion change sequence, and read the heart rate and blood oxygen saturation recorded by the bedside monitor. Extract the renal perfusion change records at the same time point as the difference in lactate change rate and perform matching and comparison. Count the number of all time points that meet the preset risk conditions and obtain the number of times the multi-organ dysfunction warning is triggered.

2. The method for early warning of multiple organ dysfunction in patients with severe burns according to claim 1, characterized in that, The continuous monitoring record sequence includes a monitoring timeline, multi-source indicator sites, and patient status segments; the lactate change rate sequence includes rate fluctuation characteristics, rate inflection points, and rate sustained segments; the lactate change rate difference includes the magnitude of metabolic acceleration, the magnitude of metabolic deceleration, and the intensity of directional shift; the renal perfusion change sequence includes a perfusion stability phase, a perfusion fluctuation phase, and a perfusion decline phase; the number of multi-organ dysfunction warning triggers includes the frequency of risk triggers, the period of concentrated risk, and the degree of risk accumulation.

3. The method for early warning of multiple organ dysfunction in patients with severe burns according to claim 1, characterized in that: The process of calculating the rate of change of lactic acid is as follows: read the time data of the monitoring device corresponding to two adjacent lactic acid concentration records in the lactic acid concentration time data and calculate the time difference. Divide the lactic acid concentration difference between the two adjacent lactic acid concentration records by the time difference to obtain the rate of change of lactic acid in a single time interval.

4. The method for early warning of multiple organ dysfunction in patients with severe burns according to claim 1, characterized in that: All lactate concentration records are sorted from smallest to largest according to the time data of the monitoring equipment. Then, the lactate change rate of adjacent lactate concentration records is calculated one by one and arranged in the order of the time data of the monitoring equipment to form the lactate change rate sequence.

5. The method for early warning of multiple organ dysfunction in patients with severe burns according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Acquire heart rate, mean arterial pressure, respiratory rate and blood oxygen saturation monitoring records collected by bedside monitors in burn intensive care unit, extract the internal time stamp field of the monitoring records and form a timestamp sequence, perform sorting processing on the timestamp sequence and rearrange the vital signs monitoring records according to the sorting results, and at the same time establish a corresponding data frame index structure to obtain the vital signs time series matrix. S102: Obtain the lactate concentration record from the blood gas analyzer and the urine volume monitoring record from the urinary catheter monitoring device, extract the record time identifier field and call the vital signs time series matrix as a time reference, perform time alignment judgment on the lactate concentration record and the urine volume monitoring record, map the matching record to the corresponding index position and write it into a unified data frame structure to obtain the multi-source monitoring data index structure. S103: Based on the multi-source monitoring data index structure, extract the monitoring data of heart rate, mean arterial pressure, respiratory rate, blood oxygen saturation, lactate concentration and urine output per unit time at the index position, integrate the data frames according to the index order and establish a continuous data linked list structure, set placeholders for missing field positions and form a continuously arranged data structure to obtain the continuous monitoring record sequence of burn patients.

6. The method for early warning of multiple organ dysfunction in patients with severe burns according to claim 1, characterized in that, The specific steps of S2 are as follows: S201: Based on the continuous monitoring record sequence of the burn patient, read the lactate concentration records of adjacent time nodes, extract the lactate concentration numerical field of the time node and the corresponding time record and establish a sequential index structure, perform difference value calculation for lactate concentration records at adjacent index positions, and perform time interval calculation in combination with the corresponding time record, and associate and map the difference value record with the time interval record to form a change rate expression data frame structure, and obtain the lactate rate expression matrix. S202: Extract the index position change rate expression data frame according to the lactate rate expression matrix, perform change rate calculation processing on the lactate concentration difference record and time interval record within each data frame, and associate the change rate record with the corresponding time node index to form a continuously arranged change rate data frame structure. At the same time, establish a sequence identification structure based on the time index to obtain the lactate rate index set. S203: Extract all change rate records based on the lactic acid rate index set, perform sequential sorting processing on the time node index corresponding to the change rate record, serialize and integrate the sorted change rate records according to the time index, and establish a continuous data chain structure to obtain the lactic acid change rate sequence.

7. The method for early warning of multiple organ dysfunction in patients with severe burns according to claim 1, characterized in that, The specific steps for S3 are as follows: S301: Read the lactate change rate record of continuous time interval according to the lactate change rate sequence, extract the change rate numerical field of continuous time interval and establish a sequential index structure, perform numerical comparison processing on the change rate record of adjacent intervals, and mark the direction of the change trend of the time interval according to the comparison result. Associate the direction mark record with the time interval index and form a direction expression data frame structure to obtain the rate direction mark matrix. S302: Extract time interval direction identifier records based on the rate direction identifier matrix, perform direction statistical processing on continuous time interval direction identifier data frames, classify and organize different direction identifier records, establish an association expression structure with the statistical results and the corresponding time interval index to form a trend expression data frame set, and generate a metabolic trend expression matrix. S303: Extract time interval trend expression records based on the metabolic trend expression matrix, perform change difference value calculation processing on continuous time intervals in combination with the change rate numerical field, and arrange and organize the difference records according to the time order to establish a continuous change difference data sequence structure to obtain the lactate change rate difference value.

8. The method for early warning of multiple organ dysfunction in patients with severe burns according to claim 1, characterized in that, The specific steps of S4 are as follows: S401: Based on the continuous monitoring record sequence of the burn patient, read the urine volume record and mean arterial pressure record per unit time, extract the urine volume value field and mean arterial pressure value field of the time node and establish a unified time index structure, perform urine volume change rate calculation for adjacent time node urine volume records, and associate the change rate record with the corresponding time node index to form a urine volume change expression data frame structure, and obtain the urine volume rate expression matrix. S402: Extract the urine volume change rate record at the time node according to the urine volume rate expression matrix, and simultaneously extract the mean arterial pressure record at the corresponding time node. Compare the urine volume change rate record and the mean arterial pressure record with the preset threshold respectively, and establish a renal perfusion status identifier based on the comparison result. Associate the status identifier record with the corresponding time node index to generate a renal perfusion status record set. S403: Based on the renal perfusion status record set, extract all time node renal perfusion change records, perform time-order sorting processing on the time node status records, and serialize and integrate the records according to the time index to establish a continuous change data sequence structure, thereby obtaining the renal perfusion change sequence.

9. The method for early warning of multiple organ dysfunction in patients with severe burns according to claim 1, characterized in that, The specific steps of S5 are as follows: S501: Based on the difference in lactate change rate and the renal perfusion change sequence, read the heart rate and blood oxygen saturation monitoring information recorded by the bedside monitor in the burn intensive care unit, extract the heart rate monitoring value and blood oxygen saturation monitoring value at the time node and establish a unified time index structure, perform circulatory load value calculation on the heart rate monitoring value and blood oxygen saturation monitoring value, and associate the circulatory load value with the corresponding time node index to form a circulatory state expression data frame structure, and obtain the circulatory load record set; S502: Extract the time node cyclic load value records according to the cyclic load record set, and simultaneously extract the time node records corresponding to the lactate change rate difference and renal perfusion change sequence. Perform risk condition comparison processing on the cyclic load value records, lactate change rate difference and renal perfusion change records, and establish a risk status identification structure based on the comparison results. Associate the risk status identification with the time node index to generate an organ risk identification sequence. S503: Based on the organ risk identifier sequence, extract all time node risk status identifier records, perform statistical counting processing on identifiers that meet preset risk conditions, and arrange and aggregate the risk identifier records in sequence according to the time node index to establish a risk trigger counting data structure and obtain the number of multi-organ dysfunction warning triggers.

10. The method for early warning of multiple organ dysfunction in patients with severe burns according to claim 1, characterized in that: The lactate change rate difference is obtained by reading two adjacent lactate change rate records and calculating the numerical difference between the two lactate change rate records, and then arranging them according to the time data of the monitoring device to form a lactate change rate difference record for the corresponding time interval.

Images