Generation device

Abstract

To solve the difficulty in reducing the time and effort during summary generation.SOLUTION: A generation device comprises: an acquisition unit for acquiring medical data in an electronic medical chart and a summary of the medical data; and a prompt generation unit for generating a prompt on the basis of the medical data and the summary acquired by the acquisition unit, which is an indication for a trained model used for generating a summary corresponding to an update part of the medical data after some data is added.SELECTED DRAWING: Figure 14

Description

[Technical field]

[0001] The present invention relates to a generating device, a generating method, and a program. [Background technology]

[0002] There is a known technology called electronic medical records, which allows medical treatment progress to be recorded in an electronic system.

[0003] An example of such a technique is, for example, Patent Document 1. For example, Patent Document 1 describes an electronic medical record system linked with a related system that requires the output of patient information. According to Patent Document 1, the electronic medical record system includes a patient information storage unit in which patient information is stored as an electronic medical record for each patient. The electronic medical record system also accepts a selection from a user for a patient name among the patient information displayed on a terminal by the related system, acquires patient information corresponding to the patient whose selection was accepted from the patient information storage unit, and displays the medical record information of the patient as the acquired patient information. Patent Document 1 also discloses that the medical record information includes medical information such as records in the SOAP format ("S (subjective): subjective information", "O (objective): objective information", "A (assessment): evaluation", and "P (plan): plan (treatment)"). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2022-151362 A [Non-patent literature]

[0005] [Non-Patent Document 1] Yongchao Zhou et al. “Large Language Models Are Human-Level Prompt Engineers” [Retrieved June 12, 2020], Internet<URL:https: / / arxiv.org / abs / 2211.01910> Summary of the Invention [Problem to be solved by the invention]

[0006] An electronic medical record such as that described in Patent Document 1 can take a long time to load. Therefore, in order to address such issues, it is common to record a summary, which is a summary of the SOAP data in past medical information, in addition to the SOAP data for each consultation.

[0007] In many cases, such summaries are created by the doctor himself / herself. This makes it difficult to reduce the amount of work required to create a summary.

[0008] Therefore, an object of the present disclosure is to provide a generation device, a generation method, and a program that can solve the above-mentioned problems. [Means for solving the problem]

[0009] In order to achieve this objective, the generating device in the present disclosure comprises: An acquisition unit that acquires medical data and a summary of the medical data in the electronic medical record; a prompt generation unit that generates a prompt, which is an instruction for a trained model used when generating a summary corresponding to an updated portion of the medical data to which data has been added, based on the medical data and the summary acquired by the acquisition unit; The structure is as follows.

[0010] In addition, the production method in the present disclosure includes: An information processing device, Obtaining medical data in an electronic medical record and a summary of the medical data; A prompt is generated based on the acquired medical data and the summary, which is an instruction to the trained model to be used when generating a summary corresponding to the updated portion of the medical data to which data has been added. The structure is as follows.

[0011] In addition, the program in this disclosure is In the information processing device, Obtaining medical data in an electronic medical record and a summary of the medical data; A prompt is generated based on the acquired medical data and the summary, which is an instruction to the trained model to be used when generating a summary corresponding to the updated portion of the medical data to which data has been added. It is a program for implementing processing.

[0012] In addition, the generating device in the present disclosure is a target data acquisition unit that acquires medical data to which data has been added as target data for generating a summary; a summary generation unit that generates a summary of the medical data acquired by the data acquisition unit for generating ... have The structure is as follows. Effect of the Invention

[0013] According to the above-mentioned configurations, the effort required for creating a summary can be reduced. [Brief description of the drawings]

[0014] [Figure 1] 1 is a diagram showing an overview of a medical support system according to the present disclosure. [Diagram 2]1 is a block diagram showing an example of the configuration of a prompt generating device. [Diagram 3] FIG. 2 is a diagram showing an example of information included in electronic medical record information. [Figure 4] FIG. 11 is a diagram illustrating an example of information included in prompt information. [Diagram 5] 11 is a diagram illustrating an example of processing by a prompt generating unit; FIG. [Figure 6] 11 is a diagram illustrating an example of processing by a prompt generating unit; FIG. [Figure 7] FIG. 2 is a block diagram showing a configuration example of a summary generating device. [Figure 8] 11 is a diagram for explaining an example of processing by a summary generation unit; FIG. [Figure 9] 10 is a flowchart showing an example of the operation of the prompt generating device. [Figure 10] 13 is a flowchart showing an example of the operation of the summary generating device. [Figure 11] FIG. 13 is a diagram illustrating another example of the configuration of the medical support system. [Figure 12] FIG. 1 is a diagram illustrating an example of a prompt management device. [Figure 13] FIG. 2 is a diagram illustrating an example of a hardware configuration of a generating device according to the present disclosure. [Figure 14] FIG. 2 is a block diagram showing an example of the configuration of a generating device. [Figure 15] 13 is a flowchart illustrating an example of the operation of the generating device. [Figure 16] FIG. 2 is a block diagram showing an example configuration of a second generating device in the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] [First embodiment] A configuration example of the medical support system 100 in the present disclosure will be described with reference to FIG. 1 to FIG. 12. FIG. 1 is a diagram showing an overview of the medical support system 100. FIG. 2 is a block diagram showing a configuration example of the prompt generation device 200. FIG. 3 is a diagram showing an example of information included in the electronic medical record information 241. FIG. 4 is a diagram showing an example of the prompt information 242. FIGS. 5 and 6 are diagrams for explaining an example of the processing of the prompt generation unit 252. FIG. 7 is a block diagram showing an example of the configuration of the summary generation device 300. FIG. 8 is a diagram showing an example of the processing of the summary generation unit 353. FIG. 9 is a flowchart showing an example of the operation of the prompt generation device 200. FIG. 10 is a flowchart showing an example of the operation of the summary generation device 300. FIG. 11 is a diagram showing another configuration example of the medical support system 100. FIG. 12 is a diagram showing an example of the prompt management device 400. In this disclosure, the drawings may be associated with one or more embodiments.

[0016] In the first embodiment of the present disclosure, a medical support system 100 is described that generates a prompt, which is an instruction for a trained model, and generates a hash summary corresponding to an updated portion of SOAP (Subjective Objective Assessment Plan) data using the generated prompt. For example, the medical support system 100 generates a prompt used when generating a hash summary corresponding to an added portion of SOAP data, based on SOAP data and a hash summary that is a summary of the SOAP data. As an example, the medical support system 100 can generate a prompt using a technique such as that described in Non-Patent Document 1. In addition, the medical support system 100 can accept a doctor's confirmation of the generated prompt and corrections according to the confirmation results.

[0017] Furthermore, the medical support system 100 can generate a hash summary corresponding to an updated portion of the SOAP data by using the SOAP data to which data has been added according to medical treatment, etc., and the above-mentioned prompt. As an example, the medical support system 100 may generate a hash summary corresponding to an updated portion of the SOAP data by using a model that has been learned in advance.

[0018] Furthermore, the medical support system 100 can manage the generated prompts. The generated prompts are less likely to be personal information. Therefore, by making it possible to collectively manage and utilize the prompts generated by multiple devices, the generated prompts can be utilized efficiently.

[0019] In this disclosure, SOAP data refers to medical data written in SOAP format in electronic medical records, etc. In SOAP format medical data, medical records are written according to four items: subjective information (S), objective information (O), assessment (A), and treatment plan (P). As mentioned above, the general framework of SOAP data is set. However, the specific method of writing and the content may vary greatly depending on the facility, department, doctor, etc.

[0020] Moreover, the hash summary refers to a summary of the SOAP data generated based on the SOAP data. In the present disclosure, the hash summary is defined as a summary of the SOAP data, since it is a summary written in the upper part (commonly called the hash part) of the part of the electronic medical record where the SOAP is written. However, the hash summary may be a summary of the SOAP data written in any part other than the hash part exemplified above. For example, the hash summary is added each time the SOAP data is added. Therefore, it is considered that the hash summary can be a label for the SOAP data. In other words, the combination of the SOAP data and the hash summary can be regarded as the teacher data. For example, the generation of the hash summary corresponding to the updated part of the SOAP data described above may be realized by training a localized large-scale language model (LLM) or any other language model using the teacher data.

[0021] In addition, in the present disclosure, a prompt refers to an instruction to be input to a model used when generating the above-mentioned hash summary. The prompt may include an instruction, command, condition, constraint, etc. to be input to the model. As described above, a large-scale language model, etc. is used in the present disclosure. Therefore, the prompt may be an instruction using a language. In other words, the prompt may be a sentence using Japanese, English, etc.

[0022] Fig. 1 shows an example of the overall configuration of a medical support system 100. With reference to Fig. 1, the medical support system 100 includes, for example, a prompt generation device 200 and a summary generation device 300. As shown in Fig. 1, the prompt generation device 200 and the summary generation device 300 can be connected to each other via wired or wireless communication so as to be able to communicate with each other.

[0023] The configuration of the medical support system 100 is not limited to the above example. For example, the medical support system 100 may include a configuration other than the above example, such as including a plurality of electronic medical record terminals. The medical support system 100 may be composed of a single information processing device having a function as the prompt generation device 200 and a function as the summary generation device 300.

[0024] The prompt generation device 200 is an information processing device that generates a prompt based on SOAP data and a hash summary. Fig. 2 shows an example configuration of the prompt generation device 200. Referring to Fig. 2, the prompt generation device 200 has, as main components, for example, an operation input section 210, a screen display section 220, a communication I / F section 230, a storage section 240, and a calculation processing section 250.

[0025] 2 illustrates an example in which the functions of the prompt generation device 200 are realized using one information processing device. However, at least some of the functions of the prompt generation device 200 may be realized using multiple information processing devices, for example, on the cloud. Furthermore, the prompt generation device 200 may not include some of the components exemplified above, such as not having the operation input unit 210 or the screen display unit 220, or may have a configuration other than those exemplified above.

[0026] The operation input unit 210 is made up of operation input devices such as a keyboard, a mouse, etc. The operation input unit 210 detects operations by a doctor or the like who operates the prompt generation device 200 and outputs the operations to the arithmetic processing unit 250.

[0027] The screen display unit 220 is composed of a screen display device such as a liquid crystal display, an organic EL (electro-luminescence) display, etc. The screen display unit 220 can display various information stored in the storage unit 240 on the screen in response to an instruction from the arithmetic processing unit 250.

[0028] The communication I / F unit 230 is composed of a data communication circuit etc. The communication I / F unit 230 performs data communication with external devices such as an electronic medical record terminal and a summary generation device 300 connected via a communication line.

[0029] The storage unit 240 is a storage device such as a hard disk or a memory. The storage unit 240 stores processing information and a program 243 required for various processes in the arithmetic processing unit 250. The program 243 is read into the arithmetic processing unit 250 and executed to realize various processing units. The program 243 is read in advance from an external device or a recording medium via a data input / output function such as the communication I / F unit 230, and is stored in the storage unit 240. Main information stored in the storage unit 240 includes, for example, electronic medical record information 241, prompt information 242, and the like.

[0030] The electronic medical record information 241 includes SOAP data, a hash summary, etc. Various information included in the electronic medical record information 241 is acquired in advance using any means, such as accepting an input using the operation input unit 210, or the reception unit 251 accepting an input from an external device such as an electronic medical record terminal via the communication I / F unit 230, and is stored in the storage unit 240. Various information included in the electronic medical record information 241 may be updated in response to the reception unit 251 accepting an input from an external device such as an electronic medical record terminal via the communication I / F unit 230.

[0031] Fig. 3 shows an example of information included in the electronic medical record information 241. For example, as shown in Fig. 3, the electronic medical record information 241 includes SOAP data and a hash summary. As shown in Fig. 3, the SOAP data can include data corresponding to one or more medical examinations. As an example, data can be added to the SOAP data each time a medical examination is performed. Furthermore, the hash summary is a summary of the SOAP data. The hash summary can be added in response to the addition of SOAP data.

[0032] In the electronic medical record information 241, the SOAP data, the hash summary, etc. may be associated with attribute information indicating the attributes of the person who input the SOAP data or the hash summary. For example, the attribute information may indicate at least a part of the attributes of the doctor, the medical department, the facility, etc. The attribute information may indicate attributes other than those exemplified above.

[0033] The prompt information 242 includes a prompt, which is an instruction to be input to the model. The prompt information 242 is updated in response to, for example, the approval unit 255 (described later) receiving approval from a doctor or the like. The prompt information 242 may be updated in response to, for example, the prompt generation unit 252 generating a prompt.

[0034] FIG. 4 shows an example of the prompt information 242. As exemplified in FIG. 4, in the prompt information 242, for example, identification data and prompt data are associated with each other. Here, the identification data is information for identifying prompt data including one or more prompts. For example, the prompt generating unit 252 can generate a prompt for each arbitrary attribute, such as for each doctor, each medical department, or each facility. The identification data may be added for identifying the prompt generated for each attribute. For example, as described above, the prompt information 242 may include prompt data for each attribute.

[0035] The arithmetic processing unit 250 has an arithmetic device such as a CPU (Central Processing Unit) and its peripheral circuits. The arithmetic processing unit 250 reads and executes a program 243 from the storage unit 240, thereby implementing various processing units by making the above hardware and the program 243 work together. Major processing units implemented by the arithmetic processing unit 250 include, for example, a reception unit 251, a prompt generation unit 252, a display unit 253, a correction reception unit 254, an approval unit 255, a verification unit 256, and an output unit 257.

[0036] In addition, the arithmetic processing unit 250 may have a GPU (Graphics Processing Unit), a DSP (Digital Signal Processor), an MPU (Micro Processing Unit), an FPU (Floating point number Processing Unit), a PPU (Physics Processing Unit), a TPU (Tensor Processing Unit), a quantum processor, a microcontroller, or a combination of these, instead of the above-mentioned CPU.

[0037] The reception unit 251 (acquisition unit) receives SOAP data, hash summaries, and the like from an external device such as an electronic medical record terminal. At this time, the reception unit 251 may receive attribute information indicating the attributes of the person who input the SOAP data or hash summary from the external device such as an electronic medical record terminal, along with the SOAP data or hash summaries. The reception unit 251 also stores the received SOAP data, hash summaries, and the like in the storage unit 240 as electronic medical record information 241. The reception unit 251 may receive the SOAP data, hash summaries, and the like at any time.

[0038] The prompt generator 252 generates a prompt to be used when generating a hash summary corresponding to an additional portion of the SOAP data, based on the SOAP data and a hash summary that is a summary of the SOAP data. For example, as illustrated in FIG. 5, the prompt generator 252 can generate one or more prompts based on a combination of the SOAP data and the hash summary.

[0039] For example, the prompt generator 252 generates one or more prompts by inputting the SOAP data into a trained model such as a localized LLM. The prompt generator 252 also generates one or more prompts by applying the above-mentioned The SOAP data and the generated prompt are input. As a result, the prompt generation unit 252 acquires a generated hash summary that is generated in response to the input of the SOAP data and the generated prompt. Thereafter, the prompt generation unit 252 calculates a predetermined score by calculating the similarity between the generated hash summary and a previously acquired hash summary. That is, the prompt generation unit 252 can calculate a score indicating how close the generated hash summary generated using the generated prompt is to the correct answer. The prompt generation unit 252 may select a prompt based on the calculated score, such as selecting a prompt with a score equal to or greater than a predetermined threshold value from among the generated prompts. The prompt generation unit 252 may generate the generated hash summary by transmitting the generated prompt to the summary generation device 300 described later, or by acquiring a model used in generating the hash summary from the summary generation device 300 in advance. In addition, in the present disclosure, the method used by the prompt generation unit 252 in calculating the score is not particularly limited.

[0040] For example, as described above, the prompt generation unit 252 may generate a prompt using a method such as that described in Non-Patent Document 1. Non-Patent Document 1 describes a technology called APE (Automatic Prompt Engineer) that creates a list of prompts to be used when outputting a result according to an input, based on an input sentence and a result. In the case of the present disclosure, the input sentence corresponds to SOAP data, and the result corresponds to a hash summary. As described in Non-Patent Document 1, by using APE, it is possible to generate multiple prompts and calculate a score for each generated prompt.

[0041] 6, by inputting a combination of SOAP data and hash summaries for each attribute, the prompt generation unit 252 can generate a prompt for each attribute, such as for each doctor or each medical department. The contents of the SOAP data and hash summaries may differ significantly depending on the attributes, such as for each doctor or each medical department. By generating a prompt for each attribute, it is possible to more accurately infer a hash summary according to the attribute.

[0042] The display unit 253 displays the prompt generated by the prompt generation unit 252 on the screen display unit 220. The display unit 253 may display information indicating the calculated score, attributes, and the like on the screen display unit 220 together with the prompt.

[0043] The correction receiving unit 254 receives corrections to the prompt generated by the prompt generating unit 252 in response to an operation by the doctor or the like on the operation input unit 210 or the like. As described above, in the case of the present disclosure, the prompt is an instruction using language. Therefore, the doctor or the like who has confirmed the prompt displayed on the screen display unit 220 can easily confirm whether the prompt is in line with his or her intention. In addition, the doctor or the like who has confirmed the prompt displayed on the screen display unit 220 can easily correct the prompt to better suit his or her intention.

[0044] The approval unit 255 receives approval from the doctor or the like for the prompt generated by the prompt generation unit 252 in response to an operation by the doctor or the like on the operation input unit 210 or the like. For example, a doctor or the like who has confirmed the prompt displayed on the screen display unit 220 may approve the prompt using the approval unit 255 after correcting the prompt via the correction acceptance unit 254. Furthermore, upon receiving the approval, the approval unit 255 stores the approved prompt in the storage unit 240 as prompt information 242. Note that when a prompt is generated for each attribute, the approval unit 255 may store in the storage unit 240 identification data corresponding to the attribute in association with the generated prompt.

[0045] The verification unit 256 generates a hash summary according to the generated prompt by inputting the prompt generated by the prompt generation unit 252 to a model used when generating a hash summary. The verification unit 256 may obtain in advance the model used when generating a hash summary from the summary generation device 300, and generate the hash summary in a sandbox.

[0046] Furthermore, the verification unit 256 can display the results of the verification on the screen display unit 220. In this case, the doctor or the like may correct the prompt after checking both the prompt displayed on the screen display unit 220 and the results of the verification. In other words, the correction receiving unit 254 can accept a correction to the prompt made by the doctor or the like after checking the results of the verification by the verification unit 256. Similarly, the approval by the approval unit 255 may also be accepted after the verification by the verification unit 256.

[0047] The output unit 257 transmits the prompt included in the prompt information 242 to an external device. For example, the output unit 257 can transmit the prompt included in the prompt information 242 to an external device such as the summary generation device 300 in response to an instruction from the summary generation device 300. The output unit 257 may transmit the prompt together with the identification data. Note that the output unit 257 can transmit all the prompts included in the prompt information 242 to an external device such as the summary generation device 300 in response to an instruction from the summary generation device 300. The output unit 257 may transmit only some of the prompts included in the prompt information 242 to an external device, such as transmitting only the prompt associated with an attribute according to the instruction in response to an instruction from the summary generation device 300.

[0048] The above is an example of the configuration of the prompt generation device 200. Note that the prompt generation device 200 may be configured with a part of the configuration exemplified in Fig. 2, for example, without at least one of the correction receiving unit 254, the approval unit 255, and the verification unit 256.

[0049] The summary generation device 300 is an information processing device that generates a hash summary corresponding to an updated portion of the SOAP data, using the SOAP data to which data has been added in response to a medical examination or the like, and the prompt generated by the prompt generation device 200. Fig. 7 shows an example of the configuration of the summary generation device 300. Referring to Fig. 7, the summary generation device 300 has, as its main components, for example, an operation input section 310, a screen display section 320, a communication I / F section 330, a storage section 340, and an arithmetic processing section 350.

[0050] The summary generation device 300 may have the same modified examples as the above-described prompt generation device 200. The configurations of the operation input unit 310, screen display unit 320, and communication I / F unit 330 may be the same as those of the above-described operation input unit 210, screen display unit 220, and communication I / F unit 230. Therefore, descriptions of the operation input unit 310, screen display unit 320, and communication I / F unit 330 will be omitted.

[0051] The storage unit 340 is a storage device such as a hard disk or a memory. The storage unit 340 stores processing information and programs 343 necessary for various processes in the arithmetic processing unit 350. The programs 343 are read into the arithmetic processing unit 350 and executed to realize various processing units. The programs 343 are read in advance from an external device or a recording medium via a data input / output function such as the communication I / F unit 330, and are stored in the storage unit 340. Main information stored in the storage unit 340 includes, for example, prompt information 341, generation target information 342, and the like.

[0052] The prompt information 341 includes a prompt generated by the prompt generation device 200. Identification data and a prompt may be associated with each other in the prompt information 341. The prompt information 341 is updated when the prompt acquisition unit 351 acquires a prompt from the prompt generation device 200, for example.

[0053] The generation target information 342 includes data for which a hash summary is to be generated. For example, the generation target information 342 may include, as data for which a hash summary is to be generated, SOAP data to which data has been added in response to a new medical treatment, and a hash summary generated based on the SOAP data prior to the addition of the data. The generation target information 342 may include only SOAP data to which data has been added in response to a new medical treatment, as data for which a hash summary is to be generated. The generation target information 342 is acquired and updated by the generation target data acquisition unit 352 accepting input of data from an external device such as an electronic medical record terminal via the communication I / F unit 330.

[0054] The arithmetic processing unit 350 has an arithmetic device such as a CPU and its peripheral circuits. The arithmetic processing unit 350 reads and executes a program 343 from the storage unit 340, thereby implementing various processing units by making the above-mentioned hardware and the program 343 work together. Examples of main processing units implemented by the arithmetic processing unit 350 include a prompt acquisition unit 351, a generation target data acquisition unit 352, a summary generation unit 353, and an output unit 354. Note that the arithmetic processing unit 350 may have a GPU or the like instead of the above-mentioned CPU.

[0055] The prompt acquiring unit 351 acquires from the prompt generation device 200 a prompt generated by the prompt generation device 200. The prompt acquiring unit 351 may acquire identification data and a prompt. The prompt acquiring unit 351 also stores the acquired prompt in the storage unit 340 as prompt information 341. The prompt acquiring unit 351 may acquire a prompt from the prompt generation device 200 at any time.

[0056] The generation target data acquisition unit 352 acquires data for which a hash summary is to be generated from an external device such as an electronic medical record terminal. For example, the generation target data acquisition unit 352 acquires, as data for which a hash summary is to be generated, SOAP data to which data has been added in response to a new medical treatment, and a hash summary generated based on the SOAP data before the data was added, as data for which a hash summary is to be generated. The generation target data acquisition unit 352 may acquire, as data for which a hash summary is to be generated, SOAP data to which data has been added in response to a new medical treatment. The generation target data acquisition unit 352 may also acquire information indicating attributes together with the data for which a hash summary is to be generated. Thereafter, the generation target data acquisition unit 352 stores the acquired data for which a hash summary is to be generated in the storage unit 340 as generation target information 342.

[0057] 8, the summary generation unit 353 generates a hash summary corresponding to the updated portion of the SOAP data, based on the data for which a hash summary is to be generated, acquired by the target data acquisition unit 352, and the prompt acquired by the prompt acquisition unit 351. The summary generation unit 353 may generate a hash summary including a portion corresponding to the updated portion of the SOAP data, in response to the above input.

[0058] As an example, the summary generation unit 353 has a model generated by training a localized large language model (LLM) or any other language model using a combination of SOAP data and hash summaries as teacher data. In other words, the summary generation unit 353 has a model trained to output a hash summary in response to an input of SOAP data or a combination of SOAP data with added data and a hash summary before the data is added. For example, the summary generation unit 353 inputs the SOAP data with added data and a hash summary generated based on the SOAP data before the data is added, which are data for which a hash summary is to be generated, to the model, and also inputs a prompt to the model. In response to this, the summary generation unit 353 can generate a hash summary corresponding to the updated part of the SOAP data as an output from the model in response to the above input.

[0059] The summary generation unit 353 may identify a prompt based on the information indicating the attributes acquired by the generation target data acquisition unit 352, and input the identified prompt to the above-mentioned model. The prompt to be input may be selected according to an operation on the operation input unit 310, in addition to the above-mentioned example. In other words, the configuration may be such that a doctor or the like selects a prompt to be input from among the prompts included in the prompt information 341, and the selected prompt is input to the model.

[0060] The summary generation unit 353 may have a model for each attribute by performing training using teacher data for each attribute, or may have a model that does not depend on attributes by performing training regardless of attributes. That is, the summary generation unit 353 may have only one model, or may have multiple models. In addition, when the summary generation unit 353 has multiple models, the summary generation unit 353 may select a model to be used according to information indicating the attributes acquired by the generation target data acquisition unit 352.

[0061] The output unit 354 outputs the hash summary etc. generated by the summary generation unit 353. For example, the output unit 354 can display the hash summary etc. on the screen display unit 320, or transmit it to an external device such as an electronic medical record terminal via the communication I / F unit 330.

[0062] The above is an example of the configuration of the summary generation device 300. Next, an example of the operation of the prompt generation device 200 and the summary generation device 300 will be described with reference to Figures 9 and 10. First, an example of the operation of the prompt generation device 200 will be described with reference to Figure 9.

[0063] Fig. 9 is a flowchart showing an example of the operation of the prompt generation device 200. Referring to Fig. 9, the reception unit 251 receives SOAP data, a hash summary, and the like from an external device such as an electronic medical record terminal (step S101). At this time, the reception unit 251 may receive, together with the SOAP data, the hash summary, and the like, attribute information indicating the attributes of the person who input the SOAP data or the hash summary from the external device such as an electronic medical record terminal.

[0064] The prompt generating unit 252 generates a prompt to be used when generating a hash summary corresponding to an added portion of the SOAP data, based on the SOAP data and a hash summary that is a summary of the SOAP data (step S102). The prompt generating unit 252 may generate a prompt for each attribute, such as for each doctor. For example, the prompt generating unit 252 can generate a prompt using a technique such as that described in Non-Patent Document 1.

[0065] The display unit 253 displays the prompt generated by the prompt generation unit 252 on the screen display unit 220 or the like (step S103). The display unit 253 may display information indicating the calculated score, attributes, and the like on the screen display unit 220 together with the prompt.

[0066] The correction receiving unit 254 receives a correction to the prompt generated by the prompt generating unit 252 in response to an operation by the doctor or the like on the operation input unit 210 or the like (step S104). As described above, in the case of the present disclosure, the prompt is an instruction using language. Therefore, a doctor or the like who checks the prompt displayed on the screen display unit 220 can easily check whether the prompt is in line with his or her intention.

[0067] The approval unit 255 receives approval from the doctor or the like for the prompt generated by the prompt generation unit 252 in response to an operation by the doctor or the like on the operation input unit 210 or the like (step S105). Furthermore, upon receiving approval, the approval unit 255 stores the approved prompt as prompt information 242 in the storage unit 240. Note that when a prompt is generated for each attribute, the approval unit 255 may store identification data according to the attribute and the generated prompt in the storage unit 240 in association with each other.

[0068] The above is an example of the operation of the prompt generation device 200. Note that at least a part of the processes from steps S103 to S105 may be omitted. Furthermore, the prompt generation device 200 may be configured to verify the prompt.

[0069] Next, an example of the operation of the summary generation device 300 will be described with reference to Fig. 10. Fig. 10 is a flowchart showing an example of the operation of the summary generation device 300. Referring to Fig. 10, the generation target data acquisition unit 352 acquires data for which a hash summary is to be generated from an external device such as an electronic medical record terminal (step S201). For example, the generation target data acquisition unit 352 acquires, as data for which a hash summary is to be generated, SOAP data to which data has been added in response to a new medical treatment, etc., and a hash summary generated based on the SOAP data prior to the addition of the data.

[0070] The summary generation unit 353 generates a hash summary corresponding to the updated portion of the SOAP data, based on the data for which a hash summary is to be generated, acquired by the target data acquisition unit 352, and the prompt acquired by the prompt acquisition unit 351 (step S202). The summary generation unit 353 may perform the above generation using a model that is trained in advance using a combination of the SOAP data and the hash summary as training data. The summary generation unit 353 may also be configured to select a prompt to be input based on information indicating attributes, etc.

[0071] The output unit 354 outputs the hash summary generated by the summary generation unit 353 (step S203). For example, the output unit 354 can display the hash summary on the screen display unit 320 or transmit it to an external device such as an electronic medical record terminal via the communication I / F unit 330.

[0072] The above is an example of the operation of the summary generation device 300.

[0073] As described above, the prompt generation device 200 has a reception unit 251 and a prompt generation unit 252. With this configuration, it is possible to generate a prompt to be used when generating a hash summary corresponding to an added portion of the SOAP data, based on the SOAP data and hash summary received by the reception unit 251. As a result, it becomes possible to generate a hash summary using the generated prompt. This makes it possible to reduce the effort required when generating a hash summary.

[0074] Furthermore, the prompt generation unit 252 can be configured to generate a prompt for each attribute. The contents of the SOAP data and hash summary may vary significantly depending on the attributes, such as for each doctor or each medical department. By generating a prompt for each attribute, it is possible to more accurately infer a hash summary according to the attribute.

[0075] Furthermore, the prompt generation unit 252 generates a prompt, which is an instruction using language. Therefore, a doctor or the like who checks the prompt generated by the prompt generation unit 252 can easily check whether the prompt is in line with his / her intention. Furthermore, the doctor or the like can easily modify the prompt to better suit his / her intention according to the result of the check.

[0076] Moreover, summary generation device 300 has a prompt acquisition section 351, a generation target data acquisition section 352, and a summary generation section 353. With this configuration, summary generation section 353 can generate a hash summary corresponding to an updated portion of the SOAP data, using the prompt acquired by prompt acquisition section 351 and the generation target data acquired by generation target data acquisition section 352. This makes it possible to reduce the effort required for generating a hash summary.

[0077] The configuration of the medical support system 100 is not limited to the example shown in Fig. 1. For example, in the present disclosure, the prompts generated by the prompt generation device 200 are instructions for a model, and do not have much personal information aspect. For this reason, as shown in Fig. 11, the medical support system 100 may have a prompt management device 400 that collectively manages the prompts generated by multiple prompt generation devices 200.

[0078] 11, the prompt management device 400 is an information processing device that acquires prompts from multiple communicatively connected prompt generation devices 200. For example, if the medical assistance system 100 has the prompt management device 400, the output unit 257 of the prompt generation device 200 may output prompt information 242 to the prompt management device 400. The prompt management device 400 can also transmit prompts that it manages to multiple communicatively connected summary generation devices 300 and other external devices such as electronic medical record terminals.

[0079] The medical support system 100 including the prompt management device 400 can centrally manage and utilize the prompts generated by each prompt generation device 200, as shown in Fig. 12. As a result, the medical support system 100 can more efficiently utilize the generated prompts.

[0080] Furthermore, when using the prompt generation device 200 or summary generation device 300 described in this disclosure, it is preferable that the hash summary corresponding to the SOAP data before data is added is standardized to some extent. For example, such standardization may be achieved by establishing some rules for the description method used when generating a hash summary corresponding to the SOAP data before data is added.

[0081] [Second embodiment] Next, the generating device 500 will be described with reference to Fig. 13 to Fig. 16. Fig. 13 is a diagram showing a hardware configuration example of the generating device 500. Fig. 14 is a block diagram showing a configuration example of the generating device 500. Fig. 15 is a flowchart showing an operation example of the generating device 500. Fig. 16 is a block diagram showing a configuration example of a generating device 600 which is a second generating device in the present disclosure.

[0082] In the second embodiment of the present disclosure, a generating device 500 that generates a prompt based on medical data such as SOAP data and a summary such as a hash summary will be described. Also, a generating device 600 that generates a summary using the prompt generated by the generating device 500 will be described. Fig. 13 shows an example of the hardware configuration of the generating device 500. Referring to Fig. 13, the generating device 500 has the following hardware configuration, as an example. ·CPU (Central Processing Unit) 501 (computing unit) ROM (Read Only Memory) 502 (storage device) RAM (Random Access Memory) 503 (storage device) Programs 504 loaded into RAM 503 A storage device 505 for storing a group of programs 504 A drive device 506 that reads and writes from a recording medium 510 outside the information processing device A communication interface 507 for connecting to a communication network 511 outside the information processing device Input / output interface 508 for inputting and outputting data Bus 509 connecting each component

[0083] 14. The program group 504 is stored in advance in, for example, the storage device 505 or the ROM 502, and is loaded into, for example, the RAM 503 by the CPU 501 for execution as required. The program group 504 may be supplied to the CPU 501 via the communication network 511, or may be stored in advance in the recording medium 510, and the drive device 506 may read out the programs and supply them to the CPU 501.

[0084] 13 shows an example of a hardware configuration of the generating device 500. The hardware configuration of the generating device 500 is not limited to the above-mentioned case. For example, the generating device 500 may be configured with a part of the above-mentioned configuration, such as not having the drive device 506. Furthermore, the CPU 501 may be a GPU as exemplified in the first embodiment.

[0085] The acquiring unit 521 acquires medical data in the electronic medical record and a summary of the medical data. For example, the acquiring unit 521 may acquire the SOAP data described in the first embodiment as the medical data. The acquiring unit 521 may also acquire the hash summary described in the first embodiment as the summary. As an example, the acquiring unit 521 can acquire the medical data and the summary from an external device such as an electronic medical record terminal.

[0086] The prompt generating unit 522 generates a prompt to be used when generating a summary corresponding to an updated portion of the medical data to which data has been added. In the present disclosure, the prompt generating unit 522 can generate a prompt based on the medical data and the summary acquired by the acquiring unit 521. For example, the prompt generating unit 522 may generate an instruction using a language such as Japanese as the prompt.

[0087] The above is an example configuration of the generating device 500. Next, an example operation of the generating device 500 will be described with reference to FIG.

[0088] Fig. 15 is a flowchart showing an example of the operation of the generating device 500. Referring to Fig. 15, the acquiring unit 521 acquires medical data in an electronic medical record and a summary of the medical data (step S301). For example, the acquiring unit 521 can acquire the medical data and the summary from an external device such as an electronic medical record terminal.

[0089] The prompt generation unit 522 generates a prompt, which is an instruction for the trained model to be used when generating a summary corresponding to the updated portion of the medical data to which data has been added, based on the medical data and the summary acquired by the acquisition unit 521 (step S302). For example, the prompt generation unit 522 may generate the prompt using the technology described in Non-Patent Document 1.

[0090] The above is an example of the operation of the generating device 500.

[0091] As described above, the generating device 500 includes an acquiring unit 521 and a prompt generating unit 522. With this configuration, the prompt generating unit 522 can generate a prompt based on the medical data and summary acquired by the acquiring unit 521. As a result, it becomes possible to generate a summary using the generated prompt. This makes it possible to reduce the effort required for generating a summary.

[0092] The generating device 500 described above can be realized by incorporating a predetermined program into an information processing device such as the generating device 500. Specifically, a program according to another embodiment of the present disclosure is a program for acquiring medical data in an electronic medical record and a summary of the medical data in an information processing device such as the generating device 500, and generating a prompt, which is an instruction to a trained model used when generating a summary corresponding to an updated portion of the medical data to which data has been added, based on the acquired medical data and summary.

[0093] In addition, the generation method executed by an information processing device such as the above-mentioned generating device 500 is a method in which the information processing device such as the generating device 500 acquires medical data and a summary of the medical data in the electronic medical record, and generates a prompt, which is an instruction to the trained model, used when generating a summary corresponding to an updated portion of the medical data to which data has been added, based on the acquired medical data and summary.

[0094] A program having the above-described configuration, or a computer-readable recording medium having a program recorded thereon, or a generation method, etc., can achieve the same action and effect as the above-described generation device 500, and therefore can achieve the above-described objective of the present disclosure.

[0095] Moreover, the generating device 600 in the present disclosure is an information processing device that generates a summary using the prompt generated by the generating device 500. The hardware configuration of the generating device 600 may be similar to that of the generating device 500 illustrated in Fig. 13. Moreover, the generating device 600 can realize the functions of a generation target data acquiring unit 621 and a summary generating unit 622 illustrated in Fig. 16 by a CPU acquiring a group of programs and executing the programs.

[0096] The target data acquisition unit 621 acquires the medical data to which data has been added as target data for generating a summary. For example, the target data acquisition unit 621 may acquire the target data from an external device such as an electronic medical record terminal.

[0097] The summary generating unit 622 has a model that has been trained using a combination of medical data in the electronic medical record and a summary of the medical data as teacher data. The summary generating unit 622 inputs the medical data acquired by the generation target data acquiring unit 621 and a prompt that has been generated in advance based on the medical data in the electronic medical record and the summary of the medical data to the model, thereby generating a summary of the medical data acquired by the generation target data acquiring unit 621. The prompt may be generated by the generating device 500, for example.

[0098] Thus, the generating device 600 has the generation target data acquisition unit 621 and the summary generating unit 622. With this configuration, the summary generating unit 622 can generate a summary based on the medical data and the prompt acquired by the generation target data acquisition unit 621. This can reduce the effort required for generating a summary.

[0099] The generating device 600 described above can be realized by incorporating a predetermined program into an information processing device such as the generating device 600. Specifically, a program according to another embodiment of the present disclosure is a program for realizing a process of acquiring, in an information processing device such as the generating device 600, medical data to which data has been added as target data for generating a summary, and generating a summary of the acquired medical data by inputting the acquired medical data and a prompt generated in advance based on the medical data and the summary of the medical data in the electronic medical record to a model that has been trained using a combination of the medical data in the electronic medical record and the summary of the medical data as teacher data.

[0100] Furthermore, a generation method executed by an information processing device such as the generating device 600 described above is a method in which the information processing device such as the generating device 600 acquires medical data with added data as target data for generating a summary, and generates a summary of the acquired medical data by inputting the acquired medical data and a prompt that has been generated in advance based on the medical data and the summary of the medical data in the electronic medical record to a model that has been trained using a combination of the medical data in the electronic medical record and a summary of the medical data as training data.

[0101] A program having the above-described configuration, or a computer-readable recording medium having a program recorded thereon, or a generation method, or a generation device 600, etc., can achieve the same action and effect as the above-described generation device 500, and therefore can achieve the above-described objective of the present disclosure.

[0102] <Additional Notes> A part or all of the above-described embodiments may be described as follows: The generating device and the like in the present disclosure are outlined below. However, the present disclosure is not limited to the following configuration.

[0103] (Appendix 1) An acquisition unit that acquires medical data and a summary of the medical data in the electronic medical record; a prompt generation unit that generates a prompt, which is an instruction for a trained model used when generating a summary corresponding to an updated portion of the medical data to which data has been added, based on the medical data and the summary acquired by the acquisition unit; have generator. (Appendix 2) the prompt generation unit is configured to generate the prompt by inputting the medical data acquired by the acquisition unit into a pre-trained model; The prompt generation unit generates the prompt for each attribute by inputting medical data to a model for each predetermined attribute. 2. The generating device of claim 1. (Appendix 3) a display unit that displays the prompt generated by the prompt generating unit; a correction receiving unit that receives a correction to the prompt displayed by the display unit; have 3. The generating device of claim 1 or 2. (Appendix 4) a verification unit that generates a summary corresponding to the prompt by inputting the prompt generated by the prompt generation unit and the medical data used to generate the prompt into a trained model used to generate a summary corresponding to an updated portion of the medical data to which data has been added; The modification receiving unit receives a modification to the prompt verified by the verifying unit. 4. The generating device of claim 3. (Appendix 5) a target data acquisition unit that acquires medical data to which data has been added as target data for generating a summary; a summary generation unit that generates a summary corresponding to an updated portion of the medical data acquired by the generation target data acquisition unit, based on the prompt generated by the prompt generation unit and the medical data acquired by the generation target data acquisition unit; have 5. The generating device according to any one of claims 1 to 4. (Appendix 6) The summary generation unit selects the prompt to be input from the prompts generated by the prompt generation unit based on the information indicating the attribute, and generates a summary corresponding to an updated portion of the medical data acquired by the generation target data acquisition unit based on the selected prompt and the medical data acquired by the generation target data acquisition unit. 6. A generating device as described in Appendix 5. (Appendix 7) An output unit that outputs the prompt generated by the prompt generating unit to a management device that manages the prompt. 7. The generating device according to any one of claims 1 to 6. (Appendix 7-1) The acquiring unit acquires, as medical data in an electronic medical record, SOAP data, which is medical data described in a SOAP (Subjective Objective Assessment Plan) format, and acquires a summary of the SOAP data as a summary of the medical data. A generating device according to any one of claims 1 to 7. (Appendix 7-2) The prompt generation unit generates the prompt indicating an instruction as a sentence using language. A generating device according to any one of claims 1 to 7-1. (Appendix 8) An information processing device, Obtaining medical data in an electronic medical record and a summary of the medical data; A prompt is generated based on the acquired medical data and the summary, which is an instruction to the trained model to be used when generating a summary corresponding to the updated portion of the medical data to which data has been added. Generation method. (Appendix 8-1) When generating the prompt, the acquired medical data is input into a pre-trained model to generate the prompt; When generating the prompt, the medical data is input to the model for each predetermined attribute, thereby generating the prompt for each attribute. 10. The method of generating the same as in claim 8. (Appendix 8-2) Displaying the generated prompt, Accepts corrections to the above prompt that was displayed A method for generating the same according to claim 8 or claim 8-1. (Appendix 8-3) a verification process for generating a summary corresponding to the prompt by inputting the generated prompt and the medical data used to generate the prompt into a trained model used to generate a summary corresponding to the updated portion of the medical data to which data has been added; Accept corrections to the verified prompt The generation method described in Appendix 8-2. (Appendix 8-4) Acquire the medical data to which the data has been added as target data for generating a summary; A summary is generated based on the generated prompt and the medical data acquired as the data to be generated, the summary corresponding to an updated portion of the medical data acquired as the data to be generated. A method for generating a signal according to any one of claims 8 to 8-3. (Appendix 8-5) When generating a summary, a prompt to be input is selected from the prompts generated based on the information indicating the attributes, and a summary corresponding to an updated portion of the medical data acquired as the data to be generated is generated based on the selected prompt and the medical data acquired as the data to be generated. The method of generation described in Appendix 8-4. (Appendix 8-6) The generated prompt is output to a management device that manages the prompt. A method for generating a signal according to any one of claims 8 to 8-5. (Appendix 8-7) As medical data in an electronic medical record, SOAP data, which is medical data written in SOAP (Subjective Objective Assessment Plan) format, is obtained, and a summary of the SOAP data is obtained as a summary of the medical data. A method for generating a signal according to any one of claims 8 to 8-6. (Appendix 8-8) When generating the prompt, generate the prompt indicating the instruction as a sentence using language. A method for generating a signal according to any one of claims 8 to 8-7. (Appendix 9) In the information processing device, Obtaining medical data in an electronic medical record and a summary of the medical data; A prompt is generated based on the acquired medical data and the summary, which is an instruction to the trained model to be used when generating a summary corresponding to the updated portion of the medical data to which data has been added. A program for implementing processing. (Appendix 9-1) When generating the prompt, the acquired medical data is input into a pre-trained model to generate the prompt; When generating the prompt, the medical data is input to the model for each predetermined attribute, thereby generating the prompt for each attribute. 9. The program described in Appendix 9. (Appendix 9-2) Displaying the generated prompt, Accepts corrections to the above prompt that was displayed 2. A program as described in Appendix 9 or Appendix 9-1. (Appendix 9-3) a verification process for generating a summary corresponding to the updated portion of the medical data to which data has been added, by inputting the generated prompt and the medical data used to generate the prompt into a trained model used to generate a summary corresponding to the updated portion of the medical data to which data has been added, and Accept corrections to the verified prompt The program described in Appendix 9-2. (Appendix 9-4) Acquire the medical data to which the data has been added as target data for generating a summary; A summary is generated based on the generated prompt and the medical data acquired as the data to be generated, the summary corresponding to an updated portion of the medical data acquired as the data to be generated. 2. A program according to any one of claims 9 to 9-3. (Appendix 9-5) When generating a summary, a prompt to be input is selected from the prompts generated based on the information indicating the attributes, and a summary corresponding to an updated portion of the medical data acquired as the data to be generated is generated based on the selected prompt and the medical data acquired as the data to be generated. The program described in Appendix 9-4. (Appendix 9-6) The generated prompt is output to a management device that manages the prompt. 2. A program according to any one of claims 9 to 9-5. (Appendix 9-7) As medical data in an electronic medical record, SOAP data, which is medical data written in SOAP (Subjective Objective Assessment Plan) format, is obtained, and a summary of the SOAP data is obtained as a summary of the medical data. 2. The program according to claim 1, wherein the first and second subparagraphs are each a first subparagraph and a second subparagraph and a third subparagraph are each a second subparagraph. (Appendix 9-8) When generating the prompt, generate the prompt indicating the instruction as a sentence using language. 2. A program according to any one of claims 9 to 9-7. (Appendix 10) a target data acquisition unit that acquires medical data to which data has been added as target data for generating a summary; a summary generation unit that generates a summary of the medical data acquired by the data acquisition unit for generating ... have generator.

[0104] The programs described in the above embodiments and appendices may be stored in a storage device or a computer-readable recording medium, such as a portable medium such as a flexible disk, an optical disk, a magneto-optical disk, or a semiconductor memory.

[0105] Although the present disclosure has been described above with reference to the above-mentioned embodiments, the present disclosure is not limited to the above-mentioned embodiments. Various modifications that can be understood by a person skilled in the art can be made to the configuration and details of the present disclosure within the scope of the present disclosure. Furthermore, each embodiment can be appropriately combined with other embodiments. [Explanation of symbols]

[0106] 100 Medical Support System 200 Prompt Generator 210 Operation input section 220 Screen display section 230 Communication I / F section 240 Storage section 241 Electronic medical record information 242 Prompt Information 243 Programs 250 Processing Unit 251 Reception 252 Prompt Generation Unit 253 Display section 254 Correction Reception Department 255 Approval Department 256 Verification Department 257 Output section 300 Summary Generator 310 Operation input section 320 Screen display section 330 Communication I / F section 340 Storage section 341 Prompt Information 342 Information to be generated 343 Programs 350 Processing section 351 Prompt Acquisition Department 352 Generation target data acquisition unit 353 Summary Generation Unit 354 Output section 400 Prompt Management Device 500 generator 501 CPU 502 ROM 503 RAM 504 Programs 505 Storage device 506 Drive Unit 507 Communication Interface 508 Input / Output Interface 509 Bus 510 Recording media 511 Communication Network 521 Acquisition Department 522 Prompt Generation Unit 600 generator 621 Generation target data acquisition unit 622 Summary generator

Claims

1. An acquisition unit that acquires medical data and a summary of the medical data in the electronic medical record; a prompt generation unit that generates a prompt, which is an instruction for a trained model used when generating a summary corresponding to an updated portion of the medical data to which data has been added, based on the medical data and the summary acquired by the acquisition unit; have generator.

2. the prompt generation unit is configured to generate the prompt by inputting the medical data acquired by the acquisition unit into a pre-trained model; The prompt generation unit generates the prompt for each attribute by inputting medical data to a model for each predetermined attribute. The generating device of claim 1 .

3. a display unit that displays the prompt generated by the prompt generating unit; a correction receiving unit that receives a correction to the prompt displayed by the display unit; have The generating device of claim 1 .

4. a verification unit that generates a summary corresponding to the prompt by inputting the prompt generated by the prompt generation unit and the medical data used to generate the prompt into a trained model used to generate a summary corresponding to an updated portion of the medical data to which data has been added; The modification receiving unit receives a modification to the prompt verified by the verifying unit. The generating device according to claim 3 .

5. a target data acquisition unit that acquires medical data to which data has been added as target data for generating a summary; a summary generation unit that generates a summary corresponding to an updated portion of the medical data acquired by the generation target data acquisition unit, based on the prompt generated by the prompt generation unit and the medical data acquired by the generation target data acquisition unit; have The generating device of claim 1 .

6. The summary generation unit selects the prompt to be input from the prompts generated by the prompt generation unit based on the information indicating the attribute, and generates a summary corresponding to an updated portion of the medical data acquired by the generation target data acquisition unit based on the selected prompt and the medical data acquired by the generation target data acquisition unit. A generating device according to claim 5.

7. An output unit that outputs the prompt generated by the prompt generating unit to a management device that manages the prompt. The generating device of claim 1 .

8. An information processing device, Obtaining medical data in an electronic medical record and a summary of the medical data; A prompt is generated based on the acquired medical data and the summary, which is an instruction to the trained model to be used when generating a summary corresponding to the updated portion of the medical data to which data has been added. Generation method.

9. In the information processing device, Obtaining medical data in an electronic medical record and a summary of the medical data; A prompt is generated based on the acquired medical data and the summary, which is an instruction to the trained model to be used when generating a summary corresponding to the updated portion of the medical data to which data has been added. A program for implementing processing.

10. a target data acquisition unit that acquires medical data to which data has been added as target data for generating a summary; a summary generation unit that generates a summary of the medical data acquired by the data acquisition unit for generating ... have generator.

Images