CN111599427B - Recommendation method and device for unified diagnosis, electronic equipment and storage medium - Google Patents
Recommendation method and device for unified diagnosis, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN111599427B CN111599427B CN202010409050.0A CN202010409050A CN111599427B CN 111599427 B CN111599427 B CN 111599427B CN 202010409050 A CN202010409050 A CN 202010409050A CN 111599427 B CN111599427 B CN 111599427B
- Authority
- CN
- China
- Prior art keywords
- diagnosis
- disease
- cluster
- similarity
- unified
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H15/00—ICT specially adapted for medical reports, e.g. generation or transmission thereof
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Theoretical Computer Science (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention relates to the technical field of disease auxiliary diagnosis in health management, in particular to a recommendation method and device for unified diagnosis, electronic equipment and a storage medium. Determining a diagnostic code and cluster C based on the diagnostic code of the subject k Similarity measure results of the primary/secondary canonical disease diagnosis coding set in (1); the cluster C k Is K clusters divided according to the similarity between the diagnostic information of N subjects; selecting a cluster to which the similarity value greater than a preset threshold belongs, and acquiring a disease co-occurrence mode corresponding to each cluster from a unified diagnosis library; and carrying out visual analysis on the ICD body structure by combining the acquired disease co-occurrence mode with the diagnosis codes, identifying the nearest father node according to the semantic relation among the diagnosis codes under the same branch, obtaining the disease type, and reordering to obtain the unified diagnosis of the examinee, thereby solving the technical problem that the unified diagnosis results given by different main inspectors have larger difference for the same examinee.
Description
Technical Field
The invention relates to the technical field of disease auxiliary diagnosis in health management, in particular to a recommendation method and device for unified diagnosis, electronic equipment and a storage medium.
Background
The health examination report is a medical document given to the examinee by a health management (examination) organization, and comprises a health examination report first page, a main examination report, a physical examination record, a laboratory and medical image examination report and the like. The main examination report is used as a core component of the examination report, no specific implementation details can be provided for reference at present, and different examination organizations all over the country have different writing modes and different quality. The unified diagnosis is the main content of the main examination report, and the main examiners should reasonably classify the disease system according to the basic principles of clinical diagnosis thinking, especially the theory of monarch, and use one disease to summarize or explain various clinical manifestations of the disease as much as possible, so that the reasonable classification can make the main examination report clear. Currently, the main examination reports of each physical examination organization are mostly finished by related software, and the diagnosis results of diseases are mostly listed as various positive examination results.
In actual work, a primary examining physician usually needs to strictly check basic information of an examined person, various examination results, summarize primary and secondary diagnosis and positive findings, and give corresponding diagnosis results by combining self medical knowledge.
In practice, the inventors found that the above prior art has the following disadvantages:
due to the differences of the professional levels of the main examination doctors, the different examination items of the examined persons, the main/secondary diagnosis and the variation of positive signs, the existing unified diagnosis method only depending on the professional knowledge of the main examination doctors is time-consuming and labor-consuming, low in efficiency and long in time-consuming, and meanwhile, for the physical examination conclusion of the same examined person, the unified diagnosis results given by different main examination doctors are greatly different.
Disclosure of Invention
In order to solve the above technical problems, an object of the present invention is to provide a recommendation method, apparatus, device and storage medium for unified diagnosis, wherein the adopted technical solutions are as follows:
in a first aspect, an embodiment of the present invention provides a recommendation method for unified diagnosis, where the method includes the following steps:
determining a diagnostic code and cluster C based on the diagnostic code of the subject k Similarity measure results of the primary/secondary canonical disease diagnosis coding set in (1); the cluster C k Is K clusters divided according to the similarity between the diagnostic information of N subjects;
selecting a cluster to which the similarity value greater than a preset threshold belongs, and acquiring a disease co-occurrence mode corresponding to each cluster from a unified diagnosis library;
and carrying out visual analysis on the ICD body structure by combining the acquired disease co-occurrence mode with the diagnosis codes, identifying the nearest father node according to the semantic relation among the diagnosis codes under the same branch, acquiring the disease type and reordering to acquire the unified diagnosis of the examinee.
In a second aspect, an embodiment of the present invention provides a recommendation apparatus for unified diagnosis, where the apparatus includes:
a similarity metric module for determining a diagnostic code of the subject from the diagnostic code and cluster C k Similarity measure results of the primary/secondary canonical disease diagnosis coding set in (1); the cluster C k Is K clusters divided according to the similarity between the diagnostic information of N subjects;
the matching detection module is used for selecting the cluster to which the similarity value greater than the preset threshold value belongs and acquiring a disease co-occurrence mode corresponding to each cluster from a unified diagnosis library; and
and the recommending module is used for performing visual analysis on the ICD body structure by combining the acquired disease co-occurrence mode with the diagnosis codes, identifying a nearest father node according to the semantic relation between the diagnosis codes under the same branch, acquiring the disease type and reordering to obtain the unified diagnosis of the examinee.
In a third aspect, an embodiment of the present invention provides an electronic device, including:
a processing unit;
a memory unit for storing processing unit executable instructions;
wherein the processing unit is configured to: performing the method of any one of the above.
In a fourth aspect, an embodiment of the present invention provides a storage medium, in which computer-readable program instructions are stored, and when the program instructions are executed by a processing unit, the method described in any one of the above is implemented.
The invention has the following beneficial effects:
according to the embodiment of the invention, the similarity measurement result of the diagnosis code of the examinee and the primary/secondary typical disease diagnosis code set is used for acquiring the disease co-occurrence mode corresponding to each cluster from the unified diagnosis library, and the acquired disease co-occurrence mode is combined with the diagnosis code to perform visual analysis on the ICD body structure, so that the disease type is obtained and the unified diagnosis of the examinee is obtained by reordering. Aiming at the diagnosis of various diseases and positive signs of the examined person, the method provides auxiliary reference for the main examining physician to quickly and automatically determine the unified diagnosis of the main diseases, and can enable the examined person to know the main problems and the secondary problems existing in the current self health according to the unified diagnosis and the sequencing result thereof. The technical problem that unified diagnosis results given by different main inspectors have large difference for the same examinee is solved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flow chart of a recommendation method for unified diagnosis according to an embodiment of the present invention;
FIG. 2 is a flow chart of a method for building a unified diagnostic library according to an embodiment of the present invention;
FIG. 3 is a flowchart of a method for obtaining a primary/secondary canonical disease diagnosis code set according to an embodiment of the present invention;
FIG. 4 is a flowchart illustrating an embodiment of a method for building a unified diagnosis library and a method for recommending unified diagnosis;
FIG. 5 is a partial structure of the ICD-10 disease classification architecture provided by one embodiment of the present invention;
FIG. 6 is a schematic diagram of the structure of an exemplary disease diagnosis coding set for 3 clusters according to another embodiment of the present invention;
FIG. 7 is a schematic structural diagram of a typical disease co-occurrence pattern for cluster 3 provided in accordance with another embodiment of the present invention;
FIG. 8 is a schematic structural diagram of a unified diagnosis for the patient according to another embodiment of the present invention;
FIG. 9 is a block diagram of a recommendation apparatus for unified diagnosis according to an embodiment of the present invention;
fig. 10 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
To further illustrate the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description of a method, an apparatus, an electronic device and a storage medium for unified diagnosis according to the present invention with reference to the accompanying drawings and preferred embodiments thereof will be made in detail. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
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.
The following describes in detail a specific scheme of a recommendation method, a recommendation device, an electronic device, and a storage medium for unified diagnosis according to embodiments of the present invention with reference to the accompanying drawings.
Referring to fig. 1, which shows a flowchart of a recommendation method for a unified diagnosis provided by an embodiment of the present invention, in order to solve the technical problem that unified diagnosis results given by different main physicians have great differences, the embodiment of the present invention recommends a reasonable unified diagnosis result through similarity measurement of disease diagnosis codes, visual analysis and evaluation by the main physicians, and the recommendation method for a unified diagnosis includes the following steps:
step S001, according to the diagnosis code of the subject, determining the diagnosis code and the cluster C k Similarity measure results of the primary/secondary canonical disease diagnosis coding set in (1); the cluster C k Are K clusters divided according to the similarity between the diagnostic information of N subjects.
And S002, selecting the cluster to which the similarity value greater than the preset threshold value belongs, and acquiring the disease co-occurrence mode corresponding to each cluster from a unified diagnosis library.
And S003, combining the acquired disease co-occurrence mode with the diagnosis codes to perform visual analysis on the ICD body structure, identifying a nearest father node according to the semantic relation among the diagnosis codes under the same branch, obtaining the disease type and reordering to obtain the disease co-occurrence mode of the examinee.
Specifically, in this embodiment, after a new subject performs a health examination, a main examining physician determines a series of main/secondary diagnoses and diagnosis codes of positive findings according to the current examination result of the subject, and first calculates the similarity between the subject and the main/secondary typical disease diagnosis code set of each cluster according to a similarity measurement method to obtain a set of similarity measurement results. And sorting the similarity measurement results, identifying the highest similarity value or the similarity value ranked in the top five, and giving the result of the cluster to the examinee, namely recommending the typical disease co-occurrence mode in the unified diagnosis library to a main examination physician. And carrying out visual analysis on the disease diagnosis code of the examinee and the recommended typical disease co-occurrence mode in the ICD body structure, automatically identifying the nearest father node according to the conceptual semantic relationship between the disease diagnosis codes under the same branch, defining a new disease type and the sequence thereof, and determining the unified diagnosis of the examinee. The main examining physician carries out operations such as correction, classification, evaluation and definition on the obtained disease co-occurrence mode according to the self-domain knowledge, and further obtains a unified diagnosis result aiming at the physical examination of the examined person.
In summary, according to the embodiment of the present invention, the similarity measurement result between the diagnosis code of the subject and the primary/secondary typical disease diagnosis code set is used to obtain the disease co-occurrence pattern corresponding to each cluster from a unified diagnosis library, and perform visual analysis on the obtained disease co-occurrence pattern in combination with the diagnosis code in the ICD body structure, so as to obtain the disease type and reorder the disease co-occurrence pattern of the subject. Aiming at the diagnosis of various diseases and positive signs of the examined person, the method provides auxiliary reference for the main examining physician to quickly and automatically determine the unified diagnosis of the main diseases, and can enable the examined person to know the main problems and the secondary problems existing in the current self health according to the unified diagnosis and the sequencing result thereof. The technical problem that unified diagnosis results given by different main inspectors have large difference for the same examinee is solved.
Referring to fig. 2 to 4, as a preferred embodiment of the present invention, taking the ICD-10 disease classification system as an example, the method for constructing the unified diagnosis library in step S002 includes the following steps:
step S201, the primary examination report of the examinee is mapped to the ICD disease system and is represented by a disease diagnosis code.
Specifically, the current primary/secondary diagnosis and positive signs of the subject are expressed in the form of a disease diagnosis code (e.g., ICD-10) in the normative primary examination report. Typically, a subject is labeled as a collection of disease codes with a certain ordering, with the leading diagnostic code representing the subject's primary disease type or positive characteristics. The diagnostic information of the subject is defined as:
D i ＝{(d 1 ,Seq(d 1 )),(d 2 ,Seq(d 2 )),L}
wherein d is 1 Representing the disease code of the subject, seq (d) 1 ) Representing the order of disease codes of the subject.
Step S202, according to the disease diagnosis code, on the basis of the ICD disease classification system, a disease code body structure is constructed.
In order to measure the similarity of the diagnostic information of the examinees, a disease code body structure (tree structure) needs to be constructed on the basis of the ICD disease classification system. Taking ICD-10 as an example, please refer to FIG. 5, which shows the local structure of the ICD-10 disease classification system, comprising 22 chapters (level-1 at the first level), 262 chapters (level-2 at the second level), 2051 categories with 3 bits (level-3 at the third level), 9505 sub-categories with 4 bits (level-4 at the fourth level), and 22908 extended codes with 6 bits (level-5 at the fifth level).
Step S203, similarity measurement of semantic relation of the disease diagnosis code in the disease code ontology structure.
The ICD-10 disease classification system is a hierarchical tree with classification system, the diagnosis codes under the same branch have certain similarity, the embodiment of the invention adopts a similarity measurement method based on semantic relation, and the similarity measurement method comprises information quantity measurement of the disease diagnosis codes, similarity measurement between the disease diagnosis codes and similarity measurement between disease diagnosis code sets.
The method for measuring the coding information of the disease diagnosis is specific, in an ICD-10 disease classification system, each code represents a concept, semantic similarity exists between classification concepts, and the concepts under the same branch are more similar to the concepts of different branches in semantic similarity. The embodiment of the invention adopts a level depth measurement method, namely, each layer is endowed with a specific numerical value, and the numerical value is larger when the concept level is deeper. Coding d for an ICD-10 1 The amount of information is defined as:
IC(d 1 )＝path(d 1 →r)
wherein r represents the root node of the ICD-10 disease classification system, and path (·) is defined as the code d encoded from ICD-10 1 Go toThe path length of the root node r. Therefore, the information amount of the root node is 0, the information amount of the first layer chapter is 1, the information amount of the second layer section is 2, the information amount of the third layer category is 3, the information amount of the fourth layer sub-category is 4, and the information amount of the fifth layer extension code is 5.
The similarity measurement method for the disease diagnosis codes is specifically used for measuring the similarity between two codes according to the information measurement of the disease diagnosis codes. The embodiment of the invention adopts a method of a nearest father node of two semantic concepts to calculate the similarity, and the similarity between two codes is defined as follows:
wherein d is 1 And d 2 Coding for two disease diagnostics in the ICD-10 disease Classification System, LCA (d) 1 ,d 2 ) Defined as the disease diagnosis code d 1 And disease diagnosis code d 2 If d is the nearest parent node of 1 ＝d 2 Then LCA (d) 1 ,d 2 )＝d 1 ＝d 2 ，IC(LCA(d 1 ,d 2 ))＝IC(d 1 )＝IC(d 2 ) (ii) a If d is 1 ≠d 2 And LCA (d) 1 ,d 2 ) = root node, IC (LCA (d) 1 ,d 2 ))＝0。
For the similarity measurement method between disease diagnosis code sets, specifically, in a health examination report, a subject is usually diagnosed as a set of disease diagnosis codes, and the similarity of disease diagnosis information of the subject is represented by measuring the similarity of two disease diagnosis code sets. The embodiment of the invention measures the similarity between the code sets by considering the most similar code pair-wise mean value, and assumes that the diagnostic information of a subject i and a subject j is respectively defined as Di i ′＝{d i1 ,d i2 ,L,d ig L } and Di' j ＝{d j1 ,d j2 ,L,d jh L (here, the order of the disease diagnosis codes is not considered), the similarity of the diagnosis information of the subject i and the subject j is defined as:
wherein, | Di' i L is the number of disease diagnosis codes in the disease diagnosis information of the detected person i, | Di' j I is the number of disease diagnosis codes in the disease diagnosis information of the examined person j, d ig Coding for the g-th disease diagnosis of subject i, d jh H disease diagnostic code for subject j.
And (3) carrying out pairwise similarity measurement on the diagnosis information of all physical examination crowd to obtain a similarity matrix S of the physical examination crowd.
Step S204, clustering the subjects according to the similarity measurement result to obtain a cluster C K 。
On the basis of the similarity matrix of the diagnosis information of the physical examination crowd, similar examinees are clustered into a class by using a clustering algorithm, so that the examinees in a cluster are similar, and the examinees outside the cluster are dissimilar.
The clustering algorithm is an effective unsupervised learning algorithm and mainly comprises a partitioning method, a hierarchical method, a density-based method and the like, and the most common methods are K-means, hierarchical clustering, density peak-based clustering and the like. The embodiment of the invention takes AP clustering (affine propagation clustering) as an example, and divides physical examination groups into different clusters.
In AP clustering, the selection of the number of clusters is controlled by a preference parameter p value, and the embodiment of the invention adopts a preference coefficient p c To control the number of clusters:
p＝mean(S)-p c gN
wherein S is a similarity matrix of diagnostic information of all physical examination people, and N is the number of examined people.
Dividing the N subjects into K clusters (C) according to the disease diagnosis information of the subjects 1 ,C 2 ,…,C K ) The number of subjects per cluster is defined as:
wherein, C k (Di′ j ) The representation is divided into clusters C K J, E (C) of the subjects in (1) K ) Represents a cluster C k Is representative of the population of (a). λ (g) is an indicative function, when subject j is divided into clusters C k In, C k (Di′ j )＝E(C k )，λ(C k (Di′ j ),E(C k ) ) =1, otherwise λ (C) k (Di′ j ),E(C k ) ) =0. In other words, when the subject j is divided into clusters C k Class tag E (C) assigned to the cluster K ) That is, the representative population of the cluster indicates that the subject is divided into the cluster.
In step S205, a primary/secondary typical disease diagnosis code set is obtained.
Referring to fig. 3, the method for obtaining the primary/secondary canonical disease diagnosis code set includes the following steps:
step S301, in the cluster C k The people with similarity greater than the threshold of similarity with the subject are selected as the core physical examination people.
Because the ICD-10 disease classification system has complex semantic relation and obvious difference exists between codes in different layers, the ICD-10 disease diagnosis codes simply represented by people cannot completely describe the common characteristics of the clusters. The embodiment of the invention adopts a method of defining a core area from each cluster, selecting a subject with large representative similarity with the cluster population, analyzing the diagnosis information of the subject and extracting high-frequency disease diagnosis codes to represent the common characteristics of the cluster.
Definition of a Cluster C k The core area of (a), the selection of the core physical examination population:
Core k ＝{Di′ j |S(Di′ j ,E(C k ))≥τ k }
wherein, E (C) k ) Represents a cluster C k Is a representation of the population of (1), tau k Is a similarity threshold set in advance.
In step S302, the diagnostic codes of the subjects whose occurrence frequency is greater than the frequency threshold are selected from the core physical examination population to form a typical disease diagnostic code set.
Disease diagnosis code d h In cluster C k The frequency of occurrence in (a) is defined as:
wherein, | Core k I is a cluster C k The number of core physical examination people. λ (g) is an indicative function, when the disease diagnosis codes for d h In the case of a diagnostic code belonging to the j-th subject, lambda (d) h ,Di′ j ) =1, otherwise λ (d) h ,Di′ j )＝0。
As an example, cluster C k Of these, 100 subjects had high blood pressure (d) in all 50 subjects h ) With this diagnostic code, a return probability of 50/100=0.5 is obtained.
Defining a cluster C by setting a threshold delta for the frequency of occurrence of disease diagnostic codes k Typical disease diagnostic code set of (a):
TICDS k ＝{d h |Frequency k (d h )＞δ}
step S303, calculating a cluster C k Average order of each typical disease diagnostic code in (1). The smaller the average order of typical disease diagnosis codes, the greater the probability of becoming a primary diagnosis.
Definition of a Cluster C k Average order of each typical disease diagnosis code in (1):
wherein H' is the number of typical disease diagnosis coding sets,
coding for typical disease diagnosis d h Diagnostic information Di of the subject j j In the order of (1).
Step S304, according to the average sequence, the cluster C is processed k The typical disease diagnosis coding set extracted in the step (c) is repeatedNew sorting to obtain cluster C k The major disease diagnosis type and the minor disease diagnosis type.
By aligning clusters C k The extracted typical disease diagnosis code set is reordered, and the main disease diagnosis type and the secondary disease diagnosis type in the cluster can be identified.
Defining a ranking function Rank (g) to obtain a primary/secondary typical disease diagnosis coding set with a sequence characteristic:
Average k (d 1 ) The return value of the function is positive.
As an example, if there are a, B and C,3 diseases, and the corresponding average is 5.3, 2.5, 3.6, respectively, after reordering by Rank function, the order becomes B, C, a, and the more the ranking is, the more likely the primary diagnosis is.
Step S206, displaying a cluster C in the ICD body structure k According to the concept semantic relationship between the disease diagnosis codes under the same branch, the main/secondary typical disease diagnosis code set identifies the nearest father node to obtain a new disease type, and the new disease types are reordered to obtain a typical disease co-occurrence mode.
Specifically, the cluster C is displayed in an ICD-10 disease classification system by combining with a visual analysis technology k According to the semantic relation between the diagnostic codes of the diseases under the same branch, the nearest father node is automatically identified, and a new disease type e is defined i I.e. by
e i ＝LCA(d 1 ,d 2 ,L,d m )
Wherein d is 1 、d 2 ……d m Coding for disease diagnosis in the m ICD-10 system, LCA (d) 1 ,d 2 ,L,d m ) To code d 1 、d 2 ……d m The nearest parent node of.
Thus, cluster C k Major/minor typical disease diagnosis algorithmThe code set is converted into a set of new disease types { e } 1 ,e 2 ,…,e m0 And obtaining a typical disease co-occurrence mode, namely a primary result of unified diagnosis through reordering.
And step S207, receiving the auxiliary operation of the main examining physician, and adjusting the co-occurrence mode of the typical diseases to obtain a unified diagnosis library.
And the main examining physician carries out operations such as correction, classification, evaluation, definition and the like on the identified typical disease co-occurrence mode according to the self-domain knowledge to obtain a final unified diagnosis result.
By summarizing and outlining typical disease co-occurrence patterns of all clusters, a typical disease co-occurrence pattern library for the selected physical examination population, namely a unified diagnosis library, is constructed.
The feasibility of the experimental result is verified by taking the clinical diagnosis data of the electronic medical record as an example.
The diagnosis information of 4418 uremic patients in the MIMIC-III data set disclosed was selected and expressed in the form of ICD-9 codes (pure numeric representation, no letters), and 3 clusters were finally determined by similarity measurement of diagnosis information, AP clustering, extraction of primary/secondary typical disease diagnosis code sets, wherein the typical disease diagnosis code set is shown in fig. 6.
Taking the patient diagnosis information in cluster 3 as an example, the typical disease diagnosis code set and the sequence thereof are visually analyzed in the ICD-9 body structure, as shown in fig. 7, wherein a unified diagnosis is urogenital disease, circulatory disease, and acute respiratory failure, and a secondary diagnosis (complication) is sepsis, anemia, and the like.
For a new admitted patient whose diagnostic codes are 519.09, 518.81, 491.21, 38.9, 995.92, 785.52, 584.9, 482.1, 427.31, 519.19 respectively, similarity measures of the patient's diagnostic information and a typical disease diagnostic code set of 3 clusters were calculated, the similarity with cluster 3 was found to be the greatest, and a unified diagnosis of cluster 3 was recommended to the patient, so that the unified diagnosis of the patient is a fusion of the cluster 3 diagnosis and the current examination diagnosis, i.e., the primary diagnosis is respiratory disease, urinary system disease, and circulatory system disease, and the secondary diagnosis (complications) is sepsis, anemia, etc., as shown in fig. 8.
Referring to fig. 9, based on the same inventive concept as the above method embodiment, another embodiment of the present invention further provides a recommendation apparatus for unified diagnosis, where the apparatus includes a similarity measurement module 901, a matching detection module 902, and a recommendation module 903.
In particular, the similarity measure module 901 is used for determining a diagnosis code and a cluster C according to the diagnosis code of the subject k Similarity measure results of the primary/secondary canonical disease diagnosis coding set in (1); the cluster C k Are K clusters divided according to the similarity between the diagnostic information of N subjects. The matching detection module 902 is configured to select a cluster to which the similarity value greater than the preset threshold belongs, and obtain a disease co-occurrence pattern corresponding to each cluster from a unified diagnosis library. The recommending module 903 is configured to perform visual analysis on the obtained disease co-occurrence mode in combination with the diagnostic codes in the ICD body structure, identify a nearest parent node according to a semantic relationship between the diagnostic codes under the same branch, obtain a disease type, and reorder the disease co-occurrence mode to obtain the disease co-occurrence mode of the examinee.
Preferably, the device further comprises a unified diagnosis library construction module, and the unified diagnosis library construction module comprises a typical disease co-occurrence pattern acquisition module and a unified diagnosis library generation module. Specifically, the typical disease co-occurrence mode acquisition module is configured to display the primary/secondary typical disease diagnosis code sets in the ICD body structure, identify the nearest father node according to a semantic relationship between disease diagnosis codes under the same branch to obtain a new disease type, and reorder the new disease types to obtain the typical disease co-occurrence mode. And the unified diagnosis library generation module is used for receiving the auxiliary operation of the main examination physician and adjusting the typical disease co-occurrence mode to obtain a unified diagnosis library.
Preferably, the device further comprises a main/secondary typical disease diagnosis code set acquisition module, and the main/secondary typical disease diagnosis code set acquisition module comprises a core physical examination crowd determination module and a typical disease diagnosis code set generation moduleThe device comprises an average sequence calculation module and an acquisition module. Specifically, the core physical examination crowd determination module is used for the cluster C k The people with similarity greater than the similarity threshold value with the detected person are selected as the core physical examination people. The typical disease diagnosis code set generation module is used for selecting the diagnosis codes of the subjects with the frequency greater than the frequency threshold value in the core physical examination crowd to form a typical disease diagnosis code set. The average order calculation module is used for calculating a cluster C k Average order of each typical disease diagnostic code in (1). The acquisition module is used for aligning the cluster C according to the average sequence k Reordering the extracted typical disease diagnostic code set to obtain cluster C k The major disease diagnosis type and the minor disease diagnosis type.
Preferably, the apparatus further comprises a similarity measure module, wherein the similarity measure module comprises an information measure module of the disease diagnosis codes, a similarity measure module between the disease diagnosis codes, and a similarity measure module between the disease diagnosis code sets.
Referring to fig. 10, fig. 10 shows a schematic structural diagram of a possible electronic device according to the above embodiment. The electronic device may include: a processing unit 1001, a storage unit 1002, and a communication unit 1003. The processing unit 1001 may be arranged in communication with the storage unit 1002. The storage unit 1002 is configured to store executable instructions and/or program codes of the processing unit 1001, and the like, wherein the processing unit is configured to execute a recommendation method for unified diagnosis provided by any one of the method embodiments described above. The communication unit 1003 is configured to support communication between the electronic device and other network entities to implement functions such as data interaction, for example, the communication module 1003 supports communication between the electronic device and other intelligent terminals to implement a data interaction function.
The processing unit 1001 may be a processor or a controller, among others. The communication module 1003 may be a transceiver, an RF circuit or a communication interface, etc. The storage unit 1002 may be a memory.
Fig. 10 is only one possible implementation manner of the embodiment of the present application, and in practical applications, the electronic device may further include more or less components, which is not limited herein.
It should be noted that the electronic device may be a server or an intelligent terminal, and the intelligent terminal may be a computer, a tablet computer, or a smart phone.
The embodiment of the present invention further provides a storage medium, where the storage medium stores computer-readable program instructions, and when the program instructions are executed by a processing unit, the recommendation method for unified diagnosis provided in any one of the above embodiments is implemented. For example, the computer readable storage medium may be a Read-Only Memory (ROM), a Random Access Memory (RAM), a Compact Disc Read-Only Memory (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, and the like.
It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, as for the device and server embodiments, since they are substantially similar to the method embodiments, the description is simple, and the relevant points can be referred to the partial description of the method embodiments.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (4)
1. A method for recommending unified diagnosis, the method comprising the steps of:
determining a diagnostic code and cluster C based on the diagnostic code of the subject k Similarity measure results of the primary/secondary canonical disease diagnosis coding set in (1); the cluster C k Is K clusters divided according to the similarity between the diagnostic information of N subjects;
selecting the cluster to which the similarity value greater than a preset threshold value belongs, and acquiring a disease co-occurrence mode corresponding to each cluster from a unified diagnosis library;
combining the acquired disease co-occurrence mode with the diagnosis codes to perform visual analysis on the ICD body structure, identifying the nearest father node according to the semantic relation among the diagnosis codes under the same branch, acquiring the disease type and reordering to acquire the unified diagnosis of the examinee;
the construction method of the unified diagnosis library comprises the following steps:
displaying the primary/secondary typical disease diagnosis code sets in the ICD body structure, identifying a nearest father node according to the semantic relationship among the disease diagnosis codes under the same branch to obtain a new disease type, reordering the new disease types to obtain a typical disease co-occurrence mode, receiving the auxiliary operation of a main examination doctor, and adjusting the typical disease co-occurrence mode to obtain a unified diagnosis library;
the method for obtaining the primary/secondary typical disease diagnosis coding set comprises the following steps:
in cluster C k The people with similarity greater than the similarity threshold value with the detected person are selected as the core physical examination people, and the frequency of the selection among the core physical examination people is largeThe diagnostic code of the subject at the frequency threshold constitutes a typical disease diagnostic code set, and a cluster C is calculated k According to which cluster C is coded k Reordering the extracted typical disease diagnostic code set to obtain cluster C k Major disease diagnosis type and minor disease diagnosis type;
the similarity measure includes an information content measure of disease diagnosis codes, a similarity measure between disease diagnosis codes, and a similarity measure between disease diagnosis code sets.
2. A recommendation device for unified diagnosis, the device comprising:
a similarity metric module for determining a diagnostic code of the subject from the diagnostic code and cluster C k Similarity measure results of the primary/secondary canonical disease diagnosis coding set in (1); the cluster C k Is K clusters divided according to the similarity between the diagnostic information of N subjects;
the matching detection module is used for selecting the cluster to which the similarity value greater than the preset threshold value belongs and acquiring a disease co-occurrence mode corresponding to each cluster from a unified diagnosis library;
the recommendation module is used for performing visual analysis on the ICD body structure by combining the acquired disease co-occurrence mode with the diagnosis codes, identifying a nearest father node according to the semantic relation among the diagnosis codes under the same branch, obtaining the disease type and reordering to obtain the unified diagnosis of the examinee;
a unified diagnostic library building module, the unified diagnostic library building module comprising: the ICD comprises a typical disease co-occurrence mode acquisition module, a unified diagnosis library generation module and a unified diagnosis library generation module, wherein the typical disease co-occurrence mode acquisition module is used for displaying the main/secondary typical disease diagnosis code sets in an ICD body structure, identifying a nearest father node according to the semantic relationship between the disease diagnosis codes under the same branch to obtain a new disease type, and reordering the new disease types to obtain a typical disease co-occurrence mode;
a primary/secondary canonical disease diagnosis coding set acquisition module, the primary/secondary canonical disease diagnosis coding set acquisition module comprising: a core physical examination crowd determination module for determining a physical examination crowd in a cluster C k The system comprises a typical disease diagnosis code set generation module, an average sequence calculation module and a cluster C calculation module, wherein the typical disease diagnosis code set generation module is used for selecting the diagnosis codes of the examinees with the similarity greater than a similarity threshold value from the core physical examination crowd to form a typical disease diagnosis code set, and the diagnosis codes of the examinees with the occurrence frequency greater than a frequency threshold value from the core physical examination crowd k An average order of each of the typical disease diagnosis codes, an obtaining module for obtaining clusters C according to the average order k Reordering the extracted typical disease diagnostic code set to obtain cluster C k Major disease diagnosis type and minor disease diagnosis type;
a similarity measurement module comprising an information measurement module of disease diagnosis codes, a similarity measurement module between disease diagnosis codes, and a similarity measurement module between disease diagnosis code sets.
3. An electronic device, comprising:
a processing unit;
a memory unit for storing processing unit executable instructions;
wherein the processing unit is configured to: performing the method of claim 1.
4. A storage medium having computer-readable program instructions stored therein, which when executed by a processing unit, implement the method of claim 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010409050.0A CN111599427B (en) | 2020-05-14 | 2020-05-14 | Recommendation method and device for unified diagnosis, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010409050.0A CN111599427B (en) | 2020-05-14 | 2020-05-14 | Recommendation method and device for unified diagnosis, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111599427A CN111599427A (en) | 2020-08-28 |
| CN111599427B true CN111599427B (en) | 2023-03-31 |
Family
ID=72192212
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010409050.0A Active CN111599427B (en) | 2020-05-14 | 2020-05-14 | Recommendation method and device for unified diagnosis, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111599427B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112017788B (en) * | 2020-09-07 | 2023-07-04 | 平安科技（深圳）有限公司 | Disease ordering method, device, equipment and medium based on reinforcement learning model |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWM478423U (en) * | 2013-11-29 | 2014-05-21 | Univ Southern Taiwan Sci & Tec | Patient disease diagnosis and exploration system |
| CN104298344A (en) * | 2013-07-16 | 2015-01-21 | 精工爱普生株式会社 | Information processing apparatus, information processing method, and information processing system |
| CN106934235A (en) * | 2017-03-09 | 2017-07-07 | 中国科学院软件研究所 | Patient's similarity measurement migratory system between a kind of disease areas based on transfer learning |
| CN106951684A (en) * | 2017-02-28 | 2017-07-14 | 北京大学 | A kind of method of entity disambiguation in medical conditions idagnostic logout |
| CN108154928A (en) * | 2017-12-27 | 2018-06-12 | 北京嘉和美康信息技术有限公司 | A kind of methods for the diagnosis of diseases and device |
| WO2019132685A1 (en) * | 2017-12-29 | 2019-07-04 | Общество С Ограниченной Ответственностью "Интеллоджик" | Method and system for supporting medical decision making |
| CN110299209A (en) * | 2019-06-25 | 2019-10-01 | 北京百度网讯科技有限公司 | Similar case history lookup method, device, equipment and readable storage medium storing program for executing |
| CN110660459A (en) * | 2019-08-30 | 2020-01-07 | 腾讯科技（深圳）有限公司 | Method, device, server and storage medium for controlling medical record quality |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016022438A1 (en) * | 2014-08-08 | 2016-02-11 | Icahn School Of Medicine At Mount Sinai | Automatic disease diagnoses using longitudinal medical record data |
-
2020
- 2020-05-14 CN CN202010409050.0A patent/CN111599427B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104298344A (en) * | 2013-07-16 | 2015-01-21 | 精工爱普生株式会社 | Information processing apparatus, information processing method, and information processing system |
| TWM478423U (en) * | 2013-11-29 | 2014-05-21 | Univ Southern Taiwan Sci & Tec | Patient disease diagnosis and exploration system |
| CN106951684A (en) * | 2017-02-28 | 2017-07-14 | 北京大学 | A kind of method of entity disambiguation in medical conditions idagnostic logout |
| CN106934235A (en) * | 2017-03-09 | 2017-07-07 | 中国科学院软件研究所 | Patient's similarity measurement migratory system between a kind of disease areas based on transfer learning |
| CN108154928A (en) * | 2017-12-27 | 2018-06-12 | 北京嘉和美康信息技术有限公司 | A kind of methods for the diagnosis of diseases and device |
| WO2019132685A1 (en) * | 2017-12-29 | 2019-07-04 | Общество С Ограниченной Ответственностью "Интеллоджик" | Method and system for supporting medical decision making |
| CN110299209A (en) * | 2019-06-25 | 2019-10-01 | 北京百度网讯科技有限公司 | Similar case history lookup method, device, equipment and readable storage medium storing program for executing |
| CN110660459A (en) * | 2019-08-30 | 2020-01-07 | 腾讯科技（深圳）有限公司 | Method, device, server and storage medium for controlling medical record quality |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111599427A (en) | 2020-08-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10929420B2 (en) | Structured report data from a medical text report | |
| US11881293B2 (en) | Methods for automatic cohort selection in epidemiologic studies and clinical trials | |
| CN110051324B (en) | Method and system for predicting death rate of acute respiratory distress syndrome | |
| CN110291555B (en) | Systems and methods for facilitating computational analysis of health conditions | |
| RU2459244C2 (en) | Clinician-driven example-based computer-aided diagnosis | |
| US20220084633A1 (en) | Systems and methods for automatically identifying a candidate patient for enrollment in a clinical trial | |
| CN111243753B (en) | Multi-factor correlation interactive analysis method for medical data | |
| WO2018073271A1 (en) | Systems, methods, and apparatus for linking family electronic medical records and prediction of medical conditions and health management | |
| CN111951965A (en) | Panoramic health dynamic monitoring and predicting system based on time sequence knowledge graph | |
| CN114582496A (en) | Common gynecological disease prediction model construction method and prediction system | |
| CN114078593A (en) | Clinical decision support | |
| CN111599427B (en) | Recommendation method and device for unified diagnosis, electronic equipment and storage medium | |
| CN114191665A (en) | Method and device for classifying man-machine asynchronous phenomena in mechanical ventilation process | |
| CN111640517B (en) | Medical record coding method and device, storage medium and electronic equipment | |
| CN108154919B (en) | Hospital department information processing method and system | |
| CN113707326B (en) | Clinical early warning method, early warning system and storage medium | |
| CN110610766A (en) | Apparatus and storage medium for deriving probability of disease based on symptom feature weight | |
| CN114201613B (en) | Test question generation method, test question generation device, electronic device, and storage medium | |
| JP2021507392A (en) | Learning and applying contextual similarities between entities | |
| CN115762698B (en) | Medical chronic disease examination report data extraction method and system | |
| CN117690584B (en) | Intelligent AI-based chronic disease patient management system and method | |
| CN113223698B (en) | Emergency hierarchical processing method and device, electronic equipment and storage medium | |
| CN115600091B (en) | Classification model recommendation method and device based on multi-modal feature fusion | |
| CN116759079B (en) | Bleeding transformation judging method, device, medium and terminal based on multi-feature fusion | |
| CN116936107B (en) | Cardiac data risk analysis method, system and medium based on convolutional neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |