A neurosurgical patient health data intelligent management method

Abstract

The present application relates to the technical field of electronic medical record data management, in particular to a neurosurgery patient health data intelligent management method. The present application first acquires the electronic medical record of the current patient; further acquires the body state score of the patient at each time point according to the deviation characteristics of the health data of each data dimension; further acquires the importance parameter of each treatment node and screens out the important nodes according to the change characteristics of the body state score; further acquires the correction correlation coefficient of each important node and each data dimension; further acquires the value coefficient of the health data of each dimension at each time point according to the time interval of each important node and the current time point, the corresponding correction correlation coefficient and the importance parameter, and the deviation characteristics of the health data of each dimension; and stores the health data based on the value coefficient, optimizes the storage mode of the health data, and relieves the management pressure.

Description

Technical Field

[0001] This invention relates to the field of electronic medical record data management technology, and specifically to an intelligent management method for the health data of neurosurgical patients. Background Technology

[0002] Existing methods for managing patient health data typically utilize electronic medical records, which digitally record and aggregate patients' health data, including medical records, surgical records, laboratory test results, and medication records. This data can be shared across medical institutions, ensuring that patients' health data can be accessed by appropriate healthcare providers anytime, anywhere.

[0003] However, because neurosurgical diseases require a long time from the start of treatment to full recovery, the amount of health data for each patient increases over time. This, coupled with the increasing number of patients, puts increasing pressure on the healthcare system to manage patient health data. Furthermore, the large amount of data generated during long-term treatment also leads to excessive pressure on data analysis and processing, and the transmission and access of large amounts of data puts strain on network bandwidth. Summary of the Invention

[0004] To address the technical problem of excessive data management burden on neurosurgical patients' health data, the present invention aims to provide an intelligent management method for neurosurgical patients' health data. The specific technical solution adopted is as follows:

[0005] Obtain the current patient's electronic medical record; the electronic medical record contains multidimensional health data at multiple time points, records of multiple treatment nodes, and corresponding operation tags;

[0006] At each time point, based on the deviation characteristics of the health data of each data dimension from the preset standard range, the patient's physical status score at each time point is obtained; based on the change characteristics of the physical status score, the importance parameter of each treatment node is obtained; important nodes are selected based on the importance parameter.

[0007] Based on the interval characteristics between the important nodes and the importance parameter of each important node, the determination time domain interval of each important node is obtained; within the determination time domain interval of each important node, based on the drastic fluctuation characteristics of the health data of each data dimension, the correlation coefficient between each important node and each data dimension is obtained; based on the correlation coefficients between all important nodes and each data dimension under each type of operation label and the corresponding correlation coefficients, the corrected correlation coefficient between each important node and each data dimension is obtained.

[0008] Based on the time interval between each important node and the current time point, combined with the corresponding corrected correlation coefficient and importance parameter, and the deviation characteristics of the health data in each dimension, the value coefficient of the health data in each dimension at each time point is obtained; the health data is stored based on the value coefficient.

[0009] Furthermore, the method for obtaining the physical condition score includes:

[0010] A physical condition score is obtained through a formula for calculating the physical condition score; the formula for calculating the physical condition score includes:

[0011] ;

[0012] Where t represents the sequence number of the time point; represents the physical condition score at time point t; J represents the number of data dimensions; j represents the index of the data dimension; This represents the maximum value within the preset standard range of the j-th data dimension; This represents the minimum value within the preset standard range of the j-th data dimension; This represents the median of the preset standard range for the j-th data dimension; This represents the data value of the j-th data dimension at time point t; This indicates taking the absolute value.

[0013] Furthermore, the method for obtaining the importance parameter includes:

[0014] The absolute value of the difference between the body state scores of adjacent time points on both sides of each treatment node is taken as the degree of body change at each treatment node.

[0015] The importance parameter of each treatment node is obtained through the calculation formula of the importance parameter; the calculation formula of the importance parameter includes:

[0016] ;

[0017] Where i represents the time sequence number of the treatment node; This represents the importance parameter of the i-th treatment node; This represents the time interval between the (i-1)th treatment node and the ith treatment node; This indicates the degree of bodily change at the (i-1)th treatment node; This represents the time interval between the (i+1)th treatment node and the ith treatment node. This represents the degree of physical change at the (i+1)th treatment node; This represents the linear normalization function.

[0018] Furthermore, the method for obtaining the important nodes includes:

[0019] Treatment nodes whose importance parameter is greater than a preset importance threshold are marked as important nodes.

[0020] Furthermore, the method for obtaining the time domain interval includes:

[0021] Obtain the average of the time intervals between all adjacent pairs of important nodes, and use this as the base value for the interval;

[0022] The sum of the product of the importance parameter of each important node and the interval base value and the interval base value is used as the interval range length of each important node. The determination time domain interval of each important node is constructed with the recording time point of each important node as the center and the interval range length.

[0023] Furthermore, the method for obtaining the correlation coefficient includes:

[0024] Select any of the aforementioned important nodes as the target node; select any data dimension as the target dimension; within the determination time domain interval of the target node, obtain a data sequence composed of the health data of the target dimension; the health data in the data sequence are arranged in chronological order;

[0025] The product of the mean absolute deviation and the mean frequency of the data sequence is normalized and used as the correlation coefficient between the target node and the target dimension.

[0026] Furthermore, the method for obtaining the corrected correlation coefficient includes:

[0027] When the correlation coefficient between the target node and the target dimension is greater than a preset correlation threshold, it is determined to be a strong correlation;

[0028] Among all the treatment nodes of the same type of operation label of the target node, the probability that the target dimension is determined to be strongly correlated is obtained as the correlation probability;

[0029] The product of the association probability and the association coefficient corresponding to the target node is used as the modified association coefficient of the target node.

[0030] Furthermore, the method for obtaining the value coefficient includes:

[0031] The reference coefficients for each health data point in each data dimension within the determination time domain interval of each treatment node are obtained using the reference coefficient calculation formula; the reference coefficient calculation formula includes:

[0032] ;

[0033] in, The sequence number represents the time sequence number of the important node; j represents the sequence number of the data dimension; k represents the sequence number of the health data. Indicates the first Within the time domain interval for determining important nodes, the reference coefficient of the k-th health data under the j-th data dimension; Indicates the first The time interval between each important node and the current time point; Indicates the first Importance parameters of each important node; Indicates the first The corrected correlation coefficient between each important node and the j-th data dimension; Indicates the first Within the time domain interval for determining important nodes, the data value of the kth health data in the jth data dimension; Represents a linear normalization function;

[0034] The largest reference coefficient corresponding to the health data of each dimension at each time point is used as the value coefficient of the health data of each dimension at each time point.

[0035] Furthermore, the method for storing the health data based on the value coefficient includes:

[0036] Health data with a value coefficient less than a preset value threshold are cold-stored; health data with a value coefficient greater than or equal to the preset value threshold are directly stored.

[0037] Furthermore, when a patient has multiple electronic medical records, the value coefficient of the health data at each time point and for each dimension in each electronic medical record is obtained, and the health data is stored based on the value coefficient.

[0038] The present invention has the following beneficial effects:

[0039] This invention first acquires the current patient's electronic medical record, providing a basis for subsequent intelligent management of the patient's health data; it then acquires the patient's physical status score at each time point, characterizing the patient's physical status at the corresponding time point, facilitating subsequent analysis of changes in the patient's physical status; it further acquires the importance parameters of each treatment node, reflecting the impact characteristics of the treatment node on the patient's body and characterizing the importance of the treatment node; it further filters important nodes based on the importance parameters, facilitating the focus on important data; it further acquires the judgment time domain interval of each important node, providing a time range constraint, focusing on health data within key time periods, and avoiding redundant processing of the entire dataset; and it further acquires the correlation coefficient between each important node and each data dimension, quantifying the association between important nodes and health data in each dimension. The invention further refines the correlation coefficients from the overall perspective of the correlation between important nodes and each data dimension, obtaining a revised correlation coefficient between each important node and each data dimension to improve the accuracy and reliability of the revised correlation degree. Furthermore, based on the time interval between each important node and the current time point, combined with the corresponding revised correlation coefficients and importance parameters, and the deviation characteristics of health data for each dimension, the invention obtains the value coefficients of health data for each dimension at each time point from multiple perspectives, including time interval, importance of important nodes, correlation between important nodes and health data, and abnormal fluctuations in the health data itself. This quantifies the reference value of each health data point, providing a basis for optimizing the storage method of health data. Finally, the health data is stored based on the value coefficients to alleviate the management pressure of health data. This invention analyzes changes in the patient's physical condition, accurately assessing the value coefficients of health data from multiple perspectives, including time interval, importance of important nodes, correlation between important nodes and health data, and abnormal fluctuations in the health data itself, thereby optimizing the storage method and alleviating the management pressure of health data. Attached Figure Description

[0040] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 A flowchart illustrating an intelligent management method for neurosurgical patient health data provided in one embodiment of the present invention;

[0042] Figure 2 This is a flowchart of a method for obtaining a modified correlation coefficient, provided as an embodiment of the present invention. Detailed Implementation

[0043] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an intelligent management method for neurosurgical patient health data proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0044] 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 invention pertains.

[0045] The following description, in conjunction with the accompanying drawings, details a specific scheme for an intelligent management method for neurosurgical patient health data provided by the present invention.

[0046] Please see Figure 1 The diagram illustrates a flowchart of an intelligent management method for neurosurgical patient health data according to an embodiment of the present invention, specifically including:

[0047] Step S1: Obtain the current patient's electronic medical record; the electronic medical record contains multidimensional health data at multiple time points, records of multiple treatment nodes, and corresponding operation tags.

[0048] In this embodiment of the invention, the electronic medical record of the current neurosurgical patient is first obtained to provide a basis for subsequent intelligent management of the patient's health data. The electronic medical record contains multidimensional health data at multiple time points during the patient's treatment process to reflect the patient's physical condition. It also includes records of multiple treatment nodes and corresponding operation tags, and records the treatment plan.

[0049] In one embodiment of the present invention, the electronic medical record includes at least the patient's basic personal information, the creation time of each page of the medical record, the doctor's diagnosis time, the name and dosage of the medication used, and the drug allergy status; the multidimensional health data includes at least: blood pressure, heart rate, body temperature, respiratory rate, Glasgow Coma Scale (GCS) score, and intracranial pressure (ICP); the treatment nodes include at least surgical nodes, follow-up nodes, and recurrence nodes; the record of the treatment node includes at least the recording time and detailed plan; and operation labels are assigned according to the operation category of the treatment node, such as the operation label 1 for medication prescription and the operation label 2 for surgical operation.

[0050] It should be noted that the patient's basic personal information can be set according to actual needs, and should at least include identity, gender, age and weight information; in other embodiments of the present invention, the implementer can add or delete the content contained in the electronic medical record, but it is necessary to ensure that each piece of information has a recording time point.

[0051] It should be noted that, considering that patients may visit multiple hospitals and there may be multiple electronic medical records, in one embodiment of the present invention, the electronic medical records are analyzed and intelligently stored one by one, and the intelligent management method of each electronic medical record is consistent. Here, only one electronic medical record is described as an example. In another embodiment of the present invention, the implementer may also sort each page of the medical record in multiple electronic medical records according to the time order, merge them into one electronic medical record for analysis and intelligent storage.

[0052] Step S2: At each time point, based on the deviation characteristics of the health data of each data dimension from the preset standard range, obtain the patient's physical status score at each time point; based on the change characteristics of the physical status score, obtain the importance parameter of each treatment node; and filter out important nodes based on the importance parameter.

[0053] Given that different treatment milestones vary significantly in their importance to patients during treatment, and that the changes in a patient's physical condition before and after a treatment milestone reflect the impact of that milestone on the patient's body, it is necessary to assess the patient's physical condition in order to analyze these changes.

[0054] Considering the deviation characteristics of health data from the preset standard range, which reflect the differences between health data and the normal range, and thus reflect the patient's physical condition, at each time point, based on the deviation characteristics of health data for each data dimension from the preset standard range, a physical condition score is obtained for the patient at each time point, representing the patient's physical condition at the corresponding time point. Based on the change characteristics of the physical condition score, an importance parameter for each treatment node is obtained, representing the importance of the treatment node. Then, based on the importance parameter, important nodes that have a significant impact on the patient's body are selected, making it easier to focus on important data.

[0055] Preferably, in one embodiment of the present invention, considering that the greater the difference between the health data of each dimension and the median of the preset standard range, the greater the deviation of the health data and the more obvious the deviation characteristics; and considering that the order of magnitude of health data in different dimensions may differ, it is also necessary to standardize the difference between the health data and the median of the preset standard range, and to integrate the deviation characteristics of health data in all dimensions, thereby constructing a calculation formula for the physical condition score; the physical condition score is obtained through the calculation formula; the calculation formula for the physical condition score includes:

[0056] ;

[0057] Where t represents the sequence number of the time point; represents the physical condition score at time point t; J represents the number of data dimensions; j represents the index of the data dimension; This represents the maximum value within the preset standard range of the j-th data dimension; This represents the minimum value within the preset standard range of the j-th data dimension; This represents the median of the preset standard range for the j-th data dimension; This represents the data value of the j-th data dimension at time point t; This indicates taking the absolute value.

[0058] The formula for calculating the physical condition score uses the absolute value of the difference to represent the difference between health data and the median of the preset standard range. The larger the value, the greater the difference between the health data and the median of the corresponding preset standard range, and the more obvious the deviation characteristics of the health data. This represents half the length of the preset standard range, used for... Standardize the data to eliminate the influence of order of magnitude and unit on health data from different dimensions.

[0059] It should be noted that the preset standard range for each dimension of health data is the normal range for each type of health data, based on existing data, and will not be elaborated further here.

[0060] Preferably, in one embodiment of the present invention, considering that the greater the difference in body state scores before and after a treatment node, the greater the degree of body change, the absolute value of the difference in body state scores between adjacent time points on both sides of each treatment node is taken as the degree of body change at each treatment node.

[0061] Considering that the greater the degree of bodily change and the shorter the time interval between adjacent treatment nodes for a given treatment node, the stronger the urgency and continuity of this node during treatment, and the greater its importance, a formula for calculating importance parameters is constructed based on this. The importance parameters for each treatment node are obtained through this formula. The formula for calculating importance parameters includes:

[0062] ;

[0063] Where i represents the time sequence number of the treatment node; This represents the importance parameter of the i-th treatment node; This represents the time interval between the (i-1)th treatment node and the ith treatment node; This indicates the degree of bodily change at the (i-1)th treatment node; This represents the time interval between the (i+1)th treatment node and the ith treatment node. This represents the degree of physical change at the (i+1)th treatment node; This represents the linear normalization function.

[0064] In the formula for calculating the importance parameter, and The larger the value, the smaller the time interval between the i-th treatment node and its two adjacent treatment nodes, and the greater the degree of physical change in the adjacent treatment nodes, reflecting the greater importance of the i-th treatment node and the larger the importance parameter.

[0065] It should be noted that when calculating the degree of physical change, the nearest physical state scores on both sides of the treatment node in the time domain are taken as the basis for calculation.

[0066] Preferably, in one embodiment of the present invention, treatment nodes with an importance parameter greater than a preset importance threshold are marked as important nodes.

[0067] As an example, the preset importance threshold is 0.6.

[0068] Step S3: Based on the interval characteristics between important nodes and the importance parameters of each important node, obtain the judgment time domain interval for each important node; within the judgment time domain interval of each important node, based on the drastic fluctuation characteristics of the health data of each data dimension, obtain the correlation coefficient between each important node and each data dimension; based on the correlation coefficients of all important nodes under each operation label and each data dimension, and combined with the corresponding correlation coefficients, obtain the corrected correlation coefficient between each important node and each data dimension.

[0069] In this embodiment of the invention, considering that the health data of patients during long-term treatment contains repetitive data with low reference value, the storage method is adjusted based on the value coefficient of the health data to alleviate the pressure of data management.

[0070] When health data in a certain dimension fluctuates significantly before and after a critical node, the changes in the patient's physical condition are more likely to be caused by fluctuations in health data in that dimension, indicating a stronger correlation between that data dimension and the critical node. Since the entire treatment process spans a long period of time and includes multiple treatment stages, it is necessary to determine the analysis time domain range for each critical node in order to analyze the relevant features.

[0071] Considering that the interval characteristics between important nodes reflect the time span and rhythm of changes in patients' health status, and the importance parameter of each important node characterizes the degree of importance of the important node and reflects the scale of its time-domain influence, the judgment time domain interval of each important node is obtained based on the interval characteristics between important nodes and the importance parameter of each important node. This provides a constraint on the time range, focuses on health data within key time periods, and avoids redundant processing of the full dataset. Within the judgment time domain interval of each important node, the correlation coefficient between each important node and each data dimension is obtained based on the drastic fluctuation characteristics of health data in each data dimension. This quantifies the correlation strength between important nodes and health data in each dimension, which helps in the subsequent evaluation of the value coefficient of health data.

[0072] Preferably, in one embodiment of the present invention, the average value of the time interval between all two adjacent important nodes is obtained as the basic value of the interval, which reflects the normal time interval between important nodes and provides a basic value for obtaining the determination time domain interval;

[0073] Considering that the larger the importance parameter of an important node, the more important the node is, reflecting that the node has a greater impact on the patient's health status, and the larger the time range that needs to be analyzed, in order to more comprehensively assess the long-term impact of the important node on the patient's health status, the product of the importance parameter of each important node and the interval base value, and the sum of the interval base value, are used as the interval range length of each important node. Taking the recording time point of each important node as the center, the judgment time domain interval of each important node is constructed based on the interval range length.

[0074] Preferably, in one embodiment of the present invention, any important node is selected as the target node; any data dimension is selected as the target dimension, which facilitates analysis one by one; within the determination time domain interval of the target node, a data sequence composed of health data of the target dimension is obtained; the health data in the data sequence is arranged in chronological order, and the health data of the target dimension within the determination time domain interval of the target node is intuitively represented by the data sequence, which is more conducive to analyzing the drastic fluctuation characteristics of health data;

[0075] Considering that the larger the mean absolute deviation of the data sequence, the more the data deviates from the average value and the more drastic the fluctuation; the larger the average frequency, the stronger the volatility and the more drastic the fluctuation; the more drastic the fluctuation of health data, the greater the impact of important nodes on the health data of the target dimension, and the greater the correlation between the two, the product of the mean absolute deviation and the average frequency of the data sequence is normalized and used as the correlation coefficient between the target node and the target dimension.

[0076] It should be noted that mean absolute deviation is an existing technology, which is normalized by linear normalization. In one embodiment of the present invention, the frequency of the data sequence is obtained by arbitrary sampling Fourier transform (ASFT) as the average frequency. Implementers may also use other methods such as anti-leakage Fourier transform (ALFT) to obtain the average frequency of the data sequence. These are all technical means well known to those skilled in the art and will not be described in detail here.

[0077] Considering that key nodes with the same operation label belong to the same type of treatment node, all key nodes with the same operation label have similar effects in the treatment process, thus exhibiting similar correlation characteristics for the same data. Therefore, based on the correlation coefficients of all important nodes under each type of operation label with each data dimension, and combined with the corresponding correlation coefficients, a corrected correlation coefficient is obtained between each important node and each data dimension. From the perspective of the overall correlation between important nodes of the same type and each data dimension, the correlation coefficient is corrected to improve the accuracy and reliability of the corrected correlation degree.

[0078] Preferably, in one embodiment of the present invention, the method for obtaining the modified correlation coefficient includes:

[0079] Please see Figure 2 The flowchart illustrates a method for obtaining a modified correlation coefficient according to an embodiment of the present invention, specifically including:

[0080] Step S301: When the correlation coefficient between the target node and the target dimension is greater than the preset correlation threshold, it is determined to be a strong correlation.

[0081] Considering that a larger correlation coefficient indicates a greater degree of correlation, a strong correlation is determined when the correlation coefficient between the target node and the target dimension is greater than the preset correlation threshold; as an example, the preset correlation threshold is 0.6.

[0082] Step S302: Among all treatment nodes with the same operation label of the target node, obtain the probability that the target dimension is determined to be strongly correlated as the correlation probability.

[0083] Considering that under a certain type of operation label, the greater the probability that important nodes and target dimensions are determined to be strongly correlated, it indicates a greater degree of correlation between important nodes and target dimensions from an overall perspective; therefore, among all treatment nodes of the same type of operation label for the target node, the probability that the target dimension is determined to be strongly correlated is taken as the correlation probability.

[0084] Step S303: The product of the association probability and association coefficient corresponding to the target node is used as the corrected association coefficient of the target node.

[0085] Since the correlation coefficient represents the degree of correlation between the target node and the target dimension from a local perspective, we further integrate the correlation probability from the overall correlation perspective and use the product of the correlation probability and the correlation coefficient corresponding to the target node as the corrected correlation coefficient of the target node.

[0086] Iterate through all target nodes and target dimensions to obtain the corrected correlation coefficient between each important node and each data dimension.

[0087] Step S4: Based on the time interval between each important node and the current time point, combined with the corresponding corrected correlation coefficient and importance parameter, and the deviation characteristics of health data in each dimension, obtain the value coefficient of health data in each dimension at each time point; store the health data based on the value coefficient.

[0088] Considering the time interval between important nodes and the current time point, the time domain reflects the reference value of multidimensional health data within the time domain interval for judging important nodes to the current time point; the importance parameter represents the importance of important nodes, and the corrected correlation coefficient represents the correlation between important nodes and health data of each data dimension. Combining the two reflects the importance of each health data point to the current time point; while the deviation characteristics of health data reflect abnormal fluctuations in health data, representing the degree of attention required for health data.

[0089] Therefore, based on the time interval between each important node and the current time point, combined with the corresponding corrected correlation coefficient and importance parameter, as well as the deviation characteristics of health data in each dimension, the value coefficient of health data in each dimension at each time point is obtained, the reference value of each health data point is quantified, and a basis is provided for optimizing the storage method of health data and alleviating management pressure.

[0090] Preferably, in one embodiment of the present invention, considering that the larger the time interval between the important node and the current time point, the more treatment stages exist between the important node and the current time point, the worse the timeliness of the important node, and the less referential the health data within the corresponding judgment time interval; the larger the importance parameter, the more important the important node is to the patient, the more important the health data within the corresponding judgment time interval, and the greater the referential value; the larger the corrected correlation coefficient, the greater the correlation between the important node and the health data of the corresponding data dimension, the more the health data of the corresponding dimension is a key indicator of the patient's physical condition change, and the greater the referential value; at the same time, the greater the difference between the health data and the median of the preset standard range of the corresponding dimension, the more obvious the deviation characteristics of the health data, the more it reflects the abnormal state of the patient's body, and the greater the referential value.

[0091] Based on this, a formula for calculating the reference coefficient is constructed. This formula is used to obtain the reference coefficient for each health data point in each data dimension within the time domain interval of each treatment node. The formula for calculating the reference coefficient includes:

[0092] ;

[0093] in, The sequence number represents the time sequence number of the important node; j represents the sequence number of the data dimension; k represents the sequence number of the health data. Indicates the first Within the time domain interval for determining important nodes, the reference coefficient of the k-th health data under the j-th data dimension; Indicates the first The time interval between each important node and the current time point; Indicates the first Importance parameters of each important node; Indicates the first The corrected correlation coefficient between each important node and the j-th data dimension; Indicates the first Within the time domain interval for determining important nodes, the data value of the kth health data in the jth data dimension; Represents a linear normalization function;

[0094] The formula for calculating the reference coefficient uses the reciprocal to negatively correlate the time interval, adjusting the logical relationship between the time interval and the reference coefficient. The larger the value, the less reliable it is. The smaller the value, the better; the difference between the health data and the median of the preset standard range is expressed by the absolute value of the difference, and by using... right Standardize the data to represent deviations from the health data. The larger the value, the more relevant it is to the overall picture. The larger the value, the more accurate the reference value of each health data point in each data dimension can be. This is achieved by multiplying various parameters and adjusting the range of reference coefficients through linear normalization, thus providing a basis for optimizing storage methods.

[0095] Considering that there may be overlap in the time domains of judgment between different important nodes, and that multidimensional health data at some time points may fall within multiple time domains simultaneously, since the largest reference coefficient represents the maximum reference value of health data, the largest reference coefficient corresponding to each dimension of health data at each time point is used as the value coefficient of each dimension of health data at each time point.

[0096] It should be noted that since the time domain range for determining key nodes is greater than the basic value of the interval, and the time domain ranges for determining all key nodes are connected, they can usually include multidimensional health data at all time points. For multidimensional health data that is not within the time domain range for determining any key node, it indicates that this multidimensional key data may be routine data that is not directly related to key treatment nodes, and the value coefficient is set to 0.

[0097] After characterizing the reference value of health data by the value coefficient, the health data can be stored based on the value coefficient, optimizing the storage method, alleviating the management pressure of health data, and making it more convenient for health data retrieval and analysis.

[0098] Preferably, in one embodiment of the present invention, considering that the smaller the value coefficient, the smaller the reference value of the corresponding health data, and that cold storage can store data in a storage medium with low access frequency and low cost, thereby not occupying the valuable resources of the main storage system, making the overall storage management more efficient; at the same time, cold storage can record and save data for a long time to meet the storage time requirements of medical data; therefore, health data with a value coefficient less than the preset value threshold are cold stored; and health data with a value coefficient greater than or equal to the preset value threshold are directly stored.

[0099] As an example, the preset value threshold is 0.4; when the stored health data reaches the storage time limit, the health data can be deleted to free up storage space.

[0100] By adopting different storage models and leveraging the advantages of cold storage, we can reduce the storage costs and space requirements of health data with low reference value, enable rapid access to important data, improve overall storage efficiency, alleviate the pressure of health data management, and enhance data management flexibility.

[0101] In summary, to address the technical problem of excessive data management pressure on neurosurgical patients' health data, this invention proposes an intelligent management method for neurosurgical patients' health data. This invention first acquires the current patient's electronic medical record; then, based on the deviation characteristics of health data in each data dimension, it obtains the patient's physical status score at each time point; further, based on the change characteristics of the physical status score, it obtains the importance parameter of each treatment node and filters out important nodes; further, it obtains the corrected correlation coefficient between each important node and each data dimension; further, based on the time interval between each important node and the current time point, combined with the corresponding corrected correlation coefficient and importance parameter, and the deviation characteristics of health data in each dimension, it obtains the value coefficient of health data in each dimension at each time point; and finally, it stores the health data based on the value coefficient, optimizing the storage method and alleviating management pressure.

[0102] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0103] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for intelligent management of health data of neurosurgical patients, characterized in that, The method includes: Retrieve the current patient's electronic medical record; the electronic medical record contains multidimensional health data at multiple time points, records of multiple treatment nodes, and corresponding operation tags; At each time point, based on the deviation characteristics of health data in each data dimension from the preset standard range, the patient's physical status score at each time point is obtained; based on the change characteristics of the physical status score, the importance parameter of each treatment node is obtained; important nodes are selected based on the importance parameter. Based on the interval characteristics between important nodes and the importance parameters of each important node, the judgment time domain interval of each important node is obtained; within the judgment time domain interval of each important node, based on the drastic fluctuation characteristics of health data of each data dimension, the correlation coefficient between each important node and each data dimension is obtained; based on the correlation coefficient between all important nodes under each type of operation label and each data dimension, combined with the corresponding correlation coefficient, the corrected correlation coefficient between each important node and each data dimension is obtained. Based on the time interval between each important node and the current time point, combined with the corresponding corrected correlation coefficient and importance parameter, and the deviation characteristics of health data in each dimension, the value coefficient of health data in each dimension at each time point is obtained; Methods for obtaining the value coefficient include: The reference coefficient is obtained for each health data point in each data dimension within the time domain interval of each treatment node, using the formula for calculating the reference coefficient. The formula for calculating the reference coefficient includes: ； in, The sequence number represents the time sequence number of the important node; j represents the sequence number of the data dimension; k represents the sequence number of the health data. Indicates the first Within the time domain interval for determining important nodes, the reference coefficient of the k-th health data under the j-th data dimension; Indicates the first The time interval between each important node and the current time point; Indicates the first Importance parameters of each important node; Indicates the first The corrected correlation coefficient between each important node and the j-th data dimension; Indicates the first Within the time domain interval for determining important nodes, the data value of the kth health data in the jth data dimension; Represents a linear normalization function; The maximum reference coefficient corresponding to the health data of each dimension at each time point is used as the value coefficient of the health data of each dimension at each time point. Health data is stored based on a value coefficient. Health data with a value coefficient less than a preset value threshold is cold-stored, while health data with a value coefficient greater than or equal to the preset value threshold is stored directly.

2. The intelligent management method for health data of neurosurgical patients according to claim 1, characterized in that, Methods for obtaining a physical condition score include: A physical condition score is obtained through a formula for calculating the physical condition score; the formula for calculating the physical condition score includes: ； Where t represents the sequence number of the time point; represents the physical condition score at time point t; J represents the number of data dimensions; j represents the index of the data dimension; This represents the maximum value within the preset standard range of the j-th data dimension; This represents the minimum value within the preset standard range of the j-th data dimension; This represents the median of the preset standard range for the j-th data dimension; This represents the data value of the j-th data dimension at time point t; This indicates taking the absolute value.

3. The intelligent management method for health data of neurosurgical patients according to claim 1, characterized in that, Methods for obtaining importance parameters include: The absolute value of the difference between the body state scores of adjacent time points on both sides of each treatment node is taken as the degree of body change at each treatment node. The importance parameters for each treatment node are obtained through the calculation formula for the importance parameters; the calculation formula for the importance parameters includes: ； Where i represents the time sequence number of the treatment node; This represents the importance parameter of the i-th treatment node; This represents the time interval between the (i-1)th treatment node and the ith treatment node; This indicates the degree of physical change at the (i-1)th treatment node; This represents the time interval between the (i+1)th treatment node and the ith treatment node. This represents the degree of physical change at the (i+1)th treatment node; This represents the linear normalization function.

4. The intelligent management method for health data of neurosurgical patients according to claim 3, characterized in that, Methods for obtaining important nodes include: Treatment nodes whose importance parameter is greater than a preset importance threshold are marked as important nodes.

5. The intelligent management method for health data of neurosurgical patients according to claim 1, characterized in that, Methods for determining the time domain interval include: Obtain the average of the time intervals between all adjacent important nodes, and use it as the base value for the interval; The product of the importance parameter of each important node and the basic value of the interval, and the sum of the product and the basic value of the interval, are used as the interval range length of each important node. The judgment time domain interval of each important node is constructed with the recording time point of each important node as the center and the interval range length.

6. The intelligent management method for health data of neurosurgical patients according to claim 1, characterized in that, Methods for obtaining correlation coefficients include: Select any important node as the target node; select any data dimension as the target dimension; within the time domain interval for determining the target node, obtain the data sequence consisting of health data of the target dimension; the health data in the data sequence are arranged in chronological order; The product of the mean absolute deviation and the mean frequency of the data sequence is normalized and used as the correlation coefficient between the target node and the target dimension.

7. The intelligent management method for health data of neurosurgical patients according to claim 6, characterized in that, Methods for obtaining the corrected correlation coefficient include: When the correlation coefficient between a target node and a target dimension is greater than a preset correlation threshold, it is determined to be a strong correlation. Among all treatment nodes with the same operation label of the target node, the probability that the target dimension is determined to be strongly correlated is taken as the correlation probability. The product of the association probability and association coefficient corresponding to the target node is used as the corrected association coefficient of the target node.

8. The intelligent management method for health data of neurosurgical patients according to claim 1, characterized in that, When a patient has multiple electronic medical records, the value coefficient of health data for each dimension at each time point is obtained in each electronic medical record, and the health data is stored based on the value coefficient.

Images