Anesthesiology clinical data management method and system

Abstract

The present invention provides a method and system for clinical data management of anesthesiology, and relates to the field of data processing technology. The method includes: obtaining the vital sign monitoring data of multiple sample patients in the target anesthesia stage and performing feature extraction, constructing a first vital sign parameter feature vector for each sample patient; clustering multiple sample patients to obtain multiple reference groups, and determining the vital sign reference vector for each reference group; obtaining the individual difference data of each sample patient, and determining the individual difference reference vector for each reference group; obtaining the target individual difference data and target vital sign monitoring data of the target patient, generating the second individual feature vector of the target patient and screening out the target group; calculating the vital sign difference factor of the target patient and generating the vital sign monitoring result of the target patient. The present invention achieves the improvement of the comprehensiveness and accuracy of evaluating the patient's status.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and system for managing clinical data in anesthesiology. Background Art

[0002] Anesthesiology is a vital area in the modern medical system, responsible for ensuring the safety and comfort of patients during surgery. During anesthesia, one of the commonly used monitoring methods is the bispectral index (BIS), which is a value obtained by analyzing and processing electroencephalogram (EEG) signals. BIS can reflect the excitation or inhibition state of the cerebral cortex and sedation and hypnosis information, and is one of the most widely used anesthesia depth monitoring tools in clinical practice.

[0003] However, EEG signals may be subject to various interferences in clinical applications, such as electromagnetic interference, muscle activity interference, and external device interference. These interferences increase the noise of EEG signals, which may lead to errors in recognition results, affecting the accuracy and reliability of signal analysis, and thus affecting the accuracy of anesthesia depth prediction based on BIS values. Therefore, in this case, comprehensive analysis combined with other vital sign parameters of the patient can improve the accuracy of recognition results. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a clinical data management method and system for the Department of Anesthesiology, which utilizes the synergistic effect of multiple vital sign data to analyze whether there are any abnormalities in the collected vital sign monitoring data of anesthetized patients, thereby providing data support for a more comprehensive and accurate assessment of the patient's condition, reducing the errors caused by single signal interference, and improving the accuracy of the overall analysis results.

[0005] In a first aspect, the present invention provides a method for managing clinical data of anesthesiology, comprising:

[0006] Acquire the vital sign monitoring data of multiple sample patients at the target anesthesia stage, wherein the vital sign monitoring data includes vital sign monitoring sequences corresponding to multiple vital sign parameters;

[0007] Extract features from the vital sign monitoring data of each sample patient, calculate the mean corresponding to each vital sign parameter, and construct the first vital sign parameter feature vector of each sample patient;

[0008] Clustering multiple sample patients based on the DBSCAN clustering algorithm and multiple first physical sign parameter feature vectors to obtain multiple reference groups, and determining the physical sign reference vector of each reference group according to multiple core points corresponding to each reference group;

[0009] Obtaining individual difference data of each sample patient, performing data cleaning on the individual difference data and constructing the first individual feature vector of each sample patient, and determining the individual difference reference vector of each reference group according to the first individual feature vector;

[0010] Obtaining target individual difference data and target vital sign monitoring data of a target patient, generating a second individual characteristic vector of the target patient according to the target individual difference data, and screening out a target group from multiple reference groups based on the second individual characteristic vector;

[0011] A second vital sign parameter feature vector of the target patient is generated according to the target vital sign monitoring data of the target patient, a vital sign difference factor of the target patient is calculated based on the vital sign reference vector of the target group, and a vital sign monitoring result of the target patient is generated according to the vital sign difference factor.

[0012] Preferably, determining the individual difference reference vector of each reference group according to the first individual feature vector comprises:

[0013] For any reference group, determining a group center point of the reference group according to multiple core points of the reference group, including calculating the mean of the first individual feature vectors of the multiple core points to determine the group center point;

[0014] Calculate the distances between multiple core points and the center point of the group respectively, determine the weighting factor of each core point according to the distance between the core point and the center point of the group, and determine the individual difference reference vector of the reference group according to the weighting factor of each core point and the first individual feature vector:

[0015] ;

[0016] In the formula, is the individual difference reference vector of the reference group, For the reference group The first volume feature vector of the core points, For the reference group The weighting factor of the core points, For the reference group The distance between the core point and the center point of the group, is a constant term, is the number of core points in the reference group.

[0017] Preferably, determining the vital sign reference vector of each reference group according to a plurality of core points corresponding to each reference group includes:

[0018] The neighborhood density of multiple core points in each reference group is calculated respectively, and the first vital sign parameter feature vectors corresponding to the multiple core points in each reference group are respectively fused according to the neighborhood density of the core points to generate the vital sign reference vector of each reference group. For any vital sign reference vector of a reference group:

[0019] ;

[0020] In the formula, is the reference vector of the reference group’s physical signs, For the reference group The first vital sign parameter eigenvector of the core points, For the reference group The neighborhood density of the core points is is the number of core points in the reference group.

[0021] Preferably, generating the physical sign monitoring results of the target patient according to the physical sign difference factor includes:

[0022] Determine the difference reference range corresponding to the target group from the difference factor reference list. The difference factor reference list includes the difference reference range corresponding to each reference group. Judge whether the physical sign difference factor of the target patient belongs to the difference reference range corresponding to the target group. If so, mark the target physical sign monitoring data as normal data; otherwise, mark the target physical sign monitoring data as abnormal data.

[0023] Preferably, the difference factor reference list also includes:

[0024] For any reference group, the distance between each core point in the reference group and the reference sign reference vector of the reference group is calculated to obtain the distance data of the reference group;

[0025] The distance data of the reference group are fitted based on the normal distribution model, and the probability density function of the distance data is calculated;

[0026] Calculate the upper and lower limits of the cumulative distribution function of the normal distribution model in the preset confidence interval, and determine the reference range of the difference of the reference group according to the upper and lower limits;

[0027] The difference reference ranges corresponding to multiple reference groups are calculated, and a difference factor reference list containing the difference reference ranges corresponding to each reference group is constructed.

[0028] Preferably, screening out a target group from a plurality of reference groups based on the second individual feature vector comprises:

[0029] The cosine similarity between the second individual feature vector and the individual difference reference vectors corresponding to multiple reference groups is calculated, and multiple reference groups whose cosine similarity is greater than a preset similarity threshold are screened out, and the reference group with the largest cosine similarity is recorded as the target group.

[0030] A second aspect provides an anesthesiology clinical data management system, which is used to implement an anesthesiology clinical data management method, including:

[0031] A data acquisition module is used to acquire the vital sign monitoring data of multiple sample patients at the target anesthesia stage, wherein the vital sign monitoring data includes vital sign monitoring sequences corresponding to multiple vital sign parameters;

[0032] The physical sign parameter analysis module is used to extract features from the physical sign monitoring data of each sample patient, calculate the mean value corresponding to each physical sign parameter, and construct the first physical sign parameter feature vector of each sample patient;

[0033] A sample patient classification module is used to cluster multiple sample patients based on the DBSCAN clustering algorithm and multiple first physical sign parameter feature vectors to obtain multiple reference groups, and determine the physical sign reference vector of each reference group according to multiple core points corresponding to each reference group;

[0034] An individual difference analysis module is used to obtain individual difference data of each sample patient, clean the individual difference data and construct the first individual feature vector of each sample patient, and determine the individual difference reference vector of each reference group according to the first individual feature vector;

[0035] A target group screening module is used to obtain target individual difference data and target vital sign monitoring data of a target patient, generate a second individual feature vector of the target patient according to the target individual difference data, and screen out a target group from multiple reference groups based on the second individual feature vector;

[0036] The vital sign monitoring and analysis module is used to generate a second vital sign parameter feature vector of the target patient based on the target vital sign monitoring data of the target patient, calculate the vital sign difference factor of the target patient based on the vital sign reference vector of the target group, and generate the vital sign monitoring result of the target patient according to the vital sign difference factor.

[0037] Preferably, the physical sign monitoring and analysis module generates physical sign monitoring results of the target patient according to the physical sign difference factors, including:

[0038] Determine the difference reference range corresponding to the target group from the difference factor reference list. The difference factor reference list includes the difference reference range corresponding to each reference group. Judge whether the physical sign difference factor of the target patient belongs to the difference reference range corresponding to the target group. If so, mark the target physical sign monitoring data as normal data; otherwise, mark the target physical sign monitoring data as abnormal data.

[0039] The beneficial effects of the present invention are as follows:

[0040] The present invention combines the BIS value to conduct a comprehensive data analysis of multiple vital signs data, effectively reducing the error caused by single signal interference and improving the accuracy of the overall analysis results; fully considering the individual difference data of patients, making the evaluation of the anesthetic state more accurate; finely classifying patients to ensure the representativeness and stability of the reference group, which helps to more accurately evaluate the anesthetic state of the target patient, and intelligently processing the data to make different features comparable, thereby improving the efficiency and accuracy of data analysis. When the comprehensive analysis results show that the patient's vital signs are abnormal, a warning can be issued to medical staff in a timely manner, providing data support so that medical staff can conduct a more detailed analysis of the patient's anesthetic state, thereby improving the comprehensiveness and accuracy of the evaluation of the patient's state. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 The figure is a flowchart of a method for managing clinical data of anesthesiology in one of the embodiments of the present invention.

[0042] Figure 2 This is a schematic diagram of the structure of an anesthesiology clinical data management system in one of the embodiments of the present invention. DETAILED DESCRIPTION

[0043] In order to make those skilled in the art better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0044] Figure 1 A schematic diagram showing a process flow of a method for managing clinical data in anesthesiology in one embodiment of the present invention is shown in FIG. Figure 1 , the method comprises the following steps:

[0045] S1. Obtaining vital sign monitoring data of multiple sample patients at a target anesthesia stage, wherein the vital sign monitoring data includes vital sign monitoring sequences corresponding to multiple vital sign parameters.

[0046] It is worth noting that anesthetic drugs will affect the brain electroencephalogram (EEG), and the anesthesia process has a strong correlation with EEG. Monitoring the patient's bispectral index (BIS) is one of the most widely used methods for evaluating the depth of anesthesia. The bispectral index (BIS) is often generated by analyzing the patient's EEG signal, and the patient's anesthesia depth is determined based on the bispectral index (BIS). For example, when the BIS value is maintained between 40-60, it means that the patient is at an appropriate depth of anesthesia. However, in actual clinical applications, the acquisition of EEG signals may be interfered by factors such as electromagnetic and muscle activity, so that there is noise in the collected EEG signal, thereby affecting the accuracy of anesthesia depth prediction. In clinical practice, anesthesiologists will also use the patient's heart rate, blood pressure, respiratory rate and other signs as the basis for anesthesia depth analysis. However, this method relies on the personal clinical experience of the anesthesiologist, and there are one-sidedness, subjectivity and limitations. A clinical data management method for anesthesia department provided in an embodiment of the present invention combines the patient's BIS value and multiple sign data to conduct a comprehensive analysis of clinical data to improve the accuracy of identifying the patient's anesthesia state.

[0047] In this embodiment, by analyzing the vital sign monitoring data of multiple sample patients in the target anesthesia stage and mining the historical data, data support is provided for identifying whether there are abnormalities in the patient's vital sign monitoring data, so as to assist professionals in analyzing the patient's anesthesia depth. Among them, the sample patients are specifically patients who are in a suitable anesthesia state and have no abnormalities during the anesthesia process. For example, the target anesthesia stage is a stage in which the BIS value of the sample patient is maintained between 40-60. The vital sign monitoring data includes a vital sign monitoring sequence corresponding to multiple vital sign parameters of multiple sample patients, wherein the multiple vital sign parameters include but are not limited to heart rate (HR), blood pressure (BP), respiratory rate (RR), blood oxygen saturation (SpO2) and other data. By analyzing the change rules of multiple vital sign parameters of patients in a suitable anesthesia state during anesthesia determined based on the BIS value, data feature mining is performed to provide data reference for the anesthesia state analysis of other patients.

[0048] S2. Extract features from the vital sign monitoring data of each sample patient, calculate the mean corresponding to each vital sign parameter, and construct the first vital sign parameter feature vector of each sample patient.

[0049] In this embodiment, the overall level of the data is reflected by calculating the mean value corresponding to each vital sign parameter, so as to extract the characteristic value of each vital sign parameter, and construct multiple first vital sign parameter characteristic vectors according to the characteristic values ​​corresponding to multiple vital sign parameters.

[0050] It is worth noting that in order to facilitate subsequent data analysis and make different features have the same dimension and comparability, multiple first vital sign parameter feature vectors can be standardized, such as Z-score standardization. Data standardization is a technical means well known to technicians in this field and will not be repeated here.

[0051] S3. Clustering multiple sample patients based on the DBSCAN clustering algorithm and multiple first vital sign parameter feature vectors to obtain multiple reference groups, and determining the vital sign reference vector of each reference group according to multiple core points corresponding to each reference group.

[0052] In this embodiment, for multiple sample patients, the DBSCAN clustering algorithm is used for clustering, and only the differences between the physical sign parameters of different sample patients are considered to obtain multiple reference groups, wherein the multiple reference groups obtained based on the DBSCAN clustering algorithm respectively include information such as core points and boundary points of each reference group. For multiple reference groups, multiple core points in each reference group are respectively represented, and the physical sign reference vector representing the overall characteristics of multiple physical sign parameters of the reference group is determined according to the first physical sign parameter feature vector of the multiple core points.

[0053] S4. Obtain individual difference data of each sample patient, perform data cleaning on the individual difference data and construct the first individual feature vector of each sample patient, and determine the individual difference reference vector of each reference group according to the first individual feature vector.

[0054] In this embodiment, individual difference data include physiological characteristic data, anesthetic drug data, medical history data, etc. Patients of different age groups, gender differences, height and weight differences, etc. may all lead to differences in individual anesthetic reactions, resulting in different individuals having different sensitivities and metabolic rates to anesthetic drugs. Factors such as the type of anesthetic drugs, previous medical history (such as hypertension, diabetes, heart disease, etc.) will also affect the individual's response to anesthetic drugs and physical sign parameters. For the individual difference data of multiple sample patients, after data cleaning, the first individual feature vector corresponding to each sample patient is constructed to extract the individual difference information of different sample patients. For different reference groups, the differences in individual factors lead to certain commonalities between some individuals in the reference groups. By analyzing the first individual feature vectors corresponding to multiple core points in each reference group, an individual difference reference vector representing the commonality of individual differences in each reference group is generated.

[0055] S5. Obtain target individual difference data and target vital sign monitoring data of the target patient, generate a second individual feature vector of the target patient according to the target individual difference data, and screen out a target group from multiple reference groups based on the second individual feature vector.

[0056] In this embodiment, for the target individual difference data of the target patient collected, a second individual feature vector representing the feature information of the target individual difference data is extracted, and by analyzing the correlation between the second individual feature vector and the individual difference reference vectors corresponding to the multiple reference groups, the reference group with the strongest correlation is determined and recorded as the target group. Exemplarily, the cosine similarity between the second individual feature vector and the individual difference reference vectors corresponding to the multiple reference groups is calculated, and the reference group with the greatest similarity is screened out and recorded as the target group.

[0057] It is worth mentioning that in order to improve the reference value of the data and the feasibility of data analysis, as an optional implementation plan, in the process of screening the target group, multiple reference groups can be further limited based on a preset similarity threshold, and the reference group with the largest cosine similarity is screened out from multiple reference groups whose cosine similarity is greater than the preset similarity threshold and recorded as the target group. The significance of this is that considering that the correlation between some outliers and multiple reference groups is relatively low, the feasibility of data analysis can be improved in this way.

[0058] S6. Generate a second vital sign parameter feature vector of the target patient based on the target vital sign monitoring data of the target patient, calculate the vital sign difference factor of the target patient based on the vital sign reference vector of the target group, and generate the vital sign monitoring result of the target patient based on the vital sign difference factor.

[0059] In this embodiment, for the target vital sign monitoring data of the target patient, a second vital sign parameter characteristic vector characterizing the clinical characteristic information of multiple vital sign parameters of the target patient is extracted, and the difference in clinical characteristic information between the reference group with the greatest commonality with the individual difference data of the target patient and the target patient is analyzed. Specifically, the difference between the second vital sign parameter characteristic vector and the vital sign reference vector of the target group is analyzed, and the vital sign difference factor of the target patient is calculated to characterize the degree of difference between the clinical characteristic information between the target patient and multiple patients when there is a large correlation between the individual difference information, and the abnormality of the target patient's target vital sign monitoring data is determined through a pre-set evaluation method, and the vital sign monitoring results of the target patient are generated, thereby providing data support for the anesthesia depth analysis of the target patient.

[0060] In a specific implementation process, determining the vital sign reference vector of each reference group according to the multiple core points corresponding to each reference group in step S3 specifically includes:

[0061] Calculate the neighborhood density of multiple core points in each reference group respectively. Specifically, according to the neighborhood radius 𝜖 and the minimum number of samples used when clustering multiple sample patients using the DBSCAN clustering algorithm, calculate the number of points of each core point in the neighborhood 𝜖 to obtain the neighborhood density of each core point. Take the neighborhood density of each core point as the fusion reference, and perform feature fusion on the first vital sign parameter feature vectors corresponding to the multiple core points in each reference group according to the neighborhood density of the core point to generate a vital sign reference vector for each reference group.

[0062] Specifically, for any reference group's physical sign reference vector, the following formula is used for calculation:

[0063] ;

[0064] In the formula, is the reference vector of the reference group’s physical signs, For the reference group The first vital sign parameter eigenvector of the core points, For the reference group The neighborhood density of the core points is is the number of core points in the reference group.

[0065] The above method can be used to calculate the physical sign reference vectors corresponding to multiple reference groups, which are used to characterize the overall characteristics of multiple physical sign parameters corresponding to each reference group.

[0066] In a specific implementation process, determining the individual difference reference vector of each reference group according to the first individual feature vector in step S4 specifically includes:

[0067] Taking any reference group as an example, the group center point of the reference group is determined according to multiple core points of the reference group, wherein the mean of the first body feature vectors of the multiple core points is calculated, and the vector formed by the mean of the first body feature vectors of the multiple core points is used as the group center point.

[0068] After determining the center point of the group, the distances between multiple core points and the center point of the group are calculated respectively. For example, based on calculating the Euclidean distance between multiple core points and the center point of the group, the weighting factor corresponding to each core point is determined according to the Euclidean distance. The weighted analysis of multiple core points is achieved through the weighting factor to obtain an individual difference reference vector that characterizes the commonality of individual difference characteristics of the reference group.

[0069] Specifically, the weighting factor of each core point is determined according to the distance between the core point and the center point of the group, and the individual difference reference vector of the reference group is determined according to the weighting factor of each core point and the first individual feature vector:

[0070] ;

[0071] In the formula, is the individual difference reference vector of the reference group, For the reference group The first volume feature vector of the core points, For the reference group The weighting factor of the core points, For the reference group The distance between the core point and the center point of the group, is a constant term, is the number of core points in the reference group.

[0072] Specifically, during the weighting process, the closer the distance between the core point and the center point of the group is, the larger the corresponding weighting factor is, so that the individual difference reference vector can better represent the common characteristics between multiple core points in the reference group. The individual difference reference vector of each reference group can be calculated in the above manner.

[0073] In a specific implementation process, generating the physical sign monitoring results of the target patient according to the physical sign difference factor in step S6 specifically includes:

[0074] Determine the difference reference range corresponding to the target group from the difference factor reference list, wherein the difference factor reference list includes the difference reference range corresponding to each reference group. After extracting the difference reference range corresponding to the target group, determine whether the target patient's sign difference factor belongs to the difference reference range corresponding to the target group. If so, it is considered that the collected target sign monitoring data does not have an abnormality, that is, it does not deviate from the change law of the sign parameters of multiple sample patients, and the target sign monitoring data is marked as normal data. Otherwise, it is considered that the collected target sign monitoring data has an abnormality, and there is a certain difference from the change law of the sign parameters of multiple sample patients. In this case, the target sign monitoring data is marked as abnormal data, and finally the sign monitoring result of the target patient is obtained, which can provide data reference for professional physicians to analyze the patient's anesthesia depth.

[0075] As an optional implementation scheme, the difference factor reference list is determined in the following manner:

[0076] Taking any reference group as an example, the distance between each core point in the reference group and the reference sign reference vector of the reference group is calculated to obtain the distance data of the reference group.

[0077] Among them, the distance between each core point in the reference group and the vital sign reference vector of the reference group is calculated. Specifically, the distance calculation formula used in analyzing the difference between the second vital sign parameter characteristic vector and the vital sign reference vector of the target group can be used for calculation. For example, the Euclidean distance calculation formula is used to measure the difference between the second vital sign parameter characteristic vector and the vital sign reference vector of the target group to obtain the vital sign difference factor of the target patient, and the Euclidean distance is used to characterize the distance between each core point in the reference group and the vital sign reference vector of the reference group.

[0078] After calculating the distance data of the reference group, the distance data of the reference group is fitted based on the normal distribution model. Specifically, the mean and standard deviation involved in the normal distribution are estimated, and the probability density function of the distance data is calculated based on the estimated mean and standard deviation. Then, the cumulative distribution function of the normal distribution model is calculated based on the estimated mean and standard deviation. The upper and lower limits of the cumulative distribution function of the normal distribution model in a preset confidence interval are calculated to determine the reference range of differences in the reference group, wherein the preset confidence interval is, for example, a 95% confidence interval.

[0079] In the above manner, the difference reference ranges corresponding to multiple reference groups can be calculated to more accurately measure the correlation between two objects, and a difference factor reference list containing the difference reference ranges corresponding to each reference group can be constructed.

[0080] An embodiment of the present invention provides a method for managing clinical data in the Department of Anesthesiology. It combines BIS values ​​to conduct a comprehensive data analysis of multiple vital signs data, effectively reducing the error caused by single signal interference and improving the accuracy of the overall analysis results. It fully considers the individual difference data of patients to make the assessment of the anesthetic state more accurate. It classifies patients in detail to ensure the representativeness and stability of the reference group, which helps to more accurately assess the anesthetic state of the target patient, and processes the data intelligently to make different features comparable, thereby improving the efficiency and accuracy of data analysis. When the comprehensive analysis results show that the patient's vital signs are abnormal, a warning can be issued to medical staff in a timely manner, providing data support so that medical staff can conduct a more detailed analysis of the patient's anesthetic state, thereby improving the comprehensiveness and accuracy of the assessment of the patient's state.

[0081] Figure 2 FIG. 1 is a schematic diagram showing the structure of an anesthesiology clinical data management system in one embodiment of the present invention, see Figure 2 , the system comprises:

[0082] A data acquisition module is used to acquire the vital sign monitoring data of multiple sample patients at the target anesthesia stage, wherein the vital sign monitoring data includes vital sign monitoring sequences corresponding to multiple vital sign parameters;

[0083] The physical sign parameter analysis module is used to extract features from the physical sign monitoring data of each sample patient, calculate the mean value corresponding to each physical sign parameter, and construct the first physical sign parameter feature vector of each sample patient;

[0084] A sample patient classification module is used to cluster multiple sample patients based on the DBSCAN clustering algorithm and multiple first physical sign parameter feature vectors to obtain multiple reference groups, and determine the physical sign reference vector of each reference group according to multiple core points corresponding to each reference group;

[0085] Wherein, the physical sign reference vector of each reference group is determined according to a plurality of core points corresponding to each reference group, including:

[0086] The neighborhood density of multiple core points in each reference group is calculated respectively, and the first vital sign parameter feature vectors corresponding to the multiple core points in each reference group are respectively fused according to the neighborhood density of the core points to generate the vital sign reference vector of each reference group. For any vital sign reference vector of a reference group:

[0087] ;

[0088] In the formula, is the reference vector of the reference group’s physical signs, For the reference group The first vital sign parameter eigenvector of the core points, For the reference group The neighborhood density of the core points is is the number of core points in the reference group.

[0089] An individual difference analysis module is used to obtain individual difference data of each sample patient, clean the individual difference data and construct the first individual feature vector of each sample patient, and determine the individual difference reference vector of each reference group according to the first individual feature vector;

[0090] Wherein, the individual difference reference vector of each reference group is determined according to the first individual feature vector, including:

[0091] For any reference group, determining a group center point of the reference group according to multiple core points of the reference group, including calculating the mean of the first individual feature vectors of the multiple core points to determine the group center point;

[0092] Calculate the distances between multiple core points and the center point of the group respectively, determine the weighting factor of each core point according to the distance between the core point and the center point of the group, and determine the individual difference reference vector of the reference group according to the weighting factor of each core point and the first individual feature vector:

[0093] ;

[0094] In the formula, is the individual difference reference vector of the reference group, For the reference group The first volume feature vector of the core points, For the reference group The weighting factor of the core points, For the reference group The distance between the core point and the center point of the group, is a constant term, is the number of core points in the reference group.

[0095] A target group screening module is used to obtain target individual difference data and target vital sign monitoring data of a target patient, generate a second individual feature vector of the target patient according to the target individual difference data, and screen out a target group from multiple reference groups based on the second individual feature vector;

[0096] A physical sign monitoring and analysis module is used to generate a second physical sign parameter feature vector of a target patient based on the target physical sign monitoring data of the target patient, calculate a physical sign difference factor of the target patient based on a physical sign reference vector of the target group, and generate a physical sign monitoring result of the target patient based on the physical sign difference factor;

[0097] Among them, the physical sign monitoring results of the target patients are generated according to the physical sign difference factors, including:

[0098] Determine the difference reference range corresponding to the target group from the difference factor reference list. The difference factor reference list includes the difference reference range corresponding to each reference group. Judge whether the physical sign difference factor of the target patient belongs to the difference reference range corresponding to the target group. If so, mark the target physical sign monitoring data as normal data; otherwise, mark the target physical sign monitoring data as abnormal data.

[0099] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A method for managing clinical data of anesthesiology, characterized in that: include: Acquire the vital sign monitoring data of multiple sample patients at the target anesthesia stage, wherein the vital sign monitoring data includes vital sign monitoring sequences corresponding to multiple vital sign parameters; Extract features from the vital sign monitoring data of each sample patient, calculate the mean corresponding to each vital sign parameter, and construct the first vital sign parameter feature vector of each sample patient; Clustering multiple sample patients based on the DBSCAN clustering algorithm and multiple first physical sign parameter feature vectors to obtain multiple reference groups, and determining the physical sign reference vector of each reference group according to multiple core points corresponding to each reference group; Obtaining individual difference data of each sample patient, performing data cleaning on the individual difference data and constructing the first individual feature vector of each sample patient, and determining the individual difference reference vector of each reference group according to the first individual feature vector; Obtaining target individual difference data and target vital sign monitoring data of a target patient, generating a second individual characteristic vector of the target patient according to the target individual difference data, and screening out a target group from multiple reference groups based on the second individual characteristic vector; Generating a second physical sign parameter feature vector of the target patient according to the target physical sign monitoring data of the target patient, calculating a physical sign difference factor of the target patient based on a physical sign reference vector of the target group, and generating a physical sign monitoring result of the target patient according to the physical sign difference factor; Determine the individual difference reference vector of each reference group according to the first individual feature vector, including: For any reference group, determining a group center point of the reference group according to multiple core points of the reference group, including calculating the mean of the first individual feature vectors of the multiple core points to determine the group center point; Calculate the distances between multiple core points and the center point of the group respectively, determine the weighting factor of each core point according to the distance between the core point and the center point of the group, and determine the individual difference reference vector of the reference group according to the weighting factor of each core point and the first individual feature vector: ； In the formula, is the individual difference reference vector of the reference group, For the reference group The first volume feature vector of the core points, For the reference group The weighting factor of the core points, For the reference group The distance between the core point and the center point of the group, is a constant term, is the number of core points in the reference group; The physical sign reference vector of each reference group is determined according to the multiple core points corresponding to each reference group, including: The neighborhood density of multiple core points in each reference group is calculated respectively, and the first vital sign parameter feature vectors corresponding to the multiple core points in each reference group are respectively fused according to the neighborhood density of the core points to generate the vital sign reference vector of each reference group. For any vital sign reference vector of a reference group: ； In the formula, is the reference vector of the reference group’s physical signs, For the reference group The first vital sign parameter eigenvector of the core points, For the reference group The neighborhood density of the core points is is the number of core points in the reference group.

2. A method for managing clinical data of anesthesiology according to claim 1, characterized in that: Generate physical sign monitoring results for target patients based on physical sign difference factors, including: Determine the difference reference range corresponding to the target group from the difference factor reference list. The difference factor reference list includes the difference reference range corresponding to each reference group. Judge whether the physical sign difference factor of the target patient belongs to the difference reference range corresponding to the target group. If so, mark the target physical sign monitoring data as normal data; otherwise, mark the target physical sign monitoring data as abnormal data.

3. A method for managing clinical data of anesthesiology according to claim 2, characterized in that: For the difference factor reference list, also include: For any reference group, the distance between each core point in the reference group and the reference sign reference vector of the reference group is calculated to obtain the distance data of the reference group; The distance data of the reference group are fitted based on the normal distribution model, and the probability density function of the distance data is calculated; Calculate the upper and lower limits of the cumulative distribution function of the normal distribution model in the preset confidence interval, and determine the reference range of the difference of the reference group according to the upper and lower limits; The difference reference ranges corresponding to multiple reference groups are calculated, and a difference factor reference list containing the difference reference ranges corresponding to each reference group is constructed.

4. The method for managing clinical data of anesthesiology according to claim 1, characterized in that: The target group is screened out from multiple reference groups based on the second individual feature vector, including: The cosine similarity between the second individual feature vector and the individual difference reference vectors corresponding to multiple reference groups is calculated, and multiple reference groups whose cosine similarity is greater than a preset similarity threshold are screened out, and the reference group with the largest cosine similarity is recorded as the target group.

5. A clinical data management system for anesthesiology, the system being used to implement a clinical data management method for anesthesiology as described in any one of claims 1 to 4, characterized in that: include: A data acquisition module is used to acquire the vital sign monitoring data of multiple sample patients at the target anesthesia stage, wherein the vital sign monitoring data includes vital sign monitoring sequences corresponding to multiple vital sign parameters; The physical sign parameter analysis module is used to extract features from the physical sign monitoring data of each sample patient, calculate the mean value corresponding to each physical sign parameter, and construct the first physical sign parameter feature vector of each sample patient; A sample patient classification module is used to cluster multiple sample patients based on the DBSCAN clustering algorithm and multiple first physical sign parameter feature vectors to obtain multiple reference groups, and determine the physical sign reference vector of each reference group according to multiple core points corresponding to each reference group; An individual difference analysis module is used to obtain individual difference data of each sample patient, clean the individual difference data and construct the first individual feature vector of each sample patient, and determine the individual difference reference vector of each reference group according to the first individual feature vector; A target group screening module is used to obtain target individual difference data and target vital sign monitoring data of a target patient, generate a second individual feature vector of the target patient according to the target individual difference data, and screen out a target group from multiple reference groups based on the second individual feature vector; The vital sign monitoring and analysis module is used to generate a second vital sign parameter feature vector of the target patient based on the target vital sign monitoring data of the target patient, calculate the vital sign difference factor of the target patient based on the vital sign reference vector of the target group, and generate the vital sign monitoring result of the target patient according to the vital sign difference factor.

6. The clinical data management system for anesthesiology according to claim 5, characterized in that: For the physical sign monitoring analysis module, the physical sign monitoring results of the target patient are generated according to the physical sign difference factors, including: Determine the difference reference range corresponding to the target group from the difference factor reference list. The difference factor reference list includes the difference reference range corresponding to each reference group. Judge whether the physical sign difference factor of the target patient belongs to the difference reference range corresponding to the target group. If so, mark the target physical sign monitoring data as normal data; otherwise, mark the target physical sign monitoring data as abnormal data.

Images