An information sharing method for pediatric clinical nursing system

Abstract

This invention relates to the field of medical information processing technology, specifically to an information sharing method for a pediatric clinical nursing system. It involves acquiring complete medical record sample data including age, symptom results, treatment duration, and medication dosage. A comprehensive analysis of the similarity between age and symptom results is conducted to determine differential characteristic values ​​and perform initial clustering, forming clusters. Given the differences in symptom responses and treatment needs among children of different age groups, the distribution and impact characteristics of symptoms are analyzed starting from the symptom results. A symptom to be analyzed is selected, and its distribution density within the cluster and treatment duration are combined to determine salient indicators reflecting distribution characteristics; simultaneously, its frequency of occurrence, differential characteristic values, treatment duration, and medication dosage are analyzed to determine impact indicators. Combining impact and salient indicators, clinical indicators are determined to accurately characterize individual child traits. Cluster analysis based on clinical indicators yields more accurate clustering results, improving the accuracy of information sharing.

Description

Technical Field

[0001] The present invention relates to the technical field of medical information processing, and in particular to an information sharing method of a pediatric clinical nursing system. Background Art

[0002] In the field of pediatric clinical nursing, the health management and treatment process of children rely on the analysis and use of a large amount of medical record data. With the development of medical informatization, more and more pediatric medical institutions have begun to explore the use of information technology to improve the efficiency and accuracy of medical record management and sharing, thereby guiding clinical decision-making and health management of children.

[0003] In terms of information sharing in pediatric clinical care, existing technologies will perform cluster analysis on all information to further achieve information sharing. Specifically, the clustering basis is usually based on the diagnosis results in the medical record data; however, in pediatric clinical practice, the conditions of children are often complex and changeable, and children of different age groups may have large differences in their reactions and treatment needs to the same disease. Therefore, if the medical record data is clustered only based on the diagnosis results, the accuracy of the cluster analysis will be low due to ignoring individual differences, which will in turn affect the accuracy of information sharing. Summary of the invention

[0004] In order to solve the technical problem that in pediatric clinical practice, the conditions of children are often complex and changeable, and the reactions and treatment needs of children of different age groups to the same disease may also vary greatly, if the medical record data is clustered only based on the diagnosis results, the accuracy of cluster analysis will be low due to ignoring individual differences, thereby affecting the accuracy of information sharing, the purpose of the present invention is to provide an information sharing method for a pediatric clinical nursing system, and the technical scheme adopted is as follows:

[0005] Obtaining sample medical record data of all children in the pediatric department, wherein the sample medical record data includes age, symptom results, treatment cycle of each symptom, and medication dosage;

[0006] Comprehensively analyze the differences between the ages and symptoms of all medical record sample data, determine the difference characteristic values ​​between the medical record sample data, and then perform cluster analysis on all medical record sample data to obtain cluster clusters;

[0007] Select any one symptom as the symptom to be analyzed; determine the prominence index of the symptom to be analyzed in all clusters based on the distribution density of the medical record sample data and the treatment cycle of the medical record sample data with the symptom to be analyzed; determine the impact degree value of the symptom to be analyzed according to the number of occurrences of the symptom to be analyzed in all clusters; analyze the difference characteristic values, treatment cycles and medication dosages between all medical record sample data with the symptom to be analyzed, and merge them with the impact degree value of the symptom to be analyzed to determine the impact degree index of the symptom to be analyzed;

[0008] In each medical record sample data, the clinical index of each medical record sample data is determined by combining the impact index and the prominence index of the disease; cluster analysis is performed on all medical record sample data based on their clinical indexes, and information is shared.

[0009] Furthermore, the method for obtaining the difference characteristic value includes:

[0010] Based on Word2vec, the text vector of the disease results in each medical record sample data is obtained;

[0011] For any two medical record sample data, the cosine similarity between the text vectors corresponding to the two medical record sample data is calculated, and the cosine similarity is numerically adjusted to obtain a similarity factor;

[0012] The absolute value of the difference between the age values ​​corresponding to the two medical record sample data is used as the difference factor;

[0013] The sum of the value after negative correlation mapping of the similarity factor between the two medical record sample data and the difference factor is used as the difference characteristic value between the two medical record sample data.

[0014] Furthermore, the method for obtaining the clustering clusters includes:

[0015] A first optimal K value is preset, and all medical record sample data are clustered based on the K-means clustering algorithm and the first optimal K value to obtain cluster clusters, wherein the distance metric is the difference characteristic value between two medical record sample data.

[0016] Furthermore, the method for obtaining the prominence index includes:

[0017] In each cluster, based on the difference characteristic values ​​between the medical record sample data and the number of the medical record sample data, the distribution density of the medical record sample data is determined;

[0018] In each cluster, the mean of the treatment cycles of all the medical record sample data with the disease to be analyzed is calculated as the cycle mean, and the value after negative correlation mapping between the maximum value of the treatment cycle of all the medical record sample data with the disease to be analyzed and the absolute value of the difference between the cycle mean is used as the prominent factor of the disease to be analyzed in each cluster;

[0019] The product of the distribution density of the medical record sample data in each cluster and the corresponding prominence factor is used as the prominence characteristic value of the disease to be analyzed in each cluster;

[0020] The normalized value of the mean of the prominent feature values ​​of the disease to be analyzed in all clusters is used as the prominence index of the disease to be analyzed.

[0021] Furthermore, the method for obtaining the distribution density includes:

[0022] In each cluster, the value after negative correlation mapping of the mean of the difference characteristic values ​​between any two medical record sample data is multiplied by the number of medical record sample data in each cluster, and the obtained product is used as the distribution density of the medical record sample data in each cluster.

[0023] Furthermore, the method for obtaining the impact degree value includes:

[0024] In each cluster, the number of medical record sample data is counted as a quantity factor, and the number of medical record sample data with the disease to be analyzed is counted as a quantity parameter;

[0025] Arrange all clusters in ascending order according to the quantity factor to obtain a sorted sequence;

[0026] The ratio of the maximum number parameter to the number factor of the corresponding cluster is used as the relative frequency value of the disease to be analyzed;

[0027] In the sorting sequence, the ratio of the sequence number value of the cluster corresponding to the maximum number parameter to the total number of clusters is used as the concentration value of the disease to be analyzed;

[0028] The sum of the relative frequency value and the concentration value of the disease to be analyzed is taken as the influence degree value of the disease to be analyzed.

[0029] Furthermore, the method for obtaining the influence index includes:

[0030] Determine the impact adjustment weight of the disease to be analyzed based on the difference characteristic values, treatment cycles and medication dosages among all the medical record sample data with the disease to be analyzed;

[0031] The product of the impact adjustment weight of the disease to be analyzed and the impact degree value is used as the adjustment value, and the sum of the adjustment value and the impact degree value is normalized to a value that is used as the impact degree index of the disease to be analyzed.

[0032] Furthermore, the method for obtaining the impact adjustment weight includes:

[0033] All medical record sample data with the disease to be analyzed are taken as a set;

[0034] In the set, the mean of the difference feature values ​​between any two medical record sample data is calculated as the difference feature mean;

[0035] Add the mean of the treatment cycle and the mean of the medication dosage of all the medical record sample data in the set, and use the sum as the comprehensive index;

[0036] The value obtained by normalizing the product of the mean value of the difference feature and the comprehensive index is used as the impact adjustment weight.

[0037] Furthermore, the method for obtaining the clinical indicators includes:

[0038] The product of the prominence index and the impact index of the symptom to be analyzed is taken as the severity factor of the symptom to be analyzed;

[0039] For any medical record sample data, the value obtained by normalizing the accumulated sum of the severity factors of all symptoms in the medical record sample data is used as the clinical indicator of the medical record sample data.

[0040] Furthermore, cluster analysis is performed on all medical record sample data according to clinical indicators of all medical record sample data, and information sharing is performed, including:

[0041] Preset a second optimal K value, perform cluster analysis on all medical record sample data based on the K-means clustering algorithm and the preset second optimal K value, and obtain a clustering result, wherein the distance metric is the absolute value of the difference between the clinical indicators of the medical record sample data;

[0042] The clustering results are uploaded to the medical resource scheduling system for information sharing.

[0043] The present invention has the following beneficial effects:

[0044] First, the medical record sample data of all children in pediatrics are obtained. The medical record sample data includes age, symptom results, treatment cycle and medication dosage, which ensures the diversity and integrity of the data and provides a solid foundation for subsequent cluster analysis and information sharing. In order to more effectively distinguish the disease characteristics of different age stages, the similarities between age and symptom results in all medical record sample data are comprehensively analyzed, and the difference characteristic values ​​between the medical record sample data are determined to perform initial clustering and obtain cluster clusters. In view of the fact that children of different age groups may have large differences in their reactions and treatment needs to the same disease, in the present invention, we can start with the disease results and analyze all the medical record sample data to determine the distribution and impact characteristics of each disease. Select any disease as the disease to be analyzed, and then determine the prominence index of the disease to be analyzed based on the distribution density of the medical record sample data in the cluster and the treatment cycle of the medical record sample data with the disease to be analyzed. The prominence index is used to reflect the distribution characteristics of the disease to be analyzed in all medical record sample data. At the same time, because the severity of the disease is different, the symptoms of the children will also be different. Therefore, the number of occurrences of the disease to be analyzed in the cluster, the difference characteristic values ​​between the medical record sample data with the disease to be analyzed, the treatment cycle and the dosage of the medicine are analyzed to comprehensively judge the influence index of the disease to be analyzed. The influence index is used to characterize the impact of the disease to be analyzed on the children. There may be multiple diseases in each medical record sample data, so the clinical indicators of each medical record sample data are determined by combining the impact index and prominence index of the disease. The clinical indicators can more accurately characterize the individual characteristics of the children corresponding to each medical record sample data. Finally, clustering analysis of all medical record sample data based on clinical indicators can obtain more accurate clustering results, so that the medical record sample data of children with the same conditions can be classified into one category, which can greatly improve the accuracy of information sharing when sharing information, and promote the efficient integration and utilization of medical record data. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0046] Figure 1 A method flow chart of an information sharing method for a pediatric clinical nursing system provided by one embodiment of the present invention;

[0047] Figure 2 A method flow chart of a method for obtaining a prominence index provided by an embodiment of the present invention;

[0048] Figure 3 A flow chart of a method for obtaining an influence index provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the information sharing method of a pediatric clinical nursing system proposed by the present invention, its specific implementation method, structure, features and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0050] Unless defined otherwise, 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 belongs.

[0051] The specific scheme of the information sharing method of the pediatric clinical nursing system provided by the present invention is described in detail below with reference to the accompanying drawings.

[0052] See also Figure 1 , which shows a method flow chart of an information sharing method of a pediatric clinical nursing system provided by an embodiment of the present invention, the method comprising the following steps:

[0053] Step S1: Obtain sample medical record data of all children in the pediatric department, wherein the sample medical record data includes age, symptom results, treatment cycle of each symptom, and medication dosage.

[0054] In the field of pediatric clinical nursing, the health management and treatment process of children rely on the analysis and utilization of a large amount of medical record data. Therefore, cluster analysis is performed on the medical record sample data of children, so that medical record sample data with similar conditions or symptom reactions are classified into one category. Information sharing can help improve the efficient integration and utilization of medical record data.

[0055] First, obtain the sample medical record data of all children in pediatrics from the medical database. Since the treatment plans for children of different age groups may be different, the sample medical record data mainly includes the age of the child (this helps to understand the differences in disease manifestations, treatment responses and drug metabolism among children of different age groups), the diagnosed symptoms (disease name), the treatment cycle of each disease and the dosage of the medicine. For example, the medical record information of a child is: the child's name is Xiao Ming, age 3, suffering from acute upper respiratory tract infection, which lasted for 5 days, taking antibiotic A, 100 mg twice a day. The medical record information is segmented based on the Jieba word segmentation tool, and the segmentation result is: child / name / Xiao Ming / , / age / 3 years old / , / suffering from / acute upper respiratory tract infection / , / medication / 5 days / , / taking / antibiotic A / , / twice / daily / each time / 100 mg / , and then based on text content recognition and calculation, the age information can be extracted: 3 years old; disease result: acute upper respiratory tract infection; treatment cycle: 5 days; medication dosage 1000 mg, which constitutes the medical record sample data of a child.

[0056] It should be noted that the method of extracting keywords based on the Jieba word segmentation tool is a well-known technology and will not be elaborated here; the collection and acquisition of medical information data in the embodiments of the present invention are all authorized by relevant users, and the process does not violate relevant laws and regulations and does not violate public order and good customs.

[0057] Step S2: Comprehensively analyze the differences between the ages and symptom results in all medical record sample data, determine the difference characteristic values ​​between the medical record sample data, and then perform cluster analysis on all medical record sample data to obtain cluster clusters.

[0058] Even for the same disease, its specific manifestations and treatment methods may vary significantly in different age groups. Therefore, in an embodiment of the present invention, in order to more accurately analyze individual differences, the similarities between the ages and disease results in all medical record sample data can be comprehensively analyzed to determine the difference characteristic values ​​between the medical record sample data, thereby providing a clustering basis for the subsequent cluster analysis process.

[0059] Preferably, in one embodiment of the present invention, the method for obtaining the difference characteristic value includes:

[0060] Word2vec is a commonly used word embedding technology that can map words or phrases in a text into a high-dimensional vector space. Therefore, the text vector of the symptom results in each medical record sample data is obtained based on Word2vec to facilitate the subsequent calculation of the difference feature value.

[0061] For any two medical record sample data, the cosine similarity between the text vectors corresponding to the two medical record sample data is calculated, and the cosine similarity is numerically adjusted to obtain the similarity factor. The formula model of the similarity factor includes:

[0062]

[0063] in, Represents the similarity factor between any two medical record sample data; Represents the cosine similarity between the text vectors corresponding to any two medical record sample data.

[0064] In the formula model of the similarity factor, the value range of cosine similarity is [-1, 1], and the closer it is to 1, the more similar the two text vectors are, and the closer it is to -1, the less similar the two text vectors are. Due to the existence of a negative value interval, in order to facilitate calculations in subsequent processes, in this embodiment of the present invention, the value range of cosine similarity is adjusted using the above formula to obtain the similarity factor, so that the value range of the similarity factor is [0, 1], and similarly, the larger the value, the more similar the two text vectors are, and the smaller the value, the less similar the two text vectors are.

[0065] Then, the age information corresponding to the two medical record sample data is analyzed, and the absolute value of the difference between the age values ​​corresponding to the two medical record sample data is calculated. The smaller the absolute value of the difference, the closer the ages are, and the more likely they are to have similar symptom manifestations and treatment plans. Therefore, the absolute value of the difference is used as the difference factor between the two medical record sample data. The larger the difference factor, the more inconsistent the age information of the two medical record sample data is.

[0066] Finally, the sum of the value of the similarity factor after negative correlation mapping between the two medical record sample data and the difference factor is used as the difference characteristic value between the two medical record sample data. The larger the difference characteristic value, the more inconsistent the basic information (age and disease results) between the two medical record sample data; conversely, the smaller the difference characteristic value, the more consistent the basic information between the two medical record sample data. The negative correlation mapping processing here can be used Function, you can also use Function, where It represents an exponential function with the natural constant e as the base, and x represents the independent variable.

[0067] It should be noted that the process of obtaining text vectors based on Word2vec is a well-known technology and will not be described in detail here.

[0068] At this point, among all the medical record sample data, the difference characteristic value between any two medical record sample data can be calculated based on the above process, and then the difference characteristic value can be used as the basis for clustering, so as to perform cluster analysis on all the medical record sample data and obtain cluster clusters, so that the basic information of the medical record sample data in each cluster cluster (age and comorbidity results) is in a relatively consistent state.

[0069] Preferably, in one embodiment of the present invention, the method for obtaining clusters includes:

[0070] The first optimal K value is preset, and all medical record sample data are clustered based on the K-means clustering algorithm and the first optimal K value to obtain cluster clusters, where the distance metric is the difference characteristic value between two medical record sample data. After cluster analysis, the medical record sample data with relatively consistent basic information can be classified into a cluster cluster.

[0071] It should be noted that the first optimal K value can be set according to the implementation scenario. In the embodiment of the present invention, if it is set to 5, it can also be calculated based on the elbow method and other methods; the elbow method and K-means clustering algorithm are both well-known technologies and will not be described in detail here.

[0072] Step S3: Select any one disease as the disease to be analyzed; in all clusters, determine the prominence index of the disease to be analyzed based on the distribution density of the medical record sample data and the treatment cycle of the medical record sample data with the disease to be analyzed; determine the influence degree value of the disease to be analyzed according to the number of occurrences of the disease to be analyzed in all clusters; analyze the difference characteristic values, treatment cycles and medication dosages between all medical record sample data with the disease to be analyzed, and merge them with the influence degree value of the disease to be analyzed to determine the influence degree index of the disease to be analyzed.

[0073] After preliminary clustering of all medical record sample data, the medical record sample data in each cluster have certain similarities in age and symptom results. In order to further analyze individual differences, in an embodiment of the present invention, based on the clusters, the prominence index and influence index of each symptom are analyzed to reflect the prominence of each symptom in all medical record sample data and the impact it produces. Specifically, the prominence index can mainly analyze the distribution density of the medical record sample data in the clusters and the treatment cycle of the medical record sample data with a certain symptom, thereby characterizing the prevalence of the symptom in the medical record sample data; the influence index can measure the overall impact of a certain symptom in the medical record sample data by analyzing the number of occurrences of a certain symptom in the clusters, the difference characteristic values ​​between the medical record sample data with the symptom, the treatment cycle and the dosage of the medication.

[0074] For the convenience of explanation and illustration, any one disease is selected as the disease to be analyzed, and the prominence index and influence index of the disease to be analyzed are obtained to explain the specific calculation process of these two indicators.

[0075] Preferably, in one embodiment of the present invention, the method for obtaining the prominence index includes:

[0076] See also Figure 2 , which shows a method flow chart of a method for obtaining a prominence index in one embodiment of the present invention, the method comprising the following steps:

[0077] Step S301: In each cluster, the distribution density of the medical record sample data is determined based on the difference characteristic values ​​between the medical record sample data and the number of the medical record sample data.

[0078] In each cluster, the difference characteristic value between any two medical record sample data can be used to measure the consistency between the two medical record sample data. The smaller the difference characteristic value, the more consistent it is. Therefore, it can be regarded as that at the information consistency level, the distribution of medical record sample data in the cluster is more compact, which can be used as a parameter for calculating distribution density. The quantity characteristic can be used as a direct factor to measure the distribution density. The larger the number, the greater the distribution density.

[0079] Therefore, the value after negative correlation mapping of the mean of the difference characteristic values ​​between any two medical record sample data is multiplied by the number of medical record sample data in each cluster, and the obtained product is used as the distribution density of the medical record sample data in each cluster. The mean of the difference characteristic values ​​between any two medical record sample data is calculated. The smaller the mean, the higher the consistency between the medical record sample data in the cluster. Therefore, the mean is negatively correlated and mapped to achieve logical relationship correction. The larger the value after processing, the higher the representation consistency and the more compact the distribution. Therefore, the processed value is multiplied by the number of medical record sample data in the cluster to obtain the distribution density. The larger the distribution density, the more compact the distribution of the medical record sample data in the cluster. The negative correlation mapping processing here can be used. Function, where It represents an exponential function with the natural constant e as the base, and x represents the independent variable.

[0080] Step S302: In each cluster, the treatment periods of all medical record sample data with the disease to be analyzed are integrated to obtain the prominent factors of the disease to be analyzed.

[0081] The length of the treatment cycle can reflect the severity of the disease or the difficulty of treatment. Therefore, in each cluster, the mean of the treatment cycle of all medical record sample data with the disease to be analyzed is calculated as the cycle mean. The larger the cycle mean, the more difficult the treatment of the disease to be analyzed may be. Then, in each cluster, the absolute value of the difference between the maximum value and the cycle mean of the treatment cycle of all medical record sample data with the disease to be analyzed is calculated. The smaller the value, the closer the general level of the treatment cycle of the medical record sample data with the disease to be analyzed is to the maximum value on the basis of relatively consistent basic information of the medical record sample data. This can further explain that the similarity between the medical record sample data in the cluster is high, and the prominence of the disease to be analyzed in the cluster is strong. The value is negatively correlated and mapped to achieve logical relationship correction, so as to obtain the prominence factor of the disease to be analyzed in each cluster. The larger the prominence factor, the more difficult the treatment of the disease to be analyzed is. It is used as an indicator to evaluate the difficulty of treating the disease to be analyzed in the cluster. The negative correlation mapping process here can be used Function, where It represents an exponential function with the natural constant e as the base, and x represents the independent variable.

[0082] Step S303: The distribution density of the medical record sample data in all clusters and the prominence factor of the disease to be analyzed in all clusters are integrated to obtain the prominence index of the disease to be analyzed.

[0083] After the above steps, each cluster corresponds to a distribution density, and the disease to be analyzed corresponds to a prominent factor in each cluster; the product of the distribution density of the medical record sample data in each cluster and the corresponding prominent factor is used as the prominent characteristic value of the disease to be analyzed in each cluster. The larger the prominent characteristic value, the more significant the comprehensive characteristics of the disease to be analyzed in the cluster.

[0084] Finally, the mean of the prominent feature values ​​of the disease to be analyzed in all clusters is normalized as the prominence index of the disease to be analyzed. The larger the prominence index, the more significant the distribution characteristics of the disease to be analyzed in the medical record sample data. Normalization is a technical means well known to those skilled in the art, and the normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0085] The medical record sample data in the same cluster has relatively consistent disease results and age characteristics, but due to the different severity of the disease, the specific symptoms will also be different. Therefore, understanding the impact of each disease will help improve classification accuracy in the subsequent information sharing process.

[0086] The impact can be represented by the number of occurrences of the disease to be analyzed in the cluster, the difference characteristic values ​​between all medical record sample data with the disease to be analyzed, the treatment cycle and the dosage of the medicine, so as to obtain the impact index of the disease to be analyzed.

[0087] Preferably, in one embodiment of the present invention, the method for obtaining the influence index includes:

[0088] See also Figure 3 , which shows a method flow chart of a method for obtaining an influence index in one embodiment of the present invention, the method comprising the following steps:

[0089] Step S311: determining the impact value of the disease to be analyzed according to the number of occurrences of the disease to be analyzed in all clusters.

[0090] In each cluster, the number of medical record sample data is counted as a quantity factor, and the number of medical record sample data with the disease to be analyzed is counted as a quantity parameter. At this time, each cluster corresponds to a quantity factor and a quantity parameter.

[0091] All clusters are sorted in ascending order according to the quantity factor to obtain a sorted sequence, that is, all clusters are sorted in ascending order according to the number of medical record sample data contained.

[0092] The ratio of the maximum number parameter to the number factor of the corresponding cluster is taken as the relative frequency value of the disease to be analyzed. The relative frequency value represents the frequency of occurrence of the disease to be analyzed in the cluster with the most diseases to be analyzed. The larger the value, the higher the frequency of occurrence of the disease to be analyzed, which implies that the prevalence of the disease to be analyzed is higher.

[0093] In the sorting sequence, the ratio of the ordinal value of the cluster corresponding to the maximum number parameter to the total number of clusters is used as the concentration value of the disease to be analyzed. The concentration value can be used to measure the concentration of the disease to be analyzed in all medical record sample data. The larger the ordinal value of the cluster corresponding to the maximum number parameter, the higher the degree of concentration of the disease to be analyzed in the cluster, reflecting the concentration of the disease to be analyzed in all medical record sample data. The larger the concentration value, the higher the impact of the disease to be analyzed, that is, the more common the disease to be analyzed is.

[0094] Finally, the sum of the relative frequency value and the concentration value of the disease to be analyzed is taken as the impact value of the disease to be analyzed. The larger the impact value, the higher the impact of the disease to be analyzed.

[0095] Step S312: Determine the impact adjustment weight of the disease to be analyzed based on the difference characteristic values, treatment cycles, and medication dosages among all the medical record sample data with the disease to be analyzed.

[0096] Since a symptom result in a medical record sample data may contain multiple symptoms, the medical record sample data is first divided according to the type of symptoms, and all the medical record sample data with symptoms to be analyzed are taken as a set.

[0097] In the set, the difference characteristic value between any two medical record sample data reflects the consistency of the basic information of the medical record sample data. The mean of the difference characteristic values ​​between any two medical record sample data in the set is taken as the difference characteristic mean. The larger the difference characteristic mean is, the higher the complexity of the medical record sample data of the disease to be analyzed is, which means that the possibility of the combination of the disease to be analyzed and other types of diseases is higher, and the impact range will be wider.

[0098] The mean of the treatment cycle and the mean of the medication dosage of all the medical record sample data in the set are added together, and the sum is used as the comprehensive index. The larger the comprehensive index is, the greater the difficulty and complexity of the treatment of the medical record sample data of the disease to be analyzed are, which also reflects that the impact of the disease to be analyzed is greater.

[0099] Finally, the product of the difference feature mean and the comprehensive index is normalized to obtain the value as the impact adjustment weight. The larger the impact adjustment weight, the higher the impact weight of the disease to be analyzed. Normalization is a technical means well known to those skilled in the art, and the normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0100] Step S313: The impact degree value of the disease to be analyzed and the impact adjustment weight are integrated to determine the impact degree index of the disease to be analyzed.

[0101] The product of the impact adjustment weight of the disease to be analyzed and the impact value is used as the adjustment value. The larger the adjustment value, the more common the disease to be analyzed is in the medical record sample data, and the greater the possibility of co-occurrence with other diseases. At the same time, the medical record sample data with the disease to be analyzed will also require a higher degree of treatment difficulty, so it is considered that the impact of the disease to be analyzed is greater. Therefore, the sum of the adjustment value and the impact value is finally normalized as the impact index of the disease to be analyzed. The larger the impact index, the higher the impact of the disease to be analyzed. Normalization is a technical means well known to those skilled in the art, and the normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0102] At this point, based on the above steps, the prominence index and influence index of each symptom in the symptom results can be calculated.

[0103] Step S4: In each medical record sample data, the clinical index of each medical record sample data is determined by combining the impact index and the prominence index of the disease; cluster analysis is performed on all medical record sample data according to the clinical indexes of all medical record sample data, and information is shared.

[0104] When a child is sick, multiple symptoms may occur at the same time, and each symptom corresponds to a prominence indicator and an influence indicator, which are used to characterize the degree of influence of each symptom and its distribution in all medical record sample data. Therefore, in each medical record sample data, the clinical indicators of each medical record sample data can be determined based on the influence indicators and prominence indicators of all the symptoms contained. The clinical indicators can more accurately reflect the information contained in each medical record sample data, and take into account factors such as individual differences. Therefore, in the subsequent steps, the clustering results obtained by clustering all medical record sample data according to their clinical indicators will be more accurate, and when information sharing is carried out on this basis, the efficient integration and utilization of medical record data can be effectively improved.

[0105] Preferably, in one embodiment of the present invention, the method for obtaining clinical indicators includes:

[0106] The product of the prominence index and the impact index of the disease to be analyzed is taken as the severity factor of the disease to be analyzed. The severity factor can be used to evaluate the comprehensive situation of the disease to be analyzed in all medical record sample data, including the degree of impact and distribution characteristics.

[0107] Then, for any medical record sample data, the accumulated sum of the severity factors of all the disease results contained in the medical record sample data is normalized and used as the clinical index of the medical record sample data. The clinical index provides a more comprehensive and accurate evaluation index for the medical record sample data. The larger the value, the higher the severity of the disease in the medical record sample data, and the greater the impact. Normalization is a technical means well known to those skilled in the art, and the normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0108] After quantifying the clinical indicator value of each medical record sample data, cluster analysis and information sharing can be performed on all medical record sample data based on the clinical indicators.

[0109] Preferably, in one embodiment of the present invention, cluster analysis is performed on all medical record sample data according to clinical indicators of all medical record sample data, and information sharing is performed, including:

[0110] Cluster analysis helps to discover potential patterns or groups in the data, so the second optimal K value is preset, and cluster analysis is performed on all medical record sample data based on the K-means clustering algorithm and the preset second optimal K value to obtain clustering results, where the distance metric is the absolute value of the difference between the clinical indicators of the medical record sample data.

[0111] The clustering results can provide valuable reference information for doctors and can also help hospitals or medical institutions optimize resource allocation. Therefore, the clustering results can be uploaded to the medical resource scheduling system for information sharing, which can realize the sharing and exchange of medical information and provide doctors with more comprehensive medical record information.

[0112] It should be noted that the second optimal K value can be set according to the implementation scenario. In the embodiment of the present invention, if it is set to 5, it can also be calculated based on the elbow method and other methods; the elbow method and K-means clustering algorithm are both well-known technologies and will not be described in detail here.

[0113] In summary, the embodiment of the present invention first obtains the medical record sample data of all children in pediatrics. The medical record sample data includes age, symptom results, treatment cycle and medication dosage, which ensures the diversity and integrity of the data and provides a solid foundation for subsequent cluster analysis and information sharing. In order to more effectively distinguish the disease characteristics of different age stages, the similarities between age and symptom results in all medical record sample data are comprehensively analyzed, and the difference characteristic values ​​between the medical record sample data are determined to perform initial clustering and obtain cluster clusters. In view of the fact that children of different age groups may have large differences in their responses and treatment needs to the same disease, in an embodiment of the present invention, it is possible to start with the disease results and analyze all medical record sample data to determine the distribution and impact characteristics of each disease. Select any disease as the disease to be analyzed, and then determine the prominence index of the disease to be analyzed based on the distribution density of the medical record sample data in the cluster and the treatment cycle of the medical record sample data with the disease to be analyzed. The prominence index is used to reflect the distribution characteristics of the disease to be analyzed in all medical record sample data. At the same time, because the severity of the disease is different, the symptoms of the children will also be different. Therefore, the number of occurrences of the disease to be analyzed in the cluster, the difference characteristic values ​​between the medical record sample data with the disease to be analyzed, the treatment cycle and the dosage of the medicine are analyzed to comprehensively judge the influence index of the disease to be analyzed. The influence index is used to characterize the impact of the disease to be analyzed on the children. There may be multiple diseases in each medical record sample data, so the clinical indicators of each medical record sample data are determined by combining the impact index and prominence index of the disease. The clinical indicators can more accurately characterize the individual characteristics of the children corresponding to each medical record sample data. Finally, clustering analysis of all medical record sample data based on clinical indicators can obtain more accurate clustering results, so that the medical record sample data of children with the same conditions can be classified into one category, which can greatly improve the accuracy of information sharing when sharing information, and promote the efficient integration and utilization of medical record data.

[0114] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0115] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. An information sharing method for a pediatric clinical nursing system, characterized in that: The method comprises: Obtaining sample medical record data of all children in the pediatric department, wherein the sample medical record data includes age, symptom results, treatment cycle of each symptom, and medication dosage; Comprehensively analyze the differences between the ages and symptoms of all medical record sample data, determine the difference characteristic values ​​between the medical record sample data, and then perform cluster analysis on all medical record sample data to obtain cluster clusters; Select any one symptom as the symptom to be analyzed; determine the prominence index of the symptom to be analyzed in all clusters based on the distribution density of the medical record sample data and the treatment cycle of the medical record sample data with the symptom to be analyzed; determine the impact degree value of the symptom to be analyzed according to the number of occurrences of the symptom to be analyzed in all clusters; analyze the difference characteristic values, treatment cycles and medication dosages between all medical record sample data with the symptom to be analyzed, and merge them with the impact degree value of the symptom to be analyzed to determine the impact degree index of the symptom to be analyzed; In each medical record sample data, the clinical index of each medical record sample data is determined by combining the impact index and the prominence index of the disease; cluster analysis is performed on all medical record sample data according to the clinical indexes of all medical record sample data, and information is shared; The method for obtaining the prominence index includes: In each cluster, based on the difference characteristic values ​​between the medical record sample data and the number of the medical record sample data, the distribution density of the medical record sample data is determined; In each cluster, the mean of the treatment cycles of all the medical record sample data with the disease to be analyzed is calculated as the cycle mean, and the value after negative correlation mapping between the maximum value of the treatment cycle of all the medical record sample data with the disease to be analyzed and the absolute value of the difference between the cycle mean is used as the prominent factor of the disease to be analyzed in each cluster; The product of the distribution density of the medical record sample data in each cluster and the corresponding prominence factor is used as the prominence characteristic value of the disease to be analyzed in each cluster; The normalized value of the mean of the prominent feature values ​​of the disease to be analyzed in all clusters is used as the prominence index of the disease to be analyzed; The method for obtaining the influence index includes: Determine the impact adjustment weight of the disease to be analyzed based on the difference characteristic values, treatment cycles and medication dosages among all the medical record sample data with the disease to be analyzed; The product of the impact adjustment weight of the disease to be analyzed and the impact degree value is used as the adjustment value, and the sum of the adjustment value and the impact degree value is normalized as the impact degree index of the disease to be analyzed; The method for obtaining the clinical indicators includes: The product of the prominence index and the impact index of the symptom to be analyzed is taken as the severity factor of the symptom to be analyzed; For any medical record sample data, the value obtained by normalizing the accumulated sum of the severity factors of all symptoms in the medical record sample data is used as the clinical indicator of the medical record sample data.

2. The information sharing method of a pediatric clinical nursing system according to claim 1, characterized in that: The method for obtaining the difference characteristic value includes: Based on Word2vec, the text vector of the disease results in each medical record sample data is obtained; For any two medical record sample data, the cosine similarity between the text vectors corresponding to the two medical record sample data is calculated, and the cosine similarity is numerically adjusted to obtain a similarity factor; The absolute value of the difference between the age values ​​corresponding to the two medical record sample data is used as the difference factor; The sum of the value after negative correlation mapping of the similarity factor between the two medical record sample data and the difference factor is used as the difference characteristic value between the two medical record sample data.

3. The information sharing method of a pediatric clinical nursing system according to claim 1, characterized in that: The method for obtaining the clusters includes: A first optimal K value is preset, and all medical record sample data are clustered based on the K-means clustering algorithm and the first optimal K value to obtain cluster clusters, wherein the distance metric is the difference characteristic value between two medical record sample data.

4. The information sharing method of a pediatric clinical nursing system according to claim 1, characterized in that: The method for obtaining the distribution density includes: In each cluster, the value after negative correlation mapping of the mean of the difference characteristic values ​​between any two medical record sample data is multiplied by the number of medical record sample data in each cluster, and the obtained product is used as the distribution density of the medical record sample data in each cluster.

5. The information sharing method of a pediatric clinical nursing system according to claim 1, characterized in that: The method for obtaining the impact degree value includes: In each cluster, the number of medical record sample data is counted as a quantity factor, and the number of medical record sample data with the disease to be analyzed is counted as a quantity parameter; Arrange all clusters in ascending order according to the quantity factor to obtain a sorted sequence; The ratio of the maximum number parameter to the number factor of the corresponding cluster is used as the relative frequency value of the disease to be analyzed; In the sorting sequence, the ratio of the sequence number value of the cluster corresponding to the maximum number parameter to the total number of clusters is used as the concentration value of the disease to be analyzed; The sum of the relative frequency value and the concentration value of the disease to be analyzed is taken as the influence degree value of the disease to be analyzed.

6. The information sharing method of a pediatric clinical nursing system according to claim 1, characterized in that: The method for obtaining the impact adjustment weight includes: All medical record sample data with the disease to be analyzed are taken as a set; In the set, the mean of the difference feature values ​​between any two medical record sample data is calculated as the difference feature mean; Add the mean of the treatment cycle and the mean of the medication dosage of all the medical record sample data in the set, and use the sum as the comprehensive index; The value obtained by normalizing the product of the mean value of the difference feature and the comprehensive index is used as the impact adjustment weight.

7. The information sharing method of a pediatric clinical nursing system according to claim 1, characterized in that: The cluster analysis of all medical record sample data based on clinical indicators of all medical record sample data and information sharing include: Preset a second optimal K value, perform cluster analysis on all medical record sample data based on the K-means clustering algorithm and the preset second optimal K value, and obtain a clustering result, wherein the distance metric is the absolute value of the difference between the clinical indicators of the medical record sample data; Upload the clustering results to the medical resource scheduling system for information sharing.

Images