Accurate medication aid decision-making method and system

By constructing the patient's entity and attribute relationship diagram in the clinical decision support system and using the relationship tree to calculate the similarity, the problem of inaccurate outline diagram construction in the existing technology is solved, and the accuracy and comprehensiveness of drug use decisions is improved, helping Doctors make more reasonable medication plans.

CN118969178AInactive Publication Date: 2024-11-15BEIJING FRIENDSHIP HOSPITAL CAPITAL MEDICAL UNIV
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202411452638.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-17
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

In the drug use decision-making, the existing clinical decision-making support system does not consider the entity attribute relationship during the contour map construction process, resulting in the contour map not accurately reflecting the medical record information, and the similarity calculation accuracy is low, which cannot assist doctors in making accurate drug use decisions.

Method used

The entity and attribute relationship of each patient is constructed in the knowledge base, a contour map is generated, and the distance between the graphs is calculated through the relationship tree to achieve a comprehensive feature similarity comparison between the contour maps, and assist doctors in making medication decisions.

Benefits of technology

Through multiple set calculations of relationship trees, the accuracy of similarity comparison between outline maps is improved, helping doctors make more accurate and comprehensive drug use decisions, and reducing the risk of unreasonable drug abuse and use.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118969178A_ABST
    Figure CN118969178A_ABST
Patent Text Reader

Abstract

According to the accurate medication aid decision-making method and system provided by the invention, in the construction process of the knowledge base, the contour map for each patient can be constructed according to the entity and attribute relationship of each patient in the obtained resources, and the known entity and attribute relationship are input into the medication aid decision-making system; the concept of the relation tree is introduced when the inter-graph distance between the contour graph of a new doctor-seeing patient and the contour graphs of the multiple patients in the knowledge base is calculated in the auxiliary decision making process, the contour graphs are converted into multiple sets of the relation tree corresponding to the contour graphs, and therefore the similarity between the doctor-seeing contour graphs can be obtained; according to the method, a doctor can refer to a plurality of doctor-seeing records with the highest similarity to make a decision for accurate medication of the doctor, and due to the fact that multiple sets of the relation tree can reflect more contour map features, the similarity is determined according to the comprehensive features of the contour map, and more accurate and comprehensive boosting significance is provided for the medication decision made by the doctor.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of medical decision-making assistance technology, and more specifically, to a precise medication decision-making assistance method and system. Background Art

[0002] In recent years, with the continuous development of medical care, precision medicine in the treatment of diseases has received more and more attention, especially for infants and young children. Infancy is an important stage of human life, and their growth and development directly affect the individual's health level and quality of life. In addition, due to factors such as immature liver and kidney function, imperfect blood-brain barrier, and low plasma protein binding rate in infants and young children, precision medicine plays a vital role in optimizing infant medicine. However, the current medication in the treatment process is generally still based on the personal medical experience of the attending physician. There are many drug abuses and irrational drug use. In many cases, it is even similar to finding a suitable medication plan through trial and error, which increases the patient's medication risk and may cause irreversible adverse effects on their long-term health development. Against the backdrop of digital transformation, digital and intelligent dual-drive uses data and intelligent technology to drive decision-making. It is a higher-level transformation based on digitalization and is reshaping the medical industry. Smart medicine with digital and intelligent dual-drive as the core is booming. Clinical decision support system (DSS) as a part of artificial intelligence The purpose of this paper is to provide clinical decision support to clinical physicians and nursing staff and assist clinical diagnosis and treatment. The similarity query of electronic medical records is a key link in the decision support system, which can determine the accuracy of similarity query and thus determine the accuracy of auxiliary diagnosis. In the prior art, contour similarity calculation is a commonly used method in the field of computer vision and image processing, which is used to compare the similarity between two images. Some scholars have tried to apply contour similarity calculation to medical record similarity query for clinical decision support. However, the attribute connection relationship between entities is not considered in the process of constructing the medical contour map, resulting in the constructed medical contour map not being able to accurately reflect the medical record information. Moreover, the similarity calculation based on the minimum distance between two contour maps in the prior art is still used to measure the similarity. The minimum distance can only represent limited features, and the overall feature comparison of the medical record cannot be achieved, which makes the similarity query based on the contour map in the decision support less accurate, resulting in the auxiliary decision reference given by the boosting auxiliary system being of little significance, and unable to assist doctors in making accurate medication decisions. Summary of the invention

[0003] In view of the above problems, the present invention proposes a method and system for assisting decision-making in accurate medication. In the construction of a knowledge base, the method constructs a profile map of the entity and attribute relationship of each patient. When there is a new patient, the doctor generates a profile map of the current diagnosis and treatment by inputting the entity and attribute relationship of the patient into the accurate medication assistance decision-making system. In the process of generating the diagnosis and treatment profile map, not only the entity but also the attribute relationship between the entities is considered. In the auxiliary decision-making process, the concept of relationship tree is introduced when calculating the distance between the profile map of the new patient and the profile maps of multiple patients in the knowledge base, and each profile map is converted into a multiple set of its corresponding relationship tree. The distance between the graphs is calculated based on the multiple sets of the relationship tree, so that the similarity between the diagnosis and treatment profile maps can be obtained, and the similarity is determined according to the comprehensive characteristics of the profile map. According to the similarity, multiple medical records with the highest similarity are selected, and the doctor is helped to make a medication decision after comprehensive reference by consulting their entity and attribute relationship. Thus, the comparison of similarity is more accurate, and a more accurate and comprehensive boosting significance is given to the medication decision made by the doctor.

[0004] To achieve the above object, the present invention proposes the following technical solutions:

[0005] In a first aspect, the present invention provides a method for assisting decision-making in accurate medication, comprising:

[0006] S1. Build a knowledge base;

[0007] The construction of the knowledge base specifically includes the following processes:

[0008] S11. Data collection;

[0009] S12, data preprocessing;

[0010] S13, extracting entity and attribute relationships from the preprocessed data;

[0011] The entities include disease categories, symptoms, drugs and their attributes, and patient attribute information, wherein the patient attribute information includes the patient's age, gender, height, and weight; the attribute relationships include the relationship between the patient attribute information and symptoms, the relationship between symptoms and disease categories, and the relationship between disease categories and corresponding drugs;

[0012] S14, constructing a medical visit profile for each patient based on the extracted entity and attribute relationships;

[0013] Specifically: The patient's visit profile is defined as a six-tuple , where V is a finite set of vertices, is the set of directed edges, is a finite set of vertex labels, is a finite set of directed edge labels, express: is the assignment function that assigns a label to a vertex, express is the assignment function for assigning labels to a directed edge;

[0014] The entities extracted in step S13 are represented as vertices in the patient profile graph, and the attribute relationships between the entities are represented as directed edges between them;

[0015] S2. Input the relevant information of the new patient;

[0016] The relevant information includes personal attribute information, disease symptoms, medical history and other information; the known information of the patient is input by the doctor;

[0017] S3, construct a medical profile map for new patients;

[0018] Based on the method of constructing the contour map in step S14, constructing a visit contour map for the new patient based on known information of the new patient;

[0019] S4, calculating the similarity between the new patient's visit profile and the patient's visit profile in the knowledge base;

[0020] Specifically, a relationship tree is defined. A relationship tree is an attributed, single-level, rooted and directed tree that can be represented by a five-tuple. To indicate that is the root vertex, L is the set of leaves, D is the set of directed edges from the root vertex to the leaves, is the assignment function for assigning labels to leaf vertices, is an assignment function for assigning labels to edges. There is a directed edge from the root vertex v to each vertex in the leaf L;

[0021] For a given patient visit profile graph f, for any vertex in f , generate a corresponding relationship tree = ,in is the vertex The corresponding set of leaves, From the vertex The set of directed edges to the corresponding leaf, Give the vertex The corresponding leaf vertex assignment function is is given from the vertex The distribution function for assigning labels to the directed edges corresponding to the leaves; therefore, a contour graph can be mapped to a multiset of relationship trees, which is called the tree representation of the contour graph f, denoted by T(f); the distance calculation expression between two relationship trees T1 and T2 is:

[0022] ; (1)

[0023] in, Represents two vertices and The distance between

[0024] ; (2)

[0025] ; (3)

[0026] ; (4)

[0027] ; (5)

[0028] in is a multiset of directed edge labels in set D, is the set of vertex labels in set L;

[0029] The multiset of the relationship tree of the contour images of two patients is the distance between the two contour images. For any two patient contour images f1 and f2, their relationship tree multisets T(f1) and T(f2) are obtained respectively. The distance between the contour images f1 and f2, that is, the similarity expression between the contour images f1 and f2 is:

[0030] ; (6)

[0031] Where P is a mapping relationship: T(f1) T(f2); Multiset representing a relationship tree Relationship tree in Multisets in a relational tree The mapping results in Representation; existence ,in It can be calculated according to formulas (1)-(5);

[0032] S5, sorting the similarities and selecting the medical records whose similarities are greater than a set threshold;

[0033] S6, assisting in making accurate medication decisions for new patients based on the medical records selected in step S5;

[0034] Since a large number of medical consultation instances and medical cases in textbooks and medical books are collected in the process of building the knowledge base, multiple records with the highest similarity to the new medical record can be automatically found in the similarity comparison process, so that doctors can have more similar cases to refer to when making medication decisions for current patients, rather than relying solely on personal medical experience, thereby helping to make more accurate medication decisions;

[0035] Furthermore, the present invention proposes to determine the distance between contour images based on the distance between multiple sets of relationship trees, thereby determining the similarity between patient medical records, making the search for similarity more efficient and better applicable to clinical medicine.

[0036] In a second aspect, the present invention provides a precise medication decision-making support system, comprising:

[0037] A knowledge base construction module is used to construct a knowledge base;

[0038] The knowledge base construction module specifically includes the following units:

[0039] Data collection unit, used for data collection;

[0040] A data preprocessing unit, used for data preprocessing;

[0041] Specifically including data cleaning and standardization;

[0042] Information extraction unit, extracting entity and attribute relationships from preprocessed data;

[0043] The entities include disease categories, symptoms, drugs and their attributes, and patient attribute information, wherein the patient attribute information includes the patient's age, gender, height, and weight; the attribute relationships include the relationship between the patient attribute information and symptoms, the relationship between symptoms and disease categories, and the relationship between disease categories and corresponding drugs;

[0044] A first profile map construction unit is used to construct a medical consultation profile map for each patient based on the extracted entity and attribute relationships;

[0045] Specifically: The patient's visit profile is defined as a six-tuple , where V is a finite set of vertices, is the set of directed edges, is a finite set of vertex labels, is a finite set of directed edge labels, express: is the assignment function that assigns a label to a vertex, express is the assignment function for assigning labels to a directed edge;

[0046] The entities extracted from the step information extraction unit are represented as vertices in the patient profile graph, and the attribute relationships between the entities are represented as directed edges between them;

[0047] An input module is used to input relevant information of new patients;

[0048] The relevant information includes personal attribute information, disease symptoms, medical history and other information; the known information of the patient is input by the doctor;

[0049] A second profile map building module is used to build a visit profile map for a new patient;

[0050] Based on the same method as the first profile construction unit constructing the profile, constructing a visit profile for the new patient based on the known information of the new patient;

[0051] A similarity calculation module is used to calculate the similarity between the medical profile of a new patient and the medical profile of a patient in the knowledge base;

[0052] Specifically, a relationship tree is defined. A relationship tree is an attributed, single-level, rooted and directed tree that can be represented by a five-tuple. To indicate that is the root vertex, L is the set of leaves, D is the set of directed edges from the root vertex to the leaves, is the assignment function for assigning labels to leaf vertices, is an assignment function for assigning labels to edges. There is a directed edge from the root vertex v to each vertex in the leaf L;

[0053] For a given patient visit profile graph f, for any vertex in f , generate a corresponding relationship tree = ,in is the vertex The corresponding set of leaves, From the vertex The set of directed edges to the corresponding leaf, Give the vertex The corresponding leaf vertex assignment function is is given from the vertex The distribution function for assigning labels to the directed edges corresponding to the leaves; therefore, a contour graph can be mapped to a multiset of relationship trees, which is called the tree representation of the contour graph f, denoted by T(f); the distance calculation expression between two relationship trees T1 and T2 is:

[0054] ; (1)

[0055] in, Represents two vertices and The distance between

[0056] ; (2)

[0057] ; (3)

[0058] ; (4)

[0059] ; (5)

[0060] in is a multiset of directed edge labels in set D, is the set of vertex labels in set L;

[0061] The multiset of the relationship tree of the contour images of two patients is the distance between the two contour images. For any two patient contour images f1 and f2, their relationship tree multisets T(f1) and T(f2) are obtained respectively. The distance between the contour images f1 and f2, that is, the similarity expression between the contour images f1 and f2 is:

[0062] ; (6)

[0063] Where P is a mapping relationship: T(f1) T(f2); Multiset representing a relationship tree Relationship tree in Multisets in a relational tree The mapping results in Representation; existence ,in It can be calculated according to formulas (1)-(5);

[0064] A screening module is used to sort the similarities and select the medical records whose similarities are greater than a set threshold;

[0065] Since there are multiple patient visit profiles in the knowledge base, multiple similarities between the multiple visit profiles and the new visit patient profile can be calculated, and the multiple similarities are sorted from high to low, and the visit records with similarities greater than the set threshold are selected. The greater the similarity, the greater the boosting significance that the visit record can provide, and the greater the auxiliary reference role it plays when doctors make accurate medication decisions;

[0066] The output module is used to assist in making accurate medication decisions for new patients based on the medical records selected by the screening module;

[0067] Since a large number of medical consultation instances and medical cases in textbooks and medical books are collected in the process of building the knowledge base, multiple records with the highest similarity to the new medical record can be automatically found in the similarity comparison process, so that doctors can have more similar cases to refer to when making medication decisions for current patients, rather than relying solely on personal medical experience, thereby helping to make more accurate medication decisions;

[0068] Furthermore, the present invention proposes to determine the distance between contour images based on the distance between multiple sets of relationship trees, thereby determining the similarity between patient medical records, making the search for similarity more efficient and better applicable to clinical medicine.

[0069] Compared with the prior art, the present invention has the following beneficial effects:

[0070] The invention proposes a precise medication decision-making support method and system. In the process of building a knowledge base, a profile map for each patient can be constructed based on the entity and attribute relationship of each patient in the acquired resources. For a new patient, its known entity and attribute relationship are input into the medication decision-making support system to generate a profile map of its current diagnosis. In the process of generating the profile map, not only the entity but also the attribute relationship between the entities is considered. In the process of calculating the distance between the profile map of the new patient and the profile maps of multiple patients in the knowledge base, a relationship tree (Relationsh Based on the concept of ip-Tree, each contour graph is converted into a multiple set of its corresponding relationship tree, and the distance between graphs is calculated based on the multiple sets of relationship trees, so as to obtain the similarity between the medical profile graphs. The doctor can refer to the top multiple medical records with the highest similarity to make a decision on the precise medication for this medical visit. Since the multiple sets of relationship trees can reflect more contour graph features, the similarity can be determined based on the comprehensive features of the contour graph. The method of calculating the distance between graphs based on the multiple sets of relationship trees of contour graphs and then obtaining the similarity is more accurate, which provides a more accurate and comprehensive boost to the medication decisions made by doctors. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 A flowchart of a method for assisting decision-making in precise medication according to an embodiment of the present invention;

[0072] Figure 2 is an example of the distance between the patient visit profile graphs f1 and f2 according to an embodiment of the present invention;

[0073] Figure 3It is a structural diagram of a precise medication decision support system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0074] Now, exemplary embodiments of the present invention are described with reference to the accompanying drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. These embodiments are provided to disclose the present invention in detail and completely and to fully convey the scope of the present invention to those skilled in the art. The terms used in the exemplary embodiments shown in the accompanying drawings are not intended to limit the present invention. In the accompanying drawings, the same units / elements are marked with the same reference numerals.

[0075] Unless otherwise specified, the terms (including technical terms) used herein have the commonly understood meanings to those skilled in the art. In addition, it is understood that the terms defined in commonly used dictionaries should be understood to have the same meanings as those in the context of the relevant fields, and should not be understood as idealized or overly formal meanings.

[0076] Embodiment 1:

[0077] The present invention proposes a method for assisting decision-making in accurate medication, such as Figure 1 As shown, including:

[0078] S1. Build a knowledge base;

[0079] In the process of precision medication decision-making, collecting comprehensive and accurate data to build a semantically rich knowledge base is one of the most critical steps. In view of the differences and formats of data obtained from various resources, data preprocessing is required after data acquisition, including data cleaning and standardization. During the data cleaning process, the consistency of the data is checked, duplicate records and low-quality data are removed, and then the data is converted into a unified standard format through standardization. Finally, entity and attribute relationships are extracted based on the standardized data, where entities include disease categories, symptoms, drugs and their attributes, and patient attribute information, where the patient attribute information includes the patient's age, gender, height and weight; attribute relationships include the relationship between patient attribute information and symptoms, the relationship between symptoms and disease categories, and the relationship between disease categories and corresponding drugs, which can reflect the process of diagnosis and treatment; the collection and processing of the above data are limited to protecting the privacy of patients.

[0080] Therefore, it can be determined that the construction of the knowledge base specifically includes the following processes:

[0081] S11. Data collection;

[0082] S12, data preprocessing;

[0083] Specifically including data cleaning and standardization;

[0084] S13, extracting entity and attribute relationships from the preprocessed data;

[0085] The entities include disease categories, symptoms, drugs and their attributes, and patient attribute information, wherein the patient attribute information includes the patient's age, gender, height, and weight; the attribute relationships include the relationship between the patient attribute information and symptoms, the relationship between symptoms and disease categories, and the relationship between disease categories and corresponding drugs;

[0086] S14, constructing a medical visit profile for each patient based on the extracted entity and attribute relationships;

[0087] Specifically: The patient's visit profile is defined as a six-tuple , where V is a finite set of vertices, is the set of directed edges, is a finite set of vertex labels, is a finite set of directed edge labels, express: is the assignment function that assigns a label to a vertex, express is the assignment function for assigning labels to a directed edge;

[0088] The entities extracted in step S13 are represented as vertices in the patient profile graph, and the attribute relationships between the entities are represented as directed edges between them;

[0089] S2. Input the relevant information of the new patient;

[0090] The relevant information includes personal attribute information, disease symptoms, past medical history and other information; the known information of the patient is input by the doctor through the input interface;

[0091] S3, construct a medical profile map for new patients;

[0092] Based on the method of constructing the contour map in step S14, constructing a visit contour map for the new patient based on known information of the new patient;

[0093] S4, calculating the similarity between the new patient's visit profile and the patient's visit profile in the knowledge base;

[0094] Specifically, a relationship tree is defined. A relationship tree is an attributed, single-level, rooted and directed tree that can be represented by a five-tuple. To indicate that is the root vertex, L is the set of leaves, D is the set of directed edges from the root vertex to the leaves, is the assignment function for assigning labels to leaf vertices, is an assignment function for assigning labels to edges. There is a directed edge from the root vertex v to each vertex in the leaf L;

[0095] For a given patient visit profile graph f, for any vertex in f , generate a corresponding relationship tree = ,in is the vertex The corresponding set of leaves, From the vertex The set of directed edges to the corresponding leaf, Give the vertex The corresponding leaf vertex assignment function is is given from the vertex The distribution function for assigning labels to the directed edges corresponding to the leaves; therefore, a contour graph can be mapped to a multiset of relationship trees, which is called the tree representation of the contour graph f, denoted by T(f); the distance calculation expression between two relationship trees T1 and T2 is:

[0096] ; (1)

[0097] in, Represents two vertices and The distance between

[0098] ; (2)

[0099] ; (3)

[0100] ; (4)

[0101] ; (5)

[0102] in is a multiset of directed edge labels in set D, is the set of vertex labels in set L;

[0103] The multiset of the relationship tree of the contour images of two patients is the distance between the two contour images. For any two patient contour images f1 and f2, their relationship tree multisets T(f1) and T(f2) are obtained respectively. The distance between the contour images f1 and f2, that is, the similarity expression between the contour images f1 and f2 is:

[0104] ; (6)

[0105] Where P is a mapping relationship: T(f1) T(f2); Multiset representing a relationship tree Relationship tree in Multisets in a relational tree The mapping results in Representation; existence ,in It can be calculated according to formulas (1)-(5);

[0106] Preferably, the optimal solution can be obtained by using an algorithm known in the prior art, such as the Hungarian algorithm; wherein the distance between the contour images f1 and f2 is That is, it reflects the similarity between the two. By traversing, the distance between the contour map of the new patient and the contour maps of all patients in the knowledge base is calculated, so that the similarity between the contour map of the new patient and other medical records in the knowledge base can be obtained.

[0107] like Figure 2 As shown, it provides an embodiment of calculating the distance between two patient visit profile images f1 and f2, wherein Figure 2 a is the two patient visit profiles f1 and f2, Figure 2 b represents the relationship tree T(f1) The mapping relationship P between T(f2), where the solid line represents the best match between the relationship trees. By calculating, we can know the distance between the two patient visit profiles f1 and f2. =0+1=1;

[0108] S5, sorting the similarities and selecting the medical records whose similarities are greater than a set threshold;

[0109] Since there are multiple patient visit profiles in the knowledge base, multiple similarities between the multiple visit profiles and the new visit patient profile can be calculated, and the multiple similarities are sorted from high to low, and the visit records with similarities greater than the set threshold are selected. The greater the similarity, the greater the boosting significance that the visit record can provide, and the greater the auxiliary reference role it plays when doctors make accurate medication decisions;

[0110] S6, assisting in making accurate medication decisions for new patients based on the medical records selected in step S5;

[0111] Since a large number of medical consultation instances and medical cases in textbooks and medical books are collected in the process of building the knowledge base, multiple records with the highest similarity to the new medical record can be automatically found in the similarity comparison process, so that doctors can have more similar cases to refer to when making medication decisions for current patients, rather than relying solely on personal medical experience, thereby helping to make more accurate medication decisions;

[0112] Furthermore, the present invention proposes to determine the distance between contour images based on the distance between multiple sets of relationship trees, thereby determining the similarity between patient medical records, making the search for similarity more efficient and better applicable to clinical medicine.

[0113] Embodiment 2:

[0114] The present invention proposes a precise medication decision-making support system 300, such as Figure 3 As shown, including:

[0115] A knowledge base construction module is used to construct a knowledge base;

[0116] In the process of precision medication decision-making, collecting comprehensive and accurate data to build a semantically rich knowledge base is one of the most critical steps. In view of the differences and formats of data obtained from various resources, data preprocessing is required after data acquisition, including data cleaning and standardization. During the data cleaning process, the consistency of the data is checked, duplicate records and low-quality data are removed, and then the data is converted into a unified standard format through standardization. Finally, entity and attribute relationships are extracted based on the standardized data, where entities include disease categories, symptoms, drugs and their attributes, and patient attribute information, where the patient attribute information includes the patient's age, gender, height and weight; attribute relationships include the relationship between patient attribute information and symptoms, the relationship between symptoms and disease categories, and the relationship between disease categories and corresponding drugs, which can reflect the process of diagnosis and treatment; the collection and processing of the above data are limited to protecting the privacy of patients.

[0117] Therefore, it can be determined that the knowledge base construction module specifically includes the following units:

[0118] Data collection unit, used for data collection;

[0119] A data preprocessing unit, used for data preprocessing;

[0120] Specifically including data cleaning and standardization;

[0121] Information extraction unit, extracting entity and attribute relationships from preprocessed data;

[0122] The entities include disease categories, symptoms, drugs and their attributes, and patient attribute information, wherein the patient attribute information includes the patient's age, gender, height, and weight; the attribute relationships include the relationship between the patient attribute information and symptoms, the relationship between symptoms and disease categories, and the relationship between disease categories and corresponding drugs;

[0123] A first profile map construction unit is used to construct a medical consultation profile map for each patient based on the extracted entity and attribute relationships;

[0124] Specifically: The patient's visit profile is defined as a six-tuple , where V is a finite set of vertices, is the set of directed edges, is a finite set of vertex labels, is a finite set of directed edge labels, express: is the assignment function that assigns a label to a vertex, express is the assignment function for assigning labels to a directed edge;

[0125] The entities extracted from the step information extraction unit are represented as vertices in the patient profile graph, and the attribute relationships between the entities are represented as directed edges between them;

[0126] An input module is used to input relevant information of new patients;

[0127] The relevant information includes personal attribute information, disease symptoms, medical history and other information; the known information of the patient is input by the doctor through the input interface;

[0128] A second profile map building module is used to build a visit profile map for a new patient;

[0129] Based on the same method as the first profile construction unit constructing the profile, constructing a visit profile for the new patient based on the known information of the new patient;

[0130] A similarity calculation module is used to calculate the similarity between the medical profile of a new patient and the medical profile of a patient in the knowledge base;

[0131] Specifically, a relationship tree is defined. A relationship tree is an attributed, single-level, rooted and directed tree that can be represented by a five-tuple. To indicate that is the root vertex, L is the set of leaves, D is the set of directed edges from the root vertex to the leaves, is the assignment function for assigning labels to leaf vertices, is an assignment function for assigning labels to edges. There is a directed edge from the root vertex v to each vertex in the leaf L;

[0132] For a given patient visit profile graph f, for any vertex in f , generate a corresponding relationship tree = ,in is the vertex The corresponding set of leaves, From the vertex The set of directed edges to the corresponding leaf, Give the vertex The corresponding leaf vertex assignment function is is given from the vertex The distribution function for assigning labels to the directed edges corresponding to the leaves; therefore, a contour graph can be mapped to a multiset of relationship trees, which is called the tree representation of the contour graph f, denoted by T(f); the distance calculation expression between two relationship trees T1 and T2 is:

[0133] ; (1)

[0134] in, Represents two vertices and The distance between

[0135] ; (2)

[0136] ; (3)

[0137] ; (4)

[0138] ; (5)

[0139] in is a multiset of directed edge labels in set D, is the set of vertex labels in set L;

[0140] The multiset of the relationship tree of the contour images of two patients is the distance between the two contour images. For any two patient contour images f1 and f2, their relationship tree multisets T(f1) and T(f2) are obtained respectively. The distance between the contour images f1 and f2, that is, the similarity expression between the contour images f1 and f2 is:

[0141] ; (6)

[0142] Where P is a mapping relationship: T(f1) T(f2); Multiset representing a relationship tree Relationship tree in Multisets in a relational tree The mapping results in Representation; existence ,in It can be calculated according to formulas (1)-(5);

[0143] Preferably, the optimal solution can be obtained by using an algorithm known in the prior art, such as the Hungarian algorithm; wherein the distance between the contour images f1 and f2 is That is, it reflects the similarity between the two. By traversing, the distance between the contour map of the new patient and the contour maps of all patients in the knowledge base is calculated, so that the similarity between the contour map of the new patient and other medical records in the knowledge base can be obtained.

[0144] like Figure 2 As shown, it provides an embodiment of calculating the distance between two patient visit profile images f1 and f2, wherein Figure 2 a is the two patient visit profiles f1 and f2, Figure 2 b represents the relationship tree T(f1) The mapping relationship P between T(f2), where the solid line represents the best match between the relationship trees. By calculating, we can know the distance between the two patient visit profiles f1 and f2. =0+1=1;

[0145] A screening module is used to sort the similarities and select the medical records whose similarities are greater than a set threshold;

[0146] Since there are multiple patient visit profiles in the knowledge base, multiple similarities between the multiple visit profiles and the new visit patient profile can be calculated, and the multiple similarities are sorted from high to low, and the visit records with similarities greater than the set threshold are selected. The greater the similarity, the greater the boosting significance that the visit record can provide, and the greater the auxiliary reference role it plays when doctors make accurate medication decisions;

[0147] The output module is used to assist in making accurate medication decisions for new patients based on the medical records selected by the screening module;

[0148] Since a large number of medical consultation instances and medical cases in textbooks and medical books are collected in the process of building the knowledge base, multiple records with the highest similarity to the new medical record can be automatically found in the similarity comparison process, so that doctors can have more similar cases to refer to when making medication decisions for current patients, rather than relying solely on personal medical experience, thereby helping to make more accurate medication decisions;

[0149] Furthermore, the present invention proposes to determine the distance between contour images based on the distance between multiple sets of relationship trees, thereby determining the similarity between patient medical records, making the search for similarity more efficient and better applicable to clinical medicine.

[0150] It should be understood by those skilled in the art that the embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The solutions in the embodiments of the present invention may be implemented in various computer languages, for example, object-oriented programming language Java and interpreted scripting language JavaScript, etc.

[0151] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0152] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0153] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0154] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0155] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A precision medication decision-making assistance method, characterized in that: The steps include: S1. Build a knowledge base; The construction of the knowledge base specifically includes the following processes: S11. Data collection; S12, data preprocessing; S13, extracting entity and attribute relationships from the preprocessed data; The entities include disease categories, symptoms, drugs and their attributes, and patient attribute information, wherein the patient attribute information includes the patient's age, gender, height, and weight; the attribute relationships include the relationship between patient attribute information and symptoms, the relationship between symptoms and disease categories, and the relationship between disease categories and corresponding drugs; S14, constructing a medical consultation profile for each patient based on the extracted entities and attribute relationships; The patient visit profile is constructed as follows: The patient visit profile is defined as a six-tuple , where V is a finite set of vertices, is the set of directed edges, is a finite set of vertex labels, is a finite set of directed edge labels, express: is the assignment function that assigns a label to a vertex, express is the assignment function for assigning labels to a directed edge; The entities extracted in step S13 are represented as vertices in the patient profile graph, and the attribute relationships between the entities are represented as directed edges between them; S2. Input the relevant information of the new patient; S3, constructing a medical profile map for the new patient based on the profile map construction method in step S14; S4, calculating the similarity between the new patient's visit profile and the patient's visit profile in the knowledge base; For the patient visit profile f, generate its corresponding relationship tree = ,in is the vertex, is the vertex The corresponding set of leaves, From the vertex The set of directed edges to the corresponding leaf, Give the vertex The corresponding leaf vertex assignment function is is given from the vertex The assignment function for assigning labels to directed edges to their corresponding leaves, from vertices To the leaf Each vertex in has a directed edge, thus obtaining a multi-set of the relationship tree corresponding to the patient's medical profile f, represented by T(f); the distance calculation expression between the relationship trees T1 and T2 corresponding to the patient's medical profile f1 and f2 is: ; (1) in, Represents two vertices and The distance between ;(2) ;(3) ;(4) ;(5) in is a multiset of directed edge labels in set D, is the set of vertex labels in set L; The distance between the multiple sets of the relationship tree of the two patients' profile graphs is the inter-graph distance between the two profile graphs, which is also the similarity between the profile graphs. Therefore, the similarity expression between the patient's visit profile graphs f1 and f2 is: ;(6) Where P is a mapping relation: T(f1) T(f2); Multiset representing a relationship tree Relationship tree in Multisets in a relational tree The mapping results in Characterization; Based on formulas (1)-(5), solve formula (6) ; Traversing and calculating the similarity between the new patient profile and all the patient profiles in the knowledge base; S5, sorting the similarities and selecting the medical records whose similarities are greater than a set threshold; S6. Assist in making accurate medication decisions for new patients based on the medical records selected in step S5.

2. The method for assisting decision-making in precise medication according to claim 1, characterized in that: The data preprocessing includes data cleaning and standardization.

3. The method for assisting decision-making in precise medication according to claim 1 or 2, characterized in that: The relevant information of the new patient includes personal attribute information, disease symptoms, and past medical history.

4. The method for assisting decision-making in precise medication according to claim 3, characterized in that: The relevant information of the new patient is input by the attending physician through the input interface.

5. A precision medication decision-making support system, running the precision medication decision-making support method according to any one of claims 1 to 4, characterized in that: The system comprises: A knowledge base construction module is used to construct a knowledge base; The knowledge base construction module specifically includes the following units: Data collection unit, used for data collection; A data preprocessing unit, used for data preprocessing; Information extraction unit, extracting entity and attribute relationships from preprocessed data; The entities include disease categories, symptoms, drugs and their attributes, and patient attribute information, wherein the patient attribute information includes the patient's age, gender, height, and weight; the attribute relationships include the relationship between the patient attribute information and symptoms, the relationship between symptoms and disease categories, and the relationship between disease categories and corresponding drugs; A first profile construction unit, configured to construct a medical consultation profile for each patient based on the extracted entities and attribute relationships; Specifically: The patient's visit profile is defined as a six-tuple , where V is a finite set of vertices, is the set of directed edges, is a finite set of vertex labels, is a finite set of directed edge labels, express: is the assignment function that assigns a label to a vertex, express is the assignment function for assigning labels to a directed edge; The entities extracted from the step information extraction unit are represented as vertices in the patient profile graph, and the attribute relationships between the entities are represented as directed edges between them; An input module is used to input relevant information of new patients; A second profile map construction module, used for constructing a visit profile map for a new visit patient based on the same way as the first profile map construction unit constructs the profile map; A similarity calculation module is used to calculate the similarity between the medical profile of a new patient and the medical profile of a patient in the knowledge base; For the patient visit profile f, generate its corresponding relationship tree = ,in is the vertex, is the vertex The corresponding set of leaves, From the vertex The set of directed edges to the corresponding leaf, Give the vertex The corresponding leaf vertex assignment function is is given from the vertex The assignment function for assigning labels to directed edges to their corresponding leaves, from vertices To the leaf Each vertex in has a directed edge, thus obtaining a multi-set of the relationship tree corresponding to the patient's medical profile f, represented by T(f); the distance calculation expression between the relationship trees T1 and T2 corresponding to the patient's medical profile f1 and f2 is: ;(1) in, Represents two vertices and The distance between ;(2) ;(3) ;(4) ;(5) in is a multiset of directed edge labels in set D, is the set of vertex labels in set L; The distance between the multiple sets of the relationship tree of the two patients' profile graphs is the inter-graph distance between the two profile graphs, which is also the similarity between the profile graphs. Therefore, the similarity expression between the patient's visit profile graphs f1 and f2 is: ;(6) Where P is a mapping relation: T(f1) T(f2); Multiset representing a relationship tree Relationship tree in Multisets in a relational tree The mapping results in Characterization; Based on formulas (1)-(5), solve formula (6) ; Traversing and calculating the similarity between the new patient profile and all the patient profiles in the knowledge base; A screening module is used to sort the similarities and select the medical records whose similarities are greater than a set threshold; The output module is used to assist in making accurate medication decisions for new patients based on the medical records selected by the screening module.

6. The precise medication decision-making support system according to claim 5, characterized in that: The data preprocessing includes data cleaning and standardization.

7. The precise medication decision-making support system according to claim 5 or 6, characterized in that: The relevant information of the new patient includes personal attribute information, disease symptoms, and past medical history.

8. The precise medication decision-making support system according to claim 7, characterized in that: The relevant information of the new patient is input by the attending physician through the input interface.

9. A computer-readable storage medium having stored thereon one or more computer-readable computer instructions, characterized in that: The one or more computer instructions are executed by a processor to implement the steps of the method according to any one of claims 1 to 4.

10. A computer program product, comprising a computer program or instructions, characterized in that: When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Citation Information

Patent Citations

  • Medication decision support method and device based on graphic state machine

    CN112802575A

  • Personalized drug recommendation method and system based on knowledge graph

    CN117672450A

  • Personalized precise medication recommendation method and apparatus

    WO2021073277A1