Database generation method and device, equipment, storage medium and program product

Abstract

The invention provides a database generation method and device, electronic equipment, a computer readable storage medium and a computer program product, and relates to the technical field of medical data analysis, in particular to the technical field of artificial intelligence such as large language models. The method comprises the following steps: determining a data structure used for creating a database for a target disease, wherein the data structure comprises a field name; determining a corresponding field generation path type based on a field name contained in the data structure; according to the field generation path type, determining field data corresponding to the corresponding field name from the storage data; and generating a database for the target disease based on the data structure and the field data corresponding to each field name. According to the method, the threshold of database construction for the target disease is reduced, and the efficiency of database construction for the target disease is improved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of medical data analysis, specifically to the field of artificial intelligence technologies such as large language models, and especially to a database generation method, device, electronic device, computer-readable storage medium, and computer program product. Background Art

[0002] Currently, many target disease databases in medical institutions rely on offline technical personnel to process existing medical data. The constructed existing database data is then imported into the database for application. This entire process is demanding and relies on the strong collaboration of various roles. This also leads to a high frequency of problems. Every error requires offline investigation and adjustment, which is time-consuming and difficult to detect.

[0003] To this end, it is very necessary to establish a low-code database construction platform for target diseases to improve construction efficiency for use by data processing personnel. Summary of the Invention

[0004] The embodiments of the present disclosure provide a database generation method, device, electronic device, computer-readable storage medium, and computer program product.

[0005] In a first aspect, an embodiment of the present disclosure proposes a database generation method, including: determining a data structure used to create a database for a target disease, the data structure including field names; determining a corresponding field generation path type based on the field names contained in the data structure; determining field data corresponding to the corresponding field names from the input data according to the field generation path type; and generating a database for the target disease based on the data structure and the field data corresponding to each field name.

[0006] In the second aspect, an embodiment of the present disclosure proposes a database generation device, including: a first determination module, used to determine the data structure used to create a database for a target disease, the data structure including field names; a second determination module, used to determine the corresponding field generation path type based on the field names contained in the data structure; a third determination module, used to determine the field data corresponding to the corresponding field name from the input data according to the field generation path type; a generation module, used to generate the database for the target disease based on the data structure and the field data corresponding to each field name.

[0007] In a third aspect, an embodiment of the present disclosure provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can implement the database generation method described in any implementation method in the first aspect when executing the instructions.

[0008] In a fourth aspect, an embodiment of the present disclosure provides a non-transitory computer-readable storage medium storing computer instructions, which are used to enable a computer to implement the database generation method described in any implementation manner in the first aspect when executed.

[0009] In a fifth aspect, an embodiment of the present disclosure provides a computer program product comprising a computer program, which, when executed by a processor, can implement the database generation method described in any implementation manner in the first aspect.

[0010] The database generation method provided by the present disclosure determines the field generation path type of the field data corresponding to each field name through the field names in the structural data used to create the database for the target disease, and classifies each field name according to its corresponding field generation path, so as to adopt different field data generation methods for different categories, establish a path between the stored data and the data in the database for the target disease, lower the construction threshold of the database for the target disease, and improve the generation efficiency of the database for the target disease.

[0011] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Other features, objects and advantages of the present disclosure will become more apparent from a reading of the detailed description of non-limiting embodiments made with reference to the following drawings: Figure 1 is an exemplary system architecture in which the present disclosure may be applied; Figure 2 A flowchart of a database generation method provided in an embodiment of the present disclosure; Figure 3 A flowchart of a database generation method in an application scenario provided by an embodiment of the present disclosure; Figure 4 A structural block diagram of a database generation device provided in an embodiment of the present disclosure; Figure 5 A schematic structural diagram of an electronic device suitable for executing a database generation method provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0013] The following description of exemplary embodiments of the present disclosure is made in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those skilled in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description. It should be noted that the embodiments in the present disclosure and the features in the embodiments can be combined with each other unless there is a conflict.

[0014] In the technical solutions disclosed herein, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0015] In related technologies, there are two main ways to build a database for specialized diseases: one is to use traditional manual methods, where doctors collect patient medical records by themselves, then manually create fields in a spreadsheet (Excel) and fill in the data for each field; the other is for data processing personnel to perform offline operations and transfer data using scripts. After processing the existing data, they manually perform quality control on the final specialized disease database and finally enter the data into the database.

[0016] As a result, the process of building a specialized disease database lacks unified management, the cost of problem handling is high, and it is not convenient for hospital managers to carry out unified management.

[0017] Based on the above problems, the present disclosure provides a database construction solution for target diseases that can lower the threshold for database construction for target diseases and improve the efficiency of database construction for target diseases.

[0018] In all embodiments of the present disclosure, the database for the target disease can be a specialized disease database established for a certain category or a specific disease (such as cancer, diabetes, cardiovascular disease, etc.), which systematically collects and manages patient clinical data, biological sample information, imaging materials and other resources; it can also be a specialized database established according to the needs of treating a specific symptom (such as indigestion, fever, etc.), which systematically collects and manages patient clinical data, biological sample information, imaging materials and other resources.

[0019] Figure 1 An exemplary system architecture 100 is shown to which embodiments of the database generation method, apparatus, electronic device, and computer-readable storage medium disclosed herein can be applied.

[0020] like Figure 1As shown, system architecture 100 may include terminal devices 101, 102, 103, a network 104, and a server 105. Network 104 is a medium for providing communication links between terminal devices 101, 102, 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links or fiber optic cables.

[0021] Users can use terminal devices 101, 102, 103 to interact with server 105 via network 104 to receive or send messages, etc. Terminal devices 101, 102, 103 and server 105 may be installed with various applications for enabling information communication between them, such as instant messaging applications.

[0022] Terminal devices 101, 102, 103 and server 105 can be either hardware or software. When terminal devices 101, 102, 103 are hardware, they can be various electronic devices with display screens, including but not limited to smartphones, tablet computers, laptop computers, and desktop computers. When terminal devices 101, 102, 103 are software, they can be installed in the electronic devices listed above. They can be implemented as multiple software or software modules, or as a single software or software module, and are not specifically limited here. When server 105 is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or as a single server. When the server is software, it can be implemented as multiple software or software modules, or as a single software or software module, and are not specifically limited here.

[0023] Server 105 can provide various services through various built-in applications. For example, server 105 can provide a database generation application for a target disease. When running the yellow diamond disease database generation application, server 105 can achieve the following: First, it determines the data structure used to create a database for the target disease, the data structure including field names; based on the field names contained in the data structure, it determines the corresponding field generation path type; based on the field generation path type, it retrieves the field data corresponding to the corresponding field names from the incoming data; and based on the data structure and the field data corresponding to each field name, it generates a database for the target disease. It should be noted that the data structure and incoming data used to create the database for the target disease can be obtained from terminal devices 101, 102, and 103 via network 104, or they can be pre-stored locally on server 105 in various ways. Therefore, when server 105 detects that this data is already stored locally (e.g., the data structure and incoming data used to create the database for the target disease were pre-stored before processing began), it can choose to directly retrieve this data locally. In this case, exemplary system architecture 100 may also not include terminal devices 101, 102, 103 and network 104.

[0024] Because database generation requires significant computing resources and high computing power, the database generation methods provided in the subsequent embodiments of this disclosure are generally performed by a server 105 with relatively strong computing power and resources. Accordingly, the database generation apparatus is generally also located within server 105. However, it should also be noted that, when terminal devices 101, 102, and 103 also possess sufficient computing power and resources, terminal devices 101, 102, and 103 can also utilize the database generation application installed thereon to perform the aforementioned operations delegated to server 105, thereby outputting the same results as server 105. In particular, in the presence of multiple terminal devices with varying computing power, if the database generation application determines that the terminal device it is in possesses relatively strong computing power and a relatively large amount of remaining computing resources, it can allow the terminal device to perform the aforementioned operations, thereby appropriately alleviating the computing pressure on server 105. Accordingly, the database generation apparatus can also be located within terminal devices 101, 102, and 103. In this case, exemplary system architecture 100 may also exclude server 105 and network 104.

[0025] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is merely illustrative. Any number of terminal devices, networks and servers may be provided as required.

[0026] Please refer to Figure 2 , Figure 2This is a flowchart of a database generation method provided by an embodiment of the present disclosure, wherein process 200 includes the following steps: Step 201: Determine a data structure used to create a database targeting a target disease.

[0027] Specifically, in order to generate a database for a target disease, it is first necessary to determine the data structure used to create the database for the target disease. The data structure used to create the database for the target disease can also be referred to as the dataset definition used to create the database for the target disease, and refers to the components and format of the data in the database for the target disease that is ultimately generated. This data structure defines the specific data included in the generated database for the target disease, as well as the hierarchical structure of this data (for example, a three-level structure consisting of modules, forms, and fields). This data structure can generally be determined by a clinician, or based on the purpose of the database for the target disease, etc.

[0028] Optionally, the data structure may include field names. Field names refer to specialized items or attributes used to identify and describe data related to the target disease. Field names may include fields for basic patient information, such as age, gender, and family history; fields for disease diagnosis information, such as the diagnosis date and diagnostic basis (e.g., examination report number, description of imaging findings); and fields for treatment process information, such as treatment duration, treatment method, and treatment dosage.

[0029] Step 202: Based on the field names included in the data structure, determine the corresponding field generation path type.

[0030] The field generation path type refers to the type of path used to generate the field data corresponding to the field name. Field data refers to the specific data corresponding to the field name. Field data can be text, numeric, enumerated, date, address, and other types.

[0031] Specifically, different field names in the data structure have different logical definitions of the corresponding field data, and thus the methods for extracting the corresponding field data from the incoming data are also different. For example, the field data corresponding to some field names can be directly extracted from the incoming data, the field data corresponding to some field names must first be extracted from the incoming data and then logically analyzed, and the field data corresponding to some field names requires a series of calculations to obtain, and so on. Therefore, it is necessary to determine the field generation path type corresponding to each field name, that is, to classify each field name according to the field generation path by field name.

[0032] Each field name is provided with a value range definition of its corresponding field data, and the value range definition is used to determine the logical definition and / or corresponding field value options for generating the corresponding field data for the field name.

[0033] Optionally, the field generation path type corresponding to each field name may be determined based on a value range definition corresponding to each field name. Optionally, the field generation path type corresponding to each field name may be pre-set, a configuration file may be generated, and then, based on the pre-set configuration file, the field generation path type corresponding to the field name contained in the data structure may be determined.

[0034] Step 203: Generate a path type based on the field, and determine the field data corresponding to the corresponding field name from the input data.

[0035] Specifically, "incoming data" refers to the range of data included in the database for the target disease. When creating a database for the target disease, it is first necessary to filter out relevant data for the target disease corresponding to the database for the target disease from a large amount of data from various hospitals, such as the Clinical Data Repository (CDR), Hospital Information System (HIS), Clinical Information System (CIS), Laboratory Information Management System (LIS), Electronic Medical Record (EMR), and Computer Aided Design (CAD) software systems, to serve as the incoming data. Next, a field generation path corresponding to the field name used to create the database for the target disease is generated from the incoming data to obtain the corresponding field data. The incoming data can generally be determined by clinicians or based on the objectives of the database for the target disease, but this disclosure is not limited to this.

[0036] Step 204: Generate a database for the target disease based on the data structure and the field data corresponding to each field name.

[0037] Specifically, the data structure contains the dataset definition required to generate a database for the target disease, and the field data contains the specific data content of each field name in the dataset definition. Thus, based on the data structure and each field data, the database for the target disease is generated by extracting the field data corresponding to each field name from the input data.

[0038] The database generation method for the target disease provided by the embodiment of the present disclosure determines the field generation path type of the field data corresponding to each field name through the field names in the structural data used to create the database for the target disease, and classifies each field name according to its corresponding field generation path, so as to adopt different field data generation methods for different categories, establish a path between the stored data and the data in the database for the target disease, lower the construction threshold of the database for the target disease, and improve the generation efficiency of the database for the target disease.

[0039] Field generation path types include direct mapping, post-structural mapping, and logical calculation mapping. Figure 2 In step 203, the following embodiment provides a specific implementation method for determining the field data corresponding to the corresponding field name for the directly mapped field name.

[0040] In some embodiments, according to the field generation path type, the field data corresponding to the corresponding field name is determined from the input data, including: in response to the field generation path type being a direct mapping type, field matching is performed from the input data according to the corresponding field name to obtain the entity corresponding to the corresponding field name; according to the entity data corresponding to the entity, the field data corresponding to the corresponding field name is obtained.

[0041] Specifically, an entity refers to a word or phrase with descriptive meaning in the stored data, which can usually be a person's name, place name, organization name, disease name, etc.

[0042] The field generation path type is a direct mapping type, which means that the field data corresponding to the field name can be directly obtained from the input data. In this embodiment, the entity corresponding to the field name is the same as the field name.

[0043] In some embodiments, field matching from the incoming data can be performed based on the corresponding field names by using keyword retrieval, or by using a preset neural network model to directly match fields from the incoming data based on the field names. This disclosure does not limit this.

[0044] In some embodiments, the directly mapped field names include but are not limited to name, age, weight, date of birth, etc. The directly mapped field names generally correspond to outpatient records, hospitalization records, medical records, and other forms in the stored data.

[0045] The disclosed embodiment targets directly mapped field names and can directly extract the entity data of the corresponding entity from the stored data as the field data for the field name. Compared with related technologies, this method does not require complex scripts and is simple to operate, improving the efficiency of determining field data.

[0046] Regarding the above Figure 2 In step 203, the following embodiment provides a specific implementation method for determining the field data corresponding to the post-structured field name.

[0047] In some embodiments, according to the field generation path type, field data corresponding to the corresponding field name is determined from the input data, including: in response to the field generation path type being a post-structured mapping type, entities and relationships corresponding to the corresponding field names are extracted from the input data based on the corresponding field names; according to the entities and relationships, field values ​​corresponding to the corresponding field names are generated to obtain field data corresponding to the corresponding field names.

[0048] In this embodiment, the entity corresponding to the field name is directly related to the field name. For example, the entity corresponding to the field name includes part of the keywords in the field name.

[0049] Specifically, the field generation path type of post-structured mapping type means that entities and relationships related to the field name need to be extracted from the text data in the input data, and the field data corresponding to the field name can be determined based on the entities and relationships.

[0050] The method of generating field values ​​corresponding to corresponding field names according to entities and relationships specifically includes: generating field values ​​corresponding to corresponding field names according to entities and relationships and based on value domain definitions corresponding to field names.

[0051] For example, if the field name is "Hepatitis History," then the incoming data must match entities such as Hepatitis and Suspected Hepatitis. Furthermore, the incoming data must also match relationships such as "Deny" or "Yes" that correspond to these entities. The domain definition for the field name "Hepatitis History" is: if Hepatitis appears in the incoming data and the relationship is "Deny," then the output is 0; if the relationship is "Yes," then the output is 1. Therefore, according to the domain definition for this field name, if both Hepatitis and "Deny" appear, the output is 0, which serves as the field value corresponding to this field name, i.e., the field data corresponding to this field name.

[0052] The field names corresponding to the post-structured mapping type include but are not limited to disease, radiotherapy date, radiotherapy site, etc. The field names corresponding to the post-structured mapping type generally correspond to long text data such as examination reports and medical records in the stored data.

[0053] The disclosed embodiment is aimed at post-structured field names. It can extract entities and relationships corresponding to the field names from the input data, and determine the field value of the field name based on the value range definition corresponding to the field name, and finally obtain the field data of the field name, thereby improving the efficiency of determining the field data.

[0054] Regarding the above Figure 2 In step 203, the following embodiment provides a specific implementation method for determining field data corresponding to a field name mapped to a logical calculation.

[0055] In some embodiments, according to the field generation path type, field data corresponding to the corresponding field name is determined from the input data, including: in response to the field generation path being a logical calculation mapping type, multiple entities corresponding to the corresponding field name are extracted from the input data according to the corresponding field name; calculation is performed based on the entity data corresponding to the multiple entities to obtain the field data corresponding to the corresponding field name.

[0056] Specifically, the field name corresponding to the logical calculation mapping type refers to a field that requires logical combination and / or calculation of entity data of multiple entities to determine the field data corresponding to the field name. In this embodiment, the entities corresponding to the field name include at least two.

[0057] Field names corresponding to logical calculation mappings can include fields that are logical combinations of multiple fields. For example, the anticoagulation treatment flag within ten minutes after the diagnosis of ST-segment elevation myocardial infarction (STEMI) contains multiple related field names: disease name (STEMI), time (after ten minutes), and treatment method (anticoagulation treatment flag). The field data corresponding to this field name can only be obtained by logically combining multiple related field names.

[0058] Field names corresponding to logical calculation mappings can also include fields that combine complex logical combinations of multiple fields with calculations. Examples include fields such as Body Mass Index (BMI), number of treatment courses for first partial remission, and the vertical diameter of the longest lymph node. For fields that combine complex logical combinations of multiple fields with calculations, you first need to determine the entity corresponding to the field name. Then, you can calculate the corresponding field data using a preset formula.

[0059] The field names corresponding to the logical calculation mapping type generally correspond to multiple medical record data or cross-text data in the input data.

[0060] The disclosed embodiment is for field names of logical calculation mapping type, and can extract multiple entities corresponding to the field name from the input data, and based on the logical combination relationship of multiple entities in the field name corresponding to the field name, finally obtain the field data of the field name, thereby improving the efficiency of determining the field data.

[0061] Regarding the above Figure 2In step 203, when determining the field data corresponding to the corresponding field name from the incoming data based on the field generation path type, it is necessary to first determine the entity in the incoming data corresponding to the field name. The entity corresponding to the field name can be pre-configured, determined based on a pre-trained neural network model, determined based on a large language model, determined based on a knowledge graph, and so on. The following embodiment provides a specific implementation method for determining the entity in the incoming data corresponding to the field name.

[0062] In some embodiments, a preset neural network model is used to determine the entity corresponding to the field name in the input data; wherein, the neural network model is obtained through neural network training based on the field generation path type corresponding to the field name.

[0063] Specifically, when training a neural network, field names, the field generation path types corresponding to the field names, and the entities and / or relationships in the input data corresponding to the field names are used as training samples for neural network training. Based on the different field generation path types corresponding to the field names, a neural network model for determining the entities and / or relationships corresponding to the field names is trained.

[0064] Specifically, the present disclosure utilizes the latest Transformer architecture (a deep learning model architecture based on a self-attention mechanism) to design a model for neural network training, resulting in a neural network model. The Transformer architecture boasts 1 billion parameters and is capable of processing complex natural language. During the Transformer architecture neural network training process, a combination of pre-training and fine-tuning is employed, initially on a large-scale medical corpus to capture long-range dependencies within the text and enhance understanding of the unique semantics of the medical field. During the pre-training phase, the Transformer architecture model learns extensive medical knowledge and language patterns, laying a solid foundation for subsequent structured information extraction for target diseases. When processing specific tasks, the Transformer architecture model directly outputs entity sets and entity relationships from the input data in an end-to-end manner. To optimize the consistency of entities and relationships, a multi-task learning mechanism is designed within the Transformer architecture model. This mechanism coordinates the learning processes of entity recognition and relationship extraction through parameter sharing and inter-layer attention. Furthermore, the Transformer architecture model incorporates a specialized consistency constraint algorithm to reduce the propagation of cascading errors and ensure logical consistency between entities and relationships. During the Transformer architecture model training phase, in addition to basic pre-training, this publication also performed targeted fine-tuning on the Transformer architecture model. This fine-tuning utilized a small amount of high-quality annotated medical data covering a wide range of diseases and symptoms, effectively improving the model's performance on specific tasks. Through fine-tuning, the model achieved significant improvements in key metrics such as accuracy and recall for entity recognition and relation extraction.

[0065] To validate the model's effectiveness, extensive testing was conducted on multiple medical datasets. Experimental results show that compared to traditional single-task models, the trained neural network model achieves a 10%-20% improvement in both accuracy and recall for entity recognition and relation extraction.

[0066] Optionally, during the neural network training process, the training data is also enhanced, specifically including: based on the annotated data exported from the annotation platform, such as pathology reports, past medical history, surgical records, marital and reproductive history, auxiliary examinations, examination reports, family history, chief complaint, current medical history, personal history, menstrual history and other data, through the powerful functions of the data enhancement module, after parsing, cleaning, and automatic verification of manually annotated entities and relationships according to the schema, appending the prefix of the report type to the text, performing fragment truncation, keyword replacement, text splicing and other methods on some texts, the richness of the data is improved from multiple angles, thereby assisting the generalization ability of model training.

[0067] During neural network training, incremental model learning is used. Specifically, in the construction of databases for target diseases, new field names constantly appear. In these cases, traditional models usually need to be retrained or at least a large amount of data needs to be relabeled to incorporate new entity and relationship types. This requirement not only consumes a lot of manpower and time resources, but also greatly reduces the application efficiency and response speed of the model. The present disclosure adopts incremental learning as an effective solution, enabling the model to adapt to new data or labels while maintaining existing knowledge, making it an ideal choice for dealing with such problems. Among them, incremental learning is a machine learning technique that allows the model to gradually adapt to new data or new labels without forgetting existing knowledge. This learning model greatly improves the flexibility and scalability of the model, enabling it to effectively respond to changes in data and environment. The core of incremental learning is its ability to dynamically update the model without having to retrain the entire model from scratch. This not only saves computing resources for retraining, but also significantly reduces the labor cost required for data relabeling.

[0068] In some embodiments, the neural network model adopts a mask-based incremental learning mechanism, allowing the model to update only the part of the model related to the new entity or new relationship label through masking technology without completely retraining.

[0069] Regarding the above Figure 2 In step 203, when determining the field data corresponding to the corresponding field name from the input data based on the field generation path type, it is necessary not only to determine the entity corresponding to the field name in the input data, but also to determine the corresponding entity data based on the entity, thereby obtaining the field data corresponding to the field name. The following embodiment provides a specific implementation method for determining the entity data corresponding to the entity.

[0070] In some embodiments, determining entity data corresponding to an entity includes: setting a data retrieval priority for the entity according to a source of the entity; and determining the entity data of the entity based on the data retrieval priority.

[0071] Specifically, the same entity often appears in multiple places in the incoming data. For entities corresponding to directly mapped field names, such as name and age, they will appear in outpatient records, hospitalization records, and medical records in the incoming data. For entities corresponding to post-structured field names, such as hepatitis, they will appear in examination reports, laboratory reports, and medical records. For entities corresponding to logical calculation mapping field names, such as weight in BIM, they may appear in multiple medical records, and so on.

[0072] The present disclosure sets an entity's data retrieval priority based on the entity's source and determines the entity's entity data based on the data retrieval priority, thereby improving the efficiency and accuracy of determining the entity's entity data. Of course, in addition to setting an entity's data retrieval priority based on the entity's source, the entity's data retrieval priority can also be set based on the frequency of the entity's occurrence, the location of the entity's occurrence, etc., and the present disclosure does not impose any restrictions on this.

[0073] Regarding the above Figure 2 In step 201 , the data structure may generally be determined by a clinician, or may be determined based on the purpose of creating a database for a target disease, etc. A specific implementation of determining the data structure is given below.

[0074] In some embodiments, determining the data structure used to create a database for a target disease includes: determining a classification form used to create a database for a target disease, the field names included in each classification form, and the value domain definition corresponding to each field name; based on the classification form, field names, and value domain definition, determining the data structure used to create a database for a target disease.

[0075] Specifically, based on the purpose of creating a database for a target disease, the classification form used to create the database for the target disease, the field names included in each classification form, and the value range definition corresponding to each field name can be determined.

[0076] Optionally, a script can be used to determine the data structure used to create a database for the target disease based on the classification form, field name and value range definition; or the data structure used to create a database for the target disease can be obtained manually based on the classification form, field name and value range definition.

[0077] The embodiment of the present disclosure ensures the accuracy of the generated database for the target disease by accurately determining the data structure used to create the database for the target disease.

[0078] In addition, in the data structure used to create a database for a target disease, only the hierarchical structure of each data is defined, for example, which modules are included in the database for the target disease, which classification forms are included in each module, which field names are included in each classification form, etc., but there is no definition of how the various field names included in the classification form are arranged, which field names need to be set in the same row, and which fields need to be set in different rows. A specific implementation method is given in the following embodiment.

[0079] In some embodiments, it also includes: determining the identifier corresponding to the corresponding patient based on the source of the classification form; aggregating the field names contained in the classification form based on the identifier to obtain an aggregated result; and setting each field name in the aggregated result in the same row of the classification form.

[0080] Specifically, the patient identifiers corresponding to the classification form vary depending on the source. For example, if the classification form originates from an outpatient record, the corresponding patient identifiers may include the patient's identity identifier, the visit identifier, and the test / inspection identifier. For another example, if the classification form is from an inpatient record, the corresponding patient identifiers may include the patient's identity identifier, the inpatient identifier, and other identifiers.

[0081] Optionally, after determining the patient's corresponding identifier based on the source of the classification form, the unique identifier corresponding to the corresponding patient can be further determined according to the role of the classification form, and the field names contained in the classification form can be aggregated according to the unique identifier.

[0082] For example, in the classification form of outpatient records, one patient corresponds to one patient identity identifier. Each time the patient visits the hospital for treatment, he or she will have a treatment identifier. If multiple examinations or tests are performed during the treatment, each examination or test will also correspond to an examination / test identifier.

[0083] Assuming that the classification form is mainly used to record the various examinations the patient undergoes in the outpatient clinic, the patient identity identifier, medical consultation identifier, and examination / test identifier need to be used as the patient's corresponding identifiers at the same time; assuming that the classification form is mainly used to record all patients involved in the outpatient clinic, only the patient identity identifier needs to be used as the patient's unique identifier.

[0084] The embodiment of the present disclosure determines the identifier corresponding to the corresponding patient based on the source of the classification form, thereby aggregating the field names in the classification form to set the field names that meet the requirements of the classification form on the same line, thereby realizing automatic generation of the classification form.

[0085] When creating a database for a target disease, it is necessary not only to Figure 2 As described in the corresponding embodiments, it is necessary to determine the data structure used to create the database for the target disease, as well as the field data corresponding to each field name in the data structure. It is also necessary to determine in advance the enrollment data for creating the database for the target disease. The following embodiment provides a specific implementation method for determining the enrollment data for creating the database for the target disease.

[0086] In some embodiments, the method for generating a database for a target disease includes not only steps 201 to 204, but also includes: determining storage conditions; and determining storage data from the original database based on the storage conditions.

[0087] Specifically, entry criteria can be determined based on the database's purpose for the target disease. These criteria include data sources, standard diagnosis names, and keywords. Data sources include CDRs, HISs, LISs, EMRs, hand numbness systems, and medical record systems. For standard diagnoses, data can be determined based on International Classification of Diseases (ICD) diagnosis names, groups, and codes. For non-standard diagnoses, data can be determined based on keywords. Furthermore, specific requirements can be met, such as the time range within which patients are extracted, as an entry criterion.

[0088] Optionally, based on the entry conditions, the entry data corresponding to the entry conditions can be extracted from the original database by using condition tree search, event search, number search, etc. The data in the original database includes data from the hospital system corresponding to each data source.

[0089] Optionally, based on different retrieval methods, the determined warehousing conditions can also be configured as configuration fields that meet the retrieval method, and then the corresponding retrieval method is used based on the configuration field to determine the warehousing data from the original database to ensure that the obtained warehousing data is more accurate and comprehensive.

[0090] The embodiment of the present disclosure determines the storage data based on the storage conditions, which can ensure that the obtained storage data is more in line with the requirements of creating a database for the target disease, thereby ensuring that the generated data for the target disease is more accurate and improving the efficiency and accuracy of generating the database for the target disease.

[0091] After finalizing the field data for the target disease database to be generated, the database can be presented in either a form view or a patient view, depending on the needs. Single-view data is suitable for troubleshooting, while patient data facilitates a patient-level overview of data quality. When presented in the patient view, data administrators can modify database field values ​​for the target disease based on the original medical record. Upon saving, the underlying database for the target disease will automatically update.

[0092] To deepen understanding, this disclosure also combines a specific application scenario and gives a specific implementation solution by taking the database for a specific disease as an example. Figure 3 The process 300 shown specifically includes the following steps: Step 301, corresponding to module 1: warehousing condition configuration, specifically includes the following steps: Step 311: Acquire primary structured clinical data.

[0093] Step 321: Determine the patient data of the disease-specific database (ie, the database-entry data mentioned in the above embodiment) from the primary structured clinical data based on the database-entry conditions.

[0094] Specifically, the conditions for entering the specialized disease database can be configured based on the conditional tree search, the clinical data center data can be searched, the population to be included in the specialized disease database can be determined, and the search results can be used as the data range for entering the specialized disease database.

[0095] Step 302 corresponds to module 2: eCRF (Electronic Case Report Form) form setting, that is, data set field design.

[0096] Specifically, the data structure of the disease-specific database is defined through the eCRF form design module, including modules, forms, field names, types, etc.

[0097] Step 303, corresponding to Module 3: Structured Definition, specifically includes the following steps: Step 313: Determine that the field type of the disease-specific database is a direct mapping field.

[0098] Step 323: Determine that the field type of the disease-specific database is a post-structured field.

[0099] Step 333: Determine that the field type of the disease-specific database is a logical calculation field.

[0100] Step 343: Data aggregation.

[0101] Specifically, the field types of the disease-specific database fields are disassembled (L1 direct mapping, L2 post-structuring, L3 logical calculation), and data is taken based on the field value priority (supports taking values ​​from original medical records or post-structured entities and arranging priorities), and the disease-specific database fields are generated through data aggregation. This module is mainly used to establish a formation path between the disease-specific database fields and the original medical records / post-structured entity attributes.

[0102] Step 304 corresponds to module four: generation of disease-specific database data.

[0103] Specifically, based on the data set field design and the disease-specific library fields obtained through data aggregation, the disease-specific library data is generated.

[0104] Step 305, corresponding to module five: special disease data verification.

[0105] Specifically, the generated disease-specific database data can be verified, and the disease-specific database data can also be updated based on the verification results.

[0106] It should be noted that the specific details in the embodiment of the present disclosure are the same as those in the embodiment mentioned above and will not be repeated here. It should also be noted that the specific implementation steps in step 301 and step 303 in the embodiment of the present disclosure are only an optional implementation method and do not constitute a limitation on step 301 and step 303. Step 301 and step 303 can also achieve the same technical effects as them through other methods, and the present disclosure does not limit this.

[0107] The method for generating a database for a target disease provided by the present disclosure determines the field generation path type of the field data corresponding to each field name through the field names in the structural data used to create the database for the target disease, and classifies each field name according to its corresponding field generation path, thereby adopting different field data generation methods for different categories, establishing a path between the stored data and the data in the database for the target disease, lowering the construction threshold of the database for the target disease, and improving the generation efficiency of the database for the target disease.

[0108] Further references Figure 4 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a database generation device. Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0109] like Figure 4 As shown, the database generation device 400 of this embodiment may include: a first determination module 401, a second determination module 402, a third determination module 403, and a generation module 404. The first determination module 401 is configured to determine a data structure used to create a database for a target disease, the data structure including field names; the second determination module 402 is configured to determine a corresponding field generation path type based on the field names included in the data structure; the third determination module 403 is configured to determine, from the input data, field data corresponding to the corresponding field names based on the field generation path type; and the generation module 404 is configured to generate the database for the target disease based on the data structure and the field data corresponding to each field name.

[0110] In this embodiment, the database generation device 400 includes the first determination module 401, the second determination module 402, the third determination module 403, and the generation module 404. The specific processing and technical effects thereof can be referred to in the respective Figure 2 The relevant descriptions of steps 201-204 in the corresponding embodiment are not repeated here.

[0111] In some embodiments, the third determination module 403 is specifically used to respond to the field generation path type being a direct mapping type, perform field matching from the input data according to the corresponding field name, and obtain the entity corresponding to the corresponding field name in the input data; wherein the entity corresponding to the field name is the same as the field name; and obtain the field data corresponding to the corresponding field name according to the entity data corresponding to the entity.

[0112] In some embodiments, the third determination module 403 is specifically used to extract entities and relationships corresponding to corresponding field names from the input data based on the corresponding field names in response to the field generation path type being a post-structured mapping type; wherein the entity corresponding to the field name is related to the field name; and based on the entity data and relationship corresponding to the entity, generate a field value corresponding to the corresponding field name to obtain field data corresponding to the corresponding field name.

[0113] In some embodiments, the third determination module 403 is specifically used to extract multiple entities corresponding to the corresponding field names from the input data according to the corresponding field names in response to the field generation path being a logical calculation mapping type; and perform calculations based on the entity data corresponding to the multiple entities to obtain field data corresponding to the corresponding field names.

[0114] In some embodiments, the generating device 400 further includes a data retrieval module, which is configured to set a data retrieval priority of an entity according to a source of the entity; and determine entity data corresponding to the entity based on the data retrieval priority.

[0115] In some embodiments, the first determination module 401 is specifically used to determine the classification form used to create a database for the target disease, the field names included in each classification form, and the value range definition corresponding to each field name; based on the classification form, field names and value range definition, determine the data structure used to create a database for the target disease.

[0116] In some embodiments, the generating device 400 also includes a setting module, which is used to determine the identifier corresponding to the corresponding patient based on the source of the classification form; aggregate the field names contained in the classification form based on the identifier to obtain an aggregated result; and set each field name in the aggregated result in the same row of the classification form.

[0117] In some embodiments, the generating device 400 further includes a storage data determination module, which is used to determine storage conditions; wherein the storage conditions include data source, standard diagnosis name and keywords; based on the storage conditions, the storage data is determined from the original database.

[0118] In some embodiments, the generating device 400 also includes an entity determination module, which is used to use a preset neural network model to determine the entity corresponding to the field name in the input data; wherein the neural network model is obtained through neural network training based on the field generation path type corresponding to the field name.

[0119] This embodiment exists as an apparatus embodiment corresponding to the above-mentioned method embodiment. The database generation apparatus 400 provided in this embodiment determines the field generation path type of the field data corresponding to each field name by using the field names in the structural data for creating a database for the target disease, and classifies each field name according to its corresponding field generation path, thereby adopting different field data generation methods for different categories, establishing a path between the stored data and the data in the database for the target disease, lowering the construction threshold of the database for the target disease, and improving the generation efficiency of the database for the target disease.

[0120] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the database generation method described in any of the above embodiments can be implemented when the at least one processor executes them.

[0121] According to an embodiment of the present disclosure, the present disclosure further provides a readable storage medium, which stores computer instructions, and the computer instructions are used to enable a computer to implement the database generation method described in any of the above embodiments when executed.

[0122] According to an embodiment of the present disclosure, the present disclosure further provides a computer program product, which, when executed by a processor, can implement the database generation method described in any of the above embodiments.

[0123] Figure 5 A schematic block diagram of an example electronic device 500 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are provided as examples only and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0124] like Figure 5As shown, electronic device 500 includes a computing unit 501, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. RAM 503 may also store various programs and data required for the operation of electronic device 500. Computing unit 501, ROM 502, and RAM 503 are interconnected via a bus 504. An input / output (I / O) interface 505 is also connected to bus 504.

[0125] Multiple components in the electronic device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a magnetic disk, an optical disk, etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 allows the electronic device 500 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0126] The computing unit 501 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 501 performs the various methods and processes described above, such as the method for generating a database for a target disease. For example, in some embodiments, the method for generating a database for a target disease can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 508. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 500 via the ROM 502 and / or the communication unit 509. When the computer program is loaded into the RAM 503 and executed by the computing unit 501, one or more steps of the method for generating a database for a target disease described above can be performed. Alternatively, in other embodiments, the computing unit 501 can be configured to perform the method for generating a database for a target disease by any other suitable means (e.g., via firmware).

[0127] Various embodiments of the systems and techniques described above can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0128] The program code for implementing the method of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device so that when the program code is executed by the processor or controller, the functions / operations specified in the flow chart and / or block diagram are implemented. The program code can be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0129] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), optical fibers, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0130] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0131] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0132] A computer system may include a client and a server. The client and server are generally remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers and establishing a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host. This is a hosting product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosting and virtual private server (VPS) services.

[0133] According to the technical solution of the embodiment of the present disclosure, the field names in the structural data used to create the target disease database are used to determine the field generation path type of the field data corresponding to each field name, and each field name is classified according to its corresponding field generation path, so that different field data generation methods are adopted for different categories, and a path between the stored data and the data in the target disease database is established, which lowers the construction threshold of the target disease database and improves the generation efficiency of the target disease database.

[0134] It should be noted that each step in the above database generation solution can be executed by a computing device / electronic device.

[0135] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this disclosure can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved. This is not a limitation herein.

[0136] The above specific embodiments do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure shall be included within the scope of protection of this disclosure.

Claims

1. A database generation method, comprising: determining a data structure for creating a database for a target disease, the data structure including field names; Determining a corresponding field generation path type based on the field names contained in the data structure; Generate a path type based on the field, and determine the field data corresponding to the corresponding field name from the input data; The database for the target disease is generated based on the data structure and the field data corresponding to each of the field names.

2. The method according to claim 1, wherein The generating path type according to the field and determining the field data corresponding to the corresponding field name from the input data include: In response to the field generation path type being a direct mapping type, performing field matching from the input data according to the corresponding field name to obtain an entity corresponding to the corresponding field name in the input data; wherein the entity corresponding to the field name is the same as the field name; According to the entity data corresponding to the entity, field data corresponding to the corresponding field name is obtained.

3. The method according to claim 1, wherein The generating path type according to the field and determining the field data corresponding to the corresponding field name from the input data include: In response to the field generation path type being a post-structured mapping type, extracting entities and relationships corresponding to corresponding field names from the input data based on corresponding field names; wherein the entities corresponding to the field names are related to the field names; According to the entity data and relationship corresponding to the entity, a field value corresponding to the corresponding field name is generated to obtain the field data corresponding to the corresponding field name.

4. The method according to claim 1, wherein The generating path type according to the field and determining the field data corresponding to the corresponding field name from the input data include: In response to the field generation path being a logic calculation mapping type, extracting a plurality of entities corresponding to the corresponding field names from the input data according to the corresponding field names; Calculation is performed based on the entity data corresponding to the multiple entities to obtain field data corresponding to the corresponding field names.

5. The method according to any one of claims 2 to 4, further comprising: Setting the data retrieval priority of the entity according to the source of the entity; Entity data corresponding to the entity is determined based on the data retrieval priority.

6. The method according to claim 1, wherein The determining of a data structure for creating a database for a target disease includes: Determining a classification form for creating the database for the target disease, field names included in each classification form, and a value range definition corresponding to each field name; Based on the classification form, the field names and the value range definitions, a data structure used to create the database for the target disease is determined.

7. The method according to claim 6, further comprising: Determining an identifier corresponding to a corresponding patient based on the source of the classification form; Aggregating the field names contained in the classification form based on the identifier to obtain an aggregation result; The field names in the aggregation result are set in the same row of the classification form.

8. The method according to claim 1, further comprising: Determine the storage conditions; wherein the storage conditions include data source, standard diagnosis name and keywords; Based on the storage conditions, the storage data is determined from the original database.

9. The method according to any one of claims 2 to 4, further comprising: Utilize the preset neural network model to determine the entity corresponding to the field name in the input data; wherein, the neural network model is obtained through neural network training based on the field generation path type corresponding to the field name.

10. A database generation device, comprising: A first determining module is used to determine a data structure used to create a database for a target disease, wherein the data structure includes field names; A second determining module is used to determine a corresponding field generation path type based on the field name contained in the data structure; A third determining module is used to generate a path type according to the field and determine the field data corresponding to the corresponding field name from the input data; A generating module is used to generate the database for the target disease based on the data structure and the field data corresponding to each of the field names.

11. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the generation method according to any one of claims 1 to 9.

12. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the generation method according to any one of claims 1 to 9.

13. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the steps of the generating method according to any one of claims 1 to 9.

Images