Electronic medical record generation method and device, equipment and storage medium

Abstract

The embodiment of the invention provides an electronic medical record generation method and device, equipment and a storage medium. The method comprises the following steps: determining an adaptive admission record template of a current patient according to main diagnosis information and department information of the current patient; according to the diagnosis filling data in the adaptive admission record template, corresponding target record item problems are generated, and the target record item problems comprise to-be-revised item problems and missing item problems; according to the interactive reply of the target record item question, generating a corresponding admission record supplementary draft; and backfilling the admission record supplementary draft into the adaptive admission record template to obtain an electronic medical record of the current patient. According to the embodiment of the invention, dual quality management and control of the electronic medical record can be realized by matching the admission record template and the admission record supplement draft, the comprehensiveness and normalization of the electronic medical record are ensured, the comprehensive and accurate description of the electronic medical record on the illness state of the patient is enhanced, and more efficient, convenient and reliable standardized diagnosis and treatment reference is provided for doctors.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of data processing technology, and specifically to a method, device, equipment and storage medium for generating an electronic medical record. Background Art

[0002] With the deepening popularization of medical informationization, electronic medical records have become an indispensable tool for hospitals at all levels, covering all aspects of clinical practice.

[0003] In today's clinical practice, doctors are usually required to manually enter the patient's medical history, symptoms, signs, diagnosis, treatment and other related medical information according to the requirements of a fixed medical record interface and operating procedures to generate the patient's electronic medical record.

[0004] The medical record data entered into existing electronic medical records mainly rely on the subjective entry of doctors under fixed medical record interfaces and operating procedures. This not only increases the cumbersome operations of doctors when writing electronic medical records, but may also result in incomplete medical record data or entry errors, and the accuracy of electronic medical records cannot be guaranteed. Summary of the invention

[0005] The embodiments of the present application provide a method, apparatus, device and storage medium for generating an electronic medical record, so as to achieve dual quality control of the electronic medical record, ensure the comprehensiveness and standardization of the electronic medical record, enhance the comprehensive and accurate description of the patient's condition by the electronic medical record, and provide doctors with more efficient and reliable standardized diagnosis and treatment.

[0006] In a first aspect, an embodiment of the present application provides a method for generating an electronic medical record, the method comprising:

[0007] Determine an adapted admission record template for the current patient based on the primary diagnosis information and department information of the current patient;

[0008] Generate corresponding target record item questions according to the consultation filling data in the adapted admission record template, wherein the target record item questions include questions about items to be revised and questions about missing items;

[0009] Generate a corresponding supplementary draft of the admission record according to the interactive responses to the target record item questions;

[0010] The supplementary draft of the admission record is filled back into the adapted admission record template to obtain the electronic medical record of the current patient.

[0011] In a second aspect, an embodiment of the present application provides a device for generating an electronic medical record, the device comprising:

[0012] An adaptation template determination module, used to determine an adaptation admission record template for the current patient according to the primary diagnosis information and department information of the current patient;

[0013] A question generation module, used to generate corresponding target record item questions according to the consultation filling data in the adapted admission record template, wherein the target record item questions include questions about items to be revised and questions about missing items;

[0014] An admission record supplementation module, used for generating a corresponding admission record supplementation draft according to the interactive responses to the target record item questions;

[0015] The electronic medical record generation module is used to fill the admission record supplementary draft back into the adapted admission record template to obtain the electronic medical record of the current patient.

[0016] In a third aspect, an embodiment of the present application provides an electronic device, the electronic device comprising:

[0017] A processor and a memory, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to execute the method for generating an electronic medical record provided in the first aspect of the present application.

[0018] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium for storing a computer program, wherein the computer program enables a computer to execute the method for generating an electronic medical record as provided in the first aspect of the present application.

[0019] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the method for generating an electronic medical record as provided in the first aspect of the present application.

[0020] The embodiment of the present application provides a method, device, equipment and storage medium for generating an electronic medical record. First, based on the main diagnosis information and department information of the current patient, an adaptive admission record template for the current patient is determined so that the doctor can fill in the current patient's medical information in a standardized manner. Then, by analyzing the medical consultation filling data in the adaptive admission record template, a corresponding target record item question consisting of questions about items to be revised and missing items is generated. The doctor's interactive response to each target record item question is obtained to generate the corresponding admission record supplementary draft, and the admission record supplementary draft is backfilled into the adaptive admission record template to obtain the electronic medical record of the current patient, thereby completing the comprehensive quality control of the electronic medical record and avoiding omissions in the description of the patient's condition in the electronic medical record. The dual quality control of the electronic medical record is achieved by adapting the admission record template and the admission record supplementary draft, ensuring the comprehensiveness and standardization of the electronic medical record, enhancing the electronic medical record's all-round and accurate description of the patient's condition, and providing doctors with a more efficient, convenient and reliable standardized diagnosis and treatment reference. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0022] Figure 1 A flowchart of a method for generating an electronic medical record shown in an embodiment of the present application;

[0023] Figure 2 A flowchart of another method for generating an electronic medical record according to an embodiment of the present application;

[0024] Figure 3 This is a schematic diagram of an interface for an interactive process of a target record item problem shown in an embodiment of the present application;

[0025] Figure 4 This is a principle block diagram of a device for generating an electronic medical record shown in an embodiment of the present application;

[0026] Figure 5 A schematic block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0028] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0029] In order to solve the problem that electronic medical records mainly rely on the subjective input of doctors under fixed medical record interfaces and operating procedures, and the accuracy of electronic medical records cannot be guaranteed, the embodiment of the present application designs a new electronic medical record generation scheme. This scheme determines the patient's adapted admission record template based on the main diagnosis information and department information of any patient, and generates corresponding target record item questions consisting of questions about items to be revised and missing items based on the consultation data filled in the adapted admission record template. Then, by obtaining the doctor's interactive response to each target record item question, the corresponding admission record supplementary draft is generated, and the admission record supplementary draft is filled back into the adapted admission record template to obtain the current patient's electronic medical record, thereby completing the comprehensive quality control of the electronic medical record, avoiding omissions in the description of the patient's condition in the electronic medical record, and ensuring the comprehensiveness and standardization of the electronic medical record.

[0030] Figure 1 This is a flow chart of a method for generating an electronic medical record according to an embodiment of the present application. Figure 1 , the method may include the following steps:

[0031] S110, determining an adapted admission record template for the current patient based on the primary diagnosis information and department information of the current patient.

[0032] Among them, the current patient can be any patient in the medical service system who needs a doctor to diagnose and treat his condition. When the current patient visits the hospital, he usually needs to undergo various medical examinations with the assistance of a doctor, so as to obtain the results of various medical examinations of the current patient. Then, through the doctor's diagnosis and analysis of the various medical examination results of the current patient, the specific condition of the current patient can be preliminarily diagnosed, thereby obtaining the main diagnosis information of the current patient, which describes the main medical diagnosis results obtained by the doctor after the preliminary diagnosis of the current patient.

[0033] Moreover, in order to ensure the quality of diagnosis and treatment when the patient visits the hospital, the entire process of diagnosis and treatment (such as patient course information, examination and test results, doctor's orders, surgical information, nursing information, etc.) of the patient's visit to the hospital is usually fully recorded through electronic medical records. And since any patient will usually choose a doctor in the appropriate department for treatment according to their own disease type when visiting the hospital. And different departments will face a variety of different diseases, and the same department will also face a variety of different diseases. So, in order to ensure the accuracy of the patient's electronic medical record, this application can pre-set an admission record template for each disease under each department by analyzing the specific record requirements of the medical institution for the patient's condition diagnosis information under the disease. In other words, each disease under each department will correspond to an admission record template, which can be used as an electronic medical record in the initial style when the patient visits the hospital, and the final electronic medical record can be obtained by filling in the corresponding patient diagnosis and treatment information in the admission record template.

[0034] In this application, in order to ensure the efficiency and accuracy of the current patient's electronic medical record, after the current patient is treated, the main diagnosis information of the current patient can be obtained first. Then, by analyzing the main diagnosis information of the current patient, the main diagnosis result of the doctor's preliminary diagnosis of the current patient can be determined to determine the type of disease of the current patient. Moreover, this application can also determine the department information of the current patient by judging which department the current patient enters in the hospital for treatment. Thus, according to the type of disease and department information of the disease of the current patient, an admission record template that matches the type of disease and department information of the disease of the current patient can be found from multiple admission record templates pre-set for different types of diseases in different departments, as the adaptation admission record template for the current patient. Among them, the adaptation admission record template can comprehensively and normatively describe the various disease diagnosis information that needs to be recorded for the current patient in the corresponding department and the corresponding disease, so as to ensure the accuracy of the electronic medical record.

[0035] S120, generating corresponding target record item questions according to the consultation filling data in the adapted admission record template, wherein the target record item questions include questions about items to be revised and questions about missing items.

[0036] After obtaining the adapted admission record template for the current patient, the present application can display the adapted admission record template to the doctor who saw the current patient during the visit, and the doctor will fill in the specific diagnosis results of the current patient's condition after the visit to the corresponding position in the adapted admission record template to complete the complete record of the current patient's visit. Thus, in the process of filling in the adapted admission record template during the visit, the present application can continuously obtain the specific filled-in content in the adapted admission record template as the visit filling-in data in the present application. Among them, the visit filling-in data may include the specific disease diagnosis results of the visiting doctor on the current patient after this visit under various relevant disease parameters.

[0037] Since the consultation data in the adapted admission record template is a subjective description of the attending doctor's diagnosis of the patient's condition after seeing the current patient, the doctor may miss asking, miss writing, or make mistakes about the patient's condition, resulting in the adapted admission record template failing to fully record the patient's diagnosis and treatment information or the recorded content is incorrect.

[0038] Therefore, in order to ensure the comprehensive integrity of electronic medical records, this application can pre-acquire a large amount of historical medical record data of patients for each disease in each department of each medical institution, and analyze a large amount of historical medical record data for each disease in each department, and determine the various disease diagnosis information of the patient in the entire disease diagnosis process that needs to be recorded in each department and each disease, so as to determine the complete diagnosis record object of the patient in each department and each disease.

[0039] Then, after obtaining the consultation data filled in the adapted admission record template, this application can first find out the complete diagnosis record object of the current patient under the current consultation department and the disease this time. Then, by analyzing the consultation data filled in the adapted admission record template of the current patient, the recorded object of the current patient after this visit can be determined. Furthermore, by comparing and analyzing the complete diagnosis record object of the current patient under the current consultation department and the disease this time and the recorded object after this visit, it can be determined whether there are erroneous content that has been recorded and needs to be revised or missing items that have not been recorded for the current patient after this visit, thereby generating corresponding questions about items to be revised and missing items, and forming the questions about items to be revised and missing items into corresponding target record item questions, so that the current patient's electronic medical record can be continuously improved by asking the current patient's attending doctor questions about each target record item and the attending doctor responding.

[0040] S130, generating a corresponding supplementary draft of the admission record based on the interactive responses to the target record item questions.

[0041] For each target record item question, the present application can display each target record item question to the current patient's attending physician, who will answer each target record item question, thereby determining the current patient's specific revision status under incorrect record content and the specific diagnosis status under each missing item, so as to facilitate subsequent supplementary records to the adapted admission record template to ensure the comprehensiveness of the electronic medical record.

[0042] That is to say, through each target record item question, it is possible to interact intelligently with the current patient's attending physician, and obtain the specific answer content of the attending physician's answer to each target record item question as an interactive reply to the target record item question. Among them, the interactive reply to each target record item question can describe the specific real diagnosis result of the current patient under the target record item. Therefore, by integrating each target record item question and the interactive reply content of each target record item question, a corresponding supplementary draft of the admission record can be generated to describe the attending physician's true diagnosis result of the relevant condition of the current patient that was not recorded or recorded incorrectly due to missed questions or missed writing after this visit. It can be seen from this that through each target record item question, it is possible to perform record improvement operations such as adding, revising and editing the consultation data filled in the current patient's adapted admission record template to ensure the comprehensive integrity of the electronic medical record.

[0043] As an exemplary solution in the present application, for each target record item question, the present application can display a corresponding interactive window to the current patient's doctor, and continuously display each target record item question in the interactive window to support the doctor in answering each target record item question accordingly. The interactive window can be a dialogue window, and each target record item question is answered in the form of a dialogue.

[0044] Alternatively, for each target record item question, the supplementable position of each target record item question can be determined directly in the consultation data filled in the current patient's adapted admission record template, and a question comment box can be initiated at the supplementable position to display the target record item question and support the attending doctor to answer the displayed target record item question in the question comment box.

[0045] This application does not limit the interactive form of each target record item question, and supports the current patient's attending physician to answer each target record item question.

[0046] S140, fill the admission record supplementary draft back into the adapted admission record template to obtain the electronic medical record of the current patient.

[0047] After obtaining the supplementary draft of the admission record of the current patient, multiple target question-answer pairs consisting of each target record item question and its interactive reply can be extracted by parsing the supplementary draft of the admission record. Then, for each target question-answer pair, the relevant disease diagnosis information represented by each target question-answer pair can be analyzed, and the content that the attending doctor needs to modify or add in the current patient's adapted admission record template during the interaction with each target record item question can be determined as the content to be supplemented in the adapted admission record template.

[0048] Then, by analyzing the specific diagnosis types of the items to be supplemented, it is possible to determine the appropriate filling positions in the adapted admission record template, so that the items to be supplemented can be filled back into the appropriate positions in the adapted admission record template to improve the specific diagnosis results of the disease recorded in the adapted admission record template of the current patient. Then, the final improved adapted admission record template is used as the electronic medical record of the current patient to ensure the comprehensive integrity of the electronic medical record.

[0049] The technical solution provided in the embodiment of the present application first determines the adaptive admission record template for the current patient based on the primary diagnosis information and department information of the current patient, so that the doctor can fill in the current patient's medical information in a standardized manner. Then, by analyzing the medical consultation filling data in the adaptive admission record template, the corresponding target record item questions consisting of questions about items to be revised and missing items are generated. The doctor's interactive response to each target record item question is obtained to generate the corresponding admission record supplementary draft, and the admission record supplementary draft is backfilled into the adaptive admission record template to obtain the current patient's electronic medical record, thereby completing the comprehensive quality control of the electronic medical record and avoiding omissions in the description of the patient's condition in the electronic medical record. The dual quality control of the electronic medical record is achieved by adapting the admission record template and the admission record supplementary draft, ensuring the comprehensiveness and standardization of the electronic medical record, enhancing the electronic medical record's all-round and accurate description of the patient's condition, and providing doctors with a more efficient, convenient and reliable standardized diagnosis and treatment reference.

[0050] As an optional implementation scheme in the present application, in order to ensure the comprehensiveness and standardization of electronic medical records, the present application may describe the specific process of generating an electronic medical record for any patient by performing double quality control on the electronic medical record by adapting the admission record template and the admission record supplementary draft.

[0051] Figure 2 FIG. 1 is a flow chart of another method for generating an electronic medical record according to an embodiment of the present application. Figure 2 As shown, the method may include the following steps:

[0052] S210, input the main diagnosis information and department information of the current patient into the pre-built template calling model to obtain the adapted admission record template of the current patient. The template calling model is configured with multiple admission record templates for different diseases in different departments.

[0053] When the current patient visits the hospital, he can choose the corresponding doctor under the corresponding department to see the doctor, who will diagnose and analyze the current patient's various medical examination results to preliminarily diagnose the current patient's specific condition, thereby obtaining the current patient's main diagnosis information. Moreover, according to the department selected by the current patient during this visit, the current patient's department information can be determined.

[0054] Moreover, in order to efficiently and accurately match a suitable admission record template for any patient, the present application can pre-train a template call model, which can be used to call the most suitable admission record template for any patient by analyzing the primary diagnosis information and department information of any patient. It is understandable that in order to ensure the accuracy of the template call model, the template call model can be pre-configured with multiple admission record templates for different diseases in different departments, so as to understand in advance the specific recording requirements of each medical institution for the patient's condition diagnosis information in different departments and different diseases, so as to realize the accurate and comprehensive recording of the patient's condition diagnosis in the electronic medical record.

[0055] It should be noted that the template call model in this application can use a long short-term memory neural network (Long-Short Term Memory, referred to as LSTM). The LSTM network can be composed of a forget gate, an input gate, and an output gate. Among them, the forget gate is used to decide which information in the training data is discarded, the input gate is used to decide which new training data can be stored, and the output gate is used to decide which information is output.

[0056] Therefore, this application can use the LSTM network to perform corresponding deep learning on the historical diagnosis information, department information, relevant medical record data and historical admission record templates used by a large number of patients in their historical medical treatment processes to build a corresponding template calling model.

[0057] Next, this application can explain in detail the specific training process of the template calling model:

[0058] In the first step, this application can collect admission record templates for different diseases in different departments from various departments under any medical institution, as well as historical diagnosis information, department information, and complete medical record data of a large number of patients in the historical medical treatment process. Then, in order to realize the accurate calling of the adapted admission record template for any patient by the template calling model, this application can configure multiple admission record templates for different diseases in different departments in the template calling model. Moreover, the historical diagnosis information and department information of a large number of patients in the historical medical treatment process can be used as corresponding training samples, and the historical admission record templates actually used by patients in the historical medical treatment process and the complete medical record data finally generated can be used as corresponding sample labels to form the training data of the template calling model.

[0059] In order to ensure accurate training of the template calling model, the present application may also perform corresponding data processing operations such as data cleaning, preprocessing and digital identification on the above training data to ensure the reliability and training practicality of the training data used by the template calling model.

[0060] In the second step, this application can define an optimal loss function for the template call model by analyzing the content matching loss and structure loss of the template call model. The loss function can be:

[0061] L=αL similarity +βL feature +γL structure +δL order +εL regularization

[0062] Among them, L similarity It can represent the similarity loss between the model output and the sample label; L feature It can represent the feature loss between model output and sample label; L structure Can represent structural loss; L order Can represent sequence order loss; L regularization Can indicate loss of regularity.

[0063] The third step is to randomly initialize the relevant parameters of the LSTM network and the attention mechanism, and continuously input the preprocessed training data into the LSTM network. The model output of each training data is calculated through the forward propagation algorithm, and it is substituted into the loss function with the sample label to calculate the corresponding training loss. Then, the gradient information of the training loss to the model parameters can be calculated through the back-propagation algorithm, so as to update the corresponding parameters and network structure of the LSTM network to continuously reduce the training loss.

[0064] The above steps can be repeated for a large amount of training data until the LSTM network reaches convergence, and a trained template call model can be obtained.

[0065] In the fourth step, it is understood that in order to ensure accurate training of the template call model, the training data can be divided into a training set and a validation set. The training data in the training set is used to perform the above model training steps, and the training data in the validation set is used to verify the performance of the trained template call model.

[0066] Therefore, the present application can evaluate the performance of the trained template call model on the above-mentioned validation set. The specific dimensions of the performance parameters to be evaluated may include accuracy, recall, and F1 value. Then, according to the specific evaluation results of the template call model, the hyperparameters of the template call model can be adjusted, which may include the number of LSTM layers, the number of hidden units, the parameters of the attention mechanism, etc., to further improve the performance of the template call model.

[0067] In the fifth step, the template call model trained for any medical institution can be deployed to the medical system of the medical institution to realize intelligent analysis of any patient diagnosis and department information. In practical applications, new training data can be continuously collected to update and optimize the template call model online to improve the adaptability and performance of the template call model.

[0068] It should be noted that, since different medical institutions have certain differences in the admission record templates for the same department or the same disease, the template calling model is required to be able to adapt to the admission record templates of different medical institutions. However, in order to adapt the template calling model between different medical institutions, it is often necessary to re-collect training data in the corresponding medical institutions to retrain the template calling model, which is costly, time-consuming, and difficult to quickly respond to the changing needs of new medical institutions. Therefore, in order to solve the above problems, the present application can perform local configuration adjustments or incremental learning on the template calling model according to the type of change in the admission record template in the template calling model to obtain the latest template calling model.

[0069] That is, because the template format or filling requirements of the admission record templates under different medical institutions may change, or some information fields may be added to the admission record template or a new admission record template may be added. In the above two cases, different update operations can be performed on the template call model.

[0070] Specifically, the present application can compare each admission record template configured in the trained template call model with each admission record template under the medical institution that currently applies the template call model to determine what changes exist in each admission record template under the medical institution that currently applies the template call model relative to each admission record template configured in the template call model, thereby determining the type of admission record template change in the template call model. Among them, the admission record template change type can include two types: related template configuration changes represented by template format or filling requirements and template content changes represented by newly added fields or newly added templates.

[0071] In one case, if the change type of the admission record template in the template calling model belongs to the template configuration change, a corresponding template configuration change parameter can be set in the configuration file of the template calling model according to the specific configuration change in the admission record template to obtain the latest template calling model, so that the latest template calling model can be actually applied in the current medical institution. When it outputs the changed admission record template for any patient, the above model update method does not require modification of the core algorithm of the template calling model, and the cost is relatively low.

[0072] In some implementable methods, if a certain admission record template has format changes, such as font, font size, layout, etc. Then, the present application can set a corresponding parameter for changes in font, font size, layout, etc. in the configuration file of the template call model to achieve rapid adjustment of the template call model. For example, when a new medical institution requires the font of the admission record template to be changed from Songti to Heiti, it is only necessary to modify the font parameter of the admission record template to "Heiti" in the configuration file of the template call model, and the new font can be automatically applied when outputting the adapted admission record template for any patient, thereby achieving the update of the template call model.

[0073] Alternatively, if there are changes in mandatory and optional items in a hospital admission record template, data validation logic for changes in mandatory and optional items can be added in the input phase of the template calling model to ensure that user input meets the new requirements. For example, when a medical institution that newly applies the template calling model changes the "allergy history" in the hospital admission record template from an optional item to a mandatory item, a mandatory prompt for "allergy history" can be added to the input interface, and the user input can be verified. In the output phase, the format of "allergy history" can be adjusted from an optional item to a mandatory item in the adapted admission record template output for any patient, that is, the mandatory item "allergy history" can be highlighted so that users can fill it out accurately.

[0074] In another case, if the change type of the admission record template in the template call model belongs to the template content change, the present application can collect a large amount of new training data containing the newly added field information according to the changed admission record template. Then, the newly collected new training data is processed to ensure that the data quality meets the model input standard. Furthermore, the pre-processed new training data is input into the deployed template call model, so that the template call model can use the incremental learning algorithm to conduct targeted learning only for the newly added fields and their related features, avoiding excessive interference with existing knowledge.

[0075] Specifically, the template calling model can automatically adjust its internal parameters and generation logic to adapt to the requirements of the new fields by analyzing the characteristics of the new fields in the new training data. When the template calling model generates a sample admission record template containing new fields during the training process, it can be output to the maintenance personnel for preliminary evaluation. If the admission record template generated by the template calling model meets the template content change requirements, the maintenance personnel can give the model positive feedback to confirm that the learning results are valid. If the admission record template generated by the template calling model does not meet the requirements (such as incorrect time format, missing information, etc.), the maintenance personnel can provide detailed negative feedback and supplement the specific requirements or sample data so that the template calling model can be further adjusted and optimized accordingly.

[0076] Therefore, based on the training feedback of the template call model by the maintenance personnel, the template call model can be supported to perform multiple iterative learning until the output admission record template fully meets the content change requirements of the current medical institution for the admission record template. In the incremental learning process of the template call model, the template call model not only learns the knowledge of the newly added fields, but also optimizes the processing logic of the existing fields through continuous feedback adjustment, thus achieving an overall improvement of the knowledge system.

[0077] In summary, after obtaining the current patient's primary diagnosis information and department information, the present application can input the primary diagnosis information and department information into the trained template calling model, and the template calling model will conduct a comprehensive analysis of the type of disease and department information of the current patient's disease, to find out the admission record template that best matches the type of disease and department information of the current patient's disease from multiple admission record templates pre-set for different types of diseases in different departments, and output it, so as to obtain the adapted admission record template for the current patient.

[0078] S220, input the consultation data in the adapted admission record template into the pre-built medical record interaction model, identify the corresponding key symptom entities and the consultation description segmentations under each key symptom entity, and the medical record interaction model is configured with question-answer pairs constructed for any symptom entity.

[0079] After the template calling model outputs the current patient's adapted admission record template, this application can directly display the adapted admission record template to the current patient's attending physician, who will fill in the specific diagnosis of the current patient's condition after seeing the patient at the corresponding position in the adapted admission record template to complete the complete record of the current patient's visit. Thus, during the process of filling in the adapted admission record template after the visit, this application can continuously obtain the specific filled-in content in the adapted admission record template as the visit filling-in data in this application.

[0080] Considering that the consultation data filled in the adapted admission record template is the subjective description of the patient's disease diagnosis results after the doctor's visit to the current patient, the doctor may miss asking, miss writing, or write the patient's disease incorrectly, so that the adapted admission record template does not fully record the patient's disease diagnosis and treatment information or the recorded content is incorrect. Therefore, this application needs to further analyze the consultation data filled in the adapted admission record template to determine whether there is content that needs to be supplemented and corrected in the adapted admission record template by initiating multiple questions about pending revisions or missing items.

[0081] Then, in order to ensure the comprehensive integrity of electronic medical records, the present application can pre-build a medical record interaction model, which is configured with question-answer pairs constructed for any disease entity, used to describe the various disease diagnosis description information involved in the entire disease diagnosis process under the corresponding disease type.

[0082] Among them, the symptom entity can refer to the multiple dimensions of describing the patient's symptoms under any disease type. For example, the symptom entity can include but is not limited to the patient's name, chief complaint, current medical history, past medical history, etc. For each symptom entity under any disease type, this application can determine the various descriptive words used to describe the patient's condition under the symptom entity by analyzing the historical medical record data of a large number of patients under the disease type. Then, based on the semantic information represented by the various descriptive words under each symptom entity, questions with various descriptive words as answers are generated, thereby constructing a corresponding question-answer pair for the symptom entity.

[0083] Therefore, by configuring question-answer pairs built for any disease entity in the medical record interaction model, the medical record interaction model can fully understand the complete description content that needs to be recorded for the patient's disease diagnosis results under any disease entity, so as to analyze whether there are omissions or errors in the consultation data filled in the current patient's adapted admission record template.

[0084] Specifically, the case interaction model can be trained through the following steps:

[0085] The first step is to collect a large number of admission record templates for different diseases in different departments from various medical institutions or a large amount of medical literature, as well as a large number of original texts such as historical diagnosis information, department information and complete medical record data of patients in their historical medical treatment process.

[0086] Then, by performing word segmentation on the original text, the continuous text is split into corresponding word sequences, which are the basic units that the model can understand and process text. Then, according to the specific description meaning of each word under each disease type, each word can be annotated to the corresponding disease entity, so as to obtain each word under each disease entity for describing the corresponding disease diagnosis information.

[0087] It is understandable that, by performing the above processing on a large amount of original text, it is possible to ensure that more and more word segments that more comprehensively and completely describe the corresponding disease diagnosis information can be marked under each disease entity.

[0088] For example, when a patient suffers from acute gastroenteritis, when a symptom entity is the main complaint, the various participles marked under the symptom entity may include yellow watery stool, vomiting, 1 day, etc.; when a symptom entity is the current medical history, the various participles marked under the symptom entity may include no obvious cause, diarrhea, yellow watery, stool, etc.

[0089] Then, after the symptom entities are labeled, in order to support the medical record interaction model to conduct question-and-answer interactions with the attending doctors to further supplement the record content of any patient's visit, the present application can support the medical record interaction model to segment the individual participles labeled under each of the above symptom entities to construct a question for the symptom entity with the corresponding participles labeled under the symptom entity as the answer, and thus obtain a question-answer pair constructed for each symptom entity.

[0090] For example, for each word segment marked under each disease entity mentioned in the above example, the following question-answer pairs can be constructed: Question 1: "What symptoms does the patient have?", Answer 1: "Yellow watery stool, vomiting"; Question 2: "Urination in the medical history?", Answer 2: "Normal urination"; and so on.

[0091] Next, data cleaning can be performed on each of the word segments marked under each of the processed disease entities and the constructed question-answer pairs to remove possible noise and redundant information to ensure the quality of the training data.

[0092] In the second step, for each word segmentation marked under each disease entity after the above data preprocessing and the constructed question-answer pairs, the present application can introduce a sparse attention (denoted as sparse attention) module in a sparse transformation (denoted as SparseTransform) model on the basis of a generative language model to semantically process each word segmentation marked under each disease entity after the above processing and the constructed question-answer pairs again, so that the medical record interaction model can learn the semantic and grammatical information represented by each word segmentation marked under each disease entity and the constructed question-answer pairs, and use specific tags to separate the question and answer in the question-answer pair constructed under each disease entity, so as to support the medical record interaction model to generate standardized answers to any input question.

[0093] Specifically, the sparse attention module can set the attention of all tokens (denoted as tokens) marked under each disease entity to 0, except for the tokens (denoted as tokens) with a relative distance not exceeding k and a relative distance of k, 2k, 3k...

[0094] Each segmentation marked under each disease entity is regarded as a corpus information, recorded as:

[0095] coppus=(token1,token2,…,token i ,…,token n )

[0096] Among them, token i is the ith word in the above corpus, and n is the number of word tokens in the above corpus. Then, the objective function of the above generative language model can be:

[0097]

[0098] Among them, θ is the network parameter of the generative language model. The objective function is to predict the current segmentation token based on the past K segmentation tokens.

[0099] The calculation process of the above generative language model can be as follows:

[0100] h0=UW embedding +W position_encoding

[0101] h z = transformer_block(h z-1 )

[0102]

[0103] In the above formula, z is the serial number of the sparse transformation (denoted as Sparse Transform) model, and u is the predicted target token i-K+1 to token i The unique hot encoding of k word segmentation tokens. U=(u i-K ,…,u i-1 ) is token i The K word segmentation tokens are all hot-encoded sequences, and this input and output setting is used to complete the operation of the typical autoregressive formal language model.

[0104] W embedding is a matrix of word segmentation token embedding vectors, which is paired with one-hot encoding (the position marked with "1" in the one-hot encoding of a word segmentation token is the embedding vector of the word segmentation token in W embedding Therefore, UW embedding The result can be a sequence of embedding vectors of k word segmentation tokens.

[0105] W position_encoding It is the positional encoding of the input data. The dimension of the p(u) vector is the same as the one-hot encoding length of the token.

[0106] By performing the above calculations on each word segment marked under each disease entity, the above generative language model can be trained to obtain a trained medical record interaction model, which can be configured with question-answer pairs constructed for any disease entity, and the question and answer in the question-answer pair constructed under the disease entity can be separated by specific tags to support the medical record interaction model to generate standardized answers to any input question.

[0107] The third step is to use the verification text divided from the collected original text to annotate the following corpus:

[0108] {(text1,y1),(text2,y2),…,(text m ,y m ),…,(text M ,y M )}

[0109] Among them, text m is the mth verification text, y m is the class label of the mth verification text.

[0110] The objective function of the classification task is:

[0111] The individual word segments marked under each disease entity in the above verification text and the question-answer pairs constructed under the disease entity are input into the medical record interaction model for fine-tuning to ensure the accuracy of the medical record interaction model.

[0112] In summary, after obtaining the consultation data filled in the current patient's adapted admission record template, the present application can directly input the consultation data into the above-mentioned trained medical record interaction model. The medical record interaction model can perform corresponding word segmentation processing on each text content in the input consultation data, and obtain multiple word segments for describing the current patient's corresponding disease diagnosis information, which are used as the consultation description word segments in the present application.

[0113] Then, through the medical record interaction model, semantic analysis can be performed on each consultation description segmentation to determine the specific description meaning of each consultation description segmentation, so as to judge from which dimension each consultation description segmentation describes the specific diagnosis of the current patient, and identify each key disease entity represented by the consultation filling data in the adapted admission record template. Moreover, by analyzing from which dimension each consultation description segmentation describes the specific diagnosis of the current patient, the consultation description segmentation under each key disease entity can be determined.

[0114] S230, generating corresponding target record item questions according to the key question answer pairs constructed for the key symptom entities in the medical record interaction model and the consultation description segmentations under the key symptom entities.

[0115] For the consultation data filled in the adapted admission record template, after the corresponding key disease entity and the consultation description segmentation under each key disease entity are identified through the medical record interaction model, the present application can also perform corresponding semantic analysis on each consultation description segmentation under each key disease entity through the medical record interaction model to generate corresponding questions for each key disease entity, so as to determine the key question answer pair constructed for the key disease entity in the medical record interaction model, and the question in the key question answer pair can be the question generated for the key disease entity through the medical record interaction model. Then, by analyzing the answer in the key question answer pair, a corresponding standardized answer can be generated for the question, and the standardized answer can be the individual segmentations that have been annotated more, more comprehensively, and completely describe the corresponding disease diagnosis information under the key disease entity during model training. Thus, in the same manner as above, the individual segmentations that have been annotated more, more comprehensively, and completely describe the corresponding disease diagnosis information under each key disease entity can be obtained.

[0116] Then, since the marked segmentation under each key disease entity can describe the corresponding disease diagnosis information more comprehensively and completely. And the consultation description segmentation under each key disease entity can represent the recorded content of the corresponding disease diagnosis information for the current patient in the adapted admission record template. Then, for each key disease entity, the present application can compare and analyze the various consultation description segmentations under the key disease entity and the various marked segmentations under the key disease entity through the medical record interaction model to determine whether there is any missing or wrong record content under the key disease entity. Then, for the wrong record content under the key disease entity, the present application can perform semantic segmentation on it to generate the corresponding question of the item to be revised, so as to prompt the doctor whether he needs to correct the wrong record content represented by the item to be revised. And for the missing record content under the key disease entity, the present application can perform semantic segmentation on it to generate the corresponding missing item question, so as to prompt the doctor whether he needs to add a record of the missing record content represented by the missing item in the adapted admission record template.

[0117] Therefore, by combining the issues to be revised and missing items under each of the above-mentioned key disease entities, corresponding target record item questions can be generated, which is convenient for subsequent asking of each target record item question to the current patient's attending physician, and the attending physician will answer, so as to continuously improve the current patient's electronic medical record.

[0118] S240, presenting a target record item question at a location associated with the adapted admission record template to obtain an interactive reply to the target record item question.

[0119] In order to ensure convenient and intelligent interaction of the seeing doctor with respect to the target record item questions, the present application can directly determine an associated position at the display position of the adapted admission record template when detecting the seeing doctor's filling-in operation on the adapted admission record template, as the presentation position of the target record item questions, so that the seeing doctor can answer the various target record item questions while checking the seeing doctor's filling-in content in the adapted admission record template.

[0120] Therefore, the present application can present each target record item question to the current patient's attending physician in turn by displaying an interactive window at the associated position of the adapted admission record template, and the attending physician can answer each target record item question, thereby determining the current patient's specific revision status under the erroneous record content and the specific diagnosis status under each missing item, thereby obtaining the attending physician's specific reply content to each target record item question as an interactive reply to the target record item question.

[0121] In some achievable methods, different target record item questions may have different answering methods, such as single-choice answer, multiple-choice answer, text answer, etc. Then, in order to ensure the efficient and convenient interaction of the attending physician with the target record item questions, the specific presentation effect of the target record item questions can be achieved in the following ways: according to the type of the target record item question, the corresponding interactive presentation style is determined; at the associated position of the adapted admission record template, the target record item question is presented according to the interactive presentation style.

[0122] Specifically, for each target record item question, the present application can first determine whether the most suitable answer method for the target record item question is a single-choice answer, a multiple-choice answer, a text answer, etc., to determine the type of the target record item question, and thereby determine a most suitable interactive presentation style for the target record item question under this type. Under this interactive presentation style, it can support the attending doctor to conveniently answer the target record item question through simple operations, thereby simplifying the attending doctor's interactive operations on the target record item question.

[0123] Then, an interactive window can be displayed at the associated position of the adapted admission record template, and in the interactive window, each target record item question can be presented to the current patient's attending physician in turn according to the interactive presentation style of each target record item question, and the attending physician will answer each target record item question.

[0124] For example, Figure 3As shown, after detecting the doctor's filling operation on the adapted admission record template or the doctor's triggering operation on the interactive trigger control in the side area of ​​the adapted admission record template, the present application can display an interactive window to the doctor of the current patient at the associated position of the adapted admission record template of the current patient. Then, in the interactive window, for the target record item question of the single-choice type, the corresponding single-choice option can be presented in association with the presentation of the target record item question to support the doctor to answer the target record item question by triggering a certain single-choice option. For the target record item question of the multiple-choice type, the corresponding multiple-choice option can be presented in association with the presentation of the target record item question. The multiple-choice option may include a supplementary option that supports the doctor to supplement the answer to the target record item question based on the existing answer, so that the doctor can immediately correct and edit the multiple answers to the target record item question that have been provided by the medical record interaction model.

[0125] S250, generating a corresponding admission record supplementary draft according to the target record item question and the interactive reply to the target record item question.

[0126] After obtaining the interactive responses of the attending physician to the questions of each target record item, this application can integrate the questions of each target record item and the interactive responses to each target record item to generate a corresponding supplementary draft of the admission record, which is used to describe the actual diagnosis of the attending physician on the relevant medical conditions of the current patient that were not recorded or recorded incorrectly due to missed questions or missed notes after this visit. It can be seen that by intelligently interacting with the attending physician of the current patient through each target record item question, the consultation data in the adapted admission record template of the current patient can be added, revised, edited, and other record improvement operations can be performed to ensure the comprehensive integrity of the electronic medical record.

[0127] S260, when a confirmation operation for the supplementary draft of the admission record is detected, the corresponding target record item content is extracted from the supplementary draft of the admission record, and a backfill position of the target record item content in the adapted admission record template is determined.

[0128] After obtaining the supplementary draft of the admission record, in order to ensure the accuracy of the electronic medical record, the application can also present the supplementary draft of the admission record to the attending physician, who will review the interactive responses to the various target record item questions in the supplementary draft of the admission record again to determine whether the responses to a certain target record item question are incorrect. If the supplementary draft of the admission record is incorrect, the attending physician can be prompted to modify the incorrect content to regenerate a supplementary draft of the admission record.

[0129] When the draft of the admission record supplement is correct, the present application can detect the doctor's confirmation operation on the draft of the admission record supplement. Then, the present application can parse the draft of the admission record supplement to determine multiple target question-answer pairs consisting of each target record item question and its interactive reply. Then, by analyzing the relevant disease diagnosis information represented by the target question-answer pairs, it is possible to determine the content that the doctor needs to modify or add in the current patient's adapted admission record template during the interaction with each target record item question, and use this as the target record item content extracted in the present application.

[0130] Furthermore, by analyzing the specific diagnosis type of each target record item content, the content correlation degree between it and each text content in the consultation filling data in the adapted admission record template can be determined. Then, by determining the position of a text content with the highest content correlation degree with each target record item content in the adapted admission record template, the appropriate backfill position of the target record item content in the adapted admission record template can be determined.

[0131] S270, according to the backfill position, merge the target record item content into the adapted admission record template to obtain the electronic medical record of the current patient.

[0132] By judging whether the content of each target record item is newly added record content or replacement content for erroneous content that has been recorded, each target record item content can be accurately inserted into the adaptive admission record template according to the backfill position of each target record item content in the adaptive admission record template, or the erroneous content that has been recorded in the adaptive admission record template can be replaced to improve the specific disease diagnosis results recorded in the adaptive admission record template for the current patient, so as to obtain the electronic medical record of the current patient and ensure the comprehensive integrity of the electronic medical record.

[0133] The technical solution provided in the embodiment of the present application first determines the adaptive admission record template for the current patient based on the primary diagnosis information and department information of the current patient, so that the doctor can fill in the current patient's medical information in a standardized manner. Then, by analyzing the medical consultation filling data in the adaptive admission record template, the corresponding target record item questions consisting of questions about items to be revised and missing items are generated. The doctor's interactive response to each target record item question is obtained to generate the corresponding admission record supplementary draft, and the admission record supplementary draft is backfilled into the adaptive admission record template to obtain the current patient's electronic medical record, thereby completing the comprehensive quality control of the electronic medical record and avoiding omissions in the description of the patient's condition in the electronic medical record. The dual quality control of the electronic medical record is achieved by adapting the admission record template and the admission record supplementary draft, ensuring the comprehensiveness and standardization of the electronic medical record, enhancing the electronic medical record's all-round and accurate description of the patient's condition, and providing doctors with a more efficient, convenient and reliable standardized diagnosis and treatment reference.

[0134] Figure 4 This is a block diagram of a device for generating an electronic medical record according to an embodiment of the present application. Figure 4 As shown, the apparatus 400 may include:

[0135] The adaptation template determination module 410 is used to determine the adaptation admission record template of the current patient according to the primary diagnosis information and department information of the current patient;

[0136] A question generating module 420 is used to generate corresponding target record item questions according to the consultation filling data in the adapted admission record template, wherein the target record item questions include questions about items to be revised and questions about missing items;

[0137] An admission record supplementation module 430 is used to generate a corresponding admission record supplementation draft according to the interactive responses to the target record item questions;

[0138] The electronic medical record generation module 440 is used to fill the admission record supplementary draft back into the adapted admission record template to obtain the electronic medical record of the current patient.

[0139] In some implementations, the adaptation template determination module 410 may be specifically configured to:

[0140] Inputting the primary diagnosis information and department information of the current patient into the pre-built template calling model to obtain an adapted admission record template for the current patient;

[0141] Among them, the template calling model is configured with multiple admission record templates for different diseases in different departments.

[0142] In some implementations, the electronic medical record generation device 400 may further include a model updating module. The model updating module may be used to:

[0143] According to the change type of the admission record template in the template calling model, local configuration adjustment or incremental learning is performed on the template calling model to obtain the latest template calling model.

[0144] In some implementations, the question generation module 420 may be specifically configured to:

[0145] Input the consultation data in the adapted admission record template into a pre-built medical record interaction model, identify the corresponding key symptom entity and the consultation description segmentation under each key symptom entity, and configure the medical record interaction model with a question-answer pair constructed for any symptom entity;

[0146] According to the key question answer pairs constructed for the key symptom entity in the medical record interaction model and the consultation description segmentations under the key symptom entity, corresponding target record item questions are generated.

[0147] In some implementations, the admission record supplement module 430 may include:

[0148] A question presenting unit, configured to present the target record item question at a correlation position of the adapted admission record template to obtain an interactive reply to the target record item question;

[0149] The supplementary draft generating unit is used to generate a corresponding supplementary draft of the admission record according to the target record item question and the interactive reply to the target record item question.

[0150] In some implementations, the question presentation unit may be specifically configured to:

[0151] Determine a corresponding interactive presentation style according to the type of the target record item question;

[0152] At the associated position of the adapted admission record template, the target record item question is presented according to the interactive presentation style.

[0153] In some implementations, the electronic medical record generation module 440 may be specifically used to:

[0154] When a confirmation operation for the admission record supplementary draft is detected, the corresponding target record item content is extracted from the admission record supplementary draft, and a backfill position of the target record item content in the adapted admission record template is determined;

[0155] According to the backfill position, the target record item content is merged into the adapted admission record template to obtain the electronic medical record of the current patient.

[0156] In the embodiment of the present application, first, the adaptive admission record template of the current patient is determined based on the main diagnosis information and department information of the current patient, so that the doctor can fill in the current patient's medical information in a standardized manner. Then, by analyzing the medical consultation filling data in the adaptive admission record template, the corresponding target record item questions consisting of questions about items to be revised and missing items are generated. The doctor's interactive response to each target record item question is obtained to generate the corresponding admission record supplementary draft, and the admission record supplementary draft is filled back into the adaptive admission record template to obtain the current patient's electronic medical record, thereby completing the comprehensive quality control of the electronic medical record and avoiding omissions in the description of the patient's condition in the electronic medical record. The dual quality control of the electronic medical record is achieved by adapting the admission record template and the admission record supplementary draft, ensuring the comprehensiveness and standardization of the electronic medical record, enhancing the electronic medical record's all-round and accurate description of the patient's condition, and providing doctors with a more efficient, convenient and reliable standardized diagnosis and treatment reference.

[0157] It should be understood that the device embodiment and the method embodiment may correspond to each other, and similar descriptions may refer to the method embodiment. To avoid repetition, they will not be described here. Specifically, Figure 4 The device 400 shown can execute any method embodiment in the present application, and the aforementioned and other operations and / or functions of each module in the device 400 are respectively for implementing the corresponding processes in each method in the embodiment of the present application. For the sake of brevity, they will not be repeated here.

[0158] The above describes the device 400 of the embodiment of the present application from the perspective of the functional module in conjunction with the accompanying drawings. It should be understood that the functional module can be implemented in hardware form, can be implemented by instructions in software form, and can also be implemented by a combination of hardware and software modules. Specifically, the steps of the method embodiment in the embodiment of the present application can be completed by the hardware integrated logic circuit and / or software form instructions in the processor, and the steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to perform, or a combination of hardware and software modules in the decoding processor to perform. Optionally, the software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, an electrically erasable programmable memory, a register, etc. The storage medium is located in a memory, and the processor reads the information in the memory, and completes the steps in the above method embodiment in conjunction with its hardware.

[0159] Figure 5 A schematic block diagram of an electronic device according to an embodiment of the present application.

[0160] like Figure 5 As shown, the electronic device 500 may include:

[0161] The memory 510 and the processor 520, the memory 510 is used to store the computer program and transmit the program code to the processor 520. In other words, the processor 520 can call and run the computer program from the memory 510 to implement the method in the embodiment of the present application.

[0162] For example, the processor 520 may be configured to execute the above method embodiments according to instructions in the computer program.

[0163] In some embodiments of the present application, the processor 520 may include but is not limited to:

[0164] General-purpose processor, digital signal processor (DSP), application-specific integrated circuit (ASIC), field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, etc.

[0165] In some embodiments of the present application, the memory 510 includes but is not limited to:

[0166] Volatile memory and / or non-volatile memory. Among them, the non-volatile memory can be read-only memory (ROM), programmable ROM (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM) or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct RAM bus random access memory (DR RAM).

[0167] In some embodiments of the present application, the computer program may be divided into one or more modules, which are stored in the memory 510 and executed by the processor 520 to complete the method provided by the present application. The one or more modules may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program in the electronic device.

[0168] like Figure 5 As shown, the electronic device may also include:

[0169] The transceiver 530 may be connected to the processor 520 or the memory 510 .

[0170] The processor 520 may control the transceiver 530 to communicate with other devices, specifically, to send information or data to other devices, or to receive information or data sent by other devices. The transceiver 530 may include a transmitter and a receiver. The transceiver 530 may further include an antenna, and the number of antennas may be one or more.

[0171] It should be understood that the various components in the electronic device are connected via a bus system, wherein the bus system includes not only a data bus but also a power bus, a control bus and a status signal bus.

[0172] The present application also provides a computer storage medium on which a computer program is stored, and when the computer program is executed by a computer, the computer can perform the method of the above method embodiment. In other words, the present application embodiment also provides a computer program product containing instructions, and when the instructions are executed by a computer, the computer can perform the method of the above method embodiment.

[0173] When software is used for implementation, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (digital subscriber line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integration. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a digital video disc (digital video disc, DVD)), or a semiconductor medium (e.g., a solid state drive (solid state disk, SSD)), etc.

[0174] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0175] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the module is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0176] The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. For example, each functional module in each embodiment of the present application may be integrated into a processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0177] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for generating an electronic medical record, characterized in that: include: Determine an adapted admission record template for the current patient based on the primary diagnosis information and department information of the current patient; Generate corresponding target record item questions according to the consultation filling data in the adapted admission record template, wherein the target record item questions include questions about items to be revised and questions about missing items; Generate a corresponding supplementary draft of the admission record according to the interactive responses to the target record item questions; The supplementary draft of the admission record is filled back into the adapted admission record template to obtain the electronic medical record of the current patient.

2. The method according to claim 1, characterized in that Determining the adapted admission record template for the current patient according to the primary diagnosis information and department information of the current patient includes: Inputting the primary diagnosis information and department information of the current patient into the pre-built template calling model to obtain an adapted admission record template for the current patient; Among them, the template calling model is configured with multiple admission record templates for different diseases in different departments.

3. The method according to claim 2, characterized in that The method further comprises: According to the change type of the admission record template in the template calling model, local configuration adjustment or incremental learning is performed on the template calling model to obtain the latest template calling model.

4. The method according to claim 1, characterized in that: The generating of corresponding target record item questions according to the consultation filling data in the adapted admission record template includes: Input the consultation data in the adapted admission record template into a pre-built medical record interaction model, identify the corresponding key symptom entity and the consultation description segmentation under each key symptom entity, and configure the medical record interaction model with a question-answer pair constructed for any symptom entity; According to the key question answer pairs constructed for the key symptom entity in the medical record interaction model and the consultation description segmentations under the key symptom entity, corresponding target record item questions are generated.

5. The method according to claim 1, characterized in that The generating of a corresponding supplementary draft of the admission record according to the interactive response to the question of the target record item comprises: Presenting the target record item question at a location associated with the adapted admission record template to obtain an interactive reply to the target record item question; A corresponding supplementary draft of the admission record is generated based on the target record item question and the interactive reply to the target record item question.

6. The method according to claim 5, characterized in that The target record item question is presented at the associated position of the adapted admission record template, including: Determine a corresponding interactive presentation style according to the type of the target record item question; At the associated position of the adapted admission record template, the target record item question is presented according to the interactive presentation style.

7. The method according to claim 1, characterized in that Fill the admission record supplementary draft back into the adapted admission record template to obtain the electronic medical record of the current patient, including: When a confirmation operation for the admission record supplementary draft is detected, the corresponding target record item content is extracted from the admission record supplementary draft, and a backfill position of the target record item content in the adapted admission record template is determined; According to the backfill position, the target record item content is merged into the adapted admission record template to obtain the electronic medical record of the current patient.

8. A device for generating an electronic medical record, characterized in that: include: An adaptation template determination module, used to determine an adaptation admission record template for the current patient according to the primary diagnosis information and department information of the current patient; A question generation module, used to generate corresponding target record item questions according to the consultation filling data in the adapted admission record template, wherein the target record item questions include questions about items to be revised and questions about missing items; An admission record supplementation module, used for generating a corresponding admission record supplementation draft according to the interactive responses to the target record item questions; The electronic medical record generation module is used to fill the admission record supplementary draft back into the adapted admission record template to obtain the electronic medical record of the current patient.

9. An electronic device, characterized in that: include: A processor and a memory, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to execute the method for generating an electronic medical record according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: Used to store a computer program, wherein the computer program enables a computer to execute the method for generating an electronic medical record according to any one of claims 1 to 7.

Images