Artificial intelligence-based early warning method for sepsis of multi-injury patient

Abstract

The invention relates to the technical field of physiological data processing, in particular to an early warning method for sepsis of a multi-injury patient based on artificial intelligence. According to the monitoring data distribution of each index of each multi-injury patient at different monitoring moments, the high risk of each multi-injury patient at each index is obtained, and the high sensitivity of the multi-injury patient with sepsis at each index is obtained; according to the monitoring data distribution of each index of each multi-injury patient at different monitoring moments and the difference between adjacent monitoring moments, the dynamic change rate of each index of each multi-injury patient at each monitoring moment is obtained, and the sepsis prediction value of each index of each multi-injury patient is obtained; and then a sepsis risk score of each multi-injury patient is obtained, and sepsis early warning is carried out on the multi-injury patients. According to the method, the sepsis risk condition of each multi-injury patient is accurately evaluated, so that the accuracy of early warning is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of physiological data processing, and particularly relates to an early warning method for sepsis of multiple-injury patients based on artificial intelligence. BACKGROUND

[0002] Multiple-injury patients are prone to develop sepsis of systemic inflammatory response syndrome due to severe tissue damage, immune system disorder and increased risk of infection, and the disease progresses rapidly with high mortality. Early identification and appropriate treatment can improve the prognosis of sepsis patients. Therefore, it is usually necessary to comprehensively analyze the physiological state of the patient based on various physiological indicators of the multiple-injury patient combined with real-time monitoring data to achieve early warning of sepsis.

[0003] In the prior art, regression analysis is usually performed on physiological indicators and sepsis incidence to screen risk factor indicators of multiple-injury patients complicated with sepsis as the basis to establish a prediction model for sepsis. However, since the multiple-injury patients themselves can cause systemic traumatic inflammatory response due to tissue damage, the early manifestations of sepsis infection and inflammation are highly overlapped, and the related physiological indicators are also changed, resulting in insufficient specificity of the early warning and being unable to distinguish between traumatic inflammation and sepsis inflammation, thus causing errors in the judgment of sepsis occurrence in multiple-injury patients. SUMMARY

[0004] In order to solve the technical problem that the multiple-injury patients themselves can cause systemic traumatic inflammatory response due to tissue damage, the early manifestations of sepsis infection and inflammation are highly overlapped, and the judgment of sepsis occurrence in multiple-injury patients is erroneous, the purpose of the present application is to provide an early warning method for sepsis of multiple-injury patients based on artificial intelligence, and the technical solution adopted is as follows: The present application provides an early warning method for sepsis of multiple-injury patients based on artificial intelligence, which comprises: obtaining monitoring data of multiple physiological indicators of multiple-injury patients with sepsis and non-sepsis at each monitoring time; obtaining baseline deviation of each indicator of each multiple-injury patient at each monitoring time according to the monitoring data distribution of each indicator of each multiple-injury patient at different monitoring times, and obtaining high risk of each indicator of each multiple-injury patient; obtaining high sensitivity of each indicator of multiple-injury patients with sepsis according to the change of high risk of each indicator of different multiple-injury patients; obtaining dynamic change rate of each indicator of each multiple-injury patient at each monitoring time according to the monitoring data distribution of each indicator of each multiple-injury patient at different monitoring times and the difference between adjacent monitoring times; and obtaining sepsis prediction value of each indicator of each multiple-injury patient according to the dynamic change rate and baseline deviation of each indicator of different multiple-injury patients at all monitoring times, and the high sensitivity of each indicator of multiple-injury patients with sepsis. According to the sepsis prediction value and high risk of each multiple injury patient in different indicators, a sepsis risk score of each multiple injury patient is obtained, and early warning of sepsis of the multiple injury patient is performed.

[0005] Further, the method for obtaining the baseline deviation degree comprises: For any monitoring time, the monitoring data of each multiple injury patient in each indicator is compared with the standard reference range, if the monitoring data is out of the standard reference range, the baseline deviation degree of each indicator is set to 0; If the monitoring data is out of the standard reference range, the minimum value of the difference between the monitoring data and the different boundary data in the standard reference range is obtained, the boundary data corresponding to the minimum value is taken as the reference data, and the ratio of the minimum value and the reference data is taken as the baseline deviation degree of each multiple injury patient in each indicator.

[0006] Further, the method for obtaining the high risk comprises: The average value of the difference between the baseline deviation degrees of each multiple injury patient in each indicator at different monitoring times and the earliest monitoring time is obtained as the high risk of each multiple injury patient in each indicator.

[0007] Further, the method for obtaining the high sensitivity comprises: The cumulative sum of the high risk of all sepsis multiple injury patients in each indicator is obtained as the sepsis high risk level; The cumulative sum of the high risk of all multiple injury patients in each indicator is obtained as the overall high risk level; The ratio between the sepsis high risk level and the overall high risk level is obtained as the high sensitivity of the multiple injury patient with sepsis in each indicator.

[0008] Further, the method for obtaining the dynamic change rate comprises: The difference between the monitoring data of each multiple injury patient in each indicator at each monitoring time and the previous monitoring time is obtained, and the ratio between the monitoring data difference and the difference between the corresponding monitoring times is calculated as the dynamic change rate of each multiple injury patient in each indicator at each monitoring time.

[0009] Further, the method for obtaining the sepsis prediction value comprises: According to the dynamic change rate and the baseline deviation degree of each multiple injury patient in each indicator at all monitoring times, the pathological disorder degree of each multiple injury patient in each indicator is obtained; The average value of the pathological disorder degree of all sepsis multiple injury patients in each indicator is obtained as the overall pathological disorder level of the multiple injury patient with sepsis in each indicator. According to the pathological disorder degree of each multiple injury patient in each index, the overall pathological disorder level of the multiple injury patient with sepsis in each index, and the high sensitivity of the multiple injury patient with sepsis in each index, the sepsis prediction value of each multiple injury patient in each index is obtained.

[0010] Further, the pathological disorder degree obtaining method comprises: The product mean value of the dynamic change rate and the baseline deviation degree of each multiple injury patient in each index at all monitoring time points is obtained as the pathological disorder degree of each multiple injury patient in each index.

[0011] Further, the sepsis prediction value of each multiple injury patient in each index comprises: The ratio of the pathological disorder degree of each multiple injury patient in each index and the overall pathological disorder level of the multiple injury patient with sepsis in each index is obtained as the sepsis risk degree of each multiple injury patient in each index. The product of the sepsis risk degree of each multiple injury patient in each index and the high sensitivity of the multiple injury patient with sepsis in each index is obtained as the sepsis prediction value of each multiple injury patient in each index.

[0012] Further, the sepsis risk score obtaining method comprises: The high risk of different indexes according to the sepsis prediction value of each multiple injury patient in different indexes is weighted and averaged, and the sepsis risk score of each multiple injury patient is obtained.

[0013] Further, the sepsis risk score of each multiple injury patient obtained by weighting and averaging the high risk of different indexes according to the sepsis prediction value of each multiple injury patient in different indexes comprises: The sepsis prediction value of each multiple injury patient in different indexes is weighted and accumulated as a first accumulated value; The accumulated value of the sepsis prediction value of each multiple injury patient in different indexes is obtained as a second accumulated value; The ratio of the first accumulated value and the second accumulated value is obtained as the sepsis risk score of each multiple injury patient.

[0014] The present application has the following beneficial effects: According to the monitoring data distribution of each index of each multiple injury patient at different monitoring moments, the baseline deviation degree of each index of each multiple injury patient at each monitoring moment is obtained, which is helpful to evaluate the relative abnormal degree of each index and obtain the high risk of each index of each multiple injury patient, evaluate the risk degree of each index of the patient, and quantify the individualized risk characteristics; according to the high risk of each index of different multiple injury patients, the high sensitivity of each index of the multiple injury patient with sepsis is obtained, and the unique characteristics of the index are analyzed; according to the monitoring data distribution of each index of each multiple injury patient at different monitoring moments and the difference between adjacent monitoring moments, the dynamic change rate of each index of each multiple injury patient at each monitoring moment is obtained, which can capture the dynamic trend of each index and evaluate the severity and development trend of the immune state disorder; considering that the index change speed and fluctuation amplitude of different patients at different monitoring moments are different, according to the dynamic change rate and baseline deviation degree of each index of different multiple injury patients at all monitoring moments, and the high sensitivity of each index of the multiple injury patient with sepsis, the sepsis prediction value of each index of each multiple injury patient is obtained, and the reference value of each index to the occurrence of sepsis is quantified; according to the sepsis prediction value and high risk of each index of each multiple injury patient, the sepsis risk score of each multiple injury patient is obtained, and the early warning of sepsis of the multiple injury patient is carried out. The sepsis risk of each multiple injury patient is accurately evaluated, and the accuracy of early warning is improved. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0016] Figure 1 A flow chart of a multiple injury patient sepsis early warning method based on artificial intelligence provided by an embodiment of the present application.

[0017] Figure 2 A flow chart of a sepsis prediction value acquisition method provided by an embodiment of the present application. DETAILED DESCRIPTION

[0018] In order to further illustrate the technical means and effects taken by the present application to achieve the predetermined inventive purpose, the specific implementation, structure, features and effects of the early warning method for sepsis of multiple injury patients based on artificial intelligence according to the present application are described in detail as follows in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0019] 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 this application belongs.

[0020] The specific scheme of the early warning method for sepsis of multiple injury patients based on artificial intelligence provided by the present application is specifically described below in combination with the drawings.

[0021] Please refer to Figure 1 which shows the flowchart of the early warning method for sepsis of multiple injury patients based on artificial intelligence provided by one embodiment of the present application, and the specific method comprises: Step S1: acquiring monitoring data of multiple physiological indexes of multiple injury patients with sepsis and non-sepsis at each monitoring time.

[0022] In the embodiment of the present application, considering that multiple injury patients themselves will cause systemic traumatic inflammatory response due to tissue damage, which is highly overlapped with the early manifestations of sepsis, in order to avoid large errors in distinguishing traumatic inflammation and sepsis, it is necessary to analyze the dynamic changes of physiological indexes of patients to early warn sepsis; first, collect the medical data of multiple injury patients in the emergency department, including physiological index data at multiple monitoring times from admission, wherein the multiple injury patients are divided into two categories of sepsis and non-sepsis according to the existing sepsis judgment rules, the physiological indexes include various blood indexes such as lactic acid, white blood cell count, hemoglobin, pro-inflammatory indexes such as neutrophil to lymphocyte ratio and immunosuppression indexes, therefore, the monitoring data of multiple physiological indexes of multiple injury patients with sepsis and non-sepsis at each monitoring time are acquired.

[0023] It should be noted that in one embodiment of the present application, the monitoring data of multiple physiological indexes of multiple injury patients within 48 hours from admission at multiple monitoring times are acquired in the medical data, and the monitoring times are pre-set by the implementers according to relevant experience; in other embodiments of the present application, the monitoring time range can be set according to specific circumstances, which is not limited and described here.

[0024] It should be noted that, in order to facilitate subsequent data processing, the missing values in the obtained monitoring data are filled by using interpolation, mean or median, the abnormal data are eliminated or corrected, and the data quality is ensured; and the obtained monitoring data is subjected to Z-score standardization processing, so as to eliminate the influence of the dimension produced in data operation; the specific means are the technical means familiar to those skilled in the art, and will not be described here.

[0025] Step S2: According to the monitoring data distribution of each polytrauma patient at different monitoring time points of each index, the baseline deviation degree of each polytrauma patient at each monitoring time point of each index is obtained, and the high risk of each polytrauma patient at each index is obtained; according to the change of the high risk of different polytrauma patients at each index, the high sensitivity of the polytrauma patient with sepsis at each index is obtained.

[0026] For polytrauma patients, due to multiple organ injuries or immune function disorders, the monitoring data of each index will change, and the data deviation of the related indexes of the polytrauma patient after sepsis infection is more serious. The greater the deviation from the index at the earliest monitoring time point, the more clearly the dynamic deviation characteristics of the data can be understood by analyzing the monitoring data distribution of each index at different monitoring time points, reflecting the risk of disease deterioration and the immediate impact of trauma on each index. According to the monitoring data distribution of each polytrauma patient at different monitoring time points of each index, the baseline deviation degree of each polytrauma patient at each monitoring time point of each index is obtained, and the high risk of each polytrauma patient at each index is obtained.

[0027] Preferably, in an embodiment of the present application, the baseline deviation degree acquisition method comprises: For any monitoring time point, the monitoring data of each polytrauma patient at each index is compared with the standard reference range, and if the monitoring data is within the standard reference range, the baseline deviation degree of each index of each polytrauma patient is set to 0. If the monitoring data is not within the standard reference range, the minimum value of the difference between the monitoring data and the different boundary data in the standard reference range is obtained, and the boundary data corresponding to the minimum value is taken as the reference data; the ratio of the minimum value and the reference data is calculated as the baseline deviation degree of each polytrauma patient at each index.

[0028] It should be noted that, in the embodiments of the present application, the standard reference range can be obtained by the implementers according to relevant professional knowledge in advance.

[0029] Take an example. There is a standard reference range ab for any indicator. For any monitoring moment, the monitoring data x of the indicator does not exist in the standard reference range. If x is less than a, there is a minimum difference between x and the boundary data a within the corresponding standard range. The ratio of |xa| and a is calculated to obtain the baseline deviation of the corresponding indicator; if x is greater than b, there is a minimum difference between x and the boundary data b within the corresponding standard range. The ratio of |xb| and b is calculated to obtain the baseline deviation of the corresponding indicator.

[0030] Preferably, in one embodiment of the present invention, the method for obtaining high risk includes: The mean difference in baseline deviation of each indicator of each polytrauma patient between different monitoring times and the earliest monitoring time was obtained as the high risk of each polytrauma patient in each indicator.

[0031] In one embodiment of the present invention, the formula for high risk is expressed as: ; in, Indicates the The multi-injured patient High risk of an indicator; Indicates the The multi-injured patient Monitoring time Baseline deviation of each indicator; Indicates the The first patient with multiple injuries was Baseline deviation of each indicator; Indicates the number of monitoring moments.

[0032] In the high-risk formula, Indicates calculation of The multi-injured patient The difference in the baseline deviation of each indicator between the monitoring moment and the earliest monitoring moment, the greater the difference, the greater the difference in the baseline deviation of the indicator relative to the earliest monitoring moment, the greater the difference in the deviation from the earliest monitoring moment, the more likely it is an indicator of sepsis infection, and the higher the risk.

[0033] Comparing the high-risk changes of a certain indicator in polytrauma patients with sepsis and non-sepsis can reflect whether the indicator shows more obvious abnormal changes when sepsis occurs, which is more helpful in identifying sensitive indicators closely related to sepsis; based on the high-risk changes of each indicator in different polytrauma patients, the high sensitivity of each indicator in polytrauma patients with sepsis can be obtained.

[0034] Preferably, in one embodiment of the present invention, the highly sensitive acquisition method includes: obtaining the sum of the high-risk of all the multiple-injury patients with sepsis in each index as the high-risk level of sepsis; obtaining the sum of the high-risk of all the multiple-injury patients in each index as the overall high-risk level; obtaining the ratio between the high-risk level of sepsis and the overall high-risk level as the high-sensitivity of each index of the multiple-injury patients with sepsis.

[0035] In an embodiment of the present application, the formula of the high-sensitivity is expressed as: , wherein, represents the high-sensitivity of each index of the multiple-injury patients with sepsis; represents the number of the multiple-injury patients with sepsis; represents the high-risk of the i-th multiple-injury patient with sepsis in the j-th index; represents the total number of the multiple-injury patients; represents the high-risk of the i-th multiple-injury patient in the j-th index.

[0036] In the formula of the high-sensitivity, represents the sum of the high-risk of all the multiple-injury patients with sepsis in each index as the high-risk level of sepsis; represents the sum of the high-risk of all the multiple-injury patients in each index as the overall high-risk level; represents the ratio between the high-risk level of sepsis and the overall high-risk level, the greater the ratio, the greater the high-risk of the multiple-injury patients with sepsis in the j-th index, the more obvious the abnormal change, and the greater the high-sensitivity of the multiple-injury patients with sepsis in the j-th index.

[0037] Step S3: obtaining the dynamic change rate of each index of each multiple-injury patient at each monitoring time according to the distribution of the monitoring data of each index of each multiple-injury patient at different monitoring times and the difference between adjacent monitoring times, and obtaining the sepsis prediction value of each multiple-injury patient in each index according to the dynamic change rate and baseline deviation of each index of different multiple-injury patients at all monitoring times and the high-sensitivity of each index of the multiple-injury patients with sepsis.

[0038] ​​​​​​​​​​Polytrauma patients infected with sepsis will experience dramatic fluctuations in immune status in the early stages, accompanied by a more intense and sustained pro-inflammatory response than traumatic inflammation, resulting in significant changes in relevant indicators. By analyzing the monitoring data distribution of each indicator of each polytrauma patient at different monitoring times, as well as the differences between adjacent monitoring times, continuous dynamic monitoring of relevant indicators is performed, and the dynamic change rate of each indicator at each monitoring time is quantified to reflect the fluctuation of the indicator. When the difference between monitoring times is smaller, the indicator has a greater change, the dynamic change rate is also greater, and there is a greater fluctuation. Based on the monitoring data distribution of each indicator of each polytrauma patient at different monitoring times, as well as the differences between adjacent monitoring times, the dynamic change rate of each indicator of each polytrauma patient at each monitoring time is obtained.

[0039] Preferably, in one embodiment of the present invention, the method for obtaining the dynamic change rate includes: The difference in monitoring data of each indicator of each polytrauma patient between each monitoring moment and the previous monitoring moment was obtained, and the ratio of the monitoring data difference to the difference between the corresponding monitoring moments was calculated as the dynamic change rate of each indicator of each polytrauma patient at each monitoring moment.

[0040] In one embodiment of the present invention, the formula for the dynamic change rate is expressed as: ; in, Indicates the The multi-injured patient Monitoring time Dynamic change rate of an indicator; Indicates the The multi-injured patient Monitoring time Monitoring data of the indicators; Indicates the The multi-injured patient Monitoring time Monitoring data of the indicators; Indicates the The multi-injured patient Monitoring time and previous monitoring time the differences between; Indicates taking the absolute value.

[0041] In the formula of dynamic rate of change, The multi-injured patient Monitoring time and Monitoring time between The greater the difference in monitoring data of an indicator, the Monitoring time and previous monitoring time The smaller the difference, the greater the data change in a shorter period of time and the greater the dynamic change rate.

[0042] After admission for surgery, if polytrauma patients do not develop sepsis, the baseline deviation of the indicators should remain stable or gradually return to normal at multiple subsequent monitoring times. The smaller the dynamic change rate, the greater the baseline deviation. For patients with sepsis, the indicators still fluctuate greatly at multiple monitoring times and the larger the baseline deviation. The high sensitivity of each indicator in polytrauma patients with sepsis can reflect the obvious degree of abnormal changes in the indicator when sepsis occurs. The greater the high sensitivity and the greater the obviousness, the more helpful it is for analyzing the occurrence of sepsis. Therefore, by analyzing the dynamic change rate and baseline deviation of each indicator in different polytrauma patients at all monitoring times, the predictive value of sepsis for each indicator in each polytrauma patient can be more comprehensively and accurately assessed. Based on the dynamic change rate and baseline deviation of each indicator in different polytrauma patients at all monitoring times, as well as the high sensitivity of each indicator in polytrauma patients with sepsis, the predictive value of sepsis for each indicator in each polytrauma patient can be obtained.

[0043] Preferably, in one embodiment of the present invention, the method for obtaining the sepsis prediction value is as follows: Figure 2 , which shows a flow chart of a method for obtaining the predictive value of sepsis, including: Step S201: Obtain the pathological disorder degree of each indicator of each polytrauma patient according to the dynamic change rate and baseline deviation of each indicator of each polytrauma patient at all monitoring moments.

[0044] Preferably, in one embodiment of the present invention, the method for obtaining the pathological disorder degree includes: The mean product of the dynamic change rate and baseline deviation of each indicator of each polytrauma patient at all monitoring moments was obtained as the pathological disorder degree of each indicator of each polytrauma patient.

[0045] In one embodiment of the present invention, the formula for pathological disorder degree is expressed as: ; in, Indicates the The multi-injured patient The degree of pathological disorder of each indicator; Indicates the number of monitoring moments; Indicates the The multi-injured patient Monitoring time Dynamic change rate of an indicator; Indicates the The multi-injured patient The next monitoring time The baseline deviation of each indicator.

[0046] In the formula of pathological disorder, the greater the dynamic change rate, the greater the baseline deviation, the greater the fluctuation and instability of the indicator data, and the greater the pathological disorder.

[0047] Step S202: Obtain the mean value of the pathological disorder degree of each indicator for all polytrauma patients with sepsis, as the overall pathological disorder level of each indicator for polytrauma patients with sepsis.

[0048] Although there are individual differences among patients, the relevant indicators of sepsis infection show roughly the same dynamic change trend. The level of pathological disorder of all polytrauma patients with sepsis is quantified by taking the mean value, and the overall level of pathological disorder of each indicator in polytrauma patients with sepsis is evaluated. The greater the pathological disorder, the more specific and reliable the sepsis performance of each indicator.

[0049] Step S203: Obtain the sepsis prediction value of each indicator for each polytrauma patient based on the pathological disorder degree of each indicator, the overall pathological disorder level, and the high sensitivity of each indicator for polytrauma patients with sepsis.

[0050] Preferably, in one embodiment of the present invention, obtaining the sepsis prediction value of each indicator for each polytrauma patient includes: The ratio of the pathological disorder level of each polytrauma patient in each indicator to the overall pathological disorder level of each polytrauma patient with sepsis in each indicator was obtained as the sepsis risk level of each polytrauma patient in each indicator. The product of the sepsis risk level of each polytrauma patient in each indicator and the high sensitivity of each indicator in polytrauma patients with sepsis was obtained as the predictive value of sepsis for each indicator in each polytrauma patient.

[0051] In one embodiment of the present invention, the formula for predicting the value of sepsis is expressed as: ; ; in, Indicates the The multi-injured patient The predictive value of these indicators for sepsis; Indicates the The multi-injured patient The degree of pathological disorder of each indicator; Multiple trauma patients with sepsis The overall pathological disorder level of each indicator; the number of polytrauma patients indicating sepsis; Indicates the The multi-injured patient The degree of pathological disorder of each indicator; Multiple trauma patients with sepsis High sensitivity of the indicator.

[0052] In the formula for predicting the value of sepsis, Indicates that all polytrauma patients with sepsis The mean value of the pathological disorder of each indicator, that is, the overall pathological disorder level of each indicator in polytrauma patients with sepsis. The larger the mean value, the greater the overall pathological disorder level, and the greater the reliability of the sepsis reference; Indicates the The multi-injured patient The pathological disorder of the indicators and the polytrauma patients with sepsis in the first The ratio of the overall pathological disorder level of the indicators, that is, the sepsis risk level of each polytrauma patient in each indicator. The larger the ratio, the higher the sepsis risk. The multi-injured patient The pathological disorder of the indexes is higher in polytrauma patients with sepsis than in The greater the overall pathological disorder level of each indicator, the greater the corresponding sepsis risk, and the greater the high sensitivity, the more obvious the change of sepsis in the corresponding indicator, and the greater the sepsis predictive value of each indicator.

[0053] Step S4: Based on the sepsis prediction value and high risk of each polytrauma patient in different indicators, a sepsis risk score is obtained for each polytrauma patient, and an early warning of sepsis is provided for the polytrauma patient.

[0054] Each polytrauma patient exhibits different changes in physiological indicators after admission. The sepsis prediction value can reflect the relative contribution of each indicator. Therefore, by analyzing the sepsis prediction value and high-risk nature of different indicators, the sepsis risk score of each polytrauma patient can be more accurately quantified.

[0055] Preferably, in one embodiment of the present invention, the method for obtaining the sepsis risk score includes: According to the predictive value of sepsis in different indicators for each polytrauma patient, the high-risk of different indicators was weighted averaged to obtain the sepsis risk score for each polytrauma patient.

[0056] In one embodiment of the present invention, a weighted average of the high-risk of different indicators is performed based on the sepsis prediction value of each polytrauma patient in different indicators to obtain a sepsis risk score for each polytrauma patient, including: Obtain the weighted cumulative sum of the sepsis prediction values ​​of different indicators for each polytrauma patient and the high-risk score as the first cumulative value; Obtain the cumulative value of the sepsis prediction value of different indicators for each polytrauma patient as the second cumulative value; The ratio of the first cumulative value to the second cumulative value was obtained as the sepsis risk score for each polytrauma patient.

[0057] In one embodiment of the present invention, the formula for sepsis risk score is expressed as: ; in, Indicates the sepsis risk score for polytrauma patients; Indicates the The multi-injured patient The predictive value of sepsis for each indicator; Indicates the The multi-injured patient High risk of the indicator; Indicates the number of indicators.

[0058] In the sepsis risk score formula, Indicates the The first cumulative value is the weighted sum of the sepsis prediction values ​​of different indicators for each polytrauma patient on the high risk; Indicates the The cumulative value of the sepsis prediction value of different indicators in multiple trauma patients is used as the second cumulative value; the larger the ratio, the greater the sepsis prediction value, the greater the credibility of high risk, the greater the proportion in reflecting sepsis risk, the greater the high-risk type, and the larger the sepsis risk score.

[0059] Based on this, a sepsis risk score is obtained for each polytrauma patient. In another embodiment of the present invention, an early warning of sepsis is also performed for polytrauma patients, including: obtaining the average sepsis risk score of all polytrauma patients with sepsis as a first sepsis risk score level; obtaining the average sepsis risk score of all polytrauma patients without sepsis as a second sepsis risk score level; and obtaining the average of the first sepsis risk score level and the second sepsis risk score level as a sepsis warning cutoff. The smaller the sepsis risk score is below the warning cutoff, the lower the risk of sepsis. The larger the sepsis risk score is above the warning cutoff, the greater the risk of sepsis, and the more necessary early warning is, which helps to timely intervene and reduce the mortality rate of sepsis.

[0060] In summary, the application obtains the high risk of each polytrauma patient in each index according to the monitoring data distribution of each polytrauma patient in each index at different monitoring time points, and obtains the high sensitivity of the polytrauma patient with sepsis in each index; obtains the dynamic change rate of each polytrauma patient in each index at each monitoring time point according to the monitoring data distribution of each polytrauma patient in each index at different monitoring time points and the difference between adjacent monitoring time points, and obtains the sepsis prediction value of each polytrauma patient in each index; and further obtains the sepsis risk score of each polytrauma patient to early warn the polytrauma patient of sepsis. The application accurately assesses the sepsis risk of each polytrauma patient to improve the accuracy of early warning.

[0061] It should be noted that the above-mentioned embodiment sequence of the 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.

[0062] Each embodiment in the specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other. Each embodiment focuses on the difference from other embodiments.

Claims

1. An artificial intelligence-based early warning method for sepsis in polytrauma patients, characterized by: The method comprises: Obtain monitoring data of multiple physiological indicators of septic and non-septic polytrauma patients at each monitoring moment; Based on the monitoring data distribution of each indicator of each polytrauma patient at different monitoring times, the baseline deviation of each indicator of each polytrauma patient at each monitoring time was obtained, and the high risk of each polytrauma patient in each indicator was obtained; based on the high risk changes of each indicator in different polytrauma patients, the high sensitivity of each indicator in polytrauma patients with sepsis was obtained; Based on the distribution of monitoring data for each indicator at different monitoring times for each polytrauma patient and the differences between adjacent monitoring times, the dynamic change rate of each indicator at each monitoring time for each polytrauma patient was obtained. Based on the dynamic change rate and baseline deviation of each indicator at all monitoring times for different polytrauma patients and the high sensitivity of each indicator in polytrauma patients with sepsis, the predictive value of sepsis for each indicator in each polytrauma patient was obtained. Based on the sepsis prediction value and high risk of different indicators of each polytrauma patient, a sepsis risk score is obtained for each polytrauma patient, and early warning of sepsis is provided for polytrauma patients.

2. The artificial intelligence-based early warning method for sepsis in polytrauma patients according to claim 1, characterized in that: The method for obtaining the baseline deviation includes: At any monitoring moment, the monitoring data of each indicator of each polytrauma patient were compared with the standard reference range. If the monitoring data were within the standard reference range, the baseline deviation of each indicator was set to 0. If there is no standard reference range for the monitoring data, obtain the minimum difference between the monitoring data and the different boundary data within the standard reference range, and use the boundary data corresponding to the minimum value as the baseline data; calculate the ratio of the minimum difference value and the baseline data as the baseline deviation of each indicator for each polytrauma patient.

3. The artificial intelligence-based early warning method for sepsis in polytrauma patients according to claim 1, characterized in that: The method for obtaining the high risk includes: The mean difference in baseline deviation of each indicator of each polytrauma patient between different monitoring times and the earliest monitoring time was obtained as the high risk of each polytrauma patient in each indicator.

4. The artificial intelligence-based early warning method for sepsis in polytrauma patients according to claim 1, characterized in that: The highly sensitive acquisition method includes: The high-risk level of sepsis was obtained by summing the high-risk level of each indicator for all polytrauma patients with sepsis; The sum of the high-risk scores of all polytrauma patients in each indicator was obtained as the overall high-risk level; The ratio between the high-risk level of sepsis and the overall high-risk level was obtained as the high sensitivity of each indicator in polytrauma patients with sepsis.

5. The artificial intelligence-based early warning method for sepsis in polytrauma patients according to claim 1, characterized in that: The method for obtaining the dynamic change rate includes: The difference in monitoring data of each indicator of each polytrauma patient between each monitoring moment and the previous monitoring moment was obtained, and the ratio of the monitoring data difference to the difference between the corresponding monitoring moments was calculated as the dynamic change rate of each indicator of each polytrauma patient at each monitoring moment.

6. The artificial intelligence-based early warning method for sepsis in polytrauma patients according to claim 1, characterized in that: The method for obtaining the sepsis prediction value includes: According to the dynamic change rate and baseline deviation of each indicator of each polytrauma patient at all monitoring moments, the pathological disorder degree of each indicator of each polytrauma patient was obtained; The mean value of pathological disorder of each indicator in all polytrauma patients with sepsis was obtained as the overall pathological disorder level of each indicator in polytrauma patients with sepsis; Based on the pathological disorder degree of each indicator in each polytrauma patient, the overall pathological disorder level and the high sensitivity of each indicator in polytrauma patients with sepsis, the predictive value of sepsis in each indicator in each polytrauma patient was obtained.

7. The artificial intelligence-based early warning method for sepsis in polytrauma patients according to claim 6, characterized in that: The method for obtaining the pathological disorder degree includes: The mean product of the dynamic change rate and baseline deviation of each indicator of each polytrauma patient at all monitoring moments was obtained as the pathological disorder degree of each indicator of each polytrauma patient.

8. The artificial intelligence-based early warning method for sepsis in polytrauma patients according to claim 6, characterized in that: The sepsis prediction value of each indicator for each polytrauma patient is obtained, including: The ratio of the pathological disorder level of each polytrauma patient in each indicator to the overall pathological disorder level of each polytrauma patient with sepsis in each indicator was obtained as the sepsis risk level of each polytrauma patient in each indicator. The product of the sepsis risk level of each polytrauma patient in each indicator and the high sensitivity of each indicator in polytrauma patients with sepsis was obtained as the predictive value of sepsis for each indicator in each polytrauma patient.

9. The artificial intelligence-based early warning method for sepsis in polytrauma patients according to claim 1, characterized in that: The method for obtaining the sepsis risk score includes: According to the predictive value of sepsis in different indicators for each polytrauma patient, the high-risk of different indicators was weighted averaged to obtain the sepsis risk score for each polytrauma patient.

10. The artificial intelligence-based early warning method for sepsis in polytrauma patients according to claim 9, characterized in that: The high-risk of different indicators is weighted averaged according to the sepsis prediction value of each polytrauma patient in different indicators to obtain the sepsis risk score of each polytrauma patient, including: Obtain the weighted cumulative sum of the sepsis prediction values ​​of different indicators for each polytrauma patient and the high-risk score as the first cumulative value; Obtain the cumulative value of the sepsis prediction value of different indicators for each polytrauma patient as the second cumulative value; The ratio of the first cumulative value to the second cumulative value was obtained as the sepsis risk score for each polytrauma patient.

Images