A method for prognostic assessment and early intervention in patients with multiple injuries

Abstract

This invention relates to the field of medical data mining technology, specifically to a method for prognostic assessment and early intervention in patients with multiple injuries. First, by calculating injury variability and performing cluster analysis, the method systematically processes multi-dimensional dynamic data of patients to identify key areas of concern regarding their injuries. Second, by analyzing the degree of concern for injuries and changes in physiological data after early intervention, the intervention plan for each patient with multiple injuries in historical data is evaluated to determine the corresponding intervention score. Then, by combining the deviation characteristics represented by injury variability with the intervention score, a comprehensive representation of the predicted intervention score for the patient to be scored is obtained, resulting in higher accuracy of the predicted intervention score; furthermore, the efficiency of obtaining the predicted intervention score is also higher without utilizing a deep learning model.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical data mining, and particularly relates to a method for prognosis evaluation and early intervention of multiple injury patients. BACKGROUND

[0002] Multiple injury is one of the most complex and critical conditions in the field of trauma medicine, and is commonly seen in traffic accidents, high-altitude falls, industrial injuries and other serious trauma events. Such patients usually involve injuries to multiple organ systems, have the characteristics of rapid disease progression, high mortality and short intervention window. The determination of the traditional intervention scheme for multiple injury patients depends on the experience judgment of doctors and standardized scoring systems such as ISS and TRISS. Due to the complex data dimensions of multiple injury patients that need to be judged and the fact that the data itself is not intuitive, there are problems such as long response time, insufficient individualization, and high probability of misjudgment.

[0003] Considering that a deep learning model can combine multi-dimensional data for comprehensive analysis and scoring determination, the prior art usually inputs the multiple injury patient data into a trained convolutional neural network model to output a predicted intervention score for each intervention scheme. However, the training of a convolutional neural network model requires a large number of data samples, resulting in generally poor efficiency; and the physiological monitoring dimension data of multiple injury patients is usually complex, and in the case of insufficient training samples, the accuracy of the convolutional neural network model in outputting the predicted intervention score cannot be guaranteed; therefore, the accuracy and efficiency of the prior art in determining the predicted intervention score of each multiple injury patient for each intervention scheme by means of a convolutional neural network model are both low. SUMMARY

[0004] In order to solve the technical problem that the accuracy and efficiency of the prior art in determining the predicted intervention score of each multiple injury patient for each intervention scheme by means of a convolutional neural network model are both low, the purpose of the present application is to provide a method for prognosis evaluation and early intervention of multiple injury patients, and the technical solution adopted is as follows:

[0005] The first aspect of the present application provides a method for prognosis evaluation and early intervention of multiple injury patients, comprising:

[0006] obtaining the part score vector of each multiple injury patient and the pre-intervention data sequence and post-intervention data sequence of each physiological monitoring dimension in the medical database;

[0007] determining the injury severity difference according to the deviation of the part score vector and all pre-intervention data sequences between each multiple injury patient and every other multiple injury patient; and determining the injury severity attention degree according to the similarity of each multiple injury patient and the multiple injury patients under the same node in each physiological monitoring dimension after hierarchical clustering according to the injury severity difference;

[0008] determine an intervention score of each multiple-injury patient on the corresponding intervention scheme according to the degree of attention and the time-series data change of each physiological monitoring dimension of each multiple-injury patient in the historical data before and after the implementation of the corresponding intervention scheme;

[0009] determine a predicted intervention score of the patient to be scored on each intervention scheme according to the injury difference between the patient to be scored and each multiple-injury patient in the historical data and the intervention score.

[0010] Further, the injury difference acquisition process comprises:

[0011] take each multiple-injury patient as a target patient in turn, and take each multiple-injury patient other than the target patient as a comparison patient;

[0012] determine the corresponding part score difference according to the Euclidean distance between the part score vector of the target patient and the part score vector of the comparison patient;

[0013] determine the corresponding physiological data difference according to the sequence deviation between the pre-intervention data sequence of the target patient and the pre-intervention data sequence of the comparison patient;

[0014] determine the injury difference between the target patient and the comparison patient according to the product between the positive correlation mapping value of the physiological data difference and the part score difference.

[0015] Further, the physiological data difference acquisition process comprises:

[0016] In each physiological monitoring dimension, calculate the DTW distance between the pre-intervention data sequence of the target patient and the pre-intervention data sequence of the comparison patient through the dynamic time warping algorithm; and take the cumulative value of the corresponding DTW distance in all physiological monitoring dimensions as the physiological data difference between the target patient and the comparison patient.

[0017] Further, the injury difference acquisition process comprises:

[0018] cluster all multiple-injury patients according to the injury difference through the BIRCH clustering algorithm to obtain a CF clustering tree; and take the CF tree layer where the sample corresponding to each multiple-injury patient is located as the corresponding analysis layer in the CF clustering tree;

[0019] take a preset number of CF tree layers with a layer number less than and adjacent to the layer number of the analysis layer as the corresponding comparison layer; and take each multiple-injury patient and other multiple-injury patients belonging to the same node in all corresponding comparison layers as the reference patients of each multiple-injury patient;

[0020] The mean value of all data in the pre-intervention data sequence of each physiological monitoring dimension of each multiple trauma patient is taken as the corresponding pre-intervention monitoring characteristic value; the pre-intervention fluctuation degree is determined according to the distribution dispersion of all pre-intervention monitoring characteristic values corresponding to each multiple trauma patient and all corresponding reference patients; the pre-intervention monitoring deviation value of each multiple trauma patient is determined according to the deviation of the pre-intervention monitoring characteristic value of each multiple trauma patient relative to the preset standard physiological monitoring numerical range under each physiological monitoring dimension.

[0021] The local attention degree of each reference patient is determined according to the product between the negative correlation mapping value of the pre-intervention fluctuation degree of each reference patient and the pre-intervention monitoring deviation value; the trauma attention degree of each multiple trauma patient is determined according to the cumulative value of all local attention degrees corresponding to all reference patients of each multiple trauma patient.

[0022] Further, the pre-intervention fluctuation degree acquisition process comprises:

[0023] The pre-intervention fluctuation degree is determined according to the standard deviation of all pre-intervention monitoring characteristic values corresponding to each multiple trauma patient and all corresponding reference patients under each physiological monitoring dimension.

[0024] Further, the pre-intervention monitoring deviation value acquisition process comprises:

[0025] When the pre-intervention monitoring characteristic value is less than or equal to the maximum value of the preset standard physiological monitoring numerical range and greater than or equal to the minimum value of the preset standard physiological monitoring numerical range, the corresponding pre-intervention monitoring deviation value is set to 0;

[0026] When the pre-intervention monitoring characteristic value is greater than the maximum value of the preset standard physiological monitoring numerical range, the difference between the pre-intervention monitoring characteristic value and the maximum value of the preset standard physiological monitoring numerical range is taken as the pre-intervention monitoring deviation value;

[0027] When the pre-intervention monitoring characteristic value is less than the minimum value of the preset standard physiological monitoring numerical range, the difference between the minimum value of the preset standard physiological monitoring numerical range and the pre-intervention monitoring characteristic value is taken as the pre-intervention monitoring deviation value.

[0028] Further, the intervention score acquisition process comprises:

[0029] Based on the principle of obtaining the pre-intervention monitoring bias value according to the pre-intervention data sequence, the post-intervention monitoring bias value of each multiple trauma patient is determined according to the post-intervention data sequence of each multiple trauma patient in each physiological monitoring dimension; based on the principle of obtaining the pre-intervention fluctuation degree according to the pre-intervention data sequence, the post-intervention fluctuation degree of each multiple trauma patient is determined according to the post-intervention data sequence of each multiple trauma patient in each physiological monitoring dimension;

[0030] The difference between the pre-intervention fluctuation degree and the post-intervention fluctuation degree is normalized to determine the corresponding stability growth value;

[0031] Under each physiological monitoring dimension, the difference between the pre-intervention monitoring bias value and the post-intervention monitoring bias value is normalized to determine the intervention evaluation value of each multiple trauma patient after the implementation of the corresponding intervention scheme;

[0032] The intervention dimension score of each multiple trauma patient under each physiological monitoring dimension is determined by combining the intervention evaluation value, the stability growth value and the injury attention degree;

[0033] According to the cumulative value of the intervention dimension scores under all physiological monitoring dimensions, the intervention score of each multiple trauma patient in the historical data under the corresponding intervention scheme is determined.

[0034] Further, the intervention dimension score acquisition process includes:

[0035] According to the product of the intervention evaluation value, the stability growth value and the injury attention degree, the intervention dimension score of each multiple trauma patient under each physiological monitoring dimension is determined.

[0036] Further, the prediction intervention score acquisition process includes:

[0037] Each multiple trauma patient in the historical data is taken as a historical patient; the intervention scheme corresponding to each historical patient is taken as a historical scheme; the injury difference between the patient to be scored and the historical patient is taken as a historical difference;

[0038] The product of the negative correlation mapping value of the historical difference and the intervention score of the historical patient under the corresponding intervention scheme is normalized to determine the historical evaluation score between the patient to be scored and the historical patient;

[0039] According to the cumulative value of all historical evaluation scores corresponding to all historical patients and the patient to be scored under each intervention scheme, the prediction intervention score of the patient to be scored under each intervention scheme is determined.

[0040] Further, the part score vector acquisition process includes:

[0041] According to the medical database, the AIS score of each injury area of each multiple injury patient is obtained by means of AIS coding; wherein the injury area includes head and neck, face, chest, abdomen, pelvic cavity, limbs and skin;

[0042] According to the head and neck, face, chest, abdomen, pelvic cavity, limbs and skin, all injury areas of each multiple injury patient are arranged in turn to obtain a corresponding part score vector.

[0043] In a second aspect, the present application provides a system for prognosis evaluation and early intervention of multiple injury patients, the system comprising:

[0044] A data acquisition and preprocessing module is configured to obtain, in a medical database, a part score vector of each multiple injury patient and a pre-intervention data sequence and a post-intervention data sequence of each physiological monitoring dimension;

[0045] A first determination module is configured to determine a corresponding injury difference according to the deviation of the part score vector and all pre-intervention data sequences between each multiple injury patient and each other multiple injury patient, and to determine a corresponding injury attention degree according to the similarity of each multiple injury patient and multiple injury patients under the same node in each physiological monitoring dimension after hierarchical clustering according to the injury difference;

[0046] A second determination module is configured to determine an intervention score of each multiple injury patient on a corresponding intervention scheme according to the injury attention degree and the time series data change of each physiological monitoring dimension of each multiple injury patient in the historical data before and after implementation of the corresponding intervention scheme;

[0047] An intervention scheme evaluation module is configured to determine a predicted intervention score of a patient to be scored on each intervention scheme according to the injury difference between the patient to be scored and each multiple injury patient in the historical data and the intervention score.

[0048] In a third aspect, the present application provides a computer device comprising a memory and a processor. The memory is configured to store computer program code, and the processor is configured to call and run the computer program code from the memory to execute the method of the first aspect or any embodiment of the first aspect of the present application.

[0049] In a fourth aspect, the present application provides a computer program product comprising computer program code, which, when executed, performs the method of the first aspect or any embodiment of the first aspect of the present application.

[0050] In a fifth aspect, the present application provides a computer readable storage medium storing computer program codes, when the computer program codes are executed, to perform the method of the first aspect or any of the embodiments of the first aspect.

[0051] The present application has the following beneficial effects:

[0052] The present application first calculates the injury condition difference and performs clustering analysis, which can systematically process the multi-dimensional dynamic data of the patient and determine the key attention field of the patient's injury condition. Secondly, by analyzing the injury condition attention degree and the physiological data changes after early intervention, the intervention scheme of each multiple injury patient in the historical data is evaluated to determine the corresponding intervention score. Then, by combining the deviation characteristics represented by the injury condition difference and the intervention score, the predicted intervention score of the patient to be scored is comprehensively represented, so that the accuracy of the obtained predicted intervention score is higher; and without using a deep learning model, the efficiency of obtaining the predicted intervention score is also higher. BRIEF DESCRIPTION OF DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art and the advantages thereof, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained from these drawings without creative labor.

[0054] Figure 1 A flow chart of a method for prognosis evaluation and early intervention of multiple injury patients provided by an embodiment of the present application;

[0055] Figure 2 A structural diagram of a system for prognosis evaluation and early intervention of multiple injury patients provided by an embodiment of the present application;

[0056] Figure 3 A structural diagram of a computer device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0057] In order to further clarify the technical means and effects taken by the present application to achieve the predetermined inventive purpose, the following describes in detail the specific implementation, structure, features and effects of a method for prognosis evaluation and early intervention of multiple trauma patients according to the present application, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment, and the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form. In addition, the terms "first", "second" are only for descriptive purposes, and cannot be understood as implying or suggesting relative importance or implicitly indicating the number of technical features indicated. Therefore, the features with "first", "second" can be explicitly or implicitly included.

[0058] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0059] The following describes in detail the specific scheme of a method for prognosis evaluation and early intervention of multiple trauma patients according to the present application in combination with the accompanying drawings.

[0060] The present application provides a method for prognosis evaluation and early intervention of multiple trauma patients, please refer to Figure 1 which shows a flowchart of a method for prognosis evaluation and early intervention of multiple trauma patients according to an embodiment of the present application, the method comprises:

[0061] Step S101: obtaining the site score vector of each multiple trauma patient and the pre-intervention data sequence and post-intervention data sequence of each physiological monitoring dimension in the medical database.

[0062] The data of multiple trauma patients after detection or physical examination will be stored in the medical database, so the data collection of multiple trauma patients is carried out through the medical database.

[0063] Preferably, in some possible implementation manners of the embodiment of the present application, the obtaining process of the site score vector comprises:

[0064] According to the medical database, the AIS score of each injury area of each multiple trauma patient is obtained by AIS coding; wherein the injury area includes head and neck, face, chest, abdomen, pelvic cavity, limbs and skin; all injury areas of each multiple trauma patient are arranged in turn according to the head and neck, face, chest, abdomen, pelvic cavity, limbs and skin, to obtain the corresponding site score vector. It should be noted that AIS coding is a technical means known to those skilled in the art, which will not be described further herein; and the arrangement order of the injury area in the site score vector can be adjusted according to the specific implementation environment, which will not be described further herein.

[0065] Then, the data of each physiological monitoring dimension is collected. In an implementation form of the embodiment of the present application, the physiological monitoring dimensions include heart rate, respiratory rate, intracranial pressure, systolic pressure, blood oxygen saturation, and Glasgow coma score. In an implementation form of the embodiment of the present application, the heart rate data is collected by an electrocardiogram monitor, and can also be collected by a pulse oximeter. The respiratory rate data is collected by a chest and abdominal belt sensor, and can also be collected by an impedance method. The intracranial pressure data is collected by an intracerebral catheter. The systolic pressure data is collected by an arterial blood pressure monitoring device. The blood oxygen saturation data is collected by a pulse oximeter. The Glasgow coma score data is collected by a clinical assessment method. The sampling frequency of the heart rate data and the systolic pressure data is set to be collected once every 5 seconds. The blood oxygen saturation data and the respiratory rate data are set to be collected once every 1 minute. The intracranial pressure and the Glasgow coma score are collected once every 10 minutes. The length of the collection time period of each multiple trauma patient is set to be 1 hour, which can be adjusted according to the specific implementation environment. In addition, for each multiple trauma patient in the historical data, namely, in the medical database, the data of each physiological monitoring dimension collected in the collection time period before and after the implementation of the corresponding intervention scheme is collected once, respectively. Specifically, after arranging the data of each physiological monitoring dimension collected in the collection time period before the implementation of the intervention scheme in time sequence, a corresponding pre-intervention data sequence is obtained. After arranging the data of each physiological monitoring dimension collected in the collection time period after the implementation of the intervention scheme in time sequence, a corresponding post-intervention data sequence is obtained. It should be noted that the data analyzed in the embodiment of the present application is all data authorized by the user for use, and will not be described further.

[0066] Step S102: determining the injury condition difference according to the deviation of the position score vector between each multiple trauma patient and each other multiple trauma patient and all pre-intervention data sequences. After hierarchical clustering according to the injury condition difference, the injury condition attention degree is determined according to the similarity of each multiple trauma patient and the multiple trauma patients under the same node in each physiological monitoring dimension.

[0067] The position score vector comprehensively represents the trauma severity of each damage site of each multiple trauma patient, and the pre-intervention data sequence represents the physiological monitoring dimension information reflecting the physiological state of the multiple trauma patient before being processed by the intervention scheme. Therefore, for any two multiple trauma patients, the higher the similarity of the corresponding two position score vectors and pre-intervention data sequences, the smaller the injury condition difference should be, that is, the closer the trauma performance of the two multiple trauma patients.

[0068] Preferably, in some possible implementation forms of the embodiment of the present application, the injury condition difference acquisition process comprises:

[0069] Each multiple injury patient is sequentially taken as a target patient, and each multiple injury patient other than the target patient is taken as a comparative patient; by comparing the target patient and the comparative patient, the injury difference between the two multiple injury patients can be obtained.

[0070] According to the Euclidean distance between the part score vector of the target patient and the part score vector of the comparative patient, the corresponding part score difference is determined. The greater the Euclidean distance between the two part score vectors, the greater the difference between the two part score vectors, and thus the greater the part score difference, and the greater the injury difference.

[0071] According to the sequence deviation between the pre-intervention data sequence of the target patient and the pre-intervention data sequence of the comparative patient, the corresponding physiological data difference is determined; in a specific implementation manner of the embodiment, the process of obtaining the physiological data difference includes: under each physiological monitoring dimension, the DTW distance between the pre-intervention data sequence of the target patient and the pre-intervention data sequence of the comparative patient is calculated by using the dynamic time warping algorithm; and the cumulative value of the corresponding DTW distance under all physiological monitoring dimensions is taken as the physiological data difference between the target patient and the comparative patient. According to the dynamic time warping algorithm, the greater the DTW distance between two sequences, the lower the sequence similarity of the two sequences; thus the greater the physiological data difference, the greater the corresponding injury difference. Finally, according to the correlation, the product between the positive correlation mapping value of the physiological data difference and the part score difference is determined as the injury difference between the target patient and the comparative patient.

[0072] In a specific implementation manner of the embodiment, the process of obtaining the injury difference is represented by a formula as follows: wherein, is the injury difference between the target patient and the corresponding comparative patient; is the part score vector of the target patient; is the part score vector of the corresponding comparative patient; is the physiological data difference between the target patient and the corresponding comparative patient. is the Euclidean distance between the part score vector of the target patient and the part score vector of the corresponding comparative patient, that is, the part score difference; is a linear normalization function; is the number of physiological monitoring dimensions; is the DTW distance between the pre-intervention data sequence of the target patient and the pre-intervention data sequence of the corresponding comparative patient under the i-th physiological monitoring dimension; is the physiological data difference between the target patient and the corresponding comparative patient. ​​​​​​​

[0073] Since there are many different regional injury conditions in multiple injury patients, and the attention degree of each physiological monitoring data dimension is different for different positions, the application comprehensively analyzes the overall data performance of each multiple injury patient and the physiological monitoring dimension data between each multiple injury patient with similar disease performance, thereby avoiding the deviation of the injury attention degree caused by single analysis; therefore, further according to the physiological data difference, the clustering analysis algorithm is processed, thereby analyzing the injury attention degree of patients with similar injuries.

[0074] Preferably, in some possible implementation manners of the embodiments of the application, the injury attention degree acquisition process comprises:

[0075] According to the BIRCH clustering algorithm passed through the injury difference between all multiple injury patients, the CF clustering tree is obtained; in the CF clustering tree, the CF tree layer where each multiple injury patient's corresponding sample is located is taken as the corresponding analysis layer; the preset number of CF tree layers with the number of layers less than the number of layers of the analysis layer and adjacent to the number of layers of the analysis layer are taken as the corresponding comparison layer; each multiple injury patient and other multiple injury patients belonging to the same node in the corresponding all comparison layers are taken as the reference patients of each multiple injury patient. In one specific implementation manner of the embodiments of the application, the preset number is set to 3, which can be adjusted according to the specific implementation environment. For example, the CF tree layer where the sample corresponding to a certain multiple injury patient is located is the 5th layer, the analysis layer is the 5th layer, and the comparison layers are the 4th layer, the 3rd layer and the 2nd layer. The greater the difference between the number of layers of the comparison layer and the number of layers of the analysis layer, the greater the difference between the corresponding multiple injury patients, therefore, the application selects the preset number as 3, so that each multiple injury patient and the corresponding reference patient present similar injury characteristics, thereby analyzing the overall injury characteristics of all reference patients to comprehensively analyze the injury attention degree of the corresponding multiple injury patient.

[0076] The mean of all data in the data sequence of each polytrauma patient in each physiological monitoring dimension before the intervention is taken as the corresponding pre-intervention monitoring characteristic value; the pre-intervention fluctuation degree is determined according to the distribution dispersion of all pre-intervention monitoring characteristic values corresponding to each polytrauma patient and all corresponding reference patients; preferably, in some possible implementation manners of the embodiment of the present application, the pre-intervention fluctuation degree acquisition process comprises: under each physiological monitoring dimension, the pre-intervention fluctuation degree is determined according to the standard deviation of all pre-intervention monitoring characteristic values corresponding to each polytrauma patient and all corresponding reference patients. The greater the standard deviation of all pre-intervention monitoring characteristic values corresponding to all reference patients, the more dispersed the distribution of the pre-intervention monitoring characteristic values, and the lower the credibility of the corresponding deviation degree; and the data of different physiological monitoring dimensions are normalized by the standard deviation, the dimension is eliminated to standardize the data, and the robustness of the obtained injury condition attention degree is higher.

[0077] Under each physiological monitoring dimension, the pre-intervention monitoring deviation value of each polytrauma patient is determined according to the deviation of the pre-intervention monitoring characteristic value of each polytrauma patient relative to the preset standard physiological monitoring numerical range. Preferably, in one specific implementation manner of the embodiment of the present application, the pre-intervention monitoring deviation value acquisition process comprises:

[0078] When the pre-intervention monitoring characteristic value is less than or equal to the maximum value of the preset standard physiological monitoring numerical range and greater than or equal to the minimum value of the preset standard physiological monitoring numerical range, the corresponding pre-intervention monitoring deviation value is set to 0; when the pre-intervention monitoring characteristic value is greater than the maximum value of the preset standard physiological monitoring numerical range, the difference between the pre-intervention monitoring characteristic value and the maximum value of the preset standard physiological monitoring numerical range is taken as the pre-intervention monitoring deviation value; when the pre-intervention monitoring characteristic value is less than the minimum value of the preset standard physiological monitoring numerical range, the difference between the minimum value of the preset standard physiological monitoring numerical range and the pre-intervention monitoring characteristic value is taken as the pre-intervention monitoring deviation value. That is, for each polytrauma patient, the greater the deviation of the data of the corresponding physiological monitoring dimension compared with the standard range, the greater the corresponding pre-intervention monitoring deviation value, that is, the greater the deviation from the normal condition, and the greater the injury condition attention degree based on it.

[0079] The local attention degree of each reference patient is determined according to the product between the negative correlation mapping value of the pre-intervention fluctuation degree of each reference patient and the pre-intervention monitoring deviation value; the injury condition attention degree of each polytrauma patient is determined according to the cumulative value of all local attention degrees corresponding to all reference patients of each polytrauma patient. In one specific implementation manner of the embodiment of the present application, the injury condition attention degree acquisition process is expressed by a formula as: ; wherein, is the injury condition attention degree of the i th polytrauma patient in the j th physiological monitoring dimension. is the pre-intervention fluctuation degree of the i th polytrauma patient in the j th physiological monitoring dimension. The level of attention paid to injuries under each physiological monitoring dimension; For the first The total number of all reference patients corresponding to each patient with multiple injuries; For the first The first physiological monitoring dimension The first multiple injury patient corresponds to the first Pre-intervention monitoring deviation values ​​for one reference patient; For the first The first physiological monitoring dimension The degree of fluctuation before intervention for patients with multiple injuries; it should be noted that, in order to ensure that the calculation results are meaningful, when performing fractional operations in this embodiment of the invention, if the denominator is 0, a parameter adjustment factor greater than 0 needs to be added to the denominator to prevent the denominator from being 0. The value of the parameter adjustment factor shall be set by the implementer according to the actual situation, and this application does not impose any special restrictions.

[0080] It should be noted that the preset standard physiological monitoring value ranges for each physiological monitoring dimension can be obtained based on the normal range values ​​of the corresponding physiological monitoring dimension. In a specific implementation of this invention, the preset standard physiological monitoring value range for heart rate is set to 60-100 beats per minute; the preset standard physiological monitoring value range for respiratory rate is set to 12-20 beats per minute; the preset standard physiological monitoring value range for intracranial pressure is set to 5-15 mmHg; the preset standard physiological monitoring value range for systolic blood pressure is set to 90-120 mmHg; the preset standard physiological monitoring value range for blood oxygen saturation is set to 95%-100%; and the GCS score is set to 10-15 points. These can be adjusted according to the specific implementation environment, and will not be elaborated further here. The degree of injury concern is obtained by analyzing the deviation characteristics of the physiological monitoring dimension data of reference patients with similar characteristics for each multiple injury patient relative to the standard data value range. Compared with the method of analyzing the standard deviation characteristics of the physiological monitoring data of each multiple injury patient alone, this method can greatly avoid the influence of random factors, making the obtained degree of injury concern more accurate.

[0081] Step S103: Based on the level of concern for the injury and the time-series data changes of each physiological monitoring dimension before and after the implementation of the corresponding intervention plan for each multiple injury patient in historical data, determine the intervention score for each multiple injury patient in the corresponding intervention plan.

[0082] For each patient with multiple injuries in the historical data, the more normal their physiological monitoring data are after the corresponding intervention program, the better the intervention effect is. Therefore, the corresponding intervention program is scored based on the changes in physiological monitoring data before and after the implementation of the intervention program for each patient with multiple injuries in the historical data.

[0083] Preferably, in some possible implementation manners of the embodiments of the present application, the intervention score acquisition process comprises:

[0084] Based on the principle of obtaining the pre-intervention monitoring deviation value according to the pre-intervention data sequence, the post-intervention monitoring deviation value of each multiple trauma patient is determined according to the post-intervention data sequence of each physiological monitoring dimension of each multiple trauma patient; based on the principle of obtaining the pre-intervention fluctuation degree according to the pre-intervention data sequence, the post-intervention fluctuation degree of each multiple trauma patient is determined according to the post-intervention data sequence of each physiological monitoring dimension of each multiple trauma patient. That is, after replacing the basic data pre-intervention data sequence in the acquisition process of the pre-intervention monitoring deviation value and the pre-intervention fluctuation degree with the post-intervention data sequence, the post-intervention monitoring deviation value and the post-intervention fluctuation degree of each multiple trauma patient are determined by the same calculation method, and the specific calculation process is not further described here.

[0085] The difference between the pre-intervention fluctuation degree and the post-intervention fluctuation degree is normalized to determine the corresponding stability growth value. First, the data of each physiological monitoring dimension under normal body state is usually smooth, and trauma can cause the data to present high fluctuation, so the greater the difference between the pre-intervention fluctuation degree and the post-intervention fluctuation degree, the more stable the post-intervention is, and the better the intervention stability effect is; therefore, the greater the stability growth value, the greater the intervention score should be.

[0086] Under each physiological monitoring dimension, the difference between the pre-intervention monitoring deviation value and the post-intervention monitoring deviation value is normalized to determine the intervention evaluation value of each multiple trauma patient after the implementation of the corresponding intervention scheme. Similarly, the pre-intervention monitoring deviation value under normal body state is usually 0, and the closer to 0, the more likely the body state is normal; therefore, the greater the difference between the pre-intervention monitoring deviation value and the post-intervention monitoring deviation value, the more normal the body of the multiple trauma patient after the intervention is, and the better the intervention effect is; therefore, the greater the intervention evaluation value, the greater the intervention score should be.

[0087] The level of concern for injuries represents the degree of attention each multiple injury patient pays to each physiological monitoring dimension. A higher level of concern necessitates greater attention to the corresponding physiological monitoring dimension, thus requiring a higher weight when scoring that dimension. Further, by combining the intervention evaluation value, stability growth value, and level of concern for injuries, the intervention dimension score for each multiple injury patient under each physiological monitoring dimension is comprehensively determined. The process of obtaining the intervention dimension score involves multiplying the intervention evaluation value, stability growth value, and level of concern for injuries to determine the intervention dimension score for each multiple injury patient under each physiological monitoring dimension. Further, by summing the intervention dimension scores across all physiological monitoring dimensions, the intervention score for each multiple injury patient under the corresponding intervention plan in historical data is determined. By combining all physiological monitoring dimensions, the intervention effect of each multiple injury patient after implementing the corresponding intervention plan is evaluated, resulting in an intervention score that more comprehensively represents the intervention effect.

[0088] In one specific implementation of this invention, the process of obtaining the intervention score is expressed by the following formula: ;in, For the first Intervention scores of multiple injury patients in the corresponding intervention program; The number of physiological monitoring dimensions; For the first The first physiological monitoring dimension Pre-intervention monitoring deviation values ​​for patients with multiple injuries; For the first The first physiological monitoring dimension Post-intervention monitoring deviation values ​​for multiple trauma patients; For the first The first physiological monitoring dimension The degree of fluctuation in the number of patients with multiple injuries before intervention; For the first The first physiological monitoring dimension The degree of fluctuation after intervention in patients with multiple injuries; For the first The multiple injury patient in the first The level of attention paid to injuries under each physiological monitoring dimension; It is a linear normalization function; the normalization process ensures that the intervention evaluation value and stability increment do not take negative values, thus avoiding the impact on the calculation.

[0089] Step S104: Based on the differences in injury severity between the patient to be scored and the patients with multiple injuries in the historical data, as well as the intervention scores, determine the predicted intervention score for each intervention program for the patient to be scored.

[0090] The to-be-scored patient is a polytrauma patient who has not been intervened by an intervention scheme, so for the to-be-scored patient, the determination of the predicted intervention score of various intervention schemes is needed; for the to-be-scored patient, only the pre-intervention data sequence of each physiological monitoring dimension is available, so the injury difference between the to-be-scored patient and each polytrauma patient in the historical data can be calculated; the smaller the injury difference is, the closer the trauma performance between the to-be-scored patient and the corresponding polytrauma patient is, so the intervention result, that is, the intervention score of the corresponding polytrauma patient after intervention of the intervention scheme, is more referential to the to-be-scored patient; therefore, according to this feature, the intervention score of the corresponding intervention scheme is weighted by taking the injury difference between the to-be-scored patient and each polytrauma patient as the weight, and then the overall size of the intervention score of each intervention scheme after weighting is combined from the overall dimension to determine the more accurate predicted intervention score of the to-be-scored patient on each intervention scheme.

[0091] Preferably, in some possible implementation manners of the embodiment of the present application, the acquisition process of the predicted intervention score includes:

[0092] Each polytrauma patient in the historical data is taken as a historical patient; the intervention scheme corresponding to each historical patient is taken as a historical scheme; the injury difference between the to-be-scored patient and the historical patient is taken as a historical difference; the product of the negative correlation mapping value of the historical difference and the intervention score of the historical patient on the corresponding intervention scheme is normalized to determine the historical evaluation score between the to-be-scored patient and the historical patient.

[0093] Because the smaller the injury difference is, the more referential the intervention score of the corresponding historical patient on the corresponding historical scheme is, so the intervention score of the historical patient on the corresponding intervention scheme is weighted by taking the negative correlation mapping value of the historical difference as the weight, thereby obtaining the historical evaluation score of the to-be-scored patient on the intervention scheme of the corresponding historical patient that is more referential; because the types of intervention schemes are limited, each intervention scheme usually corresponds to multiple historical patients, so further, in order to analyze the predicted intervention score of the to-be-scored patient on each intervention scheme, after the historical evaluation score between the to-be-scored patient and each historical patient is calculated, the predicted intervention score of the to-be-scored patient on each intervention scheme is determined according to the cumulative value of all historical evaluation scores between all historical patients under each intervention scheme and the to-be-scored patient; so that the obtained predicted intervention score is more accurate.

[0094] In one specific implementation manner of the embodiment of the present application, the acquisition process of the predicted intervention score includes: ; wherein, is the predicted intervention score of the to-be-scored patient on the first intervention scheme; is the historical data of the first a number of multiple injury patients under an intervention scheme; for a patient to be scored a difference in injury between the first a number of multiple injury patients under the first a difference in injury between the first is a Softmax normalization function, and normalization by the Softmax normalization function can make the sum of all normalized values 1, so after negative correlation mapping and weighting, the integration is performed by the accumulation method, so as to obtain a more accurate prediction intervention score; a difference in injury between the first a number of multiple injury patients under the first a number of multiple injury patients under the first

[0095] Further, according to the acquisition process of the prediction intervention score, the prediction intervention score of each patient to be scored under each intervention scheme is determined, so as to evaluate and quantify the intervention scheme; in a specific implementation manner of an embodiment of the present application, all kinds of intervention schemes are arranged and stored in the medical database in the order of the prediction intervention score corresponding to each patient to be scored from large to small, so as to more intuitively assist the selection of the intervention scheme.

[0096] In summary, a method for prognosis evaluation and early intervention of multiple injury patients first calculates the injury difference and performs cluster analysis, can systematically process the multi-dimensional dynamic data of the patient, and determine the key attention field of the patient's injury. Secondly, by analyzing the injury attention degree and the physiological data change after early intervention, the intervention scheme of each multiple injury patient in the historical data is evaluated, and the corresponding intervention score is determined. Then, the deviation feature represented by the injury difference is combined with the intervention score, so as to comprehensively represent the prediction intervention score of the patient to be scored, so that the accuracy of the obtained prediction intervention score is higher; and without using a deep learning model, the efficiency of obtaining the prediction intervention score is also higher.

[0097] The present application also provides a system for prognosis evaluation and early intervention of multiple injury patients, please refer to Figure 2 , which shows a structure diagram of a system for prognosis evaluation and early intervention of multiple injury patients provided by an embodiment of the present application, the system comprises a data acquisition and preprocessing module 201, a first determination module 202, a second determination module 203 and an intervention scheme evaluation module 204.

[0098] The data acquisition and preprocessing module 201 is used for acquiring the part score vector of each multiple injury patient and the pre-intervention data sequence and post-intervention data sequence of each physiological monitoring dimension in the medical database;

[0099] The first determining module 202 is configured to determine a corresponding injury difference according to a bias condition corresponding to all pre-intervention data sequences of a position score vector between each polytrauma patient and each other polytrauma patient; and determine a corresponding injury attention degree according to a similarity condition in each physiological monitoring dimension between each polytrauma patient and the polytrauma patients under the same node after hierarchical clustering according to the injury difference.

[0100] The second determining module 203 is configured to determine an intervention score of each polytrauma patient on the corresponding intervention scheme according to the injury attention degree and a time sequence data change condition of each physiological monitoring dimension of each polytrauma patient in the historical data before and after implementation of the corresponding intervention scheme.

[0101] The intervention scheme evaluation module 204 is configured to determine a predicted intervention score of the patient to be scored on each intervention scheme according to the injury difference and the intervention score between the patient to be scored and each polytrauma patient in the historical data.

[0102] It should be noted that the system provided in the above embodiments is only exemplified by the division of the above functional modules, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the functions described above. In addition, the system for polytrauma patient prognosis evaluation and early intervention and the method for polytrauma patient prognosis evaluation and early intervention provided in the above embodiments belong to the same concept, and the specific implementation process is detailed in the method embodiments, which will not be described here.

[0103] The embodiments of the present application also provide a computer device, please refer to Figure 3 which shows a computer device structure schematic diagram provided by an embodiment of the present application, the computer device includes a memory 301, a processor 302 and a computer program 303 stored in the memory 301 and running on the processor 302, wherein the processor 302 executes the computer program 303, so that the computer device can execute any one of the above-mentioned polytrauma patient prognosis evaluation and early intervention methods.

[0104] The embodiments of the present application also provide a computer program product, when the computer program product runs on the computer device, so that the computer device can execute any one of the above-mentioned polytrauma patient prognosis evaluation and early intervention methods.

[0105] The embodiments of the present application also provide a computer readable storage medium, the computer readable storage medium stores computer program code, when the computer program code runs on the computer device, so that the computer device can execute any one of the above-mentioned polytrauma patient prognosis evaluation and early intervention methods.

[0106] In the embodiments provided in the present application, it should be understood that the computer device, computer program product and computer readable storage medium provided are all used to execute the corresponding method provided in the above, and thus the beneficial effects that can be achieved can refer to the beneficial effects of the method provided in the above, which will not be described here.

[0107] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0108] Each of the embodiments in the specification is described in a progressive manner, and the same and similar parts between the embodiments can be referred to each other, and each embodiment mainly describes the difference from other embodiments.

Claims

1. A method for prognosis assessment and early intervention of multiple trauma patients, characterized in that, The method comprises: According to the medical database, the AIS score of each injury area of each multiple trauma patient is obtained by means of AIS coding; wherein the injury area includes head and neck, face, chest, abdomen, pelvic cavity, limbs and skin; According to the head and neck, face, chest, abdomen, pelvic cavity, limbs and skin, all injury areas of each multiple trauma patient are arranged in turn to obtain the corresponding part score vector; the pre-intervention data sequence and the post-intervention data sequence of each physiological monitoring dimension are obtained in the medical database; Each multiple trauma patient is sequentially taken as a target patient, and each multiple trauma patient other than the target patient is taken as a contrast patient; the part score difference between the target patient and the contrast patient is determined according to the Euclidean distance between the part score vector of the target patient and the part score vector of the contrast patient; the physiological data difference between the target patient and the contrast patient is determined according to the sequence deviation between the pre-intervention data sequence of the target patient and the pre-intervention data sequence of the contrast patient; the injury condition difference between the target patient and the contrast patient is determined according to the product between the positive correlation mapping value of the physiological data difference and the part score difference; after hierarchical clustering is performed according to the injury condition difference, the injury condition attention degree is determined according to the similarity between each multiple trauma patient and the multiple trauma patients under the same node in each physiological monitoring dimension; According to the injury condition attention degree and the time series data change of each physiological monitoring dimension of each multiple trauma patient before and after the implementation of the corresponding intervention scheme in the historical data, the intervention score of each multiple trauma patient on the corresponding intervention scheme is determined; According to the injury condition difference between the patient to be scored and each multiple trauma patient in the historical data and the intervention score, the predicted intervention score of the patient to be scored on each intervention scheme is determined; The injury condition attention degree is obtained by: According to the injury condition difference between all multiple trauma patients, BIRCH clustering algorithm is used for clustering to obtain a CF clustering tree; in the CF clustering tree, the CF tree layer where the sample corresponding to each multiple trauma patient is located is taken as the corresponding analysis layer; A preset number of CF tree layers with a layer number less than the layer number of the analysis layer and adjacent to the layer number of the analysis layer are taken as the corresponding contrast layer; each multiple trauma patient and other multiple trauma patients belonging to the same node in all corresponding contrast layers are taken as the reference patient of each multiple trauma patient; The mean value of all data in the pre-intervention data sequence of each physiological monitoring dimension of each multiple trauma patient is taken as the corresponding pre-intervention monitoring characteristic value; the pre-intervention fluctuation degree is determined according to the distribution dispersion of all pre-intervention monitoring characteristic values corresponding to each multiple trauma patient and all reference patients; in each physiological monitoring dimension, the pre-intervention monitoring deviation value of each multiple trauma patient is determined according to the deviation of the pre-intervention monitoring characteristic value of each multiple trauma patient relative to the preset standard physiological monitoring numerical range. determine a local attention degree of each reference patient according to a product between a negative correlation mapping value of the intervention-before fluctuation degree of each reference patient and the intervention-before monitoring deviation value; and determine an injury attention degree of each polytrauma patient according to an accumulated value of all local attention degrees corresponding to all reference patients of each polytrauma patient.

2. A method for prognosis and early intervention of multiple trauma patients as claimed in claim 1 wherein, The process of obtaining the physiological data difference includes: In each physiological monitoring dimension, a DTW distance between the intervention-before data sequence of the target patient and the intervention-before data sequence of the contrast patient is calculated by using a dynamic time warping algorithm; and an accumulated value of the corresponding DTW distances in all physiological monitoring dimensions is taken as the physiological data difference between the target patient and the contrast patient.

3. A method for prognosis and early intervention of multiple trauma patients as claimed in claim 1 wherein, The process of obtaining the intervention-before fluctuation degree includes: In each physiological monitoring dimension, an intervention-before fluctuation degree is determined according to a standard deviation of all intervention-before monitoring characteristic values corresponding to each polytrauma patient and all corresponding reference patients.

4. The method for prognosis and early intervention of multiple trauma patients according to claim 1, wherein, The process of obtaining the intervention-before monitoring deviation value includes: when the intervention-before monitoring characteristic value is less than or equal to the maximum value of the preset standard physiological monitoring numerical range and greater than or equal to the minimum value of the preset standard physiological monitoring numerical range, the corresponding intervention-before monitoring deviation value is set to 0; when the intervention-before monitoring characteristic value is greater than the maximum value of the preset standard physiological monitoring numerical range, a difference value between the intervention-before monitoring characteristic value and the maximum value of the preset standard physiological monitoring numerical range is taken as the intervention-before monitoring deviation value; when the intervention-before monitoring characteristic value is less than the minimum value of the preset standard physiological monitoring numerical range, a difference value between the minimum value of the preset standard physiological monitoring numerical range and the intervention-before monitoring characteristic value is taken as the intervention-before monitoring deviation value.

5. The method for prognosis and early intervention of multiple trauma patients according to claim 1, wherein, The process of obtaining the intervention score includes: based on the principle of obtaining the intervention-before monitoring deviation value according to the intervention-before data sequence, an intervention-after monitoring deviation value of each polytrauma patient is determined according to the intervention-after data sequence of each polytrauma patient in each physiological monitoring dimension; and based on the principle of obtaining the intervention-before fluctuation degree according to the intervention-before data sequence, an intervention-after fluctuation degree of each polytrauma patient is determined according to the intervention-after data sequence of each polytrauma patient in each physiological monitoring dimension; a difference value between the intervention-before fluctuation degree and the intervention-after fluctuation degree is normalized to determine a corresponding stability growth value; in each physiological monitoring dimension, a difference value between the intervention-before monitoring deviation value and the intervention-after monitoring deviation value is normalized to determine an intervention evaluation value of each polytrauma patient after implementation of the corresponding intervention scheme; combined with the intervention evaluation value, the stability growth value and the injury attention degree, an intervention dimension score of each polytrauma patient in each physiological monitoring dimension is comprehensively determined; an intervention score of each polytrauma patient on the corresponding intervention scheme in the historical data is determined according to an accumulated value of the intervention dimension scores in all physiological monitoring dimensions.

6. A method for prognosis and early intervention of multiple trauma patients as claimed in claim 5 wherein, The process of obtaining the intervention dimension score includes: According to the product of the intervention evaluation value, the stability growth value and the injury concern degree, a dimension intervention score of each multiple injury patient in each physiological monitoring dimension is determined.

7. The method for prognosis and early intervention of multiple trauma patients according to claim 1, wherein, The obtaining process of the predicted intervention score comprises: Each multiple injury patient in the historical data is taken as a historical patient; the intervention scheme corresponding to each historical patient is taken as a historical scheme; the injury difference between the patient to be scored and the historical patient is taken as historical difference; The product of the negative correlation mapping value of the historical difference and the intervention score of the historical patient on the corresponding intervention scheme is normalized to determine the historical evaluation score between the patient to be scored and the historical patient; According to the cumulative value of all the historical evaluation scores corresponding to all the historical patients and the patient to be scored under each intervention scheme, the predicted intervention score of the patient to be scored on each intervention scheme is determined.

Images